276'
2, 7 D
0554 m A!

51." “ES
L?

Flow from Small Watersheds Adjacent
to the Study Reachbowf, the Gila» River
Phreatophyte Project, Arizona

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—1

ft:

 

W, t. 7:
\13' ‘i ‘ _ ‘~
*L/ .M
1 p33,: 3 1
E U

 

 

-Lflm‘gwwnfl‘

 

Flow from Small Watersheds Adjacent
to the Study Reach of the Gila River
Phreatophyte Project, Arizona

By D. E. BURKHAM

 

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976

 

 

UNITED STATES DEPARTMENT OF THE INTERIOR

THOMAS S. KLEPPE, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Director

Library of Congress Cataloging in Publication Data

Burkham, D. E. 1927—
Flow from small watersheds adjacent to the study reach of the Gila River phreatophyte project, Arizona. (Gila River phreatophyte
project)

(Geological Survey Professional Paper 655-1)

Bibliography: p. 19

Supt, of Docs. no.: I 19.162655-1 .‘

l. Stream measurementngila River watershed. I. Title. II. Series. 111. Series: United States. Geological Survey Professional
Paper 655—1.

QE75.P9 no. 655—1 [GB1225.A6] 557.3'083 [551.4'83'0979154] 75—619127

 

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, DC. 20402
Stock Number 024—001—02788-7

 

CONTENTS

Page Page
Abstract __________________________________________________ 11 Peak discharge-storm volume relations ______________________ I7
Introduction ________________________________________________ 1 Runoff ____________________________________________________ 7
Characteristics of the study area ____________________________ 1 Computation of data ____________________________________ 7
Design of network of gaging stations ________________________ 3 Storm and annual volumes ______________________________ 9
Gaging stations and observation procedures __________________ 4 Accuracy of data ________________________________________ 9
Stage-discharge relations ____________________________________ 5 Comparisons of runoff from the study basins with runoff from
Controls ______________________________________________ 5 nearby basins ____________________________________________ 15
Discharge measurements ________________________________ 5 Conclusions ________________________________________________ 19
References cited ____________________________________________ 19

ILLUSTRATIONS

Page
PLATE 1. Map showing watershed boundaries and instrument locations ____________________________________________________ In pocket
2. Photographs and graphs showing development of the alluvial fan at the mouth of a tributary to the Gila River __________ In pocket
FIGURE 1. Index map of project area __________________________________________________________________________________________ I2
2. Graph showing relation of length of longest watercourse to size of drainage basin ______________________________________ 3
3. Graph showing relation of cumulative percentage of total tributary area to number of watersheds used in computing
percentages ____________________________________________________________________________________________________ 4
4. Graph showing relation of peak discharge to storm volume for single-peak storms ______________________________________ 8
5. Hypothetical hydrograph for a single-peak flow event ________________________________________________________________ 9
6. Graph showing relation of number of days of flow to size of watershed __________________________________________________ 15
7. Map showing regions in Arizona where regression analyses have been made ____________________________________________ 16
8. Graph showing relation of mean annual runoff to size of basin ________________________________________________________ 17
TABLES
Page
TABLE 1. Storm runoff from tributaries, Gila River Phreatophyte Project, 1963—71 ______________________________________________ 110
2. Mean J une—October runoff for 1963—71, standard deviation of the mean, and sizes of watersheds for tributaries to the study
reach of the Gila River Phreatophyte Project ____________________________________________________________________ 14
3. Annual peak discharge, in cubic feet per second, for streams tributary to the study reach of the Gila River Phreatophyte
Project ________________________________________________________________________________________________________ 14

 

CONVERSION FACTORS

Factors for converting English units to the International System of units (SI) are given below to four
significant figures:

English Units Multiply By Metric Units
inches (in) 2.540 centimetres (cm)
25.40 millimetres (mm)
feet (ft) 0.3048 metres (In) ‘
304.8 millimetres (mml)
miles (mi) 1.609 kilometres (km)
square feet (ftz) .0929 square metres (m2)
acres 4047. square metres (In?)
.4047 hectares (ha)
square miles (miz) 2.590 square kilometres (km2)
feet per year (ft/year) .3048 metres per year (m/year)
acre-feet (acre-ft) 1233. cubic metres (m3)
cubic feet per second (ft3/s) .02832 cubic metres per second (m3/s)
acre-feet per day (acre-ft/day) .01430 cubic metres per day (m3/day)

111

G (13"‘5

 

GILA RIVER PHREATOPHYTE PROJECT

FLOW FROM SMALL WATERSHEDS ADJACENT TO
THE STUDY REACH OF THE GILA RIVER
PHREATOPHYTE PROJECT, ARIZONA

By D. E. BURKHAM

ABSTRACT

The Gila River in Safford Valley, southeastern Arizona, was the site
for a field study of evapotranspiration (Culler and others, 1970).
During the period of study, 1963—71, measurements of storm runoff in
summer—July through October—from tributaries along a 15-mile (24
km) reach of the Gila River were required for water budget analyses.
Most of the outflow from the 43 tributary basins, which range in size
from about 0.1 to 20 square miles (0.3 to 50 km2), resulted from
thunderstorms of small area] extent. The mean summer runoff for
1963—71 was about 1,370 acre—ft (1,690,000 m3), or 9 acre-ft per square
mile (6,900 m3 km2). The maximum summer runoff was about 3,180
acre-ft (3,920,000 m3) in 1967 and the minimum was about 130 acre-ft
(160,000,000 m3) in 1970. The largest storm outflow was about 970
acre-ft (1,200,000 m3). The largest peak discharge occurring in a
tributary stream was 8,000 cubic feet per second (230 m3/s) which
came from a 14-mi2 (36 km?) watershed. The largest peak discharge
per square mile was about 2,300 cubic feet per second (65 m3/s) which
resulted from a storm centering on a 0.79-mi2 (2.0 kmz) watershed.
The streamflow data are of poor quality.

The tributary streamflow to the study reach resulted from an
average of nine runoff storms per year. The maximum number of
runoff storms in the study reach was 31 in 1967 and the minimum was
12 in 1965. The tributary watersheds contributed to the project area
on an average of lessthan 13 days per year. For a tributary, the
average number of days of runoff per year was about 3.

INTRODUCTION

The primary purpose of this report is to present data
of storm runoff from tributaries along a 15-mile (24 km)
reach of the Gila River in southeastern Arizona for
1963—71 (fig. 1). Secondary objectives are to describe the
characteristics of flow in the tributary streams; to
describe the procedure used and problems encountered
in measuring the flow; and to compare the runoff values
obtained for the study tributaries with runoff values for
nearby basins.

The storm runoff data were required for the water
budget analyses of the Gila River Phreatophyte Project
(Culler and others, 1970). In the water budget analyses,
the amount of evapotranspiration is estimated as a
residual when all other significant quantities of inflow
and outflow have been measured. The study area of the

 

Gila River Phreatophyte Project includes three reaches
(Culler and others, 1970); however, runoff data were
obtained only for watersheds tributary to reaches 1 and
2 (pl. 1). The discussion in the section “Design of
Network of Gaging Stations” pertains to tributaries to
all the reaches; however, the discussions in the rest of
the report pertain only to tributaries to reaches 1 and 2.
This report is one of several chapters of Professional
Paper 655, which describes the environmental vari-
ables pertinent to the Gila River Phreatophyte Project.

CHARACTERISTICS OF THE STUDY AREA

The study area is near Globe, Arizona, and the
tributaries are typical of others in the Basin and Range
physiographic province (Fenneman, 1931) which drain
mountainous slopes paralleling comparatively flat wide
sediment-filled valleys (pl. 1). The composite area of the
many small basins contributing flow to reaches 1 and 2
is about 150 mi2 (390 km2). The long narrow tributary
basins range in size from about 0.1 to 20 mi2 (0.3 to 50
km2) and drain the south slopes of the Gila Mountains
on the north, and north slopes of Mt. Turnbull on the
south. For a watershed of given size, the main stream
channel of a tributary basin in the study area is about 60
percent longer than the average main channel in other
basins in the southwestern United States (fig. 2). The
slopes of the study tributaries range from about 2
percent near the Gila River to more than 40 percent on
Mt. Turnbull. Altitudes of the tributary basins range
from 2,480 feet (756 m) above sea level at the flood plain
of the Gila River to 8,200 feet (2,500 m) on Mt. Turnbull.

Near the Gila River, the ephemeral tributary streams
are entrenched in deposits of silt, sand, and gravel,
which were divided into basin fill, terrace alluvium, and
flood-plain alluvium by Davidson and Weist (in Culler
and others, 1970, p. A8). The terrace alluvium, through
which the tributary streams flow before reaching the
flood plain of the Gila River, ranges from less than half a

11

ca...

12 GILA RIVER PHREATOPHYTE PROJECT

114° 113°

112°

 

1—"u—‘——1—"—r"—r"—'37°

River

PHOENIX.

0 50 100

  
   
 
   

ARIZONA

100

150

150 MILES

Wfidd‘I—fij—I—l

200 KILOMETRES

FIGURE 1.—Index map of project area.

mile (about 0.8 km) to more than 2 miles (3.2 km) wide.
The flood-plain alluvium along the Gila River is half a
mile to 1 mile (0.8 to 1.6 km) wide. The flood-plain
alluvium along the study tributaries is of limited
extent, ranging from less than 50 to more than 300 feet
(15 to 90 m) wide.

Climatically, the study area is in the Sonoran Border
Zone (Thomas, 1962, p. 13) and is characterized by a
wide range in temperature and average annual precipi-
tation. The temperature extremes recorded at Safford,
which is 2,900 feet (880 m) above sea level, are 7° and
114°F (—14° and 46° C) (Sellers, 1960). Safford is about
35 miles (56 km) east of the study area. The long-term
average annual precipitation is about 13 inches (330
mm) in the study area. In summer the precipitation is
mainly from local convective thunderstorms, which
produce rainfall of high intensity and short duration
over small areas (Burkham, 1970). The long-term
. average of summer precipitation is about 7 inches (180
mm) and the temporal variation for summer precipita-
tion is about 40 percent. In winter the precipitation
mainly is from convergence, or frontal, storms that

 

distribute moisture over large areas. Because large
amounts of precipitation from tropical Pacific storms
are infrequent, the temporal variation of winter
precipitation is larger than that of summer precipita-
tion (Burkham, 1970). The long-term average of winter
precipitation is about 5 inches (130 mm) and the
temporal variation of winter precipitation is about 50
percent. In spring the precipitation is generally less
than 1 inch (25 mm).

The vegetation in the study area may be grouped
according to its location—on the uplands, terraces, or
flood plain. The most abundant plants on the uplands
are creosotebush (Larrea tridentata (Sesse & Mos. ex
DC.) Coville), white thorn (Acacia constricta Benth.),
catclaw (Acacia greggii A. Gray), cactus, and mesquite
(Prosopis juliflora var. velutine (Woot.) Sarg.) (Turner,
in Culler and others, 1970, p. A19—A20). Scattered
clusters of mesquite normally grow along the tributary
streams in the uplands. Mesquite communities occupy
most of the terrace deposits. Other woody perennials
growing on the terraces, according to Turner (in Culler
and others, 1970, p. A20), are catclaw, white thorn, gray

 

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ.

I3

DRAINAGE AREA, IN SQUARE KILOMETRES

 

 

   
 
 
  
  
    
    
 

 

 

0.1 1.0 10 100 1000 10,000
100i I IIIIIJlII I ll|ll | IIIIIIllI IIIIIIII [IIIIIHII IllIllll llll'llil‘l l IIIIIII I IIII‘IIII I IIIII | | IIIIIIIl I lllll
: 100 a
<0 — Curve developed by Bensen (1964) for 3‘ o:
liI : watersheds in southwestern United : El
— St t s __ E
2 - a e Lr=l.65A° _ o
3 _ =1
u; ‘ EXPLANATION ‘ ¥
‘0 Z
n: A _ LLI
8 10 E- Curve developed for the study 0 Data from Leopold : E
O ' tributaries and Miller (1956) __10 3
E : o o 000 . :2 8
l- — L=3.0A ° '5‘ Data from Moosburner —: cc
‘5‘ - (1970) _— E
A — <
5 - Data from Burkham —_ E
(“3 (1966) (I;
z 1 0 ° _— ”5
8 ' E— Data for the study tributaries : 2
u. I (see table 2) : 1.0 3
O — A L 1 u.
E ' Curve developed by Burkham (1966) Length of longest watercourse,in ‘— o
(D ' A A' for watershed in northeastern miles ‘: E
E _ New Mexico A . __ L9
-' L=1 72Ao.s4 Drainage area, in square miles E
. .1
0"] | l I | I I I | I I I | I I I | I I I I | | | I I | I I | I I I I I II I I | I I l I II I I | I L
0.01 0.10 1.0 10 100 1000 10,000

DRAINAGE AREA, IN SQUARE MILES

FIGURE 2.——Relation of length of longest watercourse to size of drainage basin.

thorn (Condaliopsis lycioides (A. Gray) Suesseng.), and
four-wing saltbush (Atriplex canescens (Pursh) Nutt.)
Saltcedar (Tamarix chinensis)1 is the dominant vegeta-
tion on the flood plain of the Gila River. Mesquite and
seepweed, however, grow along the outer edges of the
flood plain.

Most of the flow in the study tributaries is the result of
summer thunderstorms. High unit rates and volumes of
flow from the small watersheds characteristically are
produced by individual thunderstorms. The crest of a
flood from a thunderstorm is typically very sharp when
the flood reaches the flood plain of the Gila River.
Sometimes, when runoff enters a dry stretch of channel,
the flood crest disappears completely because the flow
sinks into the alluvium. During late September and
October, occasional frontal activity causes precipitation
that may produce runoff simultaneously from all the
tributary streams. The combined runoff from these
general rains and concurrent local thunderstorms often
results in relatively large water yields. Infrequently,
precipitation during winter may produce small
amounts of runoff from some of the tributaries.

Flow in the tributary streams has a high velocity, is
almost always in a supercritical state (Chow, 1959,
p. 13), carries large sediment loads, and moves large
amounts of coarse material along the channel bed. Upon
reaching the wide flat flood plain of the Gila River, most

 

1Also referred to as Tamarix pentandra and Tamarix gallica.

of the material is deposited, forming sediment mounds
or fans (pl. 2).

DESIGN OF NETWORK OF
GAGING STATIONS

The collection of tributary runoff data for three
reaches of the Gila River Phreatophyte Project (pl. 1)
began in 1962. Plugging of the stream channel of the
Gila River with debris from floods in 1962—64 made
reach 3 unsuitable for water budget studies (Culler and
others, 1970). The design of the procedure to be used to
measure flows from tributaries, however, had already
been completed before the plugging occurred. The
discussions in this section, therefore, are about
tributaries to three reaches even though complete sets
of data of tributary runoff were obtained only for
reaches 1 and 2. The area tributary to the entire study
reach of the phreatophyte project is about 260 mi2 (670
km2) and the basins range in size from less than 0.1 to 39
mi2 (0.3 to 100 km2).

The method of determining flow from tributaries was
designed on the following assumptions and criteria:

1. The water budget equation which would be used to
evaluate evapotranspiration e as a residual is

(1)

8
e = 2 xj,
j=1

 

 

I4 GILA RIVER PHREATOPHYTE PROJECT

where the components x1, . . . , x1 denote (1)
surface inflow in the Gila River, (2) surface inflow
from tributaries, (3) subsurface inflow in the
alluvium of the river and its tributaries, (4)
possible artesian inflow from underlying geologic
formations, (5) surface outflow in the river, (6)
subsurface outflow in the alluvium, (7) precipita-
tion on the flood plain, and (8) change in moisture
storage, respectively.

2. The smallest water budget period would be 2 weeks;
however, longer water budget periods may be
used.

3. Records of tributary flow at or near the flood plain of
the Gila River would be needed.

4. Significant amounts of runoff occur only in the
summer—July through October.

5. Tributaries contribute surface flow to the study
reach for only a small part of the time and this flow
is assumed to be a small part of the total water
involved in a water budget for most periods.

6. Accurate records of flow for each tributary would not
be required.

7. Difficulties in accurately measuring high rates of
discharge from tributaries would require that
water budgets, which include data for large storms
from tributaries, be omitted from study.

8. Periods of no flow in the tributaries would be
accurately defined.

Using these assumptions and criteria as guides, a
decision was reached that records of tributary flow
would be obtained by using gages equipped with
continuous-stage recorders in some of the tributary
streams and crest-stage gages in most of the remaining
tributaries. Runoff records for tributaries having
recording gages would be obtained using stage records
and stage-discharge relations. Runoff estimates at
crest-stage gages would be obtained by use of peak-
stage records, stage-discharge relations, and relations
between peak discharge and storm volume.

Only 16 continuous-stage recorders were available
for the tributary study. The streams in which these
recorders were placed were selected on the basis of basin
size, physiographic characteristics, and orientation
along the study reach of the phreatophyte project.
Because runoff increases with the size of the basin, size
was given first consideration in selecting streams that
would have a continuous-stage recorder. Recording
gages were established in the 10 largest basins, which
includes about 54 percent of the total tributary area (fig.
3). The remaining six sites were selected on the basis of
physiographic characteristics and orientation. About 59
percent of the total tributary area is included in basins
where the gages were installed (pl. 1).

Crest-stage gages were installed in 47 tributary

 

 

100 I ,
90
80

TRIBUTARY AREA
3 '5’ 8 ‘5 8 8 5‘

 

 

 

I l l I
10 20 30 40 50 60 70

NUMBER 0 F WATERSHEDS

CUMULATIVE PERCENTAGE OF TOTAL
0

0

FIGURE 3.—Relation of cumulative percentagevof total tributary area
to number of watersheds used in computing percentages. Cumula-
tive percentages were derived using the largest watersheds first and
continuing with other watersheds arranged in descending order of
Size.

streams which drained basins ranging in size from 0.1 to
8 mi2 (0.3 to 20.7 km2) (pl. 1). In all, gages were installed
in streams draining about 235 mi2 (610 km2) or about 90
percent of the composite area of all tributaries along the
study reach.

GAGING STATIONS AND
OBSERVATION PROCEDURES

The gaging stations, except for 18 and 26, were
established at the downstream ends of highway or
railroad structures near the flood plain of the Gila River
(pl. 1). The road structures ranged from relatively small
box culverts—concrete at highway sites, and wood at
railroad sites—to truss bridges spanning more than 300
feet (90 m) of tributary flood plain. Crest-stage gages 18
and 26 were anchored to stream channel banks.

A typical recording gage at a tributary site is
equipped with an analog-stage recorder mounted in a
metal shelter which is attached to a stilling well. The
stilling well is a corrugated metal pipe 12 inches (30 cm)
in diameter which is vertically attached to a highway or
railroad structure. A staff gage, graduated in feet,
tenths of feet, and hundredths of feet, serves as an
outside reference gage; the staff gage is attached to the
outside of the stilling well. A crest-stage gage mounted
near the stilling well serves as a reference gage. The
vertical and horizontal scales on the continuous-
recorder charts were, respectively, 1 inch equals 1 foot
(1 cm equals 0.12 m) and 9.6 inches (24.4 cm) equals 1
day.

The gages were serviced during the summer and
current-meter measurements of significant flows were
made whenever possible to develop stage-discharge

 

———7

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ. l5

relations; otherwise, only data necessary for indirect
measurements were obtained.

STAGE-DISCHARGE RELATIONS

The definition of a stage-discharge relation presented

one of the main problems in computing discharge in
many of the study tributaries. Control sections were
unstable for most streams

obtain.

CONTROLS

Alluvial channels with moving boundaries, such as

those of the study area, have no naturally occurring
permanent control sections. According to Chow (1959, p.
70), the term “control of flow” means*** "the establish-
ment of a definite flow condition in the channel or, more
specifically, a definite relationship between the stage
and the discharge of the flow. When the control of flow is
achieved at a certain section of the channel, this section
is a control section.” Even though there were no
permanent controls in the tributary streams, many
different time-variant conditions tended to control the
flow at the gaging stations. In this report only typical
time-variant conditions at gaging stations attached to
the downstream ends of box culverts and gaging
stations attached to the downstream ends of the bridges

are described; these are approximately the extremes of

controlling conditions for the sections of streams in the
study area. The box culverts in streams draining small
basins normally span only a small part of the tributary’s
flood plain. The bridges on streams draining large

basins, however, normally span the total flood plain of

the tributary.

Even though the culverts are rigid, control sections at
the culverts are not always stable for a full range of

flows. A relatively high flow normally goes through a
hydraulic jump as it approaches a culvert and dumps
most of its load of sand, gravel, and boulders while in the
subcritical state. The flow returns to a supercritical
state immediately before or after entering the culvert.
The flow remains in a supercritical state as it moves
through the culvert barrel and downstream. The control
section for high flows at a gaging station attached to the
downstream end of a culvert, therefore, is the culvert
entrance and barrel.

Low flows approaching a culvert may move into the
culvert barrel without going through a hydraulic jump.
The control section for low flows at a gaging station
attached to the downstream end of a culvert, therefore,
is the alluvial channel upstream from the culvert
entrance and alluvial deposits in the culvert barrel. The
alluvial material deposited near the culvert entrance by
high flows normally either is removed by maintenance

!______

and current-meter meas-
urements of discharge were difficult or impossible to

  

crews or moves slowly through the culvert barrel during
low flows; this material may affect the stage-discharge
relation of low flow during the time it is in place at the
entrance and during the time it is moving through the
culvert barrel.

Changes in the channel boundary alter the stage-
discharge relation at gages attached to bridges.
Changes in the stage—discharge relation normally occur
either gradually when low flows are dominant or
rapidly when high flows are dominant.

Relatively large adjustments in the channel bed often
occur at bridge structures; these adjustments, which
affect the stage-discharge relations, are known to result
from changes in the alluvial fans at the mouth of
tributary streams. The fans form as a result of
deposition of sediment carried by the fast-moving
tributary flow which is slowed as it reaches the wide, flat
flood plain of the Gila River (pls. 1 and 2). Development
of the fans causes progressive aggradation in the
tributary streams. Short periods of scour, however,
occasionally occur at the gage sites of some of the
streams as a result of shifts in the location of a tributary
stream on its alluvial fan. Channels on a fan are
self-formed and, like most others on aggrading alluvial
deposits, they have natural levees and are “in grade”
with the upstream channel. However, channels on fans
aggrade very rapidly, and frequently the beds of the
channel become higher than the rest of the surrounding
fan. When this occurs, the natural levees fail, usually
during high flow, and a new channel forms at a different
location and at a lower level on the fan. The lowering of
the point of discharge on an alluvial fan causes scouring
in the upstream channel. The concurrent scouring in
the upstream channel and filling on the fan continues
until the bed of the stream is again “in grade.” The cyclic
scouring occasionally lowers the channel bed at some of
the gages by as much as 1 foot (0.3 m).

DISCHARGE MEASUREMENTS

Two methods of measurement were used to determine
flow rates in the tributary streams, direct and indirect.
Direct methods refer to discharge measurements made
using current meters. Indirect methods refer to meas-
urements made using theoretical or empirical equa-
tions to determine flow rates after a flow event had
occurred. _

The indirect methods of determining discharge for a
given stage at the gaging stations were based on either
the standard-step method of determining surface
profiles (Chow, 1959), or the related slope-area method
of determining discharge (Dalrymple and Benson,
1967). Both methods are based on a theory of energy
balance within a reach and both are designed for
uniform flow in which the water surface profile and

 

 

i

16 GILA RIVER PHREATOPHYTE PROJECT

energy gradient are parallel to the streambed, and the
cross-sectional area, hydraulic radius, and depth re-

solution, the methods are assumed to be valid for the

flow and channel conditions that prevail in the

tributary streams provided that the energy losses are
properly considered.

In determining a stage-discharge relation using the
standard-step method, the following information was
required:

1. An assumed discharge for which the stage at the

gaging station was desired.

2. The water-surface altitude at a control section; the

control section would be downstream for subcrit-

ical flow and upstream for supercritical flow. The

starting altitude for the gaging stations at

culverts was critical depth for the discharge at

the culvert entrance; the starting altitude for the

bridge sites was taken to be critical depth at a

distance equal to 3 to 5 channel widths up-
stream.

3. The cross-sectional width, area, and wetted
perimeter at various sections along the reach for
all depths of flow within the range expected.
Reach length between sections also was re-
quired.

4. The Manning roughness coefficient n and eddy
losses at the various sections. The selection of
roughness coefficients was based entirely on the
factors that affect the value of roughness for
subcritical flow and on the

typical roughness coefficients for supercritical
flow for channels of various types were not
available. Boundary changes or channel rough-
ness may cause flow separation or surface
disturbances in supercritical flow that are not
typical in subcritical flow. The magnitude of the
energy losses due to these disturbances of
supercritical flow may not be properly covered by
the values of n that were selected.

Briefly, the computation steps in using the standard-
step method to determine stage for a given discharge
are:

1. A water surface altitude is known or assumed at

an upstream section for an assumed discharge.

2. A velocity head is computed for the upstream
section.

3. A water surface altitude is assumed for the
discharge at the next downstream section.

4. Intervening losses due to friction or deceleration
are computed through the reach between these
two cross sections. '

5. The energy balance is tested, and if it does not

balance, the assumed altitude at the
downstream section (step 3) is revised, and steps
4 and 5 are repeated until a balance can be
achieved with an acceptable tolerance.

6. Computations then proceed to the next
downstream section and continue through all
the subreaches until the section at the gage site
is reached.

The altitude for the gage site at which an energy
balance is obtained was used, along with the corres-
ponding assumed discharge, as a plotting point in
developing a stage-discharge relation for the gaging
station. Enough points were obtained using the proce-
dure described above to define a smooth curve.

The slope-area method of determining peak dis-
charges, which makes use of the Manning discharge
equation, is described by Dalrymple and BenSon (1967).
The Manning equation is

1.4
Q : #:4122/381/2, (2)

* discharge rate in cubic feet per second;
n = a roughness coefficient;

R = hydraulic radius in feet; equals the cross-
sectional area of flow A divided by the
cross-sectional wetted perimeter P; and

S = hydraulic gradient in feet per foot.

A stage-discharge relation for each gaging station

urements follows.

Current-meter measurements were extremely dif-
ficult to obtain in most of the tributary streams because
of the following reasons:

1. Stream gagers could not wade flows deeper than
about 1 foot (0.30 In) because of extremely high
velocities, rapid erosion of the channel bed around
the stream gager’s feet, and debris—including
trees, boulders, and diamondback rattlesnakes—
being washed downstream.

2. Most of the floods came after normal working hours
and, during flood, the high rates of flow normally
lasted less than 15 minutes. Cableways normally
used to measure high flow rates were not con-

—

 

 

——————7

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ. 17

the flows in the study tributaries and in the different
watersheds near the study area resulted from high-
intensity rainfall occurring during thunderstorms and
that single-peak flood hydrographs have similar shapes.
The values of m and n were 0.03 and 1.14 for the data
sets used in the analysis giving equation (4):

structed and equipment usually used to measure
floodflows at bridges was not obtained because of
prohibitive cost. The equipment probably could not
have been used if it had been available because the
high flows would have passed before the equipment
could have been readied to make a measurement. If
the high—flow equipment had'been available, the
extremely high velocities, shallow depths, and
debris would have presented other problems to
overcome in order to make a measurement.

3. Flow less than about 1 foot (0.3 m) deep could be
waded; however, the flow velocity could not be
determined accurately using the pygmy or small
Price current meters. The velocity of flow for
depths greater than about 0.2—0.3 (0.06-0.09 m)
was more than 4 ft/s (0.11 m/s), which is about
the maximum velocity that can be measured using
a pygmy meter. The small Price current meter was
not suited for depths of flow less than about 0.8 foot
(0.2 111), because of the shallow depths and because
of a separation of the fast-moving flow at the meter
when it was set at the proper position in the flow.
The separation of flow would result in about half of
the cups being completely free of water. The meter

 
 
 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
  
 
 
 
  
 
 
 
 
 
 
  
  

V = 0.03 (Qp)1.14. (4)

The standard error of estimate for the relation is about
0.3 log unit; this is about 75 percent which is the
average of a 100 percent positive error and a 50 percent
negative error.

According to equation (4) and the equations de-
veloped by Renard and Keppel (1966), Craig (1970), and
Aldridge and Condes de la Torre (written commun.,
1969), the peak—discharge versus storm-volume relation
is approximately linear for the type of storms producing
most of the single-peak floods in the region including
the study tributaries. By assuming that the relation-
ship in linear and noting that the storm hydrographs
have a triangular shape (fig. 5), the following equation
is developed:

undoubtedly was not rated for this type of flow V = 0.041(Qp)t, (5)
condition.
On two occasions experienced stream gagers at- in which
tempted to make wading measurements in flows deeper t = duration of significant rates of flow for a

storm, in hours.
The duration of significant rates of flow for most
single-peak floods occurring in the study tributaries is
estimated to range from 0.5 to 4 hours. When these t
values are inserted in equation (5), the equation reduces
to

than 1 foot (0.3 m). In both instances, the measurements
were not completed because the stream gagers were
washed off their feet. A few discharge measurements of
flow less than 1 foot deep were made using the small
Price current meter; however, the error in the data may
be relatively large.

PEAK DISCHARGE—

: .02 v6
STORM VOLUME RELATIONS V O Qp ()

and
The relation between peak discharge and storm V = 0.16Qp. (7)
volume for the tributary streams is assumed to be of the

form As shown in figure 4, for t=4 hours most of the plotted

points lie between the lines for equations (6) and (7).

V = m(Qp)n, (3)
RUNOFF
in WhiCh COMPUTATION OF DATA
V = volume of runoff, in acre-ft;
Qp = peak discharge, in cubic feet per second; Runoff from tributaries having recording gages was
and computed using general methods adopted by the US.
m, n = coefficient and exponent, respectively. Geological Survey. These methods are described by

Corbett and others (1943) and in standard textbooks on
the measurement of streamflow. In general, a mean
stage for a period of time is used with a stage-discharge
relation. Because the stage changes very rapidly during
most floods in the tributary streams, the mean stage and
average discharge for short increments of time were

Data of peak discharge and storm volume from
watersheds near the study tributaries were used to
evaluate m and n in equation (3) (fig. 4). Only data for
single—peak storms from watersheds having drainage
areas of less than 100 square miles (260 km2) were used
in the analysis. Assumptions were made that most of

 

é

I8 GILA RIVER PHREATOPHYTE PROJECT

STORM RUNOFF, lN CUBIC METRES

10,000

1 000

 

10,000

D DATA SOURCE
g 0 Files of Agriculture Research
8 Service, U.S. Dept. of
co Agriculture, Tucson, Ariz.
II
3:1 1000 A Files of U.S. Geological
,_ Survey, Tucson, Ariz.
LLl
u.1
LL 0
0 9
Fa 9V
3 4,»
Z
u.i
(D
D:
‘I‘ 100
o
E
0
54
<
LLI
n.

10

0.1

O/ 63.

100,000 1 ,000,000

 

100

ND

A

Q'

//

a: /
V=0.03 (Qp)’~“’
in which
V=storm volume
in acre-feet

A
A
A.

Q

and
Q =storm peak in
p .
cubic feet per
second

PEAK DISCHARGE, IN CUBIC METRES PER SECO

1.0

 

1 00 1 000 2000

STORM RUNOFF, lN ACRE-FEET
FIGURE 4.—Relation of peak discharge to storm volume for single-peak storms occurring in watershed having area less than 100 mi2 (260
km2) in southeastern Arizona.

computed. The stage-discharge relations computed
using the standard-step method (see p. I6) were used as
basic rating curves. As previously discussed, the
stage-discharge relations for the study streams are
subject to change because of frequent or continual
change in the physical features that control the flow.
Because current-meter discharge measurements were
not made, shifts in the basic stage-discharge relations
were based entirely on notes supplied by the stream
gagers describing changes in the channel’s properties.
Storm and seasonal volumes of runoff were obtained by
summing the volumes computed for the incremental
time periods.

As discussed in the preceding section, storm runoff
from tributaries having crest-stage gages was computed
discharge to an average peak-
dlscharge versus storm-volume relation. The method of
determining peak discharge from the stage recorded at
crest-stage gages is similar to that of determining
discharge at the recording gages. The average peak-
discharge to storm-volume relation shown in figure 4

was used as the basic discharge-volume curve. This
average curve is applicable for single-peak floods
produced by typical thunderstorms.

Storms occasionally occur in the study watersheds
that produce multipeak floods of relatively long
duration. The average peak discharge-volume relation
shown in figure 4 was shifted for these multipeak storms
in a manner similar to the shifting of a stage-discharge
relation for a changing control condition (Corbett and
others,

flow from the storm measured in streams having
recording gages, and a graph of equation (4) superim-
posed on a plot of peak discharges and volumes obtained
at the recording gages. The adjusted peak discharge-

gages.

The stream gagers provided values of the duration of
significant rates of flow It for a few storms occurring at
crest-stage sites. For these storms, storm runoff values

 

—-———*'

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ. 19

an average of less than 13 days a year. Thus, for 96
percent of the days of a year there was no flow in any of
the streams. For the study area, the number of days of
runoff per year did not increase greatly with an increase
in size of watershed, and, for a given stream in the area,
the average number of days of runoff per year was about
3 (fig. 6).

Annual peak discharges for the different streams are
given in table 3. An annual flood peak is the highest
instantaneous discharge rate occurring during a sea-
son. A peak discharge of about 8,000 ft3/s (226 m3/s) in
tributary 16 on August 5—6, 1967, was the largest peak
flow recorded in the tributary streams; a peak discharge
of about 7,000 ft3/s (198 m3/s) occurred in tributary 17
on the same dates. The largest peak discharge per
square mile was about 2,300 ft3/s (65.1 m3/s), which
resulted from a storm centering on tributary 28 on July
16—17, 1967.

 
  
 
 
 
 
 
 
 
 
 
 
  
 
 
  
 
 
 
 
  
 
 
 
  
  
   
  
 
  
 

   
  
   
   

EXPLANATION

 

Shape of typical flood
\ hydrograph

Triangle having approximately
\ same shape and volume as
flood hydrograph

DISCHARGE —*

ACCURACY OF DATA

The accuracy of streamflow data for the tributary
streams could not be determined directly. The state-
ments that follow, therefore, are of a general nature and
are based on the author’s experience in determining
accuracy of streamflow data for other sites (Burkham
and Dawdy, 1970).

The accuracy of discharge data depends primarily on
(1) the stability of the stage—discharge relation or, if the
control is not stable, the frequency of discharge
measurements; and (2) the accuracy of observations of
stage, measurements of discharge, and interpretation of
data. As previously discussed, the stage-discharge
relations for most of the study streams were unstable
and, furthermore, discharge measurements could not
be made. The streamflow data, therefore, are of poor
quality.

In reports of surface-water data published by the
Geological Survey there are accuracy statements which
say "* * * ‘Excellent’ means that about 95 percent of the
daily discharges are within 5 percent; ‘good’ within 10
percent; and ‘fair’ within 15 percent. ‘Poor’ means that
daily discharges have less than ‘fair’ accuracy.” Accord—
ing to these definitions, streamflow data for flashy flows
in mobile alluvial channels should never be rated better
than fair and rarely should they be rated better than
poor. The streamflow data for the study streams are
rated poor. The data of peak discharges and storm totals
for most of the runoff storms are probably within 100
percent of true values; the data of seasonal runoff
probably are within 50 percent of true values.

The data of days of no flow, which were of prime
importance to the water-budget analyses, are well
documented and are therefore considered excellent. As
previously discussed for a given stream, there was no
flow 96 percent of the days of a year.

 

TIME FROM BEGINNING OF FLOW -—*

FIGURE 5.——Hypothetical hydrograph for a single-peak flow event in
the tributary streams.

were computed using equation (5) as well as equation
(4): agreement between the two computed values was
fair. The values obtained by using equation (4),
however, were used.

STORM AND ANNUAL VOLUMES

The mean summer runoff from all the study water—
sheds for 1963—71 was about 1,370 acre-ft (1,690,000
m3) or 9 acre—ft per square mile (6,890 m3/km2) per
season (tables 1 and 2). The mean summer runoff from
watersheds tributary to reach 1 was about 750 acre-ft
(925,000 m3) and from watersheds tributary to reach 2
about 620 acre-ft (2,000,000 m3) in 1967 and the
minimum summer runoff was about 40 acre-ft (49,300
m3) in 1970. In reach 2, the maximum summer runoff
was about 2,220 acre-ft (2,740,000 m3) in 1971 and the
minimum was about 90 acre—ft (111,000 m3) in 1970.
The storm that produced the largest runoff occurred on
August 5—6, 1967, when the inflow to reach 1 was about
280 acre-ft (345,000 m3) and the inflow to reach 2 was
about 690 acre-ft (851,000 m3).

The tributary streamflow into reaches 1 and 2
resulted from an average of nine runoff storms per year
(table 1). The maximum number of runoff storms at
gaging sites in streams tributary to reach 1 was 15 in
1967, and the minimum number was 6 in 1965 and
1968. In reach 2 the maximum number of storms was 14
in 1967, and the minimum number was 6 in 1965, 1969,
and 1970. ‘

The watersheds contributed flow to reaches 1 and 2 on

 

 

_4__

 

NSmw wNH «a»: m m: mv NHN {{ V‘NH mm DH 1" mm 92 If

$me 3%
“““““ a bi"

GILA RIVER PHREATOPHYTE PROJECT

Bum

 

.02 53:3th

  

N £93m .V

rmwm$xum E ”co—5E

Nulmmm: .waoxnw Séxfieuamfixk (SEN GEO Kmmfigsfih Soxkkkets SSSW'A mamflw

 

0
1
I

 

 

1
1
I

fivwr odd 06m w‘m w‘mv mg: «.mm Wm N4» v.5 H‘m ?? w‘wm «Am NNmH m‘w Wm w‘wNH w‘mfi m.ww «\me ‘‘‘‘‘‘‘‘‘‘‘ Rack
bfloom Nam m6 m «mm «Hm .NN «\N mflm mm. m ?? b b ca m min a 4% S Nd ““““““““““ ?? H .auO

Nd may m‘a ?? ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘ H .gmmm
wfloH ‘ ?? H. ?? ?? H. ?? ?? ??? ““““ ??? mm .934

 

v‘. ?? ?? ?? ?? ‘ ? “““““ ???HN$N duo
““““““““““ m7: ‘Emm
“““““ ????? mm ‘w=<

?? o: v N

FLOW FROM SMALL WATERSHEDS TO THE GI‘LA RIVER, ARIZ.

 

mm 3 ? ? ? ? 1 ?

2: HH WW HH M: mm “H? H ? ? ? ? mm 3 m: om m: m m

S ? ? ? n HH HH HH HH mm WM E m x ? “““ ? 3% 3%
E ? ? 13 8 H ? h 3 y? S 3 mm “““““““““ J a bi.

? 2.2 3%
? ? J J “““““““““““ 2 23
y ? mm w 4‘ w m ohm ?? ? ““““““““““““ wish

““““ ?????? N35!”

‘‘‘‘‘ ????? «N hag

  

GILA RIVER PHREATOPHYTE PROJECT

2
1
I

 

mwlmm b3“
S 33,
1 53:53
2 b3“
1 111123“
1 ‘ vm 5»:

‘I

 

9N midfi

       

  

 

bwm ma: m H ‘ 111} i305

mm m. C‘ l} ‘11} ‘ ‘ l E anew
ONH h. \\ 2‘ ‘ ‘ ““1?“ m .amm
‘ mm m 1‘ ‘1 1111 \l‘ «L .m:<

mH : 1! \ ‘ a N OH 0% ON m m H “1111 ‘ ‘ mm 33”

‘ ‘ \ «um 13 ‘ 1 1} 1111‘ vmlmm 35w

5.: \\ l1 :1 11 I? 11 I} 11 H 11 l} m6 . ‘ 1 l ‘ ‘ Hm bah

 

 

 

 

m «xwfi mm m S: NSw ‘ ‘ «EN 111“; 1308
‘ ‘ M: ‘1‘} {C $kuka madamw

  

8mm
.eZ \Cwfifiirb

 

 
 
  

N @uamm .m
wodcfiaool Elmmmé «8.85% EAQRSUEQR ESQ 3va .mmfgses Soxkkkowa Exoumlé mamfiw

 

, ARIZ. 113

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER

 

 

   

11111111111111

    
 

w.mHN.N H.O<H on mean 0: 1 Hdb Pvm GNP $.05 de QHH H“. 0.: N,NH wHH ~3ch w‘ww hr «AVON mdww Hus o.N Race

w
Hdmw om Wm NH Hum 11 NH m 0% w. mN b H. m m m mAm mv w h «N on m v. 1 ‘‘‘‘‘‘‘‘‘‘‘‘ 11 H ,50
. 1111 1 H 1 1 1111 111 1111 1111111111 om .unww
1 1 1111 111 xxxxxxxx 111 mN,m:<

oHH «‘H 1111 m

w m d $11 was 3 H. H‘ a

 

111111 Haaorfi

‘‘‘‘‘ 41:5
111111111111 $19“an
in .aum
«fig .m=<
1 aém haw
1111 11111111111 m Nm—EW

11111111

1 H35
111§5m Lao
21:93

1 1 333m
fiémi
Mimi
ON 35

1 1111 m6 NH m

 

m 1 1 1111 1111 111 1 1 11 1111 111 1111 1111 1 11 1 1 1111 1111 1 1 111 1111

m. 1 1 1111 1111 1 1 1 1 1111 11 1 1 1 1111 1111 1 1 1 11 1111 m. 1 1 111 1111 1111 1 1 11 1111 11 1 1 mlv,m:<
v.0H 1 1 111 111. 1 1 1 1 1111 11 1 1 1 1111 A» w 1 11 «. 11 1 1 1 111 1111 1111 1 1 1111 1111 1 1 1 1 11111111111 H.m=<
HHm 111 1111 mN h 1 1 1111 mH v v w m N N mo 1 1 1111 111 1 1 1 1 vd 11 1 1 1 1 1 1 1 1111111111 bmmfiw
N‘ 1111 1111 v. _1 1 111 1111 1111 1111 1111 1111 1111 1111 1 1 1111 1 11 1.111 111 1111 1 1 111111111 H“ H H»
w‘ 1111 1111 1 1 1 1 1111 1 11 1 1 1111 1111 1111 1 1 1 1 1 1 1111111111 1 HN 33»
m. 1 1 1 1 1 11 1111 1111 1 1 11 111111111 1 1w3=w

 

ohmmH wN m6 m4» mmH 1111 mm pH woN >N N: mmm w H.N 111 1111 Has?

h.« 1111 1111 1 1 1 N N 1 1 N. 1111111 11111 PH ,Hawm
N‘mw 1111 111 1111 1 11 1111 OH m w Hm m N m m 1 1 1 1 1111 1111 1 1 1 1 11 1111 11 1111 11111111111 hé‘aawm
N‘wbm N 111 m cm 1111 wH w omH wm HVH mH mp PN m H H“ m mH 1 1 H mH 111 m. 1 1 1 111 PH1mH .m=<
m 1111 1111 1111 1111 1 1 1111 111 1111 1111 1111 1111 1111 1111 111 111 1111 1111 111 1 1 1111 1111 m. 11 11111111111111 NH Luz/w
foam mN w‘m mm. ow 1 1 Hg H H» H H m m 1111 1 1 1111 1111 cm Nm 1 1 92 com m 1 1 1 1 11111111 mtm‘mamH
m 1111 111 1111 1111 1 1 111 1111 N 1111 H 1111 1111 1 1 1 1 111 1111 111 111 1 1 1111 1.111 1111 1 1 1 11111 11111 mlN.m:<
NANAu 1 11 m‘ 1111 w 1 1 1 1 H N h w mH H m 1 1 1111 11 1 HuH H 111 1111 H 1111 1 1 11111111111111 om 33»
mam 1 1 1111 1111 N 1 1 1 1 H A» w 1 11 1111 1111 1 1 H 1 11 H ON w mH m MN 111 m1 1 11111 11111111 bN 31w

 

i

114 GILA RIVER PHREATOPHYTE PROJECT

TABLE 2.—Mean J une—October runoflfor1963—71, standard deviation of the mean, and sizes of watersheds for tributaries to the study

reach of the Gila River Phreatophyte Project
Drainage Mean Standard Drainage Mean Standard

 

1Tributary streams having recording gaging stations.

area Jun&0ctober deviation area June—October deviation

(square flow (acreJt of mean (square flow (acre-It of mean

Tributary miles) per season) (acre-ft) Tributary miles) per season) (acre-ft)

-_______---__________--_-__ - \K
14 0.38 1.2 1.0 36 0.44 3.4 3.7
15 .27 2.3 3.5 137 1.54 5.2 4.8
116 14.0 129 194 38 6.14 75.2 68.2
117 20.3 104 110 38.5 .22 12.5 16.8
18 .59 25.8 26.9 139 8.83 46.9 38.9
19 2.18 30.3 32.3 40 .59 6.5 6.6
120 7.81 100 218 41 2.62 24.8 36.5
21 .51 4.8 4.5 142 9.77 87.1 78.5
122 .84 1.9 4.0 43 .55 8.5 13.2
23 .73 6.1 7.1 44 .42 9.1 10.7
24 1.14 43.9 67 7 45 .40 15.8 12.3
125 18.7 71.0 170 46 1.52 22.2 13.1
26 4.40 30.2 37.4 47 4.69 44.8 42.2
127 9.22 60.3 77.3 48 1.83 60.0 64.5
28 .79 30.1 50.1 49 .11 11.0 10.4
29 .54 11.1 17.7 50 .48 19.4 22.8
130 1.81 43.7 47.4 50.5 .27 2.0 2.9
31 1.17 9.0 6.4 151 12.1 69.6 68.1
32 .70 36.8 24.0 152 .99 33.9 31.0
33 1.28 1.2 2.4 53 .81 9.0 21.5
34 .22 6.3 6.1 54 2.38 32.6 51.3

35 1.87 10.0 18.7 \

TABLE 3.—Annual peak discharge for streams tributary to the study reach of the Gila River Phreatophyte Project
[Discharge, in cubic feet per second]
Tributary 1963 1964 1965 1966 1967 1968 1969 1970 1971
O

14 -__- 20 30 20 0 30 5 20
15 _-_- 0 60 0 140 0 0 0 60
16 ---- 850 440 430 8,000 40 100 5 3,500
17 2,000 1,300 500 430 7,000 20 100 20 1,500
18 _--- 80 850 120 240 0 50 450 350
19 ---_ 360 200 50 460 60 50 100 450
20 _--_ 440 40 10 2,800 20 20 30 3,120
21 -__- 90 150 0 90 4O 0 40 90
22 ---- 65 10 50 10 0 0 70
23 --_- 20 150 0 180 40 30 0 90
24 50 500 0 0 1,700 0 0 0 1,100
25 200 190 60 0 4,000 100 30 40 400
26 0 540 700 120 720 20 180 0 90
27 60 600 1,600 0 1,800 0 100 0 1,440
28 0 240 1,700 0 1,800 60 300 O 100
29 0 0 640 40 280 100 40 0 40
30 250 460 130 0 800 10 530 20 560
31 10 80 60 150 170 290 80 10 100
32 40 500 450 670 900 360 350 60 500
33 0 120 0 0 0 0 60 0 0
34 80 120 110 130 90 0 0 0 30
35 10 230 140 510 10 0 5 0 70
36 20 40 5O 10 20 40 20 0 70
37 0 50 110 0 10 40 60 40 50
38 130 720 1,300 1,200 120 1,900 200 100 360
38.5 -__- _-__ 140 370 70 120 300 20 160
39 130 460 820 1,140 180 880 200 50 480
40 20 50 200 120 10 150 50 20 90
41 10 30 3,000 20 10 1,000 20 140 130
42 50 3,000 4,700 850 1,400 1,500 700 70 400
43 -_-- 50 350 40 50 ‘ 40 60 5 60
44 -__- 80 260 120 130 60 60 0 5
45 -__- 350 130 110 300 110 150 0 60
46 _--_ 220 230 210 230 400 280 0 60
47 _--_ 300 590 80 60 40 5 0 500
48 __-_ 160 540 540 2,330 1,000 440 50 510
49 ---_ 60 80 5O 50 8O 90 110 150
50 __-- 50 100 150 340 70 230 200 450
50.5 --_- __-_ ---- ---- _--_ _--_ -_-_ -_-- __-_
51 ---_ 520 20 210 1,200 120 750 30 330
52 _-__ 0 0 1,100 500 90 0 80 230

53 _--_ 20 20 20 100 0 0 10 480
54 _--_ 1,000 200 0 40 0 0 20 700
1A dash indicates that records were not obtained for the indicated period.

 

———————'

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ.

115

DRAlNAGE AREA, lN SQUARE KILOMETRES
1.0

.1
0

NUMBER OF DAYS FLOW OCCURRED

1.0

0.1 1.0

    

30

DRAINAGE AREA, IN SQUARE MILES

FIGURE 6.-Relation of number of days of flow to size of watershed.

COMPARISONS OF RUNOFF FROM THE STUDY
BASINS WITH RUNOFF FROM NEARBY BASINS

An investigation was made to determine whether the
runoff—producing characteristics of the study area were
representative of other basins in the region by
comparing the average annual discharge data for the
study watersheds with discharge relations previously
developed for other watersheds in the region. The
purpose of the comparison was to determine whether
the streamflow data for the study tributaries agreed as
closely with the relations as did the data used in
developing the relations.

A comparison of the study data with runoff relations
developed as part of a nationwide evaluation of the
streamflow data collection program of the Geological

Survey (Moosburner, 1970) was the primary method of

analysis used in this investigation. Comparisons of the
tributary data in the study area with runoff relations
developed for other areas by Burkham (1966, 1970)
were of secondary importance. These comparisons
offered an excellent opportunity to study prediction
errors for the different relations.
The regression equations by Moosburner (1970) were
developed using the model
y = aAbBCCd, (8)
where y is a streamflow characteristic, such as mean
annual flow; A, B, and C are physical and climatic basin
characteristics, such as drainage area or precipitation;

 

    
  
  
 

and a, b, c, and d are coefficients or exponents obtained
by regression. The following procedure was used in
deriving an equation for a region (Moosburner, 1970, p.
20—21):

1. Compute an initial regression equation.

2. Test the coefficients for statistical significance at the

95 percent confidence level.

3. Drop the characteristics that were

insignificant.

4. Compute a regression equation using only the

significant parameters.

5. Compute a standard error of regression.

6. Determine residuals—the difference between the
streamfiow characteristics determined by the
regression analysis and the streamflow charac-
teristics that are measured.

7. Plot the residuals on a map of Arizona to determine
any regional variation.

The procedures used by Burkham (1966, 1970) to
develop equations were the same as those of Moos—
burner except the regional studies described in step 7
were not made.

According to Moosburner (1970, p. 21), two regions in
Arizona had enough data so that regression analyses
could be made (fig. 7). The equations derived for regions
1 and 2 are (Moosburner, 1970, table 3)

found to be

QS:1.82 X 10‘3A0-71(PS)2.25 (9)

and

QS = 5.89 x 10—3A0-71(PS)2-08, (10)

 

 

i

116 GILA RIVER PHREATOPHYTE PROJECT

I4° ”2°
" EXPLANATION

 

area and area in which

  

 

 

 

I I O ' W
. l |
: g I .
I “ I l
l 1 Approximate boundary
I between undefined
l I I
I 36° regression analyses have
been made
g l
E i
Q
\
3 Region 1

See regression eq uation(9)

< 9. Region 2
Ir” See regression equation(10)

   

 

   

 

 

 

 

I
1 l
I o.
l
34° fig 340
A I
J < ‘2
I l “J
I Lu
-A- I .J
2
2 l m
l Lu
”’3- I “5
‘— O
‘2‘
\
Yuma G R_A A M ,
\\‘ _______.____\_
32°
\ o c H l s E
0 so 100 MILES 2
W Douglas
4
o 50 100 150 KiLOMETRES ”0°
FIGURE 7.—Regions in Arizona where regression analyses have been made. From Moosburner (1970).
in which results in acre-ft per year, equations (93 and (10) reduce
QS = seasonal mean J uly—September discharge to
in cubic feet per second; '
A = area of drainage basin in square miles, Qs=26.6A 0-71 (11)
and for region 1 and
(PS) = seasonal mean May—September precipita- QS=61.5A 0‘71 (12)
tion in inches.

The mean May-to-September precipitation for the for region 2. Equations (11) and (12) along with
study tributaries is about 7.0 inches (180 mm). When equations developed by Burkham (1966, fig. 21; 1970,
the 7.0 is inserted and when changes are made to give fig. 15), and data obtained for the study tributaries are

 

————’i

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ. 117

AREA, IN SQUARE KlLOMETRES
1.0 . 1o

 
  
 
 

2000

1000
1000

-71, W

63-71, IN ACRE-FEET
SEASO
THOUSANDS OF CUBIC METRES

100
100

NAL RUNOFF FOR YEARS 1963

10

SEASONAL RUNOFF FOR YEARS 19

0.1 1.0 10 100
AREA, lN SQUARE MILES

EXPLANATlON

(Data obtained as part of the Gila River
Phreatophyte Project)

0
For stream having continuous-stage recorders
A

Data from an Agricultural Research Service
project near Safford (Burkham, 1970, table 3)

Curves developedfrom equations by Moosburner
(1970, table 3)

1

Refers to region 1 shown in figure 6
2

 

 

Refers to region 2 shown in figure 6

//
Curve from report by Burkham (1966, fig. 21)

 

_____’-——-
Curve from report by Burkham (1970, fig. 3)

FIGURE 8.—Relati0n of mean annual runoff to size of basin.

plotted in figure 8. According to Moosburner (1970), the equation (12) is applicable.
study tributaries lie Within region 2 (fig. 7), for which The standard error of prediction (SE)p for equation

 

‘—

 

fi

I18

(12), when used to estimate mean seasonal discharge for
the study tributaries, was determined by comparing the
mean annual discharge obtained from equation (12) Q8,
with measured mean annual discharge Qm, where the
d discharge (SE)m is
mean annual discharge
were not used in the
osburner, therefore, the
rrors R m in the data for
ntrol data, are indepen-
RS. The variance of the
mean annual discharge
1 discharge, (SE )2
difference between

    
   
  
   
  
  
 
   
  

Qm for the study tributaries
regression analysis made by M0
assumption is made that the e
the study tributaries, called c0
dent of the prediction errors
difference between computed
and measured mean annua
includes the variance of the
sured discharge in the cont
discharge.

The variance of the difference between computed and
measured discharge may be estimated as follows:

23 (Q.

i:1

s—m’

.—Qm,.>2

N

in whichN is the number of sets of discharge data used
in the analysis, 1' denotes individual sets, and the other
parameters are as previously defined. The mean annual
discharges QS and Qm can be described by the equations

(SE) 2s—m

 

(13)

5

QS=QT¢RS

Qm=QTier (15)
in which QT is true mean annual discharge. Equations
(13, (14), and (15) are combined to give

(14)
and

 

N
2 (:RsiiRmf
2 i:1
(SE) S_m= N , (16)
or

(SE) 2 S_m=(SE)SZ+(SE)m 2 . (17)

The expected value of the variance is
E (SE) 2 s_m=052+o-m2=(cr) 2 s_m , (18)

errors in the control group are
independent of the prediction errors. Therefore, (752:
(0)2 s—m‘O'mZ where 02 denotes a "true” or population
variance as opposed to 82, which is estimated on the
basis of data.

if the measurement

The standard error of the mean for the measured flow
(SE)m, in percent, was determined using the equation

GILA RIVER PHREATOPHYTE PROJECT

(19)

(3mm: '/CV2 + (SE)c 2

Y

 

in which

= coefficient of variation for the annual
discharges in percent; equals the standard
deviation of the annual flows (DE), divided

by the mean of the annual flows times 100;
= the number of years used in determining

the mean of the annual discharges—serial

correlation between annual discharges is
assumed to be insignificant; and

= standard error, in percent, of an annual
flow value which is a measurement or
computational error.

A value of 100 for CV was used in equation (19) to
compute (SE)m. This value of CV is large compared to
values that have been computed for other watersheds in
Arizona (McDonald, 1960, table 1; Burkham, 1970, p.
31—32; Moosburner, written commun., 1971); however,
to be on the high side when (SE)m is determined, the
value of 100 is used for this study. The standard error of
computation for the control data (SE)C probably is no
larger than 50 percent; however, to be on the high side, a
value of 100 percent is used. When values of 100 for CV,
100 for (SE)C, and 8 years for Y are used in equation (19),
a standard error of the mean for the control data of about
50 percent is obtained.

The average coefficient

Cv

Y

(SE )c

of variation CV computed for
tributaries in which continuous recorders were main-
tained was 133. This probably is significantly larger
than it would have been if the streamflow data did not
contain large computational errors. If values of 133 for
CV and 100 for (SE)c are used in equation (19), a value of
60 for (SE)m is obtained.

An average value for (SE)S_m of 250 was computed
using equation (13) when mean annual discharges
obtained from equation (12) are compared with meas-
ured mean annual discharges. Only the data for
watersheds having recording gages were used in
determining (SE)S_m; the data include those for 12
watersheds in the study area and 3 watersheds under
study by the U.S. Agricultural Research Service
(Burkham, 1970, table 3). The standard error of
prediction 0-8 for equation (12); when estimating
discharge from the study tributaries, is about 240
percent, which was determined by using values of 250
percent for 0-S_m and 50 percent for “in in equation (18).

Many factors may have been involved in causing the
standard error of prediction to be large when equation
( 12) is used to estimate runoff from the study

 

FLOW FROM SMALL WATERSHEDS TO THE GILA RIVER, ARIZ.

watersheds. The large prediction error, however,

probably results mainly from two related reasons:

1. The basin characteristics for the watersheds of this
study which affect water yields are significantly
different from those for watersheds studied by
Moosburner; basin size and shape are known to be
significantly different; there may be other differ-
ences.

2. The region for which equation (12) is applicable is
improperly defined (fig. 7).

Prediction errors of about 100 percent are determined
for equation (11) and for the two equations by Burkham
(1966, 1970) when they are used to compute average
annual discharges for 1963—71 for the study water—
sheds. The procedure used in determining prediction
errors for these equations is the same as that described
for equation (12). The runoff data for the three
watersheds under study by the U.S. Agricultural
Research Service were not used in computing errors for
Burkham’s 1970 equation because these data were used
in developing the equation.

CONCLUSIONS

Conclusions reached as a result of this study are:

1. Feasible methods of accurately measuring flashy
streamflow moving in a supercritical state in
channels of movable boundaries currently are not
available.

2. Data obtained during the study were adequate for
the water budget studies of the Gila River
Phreatophyte Project but only because there was
no flow in any of the streams for more than 96
percent of the days of the year. Periods of no flow
are well documented and therefore considered
reliable. The data of peak discharges and storm
totals for most of the runoff flows probably are
within 100 percent of true values; the data of
seasonal runoff probably are within 50 percent of
true values.

3. Prediction errors for discharge equations developed
as part of the nationwide evaluation of the
streamflow data collection program of the U.S.
Geological Survey (Moosburner, 1970) and for
discharge equations developed for other studies
(Burkham, 1966, 1970) are 100 to 250 percent

 

119

when the equations are used to estimate average
annual discharge for basins having climatic and
basin characteristics similar to those of the study
tributaries. The climatic and basin characteristics
for the study tributaries are similar to many others
in the Basin and Range physiographic province
(Fenneman, 1931).

REFERENCES CITED

Benson, M. A., 1964, Factors affecting the occurrence of floods in the
Southwest: U.S. Geol. Survey Water-Supply Paper 1580—D, p.
D7—D72.

Burkham, D. E., 1966, Hydrology of Cornfield Wash area and effects of
land-treatment practices, Sandoval County, New Mexico, 1951—
60: U.S. Geol. Survey Water-Supply Paper 1831, 87 p.

1970, Precipitation, streamflow, and major floods at selected

sites in the Gila River drainage basin above Coolidge Dam,

Arizona: U.S. Geol. Survey Prof. Paper 655—B, p. B1—B33.

1972, Channel changes of the Gila River in Safford Valley,
Arizona, 1846—1970: U.S. Geol. Survey Prof. Paper 655—G, p.
G1—G24.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis of streamflow
data for an alluvial stream: U.S. Geol. Survey Prof. Paper 655—0,
p. 01—013.

Chow, Ven Te, 1959, Open-channel hydraulics: New York, McGraw-
Hill Book Co., 680 p.

Corbett, D. M., and others, 1943, Stream-gaging procedure, a manual
describing methods and practices of the Geological Survey: U.S.
Geol. Survey Water-Supply Paper 888, 245 p.

Craig, G. 8., Jr., 1970, Synthesizing hydrographs for small semiarid
drainage basins, in Geological Survey research 1970: US. Geol.
Survey Prof. Paper 700—D, p. D238—D243.

Culler, R. C., and others, 1970, Objectives, methods, and environ-
ment—Gila River Phreatophyte Project, Graham County,
Arizona: U.S. Geol. Survey Prof. Paper 655—A, p. A1—A25.

Dalrymple, Tate, and Benson, M. A., 1967, Measurement of peak
discharge by the slope-area method: U.S. Geol. Survey
Techniques of Water Resources Inv. book 3, chap. A2, 12 p.

Fenneman, N. M., 1931, Physiography of Western United States: New
York, McGraW—Hill Book Co., 534 p.

McDonald, J. E., 1960, Variability factors in mountain-watershed
hydrometeorology in an arid region: Arizona Acad. Sci. J our., v. 1,
no. 3, p. 89—98.

Moosburner, Otto, 1970, A proposed streamflow-data program for
Arizona: U.S. Geol. Survey open-file report, 55 p.

Renard, K. G., and Keppel, R. V., 1966, Hydrographs of ephemeral
streams in the Southwest: Am. Soc. Civil Engineers Proc., J our.
Hydraulics Div., no. 92, no. HY2, p. 33—52.

Sellers, W. D., ed., 1960, Arizona climate: Arizona Univ. Press,
Tucson, 60 p.

Thomas, H. E., 1962, The meteorologic phenomenon of drought in the
Southwest: U.S. Geol. Survey Prof. Paper 372—A, p. A1—A43.

 

 

fiGPo 690- 249—1976

 

UNITED STATES DEPARTMENT 'OF THE INTERIOR
GEOLOGICAL SURVEY

110°22'30"

3302230"

EXPLANATION

WV

_\

 

_//

Approximate boundary of Gila River Phreatophyte
Project

A

Streamflow-gaging station on Gila River (equipped
with continuous-stage recorder)

o7
Streamflow-gaging station on tributary streams
(equipped with continuous-stage recorder).
The 7 is the number assigned to the station
and to the tributary basin

.5
Crest-stage gages The 5 is the number assigned to the
station and to the tributary basin

Boundary of tributary basin where runoff data have
been obtained

 

The runoff from the shaded area is not measured
before reaching the boundary of the study reach

of the Gila River Phreatophyte Project

NOTE: Gaging stations 1 to 13 and 55 to 63
were established in 1962—63, however,
no runoff data were obtained at these
stations

 

  

 
 

 
  

“e m \S‘\ I \
C A R L/(S R E,
S A N NORMALkOOL ELEVATION 21:25]

   
 

_l\ c
CO ‘

  

u
\\

* {INAL

A“

 

    
 
  

  

\\ r ,
1 6 and % D‘s/f ,
,« , -\‘ u
; r /j «S» __
. ‘ I I]
or //I P}; a"
° 1“ cgas»
W F? £511;wa
\ _ ‘ » /*\
\ NJOGG 0,! ‘\5\§ 31/2

        

 

 

     
  

T /\
, go
\.,\ 00W“

‘\\ L ' , \\
‘VABM no \\ ex T:
,9 ' _ \\ \ f\_
H W _ \T‘D‘tfls “I
s \ x
. V T

 

 

muss

 

 

 

    
    

   

   

  
  
 
  
  
 
  

 
  
  
 

,, a L
James”

.,

  

as I, 4
NJ) ,7 ,. A:
I" $;\\\H<{} (5 \fcrflw’IJE/gjk J

v

  
  

  

\Gila Riverneafij Calval? a
L I ,V j ' l '\\ E
at [,gfrosa Sectign fi)‘

‘ I. ' ,

\\i

as:

  

34/

,Wa' 0W

 

__ (If ,
xv s _ _
as K wsIt‘_ I

0 ~_,

2 ' i ‘
,2
VV1flJ‘fl—‘fi//)/’ll f,

I; mi \
VABM \ 396 »

 

 

 

 
  
 

 

$5
,7 We

, é

 

 

    

Eartha}?

s E I
Ayn» (
, , ‘ \
\ L \ ’

 
  
  
 

 
   
  
 
 

,gl
m T?“ 93'

*L

r :QOL’VL’I 96E \

 

 

 

an \ I L‘ ,K
1% Riv‘érgn/elar Bylas

 

 

 
 
 
  
 
   
  
 
 
 
 
      
  
 
 
 
 
 
 
 
 
  
 
 
  
 
 
   
 
 
  
 
 
  
  

   
  

:3

 

 

 

 

 

 

 

[@210

 

"rm

2

 
  
  
  

     
  
    
     

 

    

   
  

 

PROFESSIONAL PAPER 655—1
PLATE 1

110001
33 '22'30”

    

5‘ A

 

 

 

(r
m
z

1:}
r3

C
x,

 

 

 

 

   
 
 
 

 

  

 

  
     
 

 

 
  
 

 
  
  
 
 

   
   

       

 

 

 

    
 

   
  
 
  

 

 

 

 

 

 

 

 

 

 

 

  
 
  

 

 

 

 

   
 

33°00
11o’i30'

Base from U.S. Geological Survey
Bylas, 1960; San Carlos Reservoir, 1962,

1262,500; and Ash Creek Ranch, Branaman E E
Spring, Mount Triplet, 1966; 124,000 3 5°
% E
E 5
E

APPROXIMATE MEAN
DECLINATION, 1976

 
 

o f ,L

 

”to”
f ‘

  

 

amass:

  

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SCALE 1:62 500

 

 

 

1 V2 0 1 2 3 4 5 MILES
I—I I—I I—i I—I I—I I—-—-——-—-———I I—-—-——I I———-——-—+
1 .5 0 1 2 3 4 5 6 7 KILOMETRES
I-II-II—II—II—I. I I——-—I I—-——I I———I

 

CONTOUR INTERVAL 40 FEET IN NORTHERN PART
OF MAP AND 80 FEET IN THE SOUTHERN PART
DATUM IS MEAN SEA LEVEL

MAP SHOWING WATERSHED BOUNDARIES AND INSTRUMENTATION LOCATIONS

MAP LOCATION

 

   

 

  

13

  

 

BM 3033/ ,

 

~ ~ 33” 00"

Interior—Geological Survey, Reston, Va.—1976 —G75039 1103 00.

 

 

UNITED STATES DEPARTMENT OF THE INTERIOR

GEOLOGICAL SURVEY

ALTITUDE,ABOVE MEAN SEA LEVEL

 

    
 

EXPLANATION
1914

------ 1947 -------
——— 1964 ———

.00..... 1968 ‘00000000

Contour interval 5 feet

A, Aerial view of the alluvial fan in 1964 (see figs. 2 and 9 in
report by Burkham, 1972); the contour lines illustrate the rate
of growth of the fan, and profile A—A' shows the altitude
above mean sea level. B, Mouth of the tributary stream in
1909, looking from the Gila River toward the railroad bridge
spanning the tributary stream; note the height of the banks as
compared to that of the wagon and team. 3’, The same view
in 1969. C, Mouth of the tributary stream in 1909, looking
from the railroad bridge toward the Gila River; the barren
condition of the flood plain was caused by the floods of
1905—6 (arrow indicates the location of an old flood-plain
remnant). C’, The same view in 1969. The bottom-land
vegetation was eradicated in 1966 to control evapotranspira—
tion (see fig. 9 in report by Burkham, 1972).

 

 

 

 

A A,
FEET «963 96“ METRES
2560 — .-//\ — 780
— 775
2540 —
— 770
2520 I I ' I 768
0 2 6 8 1o 14 16 FEET
I I I j
0 2 3 4 METRES

DISTANCE FROM REFERENCE POINT

A

PHOTOGRAPHS AND GRAPH SHOWING DEVELOPMENT OF THE

Contour shows rate of growth of fan

 

PROFESSIONAL PAPER 655-1
PLATE 2

 

 

 

 

HHIM‘HHMHEHilrrru,

llllllillllll I Elli?

l

ALLUVIAL FAN

AT THE MOUTH OF A TRIBUTARY TO THE GILA RIVER

 

 

 

Q5 I‘D
Cy

Hydraulic Effects of Changes in
Bottom-Land Vegetation on Three
Major Floods, Gila River in
Southeastern Arizona

’

RT“
3155

“GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—]

 

 

MAR 16 1976
w a no

 

Hydraulic Effects of Changes in
Bottom-Land Vegetation on Three
Major Floods, Gila River in
Southeastern Arizona

By D. E. BURKHAM

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—]

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE,WASHINGTON:1976

 

UNITED STATES DEPARTMENT OF THE INTERIOR

THOMAS S. KLEPPE, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Director

 

 

Library of Congress Cataloging in Publication Data

Burkham, D. E. 1927—
Hydraulic effects of changes in bottom-land vegetation on three major floods, Gila River in southeastern Arizona. (Gila River
Phreatophyte Project)

(Geological Survey Professional Paper 655—1)

Bibliography: p. 14

Supt. of Docs. no.: 119.161655—1

1. Gila River—Floods. 2. Hydraulics. 3. Botany~Ecology—Gila River. I. Title: Hydraulic effects of changes in bottom-land
vegetation... II. Series. 111. Series: United States Geological Survey Professional Paper 655—] .

QE75.P9 no. 655—][GB1225.A6] 557.3'08s[551.4'8] 75—619028

 

For sale by the Superintendent of Documents, US. Government Printing Office
Washington, DC. 20402
Stock Number 024—001-02763-1

 

‘4—

 

 

CONTENTS

 

 

Page Page
Abstract __________________________________________________ J1 Analyses of data—Continued
Introduction ______________________________________________ 1 Changes in altitude of bottom land ____________________ J7
Characteristics of the study reach __________________________ 2 Effects of changes in vegetation ________________________ 7
Physical setting ______________________________________ 2 Discussion of results ______________________________________ 10
Undesirable characteristics of the site __________________ 2 Hydraulic effects for the 1965 and 1967 floods __________ 10
Basic data ________________________________________________ 3 Hydraulic effects for the 1967 and 1972 floods __________ 11
Analyses of data __________________________________________ 3 Range in hydraulic parameters along the study reach ____ 12
Mean velocities and mean depths ______________________ 4 Summary and conclusions __________________________________ 13
Roughness coefficients ________________________________ 5 References ________________________________________________ 14
ILLUSTRATIONS

Page

PLATE 1. Maps and aerial photographs of the study reach of the Gila River, showing station and cross-section locations, and area of
flooding __________________________________________________________________________________________________ In pocket

2. Cross-sectional profiles and maximum stage for three floods at nine sections along the study reach of the Gila River, south;
eastern Arizona __________________________________________________________________________________________ In pocket
FIGURE 1. Photographs showing stream channel and flood plain of the Gila River in 1964 and 1967 _______________________________ J2
2. Graphs showing hydraulic characteristics at peak discharge for three floods in the study reach of the Gila River ___________ 8

TABLES

Page
TABLE 1. Hydraulic parameters for peak discharges, floods of December 1965, August 1967, and October 1962, Gila River ________ J4
2. Peak discharge, Gila River at Calva, Ariz., 1963—72 ________________________________________________________________ 5

 

CONVERSION FACTORS

 

Multiply English units By To obtain SI (metric) units
feet (ft) 0.3048 metres (m)

miles (mi) 1.609 kilometres (km)

cubic feet per second (ft3/s) .02832 cubic metres per second (m3/s)

III

 

 

 

 

GILA RIVER PHREATOPHYTE PROJECT

 

HYDRAULIC EFFECTS OF CHANGES IN BOTTOM-LAND
VEGETATION ON THREE MAJOR F LOODS,
GILA RIVER IN SOUTHEASTERN ARIZONA

By D. E. BURKHAM

ABSTRACT

Changes in bottom-land vegetation between December 1965 and
October 1972 apparently caused significant differences in stage, mean
cross-sectional velocity, mean cross-sectional depth, and boundary
roughness at peak discharges of three major floods in an 11.5-mile
(18.5 km) study reach of the Gila River. The first flood, which had a
peak flow of 39,000 ft3/s (1,100 m3/s), occurred in December 1965 when
the dense bottom-land vegetation was dormant. The second flood,
which had a peak discharge of 40,000 ft3/s (1,130 m3s), occurred in
August 1967 when the vegetation had large amounts of foliage; how-
ever, the vegetation had been eradicated in the upstream half of the
study reach prior to this flood. The third flood, which had a peak
discharge of 80,000 it 3/s (2,270 m3/s), occurred in October 1972; the
vegetation had been eradicated in the whole study reach prior to this
flood. Compared to the 1965 flood, the large amounts of foliage in the
uncleared half of the reach during the 1967 flood apparently caused a 7
percent decrease in mean velocity, a 6 percent increase in mean depth,
and an 11 percent increase in the Manning roughness coefficient at
peak stage. Compared to the 1965 flood the clearing of the study reach
apparently caused a 25 percent increase in mean velocity, a 15 percent
decrease in mean depth, and a 30 percent decrease in the Manning
roughness coefficient at peak stage in the 1967 and 1972 floods.

The mean velocities of the three peak flows were relatively low
where large parts of the flows moved across the meandering stream
channel; the Manning coefficients and the mean depths were rela-
tively large in these segments. After the first flood, scour was noted at
seven of the nine cross sections in the study reach. After the second
flood, fill was observed at all the cross sections, and, after the third
flood, scour was observed at six sections. From 1964 to 197 2, there was
a net scour at .only one section, section 7, where the mean cross-
sectional velocity was relatively large for the three floods. Effects of
changes of bottom-land vegetation on scour and (or) fill could not be
determined.

INTRODUCTION

Saltcedar (Tamarix chinensis Lourl) has created
problems along many streams in the arid and semiarid
regions of the United States. Since about 1930 the plant
has spread rapidly, consumed large amounts of water,
and, in many streams, created potential flood hazards
(Robinson, 1965, p. 1). The problems intensify as the
demand for water mounts, the need for reducing flood

1Also referred to as Tamarix pentandra and Tamarix gallica.

 

hazards grows, and at the same time the areal extent
and density of the plant increases. Management of the
saltcedar is necessary to lessen the magnitudes of the
problems. As a remedial measure saltcedar has been
eradicated along several streams in the western United
States. The effectiveness and the side effects of this
measure are not well documented.

The flood plain of the Gila River in southeastern
Arizona is an area where the vegetation has been man-
aged. The low-benefit, deep~rooted vegetation, mostly
saltcedar (Tamarix chinensis Lour) and mesquite (Pro-
sopisjuliflora var. velutine (Woot.) Sarg.), was replaced
with a beneficial short-rooted grass (Culler, 1965,
p.33—38). The saltcedar and mesquite trees are known to
increase both the resistance to flow and the stability of
the flood-plain boundary. Therefore, replacement of
these trees with grass is likely to cause changes in rates
of erosion and deposition, and to cause changes in chan-
nel width, depth, sinuosity, gradient, roughness, and
even channel location.

The main purpose of this report is to describe the
apparent differences in hydraulic characteristics of the
Gila River during three major floods owing to changes in
bottom-land vegetation. The types of change in vegeta-
tion relevant to this study are seasonal increase in
foliage and plant eradication. The hydraulic parame-
ters studied are stage, mean cross-sectional velocity,
mean cross-sectional depth, and the Manning rough-
ness coefficient at peak discharge. Changes in the mean
altitude of the bottom land as a result of the floods also
are described. The floods occurred in December 1965,
August 1967, and October 197 2, with peak discharges of
39,000, 40,000, and 80,000 ft3/s (1,100, 1,130, and 2,270
m3/s). These floods have a return interval of about 17
and 50 years, and they were the largest in the study
reach since 1917 (Burkham, 1970, figs. 16 and 23).

Discussions, descriptions, methods, and analyses pre-
sented in this report deal with averages, lumped

J1

 

 

J2

parameters, and approximations. The study reach,
basic data, and methods of determining stage, mean
cross-sectional velocity, mean cross-sectional depth,
and changes in the mean altitude of the bottom land
have been described in detail in previous reports (Culler
and others, 1970; Burkham and Dawdy, 1970; Burk-
ham, 1970; Burkham, 1972; US. Geological Survey,
1963—72); therefore, these parameters are described only
briefly in this report. Procedures used in determining
the Manning roughness coefficients and determining
differences in the study parameters, however, are de-
scribed in detail. Errors in the data were not deter—
mined, but in some cases they are discussed in a general
way.

This report is one of several chapters of a, series which
describes the environmental variables pertinent to the
Gila River Phreatophyte Project.

CHARACTERISTICS OF THE STUDY REACH
PHYSICAL SETTING

The study reach is in southeastern Arizona at the
downstream end of the Safford Valley (pl. 1). The valley
is filled with alluvial material that ranges in size from
clay to small boulders. The study reach is 11.5 mi (18.5
km) long and includes about two-thirds of the study
reach of the Gila River Phreatophyte Project (Culler
and others, 1970, p. 14). Reach 1 is defined as that part of
the study reach extending downstream from the bridge
on US. Highway 70 near Bylas, Ariz., to the railroad
bridge that spans the Gila River 2 mi (3 km)
downstream from Calva, Ariz.; reach 2 extends
downstream from the railroad bridge to the confluence
of the Gila River and Salt Creek. Reach 2 extends into
the upper part of San Carlos Reservoir (Culler and
others, 1970, p. 8). The width of bottom land inundated
by the floods studied ranges between 1,500 and 4,000 ft
(460 and 1,200 m). The stream channel is from 80 to 200
ft (24 to 61 m) wide and from 6 to 10 ft (1.8 to 3.0 m) deep
at banktop level; it is a pool-and-riffle type channel with
a slope of about 0.002. The flood plain was covered by a
dense growth of saltcedar and mesquite during the flood
of December 1965; however, this vegetation was eradi-
cated in reach 1 prior to the flood of August 1967 (fig. 1)
and in both reaches prior to the flood of October 1972.
Gaging stations at the ends of the two reaches are Gila
River near Bylas, Ariz; Gila River at Calva, Ariz.; and
Gila River near Calva, Ariz. (Burkham, in Culler and
others, 1970).

UNDESIRABLE CHARACTERISTICS OF THE SITE

Parts of the study reach were not ideal for the applica—
tion of equations in determining hydraulic characteris-
tics, especially in determining Manning roughness
coefficients. The most important factors in this regard

 

GILA RIVER PHREATOPHYTE PROJECT

were (1) the bridge at US. Highway 70 (pl. 1); (2) a dike
extending downstream from the highway bridge; (3) the
railroad bridge near Calva; (4) the varying pool level of
the San Carlos Reservoir; and (5) the changeability of
the channel boundary. Factors 1 to 4 are manmade;
factor 5 is a natural phenomenon.

The bridge at US. Highway 70 and the dike extend-
ing downstream from the bridge confined the flow dur-
ing all three floods causing relatively high cross-
sectional velocities. The dike was constructed prior to
December 1965 to protect cultivated land from flooding.
Water spilled over the dike near the bridge during each
of the floods; however, the rates of flow on the north side
of the dike are unknown. The confined flow caused scour
during the December 1965 flood along the outer edges of
the south flood plain downstream from the bridge. (See
section entitled “Discussion of Results”)

The railroad bridge probably did not significantly af—
fect the hydraulic characteristics being studied during
the 1965 and 1967 floods because the bridge spanned the
entire flood plain and the only confinement of the flow
was due to bridge pilings which are about 1 ft (0.3 m) in
diameter. However, after the north end of the bridge
was partly destroyed by fire in 1970, it was repaired by
construction of an embankment across about 850 ft (260
In) of the 1,500-ft (460 m) span. In the 1972 flood, the
embankment significantly affected the hydraulic
characteristics being studied near the bridge. (See sec-
tion entitled “Discussion of Results”)

The San Carlos Reservoir reached a relatively high
pool level in 1968 inundating a part of reach 2. The high

 

FIGURE 1.-—Stream channel and flood plain of the Gila River in 1964
and 1967. A, Looking upstream from the railroad bridge near Calva
in 1964; the size and density of saltcedar are typical for the reach. B,
Looking upstream from the railroad bridge near Calva in 1967; the
bottom-land vegetation was eradicated in 1966 in an attempt to
control evapotranspiration. The stream channel at the site is from
80 to 200 ft (24 to 60 m) wide and from 6 to 10 ft (1.8 to 3.0 m) deep at
banktop level.

 

EFFECTS OF BOTTOM-LAN D VEGETATION ON MAJOR FLOODS, GILA RIVER, ARIZONA

pool level caused deposition of sediment in the down-
stream end of the reach Which decreased the size of the
stream channel and increased the altitude of the flood
plain in several places by more than 5 ft (1.5 m). During
the recession of the lake level, the Bureau of Indian
Affairs straightened and enlarged the stream channel
downstream from reach 2 by dredging; this caused ero-
sion of the alluvial material that was deposited in the
downstream end of reach 2 during the 1968 high pool
level. During 1970—72 the stream channel in reach 2
was returning to its pre-1967 size. In October 1972, the
bed level of the stream channel was about the same as it
was in 1967; however, the channel was smaller and the
flood plain was higher.

The primary natural quality of the study reach that
affects our evaluation is the changeable character of the
channel boundary; the boundary changes with the
stresses applied. A major flood enlarges and straightens
the stream channel; the resistance to the movement of a
subsequent flood is then decreased (Burkham, 1970,
1975). Conversely, in the absence of major floods, the
size of the stream channel decreases and the resistance
to the movement of a subsequent flood increases. In
order to evaluate the effects of changes in bottom-land
vegetation on the three major floods, problems arising
from the natural modifications of the parameters being
studied had to be resolved. Discussion of changes occur-
ring between floods follows in the section “Data of Hy-
draulic Parameters;” changes occuring during floods
are described in the section “Discussion of Results.”

BASIC DATA

The hydraulic data used in this study are peak dis-
charges for the floods of 1965, 1967, and 1972; profiles of
the Gila River at cross sections along the study reach;
distances between the sections along the main path of
the floods; and mean cross-sectional velocity and mean
cross-sectional depth at peak stages at the sections (ta-
ble 1). The peak discharges for the floods of December
1965 and August 1967 were measured at the bridge on
U. S. Highway 70 near Bylas by personnel of the US.
Geological Survey (1965; 1968). The peak discharge for
the flood of October 1972 was based on an extension of
the stage-discharge relation for the Bylas gage and on a
measurement of peak discharge at a site about 50 mi (80
km) upstream. Peak stages were marked along the
south bank of the study at nine cross sections during the
floods of December 1965 and August 1967. The peak
stage for the October 197 2 flood was marked at the nine
sections within a few hours of the peak discharge. The
altitudes of the marked gages were carefully surveyed
immediately after the floods. The nine cross sections
had been surveyed and permanent horizontal and verti-
cal controls established in June 1964. The cross sec-

 

J3

tions, except section 17, were resurveyed in June 1966
and again in June 1968, except sections 13, 15, and 17,
which were resurveyed in March 1970. The nine cross
sections were surveyed again in December 1972. The
surveys of section 1 in 1966, 1968, and 1972 extended
only to the top of the dike protecting the cultivated land.
The profiles of the nine sections and the maximum stage
at the sections for the three floods are shown in plate 2.

The stream channel and flood plain of the study reach
change very slowly in the absence of major floods
(Burkham, 1972) and surveys of the cross sections im-
mediately before each of the floods were not required for
this study. No significant change in the altitude of the
flood plain was possible during the period from the June
1964 survey to the start of the flood in December 1965
because the overbank rates and amounts of flow were
small (table 2). Likewise, no significant changes in the
altitude of the flood plain were possible during the
periods between the June 1966 survey and the August
1967 flood, and between the 1968—70 surveys and the
October 1972 flood. The discharge at bankfull stage
probably was between 3,000 and 6,000 ft3/s (85 and 170
111%) from 1964 to present (1973).

Data from streamflow measurement made at or near
the nine cross sections indicate that changes in the size
of the stream channel were insignificant during the
period from the June 1964 survey to the start of the
December 1965 flood, and during the period from June
1966 survey to the start of the August 1967 flood. Except
for the changes discussed earlier in the section “Unde-
sirable Characteristics of the Site,” the size of the
stream channel probably did not change significantly
between the June 1968 survey and the start of the 1972

flood.
ANALYSES OF DATA

Analyses were made to determine (1) mean velocities
and mean depths; (2) channel-roughness coefficients; (3)
average changes in the altitude of the bottom land; and
(4) changes in the study parameters resulting from veg-
etation alteration. The basic assumptions and criteria
for these different analyses are: (1) the peak discharge
did not change significantly as a flood moved down—
stream; (2) the water surface at each cross section was
horizontal; (3) the altitude of the riverbed did not
change significantly between the time of the beginning
of a flood and the time of the peak stage, except for reach
2 downstream from cross section 15 during the 1972
flood; (4) the cross-sectional profile at section 17 defined
by the 1964 survey was used in the studies of hydraulic
characteristics for both the 1965 flood and the 1967
flood—cross section 17 was not surveyed after the 1965
flood or after a flood occurring in January 1966 (table 1) ;
(5) the cross-sectional profiles at sections 15 and 17
defined by the 1972 survey were used in determining

 

J4

the hydraulic parameters for the 1972 flood; (6) the total
flow for the three peak discharges is assumed to have
passed south of the dike at cross section 1, and (7) any
differences in stage, mean velocity, mean depth, and
roughness coefficient resulting from differences in peak
discharge for the 1965 and 1967 floods are insignificant.
Further discussions of assumptions are presented with
descriptions of the individual analyses.

GILA RIVER PHREATOPHYTE PROJECT

MEAN VELOCITIES AND MEAN DEPTHS

The mean velocity in a cross section was determined
by dividing the peak discharge rate by the cross-
sectional area at the peak stage; the mean depth was
determined by dividing the cross-sectional area by the
top width of flow at peak stage (table 1). For the 1965
and 1967 floods, the mean velocities and mean depths at
the US. Highway 70 bridge were obtained from

TABLE 1.—Hydraulic parameters for peak discharges, floods of December 1965, August 1967, and October 1972, Gila
River
[Peak discharges for the floods were 39,000, 40,000, and 80,000 fta/s]

 

 

 

 

 

 

 

 

Cr 055- Length Altitude of Cross- Top Mean‘cross- Hydraulic Mean‘cross- Roughness coefficient
section of water sectional Width sectional radius sectional
N o reach surface area (ft) depth (ft) veloc1ty
' (ft) (ft) (ftfi) (ft) (ft/s) "1 "2 "g na "
December 1965
1 2,571.8 15,600 3,380 4.6 4.6 2.5 0.057
4,850 0.00418 0.064
3 2,565.3 17,800 3,780 4.7 4.7 2.2 .073
8,000 .00742 .086
5 2,553.8 19,100 2,620 7.3 7.2 2.0 0.082 .102
8,000 .00630 .080
7 2,539.7 10,800 1,860 5.8 5.8 3.6 .062
8,000 .00223 .047
9 2,527.8 9,600 1,760 5.5 5.4 4.1 .036
6,800 .00292 .054
11 2,518.6 16,300 3,060 5.3 5.3 2.4 .081
13,700 .00615 .078
13 2,501.8 17,800 2,280 7.8 7.7 2.2 .074 .076
6,800 .00507 .071
15 2,490.9 9,410 1,520 6.2 6.2 4.1 .067
6,000 .00402 .064
17 2,482.0 14,300 1,680 8.5 8.4 2.7 .060
August 1967 1
l
1 2,572.5 12,300 2,580 4.8 4.8 3.2 J3 0.053
.00 15 .056
3 2,565.8 17,400 3,860 4.5 4.5 2.3 .059
.00600 .077
5 2,552.6 16,200 2,550 6.4 6.3 2.5 .064 .101
.00313 .056
7 2,538.4 8,540 1,880 4.5 4.5 4.7 .031
.00157 .040
9 2,528.9 11,600 1,760 6.6 6.6 3.4 .050
.00475 .069
11 2,519.3 18,200 3,110 5.9 5.8 2.2 .094
.00771 .088
13 2,502.1 18,700 2,330 8.0 8.0 2.1 .082 .082
.00580 .076
15 2,491.1 10,200 1,560 6.6 6.5 3.9 .071
.00468 .068
17 2,483.0 15,900 1,700 9.4 9.3 2.5 .066
October 1972
1 2,574.8 18,700 3,420 5.5 5.4 4.3 .043
.00124 .035
3 2,566.2 14,700 3,880 3.8 3.8 5.5 .029
.00114 .034
5 2,553.4 16,900 2,670 6.3 6.3 4.7 .031 .040
.00088 .030
7 2,538.8 8,980 1,890 4.8 4.7 8.9 .022
.00042 .021
9 2,529.8 11,800 1,750 6.7 6.7 6.8 .019
.00094 .031
11 2,518.2 13,600 2,980 4.6 4.6 5.9 .049
.00077 .028
13 2,500.7 13,200 1,950 6.8 6.7 6.1 .028 .016
.00080 .028
15 2,492.2 11,200 1,640 6.8 6.8 7.1 ____
.00102 .032
17 2,484.0 13,950 1,860 7.5 7.4 5.7 _--_

 

EFFECTS OF BOTTOM-LAND VEGETATION 0N MAJOR FLOODS, GILA RIVER, ARIZONA

current-meter measurements taken during the floods.
ROUGHNESS COEFFICIENTS
The Manning velocity equation was used as the basis

for computing the roughness coefficients given in this
report. The Manning equation for English units is

V = 1.41186 R%Sl/2, (1)
in which
V = mean velocity of flow in a cross section, in feet

per second;

R = hydraulic radius at a cross section, in feet
(equal to cross-sectional area of flow, in
square feet, divided by wetted perimeter, in
feet);

S = energy gradient; and

n = a roughness coefficient.

The Manning equation for International System Units
is

V = %R%SV2,

in which metres are the units of length for V andR, and
S and n are as previously defined.

The Manning equation was developed for uniform
flow in which the water-surface profile and energy gra-
dient are parallel to the streambed, and the area, hy-
draulic radius, and depth remain constant throughout
the reach. The equation is considered valid for
nonuniform reaches, such as that of the Gila River, if
the energy gradient, or friction slope, is modified to
reflect only the losses due to boundary friction (Barnes,
1967, p. 4). The energy equation for a reach of
nonuniform channel, in which energy is expressed as
head in feet of water, is

(h + hv)1 = (h + hv)2 + (117012 + k(Ahv)1,2, (2)
where subscripts 1 and 2 refer to cross sections at the
ends of the reach, and

h = water-surface elevation at a cross section,
in feet;

hv = velocity heac21 at a cross section, in feet

(equals a 2;, where a is a velocity head

adjustment factor and g is acceleration
due to gravity, in feet per second per sec—
ondh

hf = head loss due to boundary friction in a
reach, in feet;

k(Ahv)= head loss due to acceleration or decelera-
tion of streamflow in a contracting or ex-
panding reach, in feet;

 

J5

TABLE 2,—Peak discharge, Gila River at Calva, Ariz., 1963—72
[Peak discharge above base of 3,000 {ta/s]

 

Water year
(October 1 to
September 30)

Peak discharge
Date (ft3/s)

 

September 26, 1964
August 14, 1965
September 4, 1965
December 13, 1965
December 24, 1965
January 1, 1966
March 19, 1966
August 6, 1967
August 13, 1967
December 21, 1967
January 30, 1968
February 16, 1968
February 24, 1968
March 1, 1968
March 12, 1968
No peak above base
_. No peak above base
- August 22, 1971
October 28, 1971
August 28, 1972
September 10, 1972
October 8, 1972
October 20, 1972

3,060

 

 

(Ahv) = difference in velocity heads between sec-
tions, in feet; and
k = energy loss coefficient.

The velocity head adjustment factor, a, which is the
ratio of true velocity head to the velocity head computed
on the basis of mean velocity, was not determined for
this study. The value of a was assumed to be 1.00 at all
sections for the two floods. This assumption probably
introduced bias into the computation of 71.; however, the
bias may have been small because most of the flow was
on the flood plain where the velocity across a section
probably was fairly uniform. Furthermore, a value for
the difference in roughness coefficient as a result of
vegetation changes is a primary objective of this study,
and any bias introduced by assuming a = 1 is largely
eliminated when a difference in roughness coefficient is
computed.

The friction slope S used in the Manning equation is
defined

(hf)1.2 (Ah)1.2+(Ahu)1.2 — k(Ahv)1.2
s = = , (3)
L12 L12

where L12 is the length of the reach between two sec-
tions and hfL2 is the head loss due to boundary friction
between the two sections. The energy-loss coefficient k
is taken to be zero for contracting reaches and 0.5 for
expanding reaches. In this study, the quantities Ah U 1.2
and (kAhvh,2 are small compared to Ah 1.2 because of
relatively steep channel slopes, long reaches between
sections and no major channel contractions or expan-
sions.

When the Manning equation is used_ to determine
discharge, the quantity (1.486/n)AR%, called con-
veyance and designated K, is computed for each cross
section. In computing K, the roughness coefficient n is
assigned to the section even though it is an average
value for a reach extending upstream and downstream

 

J6

from the section. For brevity, n is referred to in this
report as the roughness coefficient for a section. In the
discharge computation, the mean conveyance in the
reach between any two sections is computed as the
geometric mean of the conveyance of the two sections
(Barnes, 1967, p. 6). The discharge equation for a two-
section reach in terms of conveyance is

Q = (K1K2S)‘/2 (4)
where Q is the discharge and S is the friction slope as
previously described.

An equatlon for the product n1n2 is obtained by com-
bining equations (3) and (4) and reversing the computa-
tion described in the preceding paragraph. The equation
for English units is

2.21

 

n1n2 = [(Rle)% A1A2][(h + hv)1

1.2

— (h + h,,)2 — (kAhv)1.2]' (5)
The product n1n2 and the geometric mean of the rough—
ness coefficient, ng(= (n1n2)1/2), were computed for the
three peak discharges for each stream length between
cross sections using equation (5) and the discharge,
water-surface profile, and the hydraulic properties pre-
viously determined for the cross sections (table 1). The
data of n1n2 were used to determine the value of n for
each of the nine cross sections for the 1965 and 1967
floods and to determine the value of n for sections 1, 3, 5,
7, 9, 11, and 13 for the 1972 flood.

Average values of the Manning roughness coefficients
for the three floods for the part of reach 1 from cross section
3 to cross section 7 and for the part of reach 2 from cross
section 11 to cross section 15 were computed using the
equation for English units that follows (table 1):

1.486 /(h+h,,)1 — (h+hv)M —(kAhU)1.2

Q\

 

”’11

L12
A1A2

L23
+
Z2Z3

+(kAhU)2.3+ + (kAhv) (M_1).M>1/2 (6)

 

 

 

L‘(M—1) - M
+

Z (M—1)ZM

where Z = ARZ/3 and other quantities are as previously
defined (Barnes, 1967, p. 6). The equation is applicable
to a multisection reach of M cross sections, which are
designated 1, 2, 3, . . . M— 1, M. For a two—section reach,
the value of na in equation (6) is the same as the value of
ng in equation (5).

 

GILA RIVER PHREATOPHYTE PROJECT

A procedure for determining the Manning roughness
coefficient n for a cross section was required for this
report. Values of n were sought so that: (1) they could be
compared directly with the hydraulic parameters mea-
sured at the different cross sections; (2) the effects of
changes in the vegetation on the hydraulic parameters
for the different cross sections could be studied; and (3)
the variability of the roughness coefficient and the
reasons for this variability could better be described.

The Manning roughness coefficient n for each cross
section can be computed from product values obtained
using equation (5) if the value of n is known for at least
one section; however, a large bias may be introduced by
assuming a value of n for one cross section and then
computing values for the remaining sections based on
this value. For example, if an assumed value of n is too
small for cross section 17, the computed value of n for
cross section 15 will be too large, the computed value of
n for 13 will be too small, and the errors will continue to
increase in magnitude as values of n are computed
further. A value of n17 was sought so that this bias
might be minimized.

The bias was minimized by using an equation for
variance, the product value computed using equation
(5) and the procedure discussed in this paragraph. The
equation for variance is

N N 2
E (ni)2_( Z ni)
i 1

 

i = 1 =
,2 = N , (7)
N - 1
in which
8 2 = variance of n for a sample;
N = number of observations of n in the sample; and
n,- = Manning roughness coefficient at a cross sec-

tion.
The following procedure was used in determining n for
the 1965 and 1967 floods at the nine cross sections:
(1) values of n were computed for all sections in terms
of n17 using the product values obtained from
equation (5);
(2) the variance of n was computed by using all val-
ues obtained in step (1) except the value for
cross section 17;
(3) the first derivative of the equation obtained in
step (2) was set equal to zero;
(4) the equation obtained in step (3) was solved for
n17 ; and
(5) values of n were computed for the remaining sec-
tions by using the n17 value obtained in step (4)
and the product values obtained using equa—
tion (5).

EFFECTS OF BOTTOM-LAND VEGETATION ON MAJOR FLOODS, GILA RIVER, ARIZONA

In brief, the procedure is based on the theory that the
variance of the sample composed of values of n for all
cross sections except section 17 is not a function of the
value of n17. Step (3) in the procedure says that the
change of variance resulting from a change in n17 is
zero. Using the same procedure, values of n were deter-
mined for the 1972 flood at all cross sections except for
sections 15 and 17; values were not determined for sec-
tions 15 and 17 because of the uncertainty of when
erosion occurred. The data of n are presented in table 1
and figure 2.

CHANGES IN ALTITUDE OF BOTTOM LAND

The average change in the altitude of the bottom land
for the period June 1964 to June 1968 for cross sections
1, 3, 5, 7, and 9 and the procedure used in determining
the change are described in another report (Burkham,
1972). Theprocedure consists of (1) plotting the mea-
sured profiles for each cross section, (2) obtaining the
vertical area between plotted profiles from the graph,
and (3) dividing the vertical area by the horizontal
length of the cross section. A positive change in altitude
indicates that a larger area of fill than of scour occurred
in the section.

EFFECTS OF CHANGES IN VEGETATION

The part of reach 1 from sections 3 to 7 and the part of
reach 2 from sections 11 to 15 were considered the best
pair of reach parts for the application of hydraulic prin-
ciples and equations. The data of hydraulic characteris-
tics for these two lengths of the study reach, therefore,
were given the most emphasis in evaluating effects of
vegetation changes. The reasons for downgrading the
data for parts of the study reach near sections 1, 9, and
17 are described in the section “Undesirable Charac-
teristics of the Site.”

Among the three peak discharges, differences in the
study parameters for sections 3 to 7 and for sections 11
to 15 are assumed to have been caused mainly by (1)
eradication of bottom-land vegetation; (2) changes in
foliage on bottom-land vegetation; (3) channel changes
caused by a previous flood; and (4) differences in peak
discharge. Events 1 and 3 are expected to cause de-
creases in stage, depth, and roughness coefficient, and
increases in velocity; events 2 and 4 are expected to
cause increases in stage, depth, and roughness
coefficient, and decreases in velocity (Chow, 1959;
Burkham, 1972, 1975). The difference in the study
parameters for the floods of 1965 and 1967 in reach 1
presumably were caused by events 1 and 3; the
difference in reach 2 was caused by events 2 and 3. The
difference in the study parameters for the floods of 1967
and 197 2 in reach 1 presumably was caused by events 3
and 4; the difference in reach 2 was caused by events 1,
3, and 4. For the 1965 and 1967 floods, the difference in

 

J7

the study parameters in reach 2 caused by event 1 is for
vegetation fully foliaged.

The method of determining the effects of vegetation
removal on the study parameters is based on an as-
sumption that the four events, described in the preced-
ing paragraph, caused independent effects. The method
is illustrated by the following equations for reaches 1
and 2, respectively:

 

(DH7 — (DH5 =(1)AHE + (DAHC (8)

 

(2)117 - (2)H5 = (2)AHF + (2)AHC (9)
in which

(1>H7 *(1)H5

average of differences in stage for the
1965 and 1967 floods at cross sec-
tions 3, 5, and 7 in reach 1, in feet;
’(1)H5 indicates peak stage at a cross
section in reach 1 for the 1965 flood
and (1)H7 indicates peak stage for
the same cross section for the 1967
flood;

 

average difference in stage for the
1965 and 1967 floods due to the
eradication of dormant vegetation
in reach 1, in feet;

 

average difference in stage for the
1965 and 1967 floods due to channel
changes caused by the 1965 flood in
reach 1, in feet;

average of differences in stage for the
1965 and 1967 floods at cross sec-
tions 11, 13, and 15 in reach 2, in
feet; (2)H5 indicates peak stage at a
crosssection in reach 2 for the 1965
flood and (2)H7 indicates peak stage
for the same cross section for the
1967 flood;

<2)H 7 ‘(2)H5

 

 

(2)AHF = average difference in stage for the
1965 and 1967 floods due to in-
creased foliage in reach 2, in feet; it
is the expected change in stage in
reach 1 due to increased foliage if
the vegetation had not been re-

moved; and

 

average difference in stage for the
1965 and 1967 floods due to channel
changes caused by the 1965 flood in
reach 2, in feet; (2)AHC is assumed to
equal (DAHC.

J8

MEAN CROSS-SECTIONAL VELOCITY,

MANNING'S I!

IN FEET PER SECOND

GILA RIVER PHREATOPHYTE PROJECT

CROSS SECTIONS

17 15 13 11 9 7 5 3 1

V V V V V V V V V

DISTANCE ALONG MAIN FLOW PATH OF FLOOD, IN KILOMETRES

 

 

 

 

 

 

 

 

 

0 ¥ 4 6 8l 10 12 14 16 18 20
10 | I I I I I I I I I I I I II I I I
A 1 Velocity obtained from current-meter measurement —
made at highway bridge on U.S. 70.
1 I I I I I I I I I I I
1 2 3 4 5 6 7 8 9 1O 11 12
2 4 6 8 10 12 14 16 18 20
0-12 I I I I I I I II I I I I I I I II I I I
0.10 — _X /,/T\\\ —
// / \\ // \{\\
/ f _____ /
0.08 — X ,x’ X \ I / \ _
/ ,,¢’ \ ,l / \ \x.
/7:”” \X \\ ’I / \\ \‘
1” , \
0.06 r \ \ x / X\ s '
0.04 ~ —
0.02 — _
B
o I I I I I I I l I I I I
O 1 2 3 4 5 6 7 8 9 10 11 12
DISTANCE ALONG MAIN FLOW PATH OF FLOOD, IN MILES
EXPLANATION
5
V Location of cross section and number
--——o———— Flood of December 1965, discharge 39,000 ft3/s (1,100 m3/s)
—-—x—— Flood of August 1967, discharge 40,000 fta/s (1,130 ma/s)

FIGURE 2,—Hydraulic characteristics at peak discharge for the floods of December 1965, August 1967, and
October 1972. The lines on the graphs are based on the plotted points and on the hydraulic properties of
the bottom land between the cross sections. The roughness coefficient n is treated as if it applied to a
section even though it is an average value for a reach extending upstream and downstream from the
section. Distance along the main flow path of flood is scaled from the map shown on plate 1.

——A—-— Flood of October 1972, discharge 80,000 fts/s (2,270 ma/s)

3.0

2.5

2.0

1.5

1.0

0.5

MEAN CROSS-SECTIONAL VELOCITY,
IN METRES PER SECOND

EFFECTS OF BOTTOM-LAND VEGETATION ON MAJOR FLOODS, GILA RIVER, ARIZONA

AVERAGE CHANGE IN ALTITUDE OF
BOTTOM LAND FOR INDICATED

MEAN CROSS-SECTIONAL
DEPTH, IN FEET

IN FEET

PERIODS,

CROSS SECTIONS

 

   

 

 

 

 

 

17 15 13 11 9 7 5 3 1
V V V V V V V V V
DISTANCE ALONG MAIN FLOW PATH OF FLOOD, IN KILOMETRES
O 2 4 6 8 1O 12 14 16 18 20
10 1' 1'1 '1 I1 1' 1' 1 '1 '1 1' 1_3.0
9 I _
—2.5
9' ‘I
I
l
I
I
: _
I —2.0
I
I
11 —
1
J -- 1.5
‘,
C 'Depths obtained from current-meter measurement
made at highway bridge on U.S. 70. —1_0
3 1 1 1 1 1 1 1 1 1 1 1 1
0 1 2 3 4 5 6 7 8 9 10 11 12
2 4 6 8 10 12 14 16 18 20
1 ' 1 ' 1 ' 1 '1 1 ' 1 ' I ' I '1 1' 1 2'0
3 — _
2 — E9 "\ 69 ~
\ / 7K — 1.5
\\ 41/
/—\\
1_ V/ \\\\ e~ //‘\ -
_ r'< \ //,' \
\ — ‘i / \
L \ / /
/ /<\___$_/\ \ ’9 / , 1
__________ \ \‘QJI/ / (x 1
o — /" "t ______ _________\. 4/ ,. ~—0
5‘ , \ 'XLLJ 1‘
‘\‘ur—/ 1
1_ /// k _
i — 1.5
2 - / _
3 — _
/ —2.0
I D
4 1 1 1 1 1 1 1 1 1 1 1 1
0 1 2 3 4 5 6 7 8 9 10 11 12

DISTANCE ALONG MAIN FLOW PATH OF FLOOD, IN MILES

EXPLANATION —Continued

AVERAGE CHANGE IN ALTITUDE OF
BOTTOM LAND FOR THE PERIOD

_____ A_———— June 1964 to June 1966
__.$.__ June 1966 to June 1968
__+__. June 1968 to December 1972

_____ June 1964 to December 1972

FIGURE 2.—Continued.

MEAN CROSS-SECTIONAL
DEPTH, IN METRES

AVERAGE CHANGE IN ALTITUDE OF
BOTTOM LAND FOR INDICATED

IN METRES

PERIODS,

J9

J10

Equation (10) is obtained by subtracting equation (9)
from equation (8):

 

((1yH7 ‘ (1)H5)"((2>H7 4 (2)115)

 

=((1)AHE ‘ (2)AHF)+((1)AHC — (2)AHcl (10)

A decrease or minus (—) change in stage is expected
for (DAHE and an increase or positive (+) change in
stage is expected for (2)AHF. The sum, (DAHC — (2)AHC ,
is assumed to be zero. When the expected criteria and
assumptions are applied to equation (10), the desired
equation that shows the difference in stage resulting
from the removal of fully foliaged vegetation is ob-
tained. The equation is:

 

 

((2)177 -<2>Hs)—(<1>H7 -<1>H5)= ‘((2)AHF+(1)AHE) (11)

A numerical value for the left side of equation (10) is
obtained by using the stage data given in table 2. The
differences in stage for the 1965 and 1967 floods in reach
1 are 0.5 ft (0.15 m) at section 3, —1.2 ft (—0.37 m) at
section 5, and — 1.3 ft (—0.40 m) at section 7; the average
of these differences is — 0.7 ft (—0.21 m). The differences
in stage for the 1965 and 1967 floods in reach 2 are 0.7 ft
(0.21 In) at section 11, 0.3 ft (0.09 In) at section 13, and
0.2 ft (0.06 m) at section 15; the average of these differ-
ences is 0.4 ft (0.12 m). A value of —1.1 ft (—0.34 m) is
obtained as an estimate of the average decrease in stage
during the 1967 flood in reach 1 resulting from the
removal of vegetation; this value was obtained by using
0.4 for (ng7 — (2)115 and —0.7 for (1)H7 — (”[15 in equation
(11).

The hydraulic parameters for the 1965 flood were
used as standards in determining percentage effects of
changes in vegetation. The vegetation was in place in
both reaches during the December 1965 flood.

Differences in the study parameters in reach 1 from
sections 3 to 7 and in reach 2 from sections 11 to 15
caused by removal of fully foliaged vegetation were
computed using equations similar to equation (11). Bas-
ically, for the 1965 and 1967 floods, the equations used
say that the effects of the removal of fully foliaged vege-
tation on the hydraulic. parameters is the average dif-
ference in the parameter for the two floods in reach 2.
For the 1967 and 1972 floods, the equations say that the
effects of removal of fully foliaged vegetation on a
parameter is the average difference in the parameter for
the two floods in reach 2 minus the average difference in
the parameter for the two floods in reach 1.

 

GILA RIVER PHREATOPHYTE PROJECT

The effects of increased foliage between December
1965 and August 1967 on the study parameters in reach
2 were not computed directly; however, the probable
effects are discussed briefly in the following section.

DISCUSSION OF RESULTS

The effects of vegetation changes between the 1965
and 1967 floods and between the 1967 and 1972 floods
are discussed separately in this section. Reasons for
variation of the different parameters along the study
reach also are presented in a separate discussion.

HYDRAULIC EFFECTS FOR THE 1965 AND 1967 FLOODS

The seasonal increase in foliage between the 1965 and
1967 floods and channel changes resulting from the
1965 flood apparently caused significant changes in the
hydraulic parameters throughout reach 2; the removal
of dormant vegetation and channel changes apparently
caused significant changes in the hydraulic parameters
throughout reach 1 (table 1; fig. 2). For the 1967 floods at
sections 11, 13, and 15 in reach 2 the stage and mean
cross-sectional depth were an average 0.4 ft (0.12 In)
higher and the mean cross-sectional velocity was an
average 0.2 ft/s (0.06 m/s) lower than the corresponding
parameters for the 1965 flood. The Manning roughness
coefficient na in the part of the study reach from sections
11 to 15 was 0.008 higher for the 1967 peak than for the
1965 peak. The magnitude of the differences in the
study parameters caused by the increase in foliage and
channel changes is questionable for cross sections 9 and
17 because of hydraulic conditions at cross section 9
(discussed on p. J3) and because of possible poor data for
cross section 17 (discussed on p. J 3). Other questions
develop from the results for section 9 because, due to the
removal of vegetation in reach 1, there was a transition
during the 1967 flood from high velocity and kinetic
energy in reach 1 to low velocity and kinetic energy in
reach 2; this transition took place near section 9. The
writer believes, however, that these differences are sig—
nificant because they are in the same direction as those
for cross sections 11, 13, and 15 (fig. 2; table 1)—that is,
the mean velocities decreased and the stage, mean
depth, and roughness coefficients increased.

Average differences in the hydraulic parameters for
the 1965 and 1967 floods at sections 3, 5, and 7 in reach 1
are as follows:

1. The stage was 0.7 ft (0.21 m) lower in 1967 than in

1965;

2. The mean cross-sectional depth was 0.8 ft (0.24 m)

lower in 1967 than in 1965; and

3. The mean cross-sectional velocity was 0.6 ft/s (0.18

m/s) greater in 1967 than in 1965.
The Manning roughness coefficient na was 0.018 less in
1967 than in 1965 in the reach from sections 3 to 7. The

EFFECTS OF BOTTOM-LAND VEGETATION ON MAJOR FLOODS, GILA RIVER, ARIZONA

magnitude of the differences in the study parameters
caused by the removal of dormant vegetation and chan-
nel changes is questionable for section 1 because of the
adverse hydraulic conditions (p. J2). Results also are
questionable for section 9 because of the transition dur-
ing the 1967 flood from high velocity and kinetic energy
in reach 1 to low velocity and kinetic energy in reach 2,
discussed above.

The effects of removing fully foliaged vegetation on
the study parameters for the 1965 and 1967 floods were
assumed to be the sum of the effects of increased foliage
on the parameters in reach 2 plus the effects of the
removal of dormant vegetation in reach 1. Based on this
premise, the removal of fully foliaged vegetation in the
study reach apparently caused a decrease of 1.1 ft (0.34
m) in stage, a decrease of 1.2 ft (0.37 m) in mean cross—
sectional depth, a decrease of 0.026 in Manning rough-
ness coefficient na, and an increase of 0.8 ft/s (0.24 m/s)
in cross-sectional velocity. Relative to the 1965 flood,
the decrease in depth was 19 percent, the decrease in
roughness coefficient 33 percent, and the increase in
velocity 29 percent. The method of obtaining a sum for
the two effects and the method of removing the effects of
channel changes from the data are presented on page
J 10.

Effects of channel changes caused by the 1965 flood on
the study parameters for the 1967 flood could not be
determined directly; the effects of removing vegetation
between floods on the channel changes which occurred
during the 1967 flood also could not be determined. In
the different cross sections surveyed, the 1965 flood
apparently caused both scour and fill in parts of the
section (pl. 1). From June 1965 to'June 1966, however,
there were larger areas of scour than fill at all the
sections surveyed except sections 3 and 13 (pl. 2). Most
of this scour probably occurred during the recession of
the December 1965 flood. The relatively large scour in
cross section 1 and the fill in cross section 3 are assumed
to have been caused indirectly by the bridge on US.
Highway 70 (pl. 1). The flood of December 1965 was the
first major flood after the construction of the bridge in
1957. The bridge apparently restricted the flow along
the left side of the flood plain causing a higher-than-
normal velocity. Scour was a direct result of the high
velocity. The large scoured area of cross section 1 from
station 2400 to station 2920 indicated by the 1966 sur-
vey (pl. 1), however, did not extend as a continuous
channel from cross section 1 to cross section 3. Ap-
parently, most of the sediment scoured from the flood
plain from the highway bridge downstream past cross
section 1 was deposited in a reach which included cross
section 3. The reason for the fill at cross section 13 is not
known.

Most of the changes in the study reach from sections 3

 

J11

to 7 and from sections 11 to 15 caused by the 1965 flood
occurred along the stream channel at bends and re-
stricted sections. Most of the peak flow from the floods
investigated in this study were contained within the
flood plain and changes in the stream channel may not
have greatly affected the study parameters for the 1967
flood. At flow rates less than about 20,000 ft 3/s (570
m3/s) the effects of the stream-channel changes proba-
bly would have been more significant.

Fill was observed at all the cross sections for the
period June 1966 to June 1968 (pl. 2). The large amount
of fill in the downstream end of the study reach after the
August 1967 flood undoubtedly was caused by a high
lake level in the San Carlos Reservoir reached during
the recession of the August 1967 flood. The large
amount of fill at cross section 9 may have been caused by
the screening effects of the saltcedar and mesquite as
the floodwater entered the uncleared part of the study
reach. A logical explanation for the large amounts of fill
in cross sections 1 and 3 is not apparent.

The sediment loads carried by the two floods may
have been a significant factor in explaining why scour
occurred during the December 1965 flood and fill
occurred during the August 1967 flood (Burkham,
1972.) Studies based on the meager data available prior
to 1905 (US. Army Corps ofEngineers, 1914, p. 30) and
on data for 1965—70 (US. Geological Survey, 1965—
1971) indicate that the sediment concentration for a
given flow rate in the winter (November through April)
in the Gila River at the head of Safford Valley is less
than 20 percent of the average concentration for the
same flow rate in the summer (July through October).
Most of the winter flow originates in mountainous
terrain where there is relatively little transportable
material. Large flows having relatively low sediment
yields are conducive to erosion, While large flows of
relatively large sediment loads are conducive to
deposition if other hydraulic conditions are favorable.

HYDRAULIC EFFECTS FOR THE 1967 AND 1972 FLOODS

The vegetation removal, the unequal peak dis-
charges, and channel changes apparently caused sig-
nificant differences in the hydraulic parameters for the
two floods throughout reach 2 (pl. 2). For the 1972 flood
at sections 11, 13, and 15 in reach 2, the stage was an
average 0.5 ft (0.15 m) lower, the mean cross-sectional
velocity an average 3.7 ft/s (1.13 m/s) higher, and the
mean cross-sectional depth an average 0.7 ft (0.21 m)
lower than the corresponding parameters for the 1967
flood. The Manning roughness coefficient na, in the part
of the study reach from sections 11 to 15 was 0.054 lower
during the 1972 flood than during the 1967 flood. The
lower average stage and depth of the 1972 flood at sec-
tions 11, 13, and 15 is of particular importance. This

J12

indicates that the combined effects of vegetation re-
moval and channel changes—effects which tend to de-
crease stage and depth—are greater than the effects of
doubling the peak discharge from 40,000 to 80,000 ft3¥s
(1,130 to 2,270 m3/s).

The stage and mean cross—sectional depth at sections
9 and 15 were higher for the 1972 flood than for the 1967
flood; the reasons for the relatively high stage and depth
at these sections are not known. The relative high stage
and depth at section 9, however, probably was caused by
confinement of the 1972 flood by the embankment at the
railroad bridge. The relatively large depth at section 15
may not be real; it may have been a computational error
if erosion of the flood plain occurred after the 1972 peak
discharge rather than before as was assumed. The rela-
tively high stage at cross section 15‘also could be ac-
counted for if most of the erosion that was measured
occurred after the peaks instead of before.

The relatively large difference in average mean
cross-sectional velocity and Manning roughness
coefficient na for the 1967 and 1972 floods in reach 2
probably results from unequal peak discharges.

Average differences in the hydraulic parameters for
the 1967 and 1972 floods at sections 3, 5, and 7 in reach 1
are as follows:

1. The stage was 0.5 ft (0.15 In) higher in 1972 than in

1967;

2. The cross-sectional velocity was 3.2 ft/s (0.98 m/s)

greater in 1972 than in 1967; and

3. The cross-sectional depth was 0.1 ft (0.03 m) lower

in 1972 than in 1967.
The Manning roughness coefficient na in the part of the
study reach from sections 3 to 7 was 0.033 lower in 197 2
than in 1967. For reasons discussed on page J3, the
magnitude of the difference in the study parameters
caused by unequal discharges and channel changes is
questionable for sections 1 and 9. 7

An apparent inconsistency exists between average
difference in stage and average difference in mean
cross-sectional depth for the 1967 and 1972 floods at
sections 3, 5, and 7; the stage increased an average 0.5 ft
(0.15 In) and the closely related average depth decreased
an average 0.1 ft (0.03 m). This inconsistency may indi-
cate that the scour at the three sections, which occurred
between June 1968 and December 1972, largely oc-
curred before the peak of the 197 2 flood instead of
afterwards as was assumed.

Based on data for the 1967 and 1972 floods the re-
moval of fully foliaged vegetation in the study reach
apparently caused a decrease of about 1.0 ft (0.30 m) in
stage, a decrease of about 0.6 ft (0.18 m) in mean cross-
sectional depth, a decrease of about 0.021 in Manning
roughness coefficient na, and an increase of about 0.8 ft/s
(0.24 m/s) in cross-sectional velocity for the 1972 flood.

 

GILA RIVER PHREATOPHYTE PROJECT

Relative to the 1965 flood, the decrease in depth was 10
percent, the decrease in roughness coefficient 27 per-
cent, and the increase in velocity 18 percent. The
method of removing the effects of unequal discharges
and channel changes from the data is presented on page
J10.

The effects of unequal discharges and the effects of
channel changes on the study parameters for the 1967
and 1972 flood could not be determined independently.
However, the combined effects of the two factors proba-
bly amounted to a 3.2 ft/s (0.98 m/s) increase in velocity
and a 0.033 decrease in the Manning roughness
coefficient na—the differences in the respective
parameters for the part of reach 1 from sections 3 t0 7.

The effects of vegetation alteration between the 1967
and 197 2 floods on channel changes in the study reach
during the 1972 flood could not be determined. Scour is
indicated for the period June 1968 to December 1972 at
all the sections except 1, 11, and 13 where fill is indi-
cated. Probable reasons for the scour at sections 15 and
1 7 have previously been discussed (p. J3, J10, J 12). The
fill at sections 11 and 13 may have been an adjustment
in the channel resulting from the high pool sedimenta-
tion at sections 15 and 17. The fill at section 1 may be an
adjustment in the channel affected by the bridge on US.
Highway 70.

RANGE IN HYDRAULIC PARAMETERS
ALONG THE STUDY REACH

The range in the different hydraulic parameters
along the study reach for the three floods was larger
than expected (pl. 2). For the 1965 flood at sections not
affected by the bridge and the reservoir the mean cross-
sectional velocity ranged from 2.0 ft/s (0.61 m/s) at sec-
tion 5 to 4.1 ft/s (1.25 m/s) at section 15 (table 1), a
difference of about 100 percent of the lower figure. The
reason for the large range is not known; however, differ-
ences in density of ve getation along the study reach may
have been a minor contributing factor; bending or lack
of bending of the saltcedar and mesquite in portions of
the study reach also may have been a minor factor. The
flow at sites where high velocity prevailed may have
been strong enough to bend the trees, resulting in a
reduction in channel friction and an increase in veloci-
ty, whereas the flow at sites Where low velocities pre-
vailed may not have bent the trees.

The large range in mean velocity, however, could not
have been entirely due to a difference in vegetation
density and bending of the trees because the complete
removal of vegetation apparently only caused about a
30-percent decrease in the Manning roughness
coefficient and a 30-percent increase in mean velocity.
Furthermore, a large range in velocity still existed in
the study reach after the vegetation had been removed.

EFFECTS OF BOTTOM—LAND VEGETATION 0N MAJOR FLOODS, GILA RIVER, ARIZONA

During the 1972 flood at sections not affected by man-
made structures, the range in the mean cross-sectional
velocity was from 4.7 ft/s (1.43 m/s) at section 5 to 8.9 ft/s
(2.71 m/s) at section 7 (table 1), a difference of about 90
percent of the lower figure. The writer assumes that the
large range in mean velocity is mainly due to differ-
ences in boundary roughness caused by the meandering
stream channel. The cross sections at which the mean
velocities were relatively high were located, where the
stream is relatively straight (pl. 1); the computed
roughness coefficients are relatively small at these
sites. The cross sections at which the mean velocities
were relatively low were located where large parts of
the flow moved across the meandering stream channel;
the computed roughness coefficients are relatively large
at these sites and the mean depths upstream from these
sites are relatively large. Much turbulence along the
stream-channel banks is known to exist when a major
flood moves across the meandering stream channel of
the Gila River (Burkham, 1972), and the roughness
coefficient in such a situation is known to be large
(Rouse, 1961, p. 593).

SUMMARY AND CONCLUSIONS

Changes in bottom-land vegetation between major
floods in December 1965, August 1967, and October
1972 significantly affected the peak-discharge major
flood parameters of stage, mean cross-sectional velocity,
channel-boundary roughness, and mean cross-sectional
depth. The peak discharges for the floods were respec-
tively 39,000, 40,000, and 80,000 ft 3/s (1,100, 1,130, and
2,270 m3/s). Changes in vegetation between floods con-
sisted of:

1. The complete eradication of trees, mainly
saltcedar and mesquite, in reach 1 between the
1965 and 1967 floods;

2. An increase in foliage in reach 2 between the 1965
and 1967 floods; and

3. The complete eradication of trees in reach 2 be-
tween the 1967 and 1972 floods.

The eradication of fully foliaged trees apparently

caused the following changes:

1. An average 1.0-ft (0.30 m) decrease in stage for the
1967 and 197 2 floods in treated areas. The com-
puted average decrease in stage is 1.1 ft (0.34 m)
for the 1967 flood in reach 1 and 1.0 ft (0.30 m) for
the 1972 flood in reach 2.

2. An average 0.6 ft/s (0.18 m/s) increase in mean
cross-sectional velocity for the 1967 and 1972
floods in treated areas; this increase is about 24
percent of the average of mean cross-sectional
velocities for the 1965 flood along the study
reach. The computed average increase in mean
cross-sectional velocity is 0.8 ft/s (0.24 m/s) for

 

J13

the 1967 flood in reach 1 and 0.5 ft/s (0.15 m/s) for
the 1972 flood in reach 2.

3. An average 0.024 decrease in Manning roughness
coefficient na for the 1967 and 1972 floods in
treated areas; this decrease was about 30 percent
of the average na for the 1965 flood. The com-
puted average decrease is 0.026 for the 1967 flood
in reach 1 and 0.021 for the 1972 flood in reach 2.

4. An average 0.9 ft (0.27 In) decrease in mean cross-
sectional depth for the 1967 and 1972 floods in
treated areas; this is about 15 percent of the
average of mean cross-sectional depths for the
1965 flood. The computed decrease in mean
cross-sectional depth is 1.2 ft (0.37 m) for the
1967 flood in reach 1 and 0.6 ft (0.18 m) for the
1972 flood in reach 2.

The increase in foliage between the 1965 and 1967

floods apparently caused the following changes:

1. An average 0.4 ft (0.12 m) increase in stage for the
1967 flood;

2. An average 0.2 ft/s (0.06 m/s) decrease in cross-
sectional velocity for the 1967 flood; this is about
7 percent of the average of mean cross-sectional
velocities in reach 2;

3. An average increase of 0.008-in Manning rough-
ness coefficient na for the 1967 flood; this is about
11 percent of the nu. for the 1965 flood in reach 2;

4. An average 0.4 ft (0.12 m) increase in depth for the
1967 flood; this is about 6 percent of the average
of mean cross-sectional depths for the 1965 flood
in reach 2.

The range in the different hydraulic parameters
along the study reach for the three floods was greater
than expected. For the 1965 flood at sections not affected
by manmade structures the range in

.1_ Mean cross-sectional velocity was from 2.0 to 4.1
ft/s (0.61 to 1.25 m/s), a difference of about 100
percent;

2. Manning roughness coefficient n was from 0.036 to
0.102, a difference of about 180 percent; and in

3. Mean cross-sectional depth was from 4.7 to 7.8 ft
(1.43 to 2.38 m), difference of about 70 percent.

For the 1972 flood at sections not affected by man-

made structures the range in

1. Mean cross-sectional velocity was from 4.7 to 8.9
ft/s (1.43 to 2.71 m/s), difference of about 90 per-
cent;

2. Manning roughness coefficient n was from 0.016 to
0.049, a difference of about 210 percent; and in

3. Mean cross-sectional depth was from 3.8 to 6.8 ft
(1.16 to 2.07 m), difference of about 80 percent.

The writer assumes that the removal of vegetation

did not greatly affect the range of the different parame-
ters because the range was not significantly different for

J14

the 1965 and 1972 floods; vegetation was in place during
the 1965 flood but it had been removed before the 1972
flood. The large range in the different parameters is
probably due mainly to differences in boundary rough-
ness caused by the meandering stream channel. The
cross sections at which the mean velocities were rela-
tively high are located where the stream is relatively
straight; the computed roughness coefficients are rela-
tively small at these sites. The cross sections at which
the mean velocities were relatively low are located
where large parts of the flow moved across the meander-
ing stream channel; the computed roughness
coefficients are relatively large at these sites and the
mean depths upstream from these sites are relatively
large.

REFERENCES CITED

Barnes, H. H., Jr., 1967, Roughness characteristics of natural
channels: U.S. Geol. Survey Water-Supply Paper 1849, 213 p.

Burkham, D. E., 1970, Precipitation, streamflow, and major floods at
selected sites in the Gila River drainage basin above Coolidge
Dam, Arizona: US. Geol. Survey Prof. Paper 655—B, p. Bl—B33.

1972, Channel changes of the Gila River in Safford Valley,

 

 

GILA RIVER PHREATOPHYTE PROJECT

Arizona, 1846-1970: US Geol. Survey Prof. Paper 655—G, p.

G1—G24.

1975, Effects of changes in an alluvial channel on the timing,
magnitude, and transformation of floodwaves, southeastern
Arizona: US. Geol. Survey Prof. Paper 655—K, in press.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis of streamflow
data for an alluvial stream: U.S. Geol. Survey Prof. Paper 655—0,
p. 01—013.

Chow, Ven Te, 1959, Open channel hydraulics: New York, McGraw-
Hill, 680 p.

Culler, R. C., 1965, The Gila River Phreatophyte Project: Arizona
State Land Dept. Ninth Ann. Watershed Symposium Proc., p.
33—38.

Culler, R. C., and others, 1970, Objectives, methods, and
environment—Gila River Phreatophyte Project, Graham Coun-
ty, Arizona: US. Geol. Survey Prof. Paper 655—A, A1—A25.

Robinson, T. W.,1965, introduction, spread and areal extent of
saltcedar (Tamarix) in the Western States: US. Geol. Survey
Prof. Paper 491—A, p. Al—A12.

Rouse, Hunter, 1961, Engineering hydraulics: New York, John Wiley
and Sons, 1039 p.

US. Army Corps of Engineers, 1914, San Carlos Irrigation Project,
Arizona: US. 63d Cong, 2d sess., H. Doc. 791, 168 p.

US. Geological Survey, Surface water supply of the United States—
Part 9, Colorado River basin: U.S. Geol. Survey Water-Supply
Papers, 1963—72.

 

lII’U.S. GOVERNMENT PRINTING OFFICE: 1975-0-690-036/56

 

 

 

UNITED STATES DEPARTMENT OF THE INTERIOR

 

PROFESSIONAL PAPER 655—]

 

 

 

 

 

 

   

 

     
     
   

    
  
   

 

   

  
   
  
  

    
    
   
   
 

    
  
 

 

 
 
  

       
 
 

 

 
    
    

 

   

 

   

GEOLOGICAL SURVEY PLATE 1
_11o’”20' “0835315.
3315’ ,, ,, ~- xv, / / , 'Lv
/ . w I , 1 i? g t 7
,1 , , 00‘ S ‘ ,' ,. \ f , '/
‘ -, \faq' , ., é , ‘ ’ 1 m, J 1 ,
, i -, , /’ "P ' A 4 I
.,/ , ,1 wigsmw Weu ' "
‘ NC/[L P x 14'
1: L1
, N 1 My
1 , _ n: 1.
H , 0 , I 1
xi. »- - I ,
BIN. 0\ ARIZONA z?
' h :?/_\\\8M 2549 5 J1 ‘ xf%/
"1 ii: —-{340 _ \
13’ " _ , , g {L River I ‘. 1%; 13!
- / ,' , . ’ ' ,n' , .' , \b 0’ '
,BMMW 1 , o s / it, ~
” / regs: ,‘ 1 /, , 1 “1313142542 a) ‘ r I' 1 y ’ Project area . Mg
’ 1., ‘x, _ V , { \\*: g ‘3 , 1 ,1 » Safford Q
. '1 _ EM 254 5X:
3 Gila River mea ‘ Tumor? '32 E
. .z -1 , f
,1 Calva, Arr/zona ,
121” _ 1 Flat “112
Dewey’flal
W911 '
1 ~ ,
W A fi 11'

10'

 

 

33"08'1flfl
110 20'

Base from U.S. Geological Survey Bylas 1262,500,
1960 and San Carlos Reservoir 1:62,500, 1962

 
 

(l

2)

11’

  

gee We‘ll U

  
 
 
  

Gila giveéinearé)
-‘ -: __ . ». / 1 “s
Bylas Arizona; , 3‘

 
  

       
   

 

weyflat/ Dam, _

 

‘ 3/, 712456

 

 

 

 

V2 0 1 4 MILES

 

 

 

1—1 1—1 1—1 1—1 1—1 ' - r».——-— r 1 1;
o

2

1 .5 0 1 2 3 4 5 6 KILOMETRES g
HHHHHl—l l~——-—-l I--—1 J E

 

 

CONTOUR INTERVAL 80 FEET (24 METRES)
DATUM IS MEAN SEA LEVEL

APPROXIMATE MEAN
DECLINATION, 1976

DECEMBER 21, 1965

110°15'

 
 
  
 

JULY 19, 1967

OCTOBER 21, 1971

    
 

Aerial photomaps by U.S. Geological Survey Scale
of aerial photographs approximately 1262,500

NOVEMBER 9, 1972

EXPLANATION

 

A

JULY 19, 1967

17

Cross section and number

 

_T'—

Center of main flow path of flood.
Number indicates distance in
miles from cross section 17

U.S. Geological Survey
streamflow-gaging station

Area inundated by the three floods Date aerial photograph was taken

MAPS AND AERIAL PHOTOGRAPHS SHOWING STUDY AREA, STREAM-GAGING STATIONS,
CROSS SECTIONS, AND THE AREA OF INUNDATION FOR THE FLOODS OF
DECEMBER 1965, AUGUST 1967, AND OCTOBER 1972,

GILA RIVER, SOUTHEASTERN ARIZONA

Interior—Geological Survey, Reston, Va.—1976—W75077

bU.S. GOVERNMENT PRINTING OFFICE: 1975-0-690-036/56

 

UNITED STATES DEPARTMENT OF THE INTERIOR
F SSIONAL PAPE _
GEOLOGICAL SURVEY PRO E P1535}: 21
DISTANCE FROM LEFT BANK REFERENCE MARKER, IN METRES DISTANCE FROM LEFT BANK REFERENCE MARKER IN METRES

500 600 700 800 0

I I I I

100 200 300 400 500 600 700 800 900 1000
I I I I

I I I '
W:___L_~_L___J___L___|____IL__J_I__:F__—t___I___l_I_.___|_|_;_‘I/

 

 

 

 

 

 

 

 

 

EARTH DIKE

 

 

 

 

 

 

IN FEET, ABOVE MEAN SEA LEVEL

CROSS SECTION 1 1

IN FEET, ABOVE MEAN SEA LEVEL
IN METRES, ABOVE MEAN SEA LEVEL

 

 

I I I I I | I I
1000 1200 1600 1800 2400 2600 2800 3000 3200 3400 3600

 

762

ALTITUDE,

CROSS SECTION 1

ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

ALTITUDE,
ALTITUDE,

 

 

I I I I I I I I I I 3‘30 5‘00 800 900 1000 1 100
1000 1200 1400 1600 1800 2200 2400 2600 2800 3200 3400 3600 3800 I I I I I I II

 

 

300 500

 

 

 

 

 

CROSS SECTION 3

, IN METRES, ABOVE MEAN SEA LEVEL
IN FEET ABOVE MEAN SEA LEVEL

— 778

 

 

I I I I I I I l I I I I I I
400 800 1000 1200 1600 1800 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200

ALTITUDE, IN FEET, ABOVE MEAN SEA LEVEL

 

ALTITUDE
ALTITUDE,

CROSS SECTION 13

ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

 

300 500 800 900 1000 12100
I I I I I I _ J I I I I I I I I I I — 756

400 1000 1200 1400 1600 1800 2000 2400 2600 3000 3200 3400 3600 3800

 

 

 

300 500 600 800
I I I I
I I I I

 

 

 

 

 

 

 

 

 

 

IN FEET, ABOVE MEAN SEA LEVEL

CROSS SECTION 5

ALTITUDE,
ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

 

I I I I | L
400 1200 1600 1800 2200 2400 2600 2800 3000

 

 

 

CROSS SECTION 15

_I
LU
>
LIJ
_I
<
LU
U}
2
<
LU
E
LU
>
o
m
<
I;
LLI
LLI
LL
3
LLI
D
D
I:
'—
_I
<

 

ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

500 800 900
— 776
I I 44 LL I I I I I
0 1000 1200 1400 1600 1800 2000 2200

 

 

300 500 600

I | I
I I | I

 

 

 

 

 

 

CROSS SECTION 7

ALTITUDE, IN FEET, ABOVE MEAN SEA LEVEL
ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

IN METRES, ABOVE MEAN SEA LEVEL

 

I I I I I I
400 1200 1400- 1600 1800 2200 2400 2600

 

 

CROSS SECTION 17

'ALTITUDE, IN FEET, ABOVE MEAN SEA LEVEL

500

I I I I I I I I I I

800 1000 1200 1400 1600 1800 2000 2200

 

ALTITUDE,

 

 

 

DISTANCE FROM LEFT BANK REFERENCE MARKER, IN FEET
EXPLANATION

Note: Vertical exaggeration X 40

 

 

CROSS~SECTIONAL PROFILE PEAK STAGE
DATA OBTAINED —_ Flood of December 1965, discharge
June 1964 39,000 ft3/S (1,100 m3 /s)

June 1966 ——___ Flood of August 1967, discharge
June 1968 40,000ft3/S(1,130 ma/S)

March 1970 —— Flood of October 1972, discharge

III) CROSS-SECTIONAL PROFILES AND MAXIMUM STAGE
FOR THE FLOODS OF DECEMBER 1965, AUGUST 1967,
AND OCTOBER 1972 AT NINE SECTIONS ALONG
43,0 650 THE STUDY REACH OF THE GILA RIVER,
.N SOUTHEASTERN ARIZONA

CROSS SECTION 9

ALTITUDE, IN FEET, ABOVE MEAN SEA LEVEL
ALTITUDE, IN METRES, ABOVE MEAN SEA LEVEL

 

 

 

Interior—Geological Survey, Reston, Va.—1976— W75077

Bus. GOVERNMENT PRINTING OFFICE: 1975-0-690-036/56

 

one:
Pb
v. was-K 7 DAYS

Effects of Changes in an Alluvial
m Channel on the Timing, Magnitude, and

CES

Transformation of Flood Waves,

Southeastern Arizona

/

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—K

 

DOCUMENTS DEPARTMENT

   
    
  
    

JUL 9 1975 MW
.,..:“;~f\'\‘l Dr M; “2,»
“BRA RV 2:ng "31:;
U ' I y! \‘
NthRSll"( C-F CALIFORNIA 1/?

, M
,1” MI,

7,, <('

t. » a . . “(9
«53.1%; Stew ‘9‘ ’

   

JUN 2 4 1976

0.3.5.1),

 

 

Effects of Changes in an Alluvial
Channel on the Timing, Magnitude, and
Transformation of Flood Waves,

Southeastern Arizona
By D. E. BURKHAM

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—K

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1976

 

 

UNITED STATES DEPARTMENT OF THE INTERIOR
THOMAS S. KLEPPE, Secretary

GEOLOGICAL SURVEY

V. E. McKerey, Director

Library of Congress Cataloging in Publication Data

Burkham, D. E., 1927—
Effects of changes in an alluvial channel on the timing, magnitude, and transformation of flood waves, southeastern Arizona.
(Gila River phreatophyte project)

(Geological Survey Professional Paper 65 5 -K)

Bibliography: p. 24—25.

Supt. of Docs. no.: I 19.16z655—K

l. Gila River—Floods. 2. Channels (Hydraulic engineering) 1. Title. II. Series: Gila River phreatophyte project. III. Series:
United States. Geological Survey Professaional Paper 655—K.

QE75.P9 no. 655—K[TC425.G5] 557.3’085l551.4’8] 75—619227

 

For sale by the Superintendent of Documents, US. Government Printing Office
Washington, DC 20402

Stock Number 024—001-02823-9

 

CONTENTS

 

Page Page
Abstract __________________________________________________ K1 Temporal changes in spatial transformation of flood waves___-K12
Introduction ______________________________________________ 1 Flood hydrographs ____________________________________ 12
Characteristics of the study reach __________________________ 3 Muskingum flood-routing method ______________________ 17
Timing and velocity of flood waves __________________________ 4 Basic concepts ____________________________________ 17
Relation between peak discharge and lag time __________ 4 Special methods ____________________________________ 18
Comparisons of temporal changes in flood-wave lag time and Routing of floods ______________________________________ 19
velocity with changes in channel parameters ______ 7 Floods of July 14—16, 1919, and September 3—5, 1925 19
Lag time __________________________________________ 7 Flood of July 23—25, 1955 __________________________ 20
Velocity __________________________________________ 7 Flood of January 11—17, 1960 ______________________ 21
Flow-boundary roughness __________________________ 10 Discussion of results __________________________________ 21
Discussion of results __________________________________ 11 Summary and conclusions __________________________________ 22
References cited ____________________________________________ 24
ILLUSTRATIONS
Page
FIGURE 1. Index map of project area and map showing study reach and location of gaging stations, Gila River ______________ K2
2—11. Graphs showing.

2. Relations between average peak discharge and lag time of the center of mass of flood waves moving through a
reach of the Gila River ____________________________________________________________________________ 5

3. Average peak discharge and differences in lag time of the center of mass of flood waves for four periods after 1927
compared with those for 1914—27 __________________________________________________________________ 7
4. Relations between average peak discharge and lag time of the peak—discharge rates of flood waves ________ 8
5. Annual floods, average stream-channel width, and lag time of center of mass of flood waves ______________ 9
6. Historical changes in the bottom land in three reaches of the Gila River ________________________________ 10
7. Average velocity of the center of mass of flood waves __________________________________________________ 11
8. Measured and synthesized floodflow, July 14—16, 1919, and September 3—5, 1925 __________________________ 13
9. Measured and synthesized floodflow, July 23—25, 1955 __________________________________________________ 14
10. Measured and synthesized floodflow, January 11—19,1960 ______________________________________________ 15

11. Relation between inflow and outflow for peak discharges of floods moving through the study reach during
1914—27, 1930—32, and 1944—65 ____________________________________________________________________ 16

TABLES

Page
TABLE 1. Streamflow-gaging stations 1n or near Safford Valley ___________________________________________________________ K3
2. Velocity of the center of mass of flood waves and approximate values of Manning 71 for selected peak discharges _________ 23

CONVERSION FACTORS

Factors for converting English units to the International System of Units (SI) are given below to
four significant figures. However, in the text the metric equivalents are shown only to the number of
significant figures consistent with the values for the English units.

English Multiply by Metric
acre-ft (acre-feet) ____________________ 1233 ________________ 1113 (cubic metres)
ft (feet) ______________________________ 30.48 ______________ cm (centimetres)
ft (feet) ______________________________ .3048 ____________ m (metres)

ft/s (feet per second) __________________

.3048 ____________ m/s (metres per second)

111

IV

English

ft3/s (cubic feet per second) ____________
mi (miles) ____________________________
mi2 (square miles) ____________________

CONTENTS

Multiply by Metric

0.02832 __________ m3/s (cubic metres per second)
1.609 ____________ km (kilometres)

2.590 ____________ km2 (square kilometres)

a.

 

GILA RIVER PHREATOPHYTE PROJECT

EFFECTS OF CHANGES IN AN ALLUVIAL CHANNEL
ON THE TIMING, MAGNITUDE, AND TRANSFORMATION
OF FLOOD WAVES, SOUTHEASTERN ARIZONA

BY D. E. BURKHAM

ABSTRACT

The stream channel of the Gila River in Safford Valley, Ariz., is
wide and straight at the end of a period in which high flows have been
dominant and is narrow and has a meander pattern at the end of a
period in which low flows have been dominant; therefore, the size and
meander pattern of the stream channel are regarded as determined by
past dominant flows. The stream-channel and flood-plain system,
when fully developed for a dominant flow, has a persistent effect on
floods. A system developed for low flows reduces flood rates; the peak
flows of flashy floods (floods that have large peak rates and small
volumes) may be reduced to bankfull discharge. A system developed
for high flows does not increase flood rates; however, streamflow from
side tributaries along the study reach may contribute more signifi-
cantly to peak rates in the Gila River when a high-flow system is in
effect than when a low-flow system is in effect. At the downstream end
of the study reach, the measured annual peak flows reflect the persist-
ence of the upstream system and, therefore, are not random in time. A
high-flow system was in effect during 1914—27, and a low-flow system
began developing after about 1930 but was not fully developed until
about 1964.

The velocity of the center of mass of flood waves that had peak
discharges of between 10,000 and 20,000 ft3/s (cubic feet per secdnd) or
283 and 566 m3/sec (cubic metres per second) during 1914—27 may
have been up to three times as much as that for the same rates during
1943—70. During 1914—27, the trend was toward a gradual increase in
velocity of the center of mass of flood waves—a decrease in lag time—
as the peak discharge increased. During 1943—70, the trend was
toward an increase in velocity of the center of mass of flood waves as
the peak discharge increased from about 500 to 4,000 ft3/s (14 to 113
m3/s), a decrease in velocity as the peak discharge increased from
about 4,000 to 20,000 ft3/s (113 to 566 m3/s) and an increase in velocity
for a peak discharge of more than 20,000 ft3/s (566 m3/s). The velocities
of the center of mass flood waves that had peak discharges of 300 to 500
ft3/s (8.5 to 14 m3/s) apparently were not significantly different for the
two periods.

Outflow rates for flood Waves mofiflhrough the study reach when
the stream channel is wide and straight can be synthesized using the
standard Muskingum flood-routing method (Carter and Godfrey,
1960) if an inflow hydrograph and an approximate value for the
coefficient K are available. The standard Muskingum method, how—
ever, is not suitable for the routing of flood waves—except possibly for
extremely small or large waves—that occur when the stream channel
is narrow and the flood plain is fully developed.

 

INTRODUCTION

In the arid and semiarid regions of the United States,
major changes in the width, depth, slope, meander pat-
tern, and boundary material of several alluvial chan-
nels have been observed since about 1850 (Olmstead,
1919; Bryan, 1926; Schumm and Lichty, 1963; Burk-
ham, 1972). The reasons for these changes have been
described for a few channels (Schumm and Lichty, 1963;
Burkham, 1972), but the adjustments in other compo-
nents of the hydrologic system that were caused by
these changes have not been described in any detail.
Major channel changes probably will cause mutual ad-
justments in water and sediment yield, in the timing
and magnitude of floods, in the surface-water and
ground-water relations, and in vegetation types and
density. A comprehensive knowledge of the temporal
and spatial changes in all components of the hydrologic
system is required before questions can be answered
about the availability, distribution, and movement of
water and the effects of human efforts to develop and
control the water resources of arid and semiarid regions.

The Gila River in the semiarid Safford Valley in
southeastern Arizona (fig. 1) is an example of an allu-
vial channel in which recent major changes in channel
width, slope, meander pattern, and bottom-land vegeta-
tion have occurred (Burkham, 1972). This report gives
descriptions of the effects on the timing, magnitude, and
transformation of flood waves caused by changes in the
Gila River in Safford Valley during 1914—70. The ap-
proach used in the study deals with lumped parameters,
averages, and trends; this report advances general ideas
derived from observations and reasonable speculations.
Although adjustments in flood waves having peak dis-
charges of more than about 10,000 ft3/s (283 m3/s) are of
primary interest in this report, adjustments in smaller
flood waves are described briefly.

K1

K2

110°

GILA RIVER PHREATOPHYTE PROJECT

 

 

STUDY REACH X
X

\

 

SOUTHERN C
PACIFIC J 13

14

   

s 0"“
SAN CARLOS six ‘1 [Q
RESER VOIR “Egng/

r./'/
./‘

 

   
   

15 .
Coolidge Dam
....... " (a.
’’’’’ o
9";

 

 
  

 

 

 

     
      

"El—LT“!
! v I
!. l: l
o . ’ l
4
2 L g!
f ,2]
long/III
in” J

_ V~W~TAPACH .
|Holbrook\._v.

San Francis-ca
iver

, . y .
‘ ._i___._. ,____;_ ........ i'g G ° I Dun an
. . is ——————— as? °

‘~. 4
._ (Tucson: ?;
32 ~.| ’ , '5“

\~ ! \‘° cocmse

‘8.
<
2

. "‘ " R Z, 1.

0 25

50 MILES

0 40 80 KILOMETRES

o 20 MILES
l l l | |
I ‘ l l l l |
o

5 1O 15 20 25 3O KILOMETRES

EXPLANATION

‘8
Streamflow—gaging station
(Operated by the US. Geological Survey)
Number (8) is gaging-station number
referred to in table 1

Base from US. Geological Survey 1 :250,000
Mesa, 1954—65; Clifton, 1954—62; Silver
City, 1954—62

FIGURE 1.—lndex map of project area and map showing extent of study reach and location of gaging stations, Gila River in Safford Valley.

The effects of channel changes on the timing, mag-
nitude, and transformation of flood waves were deter-
mined by comparing temporal changes in lag time, ve-
locity, and reservoir action with temporal changes in
the size, shape, length, and tortuosity of the stream
channel and in the vegetation on the flood plain. Lag
time is the time required for a definable part of a flood
wave to pass from the upstream end of a reach to the
downstream end. Reservoir action refers to the modifi—
cation of a flood wave by reservoir-type storage (Carter
and Godfrey, 1960, p. 85). As used herein, the term
“bottom land” refers to the area inundated during major

 

floods, and the term “flood plain” refers to the part of the
bottom land not occupied by the stream channel. The
term "stream channel” refers to the area that is gener-
ally void of vegetation and that has a definite bed in
which flowing water is confined by banks.

Temporal changes in the lag time of flood waves were
determined by comparing average relations between
flood size and lag time and sets of inflow and outflow
hydrographs for different time periods. Unless other-
wise stated, the terms “lag time” and "lag time of the
center of mass” refer to the time required for the center
of mass of a flood wave to travel between two cross

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

sections of channel, and the term “lag time of the peak
rate” refers to the time interval between the times that
the peak rate of discharge of a flood wave occurs at the
two cross sections.

Temporal changes in flood—wave velocity were deter-
mined from historical data for lag time and length of
channel. The term “velocity of the center of mass” refer
to the average velocity of the center of mass of a floofl
wave, and the term “velocity of the peak rate” refers to
the average velocity of the peak discharge

Temporal changes in reservoir action were deter-
mined by comparing sets of inflow and outflow hydrd-
graphs for different time periods. The significance df
changes in timing of flood waves and reservoir action on
flood routing was studied briefly using the Muskingum
flood-routing method (Carter and Godfrey, 1960). Flood
routing, which is the process of determining progres-
sively the timing and magnitude of a flood wave at
successive points along a river, was used as a tool in
separating the attenuation effects of reservoir action on
flood—peak rates from the attenuation effects of deple—
tion of streamflow by infiltration.

Two principal types of data were used in this study—+-
streamflow data and channel-change data. Streamflo
records from 15 US. Geological Survey gaging st ‘—
tions in or near Safford Valley were used in the analyses
(table 1; fig. 1). Most of the streamflow data used in the
analyses are from the gaging stations Gila River at head
of Safford Valley, near Solomon, Gila River at Calva,
and Gila River at or near San Carlos (table 1; fig. 1).
Data on channel changes are from a study made by
Burkham (1972). i

This report is one of several chapters of US. Geologi-
cal Survey Professional Paper 655, which describes the
environmental variables that affect evapotranspiration
in the Gila River Phreatophyte Project area (Culler and
others, 1970). ;

CHARACTERISTICS OF THE STUDY REACH

Safford Valley, which extends from the confluence of
the Gila River and Bonita Creek to Coolidge Dam, is
about 12 mi (19 km) wide and 75 mi (120 km) long
(fig. 1). The valley is filled with more than 1,000 ft (300
m) of silt, sand, and gravel. The present (1972) strearri
channel of the Gila River in Safford Valley is from 60 to
800 ft (18 to 240 m) wide, and the bottom land area is
from 1,000 to 5,000 ft (300 to 1,500 m) wide. The stream
channel, a pool-and-riffle type, has an average slope of
about 0.002. The depth to ground water in the alluviurrl
along the river generally is less than 20 ft (6 m) below
the land surface. Safford is at an altitude of 2,900 ft
(880 m) above mean sea level in the upstream end of
Safford Valley. The annual precipitation at Safford
ranges from 3.0 to 17.5 in. (76 to 444 mm) and averages

 

K3

TABLE 1.———Streamflow—gaging stations in or near Safford Valley

 

 

    
  
    

Plot No.
(See Period of record
fig. 2) Gaging station used in analysis
1 Gila River at head of Safford
Valley, near Solomon 111111111111111111 April 1914—September 1970
2 San Simon River near Solomon 1111111111 June 1931—September 1932
do ____________________ 111 May 1935—September 1970
3 Gila River at Safford 111111 111 June 1940-June 1947
__do ,,,,,,,,,,,,,,,,,,,, 111 June 1948—June 1949
_do 11111111111111111111 _ October 195&September 1965

111 June 1943-January 1944

111 July 1943—January 1944

111 June 1943—0ctober 1944

111 June 1943—September 1944

111 July 1943—December 1944

111 June 1943—October 1944

111 July 1943—October 1944

111 June l943—February 1944

111 October 1963»Septernber 1970
1. October 1929—September 1970

1.. February 1963—September 1970

________ April 1914—October 1927

4 Gila River near Thatcher1 ,,,,,
5 Gila River at Pimal 11111111111
6 Gila River near Glenbarl ,,,,,
7 Gila River near Ashurst1 11
8 Gila River at Fort Thomas1 111
9 Gila River near Geronimo1 11.
10 Gila River at Black Pointl 11111
11 Gila River at Bylas‘ 11111111
12 Gila River near Bylas2 1111
13 Gila River at Calva 11111111
14 Gila River near Calva2
15 Gila River at or near San Carlos3

 

IGaging station was operated as a art ofa study by Gatewood, Robinson, Colby, Hem and
Hal enny (1950) Data are in the fi es of the U. S Geological Survey Tucson, Ariz.
2(gag ging station was operated as a part ofa study by Culler and others (1970) Data are in
the files of the U S Geological Survey Tucson, Ariz
3The completion of Coolidge Dam in 1928 made it necessary to relocate the downstream
station from the at or near San Carlos” site to the at Calva" site.

about 8.7 in. (221 mm) (Sellers, 1960). The temperature
extremes recorded at Safford are 7° and 114°F (- 14° and
455° C) (Sellers, 1960).

The annual surface-water inflow for 1938—61 aver-
aged about 257,000 acre-ft (317,000,000 m3) in the reach
of the Gila River between the head of Safford Valley and
Calva (Burkham, 1970a, table 4); the inflow includes
230,000 acre-ft (284,000,000 m3) that enters the reach
at the head of Safford Valley gaging station. The annual
flow at the Calva gaging station was about 145,000
acre-ft (179,000,100 m3) for 1938—61; therefore, the dif-
ference of 112,000 acre-ft (138,000,000 m3) was deple-
tion by infiltration, evaporation, and diversions for irri-
gation. From November through June, streamflow is
mainly from precipitation that falls during frontal
storms, snowmelt, or outflow from ground-water stor-
age and often is a combination of the three. About 70
percent of the flow in the Gila River at the head of
Safford Valley occurs from November through June.
The flow rate during November through June may be
fairly constant for several days, and the sediment con-
centrations are relatively low. From July through Oc-
tober, streamflow is mainly from thunderstorms of
small areal extent. The flow during July through Oc-
tober is usually flashy, and sediment concentrations
generally are high.

The stream channel of the Gila River in Safford Val-
ley changed significantly from 1846 to 1970 (Burkham,
1972). From 1846 to 1904 the channel meandered
through a flood plain covered with willow, cottonwood,
and mesquite. The average width of the channel was
less than 150 ft (45 m) in 1875 and less than 300 ft (90 m)
in 1903. During 1905—17, large winter floods caused
major destruction in the flood plain, and the average
width of the channel increased to about 2,000 ft (600 m).

At the head of Safford Valley, most of the destructive

K4

floods during 1905—17 originated in mountainous ter-
rain, which does not erode easily. Therefore, the amount
of sediment supplied to the reach was less than the river
could carry at full debris-carrying capacity—the
maximum load a flood is capable of carrying (Rubey,
1937). The alluvial deposits in Safford Valley in 1905—
17 were formed primarily of easily eroded sediment.
Reconstruction of the flood plain was underway during
1918—70; the stream channel narrowed, and the average
width was less than 200 ft (60 m) in 1964. The flood plain
became densely covered with saltcedar during 1918—70.
Minor widening of the stream channel occurred as a
result of the large floods in 1941, 1965, and 1967, and
the average width of the channel was about 400 ft (120
m) in 1968.

The period of flood-plain reconstruction was charac-
terized by floods having low peak discharges relative to
those in 1905-17 and large sediment loads relative to
the debris-carrying capacity of the Gila River during
1918—70 (Burkham, 1972). Primarily, the large sedi-
ment loads came from the rapid erosion of alluvial de-
posits in the low-altitude drainage basins tributary to
the Gila River. The small floods that originated in these
tributary basins spread over the wide channel of the
Gila River, lost kinetic energy, and deposited their sed-
iment. During 1935—70 the average rate of aggradation
along the bottom land was 0.03 ft/yr (0.91 cm/yr) in the
reach from the confluence of the Gila and San Simon
Rivers to the bridge at Pima and 0.08 ft/yr (2.44 cm/yr)
in the reach from the bridge at Pima to the east bound-
ary of the San Carlos Indian Reservation. The dense
cover of saltcedar and the cultivation of the bottom land
may have contributed significantly to the rapid recon—
struction of the flood plain.

The rapid erosion in the low-altitude drainage basins
tributary to the Gila River apparently was triggered by
the large floods of 1905—17 (Burkham, 1972). The Wide-
ning of the stream channel of the Gila River decreased
the length of most of the tributary streams at their
confluences with the Gila River and increased the gra-
dients of the tributaries near the confluences. The in-
creased gradients and the large floods started severe
erosion in the tributary streams at their confluences
with the Gila River. Soon deep channels were eroded
near the Gila River; these deep channels eventually
became wide and extended far upstream (Burkham,
1972).

TIMING AND VELOCITY OF FLOOD WAVES

The timing and velocity of flood waves are dependent
on other factors in addition to the channel para-
meters—size, shape, slope, roughness, and bed forms.
For example, the timing and velocity of floods in the
study reach are dependent on the size and shape of

 

GILA RIVER PHREATOPHYTE PROJECT

the inflow waves, on the debris carried by the flood, and
on the loss of surface water within the valley. All these
parameters change continually with time and place
along the Gila River. Data are not adequate to evaluate
the reasons for all the changes; however, rational specu-
lation about reasons for the average long-term changes
or trends is possible. The average relations between lag
time and magnitude of flood waves for different time
periods were established in order to determine the aver-
age temporal changes in the timing and velocity of flood
waves of different magnitudes. The temporal changes in
the timing and velocity of flood waves then were corre-
lated with temporal channel changes of the Gila River.
Water loss affects both lag time and size of flood wave;
hopefully, bias in the relations between lag time and
flood-wave magnitude is minimized by using an average
peak discharge for the flood-magnitude variable.

In the following sections the lag time of the center of
mass of flood waves is discussed in detail, and the lag
time of peak discharge is discussed briefly. The date for
a particular flood is the date of arrival of the flood at the
upstream end of the study reach.

RELATION BETWEEN PEAK DISCHARGE AND LAG TIME

Graphs showing the relation between average peak
discharge and lag time of the center of mass of flood
waves are based on data for five time periods: 1914—27,
1930—40, 1941—50, 1951—60, and 1961~70 (fig. 2). The
inflow rates for 1914—70 measured at the head of Safford
Valley gage, the outflow rates for 1914—27 measured at
the San Carlos gages, and the outflow rates for 1930—70
measured at the Calva gage were used in developing the
relations. The inflow and outflow data are for storm
periods when the tributary inflow to the study reach was
an insignificant part of the outflow. The times when the
center of mass of a flood wave passed the ends of the
study reach were needed to determine the lag time; the
times used were when half of the total volume of the
wave had passed the gaging stations. The average peak
discharge for a flood moving through the study reach
was computed using the equation

1pm,,
2

 

Qp= (1)

in which
Q,p = average peak discharge;
I p = peak inflow; and
0p = peak outflow.
The data for average peak discharge and lag time of
the center of mass for the period 1914—27 were adjusted

before being plotted in figure 2 because the 1914~27
data are for the reach from the Solomon gaging station

EFFECTS OF CHANGES OF FLOOII WAVES, SOUTHEASTERN ARIZONA

AVERAGE PEAK DISCHARGE, IN CUBIC METRES PER SECOND

 

  
   
    

 

 

 

 

 

 
 
 
  
   

    

 

 

 

 

 

 

 

 

 

6 8 10 20 40 60 80 100 200 400 600 800 1000 2000
5°Illllll I YIIII I I III lllll' I I II I I II I III I
g I | I I I | I
40 -
3
g 30
' . 8/14/23 —
E ,6/34/25L 8/6/14 (8/81/5/233
Lu 2° ‘ - / 813/161027 -
E o . .° 7/31/25
P . :3/14/23 N .4 .9/1/25
/' 9/3/25
9/19/25 2 17 9/17/14
3 A Water years 1914-27 8/20/I23/—g—-°/'N.14/19,\ 9/2/25
_, J 8/18/21 I 10/5/14
10 I I I I I I l I I l l I I I I I I I
gasollll" [1"] IIIIIIII IIIIIII JIIIIIII‘III
40 — —
D
O \
I 30 — \ 2/18/37 .12/30/40 —
E 2/25/40 /
w. 20 _ . .\ 1914.27 9/8/33 {/2/7/37 —
g . \f - , . . 10/8/39 '2/16/37
— I O O
S" B Water years 1930—40 ' . 9:51. 2/17/36
_' 10 I l l I I I I I I | I l l I I I I l I l l
gsolllllLI'IIlIll I IIIIIII lIIIIIl I IIIII'III ll
40 — '
8 1/12/49
I 30 ,_ 1/15/49 -
2 3/8/49 //
uf 20 _ 4.3/14/41 _
E
P‘- u I 1914‘27
g C Water years 1941—50 :8/30/47 12/11/41
—| 10 I I I I I J I I I I J I I I 1 I I I I I I I
wsolllll IIIIII IIIIIIIIIIIIII‘ Illllllllll'
g 40 _ ’1/11/60 _
/
g 30 _ / "1/18/52 _
2 /1/14/52
... 9/11/58 /
w‘ I
E 20 — 3/24/54 -
F__ . 10/3/55
‘ 8/24/58
2 D Water years 1951-60 ' . : .‘TTE‘ \12,26,597’23’55
0
_l I I l l I I I I I I I I I I I I I I I I I
10
”SOIIIII I'll] IIIIIIII IIIIII‘ I lllll'll' I,
”‘ 4° - °7\
8 9/29/62 ' W 12/22/65—
I 30 _ \\ .o . //12/24/65 1/2/N
E 0 " ..\ ' 1914‘27 / '8/13/67
_ C O _
E 20 ' \\\. o X
I'- a "Kl I
o . . . ‘\—’:7“\
< E Water years 1961 —70 o '
'J 10 I I I I I I I I I I I I I I I I I I I l I I
200 1000 10,000 70,000
AVERAGE PEAK DISCHARGE, IN CUBIC FEET PER SECOND
EXPLANATION o
f ' l dun/2:5 f fI d
——"“—. Dateo arriva an posn IOnO a 00 wave
Average trend line having a relatively high average peak
1914-27 discharge
. “—0
Average trend “"6 Trend line through two or more points; the arrow
for 1914-27 indicates the chronologicai order of the floods

FIGURE 2.—Relations between average peak discharge and lag time of the center of mass of flood
waves moving through a 55-mile (88 km) reach of the Gila River in Safford Valley. Average
peak discharge for a flood wave was taken as the mean of the peak rates at the ends of the reach.

K5

K6

to the San Carlos gaging station, which is 71 mi (114
km) long, whereas the data for the other periods are for
the reach from the Solomon gaging station to the Calva
gaging station, which is 55 mi (88 km) long (fig. 1). Each
lag time was adjusted by multiplying it by the ratio
55/71. For lack of a better method, the adjusted data for
average peak discharge for 1914—27 were obtained
using the equation

(Op—1p)

Qp=Ip+ (55/71). (2)

The average trend lines in the graphs in figure 2 are
approximations of average changes in lag time as the
average peak discharge varied. The large scatter in data
points probably is a direct result of unstable channel
and flood-plain conditions, and of water loss along the
channel.

The trend during 1914—27 was toward a gradual de-
crease in lag time as the average peak discharge in-
creased (fig. 2A); the trend is indicative of flow in which
the external resistance to water movement is mainly
from the channel bottom. Most flood waves during the
period spread over a flood channel that was about 2,000
ft (600 m) wide, fairly flat in cross section, straight, and
relatively free of vegetation (Burkham, 1972). The large
scatter of data points about the trend line in figure 2A
may have resulted from differences in bedforms in the
channel, in sizes of sediment in the flow and along the
channel, and in the degree of flood-plain development.

After 1930, the trend apparently was toward a de-
crease in lag time as the average peak discharge in-
creased from about 300 to 3,000 ft3/s (8.5 to 85 m3/s) and
an increase in lag time as the average peak discharge
increased from about 5,000 to 20,000 ft3/s (142 to 566
m3/s) (fig. 2). The minimum lag times shown by the
different average trend lines for periods after 1927
ranged from 13 to 16 hours for flood waves with average
peak discharges of 3,000 to 5,000 ft3/s (85 to 142 m3/s).
The lag time taken from the average trend line for
1914—27 was about 18 hours for flood waves with the
same discharge. Therefore, the decrease in lag time
after about 1927 for flows of 3,000—5,000 ft3/s (85—142
m3/s) is assumed to be about 10 to 30 percent. Flood
waves that had average peak discharges of 10,000 to
20,000 ft3/s (283 to 566 m3s) apparently had lag times
which averaged about 14 to 16 hours in 1914—27, 20 to
30 hours in 1930—50, and 30 to 40 hours in 1951—70.

The minimum lag time shown by the trend lines for
the different periods after 1927 was assumed to have
occurred at about bankfull discharge (Linsley and
others, 1949), corresponding to the stage at which water
moves onto the flood plain. Based on discharge meas-
urements made at different times at several sites along

 

GILA RIVER PHREATOPHYTE PROJECT

the study reach, the velocity of shallow flow on the
densely vegetated flood plain is significantly less than
the velocity of bankfull flow in the stream channel.
Because of this difference, a change in the slope of the
discharge-to—lag time relation should occur at about
bankfull discharge; the change should be toward a
longer lag time as the discharge increases. For the four
periods after 1927, changes in the slope of the average
discharge-to-lag-time relations occur at average peak
discharges ranging from 3,000 to 5,000 ft3/s (85 to 142
m3/s). The bankfull discharge for short reaches near
Calva, Ariz., for 1962 and 1963 was determined to be
about 4,000 ft3/s (113 m3/s) (Burkham and Dawdy, 1970;
Culler and others, 1970). However, the range in
bankfull discharge along the reach for a given time and
the range at a given site for the period 1927—70 are
probably significantly different.

The maximum lag time for large flood waves moving
through the study reach in 1960—70 apparently oc-
curred when the average peak discharge was about
20,000 ft3/s (566 m3/s); this assumption is based on data
from floods in January 1960, September 1962, De-
cember 1965, January 1966, and August 1967 (fig. 2).
The lag times for the flood waves of January 11, 1960,
and September 29, 1962, that had average peak dis—
charges of 12,000 to 13,000 ft3/s (340 to 368 m3/s) were
from 38 to 42 hours; the trend in the relation of average
peak discharge to lag time at these rates apparently is
toward an increase in lag time with an increase in peak
discharge. The lag time for the flood wave of December
22, 1965, that had an average peak discharge of about
40,000 ft3/s (1,130 m3/s), however, was only about 32
hours, and the trend in the relation of average peak
discharge to lag time apparently is toward a decrease in
lag time with an increase in discharge. The lag-time
data for the three floods seem to indicate that a
maximum lag time is reached for floods that have aver-
age peak discharges of 15,000 to 25,000 ft3/s (425 to 705
m3/s); the same conclusion can be reached by studying
the lag-time data for the flood waves of December 1965,
January 1966, and August 1967. The lag time for the
August 1967 flood wave, however, may have been re-
duced slightly as a result of the eradication of bottom-
land vegetation in 1967 in the 5.5-mi (8.8 km) reach of
the channel from the gaging station near Bylas to the
gaging station at Calva (Burkham, 1976).

The relations between average peak discharge and
differences in lag time for the four periods after 1927 are
compared with those for 1914—27 in figure 3. The dis-
charge with the maximum negative difference in lag
time is about bankfull discharge; a negative difference
indicates a decrease in lag time. The reason for the
increase in lag times for flood waves having average
peak discharges from 300 to 500 ft3js (8.5 to 14 m3/s)

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

AVERAGE PEAK DISCHARGE, IN CUBIC
METRES PER SECOND

200 300 400

100

 

3°‘I'I'I'l‘l'

DIFFERENCE IN LAG TIME, IN HOURS

 

 

 

| | | | | | I |
o 2 4 6 81012141618

AVERAGE PEAK DISCHARGE, IN THOUSANDS
OF CUBIC FEET PER SECOND

 

20

EXPLANATION

E—A

The difference in lag time was computed by sub-
tracting the lag time obtained from curve A
from the lag time from curve E in figure 2.
Curve A in figure 2 is for the period 1914-27;
curve B is for the period 1930—40; curve C is
for the period 1941 —50; curve D is for the
period 1951—60; and curve E is for the period
1961-70. A positive difference indicates an
increase in lag time

FIGURE 3.—Average peak discharge and differences in lag time of the
center of mass of flood waves for four periods after 1927 compared
with those for 1914—27.

apparently is largely due to an increase in channel
length caused by an increase in stream-channel mean-
der. The effects of channel meander on changes in lag
time are discussed further in the section “Comparisons
of Temporal Changes in Flood-Wave Lag Time and Ve-
locity with Changes in Channel Parameters.”

As expected, the trends in the relation of average
peak discharge to lag time of the peak rate roughly
parallel the trends in the relation of average peak dis-
charge to lag time of the flood wave; however, a few
differences were noted (figs. 2, 4). For example, the
average peak discharge with the minimum peak-rate
lag time for the two periods after 1950 apparently is
smaller than the average peak discharge with the
minimum center—of—mass lag times. Also, the scatter of
data points on the different peak-lag diagrams is great-
er than the scatter on the corresponding mass-lag dia-
grams. The reasons for these differences are not known.

 

K7

The peak-rate lag time, however, is known to be more
susceptible to changes in the spatial factors that control
water movement than the center-of-mass lag time,
which probably accounts for most of the difference in the
scatter of plotted data. Because of the smaller scatter of
plotted points, the average peak discharge correspond-
ing to the minimum lag time shown by the different
mass—lag relations is assumed to be a better estimate of
the bankfull discharge for the two periods after 1950
than that shown by the peak-lag relation.

COMPARISONS OF TEMPORAL CHANGES IN
FLOOD-WAVE LAG TIME AND VELOCITY WITH
CHANGES IN CHANNEL PARAMETERS

LAG TIME

The temporal increase after 1927 in lag time for flood
waves that had average peak rates greater than
bankfull discharge apparently was caused by the nar-
rowing of the stream channel, development of a mean-
der pattern, development of a flood plain, and growth of
flood-plain vegetation (figs. 5, 6). Abrupt decreases in
lag time, however, apparently follow major floods that
widen and straighten the stream channel (fig. 5). The
large flood of December 22, 1965, undoubtedly caused a
reduction in lag times for the flood waves of December
24, 1965, January 2, 1966, and August 13, 1967. As
previously discussed, the lag time of the flood wave on
August 13, 1967, also may have been affected by the
eradication of bottom-land vegetation along 5.5 mi (8.8
km) of channel near the downstream end of the study
reach.

The length of the main path of flow (L) may have been
an important factor in changes in lag time for some
flows. Lag times vary with L, which in the study reach
changes with time and discharge. For example, most of
the flow during floods that occurred during this study
moved directly downvalley, and L for these flows was
about 55 mi (88 km)—the length of the study reach.
Flood waves that had averaged peak discharges of less
than about 5,000 ft3/s (142 m3/s), however, followed the
meandering stream channel, and L increased from
about 55 mi (88 km) in 1918 to about 66 mi (106 km) in
1964 for these floods (Burkham, 1972).

VELOCITY

The velocity of the center of mass of flood waves that
had average peak discharges of 300 to 500 ft3/s (8.5 to 14
m3/s) apparently did not change greatly between the
1914—27 and 1961—70 periods. The average velocity of
the center of mass was obtained by dividing the length
of the main flow path by lag time. The length of the main
flow path (L) and the lag time (Tm) used to compute the
velocity were 55 mi (88 km) and 26 to 28 hours for

GILA RIVER PHREATOPHYTE PROJECT

AVERAGE PEAK DISCHARGE, IN CUBIC METRES PER SECOND

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g 6 8 10 20 40 60 80 100 200 400 600 800 1,000
8 30 I I ll;l\.l I | I II I I I III III I I II IIIIII
:1: \,\\ \" Jimmy 8/17/14]
2 . 8/16/23 9/1/25
_~ 20 — - - f \ 97 2_/ 9/_2/25 -
LLI ' e en‘s/237‘ - 1 27
E ' ° 8/16/14 7/31/2545 \ ' 9/3/25
; . . 8/18/21 7/14/19
A Water years 1914—27 8/20/23
(<9 10 I I I l I I l I l l I l l l l I I I
.I
g 40 I I IIIII 1 I IIIIII I III I I'lll I I IIII I II II'III
D ._ 11 7 37 _
o 30 \ 8/ I30 4173/ /
I \\ . 9/8/33 2/16/3
z \ a >/
—. 2° ‘ . . ‘\.._ . . 10/8/3/ ‘\ 2/18/37 "
”5‘ ' .1 “h. . /\ 9/6/40
,2 ' .. E‘Q' ' .L/ 2/17/36
I

(D B Water years 1930—40 ' “‘73"
< 10 1 1 I l 1 I I I 1 1 I 1 1 1 1 1 I 1 1 1
_l
m sollI‘llI I I Illlll I III I IIIII I I III I I III‘IIIIII
cc _ ’ 12/25/40
8 40 \\ - ‘ .fl‘L‘ 1/12/49
I 30 - . -\.<\: - . 3/8/‘9.r%”\-‘(.1/15/49
z .\, . . . Ar/
'45 20 — ' ' - \\ . / 1/28/41 _
5 .' , ,\L . ‘/ o
F t" ' _"/ ° °
0 . - .§ ' \
< C Water years 1941 —50 e
“I 10 l I l I l I l I I I I} I I l I I I l L

60 I I I‘ll I I _[ Tll'lll I I II III I I I I I I III .1I/11/6lo'II I I'll I
E 50 _ /“1/14/52 ‘—
8 4o — . . // I. —
I 3° \\ . / 1/18/52
E \ ' . / 7/23/55
u; 5' ' . . . ' no /a/24/54 '10/3/55
E 20 — \o\ o l/ _
p o . / . a
o ,
5 D Water years 1951 —60 ' '-'- ' ‘.

10 l l l I I l I I I I l l I l I I I I l 1

6° lllllll I I I|II I I III I ITIII l I I Illl I III IIIIII
3 5° _ . m —
8 40 — o //"\\ ‘
I C

\ / 12/24/55M _
E 30 .. Rat. . . / \. 12/22/65
. \ . ,/ 0 1/2/66

LIJ e \ /c e
E 20 — '. \ /. 8/13/67 —
I- \\_.;)/‘ '
(<3 . C . I
-’ E Watler years 1961—70

10 I I I I I I I I I I I I I I0 I 'I 1

200 1000 50,000

AVERAGE PEAK DISCHARGE, IN CUBIC METRES oPOEFI SECOND

EXPLANATION

_,_——-—- 0"
Average ’trend line Trend line through two or more points;
the/arrow indicates the chronological
.1/12/49 order of the floods

Date of arrival and position of a flood wave having a
relatively high average peak discharge

FIGURE 4.——Re1ations between average peak discharge and lag time of the peak-discharge rates
of flood waves moving through a 55-mi (88 km) reach of the Gila River in Safford Valley.
Average peak discharge for a flood wave was taken as the mean of the peak rates at the ends
of the reach,

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

 

 

K9

 

 

 

 

 

    

 

 

 

 

 

 
   
  

 

 

 

on
N
1 >‘
6° F 2”. EXPLANATION
8
u. K0 8
_ ~ 1 V
8 140 § s\\\\\\ — 4000 8
Q E Gila River near Florence Gila River at Calva n:
2 / F
D 120 — ‘ r m
5,5 2 k\\\ II; E
3 8 Gila River at San Carlos San Francisco River — 3000 0
O . _
I g 100 T at Clifton cm
I- : D o
2 E s’ 0 z
1 m 80 — o E 8
3 5 ‘ — 2000 u; 3,1
E {E 60 — 8 3.9. g “'-
5 2 3E 23 E i”
U) m '- N '-
_ 40 - N .- .: O
o 3 < «mg age -5 3 5 — 1000 2
¥ ‘1 $126: Tgéo’fli‘e WSW.“ E Q
N :4 =E§°%;E- ' g
g 20 — a N gamma? 3
fl. 3 E 3 <u.< D <
< _‘ LL!
0 l I N R I o Q.
1880 1890 1900 1910 1920 1930 1940 1950 1960 1970
WATER YEAR From Burkharn (1972, pl. 3)
ANNUAL FLOODS
EXPLANATION
2000 P X
A - 600
. _ . _A_ . .__. I_
E From the confluence of the San Simon and I—
Q Gila Rivers to the bridge at Pima _ 500 E
- 3
E 1500 - l -—_——-X _‘.__.._ _.
m I From the bridge at Pima to the east Lu
E | \ boundary of the San Carlos __ 400 E
< I \ Indian Reservation of w
I I- : \\ O 5 3:1
0 W .
2' E1000 — l o ‘ From the bridge on US. Highway 70 near _ 300 E E
g z I Bylas to the railroad bridge that spans 5 5
L: T I the Gila River near Calva A II E
P h | \ A F
m l\ I \ ' — 200 3
-\
(“9' 500 — l \x ' \X\ \ ' A o
< I \ I A\Qfi\ . ' h <
(I I \\ I \l \\\ \ . \ m
m i \\x JI ‘xa \A o- 100 g
<>t A J A,~ ‘\x~\\ ~ <
_.__x __________ ““;JO
0 I I I I I 1 l 'l o l O 0
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970

NOTE: Lag-time data are for the 55-mile (88.5-
kilometre) reach from the gaging station near

WATER YEARS

AVERAGE STR EAM-CHANNEL WIDTH

From Burkham (1972, fig.3)

 

60
50

Solomon to the gaging station at Calva (fig. 1)
40

30

20

LAG,T|ME IN HOURS

 

l l l I l

 

 

1910

1920 1930 1940 1950 1960
WATER YEAR
LAG TIME OF CENTER OF MASS OF FLOOD WAVES

1970

FIGURE 5.—Annua1 floods, average stream-channel width, and lag time of center of mass of flood waves having average peak dis-
charges of 10,000—20,000 ft3/s (283 to 566 m3/s), Gila River in Safford Valley. Average peak discharge of flood wave was taken as

the mean of the peak rates at the ends of the reach.

 

 

  
  
  

 

 

K10
3000
2000 I
1000 I
o " .
Stream channel outside the bottom
1000 I I 1 I l I I l I 5
From the confluence of the San Simon
and Gila Rivers to the brid e at Pima
5000 _ g T 203
E
a u.|
$4000 — 2
o
< 3
z 3000 - . . . >4
1 ...212232222 100:
“<5 ":::::::::: II
D: 2000 -— .......... <3):
< . 5 o
. u:
1000 — 2 2
°"‘ 0 <f
0 — _ IE1
Stream channel outside the bottom land <
1000 I I l I I I I I l 5
From the bridge of Pima to the east boundary
of the San Carlos Indian Reservation ‘7 1°
200° _. Vegetation removed to control evapotranspiration

 

 

 

From Burkham
YEAR (1972, fig. 5)
From the bridge on US. Highway 70 near

Bylas to the railroad bridge that spans
the Gila River near Calva

FIGURE 6.—-Historical changes in the bottom land in three reaches of
the Gila River in Safford Valley.

1914—27, and 66 mi (106 km) and 28 to 33 hours for
1964—70; lag times were taken from the trend lines
shown in figure 2. The velocity of the center of mass of
flood waves that had average peak discharges of 300 to
500 ft3/s (8.5 to 14 m3/s) was about 2.9 to 3.1 ft/s (0.88 to
.94 m/s) during 1914—27 and 2.9 to 3.4 ft/s (0.88 to 1.04
m/s) during 1961—70. The increase in lag time for these
peak-flow rates, therefore, apparently was due to the
increase in length of channel because the average veloc-
ity for 1961—70 was not significantly different from the
average velocity for 1914—27.

The velocity of the center of mass of flood waves that
had average peak discharges of 3,000 to 5,000 ft3/s (85 to
142 m3/s) was about 4.2 to 4.7 ft/s (1.3 to 1.4 m/s) during
1914—27 and 6.0 to 6.9 ft/s (1.8 to 2.1 m/s) in 1961—70.
The values of L and Tm used to compute the velocity
were 55 mi (88 km) and 17 to 19 hours for 1914—27 and
66 mi (106 km) and 14 to 16 hours for 1961470; the lag
times were taken from the trend lines shown in figure 2.
The increase in velocity between 1914—27 and 1961—70
for flood waves that had average peak discharges of

 

GILA RIVER PHREATOPHYTE PROJECT

3,000 to 5,000 ft3/s (85 to 142 m3/s) apparently resulted
from narrowing of the stream channel.

Temporal changes in the velocity of the center of mass
of flood waves that had average peak discharges of
10,000 to 20,000 ft3/s (283 to 566 m3/s) are shown in
figure 7. The plotted velocities were computed using a
value of 55 mi (88 km) for L and values obtained from
the trend line in figure 5 for Tm. The velocity of flood
waves that had average peak discharges of 10,000 to
20,000 ft3/s (283 to 566 m3/s) during 1914—18 apparently
was more than three times that for the same rates in
1964.

FLOW-BOUNDARY ROUGHNESS

Most of the changes in the lag time and velocity of
flood waves that had peak discharges of more than 5,000
ft3/s (142 m3/s) undoubtedly resulted from changes in
flow-boundary roughness—roughness of the stream
channel and flood plain. A precise method for evaluat—
ing average temporal changes in lag time caused by
changes in flow-boundary roughness for rapidly varied
unsteady flows in the reach was not available for this
study. However, discussions of changes in roughness, as
represented by the Manningrroughness coefficient n,
and the effects of these changes are presented. The
Manning equation is

= if? Bias '4,
n

(3)

in which
V = mean velocity of flow at a cross section;
R = hydraulic radius at a cross section;
S = energy gradient; and
n = roughness coefficient.

The Manning equation was developed for uniform
flow, in which the water—surface profile and energy gra-
dient are parallel to the streambed, and the area, hy-
draulic radius, and depth remain constant throughout
the reach. The equation is considered valid for
nonuniform flow, such as that of the Gila River, if the
energy gradient or friction slope is modified to reflect
only the losses due to boundary friction (Barnes, 1967,
p. 4). The equation is considered valid, however, only for
flow in relatively short reaches (Dalrymple and Benson,
1967).

In this study, the Manning roughness coefficient, n, is
treated as a lumped parameter, and its value during
flow in an alluvial channel and flood plain depends on a
number of time-variant and space-variant factors. Some
of the factors in the study reach that probably exert the
greatest influence on the coefficient of roughness for
flows greater than about 5,000 ft3/s (142 m3/s) are (1)
flow—boundary roughness, (2) size and shape of stream

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

 

I
I.

w A
I l
/
/

/

I I |

o .- .1

to b in
ES PER SECOND

/
/
._7
I—
11
9.0.0.
mum
NEH

N
l
I
,

VELOCITY, IN FEET PER SECOND
/
|
o
in
Y

 

 

I

O

a
VELOCIT

I I I I I
1920 1930 1940 1950 1960
WATER YEAR

 

1 0.3
1910 1970

FIGURE 7.—Average velocity of the center of mass of flood waves that
had average peak discharges of 10,000 to 20,000 ft3/s (283 to 566
m3/s) for 1914—70, Gila River in Safford Valley. Average peak dis-
charge of a flood wave was taken as the mean of the peak rates at the
ends of the reach. Velocity was taken as the ratio of the length of the
main flow path (55 mi, 88 km) divided by the lag time of the center of
mass of the flood wave; the lag time was obtained from the trend
lines shown in figure 5.

channel and flood plain, (3) stream-channel irregularity
and alinement, (4) vegetation, (5) obstructions, (6) flow
depth and rate, (7) filling and scouring, (8) size and
concentration of sediment in flow, and (9) bed forms.
Most of these factors are interdependent. Factors 1 to 5
are known to have changed significantly during 1914—
70 (Burkham, 1972). The changes, in general, have been
in a direction that would increase the roughness coeffi-
cient. The relative influence of bed-form roughness
probably decreased as stream-channel width decreased
and as the density of vegetation on the flood plain in-
creased.

The roughness coefficient n for most of the study reach
for most floods during 1914—27 probably ranged be-
tween 0.02 and 0.04. This estimate of the probable range
in n is based on photographs and other data (Olmstead,
1919; Burkham, 1972). C. C. Jacob (US. Geological
Survey, written commun., 1916) collected data for 1916
flow rates greater than 25,000 ft3/s (705 m3/s); these
data were used to compute n values for short reaches at
the upstream and downstream ends of Safford Valley.
The n values obtained using Jacob’s data in equation (3)
were 0.024 and 0.027.

The average roughness coefficient for most of the
study reach for flows ranging from about 300 to 5,000
ft3/s (8.5 to 142 m3/s) probably ranged from 0.02 to 0.04
during 1914—70; for flows ranging from 10,000 to 40,000
ft3/s (283 to 1,130 m3/s), the value of n probably ranged
from 0.06 to 0.14 for most of the study reach during the
same period. These estimates of the probable range in n
are based on photographs and other data (Burkham,

 

K11

1972) and on the author’s observations and studies. The
author made several preliminary evaluations for n
using equation (3) for 1962—70 flows below bankfull
discharge at nine cross sections in the 15-mi (24-km)
study reach of the Gila River Phreatophyte Project area
(Culler and others, 1970); the roughness coefficient n
ranged from 0.02 to 0.04. In making the evaluations the
velocity and hydraulic radius in equation (3) were
known, and S was assumed to be equal to the slope of the
water surface. The average value of n computed for
selected reaches in the Gila River Phreatophyte Project
area (Culler and others, 1970) for peak flows of 39,000
and 40,000 ft3/s (1,100 and 1,130 m3/s)—which occurred
in 1965 and 1967, respectively—was about 0.08 (Burk-
ham, 1976).

DISCUSSION OF RESULTS

The lack of major floods in the study reach during
1918—64 led to channel changes that greatly affected the
lag time and velocity of floods. The wide stream channel
that existed in 1914 changed: the stream channel be-
came narrower, stream-channel meander increased,
natural levees developed along the stream channel,
flood-plain vegetation spread and became dense, and
large alluvial fans developed at the mouth of tributaries
(Burkham, 1972). In addition, cultivation of the flood
plain increased during the absence of large floods. After
the changes, the Gila River consisted of a stream chan-
nel in which flow below bankfull discharge—3,000 to
5,000 ft3/s (85 to 142 m3/s) after 1930—moved relatively
fast and a congested flood plain in which floodflow
moved relatively slow. The velocity of the center of mass
of flood waves that had average peak discharges of 300
to 500 ft3/s (8.5 to 14 m3/s) apparently did not change
greatly during 1914—70; the velocity was about 2.9 to 3.1
ft/s (0.88 to 0.94 m/s) during 1914—27 and 2.9 to 3.4 ft/s
(0.88 to 1.04 m/s) during 1961—70. The range in velocity
of the center of mass of flood waves that had average
peak discharges of 3,000 to 5,000 ft3/s (85 to 142 m3/s),
however, increased from 4.2 to 4.7 ft/s (1.28 to 1.43 m/s)
during 1914—27 to 6.0 to 6.9 ft/s (1.83 to 2.10 m/s) during
1961—70. The increase-during 1914—70 in velocity of
flood waves that had peak discharges of 3,000 to 5,000
ft3/s (85 to 142 m3/s) probably was the direct result of .
increased depths for these flow rates. The velocity of the
center of mass of flood waves that had average peak
discharges of 10,000 to 20,000 ft3/s (283 to 566 m3/s)
changed greatly during 1914—70 and reached a
maximum average of about 5.7 ft/s (1.74 m/s) in 1914—18
and a minimum average of about 1.8 ft/s (0.55 m/s) in
1964; the velocity for 1914—18 was three times that in
1964, presumably a result of differences in roughness
factors.

Major flood waves that occur in the study reach when

K12

the channel is narrow and meanders and when the flood
plain is covered by dense vegetation apparently cause a
reduction in lag time and an increase in velocity of
subsequent major flood waves. Large floodflows exert
great force on the channel banks and on objects in the
main flow path and enlarge the channel (Burkham,
1972). During a major flood, the main flow path is
straight downvalley, and in many places the banks of
the meandering stream channel slow the movement of
the flood wave. While the meander pattern is intact,
part of the flow is directed along the meandering stream
channel, and a large amount of turbulence develops
along the streambanks, which further slows the move-
ment of water. As a result of the stresses produced by the
turbulent forces along the streambanks and around
other stationary objects, channel changes eventually
take place—banks erode, trees are uprooted and flushed
downstream, grass is removed, alluvial fans at the
mouths of tributaries are destroyed, and dikes that pro-
tect cropland are breached. These changes result in a
wider and cleaner stream channel that is more condu-
cive to the rapid movement of major flood waves; for
example, the major flood of December 24, 1965, enlarged
the channel, and the flood waves of January 2, 1966, and
August 13, 1967, moved through the study reach at
relatively rapid velocities. The major floods during
1905—17 undoubtedly caused the lag times of the flood
waves of October 5, 1914, and July 14, 1919, to be about
14 hours instead of more than 40 hours, the lag time
that probably would have occurred if the stream chan-
nel and flood plain had remained intact during the

1905—17 floods. The eradication of bottom-land vegeta- .,

tion in a 5.5-mi (8.8-km) length of the study reach may
have contributed to the reduced lag time for the August
13, 1967, flood.

TEMPORAL CHANGES
IN SPATIAL TRANSFORMATION
OF FLOOD WAVES

Two principal types of flood-wave movement, uni-
formly progressive and reservoir action, have been de-
scribed in the literature (Carter and Godfrey, 1960,
p. 82). Uniformly progressive refers to the downstream
movement of a flood wave as a kinematic wave that does
not change in shape; uniformly progressive flood wave
movement occurs under ideal conditions in a prismatic
channel in which the variation in resistance along the
channel is small. Reservoir action refers to the trans-
formation of a flood wave caused by reservoir pondage
which results when the channel is irregular and the
variation in resistance along the channel is large. Both
types of flood-wave movement apparently occurred in
the study reach during 1914—70.

Temporal changes in the amount of transformation of

 

GILA RIVER PHREATOPHYTE PROJECT

a flood wave as the wave moves through the study reach,
and the reasons for these changes, are described in the
following sections. The effects of channel changes on the
results of flood routing using the Muskingum method
(Carter and Godfrey, 1960) also are described.

FLOOD HYDROGRAPHS

More than 250 sets of inflow and outflow hydrographs
for 1914—70 were used to study the temporal changes in
flood wave transformation in the study reach; however,
for purposes of this report, hydrographs for four
periods—July 14—16, 1919, September 3—5, 1925, July
23—25, 1955, and January 11—17, 1960—were selected
to illustrate the changes (figs. 8, 9, and 10). A brief
description of the flow in the study reach immediately
prior to the four flood events is given as background for
later discussions of the factors that cause temporal
changes in flood wave attenuation.

The flood of July 14—16, 1919, which consisted of three
waves (fig. 8), was the first large flood in more than a
year, although the daily inflow at the upstream end of
the study reach ranged from 250 to 950 ft3/s (7.1 to
27 m3/s) for 2 months prior to July 14. The sharp peaks
of the inflow waves are typical of flash floods—floods
that have large peak rates and small volumes—
produced by thunderstorms. The sharp peaks of the
outflow waves are typical of those during 1914—27.

Several floods immediately preceded the flood of Sep-
tember 3—5, 1925' (fig. 8). A flood that had a peak dis—
charge of about 10,000 ft3/s (283 m3/s) passed through
the study reach during July 31—August 1, and about 20
floods that had smaller peaks followed between August
2 and September 3. The flood of September 3—5 probably
was the result of local thunderstorm activity during a
large convergence storm or large frontal storm.

The first flood in the summer of 1955 occurred July
11—12 and had a peak discharge of about 5,000 ft3/s (142
m3/s). Prior to July 11, the flow was less than 50 ft3/s (1.4
m3/s), and after about July 15, the flow was less than 200
ft3/s (5. 7 m3/s) until the arrival of the thunderstorm that
produced the flood of July 23—25; this flood consisted of
two waves (fig. 9).

The flood of January 11—17, 1960, which was the re-
sult of rainfall and snowmelt, had the highest peak rate
of any flood since 1952 (fig. 10). A flood that had a peak
discharge of about 5,000 ft3/s (142 m3/s), however, oc-
curred on December 27—29, 1959. The flow prior to
December 27was about 20 ft3/s (0.6 m3/s), and the flow
after December 29 was less than 500 ft3/s (14 m3/s).

Flood waves during 1914—27 generally retained their
inflow shape (shape at the upstream end of study reach)
as they moved through the study reach (fig. 8), and the
flood-wave movement approximated a uniformly pro-
gressive one; however, the uniformly progressive

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

 

DISCHARGE, IN CUBIC FEET PER SECOND

 

 

 

 

 

 

      
 

 

 

 

 

16,000 I I I I I I I 450
inflow
14,000 _400
1111111111350 ‘350
12,000 — 12,000 g
g o
_ o 8
300 8 ~300 "’
(I) II
10,000 — o I: 10,000 E
E E w
LLI
- 250 8 ii — 250 E
‘0 LIJ Measured Lu
0: 1.1. §
8000— E 2 8000— o
u: a: LT:
LU D 3
_ 200 n: o _ 200 o
E. E E
E u.i ~
6000— g o 6000— (”3
a: II o:
—150 3 < 150 <
_ o 2:) _ I
Z 2 8
4 D —'
4000- g 4000_ D
100 0: —1
< 00
__ I _
0
<2
2000 — 50 D 2000 — _ 50
0111111111110 0111111 0
0 5 10152025 30354045 505560 1 5 10 15 20 25 30 35
PERIOD,At PERIOD, At
1—11'1'1'1'1-1 1'1'1'1-1-1-1
14 15 16 3 4 5
JULY 1919 SEPTEMBER 1925

EXPLANATION

The flood of July 1919 was routed assuming continuity of
flow. Factors used in the routing are At, time units of
computation or routing periods; K, slope of the storage-
weighted discharge relation in which storage =K[x1+(1—x)0];
x, a dimensionless constant that weights the inflow, 1, and
the outflow, 0, in the storage-weighted discharge relation;

y, number of subreaches of traveltime used in the routing;
and N, number of complete cycles of routing used in syn-
thesizing the outflow hydrograph

At = 1 hour = 1 period

K = 17 hours

x = 0.5

y = 17 subreaches of traveltime
N = 17 cycles of routing

The flood was routed assuming that infiltration occurred
only during the first two waves, and the flow at the begin-
ning of each of the 17 cycles of routing was depleted
according to the equation in which

qf=0.0581b,
q =infiltration rate, in cubic feet per second, and
Ib=inflow at the beginning of the cycle, in cubic
feet per second
Other factors used in the routing are as described for curve A

Estimated inflow from tributaries along the study reach

FIGURE 8.—Measured and synthesized floodflow, July 14—16, 1919, and September 3—5, 1925. The inflow was measured at the
Gila River at head of Safford Valley, near Solomon gaging station, and the outflow was measured at the Gila River at or
near San Carlos gaging station. The synthesized outflow was obtained using the inflow hydrograph and the standard
Muskingum flood-routing method (Carter and Godfrey, 1960) or the Muskingum method as modified in this report.

The flood of September 1925 was routed assuming conti-
nuity of flow. Factors used in the routing are At, time
units of computation or routing periods; K, slope of the
storage-weighted discharge relation in which storage =
K[x1+(1——x)0] ; K0, slope of the storage-weighted dis-
charge relation for the part of the flood that is not con-
tained in the stream channel—the overbank component of
the flood; Kw, slope of the storage-weighted discharge
relation for the part of the flood that 'is contained in the
stream channel—the within-bank component of the flood;
x. a dimensionless constant that weights the inflow, I, and
the outflow, 0, in the storage-weighted discharge relation;
y, number of subreaches of traveltime used in the routing;
yo, number of subreaches of traveltime used in routing of
the overbank component of the flood; yw, number of sub-
reaches of traveltime used in routing of the within-bank
component of the flood; and N,number of complete cycles
of routing used in synthesizing the outflow hydrograph

---—A—-—— ______
At = 2 hours = 1 period At = 2 hours = 1 period
K = 22 hours Ko= 22 hours
x = 0.5 Kw=18 hours
y = 11 subreaches of x =0.45
traveltime yo =11 subreaches of traveltime
N = 11 cycles of routing YW = 9 subreaches of traveltime

N =1 cycle of routing

K13

Ki4

GILA RIVER PHREATOPHYTE PROJECT

EX PLANATION

 

Measured
inflow

10,000

8000

 

6000

 

   
 

4000 Measu red

outflow

DISCHARGE, IN CUBIC FEET PER SECOND

 

2000

 

 

————A———

The flood was routed assuming continuity of flow and
assuming that the factors that affected the flood wave
movement during 1914—27 prevailed in 1955. Factors used
in the routing are At, time units of computation or
routing periods; K, slope of the storage-weighted dis-
charge relation in which storage =K[x1+(1—x)0] ; x‘, a
dimensionless constant that weights the inflow, I, and the
outflow,0, in the storage-weighted discharge relation; y,
number of subreaches of traveltime used in the routing;
and N, number of complete cycles of routing used in
synthesizing the outflow hydrograph

300

250

q

1 hour = 1 period

15 hours

0.5

15 subreaches of traveltime
15 cycles of routing

2sxa>

200

 

 

——B— _
The flood was routed assuming continuity of flow and using
1955 channel conditions. Factors used in the routing are
At, time units of computation or routing periods; K, slope
of the storage-weighted discharge relation in which storage

=K[x1+(1—x)0] ; KO , slope of the storage-weighted dis-
charge relation for the part of the flood that is not con-
tained in the stream channel—the overbank component
of the flood. Kw, slope of the storage-weighted discharge-
weighted discharge relation for the part of the flood that
is contained in the stream channel—the within~bank
component of the flood; x, a dimensionless constant

that weights the inflow, I, and the outflow,0, in the
storage-weighted discharge relation; y, number of sub-
reaches of traveltime used in the routing; yo, number of
subreaches of traveltime used in routing of the overbank
component of the flood; yw, number of subreaches of
traveltime used in routing of the within-bank component
of the flood; and N, number of complete cycles of routing
used in synthesizing the outflow hydrograph

150

100

DISCHARGE, IN CUBIC METRES PER SECOND

1 hour = 1 period

27 hours

12 hours

0.4

27 subreaches of traveltime
12 subreaches of traveltime
3 cycles of routing

. -C- .

The flood was routed assuming that the flow at the beginning
of each of the 3 cycles of routing was depleted according to
the equation in which

II II II ll II Ii II

50

 

 

qf0.0151b.
q =infiltration rate, in cubic feet per second, and
I =inf|ow at the beginning of the cycle, in cubic
b
feet per second

 

 

PERIOD,At

0
50 Other factors used in the routing are as described for curve B

 

l
24

JULY 1955

25

FIGURE 9.—Measured and synthesized floodflow, July 23—25, 1955. The inflow was measured at the Gila River at head of Safford Valley,

near Solomon gaging station, and the outflow was measured at

the Gila River at Calva gaging station. The synthesized outflow

was obtained using the inflow hydrograph and the standard Muskingum flood—routing method (Carter and Godfrey, 1960) or the Mus-

kingum method as modified in this report.

flood-wave movement only applies to floods that were
not reduced greatly by infiltration. Infiltration during
many floods in the 1914—27 period significantly reduced
the size of the floods, and in many instances the inflow
shapes were altered.

From about 1935 to 1970, floods were transformed

greatly as they moved through the study reach, proba-
bly as a result of reservoir action (figs. 9, 10). Infiltration
also may have been a cause for the change in inflow
shape.

Significant temporal changes in the attenuation of
peak rates moving through the study reach apparently

 

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA K15

EXPLANATION
‘_A__

The July 1919 flood was routed assuming conti—
nuity of flow and assuming that the factors
that affected the flood wave movement during
1914—27 prevailed in 1960. Factors used in the
routing are At, time units of computation or
routing periods; K, slope of the storage-weigh-
ted discharge relation in which storage =
K [xI+(1—x)0] ; x, a dimensionless constant that
weights the inflow, 1,and the outflow,0, in the
storage-weighted discharge relation; y, number
of subreaches of traveltime used in the routing;
and N, number of complete cycles of routing
used in synthesizing the outflow hydrograph

 

 

At = 2 hours = 1 period

K = 14 hours

x = 0.5

y = 7 subreaches of traveltime
N = 7 cycles of routing

The flood was routed assuming continuity of
flow and using 1960 channel conditions.
Factors used in the routing are A,, time units
of computation or routing periods; K, slope
of the storage-weighted discharge relation in
which=K[xI+(1—x)0] ; K0, slepe of the storage

I I I I I I I I I I weighted discharge relation for the part of the

flood that is not contained in the stream
channel—the overbank component of the

— 45° flood; KW, slope of the storage-weighted

discharge relation for the part of the flood

\ that is contained in the stream channel—the

 

Measured
16.000 ’ inflow I\

within-bank component of the flood; .70 a
‘A r\ __ 400 dimensionless constant that weights the in-
14,000 —' flow, I, and the outflow, 0, in the storage-
weighted discharge relation; y, number of

subreaches of traveltime used in the routing;
\ \ y , number of subreaches of traveltime used

I \ _ 350 in routing of the overbank component of the

12.000 “‘ I \ flood; yw, number of subreaches of travel:
\ I time used in routing of the within-bank
_ . \ component of the flood; and N, number of
I \ I \ \ _ 300 complete cycles of routing used in synthe-
sizing the outflow hydrograph

I I I, \\
I \ I/ ‘\C\\
I

I

10'000 2 hours = 1 period

60 hours

12 hours

0.4

30 subreaches of traveltime
6 subreaches of traveltime
6 cycles of routing

I \ -- 250

II \ \
\ [I Measure \ \ _200

outflow \

II II II II II II II

8000 -
W
N

 

The flood was routed assuming that the flow at the
beginning of each of the 6 cycles of routing was
depleted according to the equation in which

Qf0.0151b,

‘ — 150

I \

\ \ \I q =infiltration rate, in cubic feet per second, and
// ‘ Ib=inflow at the beginning of the cycle, in cubic

I
l
4000 — /I/ /’ \ \ I feet per second

IN CUBIC FEET PER SECOND

6000

I

DISCHARGE, IN CUBIC METRES PER SECOND

/
/

DISCHARGE,

. ’ \ I \ _ 100 Other factors used in the routing are as described
/

/, , forcurveB
“ / \ \ l I l I l

// a
O __ \
200 // -\ _5o

 

 

 

— ’l / I | l l l I l l l l I I I
0 12 24 36 48 60 72 84 96

PERIOD, At

 

 

 

'f I ' ' I ' I ' I ' I ' I - I ' I
11 12 13 14 15 16 17 18 19
JANUARY 1960

FIGURE 10.-—Measured and synthesized floodflow, January 11—19, 1960. The inflow was measured at the Gila River at head of Safford
Valley, near Solomon gaging station, and the outflow was measured at the Gila River at Calva gaging station. The synthesized outflow
was obtained using the inflow hydrograph and the standard Muskingum flood-routing method (Carter and Godfrey, 1960) or the
Muskingum method as modified in this report.

 

 

K 16 GILA RIVER PHREATOPHYTE PROJECT

accompanied changes in wave shape (fig. 11). The 1930—32, and 1944—65 and that had peak discharges at
curves in figure 11 showing the magnitude of temporal the upstream end of the study reach ranging from 5,000
change in the attenuation of peak rates were developed to 43,000 ft3/s (142 to 1,220 m3/s) were used in the
using data for floods for which peak discharges at the analysis. During 1914—27 and 1930—32, the Gila River
ends of the study reach were known (Patterson and Som— was relatively wide, straight, free of bottom-land vege-
ers, 1966). Only data for floods that occurred in 1914—27, tation and free of alluvial fans. During 1944—65, the

PEAK DISCHARGE AT HEAD OF SAFFORD VALLEY, IN CUBIC METRES PER SECOND

 

 

 

 

100 200 300 400 500 600 700 800 900 1000 1 100 1200
45 I I I I I I I | I I I I
— 1200
40 —
S 1 — 1100
8 /
w /
0’ /
E / — 1000
,n- 35 - / o
E / z
w / 0
u. / 8
g / — 900 m
m D:
3 / w
o 30 — X / 3,
LL /’ LL!
0 / — 800 0‘:
U) I-
D / Lu
2 / 2
3E / 9
8 25 - / — 700 g
I 9% / o
I— 0’ / E
Z 0’“ / ~
1- y“ ‘ / — 600 §
3 v9 / <4:
0 20 — / o
_l
U- I—
'5 x / — 500 <
9 / 0
<1: / n:
> x / <
”‘ 15 — X x / 5
5 x / — 400 g
l— X
< // M
(Ll; x x / EXPLANATION 5
I .—
E 10 — . // x 1914—27 300 l
I ’3: x i A 1930-32
8 ‘ / 0 1944-65
"‘ O
D x‘ ‘ X X: / — 200
§ x o ' /
L“ X A X /
O. 5 — . . V
._.____.// O
r‘ . — 100
$° '0
O
0 I I I I I I I |
0 5 1o 15 20 25 30 35 4o 45

PEAK DISCHARGE AT HEAD OF SAFFORD VALLEY (lNFLOW ), IN THOUSANDS OF CUBIC FEET PER SECOND

FIGURE 11.—Relation between inflow and outflow for peak discharges of floods moving through the study reach during 1914—27, 1930—32,
and 1944—65. The data points are for flood waves that inflow peak discharges ranging from 5,000 to 43,000 ft3/s (142 to 1,220 m3/s).
The inflow was measured at the Gila River at head of Safford Valley, near Solomon gaging station, and the outflow was measured at
the Gila River at or near San Carlos gaging stations during 1914—27 and at the Gila River at Calva gaging station during 1930—32
and 1944—65. The study reach was 71 miles (114 km) long during 1914—27 and 55 miles (88 km) long during 1930—32 and 1944—65.

 

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

stream channel was relatively narrow and had a mean-
dering pattern; the flood plain was densely vegetated
and there were alluvial fans on the flood plain at the
mouth of tributary streams. The stream channel and
flood plain was relatively stable in 1944—65 except in
times of rare floods. The study reach was 71 mi (114 km)
long during 1914—27 and 55 mi (88 km) long during
1930—32 and 1944—65. During 1914—70, only four floods
occurred that had inflow peak rates of more than 43,000
ft3/s (1,220 m3/s); these floods occurred in 1914—16 (fig.
5). The outflow peak rates for floods during 1914—27 and
1930—32 apparently were not significantly smaller than
the peak rates of inflow. Peak rates of less than about
13,000 ft3/s (368 m3/s) were reduced to bankfull dis-
charges between 3,000 and 5,000 ft3/s (85 to 142 m3/s) for
most floods that occurred in 1944—65, and the average
reduction in peak rates ranged from about 7,000 ft3/s
(198 m3/s) for floods having inflow peaks of 14,000 ft3/s
(396 m3/s) to about 4,000 ft3/s (113 m3/s) for floods hav-
ing inflow peaks of 44,000 ft3/s (1,250 m3/s) during the
same period. The increase in attenuation of peak flow
from 1914—27 to 1944—65 probably resulted from tem-
poral increases in reservoir action and infiltration. De-
creases in the amount of streamflow contributed by
tributary watersheds along the study reach may also
have caused some difference in peak outflow. Tributary
streamflow ponded behin’d natural levees during
1944—65 (Burkham, 1972); the natural levees were not
present during 1914—27.

MUSKINGUM FLOOD-ROUTING METHOD

The Muskingum flood-routing method was used in
the determination of the effects of channel changes on
floods because of its simplicity. The variable conditions
in the study reach, however, did not agree with the
conditions for which the method was developed. Innova-
tions in the Muskingum flood-routing method were
made to determine if the method could be altered to fit
the variable conditions in the study reach; the innova-
tions also give an indication of the validity of conclu-
sions reached concerning the effects of channel changes
on the timing and magnitude of floods. The basic con—
cepts of the Muskingum method given in the following
section are from a report by Carter and Godfrey (1960),
and the innovations in the method are described in the
section “Special Methods.” The term “standard Musk-
ingum method” refers to the Muskingum method with-
out innovations, and the term "modified Muskingum
method” refers to the Muskingum method with inno-
vations.

BASIC CONCEPTS

The Muskingum method is based on storage gener-
ated in a reach during an increment of time and on the

 

K17

law of continuity of mass. In equation form the law of
continuity becomes

— AS
0 —I _A_t’ (4)
in which
— = mean outflow during routing period;
T = mean inflow during routing period;
AS = net change in storage during routing period;
and

At = time unit of computation or routing period.

An expanded version of equation (4) is

At(01+02) = At(11+12) —(Sz—Sl),

2 2 (5)

where 0,], S, and At are as previously defined and the
subscripts identify the beginning and ending of routing
period At. The assumption that mean discharge is equal
to the simple arithmetic average of the flows at the end
points of the interval can be justified if the period is
equal to, or less than, the time of travel through the
reach and no abrupt changes in flow occur during the
routing period (Carter and Godfrey, 1960, p. 85).

In the Muskingum method storage is expressed as a
function of the weighted mean flow through the reach as
follows:

Storage =K [x1+(1—x)0], (6)
in which
I = inflow rate at a given time;
0 = outflow rate at a given time;
K = slope of storage-weighted discharge relation
that has the dimension of time; and
x = a dimensionless constant that weights the
inflow and outflow.
The Muskingum method of expressing storage as-

sumes that the storage varies linearly between the ‘up-
stream and downstream ends of the reach, that the
stage and discharge are uniquely defined at these two
places, and that K and x are sensibly constant through-
out the range in stage experienced by the flood wave
(Carter and Godfrey, 1960, p. 93).

Factor x.—The factor x (equation 6) is chosen so that
the indicated volume of storage is the same whether the
stage is rising or falling. For spillway discharges from a
reservoir, x may be shown to be zero because the reser-
voir stage and hence the storage are uniquely defined by
the outflow; therefore, the rate of inflow has a negligible
influence on the storage in the reservoir at any time. For
uniformly progressive flow, x equals 0.50, and the inflow
and the outflow are equal in weight. In this wave, no
change in shape takes place, and the peak discharge
remains unaffected. Thus, the value of x will range from
0 to 0.50, with a value of 0.25 as average for river

 

 

K18

reaches. No way is known for determining the value of x
from the hydraulic characteristics of a channel system
in the absence of discharge records.

Factor K.——The factor K has the dimension of time
and is the slope of the storage—weighted discharge rela-
tion. Generally, the value of K can be determined with
much greater ease and certainty than that of x. Equa-
tions (5) and (6) may be combined into

At(l-2—;—Il—O——2;01)

K=x(I2-11)+(1—x)(02—01) (7)

 

0r
__ (Kx—O.5At) (Kx+0.5At)
02‘ (K—Kx+0.5At) 2+(K—Kx+0.5At)Il+
(K -—Kx—0.5At)

(K—Kx+0.5At)01 (8)

or simplified as

02=C012+C111+0201, (9)

where At has the same meaning as in equation (4), x and
K have the same meaning as in equation (6), and

I 1, 12 = total instantaneous inflow to a reach
at the beginning of successive times

1 and 2;

01, 02 = instantaneous outflow at the begin-
ning of successive times 1 and 2; and

CO, Cl,

and 02 represent the fractions in equation (8).

The numerator in equation (7) is the storage incre-
ment, which is equal to the inflow minus the outflow,
whereas the denominator is the corresponding
weighted-flow increment. Equation (8) gives 02 in
terms of three routing coefficients and three known
discharges: 11,12, and 01. The routing coefficients may
be computed from known values of x andK . Equation (9)
reduces the routing procedure by the Muskingum
method to tabular multiplication and addition. The
time increment At between successive values of inflow
or outflow should be less than 2Kx to avoid negative
values of Co. For many flood-routing problems, the fac-
tor K may be assumed to be constant.

SPECIAL METHODS

Significant infiltration occurred during most flow
events analyzed in this study, and the factor K was
variable, which caused problems in using the Musking-
um flood-routing method because the method assumes
a continuity of mass and a constant K. The factor K is
about equal to the lag time of the center of mass of a

 

GILA RIVER PHREATOPHYTE PROJECT

definable part of a flood (Carter and Godfrey, 1960,
p. 93), which varies with discharge and time in the
study reach (fig. 2). The standard Muskingum flood-
routing method gave satisfactory results for floods that
had no significant losses of flow and that occurred when
the stream channel was wide and relatively free of vege-
tation. After a narrow stream channel and a densely
vegetated flood plain developed, however, the flow in the
stream channel during a flood event moved much faster
than the flow on the flood plain (fig. 3). Thus, in an
attempt to duplicate outflow hydrographs, two values of
K were used in the routing.

The inflow hydrographs for floods that occurred after
the development of the narrow stream channel were
divided at the bankfull discharge into two compo-
nents—overbank and within bank—before the hydro-
graphs were used in the routing computation. A value
for the K factor was assigned to each of the two flow
components. The value of K for the overbank component
was taken as the lag time of the peak discharge, and K
for the within-channel component was assumed to be
equal to the lag time of the front of the wave for the
within-channel flow.

Using equation (9), each component of the hydro-
graph was routed through y number of subreaches of
traveltime, K s=At, such that st is equal to the value of
K for each component (Carter and Godfrey, 1960). The
increment of routing time, At, between successive rates
of flow was chosen so that the inflow and outflow hydro-
graphs would be defined adequately.

The two flood components were united at the end of a
routing cycle and then separated again into overbank
and within—bank components; this was done to allow for
interchange of water between the two components. The
correct timing was assured when the components were
reunited by use of the relation

52=fl'

At=Ks= , (10)
yo yw

where At, K, and y are as previously defined and where
the subscripts o and w represent the overbank and
wit in-bank components. Equation (10) then becomes

yo _ IQ
yw Kw. (11)
Eq ations (10) and (11) show that Ko/Kw subreaches of
tra eltime for the overbank component correspond to
on subreach of traveltime for the within-bank compo-
ne t. The ratio of K0 toKw must be constant, and bothKO
an K w must be whole numbers in order for the routing
to e physically possible; the two flood-wave compo-
ne ts are reunited, and a routing cycle is thus com-

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

pleted each time these two constraints are met. For
example, assuming aKo of 27 and aKw of 12 for a flood
wave, the constant ratio Ko/Kw would be 27/12; 27 and
12 are divisible by 3, therefore, there would be three
cycles of routing. In the example four subreaches of
traveltime for the within-channel component corres-
pond in time to nine subreaches of traveltime for the
overbank component.

Depletion of flood volumes was assumed to be entirely
by infiltration. Infiltration was assumed to have oc-
curred during the rising and receding parts of a wave
until some critical discharge was reached, as when the
forces that restricted the infiltration were equal to the
forces that caused the infiltration. The critical dis-
charge probably was different for each flood, and there is
no known method to determine its value precisely. For
lack of a better method, the estimates of critical dis-
charge were based primarily on comparisons of inflow
and outflow hydrographs, as discussed further in the
section “Routing of Floods.” After critical discharge was
reached, some flow from bank storage returned to the
stream; no effort was made in the routings to account for
the return flow.

The infiltration rate during a cycle of routing time
was assumed, as an approximation, to be defined by the
relation

 

Qf=AIb, (12)
in which

qf = infiltration rate for a routing cycle;

I b = flow during routing period At at the begin—
ning of a routing cycle; and
VI—Vo . V1+V0 _ 2(V1—V0)

A _( N >( 2 )_ N(V1—Vo)’

in which

VI = volume of inflow to the study reach from the
beginning of a flood wave until the critical
inflow discharge is reached;

V0 = volume of floodflow leaving the study reach
from the beginning of a flood wave until
the critical outflow discharge is reached;
and

N = number of complete routing cycles.

The inflow I], was reduced by the quantity qf at the
beginning of each cycle of routing. The method of es-
timating critical discharge was different for each routed
flood; the methods used are described below.

ROUTING OF FLOODS

FLOODS OF JULY 14—16, 1919, AND SEPTEMBER 3—5, 1925
The last flood wave of the July 1919 flood and the

 

K19

wave of September 1925 moved through the 71-mi
(114-km) reach of the Gila River from the head of Saf-
ford Valley to the San Carlos gaging station without
large changes in peak discharge and flow distribution
with time (fig. 8). The rates of the first two waves of the
July 1919 flood, however, were greatly reduced by in-
filtration.

The standard Muskingum method with x=0.5 was
used in routing A (fig. 8) for the July 1919 and Sep-
tember 1925 floods. The time increment, At, necessary
to define the hydrograph was 1 hour for the July 1919
flood and 2 hours for the September 1925 flood. In each
routingA, the two floods were assumed to be contained
in a wide stream channel. A value of K of 17 hours was
used for the July 1919 flood, and 22 hours for the Sep-
tember 1925 flood. The values ofK were taken as the lag
time of the peak discharges. Tributary inflow contrib-
uted to the July 1919 flood, and the estimated rates of
tributary inflow are shown in figure 8. The estimated
inflow from tributaries was added to the routed flow
before the combined routed and tributary flows were
plotted.

RoutingB for the July 1919 flood was made using the
same procedure used in routingA, except that a correc-
tion was made for depletion of surface flow by infiltra-
tion (fig. 8). Apparently, the first two waves of the flood
were almost depleted by infiltration as the flood moved
through the study reach, and only minor flow depletion
occurred during the third wave; these factors were con-
sidered when depletion-of-flow corrections were made
and the critical discharge was assumed to have occurred
at the end of the second wave. The volume V1 in equation
(12) was computed using the relation

12
V]:

1:

[AL (13)

and the volume V0 was computed using the relation

29 29
V0: 2 OiAt— 2 (It)iAt, (14)
i=19 i=19

where the subscript i represents the number of routing
periods and It represents tributary inflow; 2 and 19 in
equations (13) and (14), respectively, refer to the per—
iods in which the flood arrived at the two measuring
sites. At period 29 during the outflow, the estimated
critical discharge was reached; the 12th period is about
the time when the inflow corresponding to the critical
outflow discharge occurred. The computed values of VI

K20

and V, are 52,000 and 17,600 ft3/s-periods (1,470 and
498 m3/s-periods), respectively. Using these values in
equation (12),

qf=0.05811,. (15)
The inflow II, at the beginning of each of the 17 cycles of
routing was reduced by the quantity qf before the rout-
ing was made. The infiltration correction for the first
cycle of routing was applied to the inflows starting at
period 2 and ending at period 12 (fig. 8). The beginning
and ending points in applying infiltration corrections
for the other cycles were lagged by the number of
periods in a cycle. The correction for the last cycle of
routing was applied to inflows starting at period 18 and
ending at period 28.

The outflow synthesized for the September 1925 flood
by routing A did not agree closely with the measured
outflow; however, the timing of the peak discharge was
satisfactory (fig. 8). Other routings of the September
1925 flood were made using different values of x in the
standard Muskingum method; the results of these rout-
ings did not indicate a significant improvement over
routing A and are not shown in figure 8.

Routing B of the September 1925 flood was made
using an x of 0.45, aKo of 22 hours (11 routing periods),
and a Kw of 18 hours (fig. 8). The depletion of flow by
infiltration apparently was minor, and no corrections
were applied. The bankfull discharge for the flood prob-
ably was 8,000 to 12,000 ft3/s (227 to 340 m3/s) (fig. 2);
bankfull discharge was estimated to be 12,000 ft3/s (340
m3/s), and the hydrograph was separated into compo-
nents at that rate.

FLOOD OFJULY 23—25, 1955

Routing A of the flood hydrograph was made using a
value ofK of 15 hours and a value ofx of 0.5 in the
standard Muskingum method (fig. 9). Routing A gives
an approximate indication of what the timing and shape
of the outflow hydrograph would be if the flood had
occurred when there was no flood plain and the stream
channel was wide and relatively straight. The K of 15
hours was taken from the trend line in figure 2A for an
average peak discharge of 11,000 ft3/s (312 m3/s), which
is the peak rate of flow at the inflow site. An x of 0.5 was
selected so there would be no change in the shape of the
hydrograph and no attenuation of the flow rates as a
result of reservoir action.

Other routings of the July 1955 flood were made using
different values of K and x in the standard Muskingum
method; however, the timing, peak rate, and distribu-
tion of synthesized outflow for these routings did not
match those of measured outflow satisfactorily, and the
results are not shown in figure 9.

 

 
   

GILA RIVER PHREATOPHY E PROJECT

R0 ting B was made using two values of K and dif-
feren values of x in the modified Muskingum method
(fig. 9 . For routing B, inflow was divided into overbank
and ithin-bank components using 4,000 ft3/s (113
m3/s) as bankfull discharge. The time increment At
neces ary to define the inflow and outflow hydrographs
was 1 hour. The value of K0, which was taken as the lag
time f the peak discharge, was 27 hours. The value of
Kw w s assumed to be about equal to the difference in
time rom the occurrence of 2,000 ft3/s (57 m3/s) on the
risin limb of the inflow hydrograph to the occurrence of
2,000 t3/s on the rising limb of the outflow hydrograph;
the di erence was 12 hours. The two flood wave compo-
nents were reunited at the ends of four subreaches of
travel ime for the within-bank component, which re-

is larger than measured outflow.
ing C was made using the same procedure used

ace flow (fig. 9). The flow depletion is assumed to
have resulted entirely by infiltration, and equation (12)
was us ed in the infiltration computation. The volume V;
was cc mputed using the relation

31

VI: 2 IiAt,

i=6

(16)

and the volume V0 was computed using the relation

43
V0: 2 am.

i=18

(17)

The flood arrived at the two measuring sites at periods 6
and 18, respectively. The recession of outflow reached
the estimated critical discharge of 1,200 ft3/s (34 m3/s) at
period 43; a value of 12, representing lag time, was
subtracted from 43 to estimate the period when the
inflow corresponding to the critical outflow discharge
occurred. The computed values of V1 and V0 are 97,400
and 80,120 ft3/s-periods (2,760 and 2,270 m3/s-periods),
respectively. Using these values in equation (12),
qf=0.0151b. (18)
The inflow 11, at the beginning of each of the three
cycles of routing was reduced by the quantity qf before
the routing was made. The infiltration c0rrection for the
first cycle of routing was applied to the inflows starting

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

at period 6 and ending at period 31 (fig. 9). The begin-
ning and ending points in applying infiltration correc-
tions for the other cycles were lagged by the number of
periods in a cycle. The correction for the last cycle of
routing was applied to inflows starting at period 14 and
ending at period 39.

FLOOD OFJANUARY 11—17, 1960

RoutingA of the flood hydrograph was made using a
value of K of 14 hours and a value of x of 0.5 in the
standard Muskingum method (fig. 10). The time incre-
ment At necessary to define the inflow and outflow
hydrographs was 2 hours. The K of 14 hours was taken
from the trend line in figure 2A for an average peak
discharge of 16,000 ft3/s (453 m3/s), which is the peak
rate of flow at the inflow site. RoutingA gives an approx-
imate indication of what the timing and shape of the
outflow hydrograph would have been if the flood had
occurred when there was no flood plain and the stream
channel was wide and relatively straight.

For routing B the inflow wave was divided into over-
bank and within-bank components using 4,000 ft3/s
(113 m3/s) as bankfull discharge (fig. 10). The value of K 0
was taken as 60 hours or 30 routing periods, and the
value of Kw was taken as 12 hours or 6 periods. Six
routing cycles were used in the analysis.

RoutingB for the January 1960 flood was made using
the factors described above and different values of x in
the modified Muskingum method (fig. 10). The synthe-
sized outflow was obtained using an x value of 0.4, and
the hydrograph compares favorably in shape and timing
with the hydrograph of measured outflow; however, the
synthesized outflow is larger than measured outflow
because no allowance was made for infiltration.

A correction was made for loss of surface flow in rout-
ing C of the flood wave (fig. 10). During the January
1960 flood there were no diversions of flow for irrigation
use (US. Geological Survey, 1961); therefore, the flow
depletion is assumed to have resulted entirely from
infiltration, and equation (12) was used in the infiltra-
tion computation. The volume V; was computed using
the relation

48

V1: 2 IiAt,

i=2

(19)

and the volume V0 was computed using the relation

(20)

 

K21

The recession of outflow reached the estimated critical
discharge of 4,000 ft3/s (113 m3/s) at period 8; a value of 6
representing lag time, in periods, was subtracted from 8
to estimate the period when inflow corresponding to the
critical outflow discharge occurred. The numbers 2, 8,
48, and 54 refer to routing periods, which are shown in
figure 10. The values of V1 and V0 are 311,500 and
239,800 ft3/s-periods (8,820 and 6,790 m3/s-periods), re-
spectively. Using these values in equation (12),

qf=0.0451b. (21)

The inflow I b at the beginning of each of the six cycles
of routing was reduced by the quantity q,» before the
routing was made. The infiltration correction for the
first cycle of routing was applied to the inflows starting
at period 2 and ending at period 48 (fig. 10). The begin-
ning and ending points in applying infiltration correc-
tions for the other cycles were lagged by the number of
periods in a cycle. The correction for the last cycle of
routing was applied to inflows starting at period 6 and
ending at period 52.

The critical discharge—the discharge at which the
infiltration became zero—apparently is larger than the
assumed discharge because the synthesized outflow is
less than 3,000 ft3/s (85 m3/s) when the measured out-
flow was 4,000 ft3/s (113 m3/s). If the critical discharge is
larger than the assumed discharge, the volume of infil-
tration (VI—V0) probably is larger than the computed
infiltration. The match of hydrographs of measured and
synthesized flows is good throughout most of the flood,
and the match is assumed to be satisfactory for this
study.

DISCUSSION OF RESULTS

The lack of change in inflow shapes of flood waves
during 1914—27 indicates little or no reservoir action.
Small translatory waves may have helped the flood
waves retain their inflow shape during 1914—27. A rapid
rise of a flood wave inan unvegetated ephemeral stream
often takes place through a succession of small surges
(Leopold and Miller, 1956, p. 4); an increase in stage
occurs when a surge overrides an earlier surge as the
flow progresses downstream. The result is a mechanism
shortening the rise time, which seems to be independent
of channel storage (Renard and Keppel, 1966, p. 47) and
reservoir action.

Large quantities of water infiltrated the sides and
bottom of the channel during most flow events during
1914-27; however, the effects of the flow depletion due
to infiltration on the transformation of flood waves are
not known. A reduction in flow volume may cause a
change in the shape of a flood wave as it passes through
a reach. Infiltration of water into a dry channel that is
not sealed by silt and clay generally takes place rapidly

K22

on the arrival of a flood wave. Under these channel
conditions, which may have prevailed during 1914—27,
a large part of the infiltration losses may have occurred
on the rising limb of the flood wave, which would cause a
decrease in the rise time (Renard and Keppel, 1966,
p. 39) and an increase in the sharpness of the flood
wave—a situation in opposition to the rounding effect
caused by reservoir action. The author found, however,
in channels sealed by silt and clay and having moving
boundaries, the largest part of the infiltration losses
generally occurred during the recession limb of the flood
wave (Burkham, 1970b, 1970c). The silt and clay is set
in motion during high streamflow velocities, and large
quantities of water infiltrate when the silt and clay seal
is removed. Large amounts of infiltration during reces-
sion have a tendency to increase the slope or sharpness
of the recession part of the flood.

The last outflow wave of the July 1919 flood was
similar in shape and magnitude to the third inflow
wave, and the synthesized flow for the third wave—
assuming a continuity of flow and an x value of 0.5—
agrees very well with the measured flow (fig. 8). The
synthesized outflow for the first two waves, however,
was much larger than the measured outflow. The syn-
thesized outflow—after a correction for infiltration was
made (fig. 8, curve B) using equation (12)—— resulted in
the peak rate for the first wave being too large and the
peak rate for the second wave being too small. Undoubt—
edly the infiltration for a given inflow rate was larger for
the first wave than for the second or third wave.

Considering all the factors previously discussed and
those discussed in the section “Flood Hydrographs,” the
author concludes that during the July 1919 flood (1) a
large part of the channel was dry and was not sealed
with a layer of fine sediment, and the void space availa-
ble to store infiltrated water was relatively small when
the flood arrived; (2) the infiltration rate for a given
inflow rate was relatively large during the first wave
but probably was about zero during the third wave; (3)
there was little if any return flow to the channel from
bank storage after the flow passed through the study
reach; and (4) using the standard Muskingum method,
the synthesized outflow for the third wave agrees closely
with measured outflow—however, the inflow and a close
approximate value of K were known.

The synthesized flow for the recession part of 'the
September 1925 flood, assuming a continuity of flow and
an x value of 0.5 (fig. 8, curve A), agrees well with the
measured flow; however, the synthesized flow for the
rising part of the flood was smaller and the peak rate
was larger than the measured flow (fig. 8). RoutingB of
the flood hydrograph is in closer agreement for the ris-
ing limb and peak rate than routing A; however, the
agreement for the recession limb is not as good. A closer

 

GILA RIVER PHREATOPHYTE PROJECT

agreement between synthesized and measured flows
probably could have been obtained by additional
curve-fitting adjustments in the flood-routing method,
but the final product would not contribute significantly
to this report.

For the September 1925 flood, the author concludes
that (1) infiltration was insignificant; (2) in general the
flood wave retained its inflow shape as it moved through
the study reach; (3) the reduction in peak discharge was
less than 7 percent of the inflow peak, and this reduction
is less than the probable error in the flood data; (4)
although the size of the channel and bankfull discharge
is unknown, a partly developed flood plain probably
existed at the time of the flood, and water in the stream
channel moved faster than water on the flood plain; and
(5) the synthesized outflow obtained using one value of
K and an x of 0.5 in the standard Muskingum method
agree with measured outflow just as well as the synthe-
sized flows obtained using two values of K and an x of
0.45 in the modified Muskingum methods. .

Efforts to synthesize the outflow for the flood of July
1955 by varying x and K in the standard Muskingum
method were unsuccessful. When the correct timing and
rate for the peak discharge were obtained, the rate and
distribution of flow during the rising and receding limbs
of the wave were incorrect. Adequate duplication of the
rapidly decreasing flow, which occurred after the peak
discharge, could only be achieved by using anx value of
0.5 in the routing, but then the rate and flow distribu-
tion for the rest of the wave were incorrect.

The shape and timing of the synthesized outflow for
the flood of July 1955 obtained in routingB agrees fairly
well with the measured outflow; therefore, the author
assumes that the innovation of the use of two values of K
in the routing of the 1955 flood was justified (fig. 9, curve
B). The infiltration function (eq. 18) used in making the
loss-of-fiow correction in routing C apparently does not
adequately describe the true streamflow-to-infiltration
relation because there is a significant difference in
shape and timing between synthesized and measured
outflow after the correction is applied.

The attenuation effects of reservoir action and infil-
tration of the flood of July 1955 reduced the peak dis-
charge to bankfull discharge (fig. 8); the attenuation of
the peak was about 60 percent of the inflow peak rate.
The attenuation caused by reservoir action cannot be
determined precisely for the flood because of the large
amount of infiltration. The amount of reduction in peak
flow for a wave caused by reservoir action is known to be
closely related to the volume of the wave, and therefore
the amount of reduction is closely related to the amount
of infiltration. If curve B in figure 9 is assumed to repre-
sent the outflow when continuity of flow existed, the
attenuation of the peak flow caused by the reservoir

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

action was about 7,900 ft3/s (223 m3/s), which leads to
the assumption that the infiltration in the reach during
the flood caused a reduction in peak discharge of 1,200
ft3/s (34 m3/s). Assuming that no attenuation of the peak
discharge would have occurred if a flood similar to the
July 1955 flood had passed through the reach during
1914—27, the effects of channel changes during 1927—55
on the attenuation of the peak discharge of the July
1955 flood amounted to about 7,900 ft3/s (223 m3/s), or 72
percent of the peak discharge.

The hydrograph of measured outflow for the flood of
January 1960 could not be duplicated adequately using
the standard Muskingum method. A good agreement
between measured and synthesized outflows was ob-
tained when two values of K and a correction for infil-
tration were used in the modified Muskingum method
(fig. 10). A significant rate of flow, presumably from
bank storage, returned to the stream channel during
the recession of the flood; no correction for return flow
from bank storage was made.

The attenuation of the peak discharge for the flood of
January 1960 was 5,000 ft3/s (142 m3/s). Assuming that
curve B in figure 10 represents the outflow hydrograph
if infiltration had not occurred, the amount of attenua-
tion caused by reservoir action would be about 1,500
ft3/s (42 m3/s). The infiltration in the reach during the
flood therefore caused a reduction in peak discharge of
about 3,500 ft3/s (100 m3/s), difference between the peak
discharge shown in curve B and the measured outflow
(fig. 10). As previously discussed, the attenuation in
peak flow caused by reservoir action is closely related to
the attenuation caused by infiltration, and there is no
known way of separating the two effects when infiltra-
tion occurs. The attenuation caused by reservoir action
is probably larger than 1,500 ft3/s (42 m3/s), which
makes the attenuation caused by infiltration smaller
than 3,500 ft3/s (100 m3/s). The difference between the
volumes of the floods of January 1960 and July 1955
probably is the main reason the attenuation of the
January 1960 flood was only 30 percent, whereas the
attenuation of the July 1955 flood was more than 60
percent. The effects of channel changes during 1914—70
on the reduction of the peak discharge of the flood of
January 1960 are assumed to have been more than
1,500 ft3/s (42 m3/s).

SUMMARY AND CONCLUSIONS

The channel changes in the Gila River during 1914—
70 caused significant differences in the timing, mag-
nitude, and transformation of flood waves in the 55-mi
(88-krn) reach of the Gila River in Safford Valley. The
channel changes consisted of (1) narrowing of the
stream channel from about 2,000 ft (600 m) to less than
300 ft (90 m), (2) development of a flood plain, stream-

 

K23

channel meander pattern, natural levees along the
stream channel, and alluvial fans at the mouths of
tributaries, and (3) spreading of dense saltcedar along
the flood plain. Except for small flood waves having
peak discharges less than about 500 ft3/s (14 m3/s), the
timing and velocity of all the waves were affected by the
channel changes. For 1914—27, the trend was toward a
gradual increase in downstream velocity of the center of
mass of flood waves as the peak discharge increased,
which is indicative of flow in wide channels where the
resistance to water movement is mainly along the bot-
tom. During 1943—70, the trend was toward an increase
in downstream velocity as the peak discharge increased
from about 500 to about 4,000 ft3/s (14 to 113 m3/s), a
decrease in velocity as the peak discharge increased
from about 4,000 to 20,000 ft3/s (113 to 566 m3/s), and an
increase in velocity for a peak discharge greater than
20,000 ft3/s (566 m3/s) (table 2). Major floods that oc-
curred when the stream channel was fully developed
and the flood plain was densely covered with saltcedar
partially cleared the channel and caused a reduction in
lag time and an increase in velocity of subsequent
floods.

The channel changes in the Gila River during 1914—
70 caused an increase in the attenuation of flood waves
as they moved through the study reach. During 1914—
27, when the channel was wide and relatively free of
vegetation, flood waves moved through the reach with-
out large changes in inflow shapes; however, significant
reductions in peak rates occurred when the infiltration
was relatively large. After the stream channel de-
veloped and the flood plain became densely vegetated,
flow in the stream channel moved at a much higher
velocity than flow on the flood plain, which resulted in
an elongation of floodwaves and a reduction in peak
discharge. For flashy floods (floods that had large inflow
peaks and small inflow volumes) the combined attenua—
tion effects of reservoir action and infiltration reduced
the peak flow to bankfull discharge—about 4,000 ft3/s
(113 m3/s) at the downstream end of the study reach.

The standard Muskingum flood-routing method gave
satisfactory results for floods that occurred when the

TABLE 2.——Vel0city of the center of mass of flood waves and approximate
values of Manning n for selected peak discharges

 

Period of record
used in analysis Peak dischargel
(water year) (fta/s)

Velocity of center of
mass of floodwave2
Approximate n value3

 

1914—27
1961—70
1914—27
1961—70
1914.27
1961—70

300—500

2 002—004
2
3,000-5,000 4
6
5
1

1

4 0.02—0.04

7 0.02—0.04

9 0.02~0.04

15,000—20,000 7 002—004
3 0.06—0.14

 

IAverage of peak discharges at the ends of the reach.

2Downstream velocity was determined by dividing the length of main flow path by lag time
of the center of mass of flood wave, which was obtained from trend lines in figure 2.

“The Manning velocity equation is V = 1.49 (R WSW, in whichR is hydraulic radius, S is
slope of energy gradient, and n is a roughness coefficient.

K24

stream channel was wide and free of vegetation and that

had no significant losses of flow; however, the standard

Muskingum method was not adequate for the routing of

floods that occurred after a narrow stream channel and

a densely vegetated flood plain developed. Innovations

in the Muskingum flood-routing method were made to

fit the variable conditions in the study reach. The
hydrographs for floods that occurred after the develop-
ment of the narrow stream channel were divided into
overbank and within-bank components before they
were routed. The overbank component consists of flow
greater than bankfull discharge, and the within-bank
component consists of flows less than bankfull dis-
charge. A factor, K, was assigned to each of the flow
components. Using the standard Muskingum equation,
each component of the hydrograph was routed through
several subreaches of traveltime. Owing to infiltration,
adjustments in the flood-routing method were made on
the assumption that the infiltration rate was linearly
related to the inflow rate for each of the subreaches of
traveltime. Infiltration is assumed to have occurred
during the rising and recession parts of a wave until the
forces that restricted the infiltration were equal to the
forces that caused the infiltration. The critical dis-
charge was estimated. The innovation in flood routing
was partly successful in that there was fair agreement
between hydrographs of measured and synthesized out-
flows.

The conclusions reached as a result of this study are
as follows:

1. The size and meander pattern of the stream channel
of the Gila River are determined by past dominant
flows. The stream channel is wide and straight at
the end of a period in which high flows were domi-
nant and is narrow and has a meander pattern at
the end of a period in which low flows were domi-
nant.

2. The stream-channel and flood-plain system, when
fully developed for a dominant flow, has a persis-
tent effect on floods. A low-flow system—developed
by and for low flows—attenuates flood peaks pass-
ing through the reach; the peak flows of flashy
floods may be reduced to bankfull discharge. A
high-flow system—developed by and for high
flows—~does not increase flood rates; however,
streamflow from side tributaries along the study
reach may contribute more significantly to peak
rates in the Gila River when a high-flow system is
in effect than when a low-flow system is in effect.

3. The downstream velocity of the center of mass of
flood waves that had peak discharges of between
10,000 to 20,000 ft3/s (283 to 566 m3/s) during
1914—27 may have been as much as three times
that for the same rates during 1943—70.

 

GILA RIVER PHREATOPHYTE PROJECT

4. A low-flow system may change rapidly to a high-flow
system when a series of major floods occurs; how-
ever, several years of low flow are required before a
high-flow system changes to a low-flow system; it
took about 50 years for the present (1970) low-flow
system to develop (Burkham, 1972).

. Annual peak flows measured at the downstream end
of the study reach reflect, among other things, the
persistent effect of the upstream system, and
therefore they are not random in time. Because of
changes in the system, the data of peak flows col—
lected at the downstream end of the study reach
during 1914—27 are from a different population
than the data ofpeak flows for the period 1943-70.

6. The stream channel of the Gila River can be widened
and straightened in an attempt to duplicate a
high—flow system; however, it will be difficult to
maintain the artificial channel unless large flows
occur. Conversely, it would be difficult to develop
and maintain a low-flow system during a period in
which large fl0ws are dominant.

7. Widening and straightening the stream channel will
increase the conveyance capacity of the Gila River;
however, the widening and straightening of the
channel may increase flood rates at the down—
stream end of the valley (See conclusions 2 and 6.)

8. Outflow rates for flood waves moving through the
study reach when a high-flow system is in effect
can be synthesized using the standard Musking-
um method if an inflow hydrograph and an approx—
imate value for Muskingum’sK are available. The
standard Muskingum method, however, is not
suitable for the routing of flood waves—except
possibly for extremely small or large waves—that
occur when a low—flow system is in effect.

01

REFERENCES CITED

Barnes, H. H., Jr., 1967, Roughness characteristics of natural chan-
nels: U.S. Geol. Survey Water-Supply Paper 1849, 213 p.

Bryan, Kirk, 1926, Pedestal rocks formed by differential erosion and
channel erosion of the Rio Salado, Socorro County, New Mexico:
U.S. Geol. Survey Bull. 790-A, 19 p.

Burkham, D. E., 1970a, Precipitation, streamflow, and major floods at
selected sites in the Gila River drainage basin above Coolidge
Dam, Arizona: US. Geol. Survey Prof. Paper 655—B, 33 p.

1970b, Depletion of streamflow by infiltration in the main

channels of the Tucson basin, southeastern Arizona: US. Geol.

Survey Water-Supply Paper 1939—B, 36 p.

1970c, A method for relating infiltration rates to streamflow

rates in perched streams, in Geological Survey research 1970:

US. Geol. Survey Prof. Paper 700—D, p. D266—D271.

1972, Channel changes of the Gila River in Safford Valley,

Arizona, 1846—1970: US. Geol. Survey Prof. Paper 655—G, 24 p.

 

 

 

EFFECTS OF CHANGES OF FLOOD WAVES, SOUTHEASTERN ARIZONA

 

1976, Hydraulic effects of changes in bottom-land vegetation
on three major floods, Gila River in southeastern Arizona: US.
Geol. Survey Prof. Paper 655—J, 14 p.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis ofstreamflow
data for an alluvial stream: U.S. Geol. Survey Prof. Paper 655—0,
13 p.

Carter, R. W., and Godfrey, R. G., 1960, Storage and flood routing:
U.S. Geol. Survey Water-Supply Paper 1543—8, p. 81—104.
Culler, R. C., and others, 1970, Objectives, methods, and
environment—Gila River Phreatophyte Project, Graham

County, Arizona: US. Geol. Survey Prof. Paper 655—A, 25 p.

Dalrymple, Tate, and Benson, M. A., 1967, Measurement of peak
discharge by the slope-area method: U.S. Geol. Survey
Techniques Water-Resources Inv., book 3, chap. A—2, 12 p.

Gatewood, J. S., Robinson, T. W., Colby, B. R., Hem, J. D., and
Halpenny, L. C., 1950, Use of water by bottom-land vegetation in
lower Safford Valley, Arizona: US. Geol. Survey Water-Supply
Paper 1103, 210 p.

Leopold, L. B., and Miller, J. P., 1956, Ephemeral streams—Hydraulic
factors and their relation to the drainage net: U.S. Geol. Survey
Prof. Paper 282—A, 37 p.

 

K25

Linsley, R. K., Kohler, M. A., and Paulhus, J. L. H., 1949, Applied
Hydrology: New York, McGraW—Hill Book Co., 689 p.

Olmstead, F. H., 1919, Gila River flood control—A report on flood
control of the Gila River in Graham County, Arizona: US. 65th
Cong, 3d sess., Senate Doc. 436, 94 p.

Patterson, J. L., and Somers, W. P., 1966, Magnitude and frequency of
floods in the United States—Part 9, Colorado River basin: U.S.
Geol. Survey Water-Supply Paper 1683, 475 p.

Renard, K. G., and Keppel, R. V., 1966, Hydrographs of ephemeral
streams in the Southwest: Am. Soc. Civil Engineers Proc., Hy-
draulics Div. Jour., v. 92, no. HY2, p. 33—52.

Rubey, W. W., 1937, The force required to move particles on a stream
bed: U.S. Geol. Survey Prof. Paper 189—E, p. 121—141.

Schumm, S. A., and Lichty, R. W., 1963, Channel widening and
flood-plain construction along Cimarron River in southwestern
Kansas: U.S. Geol. Survey Prof. Paper 352—D, p. D71—D88.

Sellers, W. D., ed., 1960, Arizona climate: Tucson, Arizona Univ.
Press, 60 p.

US. Geological Survey, 1961, Surface water supply of the United
States, 1960—Part 9, Colorado River basin: U.S. Geol. Survey
Water-Supply Paper 1713, 520 p.

fiGPo 691- 088—1976

 

 

 

 

 

 

 

 

Accuracy of Evapotranspiration Rates
Determined by the Water-Budget
Method, Gila River Flood Plain,

Southeastern Arizona

 

 

GEOLOGICAL SURVEY PROFESSIONAL‘PAPER 655—L

 

 

 
  

f‘ t
D- If. man-p Dtnn 1‘?”er
l

/‘ JUL 16 1976

g mwtm“ I
.- M331? U5 Cfitum‘mn .

129197?

.1 n 3,;

 

 

Accuracy of Evapotranspiration Rates
Determined by the Water-Budget
Method, Gila River Flood Plain,

Southeastern Arizona

By Ronald L. Hanson and David R. Dawdy

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—L

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1976

UNITED STATES DEPARTMENT OF THE INTERIOR
THOMAS S. KLEPPE, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Director

Library of Congress Cataloging in Publication Data

Hanson, Ronald Lee, 1934-
Accuracy of evapotranspiration rates determined by the water-budget method, Gila River flood plain, southeastern Arizona. (Gila
River phreatophyte project)

(Geological Survey Professional Paper 655—L)
Bibliography: p. 35
Supt. of Docs.: [ 19.162655-L

l. Evapotranspiration—Alaska—Gila River watershed—Measurement. I. Dawdy, David R., 1926— joint author. 11. Title.
111. Series. IV. Series: United States Geological Survey Professional Paper 655—L.
QE75.P9 no. 655—L[QC915.7.U5] 557.3'088[551.5'72'097917] 75-619341

 

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, DC. 20402
Stock Number 024-001—01819-1

CONTENTS

Pa e Page

Symbols __________________________________________________ 1:] Evaluation of water-budget components and errors—Continued
Metric conversion factors __________________________________ V Precipitation —————————————————————————————————————————— L10
Abstract __________________________________________________ L1 Soil-moisture content __________________________________ 15
Introduction ______________________________________________ 1 Basin-fill inflow ________________________________________ 23
Description of study area __________________________________ 2 Downvalley ground-water flow __________________________ 23
The water-budget equation ________________________________ 3 Computation of ET and total error in ET ____________________ 25
Evaluation of water-budget components and errors __________ 4 Computation of A I? and error in A W ____________________ 27
Streamflow ____________________________________________ 4 Discussion of results _______‘_ ______________________________ 33

Channel storage ______________________________________ 7 Development of equations deScribing unadjusted sampling
Tributary inflow ______________________________________ 9 error, V ________________________________________________ 34
References cited ____________________________________________ 35

ILLUSTRATIONS
Page
FIGURE 1. Map showing study area and instrumentation location ________________________________________________________ L3
2—17. Graphs showing:

2. Average relation between instantaneous discharge and error in discharge at cross sections 1 and 9 ________ 6
3. Relation between average daily discharge and wetted cross~sectional area for cross sections 1, 9, 13, and 17 8
4. Relation between average daily discharge and average error in wetted cross-sectional area ________________ 9
5. Average relation between number of precipitation gages and missing-data error of average precipitation _- 13
6. Relation between average precipitation and total adjusted sampling error ________________________________ 14
7. Average relation between number of access holes and missing—data error of moisture-content change ______ 18
8. Average relation between number of access holes and total adjusted sampling error of moisture-content change 19

9. Estimates of average relation between sampling density of access holes and total adjusted sampling error of
moisture—content change __________________________________________________________________________ 20

10. Apparent specific yield values derived from relation between average change in water level and average
moisture-content change in capillary zones ________________________________________________________ 21
11. Average relation between number of ground-water wells and missing-data error in moisture-content change ‘23
12. Average downvalley ground-water slopes through cross section 9 in winter months ______________________ 24

13. Relation between average downvalley ground-water slope through ends of each reach and width of flood plain at
downstream cross section ________________________________________________________________________ 25
14. Sources of water contributing to ET for the 1964 water year, reach 1 ____________________________________ 26
15. ET and corresponding errors in ET and outflow of Gila River at cross section 9 for 1964 water year ________ 28

16. Water-budget ET values in reach 1 before and after clearing and average change in ET as a result of
phreatophyte clearing ____________________________________________________________________________ 32

17. General relation for error in departure of average of m samples from average of n samples and error in departure
of average of m samples from estimated population mean __________________________________________ 34

TABLES

Page
TABLE 1. Evapotranspiration and water-budget components for 1964 water year, reach 1 __________________________________ L5

2. Evapotranspiration and water-budget components for the 21-day budget period 688-708 (August 18—September 7, 1964)
in reach 1 and corresponding sampling and bias errors of each component __________________________________ 5
3. Daily discharges and errors atcross section 1 for budget period 688—708 ________________________________________ 5

4. Average daily discharge (q), corresponding wetted cross-sectional area of river channel (A), and error in the area (E A) for
cross sections 1 and 9 on budget period days 688 and 708 __________________________________________________ 8
5. Tributary runoff into reach 1 during budget period 688—708 ____________________________________________________ 10

6. Area assigned to each precipitation gage in reach 1 and precipitation amounts observed at each gage during budget
period 688—708 __________________________________________________________________________________________ 10

III

IV

TABLE

>9

0‘

C

9<l>0°to

CONTENTS

Page

7. Example of precipitation data array used to compute the average precipitation for each budget period (E), the departure
of precipitation at a given gage from the average precipitation (R fl), the average departure for all budget periods

. (E), and the standard deviation of the average departure (sj) ______________________________________________ 11
8. The average departure in precipitation of each gage in reach 1 (Rj) and the standard deviation of the average departures
(Sj) computed for three precipitation ranges with each range containing k budget periods of data ______________ 12

9. Example of array of mean departures of precipitation (173m) and standard deviations of the mean departures (S m) 12

10. En“, and S ",0 values for precipitation in upper range P2151 in. (38.4 cm) in reach 1 with gages arranged in order of
increasing 3,- and in orde_r of decreasing sj ________________________________________________________________ 13

1 1. Average missing-data error (Sm) of precipitation in each precipitation range from curves in figure 5 for m = 1 to 10 gages
in reach 1. Included are the unadjusted (V) and adjusted (E15) sampling errors for m=10 gages ______________ 14

12. Soil moisture content measured in the soil and capillary zones of flood plain access holes in reach 1 for budget period days
688 and 708 ____________________________________________________________________________________________ 16

13. Average departure in moisture of each access hole (NJ) and the standard deviation of the average departure ($1) for the
soil and capillary zones of the flood plain and the soil, intermediate, and capillary zones of the terrace in reach 1 17

14. Missing-data errors (5,") from curve in figure 7 for the soil and capillary zones of the flood plain and for the soil,

intermediate, and capillary zones of the terrace, reach 1 __________________________________________________ 19
1 5. Total adjusted sampling error in moisture change as defined with a complete set of moisture-content data (m =n) for each
zone of the flood plain and terrace in reaches 1, 2, 2a, and 3 ________________________________________________ 20

16. Average apparent specific yield (S ’), number of budget periods (k) used to define S’, total adjusted sampling error in the
measurement of average water-level change (E A5), and the total adjusted sampling error of moisture change in the
capillary zone (ECAh and ETC AH): when the moisture change is derived from water-level change (equation 34) __ 22

17. Ground—water level changes (Ah) in flood plain wells of reach 1 used to compute We for budget period 688—708 _- 22

18. Average departure in ground-water level change (17]) of each terrace well in reach 1 and the standard deviation (3) of T? j
computed from 20 budget periods of water-level data collected during the 1968 water year __________________ 23

19. Evapotranspiration (ET) and total measurement error of ET (EET) for each budget period during water years 1963—
71—reaches 1, 2, 2a, and 3 ______________________________________________________________________________ 29

20. ET values obtained defore and after clearing and their corresponding sampling, and total measurement errors for
selected 14-day budget periods during June and July, reach 1 ______________________________________________ 32

SYMBOLS

constant defined by equation 23

area of reach, surface area assigned to a sample point,
wetted cross-sectional area of Gila River channel, or
subscript denoting "after clearing”

constant defined by equation 24 or subscript denoting
a bias type of error

subscript denoting “basin fill” or “before clearing”

constant defined by equation 25

subscript denoting "capillary zone of soil profile”

change in Gila River channelstorage

subscript denoting a given budget period day

total number of days in budget period

temporal variability of Gila River wetted cross-
sectional area

spatial variability of Gila River wetted cross-sectional
area

error in Gila River wetted cross-sectional area due to
error in discharge used to compute area

error in Gila River instantaneous (or average daily)
discharge

average standard error of mean daily discharge of the
Gila River for budget period

bias error, adjusted sampling error, or total measure-
ment error of water-budget component or parameter
indicated by subscript x

evapotranspiration, average evapotranspiration

average change in evapotranspiration

basin-fill inflow

downvalley ground-water inflow and outflow

ground-water level change, average ground-water
level change

 

k.

k
L

. m

szt i

AM,,—,, AM

:13

downvalley ground—water slope

subscript denoting “inflow” or “intermediate zone of
soil profile”

sample point—precipitation gage, soil-moisture ac-
cess hole, or ground-water well

total number of observations (budget periods)

length of Gila River channel

number of sample points S n

soil-moisture content in zone 2 of hole j at the end of
budget period t

change in soil-moisture content in zonez of hole j at the
end of budget period t, and average change in
soil—moisture content

total number of sample points in reach

largest even value s n

number of discharge measurements made during the
budget period

subscript denoting “outflow”

total number of permutations

precipitation and average precipitation

instantaneous or average daily discharge

Gila River inflow and outflow during a budget period

tributary inflow

approximate correlation coefficient

departure of observation at sample point j from
average ‘of observations at n sample points for
budget period t

average departure of k observations at sample point j

mean departure of observations at m sample points
from observations at n sample points for permuta-
tion v

' CONTENTS

3 subscript denoting a sampling type of error TC, TI, TS
sjl standard deviation of R,-
SET standard deviation of average ET v
8457‘" standard deviation of average change in ET
S subscript denoting “soil zone of soil profile” W
8””, standard deviation of Rm, 2
gm, average standard deviation of all Rm, computed from p e

permutations (missing-data error) E
S’ apparent specific yield p.11}.2
t subscript denoting a given budget period ‘
T transmissivity or subscript denoting “tributary” V

 

V

subscripts denoting capillary,” “intermediate,” and
“soil” zones of terrace soil profile

subscript denoting a given permutation of the sample
points

width of saturated alluvium

subscript denoting a given soil—moisture zone

expectation

residual error in ratio of E2," to Egg,"

estimate of population correlation coefficient between
sample points j 1 and jz

unadjusted sampling error for a complete set of sample
points

METRIC-ENGLISH EQUIVALENTS

 

Metric unit English equivalent

Metric unit

English equivalent

 

 

 

 

 

 

 

 

 

 

 

 

 

Length Specific combinations—Continued
millimetre (mm) = 0.03937 inch (in) litre per second (l/s) = .0353 cubic foot per second
metre (m) : 3.2 feet (ft) cubic metre per second .
kilometre (km) : .62 mile (ml) per square kilometre
[(ma/s)/km9] = 91.47 cubic feet per secondp
Area square mile [(fts/s)/mi=]

' metre per day (m/d) 3.28 feet per day (hydraulic
square metre (m2) : 10.76 square feet (ft?) CODGUCflVltY) (it/d)
square kilometre (km?) = .386 square mile (mifl) metre per kilometre _
hectare (ha) = 2.47 acres m m = L28 feet per mile (ft/ml)

kilzilr‘netli‘ci per hour 9113 f t d f
in = . 00 per secon (t/s)
Volume metre per seciond (an/s) = 3.28 feet per second
m3 : . a me re square per ay
m2}; (“gunman (c ) Z 62.831 333% $311184“) (ma/d) = 10.764 feet squared per day (ftg/d)
cubic metre (m3) = 35.31 cubic feet (ttn) ("“8”“qu
cubic metre : .00081 acre-foot (acre-ft) cubicametre per second
cubic hectometre (bm=) 2810.7 acre-feet (m /s) = 22-826 million gallons per day
litre = 2.113 pints (pt) (Mgal/d)
litre = 1.06 quarts (qt) cubic metre per minute
litre = .26 gallon (gal) (ms/min) :264.2 gallons per minute (gal/min)
cubic metre : 00026 million gallons (Mgal or litre per second (l/S) — 15-85 gallons per minute
106 ga litre per second per
cubic metre : 6.290 barrels (bbl) (1 bb1=42 gal) metre [(l/s)/ml : 4.83 gallons per minute per foot
[(gal/min)/ft]
Weight kilometre per hour
(tkm/h) d ( / ) = 2.337 mile per hqlur (mi/h)
gram (g) : 0.035 ounce. avoirdupois (oz avdp) me re per secon m S = ' m es per our
gram = .0022 pound, avoirdupois (lb avdp) gram per cubic 1 _ , a
tonne (t) : 1.1 tons, short (2 000 lb) centimetre (g/cm—) _ 62.43 pounds per cubic foot (lb/ft)
tonne = .98 ton, long (2, 240 lb) gram per square
centimetre (g/cm'-') 2 2.04.3 pounds per square foot (lb/ft”)
ifi gram per square
Spec c combinations centimetre = .0142 pound per square inch (lb/in?)
kilogram per square
centimetre (kg/c1112) — 0.96 atmosphere (atm) Temperature
kilogram per square
centimetre = .98 bar (0.9869 atm) degree Celsius ('C) = 1.8 degrees Fahrenheit (T)
cubic metre per second _ degrees Celsius
(ma/s) : 30.3 cubic feet per second (ft3/s) (temperature) =[(1.SX ”C) +32] degrees Fahrenheit

 

 

GILA RIVER PHREATOPHYTE PROJECT

ACCURACY OF EVAPOTRANSPIRATION RATES
DETERMINED BY THE WATER-BUDGET METHOD,

GILA RIVER FLOOD PLAIN,

SOUTHEASTERN ARIZONA

By RONALD L. HANSON and DAVID R. DAWDY

ABSTRACT

Evapotranspiration by phreatophytes (primarily saltcedar) was
determined by the water—budget method for 5,500 acres (2,230 ha) of
the Gila River flood plain in southeastern Arizona. The water budget
consists of 12 components including surface and subsurface flow
through the study area, precipitation on the area, and soil-moisture
changes in the unsaturated soil profile.

Nine years (1963—71) of hydrologic data were collected on four
reaches within the area. These data provided over 400 measurements
of evapotranspiration for two- or three-week periods. Midway through
the study the vegetation was removed from the flood plain. The
evapotranspiration measurements are therefore defined for both
natural vegetative cover and essentially bare-ground conditions.

This report shows how each component of the water budget was
evaluated, demonstrates the significance of each component in
relation to the total evapotranspiration, and describes the methods
used to evaluate the relative accuracy of each component.

The two most significant components of the water budget are,
generally, the Gila River inflow and outflow. One of the least
significant is tributary inflow, which occurred only 4 percent of the
time during the 9-year study. Soil-moisture change is highly
significant during periods of low streamflow and is one of the more
difficult components to measure. The ground-water flow components
are the least variable in the water budget, fluctuating only in response
to seasonal changes in the downvalley ground-water slope.

The total measurement error of each component consists primarily
of a sampling error which is dependent on the number of observation
points used to measure the component. This error is time variant,
reflecting both the variability in repetitive measurements and the
error due to missing data. Included in the total measurement error is
a bias error which gives a constant overestimate or underestimate of
the component. Only the ground-water flow components introduce
a measurable bias error, but the direction of this error is unknown
and its magnitude in relation to evapotranspiration is relatively
insignificant.

The total measurement error in evapotranspiration is not related to
the magnitude of evapotranspiration but rather to the total volume of
water moving through the area. Thus, the minimum errors occur
during the midsummer months of maximum evapotranspiration
when streamflow is low and precipitation is negligible.

Evapotranspiration rates computed for reach 1 indicate that
phreatophyte clearing reduced summer rates by nearly 45 percent.
The average computed measurement errors in summer evapotranspi-
ration rates, before and after clearing, are :59 percent and :113
percent, respectively, and the average measurement error in the

 

change in summer evapotranspiration as a result of clearing is nearly
:200 percent. These large computed measurement errors are shown
to overestimate substantially the true measurement variable in
evapotranspiration. The computed errors do give, however, a good
indication of the relative significance of each evapotranspiration
value and provide a means of selecting those values which should be
used in computing average evapotranspiration rates. Furthermore,
the results of this error analysis show that reliable estimates of
summer evapotranspiration can be determined and that a significant
difference in summer evapotranspiration could be detected as a result
of clearing phreatophytes from the flood plain.

INTRODUCTION

The determination of ET (evapotranspiration) by
phreatophytes from a flood plain by the water-budget
method requires that all significant movement of liquid
water into and out of the area be measured.

Components of the water budget include surface and
subsurface flow through the area, precipitation on the
area, and soil-moisture content changes in the unsatu-
rated zone of the area. It has been the general opinion of
most researchers that the measurement errors as-
sociated with these components are too large to provide
reliable estimates of ET—particularly when an esti-
mate of water salvage as a result of phreatophyte
removal is desired. To date (1976) few studies have been
conducted which evaluated ET from a large area
(Gatewood and others, 1950; Turner and Skibitzke,
1952; Bowie and Kam, 1968), and little is known about
the accuracy of these evaluations. In October 1962 a
nine-year water-budget study began on 5,500 acres
(2,230 ha) of the Gila River flood plain in southeastern
Arizona. The primary objective of this study was to
evaluate seasonal ET rates of phreatophytes from the
area and water salvage following removal of the
phreatophytes (Culler and others, 1970).

The study was designed and instrumented to measure
independently each component of the water budget. The

L1

L2

purposes of this report are to show how each component
of the water budget was derived, to describe the method
used in estimating the accuracy of each component, and
to demonstrate the significance of each component and
its error in the resulting ET values. The errors for most
of the water-budget components are expressed in terms
of the standard error of their measurement. For some
components, however, the measurement error can only
be approximated. The assumptions and criteria used to
arrive at these approximations were defined to provide a
resultant error which, in most instances, should exceed
the expected standard error in the measurement of the
component. The derived error of each ET value is
therefore considered to be only a relative indicator of
measurement variability in ET.

Two types of error were investigated in this anal-
ysis—the bias error and the sampling error. The bias
error is a constant time-invariant error caused by
consistent overestimates or underestimates of the true
value of the component. When evaluating the average
change in evapotranspiration (AET)—such as occurs
after clearing phreatophytes from the flood plain—this
error cancels and is thus not included in the determina-
tion of the accuracy of AE T . However, when evaluating
absolute values of ET, the bias error may, in some
instances, be a significant part of the total error in ET.

The sampling error reflects the variability in the
measurement of a water-budget component due to
insufficient sampling of the component. This error
decreases with an increase in the number of observa-
tions at the sample point or with an increase in the
number of sample points, and it increases with an
increase in the magnitude of the component. The
sampling error is time dependent for those components
in which the number of sampling points and the
magnitude changes during the study period.

The total measurement error of each ET value was
obtained from the sum of squares of the bias and
sampling errors defined for each component. Because of
independence of the components no covariance term
exists in the computation of the total measurement
error. Furthermore, this total error term is considered
to be only a relative indicator of the measurement
variability in ET and not an estimate of the expected
standard error in ET.

This study was conducted under the general super-
vision of R. C. Culler, project chief of the Gila River
Phreatophyte Project. Transformation of basic field
data into a form acceptable for analysis was performed
by R. .M. Myrick and F. P. Kipple. The authors are
indebted to the San Carlos Apache Indian Tribe and the
Bureau of Indian Affairs for the use of their lands and
facilities, respectively, to make this study.

 

GILA RIVER PHREATOPHYTE PROJECT

DESCRIPTION OF STUDY AREA

The study area includes a 15-mile (24 km) length of
the Gila River flood plain above San Carlos Reservoir in
southeastern Arizona (fig. 1). The flood plain averages 1
mile (1.6 km) in ,width and has a gradual downvalley
slope of about 1.5 ft per 1,000 ft. The water-bearing
deposits in the flood plain consist of basin-fill deposits
and alluvial deposits which overlie the basin fill. The
basin-fill deposits are more than 1,000 ft (300 m) thick
and consist of fine-grained material of low permeability.
The alluvial deposits are as much as 60 ft (20 m) thick
and consist of lenticular gravel, sand, and silt beds with
a relatively high permeability. The alluvial deposits
form the flood plain and lower terraces in the study area.
The Gila River meanders across the alluvium in a
channel averaging 110 ft (35 m) wide and 7 ft (2 m) deep.
A detailed description of the geology of the study area is
given by Weist (1971).

The depth to ground water ranges from about 5 ft (ll/2
m) near the river to more than 20 ft (6 m) near the outer
boundaries of the flood plain. Wells that penetrate
through the flood-plain alluvium into the underlying
basin fill indicate that ground water in the basin fill
flows vertically upward into the alluvium at a rate of
about 0.3 ft (0.09 m) per year or 0.1 million gallons per
day per acre (0.01 m3/s/ha). Downvalley ground-water
movement through the alluvium averages about 5.1
acre-ft per day (0.0063 hm3/day) (Hanson, 1972). This
downvalley flow is equivalent to 1.7 million gallons per
day (180,000 m3/s).

Gila River inflow to the study area is derived from
11,500 mi2 (29,800 km2) of drainage area. Most of the
streamflow results from winter and late summer
precipitation. The average discharge of the Gila River is
250 ft3/s (7.1 m3/s) but can range from no flow for a few
days in the summer to several thousand cubic feet per
second during the winter and summer storm periods.

Tributary inflow is derived from 225 mi2 (583 km2) of
drainage area bordering the study area. Annual runoff
from these tributaries is small and generally occurs

only for short periods during the summer as a result of

thunderstorms.

Precipitation occurs primarily in December and
January from large frontal storms and in July, August,
and September from short—duration high-intensity
convective storms. The average annual precipitation at
San Carlos Reservoir is 14 in. (360 mm), but annual
totals have ranged from 4.0 in. (102 mm)vto 25.8 in. (655
mm) during the period of record, 1882—1973.

Mean daily temperatures in the study area, range
from a minimum of 32°F (0°C) during the winter to a
maximum of 100°F (38°C) in the summer. Pan evapora—
tion at San Carlos Reservoir averages 97 in. (2,460 mm)

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

per year, and evaporation from the reservoir is
estimated at 70 in. (1,780 mm) per year (F. P. Kipple,
written commun., 1975).

When the study began (October 1962) nearly 70
percent of the vegetation on the flood plain was
saltcedar (Tamarix chinensis)1, with the remaining
vegetation consisting of, mesquite (Prosopis), willow
(Salix), cottonwood (Populus), seepwillow (Baccharis),
and seepweed (Suaeda). Vegetation was removed from
the flood plain in segments of several acres at different
periods during 1967—71. Virtually all the study area
was cleared of vegetation by March 1971.

THE WATER-BUDGET EQUATION
Twelve components are significant in defining ET

‘Also referred to as Tamarix pentandra and Tamaer gallical

EXPLANATION

 

Boundary of Gila River flood plain

..._.®__

Cross section and number

A
Gaging station on the Gila River

    
    
 

(

 

SAN CARLOS
RESER VOIR

 

2-inch access pipe for
V soil-moisture meter

xx: RIVER
k,—

GILA x 4-inch observation well

x)! . . .
..————-Prec1p1tat10n gage

O
l
I

0

DETAIL OF TYPICAL CROSS SECTION

 

 

 

 

L3

from the water budget of the Gila River flood plain. An
equation expressing these components is
ET=QI—QO+QT +AC +F+AIWS+

AMI+AMC+GB +GI—GO+AZTJTC, (1)

where

ET : evapotranspiration from the area,

Q] = surface inflow of the Gila River,

Q0 = surface outflow of the Gila River,

QT = surface inflow from tributaries bordering
the area,

AC = change in channel storage in the Gila
River,

13 = average precipitation on the area,

AM S = average change in moisture content in the

unsaturated soil zone located im-
mediately below the land surface,

110°

 

 

 

 

 

 

F """ _T ' " ‘T‘n—l
at l
\L‘“. A fill/K. J 35°
» V I
‘x i
\ ARIZONA .
I
°°> | o
N 9 RIVER I 34
g/ Phoenix SALT J ll
GILA RIVER 35551 I
’ STUDY AREA V11
Tucsono |
\\\ | 32°
\
\ \ \ \— .__..Ji.————|

FIGURE 1.—Map showing study area and instrumentation location.

L4

average change in moisture content in the
unsaturated intermediate zone located
between the overlying soil zone and the
underlying capillary zone,

average change in moisture content in the
capillary zone located below the inter-
mediate zone and within the zone of
water-table fluctuation,

ground-water inflow vertically upward
into the alluvium from the underlying
basin fill,

ground-water inflow downvalley through
the saturated alluvium,

ground-water outflow downvalley
through the saturated alluvium,

average lateral ground-water movement
through the capillary zone between the
flood plain and the adjacent terrace area.

mm:

The basic data used to compute each component of the
water budget were collected at the 13 cross sections
shown in figure 1. Each cross section included three
ground-water level observation wells equipped with
recorders on each side of the Gila River, an access hole
for measuring soil-moisture content adjacent to each
well, and a nonrecording precipitation gage at both ends
of each cross section (see inset showing detail of typical
cross section in fig. 1). Recording precipitation gages
were established at both ends of cross sections 1, 9, 17,
and 23. Streamflow gaging stations were established at
cross sections 1, 9, 13, 17 , and 23 to define the Gila River
inflow and outflow through reaches 1, 2, 23, and 3 as
shown in figure 1. Tributary inflow was measured at 16
continuous-recording gaging stations and 43 crest-
stage gages (not shown in fig. 1) located along the
perimeter of the study area. A more detailed discussion
of the instrumentation is given by Culler and others
(1970).

To solve for ET in equation 1 all basic field data were
first transformed into terms described by the compo-
nents in equation 1. Table 1 shows the water-budget
components obtained from the basic field data in reach 1
and corresponding ET values for the 22 budget periods
included in the 1964 water year (Oct. 1, 1963, to Sept.
30, 1964). The budget period includes either a two-week
or a three-week period, depending on when field
measurements of soil moisture were obtained. The end
date shown in column 1 of table 1 refers to the last day of
each budget period. The project day shown in column 2
refers to the ending day of the budget period referenced
from the day the study began on October 1, 1962. Table 2
shows the water-budget components, the resulting ET,
and their corresponding errors for the 21-day budget
period 688 to 708 (Aug. 18 to Sept. 7, 1964). The

 

GILA RIVER PHREATOPHYTE PROJECT

following sections of this report describe how each ET
component in table 2 Was derived and discusses the
methods used in deriving the sampling and bias errors
associated with each component.

EVALUATION OF WATER-BUDGET
COMPONENTS AND ERRORS

STREAMFLOW

The Gila River inflow (Q1) and outflow (Q0) were
obtained by summing the computed daily discharges
over the budget period at the upstream and downstream
ends of the reach, respectively. For reach 1, Q, was
obtained from the computed daily discharges at cross
section 1, and Q0 was obtained from the computed daily
discharges at cross section 9 (fig. 1). The total volume of
inflow during the 21-day period in table 2 is 1,051 acre-ft
(1.296 hm3), and the total volume of outflow is 1,107
acre-ft (1.365 hm3). Table 3 lists the daily discharges
used for obtaining the volume of inflow Q I through cross
section 1 during budget period 688—708. The difference
between Q1 and Q0 is —56 acre-ft (-0.069 hm3),
indicating a net inflow to the river through the reach.

The accuracy of the computed volume of water
passing a gaging station during a budget period is
dependent on the measurement error in discharge and
the accuracy of the stage-discharge relation defined for
the station. The channel of the Gila River is subject to
considerable scour and fill; thus, good definition of the
stage-discharge relation requires frequent discharge
measurements. Burkham and Dawdy (1970, figs. 1 1 and
12) developed curves of the relation between the
standard error in a computed instantaneous discharge
obtained from the rating curve and the frequency of
discharge measurements for the stations at cross
sections 1 and 9. The data used in their analysis was
restricted to flows below a bankfull discharge of about
4,000 ft3/s (100 m3/s). Their error curves were developed
on the assumption that the standard error of any given
measured discharge is 4 percent as indicated by Carter
and Anderson (1963, fig. 1), and they show that the
standard error in a computed instantaneous discharge
obtained from the stage-discharge relation is greater for
the summer months (July through October) than for the
winter and spring months (November through June).
Burkham and Dawdy’s error curves for the summer
months at cross sections 1 and 9 were averaged to define
the error curves shown in the semilog plot of figure 2 for
measurement frequencies of one measurement every 3
days, every 5 days, and every 12 days. Because the
curves in figure 2 were developed from only the summer
data, their application to winter flows may give
estimates of the standard error in a computed discharge
which are too high. The frequency of discharge
measurements for the Gila River during the study

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

L5

TABLE 1.—Evapotranspiration and water-budget components for 1964 water year, reach 1
[All values in acre-ft for indicated days. Flood-plain area is 1,723 acres (697 ha) ]

 

 

 

 

 

 

End Project Number ET Q, Q0 AC QT P AMS AM, A1170 GB GI Go mm
date day of days
14 293 983 —960 28 1 0 7 137 41 87 —77 46
14 582 11,960 —11,288 —42 21 96 ~24 —128 41 87 —77 —64
14 137 5,036 —5,022 —13 36 26 30 41 85 —77 —5
14 209 4,107 —3,s73 17 50 —23 —7o 41 83 —77 —46
14 525 3,491 —3,131 7 24 7 71 41 83 —77 9
14 494 2,691 —2,316 18 0 10 —17 41 82 —76 61
14 230 2,527 —2,269 -29 0 0 —18 41 83 —76 —29
14 —604 4,203 —4,474 —5 o —18 —166 41 82 —76 —191
14 —40 4,322 —4,401 7 29 —2 6 41 82 —76 -48
14 97 2,185 —2,241 33 0 9 17 41 83 —76 46
14 9 1,015 —1,108 3 48 -20 17 41 84 ~76 5
14 195 958 —898 5 35 6 21 41 85 ~76 18
14 196 977 —928 0 26 —2 7 41 85 —76 66
14 112 987 —890 0 75 —17 -51 41 85 —76 —42
20 269 1,166 —999 4 0 4 16 27 59 121 —107 —22
21 410 688 —600 8 0 o 23 127 62 128 —113 87
21 372 173 —144 7 0 0 17 162 62 131 —113 77
21 424 8 —3 0 0 5 15 145 62 132 —114 174
21 1,056 7,614 —6,839 0 14 93 6 36 62 130 —115 55
21 1,903 18,502 —17,068 98 176 272 —61 —86 62 128 —113 —7
21 513 1,051 —1,107 23 252 174 8 82 62 124 —112 —44
21 2,799 24,456 —21,821 —165 310 289 —73 —272 62 123 —113 3
TABLE 2.—Euapotranspiration and water-budget components for the 21 -day budget period 688—708 (August
18—September 7, 1964) in reach I and corresponding sampling and bias errors of each component
[All figures in acre-fl; per 21 days]
ET Q, QO AC QT F AES 4217, WC G3 a, GO Wrc
Component value ______ 513 1,051 —1,107 23 252 174 8 __ 82 62 124 — 112 —44
Sampling error ________ 334 107 ' 121 12 252 16 20 __ 42 0 0 0 138
Bias error ____________ 69 0 0 2 -_ 2 W--- 0 _- 0 62 23 21 0
Total measurement error ___________ .EET=(334 +69 ) = t 341 acre-ft per 21 days

 

TABLE 3,—Daily discharges and errors in discharge at cross section 1
for budget period 688—708

 

 

 

Date 30‘ (eq quz
(1964) (Ra/s) (dimensionless) (ft-Vs!2
August
18 __________________ 47.0 0.133 39.1
19 __________________ 29.0 .142 17.0
20 __________________ 32.0 .140 20.2
21 __________________ 55.0 .130 51.3
22 __________________ 24.0 .146 12.2
23 __________________ 18.0 .151 7.4
24 __________________ 11.0 .160 3.1
25 __________________ 9.2 .164 2.3
26 __________________ 185.0 .108 395.6
27 __________________ 20.0 .149 8.9
28 __________________ 11.0 .160 3.1
29 __________________ 11.0 .160 3.1
30 __________________ 9.1 .164 2.2
31 __________________ 9.6 .163 2.4
September
1 __________________ 10.0 .162 2.6
2 __________________ 9.6 .163 2.4
3 __________________ 9.2 .164 2.3
4 __________________ 7.9 .166 1.7
5 __________________ 7.2 .168 1.5
6 __________________ 8.2 .166 1.8
7 __________________ 7.0 .169 1.4
Totals ________ 530.0 581.6

Q, = 530.0 X 1.9835 = 1,051 acre-ft/21 days
Eq = (581.6/21)"2 = :5.26 fth = :10.4 acre-fia/day
E9! = 21 ((5.26)"’/4.2)V2 = :53.9 ft3/s/21 days = :107 acre-ft/21

days

 

 

1From figure 2 (or equation 2).

period averaged about one every 5 days; thus, the 5-day
curve in figure 2 was used in this analysis. The equation
for this curve when q g 4,000 ft3/s (113 m3/s) is

eq=0.205 — 0.043 logloq, (2)
where e is the error, expressed as a fraction of the
instantaneous discharge, q, in cubic feet per second. For
discharges above 4,000 ft3/s (113 m3/s), the error is
assumed to increase linearly by the relation

eq=—1.75 + 0.5010g10q. (3)

Equation 3 assumes that the error in flows above
bankfull stage increases from 5 percent of the discharge
at q =4,000 ft3/s (113 m3/s) to 25 percent of the discharge
at q=10,000 ft3/s (283 m3/s). Estimates of the error in
discharge during overbank flooding in the Gila River as
defined by equation 3 are believed to be high because
most of the flow in the 4,000 to 10,000 ft3/s range is
contained within the main channel where measure-
ment errors are minimal. Equation 3 is not considered
applicable to discharges above 10,000 ft3/s (283 m3/s).

Burkham and Dawdy (1970) showed that the stand-
ard error, EQ, in the volume of flow passing a station
during the budget period is

L6

GILA RIVER PHREATOPHYTE PROJECT

INSTANTANEOUS DISCHARGE (q) , IN CUBIC METRES PER SECOND

 

 

 

 

 

 

 

 

 

0I4 Oi6 0I81I0 2 [I 6 8 10 20 40 60 80 100 200
l | |
0-3 ~L\I IIIIII IIIIIIII 1IIII'IIITIII
.\\
‘ \
“3° \\
¥ \\
~ \
z "\ d3
— 02 \Zsfrn
\&930r

g \e\ entf

= . \ ’e
E eq 0.205 0.04SIog1oq \\q&en\cfl

\ _
_. 5 days \\\ eq—-1.75+0.50 log 1017
< 01 \
z .
9
'—
o
<
0:
LI.
0 lIIIlIlI I IIIIIII lllIIIII
10 100 1000 10,000

INSTANTANEOUS DISCHARGE (q), IN CUBIC FEET PER SECOND

FIGURE 2.—Average relation between instantaneous discharge and the error in discharge expressed as a fraction of the discharge for
summer (July through October) flows of the Gila River at cross sections 1 and 9.

EQ.=D(E;j/N)'”-’, (4)
where E] is the average standard error in the mean
daily discharge for the budget period expressed as a
fraction of the discharge, D is the number of days in the
budget period, and N is the number of discharge
measurements made during the budget period. Eq is
obtained from the relation

D
Eq=[2(eq,,><qd)2/D] , (5)
d=l

where qd is the average discharge for day d and eqd is the
fractional error in qd as defined by equation 2 or 3. Only
the discharge measurements provide an independent
estimate of the daily discharge—all other daily dis-
charges are computed from the stage-discharge relation
and are therefore dependent estimates. Equation 4
considers this dependency by including the number of
measurements, N =D/5, made during the budget period.
The application of equation 2 or 3 to obtain Eq assumes
that the error for a daily discharge is the same as the
error for an instantaneous discharge. This assumption
is invalid only during wide fluctuations in discharge
which occur for a few days during the winter snowmelt
runoff and the late summer thunderstorm periods.
Included in table 3 are the computations for determin-
ing the standard error of the volume of inflow through
cross section 1, EQI, during budget period 688—708. The
fractional errors, eq, shown for each daily discharge
were obtained from figure 2 (or equation 2). Applying
eqUation 5 gives the average error for the 21 daily
discharges oqu =$ 10.4 acre-ft ($0.0129 hm3) per day. If

 

one discharge measurement is made every 5 days, then
N =42 measurements during the 21-day budget period
andEQl=$ 107 acre-ft ($0.132 hm3) per 21 days. Similar
computations were made to obtain the outflow error
through cross section 9 of EQo =$ 121 acre-ft ($0.149
hm3) per 21 days. The values for EQ, and EQO are shown
in table 2 as sampling errors for water-budget compo-
nents Q I and Q0, respectively.

A comparison of the net change in discharge in reach
1 of —56 acre-ft (—0.069 hm3) forbudget period 688—708
with the average sampling error in this change,
computed as V (107)2+(121)2=$ 162 acre-ft ($0.200
hm3), emphasizes the significance of the sampling error
in the discharge components of the water budget.

No independent evaluation of the discharge error was
made for the flow at the gaging stations at cross sections
13, 17, and 23, because the flow characteristics and
measurement conditions at these stations are similar to
those at cross sections 1 and 9 and the frequency of
discharge measurements is the same. The error relation
of figure 2 was therefore applied to the flow at cross
sections 13, 17, and 23. '

Of the 12 components included in the water budget, Q1
and Q0 are generally the largest (table 1). A comparison
of the ET values in table 1 with their corresponding
discharge values suggests that a high ET coincides with
a high discharge. But, this is true only during periods of
high runoff in the summer when thunderstorms provide
sufficient moisture to satisfy the seasonally high po-
tential ET. Thus, all ET values associated with a high
runoff in table 1 reflect, in part, a large sampling error
and are therefore not considered reliable estimates of
the true ET.

The errors in the volume of discharge for those budget

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

periods containing days without streamflow were not
evaluated and the ET rates computed for those periods
were discarded.

Possible bias in the determinations of Q 1 and Q0 was
investigated for the gaging stations at cross sections 1
and 9. The difference between the measured discharge
and the computed discharge obtained from the stage-
discharge rating using the river stage observed on the
day of the measurement was plotted against the
computed discharge. The plots for both stations show a
relatively uniform distribution of points about the line
of zero difference, indicating no significant bias in the
computed discharges. A further investigation of possi-
ble bias in streamflow was made by comparing the gains
and losses in the computed flow in reach 1 for days
midway between discharge measurements with the
gains and losses on days when discharge measurements
were made. The gains and losses obtained from the
computed discharges appear to be equally distributed
among the gains and losses from the measured
discharges, thus supporting the assumption of no bias in
the computed discharges at cross sections 1 and 9. A plot
of these gains and losses with time of year for the period
before the phreatophytes were removed indicates that
the Gila River was a gaining stream during the winter
months (October through February) and a losing stream
during the summer months of high ET (March through
June). No independent evaluation of streamflow bias
was made for reaches 2a or 3.

CHANNEL STORAGE

The change in channel storage within a reach of the,

Gila River during a budget period is computed from the
average change in wetted cross-sectional areas of the
channel at the upstream and downstream ends of the
reach. This change may be expressed as

'AI +A0 A1 +A0
AC = (1—2—1——22——2) L/ 43,5601, (6)

where AC is the change in channel storage in acre-feet,
L is the length of the river channel within the reach in
feet, and A I and A0 are the wetted cross-sectional areas
of the river channel in square feet at the upstream and
downstream ends of the reach, respectively. Subscripts
1 and 2 denote the first and last day of the budget per-
iod, respectively. A positive change in AC indicates a
depletion in channel storage and a corresponding
addition to ET. A negative change in AC indicates an
increase in channel storage and a corresponding
subtraction from ET .'

A I and A0 in equation 6 were not measured for each
budget period but were derived indirectly from a

 

L7

predetermined area-discharge relation. Figure 3 shows

- the data points defining the relation between measured

discharge and corresponding wetted area using data
from selected discharge measurements at cross section 9
for water years 1963-69.

Similar relations were defined for cross sections 1, 13,
and 17. The curves defining the area-discharge relation
for each cross section all lie within the scatter of the data
defining the curve for any one cross section. Thus, only
one relation approximating the average of all of the
curves has been used to obtain the A I and A0 values in
equation 6. This relation is expressed as i

A :2_9q0.65, (7)

where q is the daily average discharge in cubic feet per
second and A is the wetted cross-sectional area in
square feet.

The change in channel storage for each budget period
was determined by solving equation 7 for A1,, A12, A01,
and A02, using the inflow and outflow discharges
computed for the beginning and ending days of the
budget period and substituting these area values in
equation 6. The discharge values and corresponding
cross-sectional areas for budget period 688—708 are
given in table 4. Substituting the A1 and A0 values in
table 4 into equation 6, where L=36,800 ft (11,200 m)
for reach 1, gives AC =23 acre-ft (0.028 hm3) per 21 days.

The accuracy of AC is dependent on the accuracy of
the A1 and A0 values in equation 6. The error in the
determination of A may be expressed as

E A=(e§+e§+e§)%, (8)

where

E A = the total sampling error in the wetted
cross-sectional area of the river channel,

el = the temporal variability of A at a given cross
section as indicated by the scatter of data
points in the q versus A relation of figure
3’

e2 = the variability of A between cross sections
as indicated by the difference in the q
versus A curves of figure 3,

e3 = the variability in A due to the error in the

discharge used in equation 7 to obtainA as
indicated by the q versus eq relation of
figure 2.

Equation 8 is applicable only if e1, eg, and e3 are
independent estimates of error. This assumption is
considered valid because no relation should be expected
to exist between these three error terms.

The temporal variability inA at cross section 9 ranges
from :10 ft2 ($0.93 m2) when q=10 ft3/s (0.28 m3/s) to
:70 ft2 (:650 m2) when q=4,000 ft3/s (113 m3/s) (fig. 3)

L8

GILA RIVER PHREATOPHYTE PROJECT

AVERAGE DAILY DISCHARGE (q), IN CUBIC METRES PER SECOND

 

 
  
  

2 4 6 8 10 20 40 60 80 100 200 400 600 8001000
I l I l I l l |
1°°°_ll I IIIIIII I IIIIIiII I Ill
'— EXPLANATION
_ 0 Data for cross section 9(1963-69 water years)

 

Average relation for all cross sections

Cross section ‘I

 

_ _ —_— —— Cross section 9

Cross section 13

 

_ ____________ Cross section 17

100

 
  
 
  
  
  

 

WETTED CROSS-SECTIONAL AREA (A), IN SQUARE FEET

 

 

 

  

— 0.8

WETTED CROSS-SECTIONAL AREA (A), IN SQUARE METRES

— 0.6

— 0.4

 

 

| I IIIIII l l

 

 
 

1 000 4000

AVERAGE DAILY DISCHARGE (q), IN CUBIC FEET PER SECOND

FIGURE 3,—Relation between average daily discharge for the budget period and wetted cross-sectional area for cross sections 1, 9, 13, and 17.

TABLE 4.—Average daily discharge (q), corresponding wetted cross-
sectional area of river channel (A), and error in the area (EA) for
cross sections 1 and 9 on budget period days 688 and 708

 

Cross section 1 Cross section 9

 

 

Budget
day
q A I E A q A0 E A
(its/s) (ftz) (ff?) (IV/S) (32) (ftz)
688 ____________ 47.0 35.4 17.5 52.0 37,8 18.2
708 ____________ 7.0 10.3 8.3 51 8,4 7.4

 

and can be expressed as e1 =4.6q0-33. The variability in
A between cross sections ranges from: 2 ft2 (:02 m2) at
q=10 ft3/s (0.28 m3/s) to :55 ft2 (:5.1 m2) at q=4,000
ft3/s (113 m3/s) (fig. 3) and can be expressed as
e2 =0.38q0-60. The error in A attributed to the error in
the computed discharge can be derived from the q versus
eq relation of figure 2. For this analysis the fractional
error, eq, is assumed to be 0.13 for all discharges, and so

 

the error in q is 0.13q. When this discharge is included
in equation 7, e3 can be expressed in terms of the
expected values:
eg = e [2.9(q+0.13q)0'65]2—[e (2.9q0‘65)]2. (9)
Applying a first order approximation of the Taylor
series expansion to equation 9 and dropping all but the
first terms in the series yields e3=0.17q0-65. Inclusion of
higher order terms in the Taylor series was found to
have no effect on the expression for eg.

The above expressions for e1, eg, and e3, when
substituted in equation 8, can be approximated by

EA=3.9q0~39, (10)

where E A is in square feet and q is in cubic feet per
second.
Figure 4 shows the relation of q to E A both as defined

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

AVERAGE MILY DISCHARGE (q). IN CUBIC METRES PER SECOND

 

 

 

 

 

 

 

 

 

 

_' 0.2 040.6031 2 4 6 810 20 406080100 200 _‘
< ZOOIIIII Ill lI'IIviIJ l l‘ [lllllllll I II II IIIIII (
Z Z
9 9
I- 100 :10 5a
8)— : :8 ml:
tom I 2-6 ”.55
33' F F - 32
o — —'4 on,
cc'“ to:
”E ' ' ‘-’<
o: / :2 03
mo - E0
ta: Pu:
gz ,1 $2
. 1o_ —_ 4
a? :/ :°'8 a“;
0:3 _ _—0.6 mg
0 ' ‘_ 0
CC: _ _ 0.4 E:
El: _ _ we:
< <
nu _ —o,2 Lu
0 0
< <
K II
“>1 111:; 1 I IIIIII I I llllll l IlIJlll—OJ :4
< 5 10 100 1000 10'000, <

AVERAGE DAILY DISCHARGE (q), IN CUBIC FEET PER SECOND

FIGURE 4.—Relation between the average daily discharge for a budget
period and the error in wetted cross-sectional area of the Gila River
channel.

by equation 8 and as defined by equation 10. Equation
10 was used to obtain the errors in the wetted
cross-sectional areas shown in table 4.

An estimate of the standard error in AC for any given
budget period can be obtained from the square root of
the sum of the variances of each A1 and A0 value in
equation 6 or

u.

L
E =—-.~ 2 +512 +122 +E2 )-
AC 2x43,560 (E A11 A01 A12 A02 (11)

where the E A terms are in square feet and E AC is in
acre-feet. Solving equation 11 for the E A values in table
4 gives E Ac=i 12 acre-ft (: 0.015 hm3) for budget period
688—708. This error is about 50 percent of the computed
Change in channel storage of +23 acre-ft (+0.028 hm3)
but is only 2 percent of the total ET of 513 acre-ft (0.633
hm3).

AC is generally not a significant component in the
water-budget equation, and greater refinement in its
computation was not considered justified. The determi-
nation of E AC in equation 11 assumes independence of
the error terms E A 1 andE A 0 both with time and between
cross sections. »

The computed increase in channel storage (—AC in
equation 1) is substantially underestimated during
periods of high discharge when low-lying portions of the
channel banks are overtopped and surface water goes
into depression storage in the many small channels and
low areas on the flood plain. ET values computed for
these periods are commonly unrealistically high and
actually indicate a large component of unmeasured
water going into storage. This may partly explain the
unrealistically high ET values for the budget periods
ending on project days 394, 666, 687, and 729 (table

 

L9

1)—all periods of high discharge in the Gila River. The
subsequent period of drainage from depression storage
immediately following a high discharge frequently
causes an underestimate of the decrease in channel
storage (+AC in equation 1) resulting in computed ET
values which are too low. Reliable field measurements
of depression storage were not possible, particularly
during high flow periods. However, the errors in
discharge of the Gila River for these high flow periods
are too large to give reliable estimates of ET and those
ET values are generally disregarded.

TRIBUTARY INFLOW

Runoff from tributaries adjacent to the study area
originates from 225 mi2 (583 km?) of drainage area. A
total of 43 tributaries draining 95 percent of the area
adjacent to reaches 1 and 2 were instrumented with
either recording-stage gages or crest-stage gages. The
stage data from the recording gages were used for
estimating runoff volumes, and the crest-stage gages
were used primarily to define periods of significant
runoff. Tributary runoff into reach 3 was not measured,
because collection of water-budget data on this reach
was discontinued before instrumentation on the
tributaries was fully established.

Normally the tributaries were not monitored during
the winter season (November through April) because
precipitation during this period is generally from
frontal storm systems which may cover a large area but
seldom produce significant flow volumes into the Gila
River. Tributary inflow to the study area was observed
during a few large winter storms, but these periods
coincide with a high discharge in the Gila River and a
corresponding large error in the water budget. The ET
values for these periods have been discarded or are
recognized as not reliable. The only significant tribu-
tary runoff observed in the project area during the
9-year study occurred from May through October. Most
of this runoff resulted from short, intense thun-
derstorms in July and August.

Tributary runoff occurred, on the average, less than 4
percent of the time, or about 13 days out of the year, and
runoff in any one tributary occurred, on the average,
only 3 days per year. Tributary runoff occurred in about
30 of the over 180 budget periods evaluated during the
nine—year study, but only 15 of these periods had runoff
volumes which were a significant part of the ET.

Estimates of the volume of tributary runoff into
reaches 1, 2, and 2a were obtained from stage-discharge
relations and peak discharge-storm volume relations
developed for each tributary by Burkham (1976). The
runoff volumes during budget period 688—708 in each of
the ‘20 tributaries bordering reach 1 (table 5) were
estimated from these relations.

L10

 

 

 

 

TABLE 5.—Tributary runofir into reach 1 during
budget period ‘ 688-—708
Tributary Runoff Tributary Runoff
number (acre‘fi) number (acre-ft) -

24 _ _ _ _ 34 7 .0
25 _ _ _ _ 35 14.0
26 _ _ _ _ 36 2 .0
27 _ _ _ _ 37 5.0
28 0.4 38 50.0
29 _ _ _ _ 38. 5 _ _ _ _
30 _ _ _ _ 39 2 .0

3 1 3.0 40 2.5
32 5.0 41 1.4
33 _ _ __ 42 120.0
Subtotal __ _ 8.4 243.9
Totall _____________________________________________ 252.3

 

Burkham (1976) indicated that definition of the
stage-discharge and peak discharge—storm volume rela-
tions are poor at best and estimated that the computed
volume from a runoff event in any one tributary may be
in error by 100 percent; however, periods when runoff
did not occur were considered to be accurately defined.
Because of the generally low volume of tributary inflow
to the study reaches and their relatively infrequent
occurrence, no evaluation was made of their standard
error. All estimates of tributary inflow to the study area
were thus assumed to be 100 percent in error. Accord-
ingly, the tributary runoff of QT=252 acre-ft (0.311
hm3) for budget period 688—708 (table 2) was assumed to
have an error of EQT=1252 acre-ft (:0.311 hm3).

PRECIPITATION

Accumulated precipitation during each budget period
was obtained from wedge gages located at the ends of
each cross section (see fig. 1). Visits to the gages were
made at two- or three-week intervals which coincided
(within two or three days) with the last day of the budget
period and with the field measurements of soil moisture.
In a few instances precipitation occurred during the
two-day period required to visit all the gages in a reach,
resulting in discrepancies in the total accumulated
precipitation between gages for the budget period.

 

GILA RIVER PHREATOPHYTE PROJECT

These occurrences were rare, and an attempt was made
to correct only the obvious discrepancies.

Each gage was assigned a portion of the total area in
the reach using a method of proportioning which closely
approximates the Thiessen method (1911). The total
accumulated precipitation for the budget period was
computed as an average weighted value from

n n
P=(Z Am) 2 A, , <12)
:1 :
wher_e
P = the average weighted precipitation for the
budget period, in inches,
Pj = the accumulated precipitation at gage j for

the budget period, in inches,

the area assigned to gage j, in acres, and
n = total number of gages in the reach.

Table 6 shows the precipitation amounts observed at
each gage in reach 1 for budget period 688—708 and the
areas assigned to each gage. The average weighted
precipitation of these 10 gages is T’=1.21 in. (30.7 mm),
or 174 acre-ft (0.215 hm3) in volume for the budget
period.

Occasionally the precipitation at a gage was not
obtained. In such instances, the precipitation was
estimated using observed data from nearby wedge
gages or from the recording gages located at the ends of
the reach. Thus, all budget periods contain a complete
set of data.

The total measurement error in the computed
average precipitation for a budget period can include
both a bias error and a sampling error. A bias error
commonly occurs when the gage is located too close to
trees, buildings, or other obstructions which interfere
with catchment in the gage. This type of error is not
considered significant for the project area as all of the
wedge gages were located in areas of ample exposure.
The measurement of precipitation by the gages may
have been slightly low during the summer months
owing to loss by evaporation from the gage; however, a
thin film of oil was maintained in each gage to minimize
evaporation, and this loss is not considered significant.

3?
||

TABLE 6.—Area assigned to each precipitation gage in reach 1 and precipitation amounts observed
at each gage during budget period 688—708

 

Gage number

 

0101 0106 0307 0312 0513 0518 07 19 0724 0925 0930
Area (acres) __________________________ 154 103 77 371 55 342 305 129 92 95
Precipitation (in.) ______________________ 1.50 1.58 1.60 1.25 1.40 1.00 1.10 1.10 1.15 1.15

 

Total area reach 1=1,723 acres

Average weighted precipitation reach 1, F=1.21 inches or 174 acre-ft.

 

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

Thus, all bias errors involved in the measurement of
precipitation are believed to be negligible.

The sampling error in the measurement of precipita-
tion may be attributed to missing data and insufficient
sampling points (gages) within the reach. Included in
the sampling error is the human error resulting from
misreading the gage and the instrument error resulting
from leakage of a damaged gage or debris falling into
the gage.

In this analysis the missing-data error was first
evaluated. The rate of change in the missing-data error
as the number of sample points increases was then
evaluated to obtain an estimate of the sampling error
for a complete set of sample points—that is, with no
missing data.

The missing-data error of the average precipitation
per budget period decreases as the number of gages
monitoring precipitation on the reach increases. Esti-
mates of this error were obtained for reach 1 by
evaluating the departure of the average precipitation
computed using all gages in the reach (n=10 gages)
from the average precipitation computed using m gages,
where m <n. This analysis was made using a total of 114
budget periods with average precipitation amounts of
0.06 in. (1.5 mm) or more. The data were subdivided into
three ranges—0.06 to 0.50 in. (1.5—12.7 mm), 0.51 to
1.50 in. (13.0—38.1 mm), and greater than 1.50 in. (38.1
mm)—and sampling errors were evaluated for each of
the three ranges. Of the 114 periods, 53 fell within the
lower range defining an average precipitation of 0.31 in.
(7.9 mm), 45 periods fell within the midrange defining
an average precipitation of 0.85 in. (21.6 mm), and 16
periods fell within the higher range defining an average
precipitation of 2.06 in. (52.3 mm).

The following six steps outline the procedure used in
evaluating the missing-data error, Em, The procedure
assumes that the observations of precipitation at each
gage are independent and that Sm: 0 when m=10
gages. The computations described below refer to the
precipitation data in a given range; these same
computations were made for each range of data:

1. Compute the average unweighted precipitation, E,
for each budget period, t, from

(13)

10
_—_ 1 .
’fi 2 PI”
j=1

where PJ-t is the accumulated precipitation at gage
j for budget period t. Table 7 gives a generalized
example of the array of PJ~t data and resulting E
values involved in this step.

2. Compute'the average departure of precipitation at

 

L11

TABLE 7. —-Example of precipitation data array used to compute the
average precipitation for each budget period (Pg), the departure of
precipitation at a given gage from the average precipitation (Rjt), the
average departure for all budget periods (Rj), and the standard
deviation of the average departure (SJ)

 

Budget period (t)
Gage

 

 

 

 

 

 

 

. 13- s-
o) 1 2 h J J
'1 P1,1 131,2 Phk
R1,1 R1,2 R1,k El 51
2 P2,1 P2; P2,k
R2,1 R2,2 R2,]? E2 32
Pn,1 P n,2 P n,k
n _
If”,l R ”’2 Rn’k Rn Sn
Ft F1 E2 ‘Fk
each gage from the average precipitation com-
puted from all gages in the reach from
k
— 2 3,, mm
t: 1
where

RJ' =the average departure of precipitation at
gage j, for k budget periods, and
Rjt=pt_ jt (15)
The RJ-t and RJ values defined by this step are
included in table ’7. The RJ- values computed for
each of the 10 gages in reach 1 are given in table 8.

3. Compute the standard deviation of each RJ value
from

£21: (3102 ” (2k 3/0712

3.1: = ’

k—1

(16)

where sJ- is the standard deviation of the average
departure in precipitation for gage j based on
precipitation data from k budget periods. Table 7
includes an example of the array of sJ- values
defined by this step, and table 8 lists the SJ values
for each of the 10 gages in reach 1.

4. For each of m = 1, . . ., 9 gages, arrange the gages into
10 unique permutations (p=10) and for each

L12 GILA RIVER PHREATOPHYTE PROJECT

TABLE 8. —A verage departure in precipitation of each gage in reach 1 (R 1) and the standard deviation of the average departure (sj) computed.

for three precipitation ranges with each range containing k budget periods of data
[All values are in inches per 21 days]

 

 

 

Precipitation range ________________________________ 0.06 to 0.50 inches 0.51 to 1.50 inches 1.51 inches or greater

Sample size ______________________________________ 4 _____ k=53 periods 11:45 periods k=16 periods

j Gage No. 1‘71' fij R“,- §j Rj 81‘
1 0101 —0.001 $0.156 -0.007 10.321 0.046 2:0.511
2 0106 —.003 .180 —.O79 .261 —.109 .384
3 0307 —.002 .140 .012 .288 .045 .439
4 0312 .005 .182 —.095 .252 .015 .344
5 0513 .002 .160 .030 .289 .033 .381
6 0518 .006 .163 —.018 .272 —.032 .567
7 0719 .002 .176 .002 .302 .001 .618
8 0724 —.030 .127 .020 .337 .032 .636
9 0925 0.19 .161.049 .342 —.001 .556

10 0930.001 .155 .086 .350 —.029 .438

 

permutation (v = 1, . . ., 10) , compute the mean TABLE 9-—Example afarray 0f mean departures OfPrecipimtion (Rm)

 

 

 

 

 

 

departure in the average precipitation at m gages and Standard deviations of the med” depa"“’es (Sm)
for all m=1, . . ., 9 by the equation pemumions (0)
Number
_ 1 m of v=1 - v=2 - ‘ - v=p
_ = -— _, gages
Rmv m ;1 R JU , (17) Em“ Sling 3'1"” Sm; . . . Emu}; Sm,p
m’rl 7%?“ 51,1 in Sh: ' ' ' £9117 81:17
-. "l: 2,1 ,1: 1.2 ' ' ' '1: Si.
where R ml, is the mean departure for the m gages 1 2 j s, R. s!” 3 ; -_ 3 p :p
in permutation v. A generalized example of the m=n-- i4... 85—... 51m, Sn—m ~~ ~ [in—741,5) sn—..,p
array of R my values defined by this step is shown In
table 9.

5. For each of the v=1, . . ., 10 permutations, compute and thus defining a "closed array” of departures.
an approximation of the standard deviation of the Chayes (1971, p. 40) showed that a closed array of
mean departure of m gages for all m=1, . . ., 9 by values having a constant mean and variance
the equation defines a correlation coefficient which is nega-

tively biased. This coefficient, r, can be approxi-
%(Szv +82 . _ . +3 31 ,U +2rs 1.v 32‘” mated as a constant value from the relation
r= 1 . (19)
+2rsl,vs3,v+- - ‘+2r31,u3m,v (18) l—n

where n is the total number of gages in the reach.
+' ' '+2rs(m—1),usm,v) V2 , For reach 1, r=—0.11 with n=10 gages. JustificaJ
tion for using equation 19 to approximate the true.
correlation coefficient was investigated by
evaluating Sm”? in equation 18 for selected
combinations of gages using an estimate of the
true correlation coefficient, pj j 2 , computed from
the relation

where S m’v is the standard deviation of the mean
departure for m gages arranged in permutation v,
sly. . .sm’v are the standard deviations in the
departures for 1 . . .m gages, respectively, in
permutation v, and r is the correlation coefficient
of precipitation measured at any two gages. k k 12
Equation 18 assumes that the o ulation mean . . _ 2 2 1/2

and variance of the departures inppfecipitation are pjl .1 2 _ Z (RjIZRjZ‘) (sz 1’ sz 2 t) ’ (20)
constant throughout the reach. Furthermore, the t: 1 t: t: 1

Rj values were computed in such a manner that . . . , ,
where Pjijz IS the correlation coeffic1ent between

n gages j1 and j2 and let and Rm are as defined by
z _j=0, (18a) equation 15. A comparison of the Smm‘ values
j = 1 calculated using pjfl'z with the S m“, values calculat-

 

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

ed using r=.—0.11 showed close agreement, thus
supporting the use of equation 19 to approximate
the corgelation coefficient.

TheRr‘mv and va values for the upper precipita-
tion range are shown in table 10 for two of the
p=10 permutations used in this analysis. The
gages are arranged in sequences of increasing SJ
(see table 8) for permutation v1 and decreasing 31'
for permutation v2.

6. Compute the average standard deviation (gm) of the
10 permutations of R my for each m using the
relation

10 p

__ 1 ___ 1/2

s..=[-,,—(2133..+ 23183.0] . <2»
U: U:

where 17m, and S my are defined in equations 17 and
18, respectively, subscript v is a given permutation
of gages, and p=10 is the _total number of
permutations. The departure, Emu, is included in
equation 21 to account for the average departure
in the precipitatiOn error; however, Hmu is of little
significance as indicated in table 10 and could have
been omitted.

The Sm values defined in equation 21 represent the
average missing-data error associated with m gages. A
plot of these values is shown in figure 5 for each of the
three precipitation ranges with a solid curve drawn
through the points to approximate the average rate of
decrease in §m as m increases. The broken curves
bounding each solid curve approximate the maximum
and minimum missing-data errors computed from the
10 permutations. __ ,, 7 7

Table 11 lists the Sm values obtained from the curves
in figure 5 for each of the three precipitation ranges
investigated. These gm) values show the expected
departure of an average precipitation value computed
from data at m gages from the average value computed

TABLE 10.—-Rmvand S},1v values for precipitation in the upper range P?
1.51 in. (38.4 mm) in reach 1 with gages arranged in order of
increasing sj and in order of decreasing SJ"

 

Gages in sequence Gages in sequence

 

 

of increasing 81' of decreasing s j
m
- .—

GES” Rmu, vai Gage Rmv, Smu:

‘ No. (in.) (in) No. (in) (in_)
1 ________________ 0312 0.015 $0.344 0724 0.032 20.636
2 .024 .166 0719 .01 .312
3 —.020 .187 0518 0 .290
4 —.022 .130 0925 0 .224
5 — 009 .114 0101 .009 .114
6 0 .149 0307 .015 .087
7 0 .124 0930 .009 .080
8 -.004 .078 0106 —.006 .041
9 -.004 .071 0513 —.002 .038

 

 

 

 

 

 

 

 

 

 

 

0-3 I I I I I I I I I
- . — G
0 2 __ 0.06<P<0.50 Inches _
\\\\\ _ 4 (n
m o 1 \ \\\\ :J:
m ' — °\° \\\\\ _— 2 I-
6 ~~~~~~ _\_o \o;; ~~~~~ g
g o | l l T‘—I-——,:p_:3> ------ o 3
z :'
7 2
TE 0-4 I I I I I I I I I - 10 2
Is. :—
E 0.3 _ o \\ 0.51<F<1.50 inches _— 8 WE
— \ x
L; \\\ — 6 z
I_. 0.2 — \\\ _ E
1' \ \\\ — 4 <
0 °\ \\ t
”J 0.1 — \ o\ \\\ - n.
I: ~\\ o\ \ — 2 —
“- _________ o\ \ _____ 8
g o I I I I I I I I “‘I o E
u- n:
I o
g 0.6 I I I I I I I I I 2
E — 14 o
< 0.5 — 0 531.51 Inches — E
I— '— 12 LIJ
3 \\ g
(.9 0'4 — \ _'-‘ 10 <
E \\ Q
"’ — o
2 0.3 ~ 0\ \\ 8 E
E o \\\ u:
\ \\ — 6 g
0-2 — ° \\ _
\\\ \°\ \\\ _ 4
\\ °\ \\\
0 1 — \\\\ o\ \\ __. 2
________ °\o.-__ _.
0 I 1 I I I I I I _I 0
0 1 2 3 4 5 6 7 8 9 10

NUMBER OF PRECIPITATION GAGES (ml

FIGURE 5.—Average relation between the number of precipitation
gages and the missing-data error of average precipitation during
the budget period for three ranges of precipitation in reach 1.

from data at 10 gages. For example, referring to table
11, the departure of the average precipitation defined
with 5 gages from the average precipitation defined
with 10 gages can be expected to have a mean standard
deviation of E5 =:L-0.10 in. (10.25 cm) when the average
precipitation is in the middle range.

The development of the curves in figure 5 assumes
that no error exists in the average precipitation
computed from 10 gages. Each curve shows, however, a
residual error ($10) when m=10 gages. This residual
error is attributed to two factors: (1) the use of a constant
r to approximate the actual correlation coefficients and
(2) differences in the standard deviations, Sj, of the
gages (table 8). These curves therefore overestimate the
expected standard deviation of an average precipitation
value computed from m <10 gages.

The primary objective of the analysis of precipitation
error is not to determine the error attributed to missing
data, as all budget periods contain a complete set of
data, but rather to define the sampling error in
precipitation computed from a complete set of data
(m=10 gages). This sampling error was estimated by
evaluating the rate of change in 3m as m approaches 10

L14

GILA RIVER PHREATOPHYTE PROJECT

TABLE 11.—Average missing-data errors (‘Smflof precipitution‘ in each precipitation range from curves in figure 5 for m=1 to 1 0 gages in reach
1. Included are the unadjusted (V)I and adjusted (E?) sampling errors for m= 10 gages

 

 

 

Average standard deviation of Emu Sampling) error
Precipitation Average Number or (from curves in fig. 5, in inches) m.
122g: ”8%??th peaods _ _ _ _ _ — _ _ — _ U A. . J Adjusted
. I 81 $2 33 S4 35 Se Sq Se 39 Sm V EP-
0.06—0.50 0.31 53 0.16 0.12 0.09 0.08 0.05 0.05 0.04 0.03 0.02 0.02 i0.054 $0.056
0.51—1.50 .85 45 .31 .20 .17 .14 .10 .09 .07 .05 .04 .03 .092 .097
>1.50 2.06 16 .50 .31 .25 .21 .16 .13 .10 .08 .06 .04 .134 .141

 

(fig. 5). This sampling error, V, when expressed in terms
of the gradient between Sm and 3-2”, is of the general
form

aV4+bV2+c=0, (22)
where
n’/2
~2 _2
a: Z (S ":82... (23)
mzl
n’/2
_2__2 _.2 _4
b: E (4SmS2m_3S2m_Sm)v (24)
m=1
n’/2
_2 _2 *6 _4_2
c: E (3s,,,32m—2szm—sms2m), (25)
m=1
and

where n’ is the largest even number sn. The deriva-
tions of equation 22 and the coefficients a, b, and c are
given in the section “Development of equations describ-
ing unadjusted sampling error, V.”

The coefficients a, b, and c were determined for each of
the three ranges of precipitation by substituting the
appropriate 3',” and §2m values of table 11 in equations
23, 24, and 25. These coefficients were then used to solve
equation 22 for V, which defines the sampling error of
precipitation for m =10 gages. This value does not
include the residual error, S10, (table 11) and is
therefore referred to as the “unadjusted” sampling error
as shown in table 11.

The total "adjusted” sampling error, Epm, in precipi-
tation for any given number of gages includes both the
missing-data error (Em) and the sampling error, V. This
total error is computed from

E13”, =(sfn + :72)? (26)

where §m is defined by equation 21 and V is defined by
equation 22. The total adjusted sampling error for a

 

complete set of data (m = 10 gages) was obtained for each
precipitation range by substituting the appropriate V
and §10 values of table 11 into equation 26. These Er
values are included in table 11 and are plotted against
their corresponding average precipitation values in
figure 6.

A curve closely approximating the relation between
these sampling errors and their respective average
precipitation values may be expressed as

Ep=0.10130>47, (27)

where E]? is the sampling error in precipitation in
inches for 10 gages and P is the average precipitation
in inches. ‘

Substituting the average precipitation for the exam-
ple period 688—708 ofF=1.21 in.(30.7 mm) (see table 6)
into equation 27 gives a total adjusted sampling error of
Ep=:0.11 in. (:2.8 mm) or :16 acre-ft (10.020 hm3).

Independent evaluations of the precipitation errors
for reaches 2, 2a, and 3 were not made, but the error
relation of figure 6 is considered applicable to these
reaches because the gage density of these reaches is
similar to that for reach 1.

AVERAGE PRECIPITATION IF}. IN MILLIMETRES

 

 

 

 

 

 

2 4 6 8 1O 20 4O 60 80100
I I I I I | I I I
z 0.20 I I I I I I I I I I I I I T z
9 4 E
._
75 E
E a
6 0.10 _ _ U
I» um
I: ' ~— 2 a: m
m ‘ E —oIoF°‘47 ‘ l E
as — F ' — 3w
I mg
g o _ - o .I
m z E a
$2 ‘ “ w:
< _ 0—
gmln. — / _ 0.8 E;
i - ~O.6 53$“;
2 _ _ E
3" 5
— .1
3: 04 <
I" I-
o E
'- 0.01 I I I I I I I I I 1 I I I
0.05 0.1 l 4

 

AVERAGE PRECIPITATION {F}. IN INCHES

FIGURE 6.—Relation between average precipitation during the budget
period and the total adjusted sampling error in the average
precipitation.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

A few conclusions can be drawn from this analysis
regarding optimum gage density and variation in the
areal distribution of precipitation on the study area.
Referring to figure 5, the missing-data error curves
show a relatively small rate of change in error as m
approaches 10 gages. This suggests that a large increase
in gage density above 10 gages on reach 1 may not
significantly reduce the error in computing average
precipitation. In fact, precipitation per budget period in
the two lower ranges can be defined from only six or
seven gages with little loss in accuracy. Also, as noted
previously, the average precipitation for a budget
period is a weighted value reflecting the fraction of total
area in the reach assigned to each gage. For purposes of
simplicity in the error analysis, precipitation at each
gage was given equal weight using equation 13. A
comparison of average precipitation values weighted by
area (equation 12) with the equally weighted values
indicates no significant differences. Equation 27, which
defines the error of an unweighted precipitation value,
is therefore considered applicable to the weighted
precipitation values in the water budget.

Finally, a correlation between precipitation on the
left bank of the flood plain with precipitation on the
right bank of the flood plain indicates that precipitation
averages slightly greater (+0.02 in. or +0.5 mm per
budget period) on the left bank. This is attributed to the
orographic position of Mount Turnbull and the Santa
Teresa Mountains which rise to an altitude of over 8,000
ft (2,400 m) 8 miles (13 km) to the south (left bank) of the
flood plain.

SOIL-MOISTURE CONTENT

The soil-moisture content was measured within two
areas of each reach: (1) the flood plain which corre-
sponds with the area for which ET is evaluated and (2)
the adjacent terrace area which extends out from the
flood plain to the contact of the saturated terrace
alluvium with the basin fill. Measurements were made
with a neutron probe at two— to three-week intervals
thus defining the water-budget periods. The difference
in the moisture content measured between the begin-
ning and end of the period defines the change in
moisture content for the budget period.

An access hole for measuring moisture content was
located within about 15 ft (4.6 m) of each ground-water
observation well in the study area. Each hole was
classified as one of the following three types: (1) river
hole located adjacent to the river, (2) flood-plain hole
located between the river and the terrace, or (3) terrace
hole located in the adjacent terrace. The river and
flood-plain holes were used to obtain the change in
moisture content in the unsaturated zone of the
flood-plain alluvium and the terrace holes were used to

 

L15

obtain the change in the unsaturated zone of the
adjacent terrace alluvium. A detailed description of the
installation of the access holes was given by Myrick in
Culler and others (1970).

Neutron probes were used to obtain moisture-content
readings in each hole at the 1xé-ft (0.15 m) and l-ft (0.30
m) depths below the land surface and at l-ft (0.30 m)
intervals throughout the remaining depth of the hole
which generally extended several feet below the
ground-water level. The change in moisture content
was determined for three zones of the profile: (1) the soil
zone extending from the land surface to 21/2 ft (0.76 m)
below land surface in the flood plain and to 5 ft (1.52 m)
below land surface in the terrace, (2) the intermediate
zone extending from the bottom of the soil zone to about
3 ft (0.9 m) above the highest observed ground-water
level, and (3) the capillary zone extending from the
bottom of the intermediate zone to the bottom of the hole
in the flood plain and to about 3 ft (0.9 m) below the
lowest observed ground-water level in the terrace. N0
intermediate zone was defined for the flood plain of
reach 1 because of the relatively shallow ground-water
level in the reach.

The change in moisture content in each of these three
zones within the ET area of the flood plain corresponds
to the water-budget components W3, A2171, and We
respectively in equation 1 and items 6, 7, and 8
respectively in table 2. The change in moisture content
in the capillary zone of the terrace corresponds to A117 TC
in equation 1 and item 12 in table 2. Moisture changes
in the soil and intermediate zones of the terrace (AMTS
and AMT] respectively) are not included in the
water-budget equation when evaluating ET from the
flood plain because the moisture in these two zones is
considered to be removed solely by the overlying terrace
vegetation which lies outside the boundaries of the ET
area. Moisture in the capillary zone of the terrace,
however, is believed to be too deep (20 to 40 ft or 6 to 12
m below land surface) to be readily extracted by the
overlying terrace vegetation. Moisture changes in the
terrace capillary zone are thus assumed to result from
changes in ground-water levels in the adjacent flood-
plain alluvium. All significant movement of water out of
the terrace capillary zone is assumed to be lateral and in
the direction of the flood plain in response to an overall
drop in ground—water levels with a general water-level
gradient towards the Gila River. All significant
movement of water into the terrace capillary zone is also
assumed to be lateral but originating from the flood
plain in response to an overall increase in ground-water
levels with a general water-level gradient away from
the river.

The average change in moisture content in a given
zone of the reach for a budget period was computed from

L16

n n
M713: 2 (W212 A1) 2 A) ’ (28)
J=1 j=1

where

AM” = average weighted change in moisture
content in zone 2: of the reach during
budget period t,

Asz = sz (t—1) * zjt . (29)

zj (t—1) .

and M zjt = measured moisture content in zonez of
hole j at the beginning (t— 1) and end
(t) of the budget period,

A j = surface area assigned to hole j, and

n = total number of access holes in the

reach.

The surface area, A-, assigned to each hole was
determined using the same approximation of the
Thiessen method that was applied in assigning areas to
the precipitation gages. When moisture-content data
were missing for an access hole, the change in moisture
content for the hole was approximated using the
average unweighted change computed from the mea-
sured access holes in the reach of the same type (river,
flood plain, or terrace) as the unmeasured hole. Table 12
gives the moisture-content data for the soil and
capillary zones of each hole in the flood plain of reach 1
measured on budget period days 688 and 708. A
negative change in moisture content indicates an

 

GILA RIVER PHREATOPHYTE PROJECT

increase of moisture in the profile (negative ET
component) during the budget period, whereas a
positive change indicates a loss of moisture in the profile
(positive ET component). The AMS and ART/C values
shown in inches of moisture-content change at the
bottom of table 12 were converted to acre-feet in table 2.
Similar computations were made to obtain the
AMTC =—44 acre-ft (—0.054 hm3) for the terrace
capillary zone as shown in table 2.

The soil-moisture data in table 12 indicate that the
amount of moisture change during a budget period is
relatively small compared to the total moisture mea-
sured in the profile. For example, the total moisture
content measured in the flood plain of reach 1 averages
about 3 in. (8 cm) in the soil zone and about 28 in. (71 cm)
in the capillary zone giving a total of 31 in. (79 cm) in the
soil profile. The measured change in this moisture for
budget period 688—708 includes 0.059 in. (0.150 cm) in
the soil zone and 0.575 in. (1.460 cm) in the capillary
zone giving a total change of 0.634 in. (1.610 cm), or only
2 percent of the total moisture measured in the reach.
Thus, a reliable estimate of this comparatively small
moisture change requires that measurements of the
total average moisture be highly accurate—-—allowing for
only a fraction of a percent error.

During periods of low streamflow in the Gila River,
the moisture-storage components A1175, A2170, and NWTC
are generally the most significant components of the

TABLE 12.——Soil-moisture content measured in the soil and capillary zones
of flood plain access holes in reach 1 for budget period days 688 and 708

[All moisture content values in inches]

 

 

 

 

 

Soil zone Capillary zone
tidols: Iéole Area)
e1 0. acres
yp ( Ms, 688 Ms, 703 AMs MC, 688 MC, 708 AMC
2 0102 99.2 2.32 2.86 —0-54 20.21 20.58 —0.37
1 0103 55.1 1.79 1.73 -06 43.30 40.28 3.02
1 0104 33.1 3.46 3.30 -16 28.99 27.84 1.15
2 0105 40.4 4.20 4.58 -.38 64.23 64.81 —.58
2- 0106 29.4 ____ 4.20 2-.05 ____ 29.57 3.13
2 0308 40.4 3.19 2.67 -52 39.43 42.09 —2.66
1 0309 36.7 2.72 2.05 .67 14.66 16.04 —1.38
1 0310 77.1 2.33 1.75 .58 38.91 38.65 .26
2 0311 172.6 2.24 2.21 .03 20.58 20.66 —.08
2 0312 121.2 8.43 8.39 .04 46.05 46.04 .01
1 0514 14.7 ____ 6.66 4.18 ____ 27.02 51.13
1 0515 40.4 ____ 4.41 4.18 ____ 18.22 51.13
1 0516 95.5 2 44 2.06 .38 29.56 28.35 1.21
2 0517 246.1 1.22 1.17 .05 23.00 22.41 .59
2 0720 213.0 2.12 2.12 0 19.72 18.79 .93
1 0721 91.8 3.17 3 11 .06 21.66 20.09 1.57
1 0722 128.6 ____ ____ 4.18 ____ 41.56 51.13
2 0926 51.4 6.87 7.08 ‘21 30.61 31.24 —.63
1 0927 40.4 1.80 1.85 —-05 22.64 20.49 2.15
1 0928 44.1 1.60 1.97 —.‘37 34.26 33.21 1.05
2 0930 51.4 2.11 2.03 .08 26.72 25.11 1.61
Total area:1,723 acres All—45:0.059 in. All—40:0.575 in.

 

1Type of access hole—1=river, 2=flood plain.
2Average measured moisture-content change in soil zone of type 2 holes.
aAverage measured moisture-content change in capillary zone of type 2 holes.

4Average measured moisturewontent change in soil zone of type 1 holes.
5Average measured moisture-content change in capillary zone of type 1 holes.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

water budget. The heterogeneity of the alluvial deposits
underlying the flood plain and terrace result in a wide
variation in the measured moisture change between
access holes as indicated by the wide range in the A273
and AMC values in table 12. A comprehensive evalua-
tion of the measurement error of soil-moisture content
was therefore undertaken to determine the reliability of
these moisture-storage components.

The sampling error is the only significant error in the
measurement of the change in moisture content. A bias
error may exist in the measurement of the total
moisture content of the soil profile because of calibra-
tion inaccuracies of the measuring equipment; but this
error is one directional and nearly constant with time
and thus essentially cancels when computing the
change in moisture content as used in the water budget.

The sampling error in the measurement of the
average change in moisture content in a reach may be
attributed to the following factors:

1. Missing moisture-content data during flood periods
when measurements could not be obtained at some
access holes.

2. Insufficient sampling points (access holes) within the
reach.

3. Improper placement of the neutron probe in the
access hole or misreading the count of returning
neutrons.

4. Variability in the count of returning neutrons.

The method for evaluating the total measurement
error in moisture change is the same as was used in the
precipitation analysis, that is, the missing-data error
was first evaluated and then the rate of change in the
missing-data error as the number of access holes
increases was evaluated to obtain an estimate of the
sampling error.

These errors were evaluated for each zone of the soil
profile in reaches 1 and 2 of the flood plain and reaches 1,
2, 2a, and 3 of the terrace. These errors were then used to
approximate the error in moisture change for reaches 2a
and 3 of the flood plain, thus providing estimates of
measurement error in moisture change far all of the
areas included in the water-budget study.

The general form of the equations used in defining the
measurement errors in moisture change is identical to
the equations previously described for defining the
precipitation error. The steps used in applying these
equations are described below. All computations refer to
the moisture-content data in a given zone; the same
computations were made for each zone.

1. For each budget period containing a complete set of
moisture—content data, compute the change in
moisture content at each access hole in the reach,
AMJ-t, as defined by equation 29 and the average
weighted change in moisture content for the reach,

 

 

 

L17

Wt, as defined by equation 28. (Note that the
array of data illustrated in table 7 for precipitation
also applies to the moisture-content data by
replacing P and I3 in the table with AM and A117,
respectively.)

2. Compute ftj fo_r each hole in the reach using equation
14 where Rj is now the average departure of
moisture change in hole j , fork budget periods and

Rjt=AMt—AMJ-t. (30)
Moisture content data from a total of k =53 budget
periods were used to obtain an R} value for each of
the 21 access holes in the flood plain of reach 1, and
data from k =66 budget periods were used to obtain
an E value for each of the nine terrace access holes
in the reach. Table 13 lists the Rj values computed
for each hole and each zone in the flood plain and
terrace of reach 1.

3. ' Compute 51- for each hole in the reach using equation
16 where sj is now the standard deviation of the

TAELE 13.—-A verage departure in moisture change of each access hole
(R1) and the standard deviation of the average departure (sJ') for the
soil and capillary zones of the flood plain and the soil, intermediate,

and capillary zones of the terrace in reach 1
[All values are in inches]

 

A. Flood plain (area=1,723 acres)
12:53 budget periods

 

 

 

 

 

 

Soil zone Capillary zone
Hole
No ’ ‘
RJ' 8.1 R1 SJ
0.495 —0.125 0.711
.184 -.001 .522
213 .052 .439
369 .095 1.014
300 » .071 .586
148 .018 .471
198 .034 .422
272 .004 .602
570 .051 1.730
918 .044 .982
255 .021 .432
491 —.038 .451
581 .064 .926
441 —.190 2.761
605 —.070 1.373
353 —.038 .693
565 .143 1.409
620 «.029 .550
199 .003 374
.301 .009 .390
.420 .012 .527
B. Terrace (area=1,855 acres)
k=66 budget periods
Soil zone Intermediate zone Capillary zone
Hole
No. ‘ — -. .
Rj Sj Rj SJ' R] 5}
0.000 0.038 0.001 0.029 0.005 0.112
.001 .060 .002 .064 —.003 .160
— 002 .045 .000 .039 007 166
003 .043 000 .055 000 208
006 .064 001 .046 012 247
002 .044 002 .033 005 070
~ 001 .032 — 001 .042 005 097
001 .128 — 003 .201 — 022 323
— 004 .034 — 001 .030 — 005 110

 

L18

average departure in moisture change in hole j
based on k budget periods. Table 13 includes the Sj
values computed for each RJ- in the flood plain and
terrace of reach 1.

4. Arrange the n access holes in the reach into five
unique permutations (p =5), and for each permuta-

' tion (v=1, . . ., 5) compute Rmv for all_m=1, . . .,
(n—l) holes using equation 17 where Rm, is now
the mean departure in moisture change for m holes
in permutation v. '

5. For each of the v=1, . . ., 5 permutations, compute
va for all m=1, . . . , (n—l) holes using equation
18 where S ml, is now the standard deviation of R mu
for m holes arranged in permutation v. As in the
precipitation analysis, the mean and variance of
the departures in moisture content were assumed
to be constant within the reach. Thus, the corre-
lation coefficient, r, included in equation 18 was
approximated by equation 19. Equation 19 gives
r=—0.05 for the 21 flood plain access holes and
r= —0.12 for the 9 terrace access holes in reach 1.

To test the reliability of r as an approximation of
the actual correlation coefficient, pj 11- 2 (see equa-
tion 20), the Smu values for selected combinations
of access holes in the flood plain of reach 1 were
computed using r= —0.05 and compared with the
va values derived using the actual correlation
coefficients, pj1 jg , computed from equation 20. The
va values derived by these two methods showed
relatively close agreement, both for the soil zone
and the capillary zone. Particularly close agree-
ment exists between the 8"“, values as m
approaches the total number of access holes in the
reach.

No attempt was made to evaluate pj 1i 2 for each
combination of the nine access holes in the terrace
of reach 1; however, r= —0.12 is considered a
reasonable approximation of the actual correla-
tion coefficients. _

6. Compute the gm of the p permutations of RM, for all
m=1, . . . . , (n—l) access holes using equation 21
Where §m is now the average standard deviation of
the p permutations for m access holes. Figure 7
shows a plot of the S—m values for the soil and
capillary zones of the flood plain (fig. 7A) and for
the soil, intermediate, and capillary zones of the
terrace (fig. 7B). The curves drawn through the
data points are estimates of the average error in
the computed change in moisture content where
data (access holes) are missing.

Table 14 lists the curve values of Sm for the flood plain
and the terrace of reach 1. As with the precipitation
error curves (fig. 5), a residual error also exists in the
moisture-content error curves when m=n holes. This

 

GILA RIVER PHREATOPHYTE PROJECT

residual error is attributed to differences in the 3]- values
of the access holes (see table 13) and the use of a constant
r to approximate each ijjZ' These curves are thus
assumed to overestimate the expected standard devia-
tion in the average moisture change.

The curves in figure 7A show that the error in the
computed average change in moisture content for the
flood plain of reach 1 does not decrease substantially
beyond about 12 holes. Thus, moisture change could
have been obtained for this area from about one-half the
access holes actually used without a significant increase
in error.

The rate of change in gm as m increases (fig. 7)
provides an estimate of the sampling error in moisture
change when the change is computed from a complete
set of data (m =n holes). To estimate this sampling error,
equations 22—25 were applied using the gm values taken
from the curves in figure 7 and listed in table 14. The

1.2

 

     
 

 

i'I I I I I I I I I _ 30
A. FLOOD PLAIN
1.0 — _— 2.5
0-8 — —— 2.0
0-5 — x Capillary zone -— 1.5
I
0.4 — ——1.0 I
°\°
\x x X\
R X
0.2 — \e N —— 0.5
Soil gonrfi\ x ngx
°\s\ ~x-._
\°\o\°
fifiW—o—o_°___
o I I I I I I I I I 0

 

 

B. TERRACE

2.5 —

2.0 -—

 

‘0
Lu
J:
o
E
E
<
E
9
Lu
0
z
<
I
o
I-
Z
“J
I—
z
o
9
I.“
g30 I I I I I I I I
,_
‘2
O
2
Ir.
0
U.
cc
0
a:
c:
LU
<
I—
<
9
0
E
U,
‘2
E

MISSING-DATA ERROR FOR MOISTURE-CONTENT CHANGE (Sm), IN CENTIMETRES

 

 

 

0 I I I
0 1 2 3 4 5 6 7 8 9

NUMBER OF ACCESS HOLES {m}

 

FIGURE 7.—Average relation between number of access holes in reach
1 and the missing-data error of moisture-content change per budget
period for (A) the soil and capillary zones of the flood plain and (B)
the soil, intermediate, and capillary zones of the terrace.

 

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

TABLE 14.-—Missing-data errors (Em) from curves in figure 7 for the soil
and capillary zones of the flood plain and for the soil, intermediate,

and capillary zone_s of the terrace, reach 1
[Sm computed from p=5 permutations]

 

A. Flood plain §m, in inches

 

 

   

 

 

 

 

Number of Soil Capillary Number of Soil Capillary

holes zone zone holes mne zone
$ 1.22 $0.23
.84 .22
.65 .20
.54 .18
.47 .17
.42 .16
.37 .15
.33 .14
.31 .13
.28 .13

.25

B. Terrace Sm, in inches
Number of Soil Intermediate Capillary

holes zone zone zone
$2.39 $2.84
1.13 1.73
.74 1.20
.58 .90
.46 .71
.38 .58
.31 .48
.26 .41
.21 .36

 

 

solution of these equations give the following V errors
for the flood plain and terrace of reach 1:

Flood plain soil zone (n=21 holes): VS =$0.118 in.
($0.300 cm)

Flood plain capillary zone (n=21 holes): VG = $0.244
in. ($0.620 cm)

Terrace capillary zone (n=9 holes): VTC = $0.776 in.
($1.97 cm)

These V values represent the unadjusted sampling
error in moisture change when m=n access holes.

As with precipitation, the total adjusted sampling
error in moisture change (Em) for any given number of
access holes includes both Em and V. Thus, the gm
values in table 14 and the corresponding V values
determined above were substituted in equation 26 to
define the total error curves shown in figure 8 for each
zone of the flood plain and terrace in reach 1.

As indicated in table 12, soil moisture at 17 of the 21
access holes in the flood plain was measured during
budget period 688—708. Entering m=17 in figure 8A
gives the total adjusted sampling errors for the flood
plain of E817=i20 acre-ft ($0.025 hm3) in the soil zone
and E017=i42 acre-ft ($0.052 hm3) in the capillary
zone. Similarly, soil moisture at eight of the nine access
holes in the terrace was measured during budget period
688—708 giving, from figure 8A, E7108 =$ 138 acre-ft
($0.170 hm3) in the terrace capillary zone. Table 2
shows each of these sampling errors under their
respective zone.

 

The procedure described above for determining the

 

150 I I I I I I I I I I
A. FLOOD PLAIN

I.

1m

l.

100 ' Capillary zone

2In

 

 

 

 

 

 

 

 

0 2 4 6 8
NUMBER OF ACCESS HOLES {In}

E
3
Lu
0 50 _ \\ Soil zone m
s r~~~~~— s
5 o I I I I I I I I I I O E o
I- 0 '2 4 6 8 10 12 14 16 18 20 O O
z 8 I— E
E E NUMBER OF ACCESS HOLES {m} a I:
2 Iu '- b-
0 n- 2 m
9 E 8 0
E 0 30° ~ I I I ”.1 o
D D —O.35 0: a
l7, 3 B. TERRACE a m
— 250 — “—0 30 m “J
g 5 ' 5 ‘L
' U)
E :- 200 — ca”"""V 1°" _—o.25 2 g
u] \ E I-
n: m \ Lu
g V.- \ \ —o.2o g E
I: g 150 - \\ —- o: 8
Lu, 2 \\\|ntermediate ‘0-15 E B
E z 100 “ \ \‘3\f g i
-J " \~——— 0.10 _ _
E \\Soil zone : no
< 50 ‘ \\\ e_.__ ‘— 0.05 2 8
u) — < z
_l ‘0 _
< I I I I _I
’5 ° :5
* .9

FIGURE 8.—Average relation between the number of access holes in
reach 1 and the total adjusted sampling error of moisture-content
change per budget period for (A) the soil and capillary zones of the
flood plain and (B) the soil, intermediate, and capillary zones of the
terrace.

total adjusted sampling error in moisture change of
each zone was also applied to the soil-moisture content
data for the flood plain and terrace of reach 2 and the
terraces of reaches 2a and 3. Table 15 lists these errors
in moisture change computed from a complete set of
data (m=n).

As indicated previously, moisture change in the soil
and intermediate zones of the terrace are not included in
the water budget (equation 1) when evaluatingE T from
the flood plain. However, the errors for these two terrace
zones were independently evaluated and are included in
table 15.

The estimated error in moisture change is frequently
as large or larger than the measured change in moisture
content for the budget period. These errors are
relatively small, however, when compared with the
total volume of moisture measured in the reach. For
example, the total volume of moisture in the soil and
capillary zones of the flood plain of reach 1 averages
about 31 in. (79 cm) or 4,450 acre-ft (5.49 hm3) (p. 16).
The total error in moisture change for the flood plain
of reach 1, assuming no missing data, is (ESZ+EC2) V2:
(182+392)1’é=$43 acre-ft ($0.053 hm3) from table 15.
Because this error is derived from two measurements of
moisture volume—one at the beginning and one at the
end of the budget period—the total error for one
measurement is V432/2 =$30 acre-ft ($0.037 hm3) or
only 0.7 percent of the total volume.

No independent evaluation of the sampling errors in
moisture change (E2) was made for the flood plain of

 

L20

GILA RIVER PHREATOPHYTE PROJECT

TABLE 15.—Total adjusted sampling error in moisture change as defined with a complete set of moisture-content data (m=n) for each zone
of the flood plain and terrace in reaches 1, 2, 2a, and 3

 

 

 

 

 

Flood Plain Terrace
Reach Area Number Density Error Area Number Density Error
(acres) of holes (acres/hole) (acre-fl/budget period) (acres) of holes (acres/hole) (acre-filbudget period)
A n A/n ES E1 EC A n A/n ETS ET! ETC
21 82 18 __ 39 1,855 9 206 34 82 132
23 100 18 13 33 1,363 8 170 42 54 88
13 106 ‘24 218 :346 992 5 198 47 49 97
17 85 120 215 “38 602 6 100 23 23 56
’Estimated from equation 31. 2Estimated from equation 32. “Estimated from equation 33.
reaches 2a and 3' rather an a roximation of the errors -- -
’ ’ PP AMC=—AhS’A, (34)

for these two areas was obtained from the previously
derived errors for the flood plain of reaches 1 and 2 and
the terrace of reach 3. These previously derived errors
are considered to be applicable to the flood plain of
reaches 2a and 3 because the areas have similar
sampling densities (table 15).

To estimate the total sampling error in moisture
change for the flood plain of reaches 2a and 3, the
previously derived EzmIvalues in figure 8 were first
expressed in terms of sampling density by converting
the m associated with each error value to the ratio A/m.
The relation between E2", and A/m for each zone was
then plotted on a semilog scale as shown in figure 9.
Finally, a straight line approximating an average
relation for each zone was drawn to estimate the
average errors in the measured moisture-content
change for the flood plain of reaches 2a and 3. Equations
for these average relations are

E» =—51+37 log10 (A/m) (31)
E1: —43+30 log‘10‘(A/m) (32)
EC=—112+78 l°g10 (A/m), (33)

where E3, E1, and EC are the errors in measured
moisture change in acre-feet for the soil, intermediate,
and capillary zones, respectively, A is the flood-plain
area of the reach in acres, and m is the number of access
holes. Equations 31—33 were solved using m=n to
obtain the total sampling errors shown in table 15 for
the flood plain of reaches 2a and 3.

Moisture-content measurements could not always be'

obtained at every access hole in a reach during a field
visit; however, ground-water levels were generally
recorded at most of the wells adjacent to the access
holes. When moisture-content data from more than half
of the access holes in a reach were missing for the budget
period, a more reliable estimate of the change in
moisture content in the capillary zone could generally
be obtained from the water-level data. The relation used
to obtain the capillary moisture-content change from
the average water-level change is

where A2170 (or AMTCD is the average moisture-content
SAMPLING DENSITV (A/m}, IN SQUARE HECTOMETRES PER HOLE

 

 

 

 

 

 

 

 

 

 

 

 

50 100 200 300 400
I I I I I I I I 8
8 60 | I I I I I I I E
E Soil zone / ///,/’ — 0.60 E
:1 4o — / / —”’ — E
E /,-r"”’ E‘51+37| (A/ ) — 0.40 8
0 / s“ ”9 10 "' :>
20> 20 ”’ m
_/ _

m — 0.20 E
E m
U)
,‘f o I I I I I I I I I 0 g;
“1 I—
W LU
u_. E
'6: E
2 6° I I I I I I I I 8
E 0.60 5
~‘ E

t
D
5'- 040 u
W Z
0 _
E — 0 20 R;
5 - a
U]

._
z 0 I I I I I I I I I 0 L29
“J <
'— I
Z 0
o I—
9 120 Z
l; — 0.14 Lu
,_
3 z
" o
9 100 — o 12 o
O — . Ii:
5 u:
z D
E 80 .— 0.10 2;
0
g s
— 0.08 2
“J 60 _ _
g a:
3 _ 0.06 E
1 EC
2 4o — m
5 — 0.04 g
" . :1
‘1 n.
'6 2° _ _— 0.02 5
l- 5)
.J
I I I I I I I I I <
o o ._
so 100 200 400 600 800 1000 I9

SAMPLING DENSITY (A/m}, IN ACRES PER HOLE

EXPLANATION

Reach 1—Flood plain
—— — —- Reach 2—FIood plain
—————— Reach 3—Terrace
Average for reaches 2a
and 3—Flood plain
FIGURE 9.—Estimates of average relation between sampling density
of access holes and total adjusted sampling error of moisture-
content change per budget period for the soil, intermediate, and
capillary zones of reaches 1, 2, 2a, and 3.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

change in the capillary zone of the flood plain (or
terrace) in acre-feet, A]: is the average change in the
ground-water levels in the flood plain (or terrace) of the
reach, in feet (positive for a rise and negative for a drop
in water level), S ’ is the apparent specific yield of the
aquifer in the zone of water-level change (dimension-
less), andA is the area of the flood plain (or terrace) in
acres. An average value of S’ was determined for both
the flood plain and terrace areas of each reach by

L21

relating the average water-level change of the area, A5,
to the correspbnding measured average moisture-
content change in the capillary zone, AIL/«or AMTC),
using budget periods containing a complete set of water
level and moisture content data. Figure 10 shows plots
of this relation for the flood plain and terrace of reach 1.
The slope of the line drawn to average the data points in
each relation defines the S ’ values used in equation 34.

The Ah versus WC I relation for some of the areas

AVERAGE WATER-LEVEL CHANGE (A77), IN MILLIMETRES

 

 
  

 

 

 

 

 

 

 

 

 

 

 

~300 -2?0 -100 0 100 200 300
I l l l
0-2 I I I I I I /° / 6"
A. FLOOD PLAIN 40
Specific yield=0.20
0.1 — y _
0 / Standard error=z 0.04 ft
or :69 acre-ft - 20 ‘0
/ ° E
I. 0 5:8 u.
{3 A, o o m'/o 0 g
LL 0 / 0 /o A
z o ,o’0 _I
_ / o .69.: —-20 E
”I ~01 — / /° / Z
Z / / _ _
o / / ~
N / / / / --40 g
>- / / o
E 0.2 J/ I I I I I I I I 50 N
_, >-
_| (I
E S
5 -_J
g 3E
u o
0 E
z
< I I I I I “‘
0.3 0
5 I I I I I I o I I I <2(
/._.
'2 B.TERRACE / / 80 5
:11 Specific yield=0.17 A” / I-
z 0.2 — ° / — 60 Z
O / / Lu
9 / I—
g o / / ° / 9
I'- 0.1 - / / O / /— Lu
‘2 ./ o / ‘ ‘I
o / ’0’ / — 20 B
E / / ° 0 °° o // Standard error=t0.08 ft 8
u 0 ’0’ o o o c or 1148 acre-ft 0 E
< / /° 0 o O O 0/ Lu
E / / o ° ° ° / o 0
> / / 0 ° ‘9 0 .°/ / ° —-20 <
< O 0 / / El
0.1 D— / -- >
0 / / <
/ / O —-40
/
/ O
0.2 — / °/ / ——-60
/ 0/
° — so
03 l l I I l | l l l
-1.0 0 8 -0.6 0 4 -0.2 0 0.2 0.4 0.6 0.8 1.0 1.2

AVERAGE WATER-LEVEL CHANGE (Ah), IN FEET

FIGURE 10.—Apparent specific yield values derived from relation between average change in water levels and average moisture-content
change in capillary zones for (A) flood plain and‘(B) terrace of reach 1.

L22

indicate a slight difference in S’ between recharge and
drainage of the aquifer; however, the variability of the
data used to define the relation and the normally small
changes in Ah did not justify defining this difference in
S ’. Table 16 lists the S’ values and the number of budget
periods, k, used to define S ’ for the flood plain and
terrace areas of each reach.

The dotted lines paralleling the average line in each
plot of figure 10 bound two-thirds of the data points,
thus approximating the standard error of the measured
moisture-content change in the capillary zone. These
standard error values include, not only the sampling
(measurement) error in moisture change for any given
Ali, but also the actual variability in moisture change
due to temporal variations in ET. Thus, the error
values in the figure cannot be compared directly with
the sampling errors, EC and ETC, given for reach 1 in
table 15.

An illustration of the application of water-level data
to compute the average moisture change in the capillary
zone of the flood plain of reach 1 for budget period
688—708 is given in table 17. The average weighted
change in ground-water levels for the 21 wells in the
flood plain is Ali: —0.386 ft (—11.8 cm). Applying
equation 34 where S ’=O.20 (table 16) and A=1,723
acres (697 hmz) gives AMC=133 acre-feet (0.164 hm3).
This value differs considerably from the AMC=82
acre-feet (0.101 hm3) computed from the capillary
moisture-content data in 17 of the 21 flood plain access
holes (see table 2). But this discrepancy can be expected
as indicated by the large standard error of the data
points (:69 acre-feet or :0.085 hm3) in the Ah versus
All-lclrelation of figure 10A.

A brief examination of the error in estimating AMTC
from the average change in water levels was made using
water-level data collected at the 10 terrace wells in
reach 1. The method of analysis was identical to that
used for evaluating the sampling errors in precipitation
and moisture change and therefore will not be described
in detail here. Water-level data collected during 20
budget periods for the 1968 water year were used- in this

analysis because the data provide a wide range in.

water-level changes. Table 18 gives the average

 

GILA RIVER PHREATOPHYTE PROJECT

departure in water-level change (I—i'j) derived from
equation 14 and the standard deviation of these
departures (SJ-D derived from equation 16 for_ each of the
10 wells. The missing-data error values, Sm, derived
from equation 21 are plotted in figure 11 and, as in the
previous evaluations, indicate a residual error of Sm
=:0.035 ft (: 1.07 cm) when m=10 wells. The applica-
tion of these Em‘ values in equations 22—25 give an
unadjusted sampling error in water-level change of
V =:0. 15 ft (:4.6 cm). Combining the residual error and
the unadjusted sampling error as in equation 26 gives
an adjusted sampling error in Ah of EN-LF:0.16 ft (:49
cm) for m=10 wells. The use of water-level data to
estimate moisture-content change in the water budget
was seldom required, and the error in Ah for the other
flood plain and terrace areas was not evaluated. The
error E A}; =:0.16 ft (: 4.9 cm) was therefore applied
throughout the study area when using Ah to estimate
AM—C‘and AMTC. Using equation 34 to express this error
in terms of acre-feet of moisture change for the capillary
zone of the terrace in reach 1 gives E A7; =: 50 acre-feet
(:0.062 hm3) where S’ in equation 34 is 0.17 (table 16)
and A is 1,855 acres (751 ha). Table 16 lists these E A;
values for each reach in the study area.

The total error of moisture change in the capillary
zone when derived from the Ali vs. All—l relation includes
not only the sampling error in Ah as defined above, but
also the sampling error in AM as defined previously for
the capillary zones of the flood plain and terrace areas of

TABLE 17.—Ground—waterlevel changes (Ah) in flood plain wells of
reach 1 used to compute AMC for budget period 688—708

 

Area Ah

Area Ah
No. (acres) (ft) No.

(acres) (ft)

 

+0.02
—.57
—.46
—.21
—.06
—.33
—.35
—.31
—.18.
—.06‘

—0.60
—.81
—.76
—.55
—.36
—.47
—.57
—.18
—,55
—.53
—.48

   

 

 

Average weighted change Ah= —0.386 R.
AMC‘: —(—0.386) X020 X 1,723 = 133 acre-it.

 

TABLE 16.—Auerage apparent specific yield (8’), number of budget periods (k) used to define S', total adjusted sampling error in the measure—
ment of average water-level change (EM—1), and the total adjusteid sampling error of moisture change in the capillary zone (ECAH and
ETCAE‘) when the moisture change is derived from water-level chpnge (equation 34)

 

 

 

Flood Plain Terrace
Rem k S’ EA; ECAE k 5’ EN"; ETCN—l
(acre-fl) (acre-it) (acre-ft) (acre-ft)
0.20 :55 +67 68 0.17 :50 :141
.26 96 101 46 . 15 33 94
i 18 40 64 3 1 17 27 101
.20 46 60 15 09 9 57

 

 

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER—BUDGET METHOD, ARIZONA

TABLE 18.—Average departure in ground-water level change (Rj) of
each terrace well in reach 1 and the standard deviation (sj) of Rj
computed from 20 budget periods of water-level data collected during
the 1968 water year.

 

Well

 

. s.

No. (fiJ) (fl)
0101 __________________________________ —0.03 10.33
0106 __________________________________ .02 .33
0307 __________________________________ —.02 .43
0312 __________________________________ —.04 .41
0513 __________________________________ .02 .18
0518 __________________________________ .05 .19
0719 __________________________________ —.01 .28
0724 __________________________________ .02 .36
0925 __________________________________ .02 .34
0930 __________________________________ - .03 .49

 

each reach. This total error may be expressed as

ECAII=(E2A}1 +152 CW2, (35)
where EAT, is given in table 16 and EC is the total
adjusted sampling error in the capillary zone when
m=n (table 15). For the terrace of reach 1 this total
error is ETCAg=(502+1322)1/2=:141 acre-feet
($0.173 hm3). The E05, and ETCN; error values were
computed from equation 35 for the capillary zone of each
reach using the EA}; values in table 16 and the EC and
ETC values in table 15. A list of these total error values
is included in table 16.

BASIN-FILL INF LOW

Basin-fill inflow (G3) to the study area is derived from
deposits of low permeability which underlie the
alluvium of the flood plain. This component moves
vertically upward-into the alluvium, and estimates of
its rate offlow range from 0.07 to 1.3 ft (0.02 to 0.40 m)

 

 

 

 

l— l"!-
o
5004 | l l I 12 5‘“
l-_ r-‘D
u: 0
gm 0 23
I; one
Q)— _10 9
gm :5
D(0903— — Do.
l- l—"’
93 —-8 a):
o o:—
55 0 EM
2“L ZED
—p—O2— ——6 —._9
IE I29:
gm gee
E; _4 52
<4 __ 4
,_ 50.1 °\ — ff 5
<‘2 o <'$
Qm \°\° —2 9m
5% ‘°‘°~u' 33
as o I l I I o 3%
2 o 2 4 6 8 10 5°

NUMBER OF WELLS {ml

FIGURE 1 1,—Average relation between number of ground-water wells
and missing—data error in moisture-content change per budget
period for capillary zone of terrace in reach 1.

 

L23

per year (Hanson, 1972, p. F27). If all basin-fill inflow
surfaced as evapotranspiration, these values would
represent an ET rate between 15 and 265 acre-ft (0.018
and 0.326 hm3) per 21 days for reach 1. The constraints
and limitations associated with the methods used to
derive these GB values make these estimates question-
able.

A subsequent analysis of the moisture movement in
the capillary zone of the deep terrace wells of reach 1
did, however, provide a better indication of basin-fill
inflow to the study area. In the analysis, only moisture
data from the winter months, when ET is minimal and
cross-valley ground-water slopes are negligible, was
evaluated. The results of this study indicate that the
basin-fill inflow is about 0.3 ft (0.09 m) per year per unit
area of flood plain or G 3:62 acre-ft (0.076 hm3) per 21
days for reach 1. This value is believed to be a more
reasonable estimate of the true rate of basin-fill inflow.

The basin-fill inflow was assumed to remain constant
throughout the year, unaffected by seasonal variations
in barometric pressure, temperature, or ground-water
levels. Because G B is considered time invariant, its
sampling error is zero as indicated under item 9 of
table 2.

No evaluation of the bias in the estimate of basin-fill
inflow was possible. Thus, it is assumed that the
estimate is 100 percent in error as indicated by the bias
error value of EGB =1- 62 acre-ft (10.076 hm3) under item
9 in table 2. This bias may be significant when
evaluating ET for a given budget period, but the bias
cancels when computing the change in ET from
before-clearing and after-clearing ET data.

DOWNVALLEY GROUND-WATER FLOW

Ground-water movement downvalley through the
upstream and downstream ends of each reach was
calculated from

G=iTWD, (36)
where

G = downvalley ground-water flow through the
alluvium in acre-feet per budget period,

1 = average downvalley gradient of the ground-
water level during the budget period
through the upstream or downstream end of
the reach,

T = transmissivity of the alluvium in acre-feet
per day per foot,

W = width of saturated alluvium at the upstream
or downstream end of the reach, in feet, and

D = number of days in the budget period.

The transmissivity, T, of the alluvium was assumed
to be constant throughout the study area and was

L24

determined by Hanson (1972, p. F27) to be 28,000 ft3 per
day per foot or 0.644 acre-ft per day per foot (2,600 m3
per day per metre). The width of saturated alluvium, W,
was determined by measuring the distance between the
points of contact of the alluvium with the basin fill at the
water table on each side of the flood plain. The
downvalley slope, i, was computed from the average
ground-water levels for the budget period measured at
the river wells and flood—plain wells on and adjacent to
the cross sections at the ends of the reach. For example,
the slope through the upstream end of reach 1 (cross
section 1 in figure 1) was computed from the average
water levels measured in the river wells and flood-plain
wells at cross sections 1 and 3. Similarly, the slope
through the downstream end of the reach (cross section
9) was computed from the average water levels in the
wells at adjacent cross sections 7 and 11. The
calculations used to obtain GI and G0 in table 2 are

G1 = 0.00158X0.644X5,800X21=124 acre-ft/Zl
days (0.153 hm3 /21 days)

G0 = 0.00148X0.644X5,600X21=112 acre—ft/21
days (0.138 hm3/21 days).

The sampling error associated with the G1 and G0
components is dependent only on the sampling error of i
in equation 36 because i is the only factor in the
equation which is measured for each budget period. The
factors T and W do not have a sampling error, because
they are considered constant with time (actually T and
W may vary slightly with large changes in water level)
and T is assumed to be constant throughout all reaches.

Seasonal variations in downvalley slope through
most cross sections are generally less than 5 percent
during periods of minimum ground-water level fluctua-
tions. Most of this variability reflects changes in the
ground-water level caused by precipitation, changes in

 

GILA RIVER PHREATOPHYTE PROJECT

the river stage and seasonal variations in ET. As a
result, no detailed evaluation of the sampling error in
slope was possible. An approximation of this error was
obtained, however, by examining the variability in the
measured downvalley slope during the winter months
when ET is negligible and the ground-water level
remains relatively stable. Figure 12 shows the average
downvalley slopes through cross section 9 for the winter
months (November through February) of water years
1964, 1965, 1967, 1969, and 1970. Variability about the
general trends in these slopes suggests that the error in
i for any given budget period is probably less than
:0.005><10—3 or :0.3 percent of the slope. An error of
:03 percent gives an average error in the downvalley
ground-water movement of :OA acre-ft ($00005 hm3)
per 21 days for the ground-water inflow and outflow
components in table 2. Because this sampling error is so
small, it is considered zero as shown under the GI and
G0 components in table 2.

Any bias error in the ground-water components, G1
and G0, is attributed solely to W and T in equation 36.
The bias error in G1 and G0 resulting from an inaccurate
determination of W was estimated to be :200 ft (:60 m)
or 4 percent of W. This estimate is probably high and is
believed to be closer to: 100 ft (: 30 m). Considering that
the downvalley ground—water inflow to reach 1 is
G1: 124 acre-ft (0.153 hm3) for budget period 688—708,
theerror in W of 4 percent gives a bias error in G1 of :5
acre-ft (:0.006 hm3).

The bias in GI and G0 attributed to using a constant
average T for all reaches was estimated by assuming
that the spatial variability in downvalley slopes, not
explained by differences in flood-plain width, was a
direct measure of the variability in T. The average
downvalley slopes between the cross sections at the ends
of each reach are plotted in figure 13 against the width

 

 

 

Lu 1.66

l 1964

3 —- __

w 1965/ *'————____._———————“\\

E 1.64 __ /// \ \_ \1
/ \

LE _/ , /.

1 O __ '

D ‘- _--—— ‘———__ \ "v~\ _

g x 1-62 - , /‘ 1969 ———————— ;::-,=—"- -—'—'~—\;~- —_-_-_—_.-=—_—_ mu:- —

0:5 \1‘

(I -— ._...

0

>' 1.60-~~ _ -__-_-__——--——-—--—--—=

1'1 ‘-\- 1970 , —/ “~-———‘""

—I ‘V’

< —— _

E

g 1.58

3 NOV DEC JAN FEB

 

 

 

 

 

 

FIGURE 12.—Average downvalley ground-water slopes through cross section 9 for winter months of the 1964, 1965, 1967,
1969, and 1970 water years.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

FLOOD-PLAIN WIDTH AT DOWNSTREAM CROSS SECTION (W),

IN METFI ES
800 1000 1200 1400
I

20 1'1 I I

1600

1 800

       
 

 

1.5 —

 

/ Note: Numbers (15-17)
are cross sections

DOWNVALLEY SLOPE (i), x 10'3

/
21 -230/ used to compute

slope

 

 

0.5 I I I
2000 3000 4000 5000

 

6000
FLOOD-PLAIN WIDTH AT DOWNSTREAM CROSS SECTION {W},
FEET

FIGURE 13.—Relation between average downvalley ground-water
slope through the upstream and downstream ends of each reach in
the study area and width of the flood plain at the downstream cross
section used to compute the slope.

of the downstream cross section used in the slope
computation. The plot indicates that the spatial
variability in slope, after the effect of differences in
cross-section width is removed, is 10.25 x 10—3 or about
18 percent of the mean slope of 1.42X10—3. Assuming
that this variability reflects the spatial variation in T,
the bias error in the ground-water components attri-
buted to using a constant T is 18 percent of the value of
the component. For the ground-water inflow component
G1=124 acre-ft (0.153 hm3), this bias error is
:0.18X 124::22 acre-ft (: 0.027 hm3). The total bias in
G, due to the bias errors in W and T is then Ea]:
V 52+222=23 acre-ft ($0.028 hm3). Similar computa-
tions show that the total bias in Go'is EGO =i2l acre-ft
(:0.026 hm3) for budget period 688—708. These bias
errors are shown under the GI and GO components of
table 2.

Because the bias error is computed as a fraction of G,
the error will approach zero as G approaches zero. It was
assumed in this analysis, however, that a bias error
always exists because of the uncertainty in the
measurement of a zero slope. Thus, a bias error of $0.8
acre-ft (10.001 hm3) per day, which corresponds to the
average measurement error in slope (fig. 13), was
arbitrarily set as the minimum total bias error.

COMPUTATION OF ET AND TOTAL ERROR IN ET
The total ET for a budget period is obtained by

 

L25

algebraically summing the 12 components of the water
budget as expressed in equation 1. For the example
budget period in table 2, this summation gives ET=513
acre-ft (0.633 hm3) per 21 days.

The components of the water budget in reach 1 for
each budget period of the 1964 water year (table 1) have
been grouped into four principal sources of water (fig.
14) to illustrate the relative significance of each source.
The algebraic summation of the bar graph values for
any given budget period in the figure gives ET in acre-ft
per 14 days. The graph of surface water sources does not
indicate the amount of discharge in the Gila River and
its tributaries but rather the loss (or gain) of flow
through the reach during the budget period. The
primary components in the surface water sources are
the Gila River inflow (Q1) and outflow (Q0). The channel
storage (AC) and the tributary inflow (Q1) components
are generally only a small part of the total surface-
water source (see also table 1). The graph indicates that
surface water is the most significant source contribut-
ing to ET during the winter and late summer months of
the 1964 water year.

The soil-moisture components (A2175, All—40, and
Ail-TC) are the most significant sources of water in the
water budget during May and June, when the contribu-
tion from surface water is minimal and ET rates are
approaching a maximum. The precipitation (13) and
ground-water sources (G3, G1, and G0) are relatively
insignificant during most of the year, with the
ground-water components contributing a nearly con-
stant 50 acre-ft (0.062 hm3) per 14 days to ET
throughout the year.

This report has shown that nine of the components
contain significant sampling errors and three of the
components contain significant bias errors. Even
though some of these components are interrelated, the
measurement of each component is based on an
independent observation. Thus, the estimate of the total
error in ET is treated as an expected value of the error
variance of each term. The total sampling error in ET
may therefore be obtained from

2 2 2
EETS = (EQ[+EQO+EQT
+E§C+E%+E§+E§+E3+E%C)Vi (37)

where EET, is the total sampling error in ET and the
error terms on the right side of the equation are as
defined previously. For budget period 688—708, this
error is EETs=i344 acre-ft (:0.412 hm3).

The total bias error in ET may be obtained from

2 2 2 V2
EETb=(EGB+EGI+EGO) ,- (38)

L26

GILA RIVER PHREATOPHYTE PROJECT

REACH 1
1964 WATER YEAR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000
I I I I I I I I I I 1853— 1.2
Surface water (OI-00+A C+OTI —
800 — +.— 1.0
“'L “A — 0.9
700 — "I _. 08
600 — —-
- 0.7
500 — --— 0.6
400 — -—— 0-5
300 — ~-_ 0'4
— 0.3
200 — —-
(>13 — 0.2
é ‘00 " II H H “— 0.1
‘ E 0 _ I—I I—I I—I I1 .—. o
9 U U U
5 -100 — _— 0-1
UJ
u, —-0.2
g -200— _
o —0.3
< I I I I I I I I I I
E -300 I
”5 200
E I PI . . .I (F) I I I I I I I I — 0.2
§ 100 _ reCIpItatIon I—I ___ 0.1
g 0 H [—1 I—I I—I I l_| I—I F‘I I—II H fl I FL I_I I 0
U
I-
L“ 300 ‘
g I I I I _ I_ _I I I I I I _ 0 3
g 200 _ Soil-moisture change (AMS'I-AMC +AM7-cl _ I
E — 0.2
1- 100 —
< — 0.1
E
U U U u
—100 — ——0.1
—0.2
-200 —
—0.3
'300 — ‘—-o.4
400 I I I | I I | I I I I 415
200
I I I I I I I I I I I - 0.2
100 _ Ground water (GB +G,-Gol __ 01
oI—II_I HHHI—II‘II—IHHHI—II—II—II—II—I I—Ifl I—II—I H 0
OCT NOVI DEC I JAN FEB I MAR I APR I MAY I JUNE I JULY I AUG ISEPT
1963 1964

WATERoBUDGET COMPONENTS, IN CUBIC HECTOMETRES PER 14 DAYS

FIGURE 14.—Graph showing sources of water contributing toI the totalET per budget period for the 1964 water year, reach 1.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA

Where E ETb is the expected bias error inE T and the error
terms on the right side of the equation are as defined
previously. As in equation 37, equation 38 also assumes
that the bias errors of each component are independent
and unknown as to direction. For budget period
688-708, EETb =:69 acre-ft (10.085 hm3).

The total measurement error in ET attributed to both
the bias and the sampling errors is

EET =(E 212Ts +E 2ET,)1/2- (39)

For budget period 688—708, this total error is EET
=V (334)2+(69)2=:341 acre-ft ($0.420 hm3), which is
66 percent of the computed ET.

Figure 15A shows the ET values computed for each
budget period in reach 1 during the 1964 water year and
the errors associated with each ET value. The brackets
bounding these values define the total measurement
error (EET) in ET. Included within these brackets are
bars indicating the error in ET attributed to the
streamfiow components Q1 and Q0 and the error
attributed to the soil-moisture change components
A173 , AM—C , and All—4T0. The hydrograph of the Gila River
at cross section 9 in figure 15B shows that the

magnitude of the streamflow errors is directly related to .

the discharge, with the largest errors occurring during
periods of highest discharge.

The seasonal trend in ET is indicated in figure 15A by
the average potential ET curve. This curve was de-
termined by using the average daily temperatures and
the number of daylight hours for the study area (Blaney
and Criddle, 1962) and by assuming that sufficient
moisture is always available to satisfy the demand for
vaporization; therefore, the curve approximates the
upper limit of ET throughout the year. Actual ET may
exceed this potential curve, however, because the curve
is only an estimate of the potential rate and does not
account for all the factors controlling ET. Most of the
water-budget E T values in figure 15A which exceed this
curve contain measurement errors that fall well below
the curve, suggesting that the measurement errors are,
in most instances, at least as large as the expected
standard error in ET. Also, those ET values with the
lowest measurement error follow, in general, the trend
defined by the potential curve. Some ET values in the
water budget are negative, but their measurement
errors are generally large and extend into the positive
ET range. In a few instances the computed measure-
ment errors do not explain large negative ET values (as
in January in fig. 15A) or unrealistically high ET
values. These outliers generally occur during periods of
high streamflow and are assumed to reflect large
unmeasured changes in the stage-discharge relations
which are not fully accounted for in the streamflow error

 

L27

analysis. They may also reflect unknown quantities of
surface water moving into or out of depression storage
as described on page 9.

One of the most important points realized from
figures 14 and 15 is that the total measurement error in
ET is dependent on the volume of water moving through
the reach and not the magnitude of ET. This is
emphasized in figure 15A by the nearly constant error
in moisti . change for each ET value reflecting not the
large variation in moisture change shown in figure 14
but rather the tr ‘ a1 volume of soil moisture measured in
the reach whim fluctuates relatively little with time.

None of the 12 water-budget components has both a
sampling error and a bias error. This circumstance is
unique to this study area and should not be expected to
occur in other areas having the same type of compo-
nents, particularly if the components are of different
hydrologic significance. For example, in areas where
ground-water movement is comparatively large, the
sampling error may also be large and contribute
significantly to the total measurement error in ET. A
bias error in any one of the water-budget components
may also become significant if the frequency of data
collection or the sampling density do not adequately
describe the temporal and spatial changes in the
component.

During the 9-year study, a total of 416 ET values were
computed from reaches 1, 2, 2a, and 3. Table 19 lists
these ET values for each budget period and each reach
and gives the sampling error (EETS) and total error
(EET) for each ET value. About 60 percent of the ET
values have a measurement error which exceeds the E T
value. However, as noted previously, the assumptions
and criteria used in this analysis give measurement
errors which, in most instances, would be expected to
exceed the standard error of estimate.

COMPUTATION OF AET AND ERROR IN AE—T

One of the principle objectives of the Gila River
Phreatophyte Project is to determine the salvage of ‘
water as defined by the change in evapotranspiration *-
following removal of phreatophytes from the flood plain.
The average change in evapotranspiration derived from
ET data obtained before and after clearing for the
June-July period is presented in this section to
illustrate both the magnitude and the measurement
variability of this ET change.

The average change (MT) is
AE—T=E_TB —ETA, (40)

where E_TB and ETA are the average evapotranspira-
tion rates for given periods of time before and after

 

L28

GILA RIVER PHREATOPHYTE PROJECT

REACH 1
1964 WATER YEAR

 

2000 I

1 500

1 000

EVAPOTRANSPIRATION AND ERRORS IN ACRE-FEET

 

— Total + error in ET

I l l l l I I I

    

EXPLANATION

 
   
 

+ error in streamflow components 0, and 00 .
+ error in soil-moisture components AMS, AMC, AMTC
Computed ET for budget period

      
 

Potential ET

 

 

 

 

 

 

U)
>.
<
O 500
E
9:
Lu
0.
0
-500
-1000
m 20,000
5
p 10,000
o
“J
(I)
$
0 9
m <
0 o
L? E
3E
S o. 1000 i
a L?
0 Lu
rd an
*- 33
3 o
0 <
E E
2
n:
< 100
:1
o
50

 

OCT
1 963

NOV DEC

 

2 acre-feet

APR MAY
1964

JAN FEB R

 

JUNE JULY AUG SEPT

B

— 2.0

1.5

1.0

  
   

10.0

5.0

1.0

0.5

0.1

   

EVAPOTRANSPIRATION AND ERRORS IN CUBIC HECTOMETRES

GILA RIVER OUTFLOW (00), AT CROSS SECTION 9

FIGURE 15.——-Graphs of (A) ET and corresponding errors in ET per budget period and (B) outflow of Gila River at cross

section 9 for the 1964 water year, reach 1.

PER 14 DAYS

IN CUBIC HECTOMETRES PER 14 DAYS

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZONA L29
TABLE 19.—Euapotranspzratlon (ET) and total. measurement error of evapotranspzratzon (EET) for each budget period during water years
1963—71—reaches 1, 2, 2a, and 3
[All values are in acre-feet per budget period]
Reach 1
1963 water year 1964 water year 1965 water year
Budget Number Budget Number Budget Number
. period of days ET EET Period of days ET EET period of da) 5 ET EET
0&19—63 ______________ 14 521 1478 10—01—63 ______________ 14 728 1727 10—19—64 _____________ 21 229 +409
04—02-63 __ _ 14 30 249 10—15—63 _ - 14 293 210 11—09—64 ___ .. 21 132 177
04—16—63 -_ _ 14 129 248 10—29—63 _ _ 14 582 1106 11—30—64 ___ __ 21 132 262
04—30—63 _ _ 14 228 215 11—12—63 _ - 14 137 496 12-21—64 _ __ 21 113 275
05—14—63 __ _ 14 — 171 194 1 1—26-63 _ - 14 209 4 19 01—11—65 __ 21 42 408
05-28—63 1. _ 14 312 193 12—10—63 — — 14 525 364 02—01-65 , _, 21 —193 776
06—11—63 -_ 1 14 389 178 12—24—63 . _ 14 494 315 02—22—65 - .1 21 465 757
06—25-63 __ 1-. 14 220 158 01—07—64 _ ~ 14 230 311 03—15—65 _ -1 21 247 655
07—0963 __ ___ 14 306 159 01—21—64 _ — 14 —604 438 04—05—65 _ A, 21 33 545
07—23—63 .- _-. 14 254 165 02—04—64 _ - 14 —40 445 04—26—65 _ __ 21 378 470
08-06-63 -- _ 14 446 598 02-18—64 - _ 14 97 296 05—17—65 _ __ 21 82 335
08—20—63 _- _ 14 851 490 03-03-64 _ — 14 9 200 06—07—65 __ __ 21 407 187
09—03—63 _- _ _ 14 4122 1997 03—17—64 _ — 14 195 192 06—28-65 __ _- 21 448 159
09—17—63 14 2556 1336 03—31—64 _ — 14 196 197 07—19-65 __ __ 21 469 164
04—14—64 - — 14 112 195 08—09-65 __ __ 21 1979 1294
05—04—64 - - 20 269 200 08—30-65 __ _ 21 265 600
05—25—64 _ - 21 410 184 09—20-65 _____________ 21 1224 1247
06—15—64 _ _ 21 372 172
07—06—64 _ A 21 424 161
07—27—64 _ - 21 1056 819
08-17-64 _ _ 21 1903 1427
09—07—64 _ _ 21 513 341
09—28-64 ______________ 21 2799 1732
1966 water year 1967 water year 1958 water year
Budget Number Budget Number Budget Number
period of days ET EET period of days ET EET period of days ET EET
10—11—65 _____________ 21 436 1206 10—03—66 _____________ 21 366 +1096 10—16-67 446 1805
11-01—65 - 21 302 192 10—24—66 _ __ 21 109 264 11—06—67 _ 159 306
11—22—65 _ 21 265 221 11—14—66 _ _ 21 350 258 11~27~67 212 302
01—24-66 - 63 12—05—66 _ __ 21 263 325 12—18—67 397 464
02—14—66 _ 21 —228 1467 12—19—66 _ __ 14 110 215 01—08—68 _
0&07—66 _ 21 1702 1464 01—16—67 _ __ 28 44 286 01-29—68 1767 2181
0&28—66 _ 21 02—06—67 _ _- 21 144 363 02-19—68
04—18-66 _ 21 02—27—67 _ - 21 105 206 03-11—68 _
05-09—66 _ 21 999 1044 03—20—67 __. - 21 -273 206 04—01—68 _ 1228 3589
05-30-66 _ 21 683 456 04—10—67 -__ _ 21 199 190 04—22—68
06—20-66 _ 21 671 228 05—01—67 _ __ 21 75 181 05—13—68 -999 1649
07—11-66 _ 21 738 188 05—22—67 ___ _ 21 221 169 06—03—68 752 882
08—01—66 _ 21 405 428 06—12—67 _ _ 21 327 162 06—24—68 403 314
08—22—66 _ 21 911 501 07—03—67 . _ 21 200 166 07—08—68 351 211
09—12—66 _____________ 21 564 633 07—24—67 1 _ 21 413 1048 07—22—68 25 173
08-14—67 _ _ 21 08—05—68 254 426
09—04—67 _ __ 21 08—19—68 —55 998
09—25—67 21 738 810 09—02—68 55 632
09—16—68 _ 348 317
09—30—68 90 178
1969 water year 1970 water year r year
Budget Number Budget Number Budget Numb: r
period of days ET EET period of days ET EET period of days ET EET
10—14—68 14 —42 :188 10—13—69 ————————————— 14 208 +169 10—05—70 568 ‘1750
10—28—68 14 161 175 10—27—69 . - 14 207 221 10—19—70 113 306
11—11—68 14 97 225 11—10—69 7 -7 14 —67 221 11—02—70 97 177
11—25—68 14 83 384 11—24—69 _ - 14 112 292 11—16—70 41 187
12—09—68 14 -31 426 12—08—69 _ _ 14 86 346 11—30—70 14 179
12—2368 14 244 327 12—22—69 _ __ 14 —87 360 12—14—70 55 184
01—06-69 14 —235 501 01-05—70 _ _ 14 106 227 12—28—70 _ 5 184
01—20—69 14 317 535 01—19—70 _ __ 14 117 247 01—11—71 126 328
02—03-69 14 78 663 02—02—70 1 - 14 98 215 01—25—71 —40 376
02-17—69 14 199 573 02—16-70 . - 14 40 208 02—08-71 91 351
03—03—69 14 105 350 03—09—70 - _ 21 127 435 02—2271 _ 356 326
03—17—69 14 99 251 03—23—70 _ _ 14 81 242 03—08—71 —8 274
03—31—69 14 O 225 04—06—70 _ _ 14 135 215 03—22—71 —88 212
04—14—69 14 56 234 04—20—70 _ _ 14 18 207 04—05—71 _ 27 186
04—28-69 14 96 213 05—04—70 _ A 14 19 204 04-19—71 _ 74 184
05—12—69 14 98 212 05—18—70 _ - 14 72 186 05—03-71 —20 180
05—26-69 14 200 191 06—01—70 _ - 14 23 177 05—17—71 41 173
06—09—69 14 138 184 06-15—70 _ _ 14 136 170 05—31-71 _ 82 169
06-23—69 14 192 176 06—29—70 _ - 14 269 166 06—14—71 93 165
07—07—69 14 228 176 07—13—70 - - 14 154 164 06—28—71 106 163
07—21—69 14 183 261 07—27—70 - _ 14 209 176 07—12—71 203 165
08—04—69 14 89 183 08—10—70 , _ 14 153 282 07—26—71 192 177
08—18—69 14 154 223 08—24—70 - _, 14 183 235 08—09—71 —728 462
09-01-69 14 —17 193 09—07—70 _ 14 262 168 08-23—71
09—15—69 14 ~233 702 09—21—70 ______________ 14 193 196 09.06.71 518 626
09—29—69 14 183 279 09—20—71 — 183 463

L30 GILA RIVER PHREATOPHYTE PROJECT

TABLE 19.—Evapotranspiration (ET) and total measurement error of evapotranspiration (Em) for each budget period during water years
1963—71——reaches 1, 2, 2a, and 3—Continued

 

 

 

 

     

   

 

 

 

    

 

 

 

 

 

  
 

Reach 2
1963 water year 1964 water year 1965 water year
Budget Number Budget Number Budget Number
period of days ET E ET period of days ET EET period of days ET EET
07—09—63 14 427 1154 10—08—63 21 —293 $749 10—12—64 _____________ 21 —409 $1495
07-23—63 _ 14 573 130 10—22—63 _ _ 14 111 967 11—02—64 _ 21 363 177
08—06—63 _ 14 —195 566 11—05—63 _ _ 14 535 543 11—23—64 _ 21 85 202
08—20—63 _ 14 ~455 468 11—19—63 _ _ 14 437 457 12—14—64 __ _ 21 145 269
09—03—63 _ 14 329 1832 12—03—63 _ _ 14 —53 360 01-04—65 __ _ 21 197 227
09—17—63 _____________ 14 -755 1275 12—17—63 _ 14 —246 322 01—25—65 _ 21 283 712
12—31—63 _ 14 97 219 02—15—65 _ 21 293 728
01—14—64 _ _ 14 27 363 03—08—65 __ - 21 328 680
01—28—64 _ _ 14 167 453 03—29—65 __ _ 21 —391 564
02—11—64 _ . 14 92 368 04—19—65 __ - 21 —65 491
02—25—64 _ - 14 104 222 05—10—65 __ _ 21 793 368
03—10—64 _ - 14 302 191 05—31-65 __ -__ 21 616 171
03-24—64 _ _ 14 103 185 06—21—65 __ _ 21 563 128
04—07—64 _ _ 14 246 159 07—12—65 __ _ 21 832 129
04—27—64 _ _ 20 143 168 08-02—65 __ _ 21 1195 946
05—18—64 _ _ 21 387 153 08-23—65 __ 21 842 821
06—08-64 _ _ 21 728 135 09—13—65 __-- 21 1188 1184
06—29—64 _ - 21 789 141
07—20—64 _ _ 21 1058 647
08—10—64 _ _ 21 1068 1257
08—31—64 _ -_ 21 537 763
09-21—64 _____________ 21 439 791
1966 water year 1967 water year 1963 water year
Budget Number Budget Number Budget Number
period of days ET EET period of days ET EET period of days ET EET
10-04—65 -._ _____ 21 763 1346 10—17—66 _____________ 795 :284 10—02—67 ______________ 21 —212 :1070
10—25—65 __ _ 21 562 148 11—07—66 _ 323 206 10—23—67 __ - 21 219 421
11—15—65 __ _ 21 250 181 11—28—66 _ 64 314 11.13—67 __ _ 21 408 291
12—06—65 __ - 21 253 290 12—12—66 _ 132 210 1204—67 _ 21 318 323
02—07—66 _- _ 60 01-09—67 _ 250 253 12—25—67 _ 21
02-28—66 __ _ 21 —838 1549 01—30—67 _ —48 355 01-15—68 _ 21
03-21—66 __ _ 21 02—20—67 _ 54 214 02-05—68 __ 21
04—11—66 __ _ 21 03—13—67 _ 225 218 02-26—68 __ 21
05—02—66 __ _ 21 03—27-67 _ 17 184 03—18—68 _ 21
05-23-66 __ _ 21 04—17—67 _ 285 163 04—08—68 __ 21
06—13-66 __ , 21 05—08—67 _ 271 \ 146 04—29—68 _ 21
07-04—66 __ _ 21 05—29—67 _ 720 135 05—20—68 _ 21
07—25—66 __ _ 21 06—19—67 _ 739 131 06—10—68 _ 21
08—15—66 __ _ 21 604 221 07—10-67 _ 974 170 07-01—68 _ 21
09—05—66 __ _ 21 460 738 07—31—67 _ 531 991 07—15—68 _ 14
09—26—66 ______________ 21 ~98 1117 08—21—67 07-29—68 _ 14
09—11—67 464 664 08—12—68 _ 14
08—26—68 __ _ 14
09—09—68 _ 14
09—23—68 _ - 14
1969 water year 1970 water year water year
Budget Number Budget Number Budget Number
period of days ET EET period of days ET EET period of days ET E ET
10—07—68 14 10—06-69 _____________ 14 194 1128 10—12-70 -— ———————— 14
10—21—68 _ 1- 14 10—20—69 14 419 122 10—26—70 __ 14 187 $143
11—04—68 14 11—03—69 _ 14 194 189 11—09—70 - 14 71 147
11—18—68 — 14 11—17—69 _ 14 344 183 11—23—70 -_ 14 177 140
12—02—68 _ 14 12—01—69 _ 14 24 312 12—07—70 __ 14 122 146
12—16—68 _ _ 14 12—15—69 _ 14 474 352 12—21—70 __ 14 110 148
12—30—68 _ 14 12—29—69 _ 14 220 243 01—04—71 _ 14 177 199
01—13—69 _ 14 01—12—70 _ 14 1 228 01-18—71 __ 14 '91 353
01—27—69 _ - 14 01—26—70 _ 14 28 198 02—01—71 __ 14 31 351
02>10—69 _ 14 02—09—70 _ 14 —62 179 02—15—71 _ 14 —134 318
02—24—69 _ 14 02-23—70 _ 14 179 174 03—01-71 _ 14 260 270
03—10—69 _ 14 03—16—70 _ 21 306 436 03—15-71 _ 14 53 226
03-24—69 - _ 14 03—30—70 _ 14 165 202 03—29-71 _ 14 8 161
04—07—69 _ 14 04—13—70 _ 14 270 179 04—12—71 14 170 146
04—21—69 _ 14 04—27—70 _ 14 143 179 04—26—71 - 14 268 143
05—05—69 _ 14 05—11—70 _ 14 257 162 05—10—71 -1- 14 125 143
05—19—69 - 14 05—25-70 - 14 272 145 05—24—71 - 14 127 135
06—02—69 _ 14 289 1139 06—08—70 _ 14 314 133 06—07—71 14 205 125
06—16-69 _ 14 527 128 06-22—70 1 14 520 124 06—21—71 _ _ 14 188 122
06—30—69 _ 14 360 119 07—06—70 - 14 583 122 07—05-71 _ _ 14 144 122
07-14—69 _ 14 510 125 07—20—70 _ _ 14 587 125 07—19—71 _ _ 14 245 154
07—28—69 _ 14 1026 183 08—03—70 _ 14 479 136 08—02—71 _ - 14 340 429
08—11—69 _ 14 721 125 08—17—70 _ _ 14 558 280 08—16-71 _ _ 14
08—25—69 _ 14 576 155 08—31—70 _ 14 505 154 08—30—71 _ - 14
09—08—69 _ 14 09—14—70 _ 14 507 137 09—13—71 A _, 14
09—22—69 14 09—28—70 14 275 156 09—27—71 ___ ______ 14

 

 

 

 

 

TABLE 19.-—Evapotranspiration (ET) and total measurement error of evapotranspiration (EET) for each budget period
1963—71—reaches 1, 2, 2a, and 3—Continued

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, AR] ZONA

L31

during water years

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

  

Reach 2a
1966 water year 1967 water year 1 968 water year
Budget Number Budget Number Budget Number
period of days ET E ET period of days ET E ET period of day 5 ET E ET
10—04—65 ............. 21 10—17—66 _____________ 21 —621 1349 10—02—67 .............. 21
10—25—65 - 21 11—07—66 _ 21 228 205 10—23—67 _ 21
11—15—65 1 21 11—28—66 _ 21 180 312 11-13—67 - 21 314 i290
12—06—65 _ 21 12—12—66 _ 14 198 209 12—04—67 _ 21 188 325
02—07—66 _ 63 01—09—67 - 28 92 255 12—25—67 _ 21
02—28—66 _ 21 01-30-67 _ 21 —125 361 01—15—68 _ 21
03—21—66 _ 21 02-20-67 _ 21 —78 219 02—05—68 21
04—11-66 _ 21 03-13-67 _ 21 102 212 02—26—68 21
05-02—66 21 03—27—67 _ 14 28 177 03-18—68 _ 21
05—23-66 ___ 21 04—17—67 - 21 195 167 04—0&68 _ 21
06—13-66 _ 21 05—08—67 _ 21 86 149_ 04—29-68 _ 21
07-04—66 ___ 21 656 1'178 05—29—67 - 21 388 137 05—20—68 _ 21
07-25-66 - 21 663 154 06—19—67 - 21 343 139 06—10—68 _ 21
08-15-66 , 21 203 223 07—10—67 _ 21 323 183 07—01—68 - 21
09—05—66 _ 21 07—31—67 - 21 07—15—68 _ 14
09—26—66 08—21—67 _ 21 07—29—68 _ 14 151 177
09—11—67 _____________ 21 08—12—68 _ 14
08—26—68 _ 14
09—09—68 _ 14
09—23—68 _ _ 14 328 173
1969 water year 1970 water year 1971 wate r year
Budget Number Budget Number Budget Number
period of days ET E ET period of days ET EET period of dag 5 ET E ET
10—07-68 _____________ 14 379 :139 10—06—69 _____________ 34 $153 10—12—70
10—21—68 _ 14 232 144 10—20—69 _ 249 128 10—26—70
11—04—68 _ 14 27 170 11—03—69 _ 45 198 11—09—70
11—18—68 _ 14 106 255 11—17—69 _ 269 186 11—23—70 55 $141
12—02—68 _ 14 230 416 12—01—69 _ -114 321 12—07—70 73 145
12—16—68 _ 14 —31 357 12—15—69 _ 262 356 12—21—70 32 147
12—30—68 _ 14 42 418 12—29—69 _ 35 248 01—04—71 138 199
01—13—69 _ 14 ~325 487 01—12—70 _ —9 228 01—18-71 ~11 351
01—27—69 _ 14 —111 567 01—26—70 _ —36 199 02—01—71 _ 216 344
02—10—69 _ 14 —706 683 02—09—70 _ —131 178 02—15—71 —237 321
02—24—69 _ 14 47 452 02-23—70 - 132 17 1 03—01—71 123 272
03-10—69 _ 14 55 251 03—16—70 - 327 425 03—15—71 _ 10 228
03—24—69 _ 14 186 207 . 03—30—70 _ 125 198 03—29—71 —7 159
04—07—69 _ 14 102 193 04—13—70 , 202 177 04—12—71 95 145
04—21—69 _ 14 141 193 04—27—70 _ 160 174 04—26—71 _ 121 143
05—05—69 _ 14 223 169 05—11—70 _ 161 162 05—10—71 77 137
05—19-69 _ _ 14 45 169 05-25—70 _ 66 145 05-24—71 112 128
06-02—69 _ 14 68 142 06—08—70 _ 111 134 06—07-71 _ 92 125
06-16—69 _ 14 218 130 06—22—70 1 153 125 06-21—71 109 121
06—30-69 ___ 14 146 128 07—06-70 . 229 122 07-05-71 77 120
07—14—69 ___ 14 187 126 07—20—70 - 180 120 07-19-71 155 131
07-28-69 14 302 165 08—03—70 _ 151 135 08—02—71 .
08-11-69 ___ 14 301 132 08—17—70 _ 169 290 03-16—71
08-25—69 _ 14 350 174 08—31—70 _ 232 149 03-30-71
09—08—69 14 09—14—70 _ 262 130 09-13-71 ._ 219 365
09—22-69 ............. 14 09—28-70 - 40 159 09-27—71
Reach 3
1964 water year 1965 water year
Budget Number Budget Number
period of days ET E ET period of days ET E ET
10—08—63 _____________ 21 10—05—64 _____________ 21
10—22—63 ___ 14 10—26—64 _ 21
11—05—63 ___ 14 11—16—64 _ 21 239 :114
11—19-63 _ ___ 14 12—07—64 A 21 46 246
12—03—63 _ _ __ 14 12-28—64 _ 21 108 209
12—17-63 ___ 14 01—18—65 - 21
12—31—63 _ __ 14 02—08—65 _ 21
01—14—64 __, 14 03—01—65 _ 21
01—28-64 ___ 14 25 +446 03—22-65 - 21
02—11—64 _ 14 36 361 04—12-65 _ 21
02-25—64 - 14 - 27 209 05—03—65 - 21 400 390
03—10—64 __- 14 47 168 05—24—65 _ 21 374 191
0&24—64 .1- 14 —34 170 06—14—65 _ 21 661 100
04—07—64 _ __ 14 70 165 07—05—65 , 21 702 83
05—11—64 1 A 34 246 190 07—26—65 _ 21
06—01—64 - 21 433 95 08-16—65 _ 21
06—22-64 _ 21 541 94 09—06—65 - 21
07—13-64 _ ___ 21 580 90 09—27—65 .............. 21
08—03—64 _ _ _ 2 1
08—24—64 __ 21

 

 

 

 

 

L32

clearing, respectively.

Table 20 gives ET values determined from reach 1
during the season of high potential ET in June and July
for the before- and after-clearing period of the study.
Included in the table are the sampling error (EETS) and
total measurement error (E ET) for each ET value. The
total bias error is nearly constant at EETb =:46 acre-ft
($0.057 hm3) per 14 days and has been omitted from the
table. The ET values for some budget periods are
obviously unreasonable and generally correspond with
a large measurement error. Thus, those values with a
total measurement error of EET>:550 acre-ft ($0.678
hm3) are not included in the subsequent computations.
The criteria used in selecting this error limitation was
arbitrarily established such that no data were accepted
in which the error exceeded the average June and July
potential ET of approximately 550 acre-ft (0.678 hm3)
per 14 days. A plot of the ET values in table 20 with
measurement errors less than :550 acre-ft ($0.67 8
hm3) is shown in figure 16.

These ET data define average 14—day rates of
ETB=320 acre-ft (0.395 hm3) before clearing and
ET A =181 acre-ft (0.223 hm3) after clearing. The
average change (reduction) in ET as a result of clearing
is AET=320—181=139 acre-ft (0.171 hm3) per 14 days
for the J une—July period in reach 1.

The standard deviation of the 12 before-clearing ET
values in table 20 is SETB=i79 acre-ft (:0.097 hm3)
or: 25 percent of ETB. The standard deviation of the 19
after-clearing ET values is SETA =i77 acre-ft
(:0095 hm3) or :43 percent of WA. The near-equal

BEFORE CLEARING

 

 

 

 

 

 

GILA RIVER PHREATOPHYTE PROJECT

TABLE 20.—ET values obtained before and after clearing and their
corresponding sampling, and total measurement errors for selected

14-day budget periods during June and July, reach 1
[All values in acre-ft per 14 days}

 

 

 

   

 

 

Before clearing After clearing
Day‘ ETB EETS EETB Day‘ ETA EETs EETA
:172 :178 :140 £151
151 158 143 154
153 159 857 859
159 165 261 267
596 598 205 211
153 163 167 173
141 152 424 426
671 673 178 184
161 171 170 176
138 149 170 176
142 153 256 261
192 200 178 183
160 170 163 170
352 357 160 166
158 164
170 176
159 165
157 163
158 165
171 177
k=12 periods k=19 periods
ETB =320 ETA=181
s— =r79 s— 2:77
E B E A
EE_ =:182 Efi =1199
s s
E—— =1189 E— =t205
E B ETA

 

1D? refers to last day of 14-day budget riod.

ZbE dand error;l values originally compute for 21~day budget period but adjusted to 14-day
u get perio .

3ET and error values excluded from computations because E ET >r550 acre-ft.

values of SETB and s ETA indicate that the variability in

the computed values of ET is a function of the total
volume of water passing through the reach—which does

AFTER CLEARING

 

 

 

 

 

500 5 _ —0.6 w
>.
° 3
E 400 — . — _ _ — 0.5 g
D ETB =320 acre-feet a:
3 " E
______ _. __._._=___.__ _ —0.4 U3
1

g; 300— o . ' T — E
+— D o __ o o {E
w 0 . ... AET=139 acre-feet ._ o 3 2
EL 0 o " . .9
E 200 — — . 0 ‘ . 8
o —. _______________ —_ -_—_— o 2 I
< _ . . ' o
E ETA =181 acre-feet . . g

I; 100 — _ O
lu -- . — 0 1 g
k.
0 Lu

JUNE JULY JUNE JULY

FIGURE 16.—Water-budget ET values in reach 1 for the before-and after-clearing periods in June and July and the average change in
ET as a result of phreatophyte clearing. Values are plotted on the day corresponding to the middle of the budget period.

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARI ONA

not change appreciably with time—and not the mag-
nitude of ET.

The average measurement errors associated with
ETB and ETA may be computed from

k
EET=kl( Z EIZCTt)l/2’
t=1

where Efi is the average total measurement error of
ET for k budget periods and EET; is the total measure-
ment error in ET for budget period t. Applying equation
41 using the total measurement errors EETB and EETA
in table 20 gives EE—TB =i 189 acre-ft ($0.233 hm3) and
E1771, =:205 acre-ft ($0.253 hm3).

Assuming independence between E_TB and E_TA, the
standard deviation of AET is 3m: VsZETB+32fiA=
V792+772 =: 110 acre-ft (:O.136 hm3) ori79 percent of
MT. Much of this variability is real, reflecting year to
year variations in the potential ET rate and in the
moisture available for evaporation and transpiration.
The remaining variation results from measurement
errors in the ET components.

The average measurement error for AET includes
only the sampling errors (Efis) computed from equa-
tion 37; the bias errors (Efib) are omitted because they
are one directional and essentially constant for all
estimates of ET and thus cancel in the error computa-
tion. The average sampling errors (Efis) of t 182 acre-ft
(:0224 hm3) before clearing and i199 acre-ft (:0.245h
m3) after clearing were obtained by substituting the
corresponding EETS error values of table 20 in equation
41. The average measurement error in AET is then
Em: V 1822+1992 =i270 acre—ft (: 0.333 hm3), which
is nearly : 200 percent of MT and exceeds the standard
deviation of ET by 21/2 times.

The fact that these measurement errors in ET and
AET are significantly greater than their standard
deviations indicates that the assumptions and criteria
used to obtain the measurement errors produce an
overestimate of the true measurement variability in
ET. These total measurement errors must therefore be
considered only an indicator of the relative significance
of each ETyalue.

No evaluation was made of the winterET rates in this
report; however, Hanson, Kipple, and Culler (1972, fig.
4) showed that the winter rates average substantially
lower than the summer rates before clearing and that no
significant change in the winter rates can be detected
after clearing. The measurement errors in ET are also
generally higher during the winter than during the
spring and early summer months as indicated in figure
15A. Estimates of ET for the winter months of typically
low rates are therefore less reliable than the summer
estimates.

(41)

 

 
 
 
 
 
 
 
  
 

L33

DISCUSSION OF RESUL S

Of the 12 components of the water b dget, the Gila
River inflow (Q1) and outflow (Q0) are generally the
most significant, reaching maximum r tes during the
winter and spring snowmelt period and uring the late
summer thunderstorm period. Tributa inflow (QT)
occured only 4 percent of the time durin the nine-year
study period, and even though some of t ese events did
produce large volumes of inflow, this component is
considered to be one of the least signifi ant during the
study period. Of the more important co ponents in the
water budget, moisture-content Chang s—AMS, AM 1,
MC, and AMTc—are the most diffic It to measure.
Moisture change, particularly in the apillary zone,
generally becomes significant during eriods of low
streamflow. Except for basin-fill infl , which was
assumed constant, the ground-water i flow (G1) and
outflow (G0) are the least variable com onents during
the year, fluctuating only in respon e to seasonal
changes in the downvalley ground-w ter slope. The
only component of any consequence in t e water budget
that was not measured in this stud is depression
storage—that water which fills side channels and
depressions in the flood plain during ove bank flooding.
The relatively infrequent occurrence of depression
storage and the difficulties in measuri g this compo-
nent did not justify including it in th water-budget
analysis.

The total measurement error for most of the
water-budget components consists pri arily of a sam—
pling error which is dependent on he number of
observation points used to evaluate the omponent. The
sampling error is time variant—refle ting both the
variability in repetitive measurement and the error
due to missing data. Included in the tota measurement
error is a bias error which reflect a consistent
overestimate or underestimate of th water-budget
component. Only the basin—fill inflow a d the ground—
water inflow and outflow componen s introduce a
measurable bias in the computation of T. Because of
the uncertainty in the estimate of the b sin—fill inflow,
the bias of this component is assumed to equal the total
basin-fill inflow. The bias in the grou d—water inflow
and outflow components reflect possibl errors in the
determination of the average transm‘ssivity for the
study area and inaccurate measurements of the width of
saturated alluvium at the inflow and outflow cross
sections of each reach.

The magnitude of the measurement error of ET is
directly related to the total volume of water moving
through the reach and not the magnitude of ET. Thus,
ET computed from a budget period of high streamflow
has a correspondingly large measurement error. Fortu-
nately, high streamflow is generally limited to the

 

L34

winter months, when ET is minimal and a few weeks in
late summer when runoff from thunderstorms occurs.
During the midsummer months of maximum ET, the
measurement errors become minimal because of low
streamflow and negligible tributary inflow and precipi-
tation.

The measurement errors of ET for the summer
periods investigated in this report (table 20) are :59
percent of the computed average before-clearing value
of ET and :113 percent of the computed average
after-clearing ET rate. The measurement error of the
average change in ET as a result of clearing is nearly
:200 percent of the computed change for these summer
periods. The measurement errors of ET and change in
ET for the winter periods are generally even greater
than for the summer periods.

The large measurement errors computed in this study
would make it appear that the ET rates derived from the
water budget do not provide reliable estimates of the
true ET rates. Most of these computed errors can be
assumed, however, to exceed the actual measurement
errors of ET and AET because the criteria used to
estimate the error of each component give values that
would be expected to exceed their standard error. This is
substantiated by a comparison of the Efi values with
the significantly lower sE—T values in table 20. Because
sfi includes both the true measurement errors in the
data and real variations reflecting year to year
differences in moisture available for evapotranspira-
tion, it is apparent that the computed measurement
errors are too high. These data show, in fact, that
reliable estimates of ET for the summer periods can be
obtained and that a significant difference inE T could be
detected as a result of clearing the phreatophytes from
the flood plain.

Even though most of the computed measurement
errors for ET probably exceed the actual measurement
errors, they do provide a good indication of the relative
significance of each ET value. These measurement
errors were used as the basis for selecting the most
reliable ET estimate in evaluating the average before
clearing and after clearing ET rates from all reaches in
the study. A discussion of the application of these
measurement errors to the evaluation of the average ET
rates will be included in a subsequent paper in this
series.

Studies have been carried out to evaluate the
variability in the ET data due to differences in moisture
available for vaporization and differences in the
potential to remove the available water. In addition, the
differences in ET between reaches due to differences in
vegetative cover has been evaluated. The results of
these studies will also be included in a subsequent
report.

 

GILA RIVER PHREATOPHYTE PROJECT

DEVELOPMENT OF EQUATIONS DESCRIBING
UNADJUSTED SAMPLING ERROR, V

An estimate of the average value of a hydrologic
variable for a given area such as precipitation ap-
proaches the population mean (P) as the number of
sample points used for the estimation increases.
Measures of most hydrologic variables are sample
realizations of the time series in which they occur and
therefore are frequently autocorrelated. Thus, the rate
at which any estimate of a given variable approaches its
mean value is unknown.

If n sample points are used to estimate the mean 13”,
the standard deviation of the departure between P" and
17", (where m<n) decreases as m approaches n in a
manner indicated by curve Sm of figure 17. In this report
Sm is referred to as the average missing-data error. A
sampling error, V, exists at m=n, and a relation
between sample size and total error which includes both
Sm and V can be described by curve E13," in figure 17. An
approximation to the shape of curve E13,” can be found by
estimating V such that

E213M=§2m+V2, (26)

where Ezfi'm is the variance of the departure between 13m
and an estimate of the population mean, and is
inversely proportional to the sample size m.

The procedure for determining Sm for all values of m
has been described previously in this report. The
purpose of this section is to describe the development of
the equations used in estimating the sampling error, V.

An approximation of V may be obtained from an
evaluation of the rate of change in Sm as m approaches
n. If the variance S2”, is defined from m samples and the

 

 

ERROR IN DEPARTURES

Zero departure A

 

 

 

 

NUMBER OF SAMPLES (m/‘——>

FIGURE 17.—General relation for error in departures of average of m
samples from average of n samples (curve Sm), and error in
departure of average of m samples from estimate of population
mean (curve E13,").

ACCURACY OF EVAPOTRANSPIRATION RATES DETERMINED BY WATER-BUDGET METHOD, ARIZ

variance S22”, is defined from 2m samples, then the ratio
§2fm + V2
§§m+v2
can be assumed that the average rate of change in the
square of this residual error will approach zero as m
increases—or expressed in equation form,

6

_.§_ 2 _6 2 i 2 _
6V 1,2+av§2,4+av§4,8+ +aV§n'/2,n'_0’

is equal to 2 with some residual error, Emfim. It

where
g 2 §%+V2 ‘ C(43)
1’2 §§+V2 ’
E §§+V2 (44)
2’4 §§+V2
§Z+V2 5
=2—_— , (4 )
‘13 S§+V2
§g.,2+\72
’ ,= _~_—_, , (46)
n /2,n SZ,+V2
n
and

n’ is the largest even value less than or equal to n.
Solving the partial differential equation
E? + V2

—2 2
s2+v

 

i 52 _%(2_

UV _ 132—. )gives

V‘Wfi—Efiwv2(4§§§§_3§?_§‘11)

+(3§§§§—2§3—§§§§)=0. (47)

By replacing each Big? term in equation 42 with a
general expression for equation 47 gives the previously
shown quadratic equation

aV4+bV2+c=0, (22,
where
n’/2
_‘ _2
022(831‘82ml (23)
m=1
n72 2 2
‘2‘ “ -2 —
b =Z(45,,, 82m _ 382m _S;ll ), (24)
m=1

.(42) ,

 

ONA L35
n'/2 _2 _2 _ __ _2
and 022(38," s,,,,—2s§,,,—s;§,s2,,, . «25)
m=1
The positive root of V in equation 22 may then be
obtained from
V2 L’lfl’fi'iqg (48)
2a

Solving for V in equation 48 using the co

efficients a, b,

and 0 obtained from equations 23—25 gives the least

squares best-fit curve for V.

REFERENCES CITED

Blaney, H. F., and Criddle, W. D., 1962, Determining consumptive use

and irrigation water requirements: US. Dept. A
Bull. 1275, 59 p.

griculture Tech.

Bowie, J. E., and Kam, William, 1968, Use of water by riparian

vegetation, Cottonwood Wash, Arizona: US.

Water-Supply Paper 1858, 62 p.

Geol. Survey

Burkham, D. E., 1976, Flow from small watersheds adjacent to the

study reach of the Gila River Phreatophyte Proje
Geol. Survey Prof. Paper 655—1, 19 p.

ct, Arizona: US.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis of streamflow

data for an alluvial stream: U.S. Geol. Survey Fr
13 p.

Carter, R. W., and Anderson, I. E., 1963, Accuracy
measurements: Am. Soc. Civil Engineers Proc.,
Jour., v. 89. no. HY4, p. 105—115.

Chayes, Felix, 1971, Ratio correlation—A manual
petrology and geochemistry: Chicago, Univ. Chi

Culler, R. C., and others, 1970, Objectives,
environment—Gila River Phreatophyte Project

of. Paper 655—C,

3f current meter
Hydraulics Div.

for students of
:ago Press, 99 p.
methods, and
Graham Coun-

ty, Arizona: US. Geol. Survey Prof. Paper 655- A, 25 p.
Gatewood, J. 5., Robinson, T. W., Colby, B. R., Hem, J. D., and

Halpenny, L. C., 1950, Use of bottom-land veg
Safi‘ord Valley, Arizona: US. Geol. Survey Wat
1103, 210 p.

Hanson, R. L., 1972, Subsurface hydraulics in the
River Phreatophyte Project, Graham County, Ar
Survey Prof. Paper 655—F, 27 p.

Hanson, R. L., Kipple, F. P., and Culler, R. C., 197

station in lower
er-Supply Paper

area of the Gila
.zona: U.S. Geol.

2, Changing the

consumptive use on the Gila River flood plain, southeastern

Arizona, in Age of changing priorities for land

Soc. Civil Engineers, Irrigation and Drainage

Conf., Proc., p. 309—330.

and water: Am.
Div. Specialty

Thiessen, A. H., 191 1, Precipitation for large areas: M onthly Weather

Rem, v. 39, p. 1082—1084.
Turner, S. F., and Skibitzke, H. E.,

1952, Use of water by

phreatophytes in 2000 foot channel between Granite Reef and

Gillespie Dams, Maricopa Co., Arizona: Am.
Trans, v. 33, no. 1, p. 66—72.

Eeophys. Union

Weist, W. G., J r., 1971, Geology and ground-water system in the Gila
River Phreatophyte Project area, Graham County, Arizona: US.

Geol. Survey Prof. Paper 655—D, 22 p.

 

 

 

 

 

Effects of Phreatophyte
Removal on Water QUality
in the Gila River Phreatophyte

Project Area, Graham County
Arizona

By R. L. LANEY

With ajsection on STATISTICAL ANALYSIS
By H. W. HJALMARSON

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—M

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1977

UNITED STATES DEPARTMENT OF THE INTERIOR
CECIL D. ANDRUS, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Director

Library of Congress Cataloging in Publication Data

Laney, R. L.

Effects of phreatophyte removal on water quality in the Gila River phreatophyte project area, Graham County, Arizona.

(Gila River phreatophyte project)

(Geological Survey Professional Paper 655—M)

Bibliography: p.

1. Water quality—Arizona—Gila River watershed. 2. Phreatophytes—Control—Arizona—Gila River watershed. 3. Plant-water
relationships.

I. Hjalmarson, H. W., joint author. 11. Title: Effects of phreatophyte removal on water quality. . . III. Series. IV. Series:
United States Geological Survey Professional Paper 655—M.

QE75.P9 no. 655—M [TD224.A7] 557.3'085 [551.4'8] 77—608027

 

For sale by the Superintendent of Documents, US. Government Printing Office
Washington, DC. 20402
Stock Number 024—001—02967—7

 

CONTENTS

 

    
 
 
 
  
 

 

 

  

 

 

 

 

 

 

 
 
 
 
  

Page Page
Factors for converting English units to International System Water quality—Continued
units (51) ................................................................................ IV Alluvial deposits ...................................................................... M10
Abstract ............ M1 Ground water in the saturated zone .............................. 10
Introduction ............. 1 Ground water in the unsaturated zone... 13
Purpose and scope... 1 Effect of phreatophyte removal on water quality of the alluvial
Location and physical setting of the report area .................. 2 deposits ............................................................................ . 15
Well-numbering system ......................................................... 2 Factors responsible for changes in water quality ................. 15
Geology and ground water... ........................... 2 Statistical analysis of effects of phreatophyte removal on
Water quality ................................................. 3 specific conductance of water in the alluvial deposits,
Gila River .................................................. 3 by H. W. Hjalmarson ........................................................ 20
Tributaries to the Gila River .................. 5 Summary ........................... 21
Basin-fill deposits ....................................................... 8 References cited ............................................................................. 23
ILLUSTRATIONS
Page
PLATE 1. Map showing location of the Gila River Phreatophyte Project area, Graham County, Ariz., cross sections, data-collection
sites, and contacts between the basin fill, the terrace alluvium, and the flood-plain alluvium .......................................... In pocket
2. Maps showing distribution of dissolved solids in the upper part of the saturated zone of the alluvial deposits, Gila River
Phreatophyte Project area, Graham County, Ariz ................................................................................................................. In pocket
FIGURE 1. Cross section showing generalized geology of the project area and ground-water movement at right angle to the Gila River.... M3
2. Graph showing mean daily discharge of the Gila River at cross sections 1 and 9 versus specific conductance of river water ...... 4
3. Diagrams showing chemical composition of water versus discharge of the Gila River at cross section 1 and chemical com-
position of water from the saturated zone of the alluvial deposits ............................................................................................ 5
4. Graphs showing specific conductance and discharge of the Gila River at cross section 9, July 25, 1966, to September 30,
1968 .................................................................................................................................................................................... 8
5. Graphs and hydrograph showmg discharge of the Gila River and specific conductance of ground water from the
saturated zone of the alluvial deposits ........................................................................................................................................ l4
6. Hydrographs and graphs showing discharge of the Gila River, ground-water altitudes, and specific conductance of ground
water from the unsaturated zone and saturated zone of the alluvial depsoits .......................................................................... 16
7. Diagrams showing chemical composition of water from the unsaturated zone of the alluvial deposits ............................ 19
8. Graph showing interaction of specific conductance of ground water between reaches 1 and 2a and periods 1 and 2 ......... 22
TABLES
Page
TABLE 1. Selected chemical analyses of water from the Gila River at cross section 1 ..................................................................................... M6
2. Chemical analyses of water from tributaries to the Gila River within the project area .. 7
3. Chemical analyses of water from the saturated zone of the basin fill ......................... 11
4. Chemical analyses of water from the saturated zone of the alluvial deposits ...... l2
5. Chemical analyses of water from the unsaturated zone of the alluvial deposits ..... 18
6. Summary of data for factorial experiment ........................................................................................................................ 22
7. Summary of two-way analysis of variance for ground-water specific-conductance measurements .................................................. 22

III

IV CONTENTS

FACTORS FOR CONVERTING ENGLISH UNITS TO
INTERNATIONAL SYSTEM UNITS (SI)

The following factors may be used to convert the English units to the International System units
($1). This report contains both the English and SI units equivalents.

 
 
  
 

Multiply English unit
Feet (ft) ..........................................................
Miles (mi) .................................
Acres .........................................
Acre-feet (acre-ft) ...................
Feet squared per day (ftY/d).
Inches (in.) .................... ‘ ...........

Cubic feet per second (its/s) .........................

By
0.3048
1.609

.4047
.001233
.0929
25.4
.02832

To obtain SI 'um't
Metres (m).
Kilometres (km).
Hectares (ha).
Cubic hectometres (hm’).
Metres squared per day (mZ/d).
Millimetres (mm).
Cubic metres per second (m3/s).

 

GILA RIVER PHREATOPHYTE PROJECT

 

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY
IN THE GILA RIVER PHREATOPHYTE PROJECT AREA,
GRAHAM COUNTY, ARIZONA

 

By R. L. LANEY

ABSTRACT

The Gila River Phreatophyte Project is in southeastern Arizona along
a 15-mile (24-kilometre) reach of the Gila River above San Carlos
Reservoir on the San Carlos Indian Reservation. Two water-bearing
units underlie the area—the basin fill and the alluvial deposits. The
basin fill is divided into two facies—a silt and sand facies and a limestone
facies—and consists of deposits of clay, silt, sand, tuff, and limestone as
much as 1,000 feet (300 metres) thick. The basin fill yields only small
quantities of water to wells. The alluvial deposits, which overlie the
basin fill along the Gila River and support most of the growth of
phreatophytes, consist of as much as 85 feet (26 metres) of silt, sand, and
gravel and yield moderate to large quantities of water to wells.

The water quality is considered in terms of four major hydrologic
sources: the Gila River, the tributaries to the Gila River, the basin fill,
and the alluvial deposits. The dissolved-solids concentrations in the Gila
River ranged from about 300 mg/l (milligrams per litre) during high
flows to about 4,800 mg/l during low flows. Most of the year sodium and
chloride were the principal ions.

Water in tributaries to the Gila River contained less than 300 mg/l of
dissolved solids and had a calcium bicarbonate to sodium bicarbonate
composition. The amountof water from this Source is small and has little
effect on the quality of water of the Gila River or the alluvial deposits.

The dissolved-solids concentrations of the basin-fill water ranged from
about 200 to 5,000 mg/l; the lowest concentrations were in the silt and
sand facies, and the highest concentrations were in the limestone facies.
Water containing less than 500 mg/l of dissolved solids had a calcium
bicarbonate to sodium bicarbonate composition; water containing more
than 500 mg/l of dissolved solids had a sodium chloride composition.

Water from the saturated zone of the alluvial deposits was a sodium
chloride type and had dissolved-solids concentrations that generally
ranged from about 3,000 to 7,000 mg/ 1. Lowest concentrations of dis-
solved solids,were in water near the river and concentrations increased
with increasing distance from the river.

Dissolved-solids concentrations in water in the saturated zone of the
alluvial deposits were monitored for 8 years, during which time the
phreatophytes were removed from the area. Evapotranspiration and
fluctuations of streamflow in the Gila River caused variations in con-
centrations of dissolved solids. Concentrations were high during pro-
longed periods of low streamflow and were low during and following
periods of high streamflow. Average dissolved-solids concentrations for
the 8-year study period ranged from about 2,600 to 5,800 mg/l. Concen-
trations of dissolved solids in water from individual wells for the same
period varied about 1,000 to 13,000 mg/ 1. Variations in dissolved-solids
concentrations in water in the saturated zone decreased with distance
from the river.

Water from the unsaturated zone had dissolved-solids concentrations
that ranged from about 600 to 14,000 mg/l. Concentrations were greater
in the upper part of the zone than near the water table during periods of
low streamflow; chemical composition of the water ranged from a

 

sodium chloride type near the water table to a calcium chloride type in
the upper part of the zone. Following periods of flushing by floodwater,
the dissolved-solids concentrations were reduced, and the water had a
calcium sodium bicarbonate composition in the upper part of the zone
and a sodium-chloride composition near the water table.
Evapotranspiration and the variations of streamflow of the Gila River
caused fluctuations in dissolved-solids concentrations in the water in the
alluvial deposits and masked the detection of changes in concentrations
caused by phreatophyte removal. The infiltration of floodwater from the
Gila River removed and diluted the salt buildup in the alluvial deposits
and, thus, removed any effects that may have been caused by phreato-
phyte removal. Statistical analyses of the specific-conductance data
indicate that the removal of phreatophytes did not significantly affect the
specific conductance of water in the alluvial deposits in the study area.

INTRODUCTION

The Gila River Phreatophyte Project was established to
determine the amount of water lost through consumptive
use by phreatophytes—mostly saltcedar and mes-
quite—and the amount of water that might be sal-
vaged by their removal. The objectives of the investiga-
tion are to describe the hydrologic and ecologic variables
in the project area and to test and develop methods for
evaluating these variables in a large area (Culler and
others, 1970, p. A2). This report describes the quality of the
water in the project area and is one of a series of US
Geological Survey Professional Papers issued in connec-
tion with the project.

PURPOSE AND SCOPE

The purpose of this study was to determine the chemical
quality of water in the project area and to define changes,
if any, in water quality caused by phreatophyte removal.
Water-quality data were collected during the period June
1964 to June 1972, and more than 200 chemical analyses of
water were used to determine the water quality in the
alluvial deposits, the basin fill, the Gila River, and the
tributaries to the Gila River. Specific conductance was
monitored on water from the Gila River for a 3-year period
and on water from more than 50 observation wells in the
alluvial deposits for the period of the study. Chemical
analyses and measurements of specific conductance were
made periodically on samples of water from the
unsaturated zone of the alluvial deposits at two sites near

M1

 

 

M2

the Gila River for part of the period. Statistical methods
were used to analyze the effect of the phreatophyte removal
on the quality of the water in the alluvial deposits.

LOCATION AND PHYSICAL SETTING
OF THE REPORT AREA

The Gila River Phreatophyte Project is in the Basin and
Range physiographic province (Fenneman, 1931) in
southeastern Arizona and comprises a l5-mile (24-km)
reach of the Gila River in the northwestern part of the
Safford Valley immediately above San Carlos Reservoir in
the San Carlos Indian Reservation (pl. 1). The Gila River
in the area occupies a broad terraced valley between two
mountain ranges where it has developed a flood plain that
ranges from about half a mile to 1 mile (0.8 to 1.6 km) wide.
Altitudes range from about 2,500 feet (760 m) above mean
sea level along the Gila River to as much as 7,000 feet
(2,100 m) above mean sea level in the nearby mountains. At
the time the study was begun, the flood plain was covered
by a dense growth of saltcedar, which derived water from
ground water at shallow depth.

The river and flood plain were divided originally into
three contiguous reaches. However, in December 1965,
backwater from San Carlos Reservoir inundated reach 3
(the lower reach) and the lower half of reach 2. Con-
sequently, data collection on reach 3 was terminated, and
reach 2 was divided into reaches 2a and 2b. Most of reach 1
was cleared of phreatophytes between January and May
1967. Reach 2a was cleared mainly between January and
May 1969, and reach 2b was cleared between September
1970 and March 1971. After each reach was cleared,
virtually no vegetation remained except for some seasonal
grasses and weeds from about March through May of each
year. '

The climate in the project area is semiarid. The mean
annual precipitation at San Carlos Reservoir, at the lower
end of the study area, averaged 14 inches (356 mm) per year
for the period of record, 1882—1971, but has ranged from
4.0 inches ( 102 mm) in 1924 to 25.8 inches (655 mm) in 1941
(Hanson, 1972b, p. 5). Most of the precipitation falls in
two distinct periods—July to September and December to
March. The summer precipitation, which results from
convection of warm moist air from the Gulf of Mexico,
generally is localized, intense, and of short duration. The
winter precipitation, which results from cyclonic storms
from the Pacific Ocean, usually is less intense and more
widespread than that in summer. Mean monthly tempera-
tures range from about 8°C (Celsius) in January to about
31°C in July (US. Weather Bureau, 1964).

WELL-NUMBERING SYSTEM

When the project was begun in 1962, 72 wells were
drilled in the alluvial deposits on a series of cross sections
at right angles to the Gila River at approximately 11/1-mile
(2-km) intervals. Three wells were drilled on each side of

 

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

the river on each cross section, and the cross sections were
identified by numbers 1, 3, 5 . . . 23 (pl. 1). In 1966 five wells
were drilled on cross section 12, which was established
between cross sections 11 and 13. Also, several additional
observation wells were drilled between the cross sections
throughout the project area.

Each well is identified by a four—digit number. The first
two digits represent the cross section that the well is on or
is immediately downstream from; the last two digits
represent an arbitrary number unique to the cross-
sections. For example, all wells on cross section 9 and
between cross sections 9 and 11 have the prefix numbers
09.

GEOLOGY AND GROUND WATER

Two major sedimentary rock units are exposed in the
project area—the basin fill and the alluvial deposits
(Weist, 1971, p. D3). The basin fill is the most widespread
sedimentary unit in the area and is exposed on the steep
sides of dissected terraces on the valley slopes. The unit is
composed mainly of material ranging from sand to clay
and limestone. A silt and sand facies and a limestone facies
were mapped in the project area (pl. 1); the silt and sand
facies consists of even-bedded well-sorted light-brown to
reddish-brown silt and light-brown very fine grained sand.
The limestone facies consists of interbedded white lime-
stone, marl, siltstone, fine-grained sandstone, tuff, and
local beds of green clay. The limestone facies is exposed
only in the northwestern part of the project area. The
thickness of the basin fill may be more than 1,000 feet (300
m).

The alluvial deposits overlie the basin fill along the
river and were divided into the terrace alluvium and the
flood-plain alluvium mainly on the basis of topographic
position (Weist, 1971, p. D7). The terrace alluvium has
been incised and filled with a channel deposit of flood-
plain alluvium (fig. 1). The flood—plain alluvium under—
lies the present-day flood plain of the Gila River and
supports most of the phreatophyte growth along the river.
The terrace alluvium consists of poorly sorted cobbles,
gravel, sand, and silt and generally is less than 40 feet (12
m) thick where it underlies the flood-plain alluvium.
Locally, the unit may be as much as 75 feet (23 m) thick.
The flood-plain alluvium consists of lenticular beds of
silt, sand, and gravel and ranges in thickness from 0 to 50
feet (0 to 15 m). The combined thickness of the alluvial de-
posits may be as much as 85 feet (26 m). In most places the
flood-plain alluvium and terrace alluvium form a
hydraulically continuous aquifer.

Water in the basin fill is under artesian pressure in the
project area and moves toward the Gila River and down
the valley. Near the river part of the water moves into the
alluvial deposits. Water levels in wells that penetrate the
basin fill beneath the alluvial deposits generally are nearer
the surface than are the water levels in nearby wells that
penetrate only the alluvial deposits. Wells tapping the

 

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

Terrace alluvium

Flood-plain alluvium

‘
‘
“‘

Basin fill

 

M3

Terrace alluvium

k
s
a:
fi
L:

Basin fill

 

FIGURE l.—Cross section showing generalized geology of the project area and ground-water movement at right angle to the Gila River.
(Geology after E. S. Davidson, written commun., 1964; Weist, 1971.)

basin fill in the project area yield only a few gallons of
water per minute, and the average transmissivity is 15 ftz/ d
(1.4 m2/d; Hanson, 1972a, p. F27).

Water is present in the alluvial deposits under water-
table conditions, and movement of the water is down the
valley. The alluvial deposits are recharged mainly by large
quantities of water from the Gila River during periods of
flooding; smaller amounts of water are recharged to the
alluvial deposits from tributary underflow and water from
the baSin fill. The depth to water in the alluvial deposits is
about 5 to 40 feet (1.5 to 12 m) below land surface, but
during periods of above average streamflow and flooding
the water level may be nearly at the ground surface along
the river. Wells in the alluvial deposits yield moderate to
large quantities of water, and the average transmissivity is
28,000 ft2/d (2,600 mZ/d; Hanson, 197221, p. F27).

WATER QUALITY

For purposes of this report the water quality is con-
sidered in terms of four major hydrologic sources—the
Gila River, the tributaries to the Gila River, the basin-fill
deposits, and the alluvial deposits. Water from each source
has identifiable chemical characteristics.

An evaluation of natural variation in water quality is
necessary to place in proper perspective any apparent
changes in water quality due to removal of vegetation.
Variations in water quality in the alluvial deposits were
most critically examined because the phreatophytes
depend on this water for their supply. The variation of
water quality of the Gila River was studied because the
river is the principal source of recharge to the alluvial
deposits. The quality of the water in the tributaries to the
Gila River and in the basin fill is discussed, but the water
from these sources has little effect on the quality of the
water in the Gila River and the alluvial deposits.

GILA RIVER

The dissolved-solids concentrations of water from the
Gila River ranged from about 300 to 4,800 mg/l based on
chemical data from more than 50 water samples that were
collected at various stages of the river and from individual
and continuous specific-conductance measurements that

 

were made from 1965 to 1968. Specific conductance is a
measure of the ability of a substance to conduct an electric
current. The presence of charged ionic species in water
increases its conductance. Thus, specific conductance
increases as dissolved-solids concentrations increase, and
specific conductance can be used to determine the approx—
imate dissolved-solids concentrations. The ratio of
dissolved solids, in milligrams per litre, to specific con-
ductance, in micromhos, for water from the Gila River is
about 0.60 (table 1). The specific-conductance measure-
ments were multiplied by this ratio to obtain an approx-
imate value of dissolved-solids concentrations when only
specific-conductance measurements were made and no
chemical analyses were available.

Specific conductance of water from the Gila River is
large during periods of low flow and is small during
periods of high flow (fig. 2). At low flow, a considerable
amount of poor quality water is supplied to the river by
ground water from the alluvial deposits and from returned
irrigation water upstream from the project area. High
flow results from snowmelt in winter and springland
localized thunderstorms in summer and is of good
chemical quality; however, water during periods of high
flow may contain large amounts of suspended sediment.

During most of the year, sodium and chloride are the
principal ions in the river water; sulfate and calcium are
the next most abundant ions. The approximate range in
concentration of the principal constituents in water from
the Gila River is shown in the following list:

 
 

  

Constituent Milligrams per litre
Chloride .......................... 65 to 2,100
Sodium plus potassium 70 to 1,300
Sulfate ............................................................. 50 to 800
Calcium .......................................... 50 to 350
Bicarbonate... ....150 to 350
Magnesium ..................................................... 10 to 120
Fluoride ......................................................... 0.8 to 2.0
Dissolved solids .............................................. 300 to 4,800

The chemical composition of the water from the Gila
River is shown in a geochemical graph that was described
by Piper (1945; fig. 3, this report). Amounts of the
principal cations and anions expressed in percentage of

M4 GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

 

 

 

 

 

 

 

50"" I I I I I I I I 14°
_ o _
— O _
. EXPLANATION
— ~ 0 0 Individual water samples at -
0 cross section 1
0 0 Mean daily specific conductance
. o . at cross section 9
1000 _ . _ 28
_ O . _
_ . _
_ 0 __
6’
_ . _
O o
0
g " ° 0 o ‘ g
o O O 8
LU __ _. U)
(I) O O I‘
I: 0 Lu
5: ° ° *
U)
m 0 o- ° I:
:1 10° _ . . _ 2.80 E
9 : O O 0 . O : E
m . U
D — . o _ E)
o __ o O O
o o o — 3
Z 0
_ _ o 0 O o
. o _ 2
Lu 0 '
g _ o ' _ m‘
o 0
<( o 0 ° 0 o I:
I O <
o — o _ I
(2 ° 0
0 t2
0
O
10 A
_ u u o _ 0.28
_ o _
1 I I I I I I C? I 0,02
0 1000 3000 5000 7000 9000

SPECIFIC CONDUCTANCE, IN MICROMHOS
PER CENTIMETRE AT 250 CELSIUS

FIGURE 2.—Mean daily discharge of the Gila River at cross sections 1 and 9 versus specific conductance of river water, 1965-68.

milliequivalents per litre are plotted in the lower
triangles. The points in the triangles are extended into the
diamond field, and the resulting intersection is repre-
sentative of the composition of the water with respect to
both cations and anions.

The composition of the water in the river changes from
a sodium chloride type at low flow to a calcium sodium
bicarbonate type at high flow (fig. 3; table 1). Sodium and
chloride make up as much as 70 percent of the total ions
during periods of low flow and less than 40 percent at high

flow. Water in the river at low flow has a similar chemical
composition as water in the alluvial deposits.

A continuous record of specific conductance is avail-
able at the gaging station on cross section 9 from July 25,
1966; to September 30, 1968 (fig. 4). During the 1968 water
year (the period October 1, 1967, through September 30,
1968) the mean discharge of the Gila River was about 300
percent above the mean annual discharge for the period of
streamflow record, which was started in 1929. During the
1967 water year (the period October 1, 1966, to September

 

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY M5

  
  
 
 
  
 
   
   
   
   
  
  
 
 

 

 

 

 

EXPLANATION

DAILY MEAN DISCHARGE, IN
CUBIC FEET PER SECOND
(CUBlC METRES PER SECOND)

()1 0—10
(0—0.28)

03 10—100
(0.28—2.8l

(.7 100—1 .000
(2.8—28)

0‘ 2 1.000-1 0,000
(28—280)

.15 >1o,ooo
(>230)

Number refers to chemical analysis
in table 1

Chemical composition of water
9 from the saturated zone of the
alluvial deposits (based on 13
chemical analyses, table 4)

Chemical composition
expressed as percentage
of milliaquivalents per
litre

 

 

 

 

 

CATlONS

ANlONS

FIGURE 3.—Chemical composition of water versus discharge of the Gila River at cross section 1 and chemical composition of water
from the saturated zone of the alluvial deposits.

30, 1967) the mean annual discharge of the river was only
82 percent of the mean for the period of record. The effects
of these above and below average discharges are shown
markedly in the record of specific conductance (fig. 4).
Specific conductance usually is greatest during June and
the early part of July, which is the time of lowest flow in
the river, and values approach 8,000 micromhos (about
5,000 mg/l of dissolved solids). However, specific con-
ductance was more than 7,000 micromhos for much of the
period February through June 1967. Specific conductance
generally is lowest from December through March.
During this period of the 1968 water year the specific con-
ductance was less than 1,000 micromhos (about 600 mg/l
of dissolved solids). The specific conductance varies
widely during the period from the middle of July to the
middle of September and reflects the rapid fluctuations in
river discharge during the summer rainy season.

The large concentrations of dissolved material in the
Gila River are derived from the sedimentary deposits that

 

fill Safford Valley upstream from the project area (Gate-
wood and others, 1950, p. 76; Hem, 1950, p. 20). Soluble
material enters the alluvial deposits and thence to the Gila
River by ground-water seepage from the basin fill and
older sedimentary rocks; dissolved material is contributed
to‘ the Gila River by direct dissolution of surficial deposits
by runoff.

TRIBUTARIES TO THE GILA RIVER

Local high-intensity rainfall causes flow in the major
tributaries to the Gila River in the project area. This flow
may add appreciable amounts of water to the Gila River
during wet years and virtually none during dry years
(Burkham, 1970, p. B16). The water is of excellent
chemical quality; dissolved-solids concentrations of 6
samples of tributary flow ranged from 60 to 282 mg/l, and
the water types were mainly calcium bicarbonate,
although the water collected from Salt Creek was a
sodium chloride type (table 2). Seven additional samples

 

 

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

M6

.iofiSnn- E nus—mahoguc =9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.. vod em; 2: o wfim New .36 84m
.09 *5 mm. m5 SN mam wd ow vm a v3 mo 2 we : cow”: mwlvwlwfl 2
$58 .. mod wm.m 2: o mo.m ooh aw; aim
BZOQE E vwciufi 5c 85m in mm. mow mmw Ev .. o; cg «a o om: vw m: on .. shod mmlwwlwg 2
.m:0~€< «o
EGB>ED .m:0m wzm igwmaocU :23 mod 24 omd o and wm.w 3: vmd
-~:u€w< ‘0 Eufitmaon 3 mdmZmEV NS oo. 02. wa wwv m4 ow vw o 2% mm 2 me w« ome mmlv pm 3
mod 3mm mm; o EN WEN mad 8N
.09 ms hm. 2w EH mmm ad mm mm 0 mm: mo : vm 02v; molilwfi .2
.358 cod wwN RA 0 EN 2.x. :4 mmfi
838me 5 39:2: “on 85m cs 5. wmo mm: 5m H; mm mm o m9 mm. X mm o3; mwlguwfi NH
.0 mod uv90 wfim mm: o flaw mfm mad $.N
Q $5 on. 2% m3 amm N wd $2 on o mum wwfi NH mm mm mm. mwwwwln 2
.naofifiw Ho
EMBQVED .mzom wan Eammfiono :23 mod «Nod mad me; o $6 om;N mod mmgm.
48a? “0 58:88 3 firm: 2 mm. 85 m2 wwv _ wd X: g c 3.“ 8 w 8 mm :m 81mm-” 2
mod 36 mvd mmN o Nmfi «56 mm; 3.x:
as mm. cam; mum mm» vi. 9w m4 www m: o 2N cm: 9 wk. 3 gm mwlmmlfi a
.0D K 8.0 $.w EN o woaN 3d aw; $6
m: mm. owe M omm 3w .. w; Em 92 o 3N wow mm ma .. omw moleN m.
.OQ . 3d mm.w~ 9% o wo.v mmN~ Exw mod
ms mm. OSN vmv . can; : A: mvv cow o 3N am“ an NE . mm: melme N.
.mE_om .. mod 8.? mm.m o 36 3.2 EN mad
328mg 5 coviufi Ho: 85m ms mm. owofi wmm own; .. h; 2.9 mww o New mmw cm 3; : HS molmzlv w
36 aod 3.x“ $.w o wmd 3.5 $6 3.3
Nd ow. 9&4“ o: owed cmm.w m6 >4 So; cww o mg mg mm :« vv om vamTS m
.maofijw mo A
hmfluzni .wmom «Em NCEEMEU 19:: mod mod mwdm 8.3 o «Ev 2.3 :d 3.:
.Azuima‘ we EoEfimQuQ 3 murfia us me. oom‘m 8m oom.m m m: cow; owm o wwm ohm mu 3N mm mm $4.73 v
:6 mod Emma“ 9.3 o 3mm :Nv mms 3.:
ms 5. 81¢ 05; 3&5 35m fin h; own; owe o omm mom ww vww Nv a «Avril: m
.mwzom 86 Nm.mm 5.2 o mw.v Emmm mmd 2.5
3303:“ E @2365 8: 85m vs S. cam.” com; $.meN . w; o:.w com o mmw owmq R: vwm . 2 moIEIo N
.m:o~€< «o
>fifiu>ED .mzom 98 32525 :35 Nod 23 Exam CNS o wwd mmdv and $.9—
._8:w< we Engages E Ezflé 5 $6 83 om? om§ _ 3 8...“ cm“ o E $2 w: gm $ 2 $-m -w #
85 Comm a N a 0 s m a. .m s w m S P a m m.
.6358 :55; )m‘ )m )M 9%., )1 )m )mW )mea )fl )mmm.m. “)3 )w
“56on mos:— nOUNU Goof E W m I'm my. (ny W ‘0 m M m. m W. W mm W w mu m , m e W: WM W m 0.9 aw... um.
min—Ex *3 2 $28 .9515 mm 32. 3.38% Ezm mx 9 ( w W m X mm. m. ( ( em m m. m. [We N
U028 3:5. .Emm llllll a I\ m w m m ‘w. w. .0
.% .UsvcoU mum—om 33339 W. ”A m.
we Guam quunm m.
FEM—Rim Saw 0: 38:2: AEV $381— 4.28qu mm 385 >355 EQESBO .m.D 3 Sara—E

5:: En WHERE—acufizfi E 3:9?“ mm 32 how 2:»: 532

N 2233.22» “a SEM 3:0 2: EEK $33 «a 8533:: :6

US: “on ”SEE—:8 E @3593 mm 3:25:38 18:55 _

R .8 95m: ham: 6!

NEE? vmtfiwmlg might

865 mm 3086 Pa: 5; flaw—«zdcuzafi “Em uh:

hi WESMEEE E menisci

M7

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

.EEEOJE E FEE—ESQ"

,m.:o::~n:ou cur—«5.8% $50 + van—Sigma x 36 mczuum 3 359:8 EN AEnmv 8:3 3303mm:

 

 

 

 

 

 

.83 :3: 33305 fl SSE . . . . .
.82 33 538m 8“ was, 5 av 2 o e: o a: :6152 9; m3 .2. Sign .md
.35 5::va :: 25:3 29 ...... . .
:nfiffimfimflbB Ecflgflsmvu NS nww Nmfi . w: NdV 9v mm. o 3 wow «mm mm ow «.5 M0530 "yogi—0:00 .o
.23: 53$:me min 8:616 aflocwwgvflmwfiumwflw
. 36V 36 mwd o ES «56 de 2; NA db >w3nmmm .w..D
m 5 mm: a .. 8 Nov 3 a c Q 2 no «N so 958: ESE 2 .m
.imvasoa
3.9V Saw wad o 3; 59m $6 ow; QNE “0 :68: hE. its
2. nwm No wwm de mm cw 0 mu 2. mg. vw .. ..... E .85 £me .m.D “a «2:0 :am .v
.Eiflmfi
335%"; 89:2: Doom: E .85 :25
933mm: ”oucuhfimv 3 3393‘ Ho “985:8 can 26
.w@ 881: ”mmmfimaa Bean—8 .. mvd and o wad . mwd cod 34 fixed Ema 035000
How 333 «0 EsoEm :59:sz as mi mm «mm NE .. 2 vw o co : 9 o Nu : ...... :w 43m :0 ummw mum; $80 .m
09:83wa 3 vocab—8%
8&1; ”mmmrwcm 829:8
Bu $83 “0 550:8 Euwua 3d vwd o mod . Ed mod mod Nos min 55 £03
.32: $05: 383:: newcom NF :2 mm co co md mm o av .. 9v fig .2 m6 .. ...... : .ow... 5253 H 5:68.320 .N
._\wE _ 22:: ._\ME 9 .
:ohom .n:o~€< *0 EMEEVED .52 :93 «o :8 0:8
hmzom vcm KwhmeoAU .mhazzu 3.0 :d $6 o $.m owd ewd mad 8N m6 428% ENG vmuzooo
.tw< «0 EwEfionD >3 mirmcax us cow o9 . www 56 v mm 0 3w w e NH 3 mm ...... om .w:< 2: :0 mwww umfimémuau ‘< A
93; E23 $05 was: a m. m. ) m. ) m ) m. ) M w ) m ) m ) m 2 .3
m RE a mos—.553: mm 2630“ .Ezm E m m M m N... w W. nHu m. m w. W m... am am mu m. S m. $9 5 535:: min
x um «I 8539358 Quits: ( m“ 0 W fix M ax w 0 m m. ( m ( m‘ ( m um :oivoze .5553
958% 3:8 83829 m 0 m m m _o 25 ,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. $6.565 mm 3083 >u>5m .53230 .m.D E mug—mam
vb: 5Q 2:u_~>_:U~::E E 3:82 fl 32 SH «Em: $32 nub: an 3355:: E 3:32 mm 2:03:38 REES? =~ 5 23m: Sam: ”Baum

33: Ufiokm 2: 25:3 SEE 3:0 2: 2 8.233%: :3: E33 *0 3.23323 Nafismaolfi 55‘?

v5 mm 585 vb: En 25133351: van 8.: 5a mEfimEE. E warn—HE

M8

1°,Dooll111111

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

 

9000

6000

4000

SPECIFIC CONDUCTANCE
(MICHOMHOS PER CENTIMETRE
AT 25“ CELSIUS)

2000

 

1111111
51525

.1111111.l

0
25

JULY AUG SEPT OCT NOV

1966

1.111111 1

1
15255152551525

DEC

.1111I

5 15 25
APR

11111

.1
5 15 255 15 25

5 15 25

FEB MAR MAV

1967

 

11,000

11111111111111

1 0,000

11111111111111111111r11

 

I llllll

1 000

1111111

1

100

 

MEAN DAILY DISCHARGE, 1N CUBIC FEET FER SECOND

 

111 I1.

..1 1t.|111.l I
51525 5152551525

AUG SEPT OCT NOV

1966

of tributary flow for which only specific conductance was
determined had specific—conductance values ranging from
205 to 570 micromhos (about 120 to 340 mg/l of dissolved
solids). The total annual tributary flow in the projectarea
is generally less than 10 percent of the flow of the Gila
River at Calva (D. E. Burkham, oral commun., 1972).
Some ground water may originate by infiltration of
tributary flow, but the amount is small and probably
occurs only during wet periods (Weist, 1971, p. D13).
Therefore, the probable small amount of recharge from
tributary flow has an insignificant effect on the water
quality of the Gila River or on the ground water in the
alluvial deposits.

BASIN-FILL DEPOSITS

The dissolved-solids concentrations in water from 16
wells tapping the basin-fill deposits ranged from about

 

11.1 .....
515 2551525

DEC

 

11.

5152551525

ll..1| l 11111 |.1..ll.11.
515255152551525 5152551525

JAN FEB MAR APR MAY JUNE JULV

FIGURE 4.—Specific conductance and discharge of the Gila River at

200 to 5,000 mg/l (table 3; pl. 1). Water from wells that
were drilled in or near the limestone facies contained more
than 3,000 mg/l of dissolved solids (wells 1962, 1963, and
2374, table 3). These large concentrations resulted from the
dissolution of soluble salts that were_deposited with the
limestone. Water in the silt and sand facies was of better
quality and usually contained less than 1,000 mg/l of
dissolved solids, except where the percentage of clay was
large. For example, the basin-fill deposits penetrated by
well 1141 contained about 85 percent silt and clay and 15
percent sand; dissolved-solids concentration of the ground
water was 2,570 mg/ 1. The deposits penetrated by well
1756 contained about 60 percent silt and clay and 40 per-
cent sand; dissolved-solids concentration of the ground
water was only 676 mg/l. The clayey parts of the silt and
sand facies contained more soluble salts than the coarser
grained parts of the facies. The presence of soluble salts

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

[lllllllllll

I1111|IIVIIIIVIIYIIIIYT

  

Illllllllll‘IVVIVIIIIVI}I1I11

M9

vlvtvllvlvvlvllll

 

 

.
o"
f
r". a
I Nor???"
we -------
I.....1....1I.....1.....1.r...I..... .“.....J. .I. .|11..11....1,..1.HI
5.1525251525515255152551525515255152551525515255152551525515255152551525
AUG SEPT OCT NOV DEC JAN FEB MAR APR MAV JUNE JULY AUG SEPT
1961 1968

312

 

llllllllllllIIIII‘llll‘lllIl'lllllllllll

llllllvllTllvvlvvll

‘T..y. 1..

230

 

lllllll

l

 

A MVUUW L 2.

.11

 

 

WMWN

1.111

515 25 5 15 25 5 15 25 5 15 25 515 255

NOV

15 255 15 25 5

AUG SEPT OCT DEC JAN FEB

1967

cross section 9, July 25, 1966, to September 30, 1968.

was indicated by the accumulation of salt crusts on many
clayey exposures of the basin-fill deposits. Dissolved-
solids concentrations generally were greater in the silt and
sand facies near the Gila River where the facies was over-
lain by saturated alluvial deposits (wells 0105, 1140, 1141,
and 1756, pl. 1). These higher concentrations may indicate
contamination by poorer quality water from the alluvial
deposits because nearby wells that did not penetrate the
saturated alluvial deposits contained water having notice-
ably lower dissolved-solids concentrations.

The water in the basin fill in the Safford Valley
upstream from the project area generally is highly
mineralized because of dissolution of soluble material
from lake beds and playa-type deposits (Gatewood and
others, 1950, p. 71; Hem, 1950, p. 20). As a result of the
artesian conditions and probable faulting in the basin fill
in this area, mineralized water moves into alluvial deposits

I|ILIIlIJlIJ|I|IIllJII|Illlll|lllllllllllllllllll

MAR

MEAN DAILY DISCHARGE, IN CUBIC METHES FER SECOND

 

 

.1..1| ..... l...1.l.. 1
15255152551525515255

15 25

15 25 515 255

APR MAY JUNE JULV AUG SEPT

1968

and is a significant source of poor quality water upstream
from and in the project area.

The water in the limestone facies is a sodium chloride
type; the water in the silt and sand facies generally is a
calcium bicarbonate type near the surrounding moun-
tains (recharge areas) and a sodium bicarbonate to
sodium chloride type near the Gila River (pl. 1). The
sodium chloride type water was obtained from wells
where the facies is overlain by saturated alluvial deposits.
Temperatures of water from the basin fill ranged from
180° to 265°C (table 3).

Water from most of the silt and sand facies would be
suitable for domestic and stock use, but water in which

sodium is the major cation may not be desirable for
irrigation. Water from the limestone facies would not be
suitable for domestic and irrigation uses nor desirable for
livestock.

 

M10

ALLUVIAL DEPOSITS

The alluvial deposits consist of flood-plain alluvium
and terrace alluvium. In discussions of water quality, no
distinction is made between the two units because both are
very permeable and form a continuous aquifer. However,
the separation of the two units is made because most of the
phreatophytes are confined to the flood-plain alluvium.

The ground-water quality in the alluvial deposits is
described in terms of the saturated zone and the
unsaturated zone. Chemical data on water from the
saturated zone were obtained from water from wells that
penetrated the water table, and the chemical data on the
water from the unsaturated zone were obtained from water
that was extracted by soil—water tubes in the alluvial
deposits above the water table.

GROUND WATER IN THE SATURATED ZONE

Dissolved-solids concentrations in water from more
than 50 shallow observation wells in the alluvial deposits
ranged from 395 to 19,200 mg/l, but most of the concen-
trations ranged between 3,000 and 7,000 mg/l. Dissolved-
solids concentrations are lowest in ground water near the
Gila River and generally increase with increasing distance
from the river. The lower concentrations are caused by
infiltrating and flushing of fresh water from the river
during periods of high flow.

The water from the alluvial deposits has a sodium
chloride composition similar to that of water from the
Gila River at low flow, but the amounts of individual
constituents in the water from the alluvial deposits are
more variable than those in river water at low flow (table 4;
fig. 3). However, the average sodium to chloride ratio of
the water is slightly less than one and is not significantly
different from that of river water.

Most of the water in the alluvial deposits in the project
area is not suitable for domestic or public supplies because
of the excessive amounts of dissolved solids. In addition,
some water contains more fluoride than the recommended
maximum concentration of 1.4 mg/l established by the
US. Public Health Service (1962). Most of the water in the
alluvial deposits could be used by livestock because of their
higher tolerance to dissolved solids in water. The large
concentrations of dissolved solids, the high percentage of
sodium, and boron concentrations greater than a few
tenths of a milligram per litre make much of the water in
the alluvial deposits undesirable for irrigation purposes.
The use of water for irrigation in the Safford Valley has
been discussed by Hem (1950).

The specific conductance of water from the wells in the
alluvial deposits was monitored from June 1964 to June
1972. Measurements were made at approximately 3-month
intervals from March 1966 to March 1970 and at 6- to 12-
month intervals for the remainder of the period. The
sampling procedure consisted of bailing or pumping
water from the wells until a constant specific-conductance

 

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

value was obtained. Dissolved-solids concentrations, in
milligrams per litre, were calculated by multiplying the
specific-conductance value by 0.64, which is the average
ratio of dissolved solids to specific conductance from more
than 50 chemical analyses of water from the alluvial
deposits.

The shallow observation wells were drilled only deep
enough into the saturated zone to bottom below the
maximum water-table decline. Well depths ranged from
about 12 feet (4 m) near the river to as much as 35 feet (1 l m)
near the outer margins of the alluvial deposits. The wells
were constructed with blank casing that was open to the
alluvial deposits only at the bottom. Along the river where
the alluvial deposits are thickest, the wells penetrate only
the upper few feet of the saturated deposits, and, because of
the manner of well construction, water-quality data from
samples taken from the wells should be considered as
point sources of data from the very upper part of the
saturated zone of the alluvial deposits. The dissolved-
solids maps prepared from this data (pl. 2) show the varia-
tions of concentrations of dissolved solids primarily in the
upper part of the saturated zone of the alluvial deposits
and not necessarily average concentrations or changes in
concentrations of dissolved solids in the total saturated
thickness of the deposits.

Two periods were selected—June 1967 and December
1968—to demonstrate the general distribution of dis-
solved solids in the upper part of the alluvial deposits and
show how differences in the magnitude of discharge in the
Gila River can cause large variations in dissolved-solids
concentrations in ground water. Phreatophytes had been
cleared from reach 1 prior to the collection of the data for
the two maps. The effects of phreatophyte removal on
dissolved-solids concentrations, if any, would be included
in the map for reach 1. Thus, the data on the maps are
intended to illustrate only distribution and magnitude of
variation of dissolved solids in ground water and do not
show changes in dissolved solids caused by clearing.
During June 1967 most of the ground water contained
3,000 to 7,000 mg/l of dissolved solids (pl. 2A). Water con-
taining the smallest amounts of dissolved solids generally
was in a zone along the central axis of the Gila River flood
plain. The streamflow in the Gila River during the year ‘
prior to this period was below average. Evapotranspira-
tion and the lack of significant freshening by floodwater
caused the dissolved-solids concentration to increase in the
alluvial deposits.

In December 1968 mostof the ground water in the upper
part of the alluvial deposits contained from 1,000 to 5,000
mg/l of dissolved solids, which is considerably lower in
concentration than in June 1967 (pl. 2B). However, the
relative distribution of dissolved solids is similar; that is,
lower concentrations are near the central axis of the flood
plain and higher concentrations are along the margins.
Streamflow in the Gila River during the year prior to

M11

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

.EQEOAaH 5 62.3533. mus—S, In HH<

.Bm:o€mu o: 28:65 ENE :ouchm 5.865 2H. .23 :owmh: 05 .9» .66 5:6 89: .HHEmHHm 60 u=Hm> In a Era—H FE

anHEmm a 63553 Sh .mulodmuoamH EHHmzHULBNB >956 HauHmoHOoO .wHH vi 3 6%: Eu .5: n6 6:: ma. .3 250a- Hu:u

In @3923 EH. 9 EGGS 6.56328 3:37.. .5: 06 35cm 65 emu—HR. .BwEuemu 6:: 39:06:85 60 £53 Hu:v 558....
6555556 56 :0638 5.86:.— 62H>EuEévflw Homes 5505 a mum: mHHom HEN EEEBHU ngHauiw< 6o .55.ng 2r?
gag—€38 605.52% 55: + 25:06:83 x 3.6 6E6?“ .3 662.558 ohm 2.:sz 3:3 «52033:

 

 

 

 

 

 

 

 

 

 

 

66.6 666 3.66 meH 6 HmHN :66 m6.NH $42
.0: m: 83 22 ES. 3.3 S 2 83 men o 83 on: «.2 Na we as mcéiégemmefi Ea
. N66 666 N6.$ m: 6 Rm: 666 mmNm N66 EN
6.6 63.6 3m 6m6.m 6H.H H N.N 6x2 m5 6 6:. 6N ¢6N.H H6 6m 6: .206... ..... ...:..oH.... m66H
N66 666 Exam N66 6 m6.mH 666 «6.66 666 mmN
m6 636 66m 665.6 66.6 H N.N 66v; mwm 6 56 BN NhN.H 56 Nm 6: ¢6IHH|6 6:98.55 N63
. 66.6 HmH 6 N66 N66 3.6 mmH
.0Q 65 6Hm.H 6NH 6B6 6B6 66m «6 6 mm SN NH Hm Ham 66H H.6nvNuNH ...... 3:66... 665
66.6 3N 6 H64. 666 whN Hz: 666
H6 6:. EN NE 6 N 6 mwN N H6 6H 66 HHV $le15 ..... .. ..... 06... $2
.>w>::m Hmu . mm RH EN 66.6 6N6
.HmoHOQO 6.3 >6 mum>Huc< N6 6% 6H 6: 6H on m6 5.6 64. HuH V6.6 I6 ........ 06... «.va
66 6 no.6 65.6 6 66.6 m66 me me 66.N
66 66K. 6m NNHQ mH6 m 6.6 6N. 6 va H NH“ 6N Hum Hm. mHN $6676 ....... .06... mwNH
.mHVHHom 602036 5 H563? .
.:H 5: 8:5 333:5 H8 :6 Na: 66.6 6H.H mNN 66.6 66.6 666 .
.HMOHSO .65 .3 39523 v6 6H6 w mmm H.N 66 wN Hum umH mmH 6.H 6.H 6.NN 661676 NwNH
Sufism H8 . 66.6 m66 $.66 6N6 6H.H mH6 Hm6 26H. 6N.H 65.6
@286 .mb 3 warm: 3: 83. we own.“ 23 3 2 83 Non mm m N: «Na 2 2 we 3: x37: ........ .8... 3.:
.mEHom 6638ch :H HvoHEHu
.E 8: muHHHm .5336 H8 v66 6m.6H mNN 6H6 maN mH6H mNH mmH
.HonOQO .65 E 8335‘ «.6 6Hm.N NmH EN; >6 6mm 6: m NwH 6X. 6H 5N 6.NN 661216 ........ .06... 63H
«66 6 26 $.6H hNH 6N6 mmN 666 66.NH $6 666
m6 666; 3. $6 $6 6 6.6 6% H6 6 6: H me 6 5H 6H 66N «61> In 9.:
66.6 66.6 66.6 $6 S6 wwN «9.6.6 NN.H 6H..H 6H;
m6 66H. HN EN 666 H» 6.6 6H HN 6 62 H 6N 5H 6N Nm 6.HN $61676 ...... .06... N656
N66 v6.6 N66N 66H 6 66.6 Hm6 ww6N 66.6 mmN
m6 66N.m H6H 66N.N NH.6 H 6.6 6664 H6 6 Sum NH 66 66 S“ 66H 6.6N «6676 Hmh6
666 H66 56 66.6 6 RN m66 $6 66.6 cm: .
NR 6mm 5H EN AQ66 m N6 6 «N 6 HmH H 6H HH 6m 56 6.NN «6:676 ............ oH... 6mg
o 65 m: o: 86 SN mod Sm $6 $3
H6 62. N HwN N66 6 N6 6HV mm 6 wNH H N6 6 6 HN 6.NN «66H; :06... Hmmo
66.6 66.6 66.6 36 6 6:” mH6 666 :6 6m.H
5H. 666 H6H 6H6 6 H. 66 «N S 6 va 6 E. 6 6N um H.616NIH. .......... :06... vao
$5955
66 950: N Eda voHnH
.53 56:3 HE=H>3HHN 65
:H HVBEOFXH 60: m5: um:
.98 :63 26 um: 5636“
:m 86 HHHH £93 65
BE £8:qu HmH>=HH~ 26 N66 «N6 66.6 6v.H 666 $26 65.: 6H 6 6m6 6:3
cwzob: HBHHHHH. $3 HH£$ 66 6H6; vN 6S H 6.6 va E. «N 6mm 9% N 6 v6anIN HEN im 33
.mwlmvmumu 8 W)H )W )W mu )m ).0u W“ \W/ m6 )rla. l Cm?»
“mmmwwfiflga mmmmum mm 3.3 mm mm mm. m.” 3. mm fl .5. .._
amoopm . mm m w i am om m (m (9 (m a (m .25 2: ::
3125M «3 w m. m e a G E m m. w :0: E535: .02
"mm m s “383235 a m m .818 .86 =35
O m 32.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

352?: «EH: 0: 6688:1375691 603265masmuux~.m:cNH~<6o EH2~>H:D .mHHom 65.. iaHEuuHD 35:519.. 60 EwEtaauQ .3 333:5

6::th 25Hm>H=Uu=HHE :_6u:oao:Hm:oC£ 35w: :uiothHHthmEEwHHZ—p. 5685:23253638 H8555 HHn :06 2:»: San: ”628:.5 mm 53.8 Pa: in w~:uHa>H:ku=:E 65m oh: 5: 2:36:65 5 mug—Ea:

H3 2.23 2: .8 BEN 68:33: 2: Sex ES: F 832$ SEEMSUIHW ”SEQ.

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

M12

553128 35.53% $50 + 233.585 x mvd MESH .5 v3:

592

3 a 85:55. 836, an :5
E8 Bu 9:sz mum—8 $3.3me

 

 

d 3d 93.: 3.2 d «we do; 3d: 8.2 9d
5.5 ddmfi Id d5.“ mm; d «l deed dmu d EN an Edd m2 2: 2 .. «6|de «5mm
:.d,mdd .ddddm v2“ mNd mud mNN Eddw Nudm and
«d ddmgm dwm; ddwdg 8d 5 dd 23: mwd; N. Nmm mm ddwd own N: w 92 $de SE
mod add 3.3 3.2 d 34. «Ad «Yum mmd ddsm
ms 8%: Swan on: mud N «A 23% end d mom 2 dwm; mm $3 mm m2 $126 $3
Ndd mod d3: odd d $.N mod dddw 8d 2;
Wm dead S 3.: 3d _ dd 2.0 d2 d z: N vvm w Nu mm 2N 51w A SE
Nod mdd wdg odd d NEW dwd wmi odd m3.
wd d3.“ 8v 3%.; Ed M d; ohm m9 d R: S Nvm cm d3 v 92 «.de .5 SE
mdd Ed 8.3; Nddw d m5. wmd 3.2: wmdw wvdm
Md 25.2 dmwd do}: we; N Nu deed dmwg d m? B and mvm Zn 2 ddw Elm A $2
mod 8d Eda 8d. d m3. 3d $.de odd :3
3. ddmN mwv dud; «Nd N I dun 2: d «.3 E m9. 3 m2 w 92 veand mm:
mod «dd 3.3 deg d :4 Ed NE“. 2d «mi
Tm dead EN dmwd odd _ wd dmw; m8 d 3 cm dmm :_ 5w v .. vouwTd vm:
d wdd wddm odd d deV Ed 2.3 mm; mod ,
.00 Nm emu.» 8N E: Ed o 3 8w 8 o wvw ad add vw :. a can Eémé $8
d odd «odd. 3.5 d odd and wddE $22 mmdfi
ms 812 Em; ddmdfl a; d w; dvmd dmw d wwv 3 wam m2 dwm wv odd «onwwud «NS
. . «dd vdd Edw Ed d mod mud cmdw mg is
Nd dmdw Nmm 2;: Ed w, wd E mvm d Sun a mew wv m: Nv ddw volmwud End
. . 8d odd mnwv 2%: d. dwd and wmdn add wvfi
.oa ms dmmm d3 dmwm odd wd NA 83 ddm d w? E can mm dmw hm d9 Kimmie :md
. . . . 5d mod 3.; dde d 5N and mwdd $4; mad“
.Etsm_~u_mo_80.mdSflfiESa. ms 8:: do; dmmd : 5d dd dead 8: d m2 3 com; d2 2% mm ddm Euvwld 38
Gem“ 3 N u 3 S m. d S S .
.25 mm mm. m; mm 9w mm mm mm mm. mm mm mu am £0”..ch :3
5a 8::— nGnu—NO 553 u 0 m m. U m. m m .0 u m m. ( m. N m crux m ( m. 0 e 0m 5 .02
9.55: In 56:5 a 8:8 s( a a (m Gm m m. m G m 5.53:8 EB
~ 85 3263: 328me m m m do can
$2656
958%

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6:: 5m Ezufiisgzzfi E sienna mm 2.2 ‘5‘ 2:»: 530—
mtwoamfi 35:2: 2: go 28m 32:33. .2: Set 5.33: x0 333:3 EQNEMQUIQ mum—«H

Tia—13m Sam. o: 386.: Aid 2253 .8382: as 3936 .acofi
u2.: BA 2523—5: 5 "5:an fl 553.538 Ramsay—«u =a k8 2:»: Sun: 628%.: mm 58.8 an: EA flue—misgamfl. “Em 2.:— hi flaw-MEWS E mug—«=3

a do 5153:: £18 “Em ism—5:0 RAE—5:3 do EuEfimauQ .3 83:5;

. EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

December 1968 was much above average, particularly
during the period December 1967 through April 1968, and
flushing by floodwater caused a reduction in the ground-
water salinity of the alluvial deposits.

The effects of variations in discharge of the Gila River
on specific conductance of ground water in the alluvial
deposits are shown in figure 5. Two periods of above
average discharge were in early 1966 and early 1968, and
one extended period of below average discharge occurred
between July 1968 and July 1971. Values of specific
conductance generally are lower after periods of above
average runoff and higher during periods of below average
runoff, although this correlation is not always evident for
short periods.

The lowest value of mean specific conductance, 4,270
micromhos, for the period June 1964 to June 1972, was
obtained in December 1964 and probably was caused by
the large amount of runoff in August and September 1964
(figs. 5A, B). The highest value, 9,200 micromhos, was
during June 1967 following a year of below average runoff
in the Gila River. In the latter half of 1967 and the first half
of 1968, stream discharge was considerably above average,
and the mean specific conductance of ground water
declined during 1968 and 1969 to a low of 6,480 micromhos
in January 1970. Values increased to 7,840 micromhos in
June 1971, which was near the end of a prolonged period
of below average stream discharge. Details of the specific-
conductance variations during the period of above average
stream discharge from August 1971 to January 1972 are not
defined.

The variations in specific—conductance values of ground
water in the alluvial deposits versus stream discharge are
even more strikingly shown by a comparison of data from
individual wells and stream discharge (figs. 53, C). About
70 percent of the wells, from which sufficient specific-
conductance data were collected, showed an inverse
relation between specific conductance and stream
discharge. The relation is best illustrated by data from
wells nearest the river and least, if at all, from wells farthest
from the river. Well 0103 is adjacent to the river and well
0102 is about 0.3 mile (0.5 km) from the river (pl. 1).
Specific-conductance data from each well show: A sharp
decrease after the above average streamflow in early 1966
and August 1967 to May 1968, and a general increase
during most of the prolonged period of below average
stream discharge from June 1968 to June 1971 . The degree
to which specific-conductance changes were detected
depended not only on the distance of the well from the
river but also on whether the well was inundated by flood-
water and how long inundation occurred.

Fluctuations and maximum and minimum of values of
specific conductance were greater from 1964 to about 1968,
which corresponded to the period of greatest fluctuation of
stream discharge for the interval 1964 to 1972 (fig. 5). The
magnitude of variation in specific conductance of ground

 

M13

water from individual wells during the period 1964 to 1972
ranged from about 1,500 to 20,000 micromhos
(approximately 1,000 to 13,000 mg/l of dissolved solids);
fluctuations of between 5,000 and 10,000 micromhos
(approximately 3,200 and 6,400 mg/l of dissolved solids)
were the most common, however.

Temperatures of ground water in the saturated zone of
the alluvial deposits ranged from about 10° to 22°C.
Temperatures generally were lower and fluctuated more
in ground water next to the river than those in ground
water farthest from the river. Temperatures fluctuated as
much as 10°C in ground water from wells next to the river,
but fluctuations generally were less than 3°C in water from
wells near the outer margins of the flood plain. Lowest
and highest annual ground-water temperatures were
recorded in March and September, respectively.

GROUND WATER IN THE UNSATURATED ZONE

Ground water from the unsaturated zone was sampled
along the Gila River near wells 0103 and 1751 where the
depth to water ranged from about 1 to 10 feet (0.3 to 3 m)
below land surface during the period of the study. As many
as five 2-inch (51-mm) diameter plastic soil-water tubes
were inserted into auger holes that ranged in depth from 3
to 7 feet (1 to 2 m) at each location. Each tube had a porous
ceramic cup on the lower end and a petcock fitted into a
rubber stopper on the upper end. Air was evacuated from
the tube, and the petcock was closed to obtain a water
sample. The time required for sufficient water to pass
through the porous cup ranged from a few hours, if the
soil around the porous cup was nearly saturated with
water, to several days or weeks, if the soil-moisture content
was low. Thus, the date when the sample was removed
from the soil-water tube (date of collection) does not
represent the actual date when the water entered the tube
from the soil; the sample may represent water that entered
the tube any time after vacuum was applied on the
previous sampling date. Although this method of
sampling water was somewhat crude, it has provided some
insight into the variations of dissolved-solids concen-
trations and chemical composition of the water in the
unsaturated zone.

The site near well 1751 was established in June 1965. In
the spring of 1966 and summer of 1967 the site was
inundated by backwater and silt from San Carlos Reser-
voir. The sampling tubes were uncovered and maintained
until February 1968 when the deposition of more than 7
feet (2 m) of silt forced the abandonment of the site. The
site at well 0103 was established in March 1965 and was
destroyed by floodwater in March 1966. The site was
reestablished in March 1967 about 50 feet (15 m) from the
original site. The data from the site near well 0103,
although not as abundant or as continuous as the data
near well 1751, indicate similar trends. Specific-
conductance values, which ranged from about 1,000 to

M14 GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

 

    
           
 
   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1964 1955 1966 1967 1968 1969 1970 1971 1972
29,000 IIlllllllllllllllIIllIlllllllllllllllllllllllllIlllllllllllllllllllllll Illllllllll lllllllllllllllllllllll
E PLANATI N —
Most of reach 1 was cleared of . . a . . 0
_ phreatophytes by this date 1!- Maxi— Maximum, mean, and minimum values of specific —
_ l ‘ mum conductance determined from 34 selected wells -
25.000 _ that were sampled consistently during the period—
_ _ D Mean of data collection _
u: _ at Indicates that data from less than 75 percent of _
g _ _ the 34 wells were available to calculate the maxi- _
E g _ MIni— mum, mean, and minimum values. Missing data
— mum - - ‘
E w 20 000 _ usually were caused by wells being mundated nv_
u " ' floodwater.
_ W ,_ _
2 0 Most of reach 2a
g 53 . was cleared of
K .- ' [7 phreatophvtes by Most of reach 2b _
(“5' < — this date was cleared of “
2 m 15,000 — phreatophvtes by —
E E _ this date —
0 m
D 2 ' _
O — _ —i
Z l-
0 Z - ‘
Lu
8 0 10,000 — —
_ a: _ _
E :4 //"\
U " _
in HH”
0. — _
”I _ _
5000 -— J _
_ A ‘ J _
A
t J -
0 l||llllllIllllllLllllllllllllllllllllllllllllllll llllllllllllllllllllllllllllllllllllllllllllllllllllII|II
[LSFECIFIC CONDUCTANCE OF GROUND WATER IN THE ALLUVIAL DEPOSITS
200'000llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllTlTlllllllllllllllllllllllllllll 246.6
I
100,000 _, 9 _ 123.3
I 2 u;
c _ w
P _ - E
“J Lu
LII ,— _
u. E
d — - 9
I U
o ui
< — 1 — I
a 9
' 71 ll l ‘l A ‘°
01 3
0 10,000 123 U
0: 8 a z
< ‘1 _
I 1 ~
0 — uJ
‘3 - <
I
- 0
‘2
_ D
1000 _ - 1.2

 

 

 

 

I U
800 ”lulu.“ n.n.n.nlu.n.n.nln.n.n.nln.H.H.Hln.n.n.ul“.”.n.nlu.n.n.u H.H.H.H 999

B. HYDROGRAPH OF GILA RIVEFI AT CROSS SECTION 9

 

lllllllllll llllI||lIlllllIIIllllll'llllllllIiIIlllllllllllllllllllllllllllllllllll'lllllllllll lllllllllll

12,000 '—

10,000 —

8000 —

  

Well 0103

   
 
 

6000 —

  

Well site 0103
inundated by
floodwatsr

   

SPECIFIC CONDUCTANCE, lN MICROMHOS
PER CENTIMETHE AT 250 CELSIU

4000 —

 

 

 

0 ”1H.H.H|”.H.H.”I“.H.H.“lulu.nlnln.n.n.nlnlnln.ulu.”.n.ulu.n.n.nlu.“tut”
1964 1965 1966 1957 1968 1969 1970 1971 1972

C. SPECIFIC CONDUCTANCE 0F GROUND WATER FROM WELLS ON CROSS-SECTION 1

FIGURE 5.—Discharge of the Gila River and specific conductance of ground water from the saturated zone of the alluvial deposits.

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

22,000 micromhos (about 600 to 14,000 mg/l of dissolved
solids), had similar variations as those of ground water in
the saturated zone (fig. 6). Values were highest from water
in the unsaturated zone following periods of prolonged
low flow in the river; concentrations were lowest after
major flood events, which flushed the accumulation of
salts from the unsaturated zone. During periods of pro—
longed low flow, not only did the specific conductance of
the water in the unsaturated zone increase, but it was
higher near the land surface than near the water table (data
from well 0103, June 1965, fig. 63, table 5; well 1751, Aug.
1965, fig. 6C, table 5).

The chemical composition of the water changed from a
sodium chloride type near the water table to a calcium
chloride type near the land surface (fig. 7). Flushing of
salts from the soil by large flood events reduced the specific
conductance and changed the water to a sodium chloride
type throughout the zone or a sodium chloride type near
the water table and a calcium sodium bicarbonate type
near the land surface (data from January 1967, figs. 6C, 7,
table 5).

Specific conductance of the soil water was as high as
22,000 micromhos (about 14,000 mg/l of dissolved solids)
at the site near well 1751 (fig. 6C) and about 15,000
micromhos (about 9,000 mg/l of dissolved solids) at the
site near well 0103 in the summer of 1965 (fig. 63). During
the winter of 1965-66 the sites were inundated by flood
water, and specific-conductance values dropped to as low
as 5,000 micromhos (about 3,000 mg/l of dissolved solids)
at site 0103 in early 1966 and to about 2,000 micromhos
(about 1,300 rng/l of dissolved solids) at site 1751 by mid-
1966. The specific-conductance values generally increased
to as much as 6,000 micromhos during the period mid-
1966 to early 1968 corresponding with a period of below
average streamflow (figs. 6A, C).

During the period mid-1966 to early 1968, zonation of
specific conductance formed in the unsaturated zone, but
the reverse of that developed prior to the flood event in the
winter of 1965—66—that is, values were higher from water
nearest the water table (fig. 6C). Data from samples
collected in early 1968, however, indicate that the near-
surface buildup of salinity in the unsaturated zone had
begun. _

The near-surface buildup of dissolved solids in water in
the unsaturated zone during prolonged periods of low
streamflow at the sites near wells 0103 and 1751 is caused
by the upward migration of water from the saturated zone
and subsequent concentration by evapotranspiration. The
reason for the change from a sodium chloride composi-
tion of soil water near the water table to a calcium chloride
composition near the ground surface is not so apparent,
but it may be caused by ion exchange of sodium from the
water for calcium in the sediments as the water moves
upward through the unsaturated zone. The calcium-rich
water might also be caused by differential rates of upward

 

M15

migration of sodium and calcium in the unsaturated zone.
Sodium may be more concentrated than calcium in the salt
crusts that form in places on the surface of the alluvial
deposits where the water table is close to the surface.
Most of the calcium chloride water contained large
amounts of bicarbonate, and the water may have been
saturated with respect to calcite at the indicated pH of the
water. For example, a sample collected on September 1,
1965, from 3 feet below land surface at the site near well
1751 contained 13,500 mg/l of dissolved solids of which
2,308 mg/l were calcium and 561 mg/l were bicarbonate
(table 5). The laboratory pH of the sample was 7.3, but this
value may not be representative because the pH could have
changed from the time of collection until the time the
laboratory measurement was made. Calcium carbonate
precipitated in the sample bottles from some samples of
soil water after standing several weeks. The large amounts
of calcium and bicarbonate in the water probably are pre-
vented from precipitating in the unsaturated zone by a
high partial pressure of carbon dioxide, which may be 10
times that in air (Bear, 1955, p. 205), caused by decay of
organic material and root transpiration. In addition,
calcite is more soluble as the salinity of water increases.

EFFECT OF PHREATOPHYTE REMOVAL
ON WATER QUALITY OF THE
ALLUVIAL DEPOSITS

An evaluation of the data presented in the previous
sections shows that the fluctuations of dissolved-solids
concentrations of water in the alluvial deposits and the
Gila River are large. The fluctuations of dissolved-solids
concentrations of water in the alluvial deposits are caused
by the large fluctuation of streamflow of the Gila River
and evapotranspiration. With this large amount of
“natural” variation in dissolved-solids concentrations the
detection of changes, if any, in water quality due to
phreatophyte removal is complex.

The factors that could cause changes in water quality in
the alluvial deposits are discussed qualitatively in the first
part of this section. Results of a statistical analysis of the
effects of phreatophyte removal on the specific con-
ductance of water in the alluvial deposits are presented in
the second part of the section.

FACTORS RESPONSIBLE FOR CHANGES
IN WATER QUALITY

The Gila River, by flushing away accumulated salts and
recharging fresh water, tends to reduce the dissolved- solids
concentrations in the alluvial deposits. Conversely, evapo-
transpiration during periods of prolonged low flow
increases the concentration of salts. If other factors were
constant, the removal of the phreatophytes from the Gila
River flood plain would reduce the amount of evapo-
transpiration, and a greater amount of water would be

M16

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

 

 

 

 

 

 

 

 

 

 

 
 

 
 
 

 

 

 

 

 

 

 

 

 

 

 

        

 

 

 

2001000 IllllIll"lllllllllllllIllllllllllllillllllll|l|llll|llIlllIll'llllll‘llllllllllll‘lllllllll‘ll llllWll||11246'6
100,000 _ f a 123.3
_ — <n
_ _ LU
CC
— .. l—
l- . LLI
3 — _ 2
1+ E
“J U
I: ~— _ u:
U I
: T1 fl 1 9
_ m
u; 10,000 A 12.3 8
0
0: : E
g _ u:
U (D
‘2 ‘ '5:
D _ _ I
U
— — ‘2
0
1000_ u . u n. “1.2
800 lllllllllllllllllllllllllllIll]lllllllllI[lIIlllllllllllll‘lllllllllllllllllIllllllllllllllllllllllllllllll 099
1964 1965 1966 1967 1968 1969 1970 1971 1972
A. HYDROGRAPH OF GlLA RIVER AT CROSS SECTION 9
_J _l
1-"; 2570 llllllllllI'lllllllllllllllllllllllllllllllllll llllllllllllllllllllllllevlllllvllllllilvIlvrll 783.3 3:
w m Land surface 1! Lu
E .I Altitude of water table E 'J
<( 256 in well 0103 2 <12
3 w 5 — ‘2, Open to alluvial ~ 781.8 W
- 1/) 0°, deposits only at E m
“4 2 «°/ bottom of casing . 2
D < e0/ \ Soil water Soil water W ‘1
3'41 2560— A, tubes tubes —780.3 0””
:5 March 1965 April1967 3’5
Pu, to to but
_|> arch 1966 June 1972 l3)
<0 Wel|0103 0
g 2555 . 778.8 <3
22looollllllllllll'lllllllIlllllllllllllllllllllllllllllllllllllll[IlllllllllIIIlllllIIlllllllllllllllllll|llllv
EXPLANATION
20,000— ° _.
SpecificConductance
March 1965 to March 1966 April 1967 to June 1972
13:000— . A 4f“: below land surface A——- 4.5 feet below land surface _
m o o 5 feet below land surface . . 5.5 feet below land surface
0 0 o 7 feet below land surface 0 o 6.5 feet below land surface
g $1 16,000 — o o 7.5 feet below land surface —
0(7) a 0—0 Well0103
1:4
out
50 14.000— § _
o «-
2m
_: S
g< 12,000— g}: _
qu if
(I —w
i—l— fig
ow 10,000— 5-; _
3E 30 o
— ..
2'; w 5
0m ":9
00 8000— ”M- _
9c: o+°
um /
6" 9 o
w 6000— \ / l o—° _
a. co /
m 0\ /°
1 o
4000— _.
.’
A/A /
\{o 5/
2000— l/" A‘ \A —
\./\:7
j ’ . ‘
llllllllllllllllllllllll lIlllllljllllllIlllllllllllllllllllllllllllllllIllllllllIlllllIllllllllllllllllll
1964 1965 1966 1967 1968 1969 1970 1971 1972

3- SPECIFIC CONDUCTANCE OF GROUND WATER NEAR WELL 0103

FIGURE 6 (above and facing page). —— Discharge of the Gila River, ground-water altitudes, and specific conductance of ground water from the
unsaturated zone and saturated zone of the alluvial deposits.

EFFECTS OF PHREATOPHYTF. REMOVAL ON WATER QUALITY

2475

M17

754.4

 

Ilvtlllllllvl

  

 

 

     
  

 

 

 

 

 

 

 

 

_| ”In"1""H””'1””'"””1""H""‘lt'a'l‘ldlgdhgc'elnn'"H" (”d
w
W >
'- > <5 11:
( ul
‘“ '5 2470— 909» #7523 l- .1
Lu at, w
u. < $,/ 2 <(
ul
_Z_ (“/i Soil water tubes 2 (n
u; z 2455 — Altitude of water table — 751.3 '. Z
O < in wall1751 g:
03- ; Well open to alluvial deposits 3 2
_ . p.
5 ul 2460 __ only at bpttom of casmg _ 749.8 I: g
< 5 -‘ 0
m ‘1 en
< Well 1751 <
2455 748.3
22,0
°° l l I I
EXPLANATION
20 000 Specific conductance of water
' u D 2 feet below land surface
A A 3 feet below land surface
18,000 _ A . 4 feet below land surface _
o o 5 feet below land surface
8 o u 6 feet below land surface
— ll
1 m 16,000— ° ° V“ 1751 ——
g 2
tn
I _, e 5
2 u: o ‘5
2 0 14,000— t 2 _
2°10 S u
_. m E g
ui 1- _ -— _
o < 12,000 E E
Z w : a
S E n g
D m 10,000 — g u. _
3 E 'g .5
Z 1' n in
O 2 o g
o l“ 3000 — 3: g _
O U U) u-
- I
E w
0 IL
w 6000-— _
n.
U)
4000— ._
zooo — \ _
o i
onunnull ........ 1.. |l|llllllll1||llllJllll| IIIIII|l||1Illllllllllllllllllllll
1964 1965 1966 1967 1968 1969 1970

C SPECIFIC CONDUCTANCE OF GROUND WATER NEAR WELL 1751

FIGURE 6 — Continued.

present in the ground-water system. The increased
amount of ground water may cause the water levels to be
higher and (or) the base flow of the Gila River to be
greater. After the phreatophytes were removed from reach
1, the evapotranspiration was reduced from 50 inches
(1,270 mm) per year to 20 inches (508 mm) per year
(Hanson, 1972b), and water levels did not decline as
rapidly during the growing season as did water levels
before removal of vegetation (F. P. Kipple, oral commun.,
1972). Base flow of the Gila River increased after clearing,
but antecedent moisture conditions before each reach was
cleared made the calculations for the actual amount of
increase very complex (R. L. Hanson, oral commun.,
1974). Removal of the phreatophytes should likewise
cause a less rapid increase in dissolved-solids concen-
trations of the ground water during the growing season.
However, upward movement and subsequent direct
evaporation of ground water from the cleared surface
occurred even after phreatophytes were removed. Soil-
moisture data from the project area indicated that, in terms
of the liquid phase, this upward movement of water in the

 

unsaturated zone probably is not significant where the
depth to the water table is more than about 6 feet (2 m)
below land surface (F. P. Kipple, oral commun., 1973;
McQueen and Miller, 1972, p. E47).

Other factors, such as infiltration of precipitation and
accumulation of salts in the soil as a result of water exuded
by saltcedar, are of minor importance in affecting water
quality. Effects of increased infiltration of rainfall on the
cleared surface probably are negligible; even though the
phreatophytes no longer are present to intercept the
precipitation, the precipitation evaporates readily from
the cleared ground surface.

Saltcedar can grow and thrive in areas of salty water by
excreting fairly concentrated solutions of salt through the
leaf surfaces by a process known as guttation or through
”salt glands” (Hem, 1950, p. 80; 1967, p. C2). One sample
)f this exuded water contained 41,100 ppm (parts per
million) of dissolved solids of which 18,200 ppm and
13,800 ppm were chloride and sodium, respectively (Hem,
1950, p. 81). The removal of the exuded water and salt from

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

M18

.EBE «— 5 3588.38“. 338.» In =<N
.3555ng 35.52% .850 + 38:38.85 x 3.6 9:35 .3 vs: E8 28 25:3 82—8 330885.

 

 

 

 

 

 

 

. No.8 8.8 8.2 No.8 8 2.2 8.2 88.8. 8.:
8.8 88.8 8N8 8N8: : N: 888 8: 8 8N8 888 N8 8N 8 8
. 8.8 88.8 88.2 :N 8 N82 8.2 8.8. 8.2
8.8 88.N 88 8N8: N 8.: 88 82 8 888 8N 8 N8N 88 8
. 8.8 8.8 No.8 No.8 8 8:: 88.2 88.8 88.8.
E 85 N8 8N: N 8.8 8: 8: 8 88 88 2. N8 8 8
. No.8 8.8 8N 88: 8 88.:N 8N2 88.N 8.8
8.8 88.N 888 2: _ 8.8 82 N8 8 88.: 8N8 88 N2 88 8
. No.8 88 8.: 88: 8 :.8N 88.: :N8 88.:
8.8 88.N $8 8: . 8.8 88 8.8 8 23 88N 8 8N 2. N .. ........ .....N82 .2 d8
. 8 8.8 8.8 88.2 8 8.2 :88 N88 8.2
E 88.8 883 88 8 8 8.8 88.: :8 8 888 88 NB 2.8 8 8
. No.8 8.8 88.88 8.8 8 $2 8.8 N82 8.88
E 88.: 2.8.8 8: 8 2 8.8 88.8 88.: 8 88 28.: 8NN 88.: 2. 8
. 8.8 8.8 8.82 88.88 8 8.2 88.8 :28 82.:
E 88.2 888.8 88 : N 8.8 2.8.8 88.: 8 E 8N8 88 88.N 88 8
. . 8.8 5.8 88.8: 8.2 8 88.N N8.N 8.8 8:: ,
8.8 82.8: 2.8 8 8N 8 N N8 888 888 8 82 88 888 8N.N 8N: N .. 882 .2 to
. 8.8 8.8 8.; 8.2 8 8.8 8.8 2.8 88.2
I 83 88 88 8 8 8.8 83 88 8 888 88 N8 88N 88 8
. 8.8 8.8 8.8 N8.8N 8 8.2 N888 8.8: :5.
E 88.8 82.8 28 N N 8.8 88.N 8N: 8 8.88 88.: N2 :8 8 8
. 8 88.8 8.2 8.8 8 8.2 N888 SNN 8.88
8.8 88.2 88.8 88 8 8 8.: 88.8 88.: 8 888 888.: NNN 88.: 8 8
. 8.8 8.8 88.82 88.8 8 2.8 2.88 88.8 2.8:
8.8 88.NN 88.: 88 2 N 8.8 88.8 88.: 8 :88 88: 8.88 88.N 8 8 ........ . .......... .. ........ 882 .2 .888
. 8.8 5.8 88.8 88.:N 8 88.: 8N.N8 88.: 2.2.
8.8 88.2 28.8 88 8 N N8 88.8 88.: 8 88 88.: N:N 888 88 8
. 88 No.8 8.8 8.: 8 8.2 :88 88.: 8.5.
8.8 88.8 8N8 88N 8 N 8.8 88.N 888 8 8N8 am N_N 88 88 8
. 8.8 8.8 8.82 8.8 8 8.: 8.88 88.8 8.88
E 88.: 82.8 28 8 N 8.8 2.8.8. 8: 8 88 883 82 88.: 88 8
. 8N8 No.8 8.8: NN.N8 8 88.2 8.8 :8 R8:
3 88.: 82 88 : E 8.8 88.8 888.: 8 $8 2.8.: 88 8N.N 8 8 .......... .. .......... 882 .8 .83.
SE :93 So: Sum
No.8 2.8 88.8 N88 8 88.8 N8.8N 8N 88.8
I 8NN.8 8N8 88.: : 8N 888 EN 8 8:. 88 88 N: 8 N
2.8 8.8 838 N82 8 88.8 8.88 8.: :8
N8 88.8 2N.N 88.8 8 E 8N N 88 8 88 28.: E 888 82 8
. 2.8 88.8 8.88 N82 8 8.8 2.8 8.2 8.88
m.B OOONW COm m 051m Na CA Owa N OVQ O V—N ®®© ®®~ ®N® CG «a ........ ..... 2.2.2.22... mmmfi .mw UCSH
.38 =83 in: 85
6°me
N H D S K“.
EEK 606 .353 m m. m m m m 9 m m m m m. >d m m M m m m m. s .88 8:
NIQ HEP—05:5 :88 88:08 :0 m ( W U. m. 50 m :0 m. m .w... W m m. 8/w m ( m. zO e 8&2: :52 8.23218 «0 Ban
0 = 83 ( a a ( ( 2 + w I. (
”Wyn—mus u E “52385 N ".1 m (d W. W m 3075 saun—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

chofia no 2:853:83 .mzom ER PamMEmAU HEB—:UWHMa‘ we 85:58an .3 meg—mam
”vb: 5Q flzufizmamzzp: E votes—w.— mm 2%: .3 Piw: .532 55: 5Q 8535:: E 3:82 3 £=~3138~N£Ewnu :m :2 85mm :35: ”@3883: 88 385 Pa: nun flau~m>m:g==E GEN 95: Non— EEEEE. E avg—«:4;
8:39.48 3.33288 2: 3 8:3 8885:8888: 8:: Sci 833 «No 88835:» Nufifiufiuld ”SEQ.

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

  
   
 
   
  
 

EXPLANATION

SITE NEAR WELL 0103
.7 June 9, 1965
SITE NEAR WELL 1751
06 August 6, 1965
.3 September 1, 1965
A4} October 1. 1965
A2 January 19, 1967

Number on data points indicates the
sampling depth below [and surface;
chemical analysis of each sample
is shown in Table 5

 

 

 

 

 

M19

 

 

 

 

 

 

 

 

CATIONS

ANIONS

FIGURE 7.—Chemical composition of water from the unsaturated zone of the alluvial deposits.

the leaves and stems by dew, precipitation, and wind
temporarily may cause large concentrations of salt to
accumulate in the upper part of the soil. The presence of
the salt in the soil was indicated by the rapid corrosion of
steel well casings and aluminum soil-moisture access
pipes at the point where they emerged from the ground.
The very rapid decrease in specific conductance of ground
water after a major flood event, especially near the river,
indicated that salt accumulations in the unsaturated zone
resulting either from exuded water or from upward migra-
tion and evaporation of water from the water table have no
major effect on the salinity of ground water in the
saturated zone of the alluvial deposits.

 

Removal of phreatophytes in the area surrounding the
observation wells caused no significant change in the
specific conductance of water in the alluvial deposits. The
general pattern of graphs of specific conductance versus
time for all wells in the alluvial deposits for which
sufficient data were available is similar, regardless of the
date of phreatophyte removal—that is, large fluctuations
occurred prior to 1968 and much smaller fluctuations
occurred after 1968. (See section on ”Water Quality,
Alluvial Deposits”) Wells on the right side of the river on
cross sections 1 and 3 were in areas that were cleared of
phreatophytes in 1964, yet the specific-conductance plots
of the wells in this area show the same general pattern

M20

before and after 1968 as those in areas that were cleared of
phreatophytes at a later date. These graphs indicate that
any changes in water quality caused by phreatophyte
removal are of less magnitude than the normal variations
in water quality that are con trolled by the effects of stream-
flow of the Gila River and of evapotranspiration and,
therefore, are not detectable.

STATISTICAL ANALYSIS OF EFFECTS OF
PHREATOPHYTE REMOVAL ON SPECIFIC
CONDUCTANCE OF WATER IN THE

ALLUVIAL DEPOSITS

By H. W. HJALMARSON

The objective of this analysis is to determine whether
the specific conductance of water in the alluvial deposits
changed as a result of the removal of phreatophytes. A
simple method of determining whether a change occurred
is the comparison of the mean of the preremoval specific-
conductance measurements with the mean of the post-
removal specific-conductance measurements, using
statistical tests. The first statistical test used is the
“Student’s” t, where the specific-conductance measure-
ments are assumed to be from normally distributed
populations. The second is a rank—sum test, where the pre-
removal and postremoval measurements are assumed to be
from continuously distributed populations that differ
only in their means. These two tests are advantageous
because of their simplicity and because all measurements
of specific conductance are used for each test. When using
these tests, however, a change in specific conductance of
ground water might be caused by flooding of the Gila
River or by prolonged periods of evapotranspiration in
absence of flooding and not caused by the removal of the
phreatophytes. Conversely, a sameness in measured
specific conductance of ground water might be the result
of offsetting actual conductance changes caused by the
phreatophyte removal with a flow change of the Gila

River. Thus, a third test known as a factorial experiment is \

used that considers the effect of a single factor, the phreato-
phyte removal.

For the first statistical test all measurements of specific
conductance can be used to test the hypothesis that the
mean of the preremoval conductance is equal to the mean
of the postremoval conductance. A two—sided “Student’s” t
test of the true difference between the means of two normal
populations is used. The means of the 342 preremoval and
302 postremoval conductance measurements were 7,410
and 7,130 micromhos, respectively. The hypothesis could
not be rejected at a statistical confidence level of 95 percent
and, thus, a change of ground-water conductance is
unlikely.

The second statistical test used is the rank-sum test of the
true difference between means of two continuous
probability distributions that might not be normal (Dixon

 

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

and Massey, 1957, p. 289). The hypothesis that the means
of the two populations are equal is rejected if the test
statistic, T’, is significantly large or significantly small
where T’ is the sum of the ranks of the 302 postremoval
conductance measurements. The computed T’ value of
96,060 is well within the 5 percent critical limits; there-
fore, the hypothesis cannot be rejected. The preremoval
and postremoval measurements of specific conductance
probably are not from different populations.

The equality—of—means hypothesis could not be rejected
using either of the tests; therefore, unequal means of the
preremoval and postremoval specific-conductance meas-
urements are unlikely. If the factors affecting ground-
water specific conductance before the removal of vegeta-
tion were the same as those after the removal, it might be
concluded that the effect of the vegetation removal on the.
specific conductance of water in the alluvial deposits was
insignificant; A change of ground-water conductance
resulting from the vegetation removal may be masked
because the factors affecting the amount of ground-water
specific conductance, such as the amounts of streamflow
in the Gila River (the main factor), were not the same
before and after the removal of the phreatophytes. The
following test considers the significance of each factor
affecting ground-water specific conductance.

In order to isolate the effect of vegetation removal, the
third test utilizes a reach of river where the phreatophytes
are removed and a reach of river where the vegetation cover
is undisturbed. These reaches of river are referred to as the
removal and control reaches, respectively. The removal
and control reaches are considered as two levels of reach;
the periods before and after the removal of the phreato-
phytes are considered as the two levels of period. Thus,
reach and period are the two factors of the experiment, and
each factor is at two levels. The combination of the levels
of each factor yields four experimental conditions of the
factorial experiment for which ground-water specific
conductance was measured. The analysis of variance tech-
nique for testing of a significant change is used, and the
mathematical model for each of the three reaches from
which the phreatophytes were removed is

Xijk =M+Ri+Pj +(RP),j +6171: ’
where
X ijk = kth measurement of conductance within
the ith reach during the jth period,
,1). = true mean effect for whole experiment,
R1. = effect of ith reach,
P j = effect of jth period,
(RP)I.]. = effect of reach—period interaction,
e ijk 2 random error present in the kth meas-
urement on the ith reach and jth period.

The error term of the model is assumed to be an
independently and normally distributed random effect
with a mean value of zero and a variance equal at all com-
binations of reaches and periods. Also, the sum of all

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

reach, period, and reach-period effects is assumed to equal
zero.

If a change in ground-water specific conductance of the
removal reach follows the removal of phreatophytes and
this change is different than a change in ground-water
specific conductance of the control reach, there is an inter-
action between the periods and reaches. This interaction is
a measure of the effect of phreatophyte removal on
ground-water specific conductance provided the
following conditions are met: (1) Movement of ground
water between the removal and control reaches is
negligible; (2) both reaches are subject to approximately
the same factors affecting ground-water specific
conductance; and (3) a change of a factor affecting ground-
water specific conductance will result in similar changes
of specific conductance in both the removal and control
reaches. Effects of the movement of ground water between
reaches on specific conductance are considered negligible
because the average velocity of the ground water moving
between the reaches is less than 400 feet (120 m) per year
based on data from Hanson (1972a, p. F4, F25). Hanson
(1972b, p. 12) estimated that the total surface and sub-
surface flow into reach 1 is about 210,000 acre-feet (259
hm’) per year, of which 95 percent is contributed by the
Gila River and only 5 percent by tributary runoff, pre-
cipitation, and ground-water inflow. In addition, the
amount of variation in specific conductance of water in
the Gila River is large, and the river has a very significant
flushing effect on the ground water near the river. Because
the flow changes of the Gila River are the principal cause
of changes in specific conductance of ground water, the
specific conductance of ground water for the reaches is
affected similarly by a change in the amount of flow of the
river. Thus, if the change of ground-water specific
conductance of the control reach is significantly different
from the change of ground-water specific conductance of
the removal reach after the phreatophytes are removed,
this difference is due to the removal of phreatophytes (fig.
8).

Reaches 1, 2a, and 2b are analyzed separately in this
analysis because the phreatophytes in each reach were
removed at different times (table 6). Reach 2a is used as the
control for reaches 1 and 2b, and reach 1 is used as the
control for reach 2a. Reach 2b is not used as a control reach
because a number of observation wells in the reach were
inundated by water from San Carlos Reservoir following
the high flows during the winter of 1967—68. Four of the
wells in the reach were inundated for at least a year. Reach
3 is not used in the analysis because of the limited amount
of data. The results of the analysis of variance and tests for
significant effects are given in table 7. The hypothesis of
interest is that there is no interaction between the reach
and period factors. Bias due to missing measurements of
specific conductance was removed by treating the data as a
case of disproportion subclass numbers (Ostle, 1960, p.

 

M21

300). For reaches 1, 2a, and 2b, the computed value of F for
the reach-period interaction is less than the value of F at
the 5-percent level of significance. Therefore, the effect of
interactions is not significant for any of the reaches, and a
significant change of ground-water specific conductance
resulting from the removal of the phreatophytes is
unlikely.

Inherent in the analysis is the effect of variable amounts
of flow in the Gila River on the measured ground-water
specific conductance. The magnitude of the conductance
change corresponding to a change of river flow varied
from well to well and therefore increased the magnitude of
the error (sijk). This increase of error could be significant
because the computed values of F given in table 7 are
inversely proportional to the computed amounts of error.
Had the flow of the Gila River been less variable, the con-
clusions reached might have been different; thus, only
judicious use of the results should be made.

A significant change of ground-water specific conduct-
ance was not detected using the three tests. The con-
clusions reached are consistent, are based on statistical
tests requiring different assumptions, and are based on a
large number of specific-conductance measurements.
Thus, the removal of the phreatophytes in the study
reaches did not result in a specific-conductance change of
ground water in the alluvial deposits that could be
ascribed solely to removal rather than to changes in river
discharge.

SUMMARY

The water quality in the Gila River Phreatophyte
Project area is considered in terms of four hydrologic
sources: the Gila River, the tributaries to the Gila River,
the basin fill, and the alluvial deposits. The dissolved-
solids concentrations and chemical composition of the
Gila River ranged from about 300 mg/l and a calcium
sodium bicarbonate type at high flows to about 4,800 mg/l
and a sodium chloride type at low flows. Most of the year
sodium and chloride are the principal ions in the river
water. The Gila River is the main source of recharge to
the alluvial deposits.

Water in tributaries to the Gila River generally
contained less than 300 mg/l of dissolved solids and was a
calcium bicarbonate type. The total amount of tributary
flow was less than 10 percent of the flow of the Gila River,
and the tributary flow had little effect on the water quality
of the Gila River or on the ground water in the alluvial
deposits.

The dissolved-solids concentrations of water from the
basin fill ranged from about 200 to 5,000 mg/l. Water from
the silt and sand facies contained 1,000 to 3,000 mg/l of dis-
solved solids in the fine—grained parts of the unit and less
than 1,000 mg/l in the coarse-grained parts of the unit.
Water from the limestone facies had the poorest quality
and contained more than 3,000 mg/l of dissolved solids.

M22

 

 

10,000 I I
LL!
‘2’ m

U)
<02
I-Im 9000—
02.:
Dow
05:0
69%
92»- 8000—
9E<
E‘tu
319m

21--
Q.
mull-u 7000—
m2;
ogr-

Z
“JDLIJ
0100
E50: 6000—
wig:
>
<

5000 l , '

 

 

 

REACH OF GILA RIVER

FIGURE 8.—Interaction of specific conductance of ground water between
reaches 1 and 2a and periods 1 and 2.

Water that contained less than 500 mg/l of dissolved solids
had a calcium bicarbonate to sodium bicarbonate com-
position, and water that contained more than 500 mg/l
of dissolved solids had a sodium chloride composition.

The ground water in the alluvial deposits, which
supports the growth of the phreatophytes in the project
area, is considered in terms of the saturated zone and the
unsaturated zone. Most of the data are from the saturated
zone. The dissolved-solids concentrations of water in the
saturated zone ranged from about 400 to 19,000 mg/ 1, but
most concentrations ranged from 3,000 to 7,000 mg/l. The
water is a sodium chloride type similar to water from
the Gila River. The lowest concentrations of dissolved
solids are near the Gila River, and concentrations increase
with increasing distance from the river.

Water from the unsaturated zone of the alluvial deposits
was sampled at two sites in the report area. Dissolved-
solids concentrations ranged from about 600 to 14,000
mg/ 1. Concentrations were greatest in the upper part of
the unsaturated zone following prolonged periods of low
flow in the river during which time the sites were not
inundated by floodwater. Inundation of the sites by flood-
water flushed the accumulations of salts from the zone
and removed the stratification of dissolved solids. The
stratification redeveloped in the zone in absence of major
river flooding. During periods when the dissolved-solids
stratification was present, the chemical composition of
the water ranged from a calcium chloride type in the
upper part of the,zone to a sodium chloride type near the
water table.

Variations in dissolved-solids concentrations in water
in the alluvial deposits are related mainly to evapo-
transpiration and the amount of flow in the Gila River.
During periods of above average streamflow, the

 

GILA RIVER PHREATOPHYTE PROJECT, ARIZONA

TABLE 6.—Summary of data for factorial experiment

 

Average specific

 

  

 

Reach Period‘ Condition Number 0‘ conductance,
of reach measurements . .
in micromhos
l .................... 1 Phreatophytes ________ 53 9,690
' 2 Cleared 98 7,980
3 . 65 6,870
4 25 7,510
2a .................. l 58 6,920
2 109 6,020
3 ' 70 6,060
4 28 6,520
2b .................. l 41 6,560
2 ' 47 8,510
3 34 8,620
4 16 8,040
Total and
average ....... 644 7,280

 

‘Period 1, March 1966 to March 1967; period 2, June 1967 to March 1969; period 3, June 1969 to
March 1971; period 4, June 1971 toJune 1971.

TABLE 7.—Summary of two~way analysis of variance for ground-water
specific-conductance measurements

 

 

 

 

 

 

 

 

 

 

Source of Degrees of Mean Computed
variation freedom square F (1-0‘05)
Reach 1
Between reaches ’
1 and 2a ................. 1 398,650,000 27.3 3.9
Between periods
1 and 2 ................... 1 118,670,000 8.] 3.9
Reach-period
interaction ............. 1 12,100,000 .8 3.9
Error ............... 314 14,600,000 ..........
Reach 2::
Between reaches
2a and 1 ................. 1 192,860,000 20.9 3.9
Between periods
2 and 3 ................... 1 20,020,000 2.2 3.9
Reach-period
interaction ............. 1 28,230,000 3.1 3.9
Error ............... 338 9,230,000 ..........
Reach 21)
Between reaches
2b and 2a ............... l 165,890,000 34.4 3.9
Between periods
3 and 4 ................... 1 690,000 , .1 3.9
Reach-period
interaction ............. 1 7,760,000 1.6 3.9
Error ............... 144 4,820,000 ..........

 

dissolved-solids concentrations of the water are low, and
subsequent recharge of this water reduces the concen-
trations in water in the alluvial deposits. The amount of
variation is dependent on the frequency, magnitude, and
duration of major flood events. The amount of variation
in dissolved-solids concentrations in water from
individual wells for the period of the investigation (8
years) ranged from about 1,000 to 13,000 mg/l, but varia-

EFFECTS OF PHREATOPHYTE REMOVAL ON WATER QUALITY

tions of between 3,200 and 6,400 mg/l were most common.
Water from wells tapping the alluvial deposits near the
river had the greatest variation in dissolved-solids concen-
trations, and water in wells farthest from the river had the
least variation. The amount of these variations may have
masked changes, if any, in dissolved-solids concen-
trations of water caused by phreatophyte removal.

The effects of the phreatophyte removal on the specific
conductance of water in the alluvial deposits were
analyzed using three statistical tests. The results of the
“Student’s” t test, a rank-sum test, and a factorial exper-
iment indicate that the removal of phreatophytes did not
significantly affect the specific conductance of water in the
alluvial deposits.

The absence of a significant change of ground-water
specific conductance for the area studied does not mean
that phreatophyte removal will not result in detectable
conductance changes of ground water in other areas. In the
reach of river studied in this report, the water in the
alluvial deposits was flushed or freshened by floodwater
from the Gila River, This flushing may have masked the
effects, if any, of the phreatophyte removal on the ground-
water specific conductance, and therefore a change in
specific conductance was not detected.

REFERENCES CITED

Bear, F. E., 1955, Chemistry of the soil: New York, Reinhold Publish—
ing Corp., 373 p.

Burkham, D. E., 1970, Precipitation, streamflow, and major floods
at selected sites in the Gila River drainage basin above Coolidge
Dam, Arizona: U.S. Geol. Survey Prof. Paper 655—B, 33 p.

Culler, R. C., and others, 1970, Objectives, methods, and environment—
Gila River Phreatophyte Project, Graham County, Arizona: U.S.
Geol. Survey Pr'of. Paper 655—A, 25 p.

 

M23

Dixon, W. S., and Massey, F. J., 1957, Introduction to statistical analy-
sis: McGraw-Hill Book Co., 488 p.

Fenneman, N. M., 1931, Physiography of Western United States: New
York, McGraw-Hill Book Co., 534 p.

Gatewood, J. 5., Robinson, T. W., Colby, B. R., Hem, J. D., and
Halpenny, L. C., 1950, Use of water by bottom-land vegetation.
in lower Safford Valley, Arizona: U.S. Geol. Survey Water-Supply
Paper 1103, 210 p.

Hanson, R. L., 1972a, Subsurface hydraulics in the area of the Gila
River Phreatophyte Project, Graham County, Arizona: U.S. Geol.
Survey Prof. Paper 655—F, 27 p.

1972b, Changing the consumptive use on the Gila River flood
plain, southeastern Arizona: Spokane, Washington, Irrig. and
Drainage Div. Spec. Conf., 21 p.

Hem, J. D., 1950, Quality of water of the Gila River basin above
Coolidge Dam, Arizona: U.S. Geol. Survey Water‘Supply Paper
1104, 230 p.

1967, Composition of saline residues on leaves and stems of
saltcedar (Tamarix pentandm Pallas): U.S. Geol. Survey Prof.
Paper 491—C, 9 p.

McQueen, I. S., and Miller, R. F., 1972, Soil moisture and energy
relationships associated with riparian vegetation near San Carlos,
Arizona: U.S. Geol. Survey Prof. Paper 655—E, 51 p.

Ostle, Bernard, 1960, Statistics in research: Ames, Iowa, Iowa State
Univ. Press, 487 p.

Piper, A. M., 1945, A graphic procedure in the geochemical interpreta-
tion of water analyses: Am. Geophys. Union Trans. 25th Ann.
Mtg., pt. 6, p. 914—928.

U.S. Public Health Service, 1962, Drinking-water standards: U.S. Pub-
lic Health Service Pub. 956, 61 p.

U.S. Weather Bureau, 1964, Decinnial census of the United States
climate; climatic summary of the United States—supplement for
1951 through 1960, Arizona: U.S. Dept. Commerce, Climatog-
raphy of the United States, no. 86—2, 95 p.

 

 

Weist, W. G., Jr., 1971, Geology and the ground-water system in the
Gila River Phreatophyte Project area, Graham County, Arizona:
U.S. Geol. Survey Prof. Paper 655-D, 22 p.

fins. GOVERNMENT PRINTING OFFICE 1977-777034/42

 

33°07’30”

 

 

PLATE 1

PROFESSIONAL PAPER 655—M

 

Xxx/w fl

. . $.39

. , ,. . . /m/I;sw \
v/ L

.
_
,J

v,

 

 

P.“ \/

of ma

‘z‘

\

 

V-._

\____L__

\

ARIZONA
75 MILES
100 KILOMETRES
INDEX MAP

22'30’

>M<ZMWH<DO
Lr

280203 2583on
III %I Cr

.\

 

' slanted figure in—

’

CARLOS
k.
\
\~’\.

on 2:55
I 1% <45 4.

N
A \
S

Upper number indicates

INDIAN

| L-\Q

!\.

RESERVATION

20 MILES

OF-WATER DATA WERE COLLECTED

middle number indicates dissolved-solids concentration

Dashed where approximately located, dotted where con-
in water, in milligrams per litre; lower number indicates well number and,

 

 

 

 

 

 

 

 

 

 

e
’ n f
S. d n 0 O
mmmn s.m nm 0 .w e
.1 wa.a1l .WS.O. n C
Sr. S 68 d e n
neo dhm. no n,o v .1 n
OprLOCd er.mfi1.m. S a
C geouo Vbh nm 1 t
Vfl e. .. O C
Wea m0 ftaeae $.10 u
mwwgesfl omemm e m d
.,h yumah SH te b V. n
a Q d o
ekpflsffi .md mmn m s c
ummlmow .mnan a u r 2
S 10 13m n e C
mabnwr nab...“ 1 p fim
f e 856 0.1.311 I p 8.1b
OIIOde Cleryn 6 [ca
nehcne.m ;.$mam.m W U fiet
.mkwwuumq wnrmcj _ Hmn
NOnfifica .awmawnm _ S en.“ .| a
SaoeMm ho nwa S T P0.8 3 m
d_dsrt0 mword T I Rde m E
R COD 6 d tgn I S m I x M
.auy Cu .st S E ea a . O
0 1h nh fu S I
0 Nfllbr tn asaeob O O aan J ,e \ LP
I 0 FVvaOWBuU SMflmsa P P wroloba pm. .H HA
R Hmmumwmm madhwm En m Rmna mm
E AflgofMHctawomba D0 Am.w.w m X
T BdmOtaSCtomMIM LpL L tm N T E
N dnfSthmafltO a e L I.nme .01D
1 naofiechsntwcw R Mr H Guth m N
LU¢a.$fivbtnlv.OQm,WV.dbS E V.m _ nmc T I
E _nen n of ledd B h N t Eocs A II
H m aroOfdnbmmdne Ut I m v...H.h..n1nrv T o:
r EMsmcMwm.-Q._.omab mum“ pTTmmws
Ume_am..em_t.n,.bf.m d m IOtsi a
F Itafi thasmcdflfo N Ae B S Lded G n
O WfamWwamTfimmmfi WWW m .m AUnSnnchmm s
0.1 St .IIdS.l>e C .T. E086
T LsdlmummSe feme A Ns N n QIc_m A
bf. OtdOfl.1I 1.1 I; .1 R d
N .AandWV. eePre 01b .0 e H se.gG
, o s a N D DP .0 Advfi
E bwLumpaEsmtaee o a y e c .11 _
M NmflLmuYhDLafidumm 1 Ed Et fl ITMOM W
I I S.1IW figmn.af T L LT. u HUS$W O
T AweAuww Lw omke _ C Lw Le c WBdm L
o
R meEunumu me .mp T E Mm mm .m 1es1 F
Pt I d.1;d V. d TRV ;
AY _ngC.mae.mFefseml C S Df Dw s ATlmr M
PE DemA .mkvuawmmtm A s y e u w; A
EV OHMRMMSNIwecmem T S Lim Lfl e Wthka E
DR moamap7flflbvmgcwe nNu mflo He W TOddW m
I
SU F T B C C W W S S
ES
TL
10
MA @o @mm v 1
C 2 N2
SI 0 82
G 1 501
D0 0 27
E
TL
IO 5
NE 2
UG ,0 DO
1
1 1

12’30”
33" 07.30”

 

 

Base from US. Geological Survey, unedited advanced prints

Geology by E. S. Davidson and W. G. Weist, Jr., 1964-67

fiUS. GOVERNMENT PRINTING OFFICE 1977-777034/42

COLLECTION SITES,

DATA—

AND CONTACTS BETWEEN THE BASIN FILL, THE TERRACE ALLUVIUM, AND THE FLOOD-PLAIN ALLUVIUM

9

S
N
m
T
C
E
S
S
S
O
R
C
A,
N
O
m
R
A
Y,
T
N
U
0
C
H
A
R
G
A.
E
R
A
T
C
E
J
O
R
P
E
T
Y
H
D...
O
T
A
E
R
H
P
R
E
V
I
R
A
m
G
E
H
T
F
O
N
m
T
A
C
O
L
G
N
I
W
O
H
S
D...
A
M

1962 and 4 NW, 1960

3

San Carlos, 3 NW and 3 NE

APPROXIMATE MEAN

SCALE 1:24 000

DECLI NATION, 1977

 

 

 

 

 

 

 

 

 

 

 

 

E
L
w

E

R

U

M

m

K

.E

0.? O
L

H 5.
J
a
h
E
1:

CONTOUR INTERVAL 40 FEET

DATUM IS MEAN SEA LEVEL

 

 

110°07’30“
nus. GOVERNMENT PRINTING OFFICE 1977—777034/42

PROFESSIONAL PAPER 655—M
d number

More than 9,000
Cross section an

 

 

EXPLANATION
DISSOLVED SOLIDS, IN MILLIGRAMS PER LITRE

 

 

 

12’30”

4 , 33a

9

k/./ 4/? I I, 1% :

ARIZONA

9

1 KILOMETRE

, GRAHAM COUNTY

 

SCALE 1:24 000
DATUM IS MEAN SEA LEVEL

A.

CONTOUR INTERVAL 40 FEET

 

 

 

 

 

 

GILA RIVER PHREATOPHYTE PROJECT AREA

S
H
S
0
P
E
D
L
m
V
U
L
L
A
E
H
T
F
O
E
N
0
Z
D
E
T
A
R
U
T
A
S
E
H
T
F
0
8m
m mm
DE
B.H
T
m
S
m
L
0
S
D
m
L
O
S
E
D
F
O
N
m
T
U
m
R
T
E
D
G
1N1.
W
O
H
S
m
A
M

IPKOZ mat:

APPROXIMATE MEAN
DECLINATION 1977

, 1960

Base from U.S. Geological Survey, unedited advanced prints

San Carlos, 3 NE, 1962 and 4 NW

I

33°12'30"§

UNITED STATES DEPARTMENT OF THE INTERIOR

GEOLOGICAL SURVEY

 

 

 

The Hydrologic History of the
San Carlos Reservoir, Arizona, 1929—71,
with Particular Reference to

Evapotranspiration and Sedimentation
By FRANK P. KIPPLE

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—N

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1977

UNITED STATES DEPARTMENT OF THE INTERIOR

CECIL D. ANDRUS, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Dirertor

Library of Congress Cataloging in Publication Data

Kipple, Frank P.
The hydrologic history of the San Carlos Reservoir, Arizona, 1929—71, with particular reference to evapotranspiration
and sedimentation.

(Gila River phreatophyte project) (Geological Survey Professional Paper 655-N)

Bibliography: p. N40.

1. Hydrology—~Arizona—~San Carlos Reservoir. 2. Evapotranspiration-—Arizona-—San Carlos Reservoir. 3. Sediments
(geology)-—Arizona--San Carlos Reservoir. I. Title: The hydrologic history of the San Carlos Reservoir... II. Series.
III. Series: United States Geological Survey Professional Paper 655-N.

QE75.P9 no. 655-N [GB705.A6] 557.3’033 [551.4’82’0979’54] 77-608085

 

For sale by the Superintendent of Documents, Us. Government Printing Office
Washington, DC. 20402
Stock Number 024-001-02976-6

CONTENTS

 

  
  
 
 
 
 
 
  

 

 

 

 

 

 

 

 

 

 

 

 

Page Page
Conversion factors ................................................................................ V Hydrology _____________________________________________________________________________________________ N14
List of symbols ........................................................... ._ V Water-budget equation 14
Abstract ................................................................... Surface flow ___________________________________________________________ __ 15
Introduction _______________________ 1 Gila River and San Carlos River inflow _________________________ 15
Purpose and scope ..................................... 1 Tributary inflow _____________________________________________________________________ 16
Acknowledgments .................................. Gila River outflow __ 18
History of the San Carlos Project ___________________ 2 Ground-water inflow 18
Definitions ........................................................ 2 Precipitation _____________________________________________________________________________________ 19
Reservoir sediment ................................................... Evaporation ..................................................................................... 19

Sediment deposition and capacity surveys ............................ 3 Water stage and surface-water storage at San Carlos
Surface-water storage losses from sediment Reservoir 22
accumulation __________________________ 5 Bank storage ._ ______ ,_ 24
Sediment distribution ____________________________________________________________________ 6 Water-budget analyses _________________________________________________________________ 27
The effect of phreatophytes on inflow channels ..... 7 Evapotranspiration _______________________________________________________________________________ 27
Sediment trap efficiency ............................................... 11 Development of surface-water storage-capacity ratings __ 31
Analysis of sediment data .......................................................... 13 Simulation of the sediment depositional process ______________ 31
Interpolation of elevation-capacity ratings between Interpretation of simulation results _______________________________________ 37
capacity surveys ........................................................................ 14 Procedure to develop ratings ______________________________________________________ 37
Summary and conclusions . .............. _ 39
References cited ____________________________________ 40

ILLUSTRATIONS
Page
FIGURE 1. Map showing reservoir boundary, location of gaging stations, and centerline of San Carlos Reservoir ............................ N3
26. Graphs showing:

2. Area curves for San Carlos Reservoir 4
3. Capacity curves for San Carlos Reservoir 5

 

4. Vertical distribution of the volume of sediment deposition with 5—ft (1.5 m)—elevation intervals at San Carlos

Reservoir for the periods 1928-47 and 1947—66 3
5. Bottom profiles of San Carlos Reservoir along principal longitudinal axis as determined by the five capacity

 

 

 

 

 

 

 

 

 

 

 

surveys 9
6—11. Photographs showing:
6. Aerial view looking downstream (west) on April 20, 1965, showing the sediment flats at the confluence of the
Gila and San Carlos Rivers within the San Carlos Reservoir 10
7. Aerial view looking upstream (north) on April 20, 1965, showing the San Carlos River on the left and the
Gila on the right 10
8. Aerial view of channel conditions of the Gila River upstream from the San Carlos River in 1935—7 years
after the completion of Coolidge Dam 10
9. Aerial view of conditions of a part of the Gila River channel above the San Carlos River in June 1962 ................. 10
10. View looking downstream (south) on August 18, 1965, at the Gila River flood plain from a point 2 mi (3.2 km)
upstream from the mouth of the San Carlos River 11
11. Flood-plain conditions after the area was inundated by the reservoir pool 11
12. Map showing extent of channel plugging in the Gila River on indicated dates 12
13. Graph of profiles along the longitudinal axis of San Carlos Reservoir showing changes in bed elevations and slope
within the reach of channel plugging 13
14. Photograph showing view of the Gila River looking north from a point 2.6 mi(4.2 km) upstream from the mouth
of the San Carlos River photographed in September 1964 13
15. Photograph showing downstream view on July 15, 1965, of a reach of the Gila River channel which had been
plugged in 1964 .. 13
16. Photograph showing aerial view of channel plugging in the Gila River on July 22, 1964 ...................................................... 13

III

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IV CONTENTS
Page
FIGURES 17—31. Graphs showing:
17. Relation between accumulated decrease in storage capacity and accumulated streamflow by capacity
survey periods ............. N15
18. Annual combined Gila River and San Carlos River inflows to San Carlos Reservoir for water years
1929—71 ........................................... 15
19. Frequency of occurrence of water-year inflow volumes, 1929—71 .......... ............... . 16
20. Mean monthly flow as a percent of mean annual for the Gila and San Carlos Rivers ________________________________________ 17
21. Relations between measured rainfall and tributary runoff for summer seasons 1964—71 from 72.6 mi2
(188 km?) of area tributary to San Carlos Reservoir . . ..... 18
22. Evaporation at San Carlos Reservoir for water years 1931—71 20
23. Five-year moving averages of pan evaporation at San Carlos Reservoir and the mean of pan evapora-
tion at Mesa, Roosevelt Lake, and at Tucson for water years 1931—71 21
24. Double-mass diagram of cumulative pan evaporation for San Carlos Reservoir and mean of Mesa,
Roosevelt Lake, and at Tucson for water years 1931—71 _. 21
25. Beginning-of-month lake stage of San Carlos Reservoir for water years 1929-71 __________________________________________________ 23
26. Number of days in which lake stage was within a particular elevation interval for water years 1929—71 .. 24
27. Percentage of time lake stage of San Carlos Reservoir equaled or exceeded a given elevation for
water years 1929—71 24
28. Percentage of time usable surface-water storage of San Carlos Reservoir equaled or exceeded a
given volume for water years 1929—71 25
29. Relations of cumulative AS B to cumulative AS R and cumulative A53 to cumulative AST for the
periods January through April 1965 and December 1967 through May 1968 __________________________________________________ 26
30. Relation between the computed change in bank-storage capacity and elevation at San Carlos Reservoir
for water years 1931—47 and 1948—71 ._ 27
31. Elevation-capacity relation for usable surface-water storage, usable bank storage, and total usable
storage in 1966, for San Carlos Reservoir __ .. 29
32. Sketch showing relative magnitude of inflow and outflow water-budget components ..................................................... 31
33—36. Graphs showing:
33. Annual evapotranspiration from exposed area of reservoir, computed by reservoir water budget __________________ 32
34. Annual depth of evapotranspiration from the exposed surface of San Carlos Reservoir for water
years 1931—71 33
35. Computed monthly evapotranspiration at San Carlos Reservoir for water years 1931—71 ________________________________ 33
36. Mean monthly evapotranspiration depths computed for 10-year periods 33
37. Sketch identifying terms used in simulation of suspended sediment distribution ............................................................ 35
38. Sketch showing distribution of suspended sediment, Ssi’ when distance, D A: was computed as greater than
the distance from inflow point (A) to the dam 36
39. Graph showing simulated sediment deposition, in percent, downstream from point of inflow ____________________________________ 37
40. Graphs showing comparison between the volumes of deposits measured and estimated for each incremental
reservoir, 1929-66, and volumes of estimated and measured deposits cumulative by 5-ft-elevation increments
upward from lowest part of reservoir, 1929—66 water years .. 38
TABLES
Page
TABLE 1. Results of surface-water storage-capacity surveys N5
2. Storage capacities, sediment deposition, and streamflow data 7
3. Volumes of sediment deposited by 5-ft-elevation intervals in San Carlos Reservoir during different periods ______________________ 7
4. Maximum and minimum volumes of water supplied by streamflow into San Carlos Reservoir for different time
durations 16
5. Mean monthly inflows for the Gila River, the San Carlos River, and for both rivers, 1930-71 .............................................. 16
6. Tributary inflow into the San Carlos Reservoir along a reach of the Gila River 17
7. Estimated tributary inflow into San Carlds Reservoir 17
8. Mean monthly discharge of the Gila River below Coolidge Dam, 1931—71 18
9. Annual (water year) precipitation at San Carlos Reservoir, 1931—71 19
10. Mean monthly precipitation and monthly extremes (1931—71) at San Carlos Reservoir .......................................................... 20

11.
12.
13.
14.
15.
16.

Evaporation at San Carlos Reservoir by water year, 1931—71 ..
Mean monthly evaporation and mean monthly pan coefficients for San Carlos Reservoir .............................. ..
Percent of time that available monthly surface-water storage was less than amount shown ................................................
Example of water budget used to determine change in bank storage
Results of procedures to determine bank storage capacity at San Carlos Reservoir

 

 

 

San Carlos Reservoir elevation-capacity tables of usable bank storage, usable surface-water storage, and

total usable storage

 

 
 
  

CONTENTS V

 

 

 

 

Page
TABLE 17. Summations of San Carlos Reservoir inflow-outflow components by water year, 1931-71 .................................................... N30
18. Water-budget summations 32
19. Total evapotranspiration for San Carlos Reservoir, by water year 33
20. Mean monthly evapotranspiration computed from 4 periods of 10 years each, and median monthly
evapotranspiration for 41 years * 34
21. Chart of notation used to identify time and inflow location of computed proportional sediment weights, Syi j .............. 34
22. Optimum values of variables from simulation of the sediment depositional procedure .......................................... ’ __________________ 3 6
23. Volumes of sediment deposits measured (S Mk) and computed (Sek) for 1937-47, and the SMk/Se;a ratios ..................... . 38

24. Estimated change in surface-water storage capacity, Z, from start of period 3 to 1942, by 5-ft—elevation increments 39
25. Computed capacity ratings of surface-water storage used for 1938 through 1942 39
26. Comparison of segments of surface-water storage capacity ratings made by curve fitting and by computer .................. 39

 

 

CONVERSION FACTORS

English Multiply by Metric (SI)
acre-feet (acre-ft) 1.233 X 10'3 cubic hectometers (hm3)
miles (mi) 1.609 kilometers (km)
feet (ft) .3048 meters (m)
cubic feet per second (fth) 2.832 x 10‘2 cubic meters per second (m3/s)
acres 4.047 X 10'3 square kilometers (km?)
pounds per cubic feet (lb/fta) 16 kilograms per cubic meter (kg/m3)
acre-feet per square mile (acre-ft/miz) 0.476 x 10'3 cubic hectometers per square kilometer (hma/kmz)
gallons per day per square foot [(gal/day)/ft2] 4.074 X 10-2 cubic meters per day per square meter [(m3/day)/m2]
inches (in.) 25.4 millimeters (mm)
parts per million (ppm) 1.001 milligrams per liter (mg/ 1)

LIST OF SYMBOLS

 

I G Gila River inflow at Calva Si Weight of sediment trapped by reservoir
13 San Carlos River inflow at Peridot E Sediment trap efficiency of a reservoir
I T Tributary inflow into the San Carlos Reservoir Cw Mean Winter sediment concentration
0 G Gila River outflow below Coolidge Dam Cs Mean summer sediment concentration
I GW Ground-water inflow Qw Winter streamflow summations
I P Precipitation input over the lake surface Qs Summer streamflow summations
0 E Evaporation from the lake surface :9 gatio of :ummer to winter colncentratlilonsf d
. . ompute va ue proportiona to weig t 0 se iment
y . . .
0 ET E::§:1t::ir;sp1ration from the exposed surface of the deposited in a reservoir
. I l t '

ASR Change in surface-water storage j £:::?;:atf s orage reserv01r
AS B Change in bank storage a The ratio of the weight of larger sediment particles
q Ground-water flow to the weight of total sediment
K H draulic conductivit D Distance estimated by simulation procedure from

y y A . . . . .
m Thickness of aquifer location of river dlscharge into the reserv01r pool
u Ground-water slope to the point where deposition is complete
w Aquifer width DB Distance from any point along DA to point where
P Rainfall distribution of sediment is complete
Q Stream discharge x Exponent of suspended sediment distribution
AS T Change in total reservoir storage S s B Computed amount proportional to weight of sediment
S Weight of sediment discharge deposited between point of inflow and any other
C Mean sediment concentration downstream point

‘ See “Water Resources Data for Arizona, Part 2, Water Quality Records, 1973" for conversion factors when sediment concentration exceeds 8,000 ppm (factor varies depending on
specific gravity of sediment and density of water).

VI

Ssk

CONTE NTS

Estimated distance from location of river discharge
into incremental reservoir i to the point where
deposition is complete

Incremental reservoir of deposition

Distance from upstream point of incremental reser-
voir k, in which simulated deposition is occurring,
to the point where deposition is complete

A computed quantity which is proportional to total
sediment weight entering an incremental reservoir
from the stream

A computed quantity which is proportional to weight
of suspended sediment inflow

A computed quantity which is proportional to weight
of suspended sediment deposited in reservoir k

 

Sdk,j

A computed quantity which is proportional to the
suspended sediment weight deposited in reservoir k
during water year j

A computed quantity which is proportional to the
sediment weight deposited in an incremental reser-
voir during a period

The sum of all 83k

The volume of sediment measured

A proportionality constant between sediment mea-
sured and sediment estimated

Estimate of the absolute sediment volume for incre-
mental reservoir k

The estimated decrease in storage for an incremental
reservoir k from the start of a period to water year j

GILA RIVER PHREATOPHYTE PROJECT

THE HYDROLOGIC HISTORY OF THE SAN CARLOS RESERVOIR,
ARIZONA, 1929-71, WITH PARTICULAR REFERENCE TO
EVAPOTRANSPIRATION AND SEDIMENTATION

By FRANK P. KIPPLE

ABSTRACT

Reservoir data records were used in an investigation of evapo-
transpiration from the land area of San Carlos Reservoir and
evaporation from the water-surface area. A water-budget analysis
indicates that the evapotranspiration loss was 11.3 percent and
the evaporation loss was 10.5 percent of the total outflow from the
reservoir during 1931-71.

The water-budget computations were used to develop ratings
relating lake stage to usable bank storage. The rating developed
for the 1948-71 period indicates that usable bank storage is
approximately 159,000 acre-ft (196 hma), or about 14 percent of
total usable storage capacity, if the reservoir is filled to the
spillway level of 2,511 ft (765 m).

A procedure was developed to simulate sediment deposition in
the reservoir. The procedure was used to estimate the change in
storage capacity between five reservoir capacity surveys made
during the period 1914-66.

INTRODUCTION
PURPOSE AND SCOPE

Once a reservoir is put into operation, a number of
progressive changes are produced which affect the
hydrology of the reservoir. Prior to inundation, water
vaporization from the reservoir area is from plant
transpiration and from soils and off-channel pond-
ing. During inundation vaporization is evaporation
from the ponded water surface. The soils and topo-
graphy are changed by the deposition of sediment
and sometimes by bank erosion. The water table
adjacent to the reservoir rises, and the bank storage
of water is increased. Vegetation on exposed parts of
the reservoir may be altered because of changes in
soils and water availability. These changes are
particularly significant for reservoirs where the
streams convey large quantities of sediment and
where fluctuations of the reservoir water level and
the water surface areas are large. The records of
inflow, outflow, surface-water storage, and sediment
deposition provide the data for evaluating some of
the changes for the San Carlos Reservoir.

 

An investigation of these changes, and of the
reservoir hydrology in general, was made to evaluate
reservoir evapotranspiration (ET) and the change in
E T from 1929 to 1971. The investigation was made as
part of the Gila River Phreatophyte Project, a study
of the hydrologic effect of phreatophyte control by
the U.S. Geological Survey (Culler and others, 1970).
The evaluation of ET is made by use of a water-
budget equation in which ET is the residual in the
equation.

Secondary objectives included investigations of
reservoir sediment deposition, lake evaporation, and
reservoir bank storage. These investigations were
essential prior to compilation of the water budget.

Data sources for this report include five surveys of
reservoir capacity which provide a history of capac-
ity change and sediment accumulation. Investiga-
tions of tributary runoff, precipitation, evapotran-
spiration, and lake evaporation were made as part of
the Gila River Phreatophyte Project and furnish
information for this report. Other sources of data are
U.S. Geological Survey surface-water records (issued
annually), Gila River Water Commissioner reports
(issued annually), log books of precipitation and pan
evaporation at Coolidge Dam, and climatic data
published by the National Weather Service (issued
annually).

ACKNOWLEDGMENTS

Much of the planning and data interpretation for
this report was contributed by R. C. Culler, project
chief of the Gila River Phreatophyte Project. Prepar-
ation of the report was under the direction of R. L.
Hanson. Assistance and cooperation were received
from personnel of the Arizona District office, U.S.
Geological Survey, and of the San Carlos Project of
the U.S. Bureau of Indian Affairs, supervised by M.
D. Young, general engineer. Computations of lake

N1

N2

evaporation by energy-budget and mass-transfer
methods were principally by J. Stuart Meyers, US.
Geological Survey.

HISTORY OF THE SAN CARLOS PROJECT

The San Carlos Project was established to provide
irrigation water to the Middle Gila District, Gila
River Basin, Ariz. The district was defined by Davis
(1897, p. 17) as “that portion from the mouth of Salt
River to The Buttes above Florence and including the
Pima Indian Reservation and the great Casa Grande
Valley.” Diversions of water from the upstream
reaches of the Gila River by farmers during the
period 1870—86 imperiled the water rights of the
Indians on the Pima Reservation. An investigation
was made by the Geological Survey to examine the
possibility of providing a firm water supply to the
Indians as reported by Davis (1897, p. ~ 71). The
construction of a dam on the Gila River at The Buttes
14 mi (23 km) east of Florence was recommended.
Numerous other feasibility studies were made during
the next 15 years, culminating in a report by the US.
Army Corps of Engineers (1914) recommending a
dam on the Gila River.

On June 7, 1924, Congress approved legislation
that authorized the Secretary of the Interior through
the Indian Service to construct a dam at the San
Carlos site as part of the San Carlos Project. The
construction of Coolidge Dam was started in J anu-
ary 1927 and completed in October 1928. Water im-
poundment began on November 15, 1928. The area
within the boundary of the reservoir is administered,
and the facilities at Coolidge Dam are operated, by
the San Carlos Project, an agency of the Bureau of
Indian Affairs.

Coolidge Dam is located in sec. 17, T. 3 S., R. 18 E.,
Gila County, Ariz., in the San Carlos Indian Reser-
vation (fig. 1). The dam is a multidomed structure
having a length, including two spillways, of 850 ft
(259 m). Each of the three domes has a span of 180 ft
(55 m) and a base thickness of 28 ft (8.5 m). The thick-
ness decreases to 4 ft (1.2 m) at the top. The dam rises
203 ft (62 m) to the spillway at elevation 2,511 ft (765
m) above mean sea level and approximately an addi-
tional 25 ft (7.6 m) to the highway on top of the dam.
A spillway is located on each side of the dam. Each
spillway has three gates 50 ft (15 m) wide and 12 ft
(3.7 m) high. Maximum storage capacity of 1,267,000
acre-ft (1,560 km3) at elevation 2,523 ft (769 m) is
reached when the gates are raised. The gates are now
inoperative in the lowered position. The maximum
safe release from spillways and outlets is 122,000
ft3/s (3,455 m3/s). The sill of the lowest outlet gate is
at elevation 2,382.63 ft (726 m), providing an oper-
ating range, outlet to spillway, of 128.37 ft (39.13 m).

The principal purpose of the reservoir is to store

 

GILA RIVER PHREATOPHYTE PROJECT

water for irrigation of 100,000 acres (405 km2) of land
within the San Carlos Project. Fifty thousand acres
(202 kmz) are Indian lands Within the Gila River
Reservation, and 50,000 acres (202 km?) are privately
owned lands in the Florence—Casa Grande Valley.
Water released from Coolidge Dam is diverted from
the Gila River channel at the Ashurst-Hayden Diver-
sion Dam 68 mi (109 km) downstream. A power plant
at Coolidge Dam contains two generators having a
combined output of 10,000 kilovolt—amperes. Power
generation is subordinate to irrigation requirements
and is stopped when irrigation demands are cur-
tailed. The reservoir is also used for recreation, and
the recreational facilities are operated under lease by
the San Carlos Apache Indian Tribe. Flood protec-
tion provided by the reservoir is an incidental
benefit.

Water in the Gilla River was adjudicated by a court
decree entered in 1935 (US. vs. Gila Valley Irrig.
Dist. et al., 1935). Briefly, the decree divides the
water between the upstream users in the Safford and
Duncan Valleys, the Gila Valley Irrigation District,
and the downstream users of the San Carlos Irriga-
tion Project, on the basis of priority of appropriation.
In addition to priority rights, upstream users are also
entitled to apportioned rights, which are dependent
on the amount of water stored in San Carlos Reser-
voir. As defined in the decree, the rights are deter-
mined as follows: On January 1 of each year, or as
soon thereafter as there is water stored in the San
Carlos Reservoir, which is available for release for
use on lands of the San Carlos Project, the Gila Water
Commissioner, who is appointed by the court to
enforce the decree, shall apportion for the ensuing
irrigation year to irrigated lands above the San
Carlos Reservoir from the natural flow of the Gila
River an amount of water equal to that stored in the
San Carlos Reservoir less losses. It is also provided
that if and when at any time, or from time to time,
during the year storage in the reservoir shall be
increased and made available to downstream users,
the Commissioner shall make further and additional
apportionments to upstream users which shall be
equivalent in amount to the newly available stored
water supply.

DEFINITIONS

The following are definitions of terms as used
throughout this report:

Water year .......................... October 1 to September 30

Surface-water storage The above-ground volume of a reservoir
capacity available to store water
Dead storage capacity, The above-ground volume of a reservoir
surface water below the invert of the lowest reser-
voir outlet, which cannot be evacuated
by gravity

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

 

    
  
 
   
 
    

Coolidge
Dam‘ San Carlos Reservoir
Gila River below

Coolidge Dam

110°30'

Base from U.Si Geological Survey
Mesa 1:250,000, 1954

EXPLANATION

Reservoir boundary

   
        

Centerline of reservoir

ARIZONA
A
Guging station

RESERVOIR
AREA

0
32X\\‘\ Tucson “I.
INDEX MAP ~~ L |
0 50 100 MILES

Irfi-‘w—vJ-v—B‘
$0100 KILOMETERS

 
   

‘4

  
 

\ Gila River

Gila River at Calva

110°15'

0 1 2 3 4 5 MILES

l—1_l_l"_l_l—l'—Ll—"—;—_l
o 1 2 3 4 5 KlLOMETERS

CONTOUR INTERVAL 200 FEET

WITH SUPPLEMENTARY CONTOURS AT 100—FOOTlNTERVALS

DATUM IS MEAN SEA LEVEL

FIGURE 1.—-Map showing reservoir boundary, location of gaging stations, and centerline of San Carlos Reservoir.

The difference between surface-water
storage capacity and dead storage ca-
pacity, defined as the volume available
for release below the stage of maxi—
mum controllable level

The below-ground volume in the banks of
a reservoir available for storage which
can be evacuated by gravity

Usable surface-water
storage capacity

Usable bank-storage
capacity

RESERVOIR SEDIMENT

SEDIMENT DEPOSITION AND CAPACITY SURVEYS

Sediment deposition in a reservoir, and the result-
ing reduction in water-storage capacity, affects
water supply and water management, installations
within the reservoir, and recreational activities.
Management can, in turn, influence the sediment
distribution and sediment compaction within the

 

reservoir and the volume of sediment which passes
through the reservoir by (1) regulating the stage, rate
of water release, and frequency of sediment wetting
and drying, (2) vegetative clearing, and (3) channel
dredging. An inventory of sediment deposition can
be obtained by use of data from reservoir-capacity
surveys. In addition, these surveys provide the data
used to establish the elevation-capacity and eleva-
tion-area ratings, which are needed for water
management.

The volume of sediment deposited in the San
Carlos Reservoir was calculated from the results of
five surveys. The first survey was made during 1914
and 1915 for the Indian Irrigation Service to deter-
mine potential storage capacity of the proposed
reservoir. The second and third surveys were com-

N4

pleted in 1935 and 1937 by the Soil Conservation
Service, the third at the request of the Gila Water
Commissioner. A fourth survey was made in 1947 by
the Corps of Engineers to assess changes in the
capacity of the reservoir, mainly resulting from
above-normal inflow in 1941—42 (Thorp and Brown,
1951). A map of the reservoir, scale 1:7,200, was
produced by photogrammetric methods from the
1947 survey. Changes during the period 1947—66
were assessed by a fifth survey made in 1966 by the
US. Geological Survey.

The survey of 1966 was made primarily to provide
better water-surface area and storage-capacity data
for the water-budget analysis of evapotranspiration.
An above-normal lake level during the summer of
1966 provided an opportunity to obtain an economi-
cal survey of the reservoir. Control points at-the ends

 

GILA RIVER PHREATOPHYTE PROJECT

of 51 range lines were established near the water’s

‘ edge, and a recording fathometer was used to obtain

a continuous record of the ground profile along each
range line. The shorelines shown on aerial photog-
raphy taken during 1966—67 were used to check
some of the topographic data. Topographic changes
within the study area of the Gila River Phreatophyte
Project above the maximum 1966 pool level were
obtained from cross-valley profiles repetitively sur-
veyed (Burkham, 1972). Elevation data were plotted
on copies of 1947 reservoir topographic maps of the
US. Army Corps of Engineers and were used to
locate 5-ft (1.5 m) contours for 1966.

Area curves from the five capacity surveys are con-
tained in figure 2. Reservoir capacities between con-
secutive 5—ft (1.5 m) contours were computed using
the elevation-area data of each survey. The curves of

WATERASURFACE AREA, IN THOUSANDS OF SQUARE HECTOMETERS

 

O 1 2 3 4 5 6 7 8
I I I I I I I I I I I | I I I | | __ 770
Top of fully raised gates, 2, 523 ft ———————————— _,—.-‘—
2520 — —
2510 _ Top of spillway crest, 2, 511 ft ——~ _':._,L_ _ _ _______ 765
2500 “ _
EXPLANATION
,.-;'/ — 760
2490 — 1966 p I/ _
------------ 1947 ..> /
2480 _ _.._.._.r_.._1937 u)" _
— 755
————— 1935
2470 — 1915 / ~

2460

2450

2440

 

2430

2420

2410

ELEVATION, IN FEET ABOVE MEAN SEA LEVEL

2400

 

  
 

,_ _ ___-.___..‘

 

Elevation at top of dead storage pool, 2, 382.63 ft

750

—— 745

— 740

735

ELEVATION, IN METERS ABOVE MEAN SEA LEVEL

~— 730

725

 

i | I I I I I I I 720

 

2360 | I I I I I I I I I
0 5 6 7 8 9 10

11 12 13 14 15

WATER-SURFACE AREA, IN THOUSANDS OF ACRES

FIGURE 2.—Area curves for San Carlos Reservoir.

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—7]

figure 3 show the elevation-capacity relations for all
surveys. Water-surface areas and surface-water
storage capacities for the surveys are listed by 5-ft
(1.5 m)—elevation increments in table 1. Much of table
1 data is from Thorp and Brown (1951, table 1).

The earlier capacity surveys were not as detailed
as later surveys because the needs were different.
The 1914—15 reservoir survey was designed to de-
scribe agricultural lands and land use on the Gila
and San Carlos River flood plains. The topography
outside the flood plain area was included in this
survey but was with less detail. More attention was
given to defining flood—plain topography in the 1935
and 1937 surveys. Not until the 1947 survey, how-
ever, was the upland topography well defined. Ca—
pacity ratings for the first three surveys were ad—
justed to the more accurate 1947 survey by Thorp and
Brown (1951, p. 9, 10) and are shown in table 1. The
errors remaining after adjustments are significant
in the analyses of this report. As an example of the
errors remaining in the ratings, it is seen in table 1
that capacities computed for 5-ft-elevation incre-
ments in 1937 were greater than for corresponding 5-
ft increments in 1935, in the range of 2,435 to 2,495 ft
(742—760 m). Scour or compaction might have pro-

 

N5

duced some of the increases, however, the increases
extend about 25 ft (7.6 m) above the maximum water
stage recorded prior to the 1937 surveys. The maxi-
mum reservoir elevation through 1937 was 2,471.56ft
(753 In), April 5, 1932. Another example of error in
the first two surveys is indicated by no measured
capacity change between surveys above certain
stages, although considerable inflow occurred when
the reservoir water level exceeded those stages. The
1935 capacity table was identical to the 1914-15 table
above 2,435 ft (742 m) elevation, yet about 668,000
acre-ft (824 hm3) of inflow occurred above this
elevation during 1928—35.

SURFACE-WATER STORAGE LOSSES FROM
SEDIMENT ACCUMULATION

The maximum surface-water storage capacity of
the San Carlos Reservoir in November 1928 was
1,266,837 acre-ft (1,562 hm3) at 2,523 ft (769 m). By
1966, sediment deposition had reduced the capacity
by 96,719 acre-ft (119 hm3) to 1,170,118 acre-ft (1,443
hm3), a 7.6 percent loss. The loss in usable surface-
water storage capacity was 72,476 acre-ft (89 hm3)for
the same period.

These storage losses indicating that the total

CAPACITY, IN CUBIC HECTOMETERS

0 100 200 300 400 500 600 700
1

253° \1 ‘1‘1‘ 1l i1 1'1}
2520—
2510—-
2500 --
2490

2480

ELEVATION, IN FEET ABOVE MEAN SEA LEVEL

 

1 1 1 1 1

I .
0 100 200 300 400 500 600

800

Elevation of maximum capacity, 2, 523 ft

, 7 , Elevation at top of dead storage pool, 2, 382.63 ft

900 1000 1100 1200 1300 1400 1500 1600

1 1 1 1 1 1
1 1 1 1 1 I

  

EXPLANATION

................ 1966

1947 a
— _,_.,_ 1937 _
———— 1935 ‘740

1915 A

 

_— 735

m 730

ELEVATION, lN METERS ABOVE MEAN SEA LEVEL

T— 725

 

F 720
1300

1 1 1 1 1 I
700 800 900 1000 1100 1200

CAPACITY, IN THOUSANDS OF ACRE—FEET

FIGURE 3.—Capacity curves for San Carlos Reservoir.

N6

GILA RIVER PHREATOPHYTE PROJECT

TABLE 1.—Results of surface-water storage-capacity surveys

 

 

 

 

 

    

1914—15 survey (adjusted)1 1935 survey (adjusted) 1937 survey (adjusted) 1947 survey 1966 survey
Elevation Capacity Capacity Capacity Capacity Capacity
above to next to next to next to next to next
mean sea lower Cumulative lower Cumulative lower Cumulative lower Cumulative lower Cumulative
level Area contour capacity Area contour capacity Area contour capacity Area contour capacity Area contour capacity
(ft) (acres) (acre-ft) (acre-ft) (acres) (acre-ft) (acre-ft) (acres) (acreft) (acre-ft) (acres) (acre-ft) (acre-ft) (acres) (acre-ft) (acre-ft)
2308 ........ Orignal bottom at dam 0 0 0 0 0 0 Original bottom at dam
2365 489 6,060 6,060 66 5 5 55 4 4 0 0 0 0 O 0
2370 3,307 9,367 428 1,103 1,108 363 932 936 6 1 1 0 0 0
2375 4,946 14,323 768 2,949 4,057 622 2,434 3,370 227 450 450 13 22 22
2380 6,523 20,846 1,127 4,709 8,766 1,021 4,069 7,439 524 1,827 2,277 174 394 414
2385 8,162 29,008 1,362 6,213 14,979 1,264 5,704 13,143 803 3,293 5,570 291 1,150 1,564
2390 9,396 38,404 1,624 7,456 22,435 1,570 7,071 20,215 1,182 4,932 10,502 599 2,180 3,744
2395 10,626 49,030 1,824 8,615 31,050 1,835 8,504 28,719 1,454 6,578 17,080 968 3,880 7,624
2400 12,347 61,377 2,253 10,174 41,224 2,214 10,108 38,827 2,083 8,796 25,876 1,508 6,140 13,764
2405 14,522 75,899 2,651 12,247 53,471 2,717 12,306 51,133 2,493 11,425 37,301 2,065 8,897 22,661
2410 16,434 92,333 3,020 14,168 67,639 3,067 14,451 65,584 2,902 13,475 50,776 2,562 11,545 34,206
2415 18,315 110,648 3,481 16,239 83,878 3,430 16,234 81,819 3,315 15,531 66,307 3,061 14,039 48,245
2420 20,559 131,207 3,922 18,497 102,375 3,797 18,060 99,879 3,741 17,630 83,937 3,319 15,945 64,190
2425 23,984 155,191 4,537 21,129 123,504 4,455 20,609 120,487 4,391 20,309 104,246 3,859 17,929 82,119
2430 25,547 180,738 5,197 24,317 147,821 5,084 23,831 144,318 5,099 23,703 127,949 4,538 20,970 103,089
2435 28,303 209,041 5,649 27,108 174,929 5,648 26,818 171,136 5,564 26,650 154,599 5,095 24,069 127,158
2440 29,735 238,776 6,250 29,735 204,664 6,296 29,846 200,982 6,143 29,256 183,853 5,839 27,314 154,472
2445 32,727 271,503 6,845 32,727 237,391 6,896 32,969 233,951 6,731 32,174 216,029 6,372 30,519 184,991
2450 35,769 307,272 7,467 35,769 273,160 7,531 36,057 270,008 7,381 35,268 251,297 7,123 33,721 218,712
2455 38,809 346,081 8,060 38,809 31 1,969 8,145 39,180 309,188 7,971 38,371 289,669 7,751 37,174 255,886
2460 42,238 388,319 8,841 42,238 354,207 8,884 42,560 351,748 8,900 42,157 331,826 8,654 40,992 296,878
2465 46,529 434,848 9,778 46,529 400,736 9,808 46,712 398,460 9,617 46,282 378,108 9,668 45,783 342,661
2470 50,688 485,536 10,501 50,688 451,424 10,516 50,801 449,261 10,522 50,332 428,439 10,460 50,308 392,969
2475 54,212 539,748 11,187 54,212 505,636 1 1,183 54,238 503,498 11,133 54,131 482,570 11,078 53,839 446,808
2480 57,677 597,425 1 1,887 57,677 563,313 11,905 57,709 561,207 1 1,832 57,405 539,975 11,690 56,914 503,722
2485 61,413 658,838 12,682 61,413 624,726 12,685 61,466 622,673 12,689 61,291 601,266 12,524 60,525 564,247
2490 65,485 724,323 13,516 65,485 690,212 13,519 65,500 688,174 13,528 65,533 666,799 13,224 64,363 628,610
2495 69,582 793,905 14,320 69,582 759,793 14,320 69,589 757,763 14,320 69,612 736,411 14,212 68,576 697,186
2500 73,914 867,819 15,250 73,914 833,708 15,250 73,914 831,677 15,250 73,914 810,325 15,250 73,914 771,100
2505 78,407 946,226 16,116 78,407 912,114 16,116 78,407 910,084 16,1 16 78,408 888,732 16,1 16 78,407 849,507
2510 82,849 1,029,075 17,027 82,849 994,963 17 ,027 82,849 992,933 17,027 82,849 971,581 17,027 82,849 932,356
2515 87,536 1,116,611 17,991 87,536 1,082,499 17,991 87,536 1,080,468 17,991 87,536 1,059,116 17,991 87,536 1,019,892
2520 92,403 1,209,014 18,974 92,403 1,174,902 18,974 92,403 1,172,872 18,974 92,403 1,151,520 18,974 92,403 1,1 12,295
2525 . 97,389 1,306,403 19,985 97,389 1,272,291 19,985 97,389 1,270,260 19,985 97,389 1,248,908 19,985 97,389 1,209,684
1,046,203 .. 1,012,091 ,,,,, 1,010,061 . 988,709 .. 949,484
, 1,266,857 . 1,232,725 ,,,,, 1,230,695 , . 1,209,343 . 1,170,118

 

 

1 Use of these tables began in 1928 with closure of dam.

'1 Area on area-elevation curves interpolated from 2425 and 2440 contours.
3 Spillway crest.

4 Spillway crest with gates fully raised.

volume of sediment deposited was 96,719 acre-ft (119
hm3) also indicate that the mean annual volume for
the period November 1928 to August 1966 was 2,553
acre-ft (3.1 hm3), which is equivalent to 0.20 percent
of the original surface-water storage capacity. Table
2 includes storage capacity data, sediment deposi-
tion amounts and mean rates of deposition, and a
comparison of sediment volume deposited to stream-
flow volume for all periods between surveys.

SEDIMENT DISTRIBUTION

The data of table 1 were used to calculate the
volumes of sediment deposition for 5-ft (1.5 m)—eleva-
tion increments for the periods 1928-47, 1947—66, and
1928—66. (The capacity tables used in 1928 were
obtained from the 1914-15 survey.) The computed
volumes are given in table 3 and are illustrated in the
vertical distribution graph of figure 4. Changes in
elevation along the centerline of the reservoir result-
ing from sediment deposition are evident from the
profiles shown in figure 5 for all surveys. The
longitudinal centerline, as established on maps of
the 1947 survey, was changed to more nearly center it

with respect to the flood plain at a distance of 42,000
to 58,000 ft (12,800 to 17,700 m) upstream from the
dam. The centerline, as revised, is shown in figure 1.
Centerline distances from the dam to intercepts of
centerline and contour lines were scaled from maps
of all surveys except the 1937 survey, for which no
maps were available. Centerline distances for the
1937 survey were based on the 1937 capacity-survey
data and the 1935 centerline distances.

Longitudinal slope of the reservoir bottom below
the 2,380-ft (725 m) contour has decreased from
0.00246 to 0.00057 during 1929—66. Above the 2,380—ft
(725 m) contour the slope has decreased from 0.00167
to 0.00138. Over the entire 22—mi (35.4 km) reservoir
length, the mean slope was 0.00188 in 1928, 0.00140
in 1935 and 1937, 0.00135 in 1947, and 0.00131 in
1966. The longitudinal slope of the San Carlos arm of
the reservoir is about 0.00275. Changes in slope due
to channel plugging are discussed in the next
section.

A large sediment accumulation occurred in the
lower parts of the reservoir, as indicated in table 3
and figures 4 and 5, which corresponds to the most

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71 N7

TABLE 2.—-Storage capacities, sediment deposition, and streamflow data

A. Storage capacities

 

 

 

 

 

 

Year of survey 1928‘ 1935 1937 1947 1966
Surface-water storage capacity in acre-ft, at 2,523 ft (gates fully raised) 1,266,837 1,232,725 1,230,695 1,209,343 1,170,118
Surface-water storage capacity in acre-ft, at 2,511 ft (spillway crest) .. 1,046,203 1,012,091 1,010,061 988,709 949,484
Usable surface-water storage capacity in acre-ft, at 2,511 ft .................. 1,021,060 1,000,916 999,231 984,875 948,584
Dead storage capacity in acre-ft, at 2,382.63 ft 25,143 11,175 10,830 3,834 900
Elevation of zero surface storage at time of capacity survey, in feet ................................................ 2,309 2 2,354 2 2,354 2,370 2,374

 

B. Sediment deposition

 

 

 

 

 

  

 

 

 

 

Period 1928—351 1935-37 1937-47 1947—66 19284561

Number of years in period 6.28 1.91 1000 19.70 37.89
Volume of sediment deposited per period, in acre-ft ............................................................................. 34,112 2,030 21,352 39,225 96,719
Mean annual deposition per period, in acre-ft ..... 5,431 1,063 2,135 1,991 2.553
Sediment deposition per period in percent of original surface-water storage capacity

at 2,523 ft . 2.69 .16 1.69 3.09 7.63
Mean annual sediment deposited per period as percent of original surface-water

storage capacity .43 .08 .17 .16 .20
Sediment volume deposited in dead storage, in acre-ft 345 6,996 2,934 24,243
Dead storage loss by sediment deposition per period, in percent 0

capacity 1.37 27.8 11.67 96.42
Stream inflow per period in acrevft3 .. 345,275 2,541,463 3,579,217 8,188,940
Ratio of sediment volume deposited to volume of stream inflow per period ..................... .0059 .0084 .0110 .0118

 

‘Beginning November 15, 1928.
zApproximate. '
“Inflow of Gila River and San Carlos River.

TABLE 3.—Volumes of sediment deposited by 5-ft-elevation inter-
vals in San Carlos Reservoir during different periods

 

. . Volume to next lower contour, in acre-ft
Reservoir elevation

 

 

 

 

(mean sea level) 1928—47 1947—66 1928-66
0 0 0
6,060 0 6,060
3,306 1 3,307
4,506 428 4,934
4,696 1,435 6,131
4,369 2,143 7,012
4,464 2,752 7,216
4,048 2,698 6,746
3,551 2,656 6,207
3,097 2,523 5,625
2,959 1,930 4,889
2,734 1,492 4,276
2,929 1,685 4,614
3,675 2,380 6,055
1 ,344 2,733 4,577
1,653 2,581 4,234
479 1,942 2,421
553 1,655 2,203
501 1,547 2,043
438 1,197 1,635
81 1,165 1,246
247 499 746
356 24 330
81 292 373
272 491 763
122 766 388
- 48 1,170 1,122
- 30 1,036 1,006
o 0 0
0 0 0
0 0 0
0 0 0
0 0 o
0 o 0
Total ...................... 57,493 39,226 96,719

 

common range of stage of the reservoir pool. As a
consequence, the surface-water dead storage capac-
ity of the reservoir was reduced 96 percent from
25,143 acre-ft (31 hm3) to 900 acre-ft (1.1 hm3) from
November 1928 to August 1966 (see table 2). The
distribution of sediment is partly regulated by the
physical features of the reservoir, such as longitu-

 

dinal slope, cross-sectional dimensions, shape, vege-
tation on exposed ground surface, and inflow chan-
nel geometry. The location of deposition is also regu-
lated by concentration and particle size of sediment
inflow, by rate of streamflow, and by the stage and
volume of water in storage.

THE EFFECT OF PHREATOPHYTES ON INFLOW CHANNELS

Phreatophytes can significantly reduce the con-
veyance of reservoir inflow channels, particularly if
the stage of the water in the reservoir fluctuates
widely. The flat fertile plain at the upstream end of
the reservoir pool has a shallow water table and is
periodically inundated, creating an ideal environ-
ment for phreatophytes such as saltcedar. Inunda-
tion may kill the plants, but the prolific seed pro-
duction and rapid growth of saltcedar quickly re-
creates a dense thicket. When these sediment flats
are exposed for extended periods of time, the salt-
cedar narrows the inflow channel by encroachment
and can eliminate a continuous channel. During the
period 1962—65, the inflow channel of the Gila River
into the San Carlos Reservoir was blocked by a
combination of conditions including encroachment
by phreatophytes on the sediment, a reduction in
channel gradient, and plugging by the deposition of
floating debris.

Figures 6 and 7 show the sediment flats formed by
deposition in the area above and below the con-
fluence of the Gila and San Carlos Rivers. The
upstream end of the reservoir pool was located in the
area shown in these photographs during much of the
period 1945—65. ‘

During the period 1935—62 the location and aline-
ment of the Gila River channel did not change signif-

N8

GILA RIVER PHREATOPHYTE PROJECT

VOLUME OF SEDIMENT DEPOSITED WITHIN 1.52-METER»
ELEVATION INCFIEMENTS, IN CUBIC HECTOMETERS

0 1 2 3 4

Elevation of maximum storage: SpiIlway
gates fully raised 2,523 ft

Elevation of spillway crest 2, 511 ft

ELEVATION, IN FEET ABOVE MEAN SEA LEVEL

LE, I I I

0 1000 2000 3000

 

4000

6 7 8 9

770

765
EXPLANATION

760
1928~47

I:

1947—66
755

750
745
740

735

ELEVATION, IN METERS ABOVE MEAN SEA LEVEL

Top of dead storage
pool 2, 38263

730

725

721

703

l I I J
5000 6000 7000 8000

VOLUME OF SEDIMENT DEPOSITED WITHIN 5-FT-ELEVATION INCREMENTS, IN ACRE-FEET

FIGURE 4,—Vertical distribution of the volume of sediment deposited within 5-ft (1.5 m)-elevation intervals at San Carlos
Reservoir for the periods 1928—47 and 1947-66.

icantly as indicated by comparing figure 8 with
figure 9. The width of the channel was appreciably
reduced and natural levees had formed during this
period, however. Figure 10 shows the levees along
the banks of the inflow channel of the Gila River
which had developed by August 1965. The river-
banks with abundant water supply and extensive

exposure to sunlight are an excellent environment
for saltcedar. This vigorous growth encroaches on
the channel and thus reduces the conveyance capa-
bility of the channel during flood flows. When flows
exceed the conveyance capacity of the channel, the
excess water overflows the banks and inundates the
adjacent flood plain. Sediment is then deposited on

N9

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

"IE/\EI'I VEIS NVBW EAOSV SHEILEIW N} 'NOILVAE'IEI

.933: 38—38 3% 05 K3 vofifiuwevv an aim fimfivafimcog 15653 2: mac? amoiwmom moimo :mm yo $595 533ml.m mane—m

mem mOwOZ<mDOIF Z_ .5710 EOEL §<mIleD m02<kw_D

 

 

 

 

 

ONF m: 0:. m0_. 00F mm 0m mm 0m mm Oh mm 00 mm Om mv 0.? mm on mN ON m_. OF
‘ _ u _ _ _ _ _ _ _ _ a _ _ _ fl _ _ _ v _ _
Much I
Orb I
mrh J
mwmr
own [I mm?
mfi II l..mmh_.l
hflmr
own I ..........
I mmmr
m? I zo_h<z<._n_xm_
OVB l
mu: l1 .\\‘\.\ ........... g Ehm Sim .3750 :mm
.\\.H\M.\.q ..... z: 32:5 .20
own I]
mmh l wants Umohzm‘. vac
| 503.3% >>0IEmmZm m>_mU
00% J a
man J [Illflfl.r...l.fllt $3 :md :80 $5.3m
| 7
||.|...r..1. |||||| 33 mnme mmuwm Umflm‘. >23 +0 QOF
OFF |
m _ _ _ w _ _ _ H _ _ _ _ _ _ _ _ _ h _ _ _
_ a _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ M _ _ _ _ _ _ _ _ _ 4 _ _ _ _ _ _
mm Vm” NM om WN ON VN NN ON me 0—. .v_. N_. 0—. w m V N

mImFm—S— ".0 mDZ<wDOIF Z. ._>_<D EOE”. §<mmemD mUZ<Fw_D

 

DOWN

ONMN

owmm

.. owmw

owmm

OOVN

ONVN

OVVN

OmVN

owvm

00mm

ONmN

OVmN

WEAEI'I VEIS NVEW EIAOBV .LEIEId NI 'NOILVAB'IE

N10

 

FIGURE 6.—Aerial view looking downstream (west) on April 20,
1965, showing the sediment flats at the confluence of the Gila
and San Carlos Rivers within the San Carlos Reservoir. The
reservoir pool, with water-surface elevation at 2,419.00 ft (737 m)
above sea level and usable contents of about 60,000 acre-ft
(74 hm3), is shown in the upper right center of the photograph.
The Gila River channel is at the lower right and the San Carlos
River channel is at the right center. The dark areas are well-
established saltcedar.

 

 

FIGURE 7.—-Aerial view looking upstream (north) on April 20,
1965, showing the San Carlos River on the left and the Gila
River on the right. The lower center part of this photograph
overlaps the right center part of figure 6. The area shown in
figures 6 and 7 includes the area periodically inundated by
reservoir water during the period 1945-65.

the flood plain because of decreased velocity of the
overbank flow. Intermittent channels and pools are
formed on the flood plain and provide a surface
irrigation for the saltcedar.

 

 

GILA RIVER PHREATOPHYTE PROJECT

 

FIGURE 8.—Aerial View of channel conditions of the Gila River
upstream from the San Carlos River in 1935—7 years after the
completion of Coolidge Dam. The upstream end of the reservoir
pool is shown in the lower left. The Gila River flood plain shown
in the photograph had been inundated by the reservoir pool
during the period 1929—35. Soil Conservation Service photo—
graph.

FIGURE 9.—Aerial view of conditions of a part of the Gila River
channel above the San Carlos River in June 1962. Flow in the
river is from upper right to lower left. The channel is continuous,
but a comparison of this photograph with figure 8 indicates that
the width has been reduced and the riverbed has been consider-
ably elevated. Extensive ephemeral off-channel ponds are
shown in the vicinity of the abandoned railroad crossing.

Figure 11 shows a saltcedar thicket after the
reservoir water has receded following an inundation

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

 

FIGURE 10.—View looking downstream (south) on August 18,
1965, at the Gila River flood plain from a point 2 mi (3.2 km)
upstream from the mouth of the San Carlos River. The former
Gila River inflow channel extends from the lower left corner
along the right side of the photograph. The natural levees and
vigorous bankside saltcedar are apparent. Deposition elevated
the channel above the surrounding flood plain. In the lower
foreground numerous channels cross the former main channel
at right angles. The water surfaces shown near the center of the
photograph are ephemeral off-channel ponds. The roadbed fill
approach to the abandoned railroad bridge shows as a dark line
in the center of the photograph.

 

FIGURE 11.—Flood-plain conditions after the area was inundated
by the reservoir pool. The water had been higher than the tree-
tops, thus killing the saltcedar. The white bands on the dead
tree stems are mud deposited during the reservoir recession.

of sufficient depth to kill the plants. The soggy
ground conditions, obvious in the photograph, are

 

N11

ideal for subsequent germination and establishment
of saltcedar growth.

The plugging of the inflow channel of the Gila
River during 1962—65 occurred within the 4.1-mi (6.6
km) reach shown in figure 12.

The deposition of sediment and reduction in slope
of the flood plain since construction of Coolidge Dam
in 1928 are shown in figure 13. The constriction
caused by the railroad fill at the abandoned railroad
bridge site was partly responsible for the exceptional
depth of sediment deposits and the resulting de-
crease in slope in this reach of flood plain. Deposition
of sediments at the mouth of the San Carlos River
may also have been responsible. The combination of
a reduction in channel width due to encroachment by
vegetation and a reduction in channel velocities due
to reduced slopes ultimately caused the deposition of
debris shown in figure 14. Logs and sediment formed
a dam practically eliminating the conveyance capa—
bilities of the channel. Saltcedar then became estab-
lished in the debris and sediment to create an
erosion-proof channel plug as shown in figure 15.

The formation of a channel plug causes channel
filling which quickly progresses upstream as indi—
cated in figure 12. The upstream end of the debris
plug deposited by a flood discharge during the period
July 15-22, 1964, is shown in figure 16. Large concen-
trations of debris are a characteristic of summer
floodflows on the Gila River. As described. by Burk-
ham (1970), these flows originate from intense indi-
vidual thunderstorms which produce high rates of
runoff from small tributary watersheds but only
rarely do they produce large rates of flow on the main
channel. The high rates on the tributaries strip
debris from the watersheds, convey the debris to the
Gila River, and thence downstream on what is
generally a moderate flow in the Gila channel. The
progress of the plugging phenomenon as shown in
figure 12 was 1.5 mi (2.4 km) in 1963,1.8 mi (2.9 km) in
1964, and 0.8 mi (1.3 km) in 1965 and is primarily the
result of summer flows.

In the summer of 1965, a channel was excavated by
dragline parallel to the plugged natural channel. The
high discharge in the Gila River during December
1965 to May 1966 raised the water level in San Carlos
Reservoir to an elevation which inundated most of
the Gila River flood plain shown in figure 12. An
excavated inflow channel has been maintained for
the Gila River since 1966.

SEDIMENT TRAP EFFICIENCY

Sediment trap efficiency is the ability of a reservoir
to retain sediment and is expressed in percent of total
incoming sediment (Gottschalk, in Chow, 1964). A

N12 GILA RIVER PHREATOPHYTE PROJECT

 

  
 

 

0 110025'
3 15' ‘
2,33 “n
3:
x *V r, ‘,
p
i 2,“ Sept. 1963
c: _
5 Oct. 1964
July 22, 1964
June 1963’
Plugging started 1962
is a
58“
i
x,
s: f: . ~,
5 i \y
I:
as. "N.

 

 

 

33 10 Base from US. Geological Survey, San Carlos
Reservoir l:62,500,1962

 

EXPLANATION

CL) ‘Il .2 MILES 60

' ' I 7* ———-o—-—

0 1 2 3 KI LOMETERs Distances along longitudinal axis of

reservoir from Coolidge Dam. in
CONTOUR INTERVAL 80 FEET thousands of feet
DOTTED LINES REPRESENT 20-FOOT CONTOURS
Oct. 1964

DATUM lS MEAN SEA LEVEL

Location of channel plug on date showu

FIGURE 12,—Extent of channel plugging in Gila River on indicated dates.

sediment trap efficiency of 83 percent was predicted ment was expected to pass through in suspension or

for San Carlos Reservoir by the US. Army Corps of by sluicing.
Engineers (1914, p. 30). The remainder Of the sedi- There are no records of sediment inflow or outflow

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

DISTANCE FROM COOLlDGE DAM,
lN KlLOMETERS

 

 

 

 

13 14 15 16 17 18 19 20 21
24601 I ‘ l l l l l ‘ ‘ l I l x J
.1 uJ
w a
L>U /// ’/ —' 748 __1
_, /
_ «1’ <
< 2450 / ~ w
% /'/ / - 746 V)
6/ / Z
2 «99 ./ <
< 2440 // ' 744 ”J
Lu — /,/ 933/ — E
E ,x” >/ / “J
LU / ,» / >
> / / // / r 742 o
O K’/ / m
m 2430 -" A/// 9.3; _ <
< // ‘ « 740 g;
{E / m
Lu 1—
LL 2420 _ / A- 738 E
Z
—_ l 2
Z l “ 736 2‘
O
,: 2410 — _ 9
<>I Abandoned railroad crossing > 734 E
s / a
. 732
L” 2400 l I l l I d
43 50 60 7O

D|STANCE FROM COOLIDGE DAM,
IN THOUSANDS OF FEET

FIGURE 13.—-Profiles along the longitudinal axis of San Carlos
Reservoir showing changes in bed elevations and slope with-
in the reach of channel plugging.

 

FIGURE 14.—View of the Gila River looking north from a point
2.6 mi (4.2 km) upstream from the mouth of the San Carlos River
photographed in September 1964. The debris in the channel is
part of a recently deposited channel plug. Flow is from right to
left.

for San Carlos Reservoir, so trap efficiency is not
known; it probably exceeds the predicted 83 percent.
Trap efficiency should be 96 percent or more at San

N13

 

FIGURE 15.—Downstream view on July 15, 1965, of a reach of the
Gila River channel which had been plugged in 1964. Saltcedar
is becoming established in the former channel bottom.

500 FEET

150 METERS

 

FIGURE 16.—Aerial view of channel plugging in the Gila
River on July 22, 1964. Flow is from right to left.

Carlos Reservoir, according to Gottschalk (in Chow,
1964, fig. 17—I—6).

ANALYSIS OF SEDlMENT DATA

The US. Army Corps of Engineers reported (1914,
p. 29—30) predictions of the volume of sediment
deposition based in part on streamflow records and
on 15 sediment samples collected from deposits

 

along the Gila River. Streamflow records of 1890 and

N14

1895-1912 indicate that the mean annual streamflow
was 346,000 acre-ft (427 hm3). In the 1914 report, the
predicted specific weight of deposited sediments was
70 lb/ft3 (1,120 kg/m3), predicted volume of deposi-
tion at 100 percent trap efficiency was 1.8 percent of
total streamflow, and the predicted trap efficiency
was 83 percent. Based on these predictions and
measured streamflow, the estimated mean annual
volume of reservoir sediment deposits was 3,740 acre-
ft (4.6 hm3).

Records of streamflow from November 1928
through August 1966 show the mean annual stream-
flow was 216,120 acre-ft (266 hm3), or 63 percent of
the predicted flow. The mean annual volume of
sediment deposition was 2,553 acre-ft (3.15 hm3), or
68 percent of the predicted volume.

A mean sediment concentration for stream dis-
charge can be estimated from measured volumetric
changes in reservoir deposits because a reservoir is a
collector for fluvial sediment moved by all transport
methods. The mean sediment concentration com-
puted from the information in the 1914 report is
14,478 ppm (14,615 mg/l). The mean concentration
from November 1928 through August 1966 is 13,684
ppm (13,808 mg/l) when computations are made
using measured streamflow and sediment deposition
and when estimates of trap efficiency, specific
weight of deposits, and specific gravity are 96 per-
cent, 70 lb/ft3 (1,120 kg/m3), and 2.65, respectively.

On a volumetric basis, sediment accumulated at
an average rate of 0.0118 (volume of sediment depos-
ited/ volume of streamflow) from November 1928 to
August 1966. The rates for the periods between
surveys, chronologically, are 0.0198, 0.0059, 0.0084,
and 0.0110, as listed in table 2.

INTERPULATION OF ELEVATION-CAPACITY
RATINGS BETWEEN CAPACITY SURVEYS

Elevation-capacity relations were defined for each
reservoir survey. Significant changes in these rela-
tions between surveys required the development of a
systematic method of interpolating changes during
the periods.

The simplest method of interpolation is to pro-rate
storage capacity change by time between consecu-
tive surveys. This method of interpolation can be
applied either to change in total storage capacity or
to change by increments of the total reservoir
storage.

Interpolation can also be made by pro-rating the
change in storage capacity according to streamflow.
Because the loss in storage capacity is the volume of
sediment deposition, this method is basically the use
of a mean sediment concentration computed for a

 

GILA RIVER PHREATOPHYTE PROJECT

period. The storage loss for any time interval is
estimated as the product of this mean concentration
and the interval streamflow. (See table 2 for period
rates for San Carlos Reservoir.) This storage loss can
be obtained graphically for San Carlos Reservoir by
using accumulated streamflow and figure 17. The
inset in figure 17 shows measured change in storage
compared to measured streamflow for each of the
periods between surveys.

The method of interpolating storage change
adopted for use in this report employs the equiva—
lency of the loss in surface—water storage capacity
and the change in sediment deposits. A procedure for
simulating deposition was developed and used to
estimate the sediment volume deposited each water
year. The development of this procedure and its
application in providing yearly capacity ratings is
described in the section “Development of Surface-
Water Storage-Capacity Ratings.”

HYDROLOGY

WATER-BUDGET EQUATION

A water budget is used in this report for presenta-
tion of an historical accounting of the hydrology of
San Carlos Reservoir for water years 1931—71. The
water-budget equation for San Carlos Reservoir is

IG+IS+IP+IGW+IT- OG‘ OE‘ GET
1' ASRi ASB = 0. (1)

The components of the water budget are identified
as follows:

IG =Gila River inflow at Calva,
IS =San Carlos River inflow at Peridot,
I P =precipitation input over the lake surface,
10 W =ground-water inflow,
IT =tributary inflow downstream from the
Gila and San Carlos River
gaging stations,
0G =Gila River outflow below Coolidge Dam,
0E =evaporation from the lake surface,
OET =evapotranspiration from the exposed
land surface of the reservoir,
ASR =change in surface-water storage, and
AS B =change in bank storage.

Neither the evapotranspiration nor the change in
bank storage was measured. Before evapotranspira-
tion could be computed by the water-budget equa-
tion, estimates of bank storage were necessary. A dis-
cussion of the evaluation of each water-budget com-
ponent is presented in the following sections.

HYDROLOGIC HISTORY OF SAN

CARLOS RESERVOIR, ARIZONA, 1929-71 N15

ACCUMULATEDSTREAMFLOW, Q, IN CUBIC KILOMETERS

 

 

 

 

 

 

 

 

0 1 2 3 4 5 6 7 8 9 10 11
120 I I | l I l I I I I I I I I II I
Accumulated streamflow, Q, in cubic kilometers
0 I 2I I ‘I 5 g — 140
110— few I I I —5o,5m 1
._ on
>‘ $33 8 a
D- I —- 40 m E;
I: 100 _ m 8 30~ 66 _ U E m
0 2’, g ‘9 3 o ~12o
E a"; a)" — so 2 5 / w
< 90 — r a. 20 — «‘3 _ 8 2 // _ 0m
U S U 1 m // <0:
I— “ = 19A — 20 c .9 , III“
gm ES 5% /// 100 BID
3 ~ —- “M
(I: 80— $310 ~10 §Z // w:
I III 8 H b ._ / a 0
I90: Egg 0 «931 I l | a // Lu 5
U 70 ~ 0 / _.
021< Q 0 1 2 3 4 /// 2%
/ .—
:1 l5 Accumulated streamflow, Q, in millions of acre-feet // 80 IELIJ 9
U) 60 — / fl 0 m
< 8 194/7c/ 966 LLI D
g 2 / / / &o\ D O
D
85; 50— /’// «91% ”—60 NE
D , I F .
0 o // a 9” EXPLANATION (I:
SI 40 — /// o“ _ 36
I'I— ,r’ \‘5' 1935 §<
<3 1935.,/ 1937 059 0 3“
J // "\04 Year of capacity survey — 40 0 <
D 30 ‘ / (3‘ -— 00
E ’ $‘Q <
8 // a?“
2 20 — / \4‘ _
/ p.—
/ 20
/
10 — / _
/
/
/
o l I I I I I I I 0
0 1 2 3 4 5 6 7 8 9

ACCUMULATED STREAMFLOW, Q, IN MILLIONS OF ACREiFEET

FIGURE 17.—Relation between accumulated decrease in storage capacity and accumulated streamflow by capacity survey
periods. Inset shows relation of storage capacity change and streamflow for each period.

SURFACE FLOW
GILA RIVER AND SAN CARLOS RIVER INFLOW

Records of streamflow into the reservoir for the
1929—71 water years were taken from US. Geological
Survey Water-Supply Papers 1313 and 1733 and from
US. Geological Survey annual state reports of
Arizona streamflow. Inflow records used in this
study are primarily those for the Gila River at Calva
and the San Carlos River at Peridot. Some of the 1929
inflow data were estimated from reservoir outflow
data and changes in reservoir storage. Additional
sources of Gila River streamflow information in-
clude reports by Burkham (1970) and the US. Army
Corps of Engineers (1914).

Figure 18 shows the total annual streamflow into
the reservoir for water years 1929—71. The stream-
flow data are included in table 17. The mean annual
inflow for this period was 214,940 acre-ft (265 hm3),
and the annual median was 159,000 acre-ft (196
hm3). The large difference between the mean and
median values is caused by infrequent years of
extremely high flow. Of the annual totals 67 percent

 

1100

 

 

I 1300

8 100° I 1200
<2: 900 1100 g
gm 800 J: 1000 3%)

I , 0
gm 7ooI 3 90° 2114
I— ".L ‘ H I’ 800 — I;
Z I; 600 Stream ow 10-year moving average I, 700 LII:
‘10 500 streamflow I I 500 I30
11] < . I I I: I-
o 400.» I - 500 <0
0:8 I I 400 1"”
< 300 /\ A I 1 a I
I 300 _
o 200 w A A I 200 0
(9 \J
D 100 \/ V 100

0 IT I I I I I I I I I I I I I I I I I I I I I I I I I ' I I I I I I I I I I I I I I I o

as; a s s 2:

mm c» m a: ma

F1- \— F .- r-l—

WATER YEAR

FIGURE 18.—Annual combined Gila River and San Carlos River
inflows to San Carlos Reservoir for water years 1929—71.

is less than the mean because of the influence of
these infrequent extreme annual totals.

Mean annual streamflow, 1929—71, into the reser-
voir was 33,450 acre-ft (41 hm3) for the San Carlos
River and 181,490 acre-ft (224 hm3) for the Gila River.

N16

The San Carlos River, with 8.6 percent of the con-
tributing area, produced 15.6 percent of the total
streamflow into the reservoir. Annual streamflow of
the San Carlos River ranged from 6.2 percent of the
total streamflow into the reservoir in 1959 to 40
percent of the total streamflow in 1956.

A generally declining trend in annual streamflow
coincided with a similar declining trend in annual
precipitation from about 1920 to 1962 (Burkham,
1970, fig. 7). The trend is not as evident during the
period of reservoir operation included in this report
(1929—71). The many years of low runoff during the
1940’s and 1950’s are distinct in the 10-year moving
average flow graphed in figure 18, but the average
annual runoff was higher near the beginning and
end of the study period. Annual streamflow exceeded
the mean only twice (1949 and 1952) during the 15-
year period 1943-57. By contrast, annual streamflow
exceeded the mean 14 times during the 43 year period
1929—71 and 5 times during the 11 year period 1958—
68. A histogram of annual inflow volumes is shown
in figure 19 for water years 1929-71.

Of special interest to users of San Carlos Reservoir
water is the probable water supply over time dura-
tions of a year or more. Table 4 shows the 1930—71
historical extremes in both maximum and minimum
supply for several time durations.

Table 5 gives the computed mean monthly inflows
for the Gila and San Carlos Rivers and their com-
bined total inflows, based on 1930—71 data. Figure 20
compares the mean monthly streamflows of the two
rivers in percent of their mean annual totals. In
general, the mean monthly discharges for both
rivers follow similar seasonal patterns. The greatest
monthly volumes typically occur during several
months of the winter season as the result of frontal
storms passing over most or all of the basin. The
period of increased runoff generally begins in
December and extends through March. The mean
Winter and spring (November to June) inflow is
155,480 acre-ft (192 hm3), but flow is highly variable
from year to year. The variability is demonstrated in
that the average deviation from the November to
June mean is 137,000 acre-ft (169 hm3), an 88 percent
deviation.

Increased streamflow from summer storms nor-
mally begins in July, peaking in August, and may
continue high into October. Individual summer
storms are smaller in areal extent than winter
storms and are more highly variable in precipitation
intensity, but the total summer streamflow is slight-
ly more consistent than the total winter and spring
streamflow. The mean streamflow from July

 

GILA RIVER PHREATOPHYTE PROJECT

INFLOW, IN CUBIC HECTOMETERS

 

 

 

 

 

 

 

 

0 200 400 600 800 1000 1200
10 l I I I I I I | I | I I l I
l l
w
n: 93 I
5 8% I .
> .21 I I
u.
0 54‘ . !
cc .
LU 4 E I g ' l
m 31 J a c
E 2% I g 8 I
z 1 5 fl
0 I l l IlI I H I I | I I I I I I I I H I
o o o o o c: o o o o o o
o o O o o o o o o O o
v— N m s: In to h 00 m o «—
lNFLOW,lN THOUSANDS OF ACRE FEET

FIGURE 19.—Frequency of occurrence of inflow volumes,
water years 1929—71.

TABLE 4,—Maximum and minimum volumes of water supplied by
streamflow into San Carlos Reservoir for different time
durations, in acre-ft

 

Duration of

 

period ........................ 1 year 3 years 5 years 7 years 10 years
Maximum inflow
for specific period
and dates of
period . 1,005,180 1,518,810 1,775,650 2,309,870 3,041,280
(1941) (1940—42) (1937—41) (1936—42) (1932—41)
Minimum inflow
for specific period
and dates of
period ., , ,, .. .. 34,740 198,880 444,140 755,440 1,301,020
(1956) (1969-71) (1944—48) (1950—56) (1944—53)

 

TABLE 5.—Mean monthly inflows for the Gila River, the San
Carlos River, and for both rivers 1930—71

Combined total

 

Mean monthly Mean monthly

 

  
 

 

Gila River San Carlos River mean monthly

Month inflow inflow inflow

(acre-ft) (acre-ft) (acre-ft)

Oct 8,587 821 9,408
Nov . ,, 5,845 809 6,654
Dec . 14,948 5,782 20,730
Jan . 26,594 5,611 32,205
Feb . 28,826 6,843 35,669
Mar . 31,669 6,307 37,976
Apr , 14,966 1,035 16,001
May 6,702 265 6,967
June ,,,,, 972 95 1,067
July 6,041 1,126 7,167
Aug ....... 23,272 3,461 26,733
Sept ..... 12,933 1,255 14,188

 

 

through October is 59,480 acre-ft (73 hm3), and the
mean deviation of 37,450 acre-ft (46 hm3) is 63 per—
cent of mean summer streamflow.

TRIBUTARY INFLOW

Tributary discharge from 390 mi2 (1,010 kmz) flows
directly into the reservoir between the stream gaging
stations (Gila River at Calva and San Carlos River at
Peridot) and the dam. It is convenient for water-
budget computations that flow seldom occurs in the
tributaries and that mean annual tributary flow is
probably less than 2 percent of the mean annual
input from all sources. Significant tributary runoff

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

     
  
  
       

/Gi|a River
\
t\ / atCalva

OF MEAN ANNUAL FLOW

‘ 1

\\ 7‘1

MEAN MONTHLY FLOW AS PERCENT

Ol‘rfiffir‘rffrfififrji 7777 Tirrfrl

- - - >

V > L) c Q C _ U) +4

o o m m 0. N 3 Cl

0 z o 2 Li 2 < E 2 3, < 3;
MONTH

FIGURE 20.—Mean monthly flow as a percent of mean annual
flow for the Gila and San Carlos Rivers.

usually occurs only during the summer storm sea—
son, July through October.

Tributary inflows from 72.6 mi2 (188 km?) of the
gaged tributary area entering the reservoir below the
Calva gaging station were determined for the sum-
mer storm seasons of 1964—71 (Burkham, 1976).
Totals of seasonal runoff from this area, and runoff
per unit area, are listed in table 6.

Several methods were used to estimate seasonal
runoff from all areas tributary to the reservoir. In the
first method, runoff was assumed to be spatially con-
stant. The seasonal runoff values per square mile for
1964—71, from table 6, were multiplied by the 390-mi2
(1,010 km2) tributary area and the results listed in
column 2 of table 7.

The tributary runoff estimates of column 3 in table
7 are results of a rainfall-runoff correlation. Rainfall
was measured for the Gila River Phreatophyte Proj-
ect in gages located near the downstream ends of the
tributary streams. The tributary runoff data is from

TABLE 6.—Tributary inflow into the San Carlos Reservoir along
a reach of the Gila River

 

Seasonal tributary runoff Mean seasonal runoff

 

Water year (acre-ft) (acre-ft/miz)

616 8.5

' 732 9.9

301 4.1

1,560 21.5

228 3.1

229 3.2

86 1.2

2,220 30.6

 

 

‘ Includes an additional gaged area of 10 mi2 (259 km").

 

N17

TABLE 7.——Estimated tributary inflow into San Carlos Reservoir

 

Estimated seasonal runoff, in acre-ft

 

 

 

     

 

 

Based on runoff From rainfall- Average 0f
measured for runoff relation, Based on 9 columns 3
Water year part of the area equation 2 acre-ft/mi2 and 4
1 (2) (3) (4) (5)

1931 . . . ,. 3,660 3,510 3,580
1932 .. 0 3,510 1,760
1933 .. 1,420 3,510 2,460
1934 . 1,800 3,510 2,660
1935 .. 4,520 3,510 4,020
1936 ,. 0 3,510 1,760
1937 . 0 3,510 1,760
1938 . 0 3,510 1,760
1939 .. 0 3,510 1,760
1940 0 3,510 1,760
1941 10,470 3,510 6,990
1942 1,320 3,510 2,410
1943 N 890 3,510 2,200
1944 . 8,640 3,510 6,070
1945 , 2,520 3,510 3,010
1946 , 6,190 3,510 4,850
1947 . 3,600 3,510 3,550
1948 . 1,650 3,510 2,530
1949 . 1,860 3,510 2,680
1950 . 1,550 3,510 2,530
1951 . 2,340 3,510 2,930
1952 . 1,720 3,510 2,620
1953 . 0 3,510 1,760
1954 . 11,460 3,510 7,430
1955 . 10,420 3,510 6,960
1956 . 0 3,510 1,760
1957 . 0 3,510 1,760
1958 , 9,330 3,510 6,420
1959 . 8,330 3,510 5,920
1960 4,370 3,510 3,940
1961 . 6,670 3,510 5,080
1962 . 2,460 3,510 2,980
1963 . 3,660 3,510 3,580
1964 . 3,320 11,730 3,510 7,620
1965 . 3,860 370 3,510 1,940
1966 1,600 13,660 3,510 8,580
1967 . 8,380 8,100 3,510 5,800
1968 1,210 0 3,510 1,760
1969 1,250 1,460 3,510 2,480
1970 470 3,620 3,510 3,560
1971 11,930 4,310 3,510 3,910
Mean ...... ‘ 3,998 2 3,758 2 3,510 2 3,631

 

‘Eight (8) year mean.
2Forty-one (41) year mean.

table 6. The regression equation relating rainfall and
runoff data was ,

Q =-0.382 + 0.093P, where P 2 4, (2)

with rainfall, P, and runoff, Q, given in inches. The
regression line and the data used to develop equation
2 are plotted in figure 21. The equation was used to
estimate runoff from the 390-mi2 (1,010 km?) tribu-
tary area using seasonal precipitation at San Carlos
Reservoir. The correlation between seasonal rainfall
measured on the Gila River Phreatophyte Project
area and seasonal rainfall at the San Carlos Reser-
voir weather station is poor. Runoff estimates in
column 3 of table 7 are probably poor partly because
of this high spatial variability in rainfall intensities.

Burkham (1974) indicated that the mean seasonal
runoff from all the study watersheds for the period
1963—71 was about 9 acre-ft/mi2 (0.004 hm3/km2).
Seasonal tributary runoff into the reservoir using

N18

SUMMER RAINFALL,P, IN MILLIMETERS

 

  

   
      

 

m o 50 100 150 200 250 300
g 0.7!— i l: l 1 l: 1 ‘1 11 j
1 E ~
3 06 f i 150 Q:
2 1 ‘V’
i 0.5» i KS
OI L 1 0;—
IL}... 0.41 Q=-o.382 + 0.0931) ' 100 g;
0 0 3' (Q and Pin inches) I 3
z ' , $4
3 _
2: 1..
‘ —
m . 3
2 0.1 f a; m
g 0‘ u 1 | °i l 1 1_,i 1 l 0
m 0 2 4 6 8 1o 12

SUMMER RAINFALL,P, lN INCHES

FIGURE 21.—Relations between measured rainfall and tributary
runoff for summer seasons 1964—71 from 72.6 mi? (188 km?) of
area tributary to San Carlos Reservoir.

this rate is 3,510 acre-ft (4.33 hm3), as shown in
column 4 of table 7.

These methods do not accurately estimate sea-
sonal tributary runoff to San Carlos Reservoir for
each year. However, the individual means at the
bottom of column 2 through column 4 of table 7
deviate less than 10 percent from the average of the
means. The close agreement of mean values suggests
that the 41-year water budget is improved by the
addition of tributary runoff estimates.

The average of the estimates from columns 3 and 4
of table 7 was selected somewhat arbitrarily to define
the seasonal totals to include in the water budget and
column 5 of table 7. The estimates in column 3
assume that a seasonal variability of precipitation is
uniform in space. The estimates in column 4 assume
runoff is uniform in space and time.

GILA RIVER OUTFLOW

About 78 percent of all inflow into San Carlos
Reservoir is released downstream into the Gila
River. Releases are based on reservoir supply and
the needs of the San Carlos Project near Coolidge,
Ariz. Records of releases are based on river-stage
data obtained at a flume 0.4 mi (0.6 km) downstream
from Coolidge Dam, at the U.S. Geological Survey
gaging station, Gila River below Coolidge Dam,
Ariz. Streamflow records have been available at this
station from 1939 to 1971 and near this location for
much of the period 1899—38.1

Discharge data for water years 1931-71 are in-

lQuoting from the 1971 annual report of U.S. Geological Survey surface-water records,
Gila River below Coolidge Dam, Arizona, “Records available—July to October 1899,
April 1900 to March 1902, July to September 1902, December 1902 to December 1904,
January to May 1905 (gage heights only), June to November 1905, August 1910 to Feb-
ruary 1911 (gage heights only), April 1914 to current year. Published as ‘at San Carlos’
1899—1911, as ‘near San Carlos’ 1914-1926, and as ‘at Coolidge Dam’ 1927-38.”

GILA RIVER PHREATOPHYTE PROJECT

cluded in the water budget. Mean monthly dis-
charges of the Gila River below Coolidge Dam are
shown in table 8 and are an indication of seasonal
demands of water for irrigation.

The peak instantaneous discharge since the dam
was constructed was 1,350 ft3/s (38 m3/s) on July 28,
1952. No flow occurred several times prior to 1938,
when the gaging station was about 0.2 mi (0.3 km)
upstream from its present location. The minimum
flow after 1938 was about 0.4 ft3/s (0.011 m3/s) which
includes the discharge of Warm Springs, a tributary
to the Gila River.

GROUND-WATER INFLOW

From 1963 to 1972, ground—water data were col-
lected and analyzed as part of the Gila River Phreato-
phyte Project. The part of the Gila River Phre-
atophyte Project area Within the boundaries of the
reservoir (fig. 1) included an 8 mi (13 km) reach of the
Gila River flood plain below the Calva gaging
station. Computations of ground-water inflow into
the reservoir were based on results of the project’s
ground-water analyses (Hanson and others, 1972;
Hanson, 1972).

The geologic water-bearing units along the Gila
River have been identified as alluvium and basin fill.
The alluvium was deposited in a trench incised in the
basin fill. Two components compose the alluvium:
flood-plain alluvium and terrace alluvium. Terrace
alluvium covers the basin fill in the trough and
extends above the flood plain on the adjoining
slopes. Flood—plain alluvium overlies the terrace
alluvium in the flood-plain region and averages
about 50 ft (15 m) in depth and 5,000 ft (1,500 m) in
width. The basin fill extends a considerable distance
up the slopes beyond the terrace alluvium.

Water enters the basin fill on the mountain slopes
and moves generally toward the flood plain. Where
the basin fill is in contact with the alluvium, suffi-
cient head exists to create a slow upward movement
of water from basin fill to the overlying alluvium.
Another ground-water source into the reservoir is

TABLE 8.-—Mean monthly discharge of the Gila River below
Coolidge Dam, 1931—71

Month

 

Discharge, in acre-ft

 

 

Oct
Nov .

Dec
Jan
Feb
Mar
Apr
May

June
July
Aug
Sept

 

 

 

 

 

 

 

 

 

 

 

 

8,580
5,980
6,640
3,160
6,920
16,330
21,140
20,800
24,050
25,880
21,840
18,110

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

downvalley flow through the saturated flood-plain
alluvium. A small amount of ground water enters the
reservoir from the alluvium deposited by tributary
streams but is considered insignificant.

Downvalley ground—water movement in the Gila
River flood plain alluvium has been computed at 5.1
acre-ft (0.0063 hm3) per day (Hanson, 1972, p. 25).
Hanson, Kipple, and Culler (1972, p. 317) estimated
basin-fill inflow along the Gila River at 0.82 acre-ft
(0.0010 hm3) per day per 1,000 acres (4.05 km2). This
is equivalent to about 0.50 acre-ft (0.00062 hm3) per
day per downvalley mile, or 10.5 acre-ft (0.0129 hm3)
per day over the 21-mi (33.8 km) reach from the Calva
streamflow station to the dam. The sum of the
alluvial inflow and basin-fill inflow is 15.6 acre-ft
(0.0192 hm3) per day in the Gila River part of the
reservoir.

Flood-plain alluvial flow (q) for the San Carlos
River at Peridot was computed from the equation
(3)
where K is the hydraulic conductivity estimated for
Gila River flood-plain alluvium at 5,200 gal/day/ft2
(212 m3/day/m2) and is assumed the same for the
flood-plain alluvium along the San Carlos River, m is
the depth of alluvium and is estimated to be 30 ft (9
m), w is the alluvial width of about 2,800 ft (850 m), u
is the slope of the downvalley ground-water surface
and is estimated to be equal to the downvalley flood-
plain slope of 0.00275. Equation 3 gives q equal to 3.7
acre-ft (0.0046 hm3) per day. The reach of the San
Carlos River from the Peridot gaging station to the
river mouth is 9.2 mi (14.8 km). Assuming that the
San Carlos River flood plain has the same basin-fill
inflow per unit area as the Gila River flood plain, the
inflow per downvalley mile per day is 0.50 acre-ft
(0.00062 hm3) times the ratio of alluvium widths
(2,800 ft/ 5,000 ft) for the two flood plains, or 0.28 acre-
ft (0.00034 hm3) per mile per day. Basin—fill inflow
along the 9.2 mi (14.8 km) San Carlos River reach is
therefore about 2.6 acre-ft (0.0032 hm3) per day. The
sum of the basin-fill and alluvial ground—water con-
tribution from the San Carlos River basin is 6.3 acre-
ft (0.0078 hm3) per day.

q=K-m-w-u,

PRECHHTATWON

Records of precipitation at San Carlos Reservoir
were obtained from National Weather Service pub-
lications and log books at Coolidge Dam. Table 9
includes annual precipitation totals for water years
1931—71. The mean annual precipitation was 13.87
in. (352 mm) and ranged from 6.46 in. (164 mm) in
1934 to 30.53 in. (775 mm) in 1941.

Mean monthly precipitation at San Carlos Reser-

 

N19

TABLE 9.——Annual (water year) precipitation, in inches, at San
Carlos Reservoir, 1931—71

VVater
year

 

VVater
year

VVater

year Precipitation Precipitation Precipitation

 

1945 .
1946 .
1947 .
1948 .
1949 .
1950 .
1951 .
1952 .
1953
1954
1955
1956 .
1957
1958

1931 .1
1932 .1
1933 .1
1934 .1
1935
1936
1937
1938 .v
1939 ,V
1940 .n
1941 .w
1942 .1
1943 .“
1944

1960 V”
1961 ..
1962 "_
1963 .m
1964 "n
1965 _“
1966 H"
1967
1968
1969
1970 a,
1971

     

 

voir for the period ranged from 0.21 in. (5.3 mm) in
May to 2.19 in. (56 mm) in August. Seasonal trends
are evident in the mean monthly precipitation data
of table 10. The monthly extremes, also listed in table
10, indicate the large variation in precipitation. The
mean precipitation for the summer season (July
through October) is 5.89 in. (150 mm), or 42 percent of
the mean annual total. Precipitation for the remain-
der of the year averaged 7.98 in. (203 mm) during the
41 years.

The volume of daily precipitation into the reservoir
was computed as the product of daily precipitation
measured at the San Carlos Reservoir weather sta-
tion and the surface area of the water in storage for
the day. Water year volumes of precipitation falling
onto the water in storage are listed in table 17. From
1931 through 1971, the total accumulated volume of
precipitation falling directly onto the water surface
of the reservoir was computed as 203,900 acre-ft (251
hm3), about 2.2 percent of reservoir input from all
sources.

EVAPORATION

Evaporation from San Carlos Reservoir is signifi-
cant because of the warm, dry environment in which
the reservoir is located. Evaporation, as used in the
San Carlos Reservoir water budget, is computed as
direct loss from the surface area of the reservoir pool.

Daily pan evaporation data for water years 1931—
71 were available for the weather station at San
Carlos Reservoir. Saturday and holiday pan evapor-
ation readings were omitted from 1930 through April
1948. The history of changes to the evaporation pan
at the dam is available in the National Weather
Service publication, “Substation History for
Arizona.”

Reservoir pool evaporation is computed as the
product of measured pan evaporation and the evapo-
ration pan coefficient. A daily volume of computed
pool evaporation is the product of daily pan evapora-
tion, the pan coefficient, and daily surface area of the

N20

GILA RIVER PHREATOPHYTE PROJECT

TABLE 10.—Mean monthly precipitation and monthly extremes (1931—71) at San Carlos Reservoir

 

Month .. Oct. Nov. Dec. Jan. Feb.

Mar. Apr. May June July Aug.

 

Mean monthly
precipitation,
in inches ..
Minimum monthly
precipitation,
in inches ............................. ,0 .0 .0 .0 .0
Maximum monthly
precipitation,
in inches ...............................

0.88 1.80 1.44

4.28 3.73 4.00 3.26

1.28 0.52 0.33 1.62

2.29 .98 1.40 4.68 5.99

 

pool. Records of daily pool evaporation volumes have
been published in annual reports of the Gila River
Water Commissioner beginning in 1936. A pan coef-
ficient of 0.7 was used by the Gila River Water Com-
missioner in computations of pool evaporation.

In December 1963, the US. Geological Survey
established a station about 350 ft (107 In) from the
San Carlos Reservoir weather station to collect
radiation and air temperature data for use in comput—
ing pool evaporation by energy-budget and mass-
transfer methods. lWind movements and water-
surface temperature data were recorded on raft-
mounted instruments at either one or two locations
on the lake; the number and location of rafts was
determined by the surface area of the pool. Stream—
flow temperature profiles (thermal surveys) of the
pool were made every 2 or 3 weeks to measure
changes in stored energy.

The energy—budget equation is based on the prin-
ciple of conservation of energy. Measurements are
made of most of the incoming and outgoing energy
components and of changes in stored energy in the
water body. The unmeasured energy remaining as a
residual of the energy-budget equation includes
energy for the evaporation process, energy of sensi-
ble-heat exchange, and energy advected by evapo-
rated water. The energy-budget method has been
described by Anderson (1954).

In the mass-transfer method, evaporation is
treated as the turbulent transport of water vapor in
the boundary layer overlying the water surface. The
method requires data of wind speed, vapor pressure
of the air, and vapor pressure of saturated air at
water-surface temperature. A detailed description of
the mass-transfer method is available in Marciano
and Harbeck (1954).

Computations of lake evaporation were made by
energy-budget and mass—transfer methods for 93
periods during 1964-71. Occasional incomplete data
reduced the number of periods of reliable data to 72.
A pan coefficient was computed for each period. The
average pan coefficient was 0.80 for the 8 years of
record.

Annual pan evaporation at San Carlos Reservoir

 

is shown in the upper portion of figure 22 for water
years 1931—71. Mean annual (water year) pan evapo-
ration was 97.3 in. (2,470 mm), and water-year
extremes ranged from the 1941 low of 83.6 in. (2,120
mm) to the high in 1939 of 111.4 in. (2,830 mm).

At San Carlos Reservoir, annual pan evaporation
appears to have followed a downward trend, as
indicated by the plot of 5-year moving averages
shown in figure 22. During this same 41-year period,
annual precipitation also indicates a decreasing
trend (Burkham, 1970). This seems contradictory
because the decrease in precipitation implied that
evaporation potential might have increased. Thus, it
was necessary to investigate further pan evapora-
tion data at San Carlos Reservoir.

Pan evaporation is affected by the conditions of
both the pan and the immediate environment, condi-
tions which can change independently of changes in
climate. In an attempt to evaluate the reliability of
the pan evaporation data at San Carlos Reservoir,
the data were compared with pan evaporation data
at other Arizona stations. Figure 23 compares the 5-
year moving average of pan evaporation at San
Carlos Reservoir with the 5-year moving average of

 

3 E
‘ JD
Z 21];
9m 0m
._LLl Fl-
<zI ”J
0:0 g;
83 0;}
< 3(-
a >5
LU

 

 

 

WATE R Y EAR
EXPLANATION

5 year moving average of pan evaporation
Adjusted annual evaporation for reservoir

FIGURE 22.——Evaporation at San Carlos Reservoir for water years
1 931—71.

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71 N21

 

. the mean of annual pan evaporation for Mesa,
{$110 Wfififlffwwffi‘fiwfiflflw Wfirflfl Em” Roosevelt Lake, and Tucson .(Univ. of Ariz.). The
(55100 ‘1} f/Vg\ __ (i 26 E5: g trend at San Carlos Reserv01r is downward for much
2 Z i San Carlos/ »~/‘*“\, .7 ”/4 24 o g E of the 1931-71 period as opposed to an upward trend
53‘ 901 “Rage/513': «ix/LCM" f M R velt nd i 22 32% at the other stations. Changes in the relation be—
3 g i: L", 7 M w ‘ eai‘ucosrgn:33.xlfiocgljexinfiz)affik f: m g g tween the San Carlos Reservoir data and the mean of
“ F 8‘ ‘ TTWS‘E‘ ‘ T‘ ‘Tf ‘3 ‘ ‘ 18‘ ‘ ‘ ‘ 53E :1 the other three stations were found to have occurred
5’2 9 E e 921’ in about 1938 and 1950 by use of a double-mass curve
WATER YEAR (fig. 24). Accordingly, three separate time periods
FIGURE 23.-—Five-year moving averages of pan evaporation at were used In computlng evaporation from the reser'
San Carlos Reservoir and the mean of pan evaporation at Mesa, VOiYZ (1) 1931-38, (2) 1939-50: and (3) 1951—71-
Roosevelt Lake, and at Tucson (Univ. of Ariz.) for water years For period 3, a pan coefficient of 0.78 was derived

 

1931—71.

CUMULATlVE PAN EVAPORATION—SAN CARLOS RESERVOlR, IN METERS
0 5 10 15 2O 25 3O 35 40 45 50 55 60 65 70 75 80 85 90 95 100

320 is " A? 'EA'T”J"#T_L’" filu—ir ' '"47"JWH'WTL’"'JF"’AT"“*"1*L %*E#%"E';T"' "'L‘IWL'T

  
   

1971 L
l \, ‘ 95
300 v 4
i k 90
l i
280 ‘; l" 85
‘ ‘ uJ
”>21 L 80 )4
< ‘L .,. A <[
_| 260 ; 1 :
5 l T 75 d
l“ 240 r a >
> 1 Lu
3,1 ; ‘— 70 (n w
0 l O D:
+- 220 l— A 0 Lu
0 LU ‘ i (I p—
cc Lu 1 P 65 . u:
< ‘L ‘ i 5,:
$2 200 E 1 60 “g E
E A, ‘ ‘ ,:
u. [I \ ‘ 5N-
O m 180 5 + 55 E
z < I ‘ <21 <t
Eu. ‘ l w u.
0 i f“ 50
2 , 160 h ~; 2‘9
l 2 ‘ Z >
8% i E 45 9 E
i: v 140 1F a 2—1 3
< Z l ‘_ Z
n: o r 40 n:
t o o
2 8 1 ‘ m a
< 3 1203 i— 35 :3
> " t 1 Lu l—
;9 0‘ 4 Z”
'l 0 Z
<( < l7 30 E <
3 1 l w
2 80 L <f— 25 E
,_ l 1 <
< i ‘ _'
_| ‘ *— 20 1)
g 60 7— l 5
‘ ‘ 3
3 i 15
o 1 l— 0
40
F m
\
20k» 5
l
0 , ,li Lnl LL..,,.L_L__i_,.Lr,_Li_Ln#LL,%L,IL.IMFJL_Mrfl LL, L 0
o 20 4o 60 so 100 120 140 160 180 200 220 240 260 280 300 320 340

CUMULATIVE PAN EVAPORATION—SAN CARLOS RESERVOIR, IN FEET

FIGURE 24.-Double-mass diagram of cumulative pan evaporation for San Carlos Reservoir and mean of Mesa, Roosevelt Lake,
and at Tucson (Univ. of Ariz.) for water years 1931-71.

 

 

N22

from computed lake evaporation at San Carlos Res-
ervoir and the average of pan evaporation for Mesa,
Roosevelt Lake, and Tucson (Univ. of Ariz.). This
coefficient was assumed applicable in periods 1 and 2
also, when computing San Carlos Reservoir evapora-
tion on the basis of pan evaporation at the other
three stations. San Carlos Reservoir pan coefficients
for periods 1 and 2 were 0.61 and 0.66, respectively,
obtained from the relation: San Carlos Reservoir pan
coefficient for a period equals computed lake evapo-
ration divided by measured pan evaporation. Periods
1 and 2 pan coefficients were applied to annual pan
evaporation measured at San Carlos Reservoir to
compute annual lake evaporation. The annual pan
evaporation, pan coefficients, and corresponding
computed lake evaporation amounts for the 41-year
study period are included in table 11. Computed lake
evaporation is compared with the San Carlos Reser-
voir pan evaporation in figure 22. The annual lake
evaporation expressed as volumetric loss is given in
table 17. Mean annual depth of evaporation was
69.95 in. (1,780 mm) and represents 10.5 percent of
the outflow from the reservoir. Mean monthly depths
of evaporation from the lake are shown in table 12.

Monthly differences between the computed pan
coefficients shown in table 12 are caused primarily
by changes in available energy. There was not
sufficient confidence in the computed monthly pan
coefficients for use in computations of lake evapora-
tion, although it is evident that the relation between
pan evaporation and lake evaporation is subject to
seasonal change.

WATER STAGE AND SURFACE-WATER STORAGE
AT SAN CARLOS RESERVOIR
Prior to January 15, 1937, water stage was deter-
mined by use of reference points of known elevation
on a series of stakes. A continuous record of lake
stage was made from January 1937 through 1971
using a water-stage recorder. Daily stage records
were applied to the elevation-capacity relations (fig.
3) to obtain daily reservoir storage contents. Daily

 

 

GILA RIVER PHREATOPHYTE PROJECT

records of stage and usable surface-water contents
are published in the annual reports of the Gila River
Water Commissioner. Usable surface-water storage
data are also contained in annual US. Geological
Survey reports. Figure 25 shows beginning-of—month
reservoir stages for water years 1929-71. Total
surface-water storage and change in surface-water
storage were obtained from the yearly elevation-
capacity ratings and are shown in table 18 by water
year. The derivations of these ratings are described
in the section “Development of Surface-Water
Storage-Capacity Ratings.”

TABLE 11.—Evaporatlon at San Carlos Reservoir by water year

 

 

 

 

 

 

TABLE 12.—Mean monthly evaporation and mean monthly pan coefficients for San Carlos Reservoir

 

Month ............................................ Oct. N ov. Dec. Jan. Feb.

 

 

Mean monthly lake

evaporation in inches,

1931—71 .................................. 5.01 2.66 1.57 1.73 2.52
Mean monthly pan

coefficients from energy

budget computations,

1964-71 ................................. .88 1.00 1.05 .74 .66
Mean monthly pan

coefficients from mass-

transfer computations,

1964—71 .................................. .90 1.10 .93 .71 .68

1931 —71
Pan Computed
evaporation lake evaporation

Water year (in.) Pan coefficient (in.)
1931 94.47 0.61 57.63
1932 96.46 .61 58.84
1933 105.45 .61 64.32
1934 109.86 .61 67.01
1935 95.64 .61 58.34
1936 . 103.00 .61 62.83
1937 . 107.54 .61 65.60
1938 . 107.56 .61 65.61
1939 . 111.35 .66 73.49
1940 . 101.83 .66 67.21
1941 . 83.64 .66 55.20
1942 . 94.34 .66 62.26
1943 . 101.13 .66 66.75
1944 . 98.73 .66 65.16
1945 . 100.81 .66 66.53
1946 . 104.80 .66 69.17
1947 . 104.20 .66 68.77
1948 . 106.18 .66 70.08
1949 .. 98.30 .66 64.88
1950 .. 98.68 .66 65.13
1951 .. 98.82 .80 79.06
1952 .. 90.55 .80 72.44
1953 .. 94.84 .80 75.87
1954 .. 93.73 .80 74.98
1955 .. 91.49 .80 73.19
1956 .. 100.35 .80 80.28
1957 .. 96.54 .80 77.23
1958 89.03 .80 71.22
1959 97.63 .80 78.10
1960 99.30 .80 79.44
1961 94.60 .80 75.68
1962 94.63 .80 75.70
1963 89.13 .80 71.30
1964 94.81 .80 75.85
1965 86.74 .80 69.39
1966 88.58 .80 70.86
1967 93.12 .80 74.50
1968 91.39 .80 73.11
1969 92.56 .80 74.05
1970 91.19 .80 72.95
1971 97.47 .80 77.98

Mar. Apr. May June July Aug. Sept.

4.42 6.57 9.03 10.58 10.28 8.46 7.12

.68 .65 .69 .75 .79 .79 .86

66 66 70 .71 83 82 79

 

 

 

 

 

INFEETABOVEMEANSEALEVEL

FlRST—OF-MONTH ELEVAT|ON,

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

 

 

 

 

 

 

 

 

 

 

 

2520
I I I I I

| I I | I | ,2 765

2500
750
2480 755
2450 750
745

2440
740

2420
735
2400 730
2380 725
2360 I I I I I I I I 1 720

1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942

2500
—-750
2480 ‘1—755
2460 ——750
——745

2440
—-740

2420
—-735

2400
——730
2380-— ——725
2360 I I I I I I I I I I I I — 720

1943 1944 1945 1945 1947 1948 1949 1950 1951 1952 1953 1954 1955 1955
2500

2480

2460

2440

2420

2400

2380

 

l

l

I

l i

I

 

l

—-760

755

750

745

740

735

730

725

 

—-720

 

6
23 O 1957

1958

1959

1960

1961

1962

1963 1964 1965
WATERYEAR

1966

1967

1968

1969

1970

1971

FIGURE 25.—Beginning-of-month lake stage of San Carlos Reservoir for water years 1929-71.

PJ23

FlRST-OF-MONTH ELEVATION, IN METERS ABOVE MEAN SEA LEVEL

 

 

N24

The highest reservoir stage of record (1929—71
water years) was 2,501.62 ft (762.49 m) above mean
sea level on March 18, 19422. Usable surface-water
storage at that elevation was 819,200 acre—ft (1,010
hma) using 1947 capacity tables. This storage is
about 83 percent of the usable surface—water storage
capacity at 2,511 ft (765 m). Other peak stages
recorded were 2,491.17 ft (759.31 In) in 1968 and
2,477.30 ft (755.08 m) in 1966. Usable surface-water
storages at these elevations were 643,324 acre-ft (793
hm3) and 417,673 acre-ft (515 hm3) using 1966
elevation-capacity tables.

The minimum stage has fallen to, or below, the
lowest outlet elevation of 2,382.63 ft (726.22 In) at
some time during 14 of the 43 years of record. At other
times, inflow was barely sufficient to keep the pool
level above the stage of zero usable surface-water
storage. The pool level was below 2,385 ft (727 m)
elevation about 15 percent of the time and can occur
at any time of the year.

A summation of the number of days in which mean
daily stage was within a prescribed elevation inter-
val was needed prior to frequency analyses of stor-
age. Figure 26 is a graph and listing of the number of
daily occurrences within each 5-ft (1.5-m) stage
interval.

The cumulative number of days for which the lake
stage was above a specified elevation is expressed as
a percentage of the total days of record in figure 27.
Figure 27 shows that stage exceeded elevation 2,439
ft (743 In) only 25 percent of the time. Elevation 2,416
ft (736 m) was exceeded 50 percent of the time, and
elevation 2,392 ft (729 m) was exceeded 75 percent of
the time.

Figure 28 is a time-storage curve which shows the
percentage of time that usable surface-water storage
was equal to or greater than a given volume, based
upon first-of—month storages. Usable surface-water
storage exceeded 176,000 acre-ft (217 hm3) only 25
percent of the time, 64,000 acre-ft (78.9 hm3) 50
percent of the time, and 11,000 acre-ft (13.6 hm3) 75
percent of the time.

Table 13 shows the usable surface-water storage
available at the beginning of each month at 25, 50,
75, and 100 percent of the time, as an indicator of
seasonal availability.

BANK STORAGE
A part of the total storage in a reservoir in addition

2 Prior to January 1, 1948, gage datum was 0.72 ft (0.22 m) below mean sea level. (See
US. Geological Survey, 1954, p. 636). No adjustments were made in this report for the
datum change except that the peak stage listed has been adjusted to datum used after
January 1948. The error in the water budget introduced by datum change is small
because the computed annual change in surface-water storage is not significantly
affected.

 

GILA RIVER PHREATOPHYTE PROJECT

2510 awn/W, 755
2500
U)
I. 2490 760 0:
L” IE
LLI
LL: 2480 755 U14
2 “J ELL]
_ > 2470 >
.5: 2m
«I _-_l
< < 2460 750 _,<
> w <1”,
0: w 2450 EU,
Lu 2 745
I— 102
2E 2440 EE
_ E _
Stu 2430 740 25
— >
:0 2420 E;
m
> 7 <50
m <( 2410 35 ><
_J Lu
0: 2400 ii
730
2390
Below 2385

 

o o o oo o o o o o o oo o
ooooooooooooo
N «LocoOanqooquw

l—I—FFFNNNN

NUMBER OF DAYS IN INTERVAL

FIGURE 26.-—-Number of days in which lake stage was within a
particular elevation interval for water years 1929—71.

LAKE STAGE, IN METERS ABOVE MEAN SEA LEVEL

 

730 735 740 745 750 755 760 765
D 100I "'7"! " ’1 I ' T VII r 'r :2 ‘ v' T"’r”r"zr*T"7"' "Iw’fi I"I:
LU I
I- 90 I
<(
0 80‘
— o 1»
gm , .
O
— 70 . \
$3 I I \\
_ U I
,_ X 60 \K
.1 “J \
< 2 50 r 7e 7,, , 7 3x 1
’_ .
,9; 40 \ 7
Lu I
I
u. 0 I I
O < 30 t’ I I \
p— ; fit 7 .. #1,", J3
I— I c c c .
2‘” 201’ 8‘ 8 8I\‘\._
8 331 E: E \\
II 10 r Q‘ Q‘ Q; \
w I s s a I \
CL 0 I,,,i,,,,,I,I,,. :,, .1 I 1 1 I1 I 1,1 I l l inirrftjaxfl
3 8 “(‘3 S E 8 I?
m v v v v to L0
N N N N N N N

LAKE STAGE, IN FEET ABOVE MEAN SEA LEVEL

FIGURE 27.—Percentage of time lake stage of San Carlos
Reservoir equaled or exceeded a given elevation for water
years 1929—71.

to surface-water storage is bank storage. In some
reservoir water budgets, the change in bank storage
has been treated as being equal to the residual of a
water-budget equation, even when the residual in-
cluded significant evapotranspiration losses,
ground-water inflow, and so forth. For this investiga-
tion the change in bank storage is considered a sepa-
rate water-budget component. The symbols for the
reservoir storage terms are as follows:

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

USABLE SURFACErWATER STORAGE,
IN CUBIC HECTOMETERS

100 200 300 400 500 600 700 800 900 1000 1100
I I l I I I I l l I I I I I I I I I
I

—O

 

  
 

 

 

D 100 l I I I I
E010 90 ~
<ww
0:: 8° -
2:3370 = -
the. 6° i ‘
< :
§;2350"y", —
33X 40 1 1 —
u'mu.I"” I i
ONO“: o E: v: _
I—_,<O 3 01.84 ........
2:00: 2: u: I ; -
L“<10 2° <1’13: 6.0:
gun— 10 a4 a; 0c: —
“J30" L3:0: '08:
l o Ilm' IN '| I I I I
0 1 2 3 4 5 6 7 8 9

USABLE SURFACE-WATER STORAGE,
IN 100,000 ACRE FEET

FIGURE 28.—Percentage of time usable surface-water storage at
San Carlos Reservoir equaled or exceeded a given volume for
water years 1929—71.

ST is the total water in storage,

SR is water contained in surface-water storage,

SB is the water contained in bank storage, and

A indicates the change in an associated stor-
age term.

Inflow and outflow from bank storage can be
computed by use of combined water-budget and
modeling methods for some reservoirs, as demon—
strated by Simons and Rorabaugh (1971). At San

N25

Carlos Reservoir insufficient ground-water data
were collected to model aquifer response. However,
by applying the water budget of the reservoir to
selected short periods during the winter, estimates of
bank storage were made.

The change in bank storage, ASB , at San Carlos
Reservoir was estimated for short budget periods by
solving a modified form of equation 1. The periods
selected included winter months when ASB was a
significant budget component and when evapotran-
spiration, OET’ was insignificant. Tributary flow
was not included because no flow was assumed in
winter (Burkham, 1974). Equation 1 as applied was

(4)

The rate of change in bank storage is dependent
upon the change in surface-water storage, ASR , so it
is important that ASR be small for a month or more
before and after the evaluation period. Table 14
illustrates the application of equation 4 in computing
A83 for the winter period January through April of
1965.

A83 was determined for each of 23 winter periods
of significant ASB increases. ASB was not deter-
mined for periods of decreasing 83 because all
periods of a significant decrease in S3 corresponded
to periods of high evapotranspiration rates.

Several procedures were used to investigate the

ASBZIG+IS+IP+IGW_ OG— OEiASR.

 

TABLE 13.—-—Percentage of time that available monthly surface-water storage was less than amount shown

Usable surface-water storage, in acre-ft

 

 

 

 

 

 

 

 

 

Percentage
of time Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept.
6,820 3,150 11,200 8,700 20,100 29,100 35,700 21,000 12,700 950 5,000 8,200
50,300 48,000 46,000 55,000 93,400 104,000 113,000 115,000 97,000 70,100 44,000 47,000
113,000 103,000 107,000 138,000 198,000 194,000 260,000 235,000 203,000 165,000 140,000 116,000
689,000 744,000 752,000 802,000 835,000 848,000 841,000 816,000 779,000 727,000 676,000 659,000
TABLE 14.-—-Example of water budget used to determine change in, bank storage
Month Surface-water Change in Change in Change in
Water surface Change storage, surface-water total storage, bank storage,
(water year 1965) stage in stage S R storage, AS R Inflow2 Outflow3 ST‘ A885
((1)1 (ft) (acre-ft) (acre-ft.) (acre—ft) (acre-ft) (acre-ft) (acre—ft)
Jan 21407447 5.88 29,790 15,464 18,986 619 18,367 2,903
Feb 1413-35 5.28 45,254 16,446 24,340 4,810 20,030 3,584
Mar 141833 .33 61700 1,069 17,327 14,715 2,612 1,543
Apr 2141896 —.45 62-769 —l,458 10,642 11,743 -1,106 352
2,418.51 61,311
Jan—Apr
totals .................................. 11.04 31,521 39,903 8,382
ASE/ft : 759 acre-ft/ft Average stage = 2,412.99 ft
AS /AS = .210 Average S = 45,550 acre-ft
B T R
ASE/AS}? = .266 Elevation corre‘
sponding to
average SR = 2,413.45 ft

 

1Values for beginning and end of month.

2Inflow is the sum of Gila River inflow, San Carlos River inflow, precipitation on water surface, and ground—water inflow.

3Outflow is the sum of Gila River outflow and evaporation.
‘AST equals inflow minus outflow.

51133 = AsT — AsR.

 

 

N26

relation between bank-storage capacity, SB , and
water-surface stage. In the first procedure, the ratio
A83/ A8}; was compared to stage, where A83 and
ASR values were from the water budgets. In the
second procedure, the ratio ASB/ AST was com-
pared to stage. AST is the change in total storage,
computed in the water budget as the difference
between inflow and outflow components. Relations
of A83 to ASR and A83 to AST are shown in figure
29 for two winter budget periods. The need for
extending the time period of the water budget past
the period of rapidly increasing surface-water stor-
age is obvious in the upper ends of the curves in
figure 29A. These curves show SR and ST decreased
during April but SB increased because of the time lag
between inflow into surface-water storage and sub-
sequent movement into bank storage. Figure 293 is
included to show the small increase in 83, when
compared to changes in SR and ST, in December
1967 and January 1968. This condition occurred be-
cause gravity drainage from bank storage was in-
complete at the start of the period. As a result, only
the February through May period of 1968 was used in
the analyses of bank storage. Figure 29A also shows
that the cumulative plots of A83 versus ASR and
AST define the ratios ASB/ AST and AS3/ ASR
from the slopes of lines drawn from the start to end of
the period.

The third and principal procedure of analyzing
bank storage at San Carlos Reservoir was based on
the relation between a change in bank storage and a
change in stage. The computed A83 for a budget
period was divided by the range in stage, giving the
rate ASB /ft. Because each rate, ASB /ft, is related to
a specific range of stage, the rate must be associated
with the stage which seems most representative of
the budget period. A representative water-surface
stage is easily obtained from either of two calcula-
tions. The first calculation simply determines the
mean of the beginning and ending stages for the
period. In the second calculation, considered better,
the mean value of the beginning and ending surface-
water storages for the period is applied to elevation
versus surface-water capacity tables to obtain the
corresponding stage. The stage for each period is
plotted against the corresponding ASB /ft of the
period to define the ratings of A83 and stage as
shown in figure 30 for San Carlos Reservoir data.
The winter stage and storage data obtained from the
water budget of the reservoir and used in the above
three procedures are tabulated in table 15.

The 1931-47 stage and storage data define one
curve of figure 30, and the other was defined by data

 

GILA RIVER PHREATOPHYTE PROJECT

of 1948—71. The shift in the rating with time reflects
an increase in bank storage due to sediment accumu-
lation in the reservoir. The data were inadequate to
define more than the two ratings shown in figure 30.
Much of the scatter exhibited by points from the
1931-47 data is due to inaccuracies in the early

CUMULATIVE CHANGES IN SURFACEVWATER
STORAGEIASRI AND TOTAL STORAGE (AST).
IN CUBIC HECTOMETERS
O 5 10 15 20 25 30 35 40 45 50 55
9 II | I I I I II II I I I I I II I
Mean ASE/AST fooiperiod Apr.
8- Mean ASE/ASK for period\\// ° / \' —

 

—|
.-

I
.1
O

I
(D
CUMULATIVE CHANGE IN BANK STORAGE

(11

(ASE), IN CUBIC HECTOMETERS

    

1_ / A. January through April 1965 _

 

 

CUMULATIVE CHANGE IN BANK STORAGE
(ASE) IN THOUSANDS OF ACRE FEET

 

0 l I I I | I I I
0 5 10 20 25 30 35 4O 45 50

CUMULATIVE CHANGES IN SURFACE-WATER
STORAGE (ASR) AND TOTAL STORAGE (AST),

IN THOUSANDS OF ACRE-FEET

O

CUMULATIVE CHANGE |NSURFACE~WATER
STORAGE (ASR) AND TOTALSTORAGE (ASTI.
IN CUBIC HECTOMETERS

 

0 100 200 300 400 500 600
5 I I I I I I I ‘ I I I I I I I
O I I I I I | I . I nJo ”May? 60
' 40 ~ / f 50
* Mar. “
, r 40
30 7 /
ASR ° - AST 7

20*

I I
M b.)
O O

CUMULATIVE CHANGE IN BANK STORAGE

   

B. December 1967
through May 1968

(ASE), IN CUBIC HECTOMETERS

.10r

 

CUMULATIVE CHANGE IN BANK STORAGE
(ASE) IN THOUSANDS OF ACRE FEET
I

 

 

0 ‘ ' I I I I I I I I 0
0 100 200 300 400 500

CUMULATIVE CHANGE IN SURFACE-WATER
STORAGE (ASR) AND TOTAL STORAGE (AST).
IN THOUSANDS OF ACREVFEET

FIGURE 29.—Relations of cumulative A83 to cumulative AS R and
cumulative AS B to cumulative ASTfor the periods (A) January
through April 1965 and (B) December 1967 through May 1968.

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

COMPUTED CHANGE IN BANK STORAGE, IN
CUBIC HECTOMETERS PER METER

4 6

0.5 2
2 51° nL'n‘F‘Tnb'miT mimr—HT‘TTrL‘J
2 500 [
1

81012

765

 
 
  
    

— 760
2490

Number (66) indicates water year
2480
2470
2460
2450

2440

— 755
n—7 750
— 745
2430

_— 740

2420 /
1931t01947\/,

rating _— 735

  

2410

RESERVOIR STAGE, IN FEET ABOVE MEAN SEA LEVEL

 

 

2400 rating
460 // __ 730
39/.‘53
2390 /° ‘
/ 57
(/ |
1
238(300 1000 3000

COMPUTED CHANGE IN BANK STORAGE,
IN ACRE-FEET PER FOOT

FIGURE 30.—Relation between the computed change in bank-
storage capacity and elevation at San Carlos Reservoir for
water years 1931—47 and 1948—71.

capacity surveys. The data points for 1931 and 1932
include an increment of water which went into
nonretrievable bank storage when the reservoir ini-
tially filled.

Table 16 includes usable bank storage capacity
ratings for the 1931-47 and 1948—71 periods based on
figure 30. It was not practical or necessary to make
an elevation-capacity rating of total bank storage
because an estimate of bank dead storage would
have been required.

The usable bank-storage capacity at the spillway
elevation of 2,511 ft (765 m) was about 152,800 acre-ft
(188 hm3) for 1931—47 and about 159,200 acre-ft (196
hm3) for 1948—71 (table 16). At this elevation, usable
bank-storage capacity is about 14 percent of total
usable storage capacity. At lower reservoir eleva-
tions, table 16 shows that usable bank storage
sometimes exceeds usable surface-water storage.

83 is never static because of the time of response
required to adjust to changes in 813- However, the
quantity of water required to place SB and SR in
equilibrium at the end of a water year is usually

RESERVOIR STAGE, lN METERS ABOVE MEAN SEA LEVEL

 

N27

small in comparison to water-year budget totals and
for expediency was assumed zero.

The fact is stressed that reservoir water avail-
ability is more than just the amount in usable
surface-water storage. Figure 31 shows this differ-
ence by comparison of 1966 ratings of usable surface-
water storage capacity and total usable storage
capacity.

WATER-BUDGET ANALYSES

The water budget utilizes the conservation of mass
equation

I- 0: AS, (5)

where I is inflow, O is outflow, and AS is change in
storage. Identification of all the I, 0, and AS com-
ponents included in the water budget of San Carlos
Reservoir is given in equation 1.

The water budget was computed by months and by
water years for the 41 years of record using equation
1. Data for all the yearly inflow and outflow budget
components except OET are recorded in table 17.
Analyses of computed 0 ET data are included in the
following section on evapotranspiration. Table 18
lists the storage values, A33 and A83, the summa-
tion of the inflow and outflow components given in
table 17, and the evapotranspiration, OET, for each
water year during the 41-year period of record.

The data in tables 17 and 18 were used to c0mpare
the magnitude of each inflow component with the
total inflow for the period 1931—71. The outflow com-
ponents were compared similarly. Gila River stream-
flow contributed 78.2 percent of the total inflow; the
San Carlos River, 14.5 percent; ground water, 3.5 per-
cent; precipitation, 2.2 percent; and tributary flow,
1.6 percent. The outflow components and percent of
total outflow are: Gila River, 78.2 percent; evapo-
transpiration, 11.3 percent; and lake surface evap-
oration, 10.5 percent. Figure 32 compares the relative
magnitude of each component.

EVAPOTRANSPIRATION

Water loss by evapotranspiration (ET) from the
exposed surface of San Carlos Reservoir occurs by
plant transpiration and by evaporation from soil,
litter, and ephemeral ponds. Water in the reservoir
area becomes available for ET by movement from
streams, ground water, reservoir surface—water stor-
age, and by direct precipitation on the exposed
surface. Annual E T computed by equation 1 is listed
in table 18 and plotted in figure 33.

Large errors in the computed ETlosses occur when
one or more of the hydrologic components of the
reservoir are in a state of rapid transition at the end
of a budget period. The annual E Tshown in figure 33

 

 

N28

GILA RIVER PHREATOPHYTE PROJECT

TABLE l5.——Results of procedures to determine bank storage capacity at San Carlos Reservoir

 

 

 

 

Starting and S‘ S‘ ‘
AS] A83 A8}; ending Mean Mean Elevation of Al 8 A B AbB
Water Period elevations elevation S mean S R
year included (acre-ft) (acre-ft) (acre-ft) (ft) (ft) (acre-ft) (ft) AS T A S R ft
. - , , 2421.76 .
1931 Now—Mar. 92,217 71,812 20,465 2435 66 2428.71 163.059 2429.20 0.222 0.285 1,472
1932 .. Dec—Apr. 323,063 270,007 53,056 33%;?) 2454.31 338,814 2456.68 .164 .196 1.600
1935 Dec—Apr. 136,282 121,428 14,854 3:22;? 2422.03 122,006 2424.65 .109 .122 520
2420.04 .
1936 . Jan—Apr. 88,530 71.804 16,726 24% 00 2427.52 137,102 2428.15 .189 .233 1,118
1937 . Jam—Apr. 264,946 239,274 25,672 3323:: 2424.12 153,975 2431.80 .097 .107 487
. , 2384.10
1939 Now—Apr. 22,826 18,513 4,313 2396 26 2390.18 19,935 2390.86 .189 .233 355
2387.96 .
1940 Jam—Mar. 49,777 40,906 8,871 2407 94 2397.95 35,185 2400.03 .178 .217 444
1941 Man-June 359,660 312.877 46,783 3:322: 2481.00 578,220 2482.04 .130 .150 1,811
1942 Sept—Mar. 215,359 190,272 25,087 32333 2495.39 753,257 2495.24 .116 .132 1,903
1946 Jan—Feb. 17,718 14,577 3,141 3233:; 2396.23 21,236 2396.69 .177 .215 360
1948 . . Jam-Apr. 18,858 12,962 5,896 gggggg’ 2389.18 10,521 2390.08 .313 .455 479
1949 Dec—Apr. 312,140 251,306 60.834 Sjgg'ig 2416.94 129,452 2430.58 .195 .242 901
1952 . .. . Dec—Mar. 196,070 155.216 40.854 3233:: 2409.94 80,936 2420.16 .208 .263 768
. r ..
1953 . Jam-Mar, 14.686 10.544 4,142 3:26;: 2390.59 9,690 2391.44 .282 .393 409
2383.19 .
1954 Mar. 51,082 36.336 14.746 2407 50 2395.34 21,043 2394.74 .289 .406 607
2382.50 .
1957 Jan-Feb. 11.715 7,422 4.293 2392 86 2387.68 5,795 2389.00 .366 .578 414
1958 Feb-Apr. 204,404 167,062 37,342 3213(3): 2430.17 133.899 2433.59 .183 .224 1,090
1960 Nov.—Mar. 196,941 163,768 33,173 23222:; 2426.64 117,198 2430.93 .168 .203 891
1962 ........ N0v.—Apr. 181,480 141.049 40,431 3233;; 2414.70 77,624 2421.96 .223 .287 897
1963 Oct—Dec. 39,560 30.557 9,003 3(1):; 2405.42 29,343 2406.40 .228 .295 657
. , 2407.47 4
1965 Jam—Apr. 39,904 31,521 8,383 241851 241299 45,550 2413.46 .210 .266 799
1966 . Nov.—May 520,735 414.683 106.052 3233;; 2430.24 210,140 2448.77 .204 .256 1,260
1968 Feb—May 320,492 279,289 41,203 3:332: 2478.15 488,088 2479.03 .129 .148 1,700

 

for 1941 is an example of this condition. This error is
compensated, however, by an error of equal magni-
tude but of opposite sign during the following year(s)
and is of no significance in the 41-year mean annual
ET rate.

Determination of the size of the “exposed surface
area of the reservoir” is a prerequisite to computing
the ET by depth. The reservoir area at 2,525 ft (770 m)
is 19,925 acres (8,064 hmz). This area excludes ap—
proximately 60 acres (24 hm?) which lie within the
reservoir boundary but are upstream from the Gila
River Calva station. Added to the reservoir area,
however, are 925 acres (374 hm?) in the San Carlos
River flood plain between the Peridot gaging station
and the reservoir boundary, giving a total of 20,850
acres (8,438 hm?) as the maximum area possible for
ET loss. At any specific time, the exposed surface

 

area available for E Tloss is 20,850 acres (8,438 hm2),
less the lake surface area.

Surface conditions on the exposed area ranged
from open bodies of shallow water to dense phreato-
phytes and from wet to very dry soil. Optimum
surface conditions for high ET exist over a large
reservoir area following a major lake stage recession
such as occurred in 1942—45.

The computed volume of monthly ET was divided
by the mean monthly exposed area providing a value
of monthly ET depth. Water year totals of these
monthly ET depths are listed in table 19 and are
plotted in figure 34. The computed mean annual
depth of ET was 1.47 ft (0.448 In).

For each month, 41 values of E T depths were avail-
able from the water budget. All monthly values were
used to indicate the most common range in ET for

 

TABLE 16.—San Carlos Reservoir elevation-capacity tables of usable bank storage, usable surface-water storage,

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

Period 1931—47

storage

Cumulative usable storage, in acre—ft

Period 1948—71

N29

and total usable

Cumulative usable storage, in acre-ft

 

Elevation Bank Surface water Total Bank Surface water Total
(fl) SB SR ST SB SR ST

0 2 0 0 0 3 0 0
770 1,735 2,505 1,006 660 1,666
2,520 6,668 9,188 3,269 2,845 6,114
4,483 13,249 17,732 5,769 6,728 12,497
6,670 22,041 28,711 8,531 12,865 21,396
9,095 33,468 42,563 1 1,581 21,749 33,330
11,795 46,941 58,736 14,931 33,308 48,239
14,820 62,474 77,294 18,581 47,343 65,924
18,185 80,102 98,287 22,531 63,284 85,815
21,945 100,411 122,356 26,806 81,217 108,023
26,095 124,115 150,210 31,431 102,202 133,633
30,670 150,760 181,430 36,431 126,269 162,700
35,720 180,020 215,740 41,831 153,572 195,403
41,270 212,197 253,467 47,606 184,059 231,665
47,295 247.463 294,758 53,706 217,811 271,517
53,720 285,832 339,552 60,131 254,977 315,108
60,495 327,993 388,488 66,906 295,976 362,882
67,620 374,278 441,898 74,031 341,707 415,738
75,120 424,606 499,726 81,531 892,073 473,604
83,020 478,736 561,756 89,431 445,822 535,253
91,320 536,141 627,461 97,731 502,827 600,558
100,020 597,432 697,452 106,431 503,363 669,794
109,145 662,965 772,110 115,556 627,710 743,266
118,720 732,575 851,295 125,131 696,272 821,403
128,770 806,491 935,261 135,181 770,200 905,381
139,345 884,898 1,024,243 145,756 848,607 994,363
150,470 967,746 1,118,216 156,881 931,456 1,088,337
152,750 984,874 1,137,624 159,161 948,584 1,107,745
162,170 1,055,286 1,217,456 168,581 1,018,996 1,187,577

 

 

 

‘Elevation of zero usable storage.
2From 1947 tables of usable surface storage.
“From 1966 tables of usable surface storage.

CAPACITY, IN CUBIC HECTOMETERS

 

 

 
  
      

O 200 400 600 800 1000 1200 1400

2520 . I I ‘ I I ‘ I I‘ I ‘ I ‘ j
19 C '
2510 (From 66 apacrty Survey) , 765
2500 .
2490 Usable surface-water 7 760
storage capacity U

L sable total
2480 / storage capacity r 755
2470 ' ,

2460 ..

2450 ’

2440 7

2430

2420

2410 -"

Shaded area equals usable
bank‘storage capacity

750

745

740

' 735

 

2400 '
. 730

 

2390 '

ELEVATION, IN FEET ABOVE MEAN SEA LEVEL
ELEVATION, IN METERS ABOVE MEAN SEA LEVEL

 

 

 

2380 I I I ' l I I I I I
0 1 2 3 4 5 6 7 8 9 10 11 12

CAPACITY, IN 100,000 ACRE-FEET

FIGURE 31,—Elevation-capacity relations for usable surface-
water storage, usable bank storage, and total usable storage
in 1966, for San Carlos Reservoir.

each month, to show the monthly extremes, and to
illustrate the seasonal ET trend. In figure 35, the

range of the central two-thirds of the values for each
month is bracketed. This approximately corresponds
to the number of values included in one standard
deviation. The seven values above and seven below
the bracketed range were plotted for each month. The
monthly medians have also been identified in figure
35. The occurrences of extreme values of computed
ET usually coincide with periods of vigorous change
in one or more water-budget components. An un-
realistic extreme resulted when bank storage had not
adjusted to the change in stage, although changes in
lake stage and changes in bank storage were as-
sumed simultaneous in water-budget computations.
The mean annual ET of the monthly medians in
figure 35 is 1.45 ft (0.442 m). This value corresponds
to the mean of the annual ET values in figure 34,
computed as 1.47 ft (0.448 m).

The seasonal trend of E Tis obvious in figure 35 in
spite of the large scatter exhibited. After excluding
extreme values (outliers) the data were examined for
possible changes in rates over the life of the reservoir
and for obtaining more realistic values of mean
monthly ET depths. The monthly data were sepa-
rated into 4 periods of 10 years each, beginning with
1931. Within a 10—year period, the maximum and
minimum extremes were omitted from the 10 values
for each month, and a mean monthly ET was

 

obtained from the 8 remaining values. Figure 36

 

N30

GILA RIVER PHREATOPHYTE PROJECT

TABLE l7.—Summations of San Carlos Reservoir inflow-outflow components by water year, 1931--71

 

Inflows for year in acre-ft

Outflows for year in acre-ft

 

 

 

Gila San Carlos Gila Total
Water year River River Precipitation Ground water Tributary Total inflow River Evaporation outflow‘
289,917 36,677 5,850 7,994 3,580 344,018 223,920 23,008 246,928
442,175 51,950 9,565 8,015 1,760 513,465 256,150 44,455 300,605
149,072 16,802 7,399 7,994 2,460 183,727 334,910 36,244 371,154
160,085 13,733 1,895 7,994 2,660 186,368 184,660 15,416 200,076
149,423 87,670 6,684 7,994 4,020 255,791 184,553 21,554 206,107
150,165 44,559 3,659 8,015 1,760 208,157 228,090 22,162 250,252
317,944 46,630 3,717 7,994 1,760 378,046 279,621 29,436 309,057
106,170 15,713 2,505 7,994 1,760 134,142 189,720 14,321 204,041
91,500 18,378 1,362 7,994 1,760 120,994 102,426 9,365 111,791
158,275 15,858 1,891 8,015 1,760 185,799 160,720 10,323 171,043
803,991 201,192 22,378 7,994 6,990 1,042,546 215,724 54,381 270,105
314,222 25,273 16,250 7,994 2,410 366,150 352,030 74,465 426,495
102,531 30,191 10,954 7,994 2,200 153,870 365,140 60,302 425,442
80,725 13,161 7,512 8,015 6,070 115,483 297,430 34,186 331,616
131,705 17,038 3,863 7,994 3,010 163,610 216,290 20,734 237,024
55,627 15,233 1,245 7,994 4,850 84,949 74,550 7,049 81,599
45,607 11,141 796 7,994 3,550 69,088 59,983 5,467 65,450
63,511 10,390 620 8,015 2,530 85,066 68,270 5,913 74,183
422,114 23,192 3,863 7,994 2,680 459,843 254,061 29,914 283,975
37,441 5,908 2,709 7,994 2,530 56,583 148,860 13,692 162,552
35,304 9,134 808 7,994 2,930 56,170 36,347 4,866 41,213
191,624 80,840 4,658 8,015 2,620 287,758 229,435 22,710 252,145
42,937 8,384 861 7,994 1,760 61,936 49,319 5,983 55,302
116,372 42,254 2,345 7,994 7,430 176,396 68,964 10,579 79,543
123,593 26,899 3,194 7,994 6,960 168,639 97,710 12,342 110,052
20,858 13,886 2,053 8,015 1,760 46,572 109,062 12,572 121,634
128,264 9,136 586 7,994 1,760 147,740 52,123 5,288 57,411
296,372 47,883 5,676 7,994 6,420 364,345 243,220 28,1 18 271,338
97,574 6,419 2,290 7,994 5,920 120,197 147,870 15,965 163,835
193,693 64,990 3,809 8,015 3,940 274,447 256,896 26,884 283,780
45,499 6,189 461 7,994 5,080 65,223 25,312 2,449 27,761
275,405 33,567 2,650 7,994 2,980 322,596 246,730 21,971 268,701
175,069 34,123 3,340 7,994 3,580 224,106 141,430 18,074 159,504
94,351 10,301 1,697 8,015 7,620 121,984 107,170 12,697 119,867
90,948 24,550 2,803 7,994 1,940 128,234 121,999 13,688 135,687
533,205 119,391 13,948 7,994 8,580 683,118 226,476 53,587 280,063
148,375 11,295 7,697 7,994 5,800 181,162 255,250 44,067 299,317
579,171 72,822 12,339 8,015 1,760 674,107 281,340 70,871 352,211
60,545 15,599 11,171 7,994 2,480 97,789 315,376 59,744 375,120
31,214 12,502 5,889 7,994 3,560 61,159 218,860 28,413 247,273
59,775 19,247 861 7,994 3,920 91,786 57,463 5,777 63,240
7,412,348 1,370,100 203,853 327,964 148,890 9,463,160 7,485,460 1,009,032 8,494,492

 

 

IDoes not include evapotranspiration losses.

shows the resulting monthly means for each 10-year
period. The seasonal trend is apparent for all four
periods, but no obvious changes in ET rates can be
detected during the life of the reservoir.

The average of the annual mean ET values com—
puted from these four periods is 1.51 ft (0.460 m),
which is slightly higher than that obtained when the
maximum and minimum extremes are included in
the computation. The monthly mean and median ET
values and the water-year means from the data
shown in figures 35 and 36 are listed in table 20.

Two characteristics can be noted about the sea-
sonal trends in the computed ET values shown in
figures 35 and 36. First, the computed ET is essen-
tially zero during the winter months from December
through February in part because zero ET was used
in the development of the bank-storage ratings.
Second, the reservoir surface-water storage normally
increases during August, and the bank storage
response to this increase does not approach equi-
librium until sometime in September. As a result,

 

August ET was often underestimated and September
ET overestimated.

Vegetation increased on the exposed areas of the
reservoir flood plain, especially following the 1941—
42 floods (Turner, 1974, fig. 4). If an increasing trend
in ET could have been conclusively shown, changes
in ET would have been correlated with the increase
in vegetation. Data used in figure 34 were examined
for an indication of increased ET, since no increasing
trend was evident in figure 36. Mean annual ET was
computed for the selected periods shown in figure 34.
Only those periods of least stage change were used,
thereby eliminating periods when soil-moisture con-
tent was high over large areas. Thus, a greater per-
centage of the total ET should have been through
transpiration and not by evaporation from exposed
areas of bare ground. Mean annual E T depths were
computed as 0.93 ft (0.28 m) from 1934 to 1940, 1.07 ft
(0.33 m) from 1945 to 1957, and 1.48 ft (0.45 In) from
1959 to 1965. This shows an apparent increase in ET,
but as figure 34 indicates, the range in annual

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

Precipitation 2.2 percent
Tributary flow 16 percent
Ground water 3.5 percent

   
  

San Carlos River
14.5 percent

Gila River
78.2 percent

INFLOW

Evaporation 10.5 percent

Evapotranspiration
11.3 percent

Gila River
78.2 percent

OUTFLOW

FIGURE 32.—Relative magnitude of inflow and outflow water-
budget components.

amounts during each period is too great to state
definitely that the increase is related to vegetation
changes.

The total exposed surface area of the reservoir
extends outside the flood plain because much of this
surface area was occasionally inundated. Generally,
however, the highest rates of ET are from the flood-
plain portion of the area. Use of the total exposed
surface area prevented making a more accurate
evaluation of vegetative ET losses.

Flood-plain vegetation in the reservoir area was
removed beginning in the 1967 water year. A com-
parison of monthly before-clearing and after-
clearing data was made for periods 1961—66 and
1967—71 to evaluate vegetative consumptive use
(transpiration). The results of these comparisons
were inconclusive, however, because the ET rates

 

N31

were too greatly affected by the large fluctuations
in lake stage from 1966 to 1971.

Although precipitation falling on the exposed area
of the reservoir is a reservoir inflow, a distinction is
made between the relation of this inflow to ET and
the relation of inflow from other sources to ET. The
ET loss computed by the water budget (table 18)
represents a loss in usable reservoir water. Normal-
ly, only a minimal amount of the input from direct
precipitation on the exposed areas is ever a part of
“usable” contents. The fact that precipitation is
distributed on the surface enhances the possibility
for immediate and total ET. Vaporization of the
precipitation depletes the energy available for vapor-
izing water from other sources, but this effect is
temporary in arid regions because of the infrequent
and limited quantity of precipitation.

All precipitation on the exposed ground area of the
reservoir is an additional ETloss. This added ET is
assumed equal to precipitation measured at the San
Carlos Reservoir weather station. ET from this
precipitation source was added to the ET computed
from the water budget of the reservoir (table 18) to
give the total ET from the reservoir water-surface
area and adjacent exposed ground area. The average
annual total ET for the 1931—71 period is 2.62 ft (0.80
m). A comparison of this value with the computed
water-budget ET values in table 20 indicates that
precipitation on the exposed area of the reservoir
contributed an average of about 1.2 ft (0.37 m) per
year to the total ET.

DEVELOPMENT OF SURFACE-WATER
STORAGE-CAPACITY RATINGS

SIMULATION OF THE SEDIMENT DEPOSITIONAL PROCESS

Interpolations of capacity changes between capac-
ity surveys were made by using a procedure which
included simulating the sediment depositional
process in the San Carlos Reservoir. The simulation
procedure was structured about three basic phases of
the depositional process:

1. Sediment inflow

2. Sediment distribution

3. Sediment compaction
Equations were developed to represent each phase.
The “best fit” values of variables in the equations
were selected by minimizing the difference between
the volumes of sediment measured and estimated.

Measurements of sediment accumulation were
available from only four separate periods between
1929 and 1966. However, volumes of sediment depos-
ited within 5-foot—elevation increments were defined
from each survey, and these volumes provided the

 

 

N32

GILA RIVER PHREATOPHYTE PROJECT

TABLE 18.—Water-budget summations

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Watersurface Surfacewater Surface-water Change in Evapo-
elevation storage storage change Bank storage bank storage Inflow Outflow‘ transpiration2
Water year (ft) (acre-ft) (acre-ft) (acre-ft) (acre-ft) (acreft) (acreft) (acre-ft)
1931 31:32: £233: -24,040 3:23: -5204 344,018 246,928 67,846
1932 ' * -146,313 ' _25,132 513,467 300,605 41,417
1933 245353 309’7” 199 033 51331 33 441 183 727 371 154 45 097
1934 7 7 7 242026 ”0’631 36,984 “3'39” 54% 186,368 200,076 28,751
1935 J 7 7 7 241135 73'6“ 414,803 12’9” 6616 255,791 206,107 8,266
1936 H 7 I 7 7 242178 108'45‘) 63,472) 19530 11,647 208,153 250252 33024
1937 H A} 240250 “‘980 416,230 ”82 6969 378,0:16 309057 13,769
1933 VVVVVVV I 5 241804 91'2“) 75,072 16'872 14,961 134,142 204,041 20,134
1939 A} 7 238826 ““38 1327 1’9“ 1357 120,994 111,791 8,519
1940 .. 238928 16’465 5189 2268 980 185800 171943 20926
1941 7 238648 “'276 —685'390 1'288 -110 998 1042546 270105 -23’947
1942 i ' rrrrrrrrrrr 2491.64 696,666 102,486 112286 14,023 366,150 426,495 56,164
1943 .. 7 248399 594’18‘) 264,618 98263 38,825 153,870 425,442 31,871
1944 245922 329562 223942 59438 37,718 115,484 331,616 45,528
1945 7777 2424'70 105'620 79,042 21'720 15,194 163,610 237,024 20,822
1946 V 239967 26578 11,367 6526 2,561 84,949 81,599 17,278
1947 7 i U N 2393'“ 15’2“ 8,951 3965 2,497 69,088 63450 15,037
1948 N 238602 6260 2,718 L468 1,384 85067 72,193 14,986
1949 i N N 238275 3’542 405,910 83 ~28‘360 459843 283975 41 ’598
1950 H N 242677 109’452 105758 281443 28,109 56,583 162,552 27,898
1951 V I 238336 3’694 '634 334 ,226 56,170 41,213 15,817
1952 2382's} 3'060 -439 ‘08 ~488 287,759 252,145 34,687
1953 332:2: 3’22? 918 5:98 586 61 :936 55302 8:138
1954 ‘ ‘ —56,443 —18,506 176,396 79,543 21,904
1955 241491 59'024 —22 312 18515 —5 503 168 639 110 052 30 272
1956 V 7 2421'“ 81'836 79,332 24919 24,120 46,573 121,634 28,891
1957 ,, , ,. . . :2382'30 2‘0“ 556,107 '10} _19,124 147,740 57,411 15,098
1958 241556 58'1” -39 877 g: 0:3 —9 161 364 345 271 333 43 970
2426.49 97,988 ‘ ,1 4 1 1 . r , 1
1:2: 22:32 3:2: 321:; 3:33 33?
1961 ' 777777 238239 133° -10'090 1‘“ 6462 65,223 27,761 20,910
1962 .. V i 239651 “'920 -2'144 6'603 .1138 322,596 268,701 50,613
1963 239857 ”'06“ {17,596 7’7“ ~10’962 224,106 159,504 16,024
1964 7 V 241518 51560 18,582 mm} 4,549 121,985 119,867 25,250
1965 240837 33’078 28,609 ”'174 10,320 128,234 135,687 31,476
1966 H i 2391‘” 4'469 —3122260 3354 «66,230 683,118 280,063 24,565
1967 246223 316’729 120990 701084 19,904 181,162 299,317 22,738
1968 I 7 i 7 2447'” 195’739 —275‘633 50’180 43,998 674,108 352,211 2,265
1969 M, 247186 471372 266,160 94’179 41,912 97,789 375,120 30,741
1970 244832 205’le 184,431 52266 40,966 61,159 247,273 39,283
2404.54 20,781 ’ 11,300 ’ ' , 1 1
1971 241227 37,542 —16,761 16 588 —5,288 91,786 63,240 6,498
1Does not include evapotranspiration losses.
2Excludes precipitation from exposed area of reservoir.
70 increment is considered a separate surface-water
’— n
E E 60 .k :3 3 storage reserv01r.
- ‘m . . . .
gfi so - —_ so 55 The sedlment inflow phase of the s1mulat10n pro-
1'—"1 40 r 50 1:!- cedure rovided estimates of wei ht of incomin
< 0 40 < L“ . . . . .
11c“ 30 —’ 0 egg sedlment by assuming that sedlment inflow 1s a
- u— — 3 n. . . . . .
310 20 g 20 $5 functlon of streamflow. The 11m1tat10n of thls as-
Z 0') Lu . . .
$2) 10 7 10 g: sumptlon IS the large range of sediment concentra-
1535 0 ‘5 0 53 tions which occur for any given water discharge.
_1 a o I l
3,3 ~10 5 7 a: 3: 3 However, for long time periods 1t 1s assumed that use
>1 '20 ' , - 51 of mean sediment concentrations is acce table if
LU l— 30 9
‘301111w1111mmHHHHHIHHHIHH adjustments are made for seasonal dlfferences 1n
“ O O O 0“ .
g g 3 g 53 concentratlon.
‘— I— I- F Fr-

WATER YEAR

FIGURE 33.—-Annual evapotranspiration from exposed area of
reservoir computed by reservoir water budget.

basis for developing and testing the simulation of the
sediment depositional process. In the simulation,
each storage unit identified by a 5-foot-elevation

 

According to the US. Army Corps of Engineers
(1914, par. 92, p. 30), the total sediment discharge in
summer averaged about 2.5 percent, by weight, of the
discharge of the water-sediment mixture. In winter,
the ratio of discharges was approximately 0.5 per-
cent, so the summer-to-winter concentration ratio
was about 5:1. Burkham (1972, p. 8) derived a similar
ratio based on 1965-70 US. Geological Survey data

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

TABLE 19.—Total euapotranspiration (ET) for San Carlos Reser-
voir, by water year

 

ET from precipitation

 

 

 

 

ET as residual of exposed
of water budget reservoir surface Total ET
Water year (ft) (ft) (ft)
3.55 1.30 4.85
3.18 1.18 4.36
2.88 1.12 4.00
1.08 .54 1.62
.20 1.64 1.84
1.81 .97 2.78
.86 1.08 1.94
.97 .98 1.95
.29 .89 1.18
.93 1.04 1.97
—3.08 2.54 -.54
7.86 1.08 8.94
2.95 .98 3.93
2.71 1125 3.96
.97 1.13 2.10
.58 .96 1.54
.57 .79 1.36
.61 .77 1.38
2.41 1.14 3.55
1.42 .93 2.35
.63 .99 1.62
1.86 1.55 3.41
.31 .90 1.21
.66 1.60 2.26
1.20 1.17 2.37
1.48 .75 2.23
.61 .93 1.54
2.25 1.66 3.91
1.78 .98 2.76
1.66 1.28 2.94
.75 .92 1.67
2.69 1.11 3.80
.62 1.24 1.86
.91 1.08 1.99
1.54 1.17 2.71
2.31 2.26 4.57
1.13 1.11 2.24
.37 1.38 1.75
2.48 1.12 3.60
2.21 1.11 3.32
.05 .74 .79
11111111 11 11‘1 11 1111f'i'1
A EXPLANATION ’ 2-0
5—!
Mean annual ET depth 1,5
for period bracketed
1.0

 

' —0.5

 

EVAPOTRANSPIRATION,1N METERS

 

7 —1.0
11 11 11 ‘17 1 111 111 1

O O 0“

L0 QB 51‘

01 m 0101

F _ ar-

 

WATER YEAR

FIGURE 34.—Annual depth of evapotranspiration from the
exposed surface of San Carlos Reservoir for water years 1931—71.

of suspended sediment for the Gila River station at
the head of Safford Valley, which is located about 50
mi (80 km) upstream from San Carlos Reservoir. The
change in this ratio from the Safford Valley site to

 

N33

 

 

 

 

 

 

 

 

 

5-01 1 1 1 1 1 1 1 1 1 1 1 1 1 1-52
1 04.4 21.7 1.20 o 1.4 —
1.0 , fl
— 0.30
0.9 _ 0
0.8 - ° —’ 0'25
0.7» ° 0 8 —
o — 0.20 w
E 0.6— ° ° 0" 5
Lu 0 o 0 0 ° 8 o o ,_
l 0.5 - ° 2 8 g -— 0 15 w
z 2 ° 0 8 g 2 g 0 - 2
‘. 0A-—° a o 2 0 g_ g
z e T 2‘
g 0.3_ o a g + _— 0.10 o
'3: a ., ° 2 ;
1:0.2—TS: 31 . . -005:
3) _ k T 2 2 l + 1’ < . E
z 0.1 1, l 1 m
< i + 1 1 2* 2 2 2
”C 0 #2 "13 . + 33 a2 a; 3 o — 0 <
1— g o 2., 2 tr
0 ° 0 1L 0 a I—
n_ -0.1— o . 0 ° 0 — O
<( 0 ° 8/2 A ° 2 n.
> °2 8 o — -0.05 <
M vo 2 — o .2 2 o — E
"0'3 ’ i 70.10
-O.4 — o _
_o.5 _ ° 0 a— —0.15
0
-09 a
—0.5 o _
1 :JS1 ~13 -03 —49 2
-501 1 1- 1 1’ 2‘1 1 I 1 1 1 T _1_52
' ' ..' c > w > ' '
‘5 S a”: 5 8 m a m g 3 E” ‘a
O z D ‘1 u. 2 < E 3 '1 <1; (g
MONTH

EXPLANATION

'4° Number indicates actual evapotranspiration, in feet
Range of 2/3 of monthly values
+ Median of monthly values

0 Monthly values outside bracketed range

2 Number indicates number of values

FIGURE 35.——Computed monthly evapotranspiration at San
Carlos Reservoir for water years 1931—71.

.0
1:.
1
1
1

 

   

 

 

 

 

 

Z‘ 2
o o
- 0.3 , _
e a
<
I 0.2 gin
._.|— —CE
aw ‘ 9*“
LLI
31.1.01 <Z[E
m E 0 HZE
5 “E
a —01» #7—0022
g 7—004<
w ‘02 1 1 1 1 1 1 1 1 : 1 1 'ODGM
A >- d C" n‘ u i. >- a: > ca o-‘a
o o 0 (u a) m Q <0 : — Q
0 Z O " u. 2 <( E 3, Pa 2 3’,
MONTH
EXPLANATION
rrrrrrrrrrrrrrrrr 1931—40
-" ---1941—50
—————— 1951460
1961—70

FIGURE 36.—Mean monthly evapotranspiration depths computed
for 10-year periods.

the reservoir is not known, nor is it known how the
ratio would be affected by using total sediment

 

 

N34

GILA RIVER PHREATOPHYTE PROJECT

TABLE 20.——Mean monthly evapotranspiration (ET) computed from 4 periods of 10 years each, and median monthly evapotranspiration
for 41 years

Computed monthly ET

(ft)

 

Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Annual

Mean monthly

from 4 periods

of 10 years

each .................. 0.123 0.092 —0.009 0.003 0.004 0.062 0.168 0.212 0.274 0.239 0.166 0.176 1,510
Median of

41 monthly

values .109 .071 .044 — .018 .005 .099 .142 .193 .300 .198 .155 .150 1.448

 

discharge instead of solely the suspended sediment
discharge. Estimates of the magnitude of deposition
in San Carlos Reservoir were made with ratios of
summer-to-winter concentrations in the range of 1:1
to 10:1 for the four periods prior to implementing the
total simulation procedure. Comparison between
estimates and measurements of deposition showed
that results using the suggested 5:1 concentration
ratio were only slightly better than those using a 1:1
ratio.

A general equation relating the weight of sediment
discharged by a stream to the stream inflow volume
during a fixed time interval is

S=CQ , (6)

where S is the weight of sediment discharged, C is the
mean sediment concentration expressed as the
weight of sediment per unit volume of inflow, and Q
is the volume of stream inflow. The weight of sedi-
ment retained by the reservoir, Si, is equal to S times
the trap efficiency, E. The weight of trapped sedi-
ment is therefore

Si=SE=ECQ. (7)

When equation 7 is expanded to include seasonal
terms for C and Q, it becomes

Si 2 13(0wa + Cst) ,

where Cw and 'C8 are the mean winter and summer
sediment concentrations, respectively, and Qw and
Qs are the winter and summer streamflow summa-
tions, respectively.

The ratio r, where r = 03/ C , was one of the
variables for which an optimum value was sought in
defining the sediment deposition equations. Substi-
tuting r into equation 8 gives

3, = ECw(Qw + res), <9)

which eliminates Cs from the equation. A value of r
was assumed for the initial trial through the simula-
tion procedure. The value of r was adjusted in subse-

(8)

quent trials to improve the results from the simula—'

tion procedure.
E and Cw were assigned values of unity because

 

both are constant during each trial of the simulation
which compares the estimated with the measured
sediment deposition. Equation 9 was modified for
these assigned values, and the resulting equation
provided estimates, designated S , which were pro—
portional to Si- The modified equation is

sy = rQs + Qw . (10)

Each daily streamflow amount was assigned to
the 5-foot-elevation increment (incremental storage
reservoir) into which streamflow occurred; the proper
incremental reservoir was determined by the lake
stage at the end of the day. Daily streamflow data
was summed according to water year, season of the
year, and incremental reservoir. The proportional
sediment weight, Sy, for each year and each incre-
mental storage reservoir was computed by inserting
the appropriate Qs and Qw sums into equation 10.
Table 21 shows a generalized chart of Syi,j into a

reservoir, where i = 1 to n designates the incremental
storage reservoirs and j = 1 to m designates the water
years.

In table 21, the proportional sediment weight for
any given year is determined by summing the Syi, j

values in the column corresponding to that year.
Similarly, a summation of Syi j values for a partic-

ular row of the table is the proportional sediment
weight for an elevation-increment of storage. The
total Sy for a period is therefore

TABLE 21.—Chart of notation used to identify time and inflow
location of computed proportional sediment weights, SYi j

 

 

 

Water year,j

1 2 0 I m -l m
3’ 1 Syn Sy1,2 ' ' Sy1,m -1 5mm
F1 V V
g; 2 S321 by2,2 ' ' Sy2,m—1 Sy2,m
'3 5 o o o o o o a
5 E . . . I I . .
E a; _ .
g “ ” 1 Syn—1,1 Syn —1,2 ' ' Syn—1, m—l b3’n-1,m
::
T n Syn,1 Syn,2 . . Syn,m—1 Symm

 

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929-71

n m
831:: 2 SyiJ' (11)
i= 1 j = 1

The second phase in the sediment depositional
process is the distribution of sediment inflow within
a reservoir. As a stream merges with pooled reservoir
water, water velocities decrease, resulting in deposi-
tion of sediment.

In this preliminary simulation the reservoir is
assumed uniform in width and bottom configura-
tion. Another assumption is that the larger sediment
particles are deposited immediately upon entry into
the surface-water storage pool of the reservoir.

The weight of larger sediment particles to the total
sediment weight was defined as a. Selection of an
optimum a was done in the simulation procedure by
comparing estimated with measured sediment depo-
sition. The S yi, j values were divided into two parts by

use of the equation .
Sy=a Sy +(1.0—a)Sy. (12)

The quantity (1.0 - a)Sy is the part of Sy made up of
the smaller (suspended) sediment particles. Some of
this quantity of smaller particles is assumed to be
deposited within the incremental storage reservoir
where the stream enters the storage pool, and the
remainder is assumed to move to lower incremental
reservoirs and is then deposited.

The computed weight of smaller particles, (1.0 —
a)Sy, was designated S5 to reduce equation symbol-
ism. 83 was exponentially distributed over the dis-
tance from the point of inflow to a point where depo-
sition is considered complete. Referring to figure 37,
the deposition of suspended sediment was distri-
buted from point A, the point of inflow, to point C,
over the distance DA. B is a point along DA, and DB
is the distance from point B to point C. The ratio of
the weight of suspended sediment passing point B to

D x
that which passed point A is (BE) , where x is the
exponent of distribution. A
In the simulation procedure, the proportional
weight of suspended sediment passing B is equal to

D x
Ss (5‘5). The difference between quantities pass-
A

ing A and B is the proportional weight of suspended
sediment deposited between A and B, designated
SSB' Accordingly,

DB

x
SSB=SS[1.0— 314)]

(13)

 

N35

 

D K1 '04))

A 5
Water level ,,— I
BBB—i ’ A .
B Suspended sediment depOSIted
in reservoir

 

 

 

C

FIGURE 37.—Sketch identifying terms used in simulation of
suspended sediment distribution.

The optimum values for the distance DA and the dis-
tribution exponent x were found in the procedure
which selected best estimates of sediment deposition.

To simulate distribution of suspended sediment
into incremental reservoirs, DA and DB of equa-
tion 13 were redesignated DAi and DBk, respec-

tively. The subscript i represents the incremental
reservoir into which sediment entered the reservoir
pool, and k represents the incremental reservoir for
which deposition is being computed. Point A is the
upstream limit of reservoir i, and point B is the up-
stream limit of reservoir k. Syi is the computed

quantity proportional to sediment weight entering
incremental reservoir 1' at point A, so the suspended
sediment fraction of Syi is 531‘. The suspended sedi-

ment which passes into reservoir k at point B is con-
D x
31:

sequently computed as Ssi( ) ~ The sediment

DA i
passing into the next lower incremental reservoir,
x

D
B k + 1 l
—— ) ., The difference between

k+1is ss(
’ lD
Ai

these two amounts, S; k .is proportional to the sus-

pended sediment weight deposited in incremental
reservoir k and is given as

x x
DB DB
3 =5 (.D_.k)-(D_k;+_1) .(14)
3k Si Al. Ai
The 531' total for each incremental reservoir was

distributed by Ssk amounts into the appropriate
incremental reservoirs by equation 14. Distribution
of Ssi started with k=1 and continued through suc-
cessively lower incremental reservoirs to either the
lowest incremental reservoir or to the end of DAi (at
point C ), whichever came first. Incremental reser-
voirs were numbered from 1, for the upper reservoir,
to n, for the lowest, and Ssi was distributed into all

incremental reservoirs from i = 1 to i = n.

 

 

N36
If the optimal value selected for 014i was greater

than the distance from point A to the dam, a part of
the Ssi quantity could not be distributed by equation

14. This remaining quantity was distributed uni-
formly over the distance from point A to the dam as
shown in figure 38.

For San Carlos Reservoir, distribution was made
by water year. Thus, within each of the four periods,
the deposition was categorized by the incremental
storage reservoir of deposition and by water year.
For incremental reservoir k and for any water year j,
designation of the computed amount deposited
was Sd k, J

The simulating process continues with the com-
paction phase after the sediment is distributed and
deposited. Lane and Koelzer (1943) presented a com-
paction equation to estimate unit weight .of sedi-
ments at a specified time following deposition. This
equation requires knowing the “in place” composi-
tion of the sediment and its specific weight 1 year
after deposition. This information was unavailable
for San Carlos Reservoir. However, the Lane and
Koelzer equation was applied to the reservoir using
estimates of percentages of sand, silt, and clay
deposited after 1 year and an estimate of the specific
weight of the deposit. The amount of compaction esti-
mated by this method was found to be small com-
pared to the probable error in the simulation of sedi-
ment deposition. Also, the addition of a compaction
equation to the simulation procedure resulted in un-
acceptable parameter values in the sediment inflow
and distribution equations. For these reasons, the
compaction of reservoir sediments was deleted from
further computations.

With deletion of the compaction phase from the
simulation, the results of only the sediment inflow
and sediment distribution phases were used in mak—
ing estimates of sediment deposited. The computed
proportional weights for distributed sediments—the
Sdk, j amounts—were summed for each incremental

reservoir and for each of the periods including the
1929—66 period. The incremental reservoir sums are
designated SRk, and the total SRk amount of a

period is S R- 813 can be expressed as

n 771
SR = 2 Z Sdk, j , (15>
k=1j=1

where the range of incremental reservoirs is from k=1
to k=n and the water years range from j = 1 to j = m.
The relation between simulated deposition and

 

GILA RIVER PHREA’i‘OPHYTE PROJECT

‘ i
i DA 1
L— Distance from inflow point to dam fli
l i
l i

 

    

Water level

 
 
 
 

Distribution of sediment
by equation 14

Uniform distribution of remaining
sediment after distribution of
first part by equation 14

FIGURE 38.—Sketch showing distribution of suspended sediment,
831., when distance DA was computed as greater than the dis-
tance from inflow point (A) to the dam.

measured deposition was established for each period
by the following equation:

SM=RATIO (SR), (16)

where SM is the total volume of sediment deposited
in the period. RATIO is a proportionality constant
whose value is affected by E and Cw of equation 9,
which had been set equal to unity for the computa-
tion procedure. RATIO also includes a unit conver-
sion to relate the measured volume of SM to SR. Each
repetition of the simulation produced a different
value for RATIO.

The estimate of absolute volume of deposited
sediment, Sek’ was made for each 5-foot incremental

reservoir by multiplying the periods RATIO times
the S Rk of the incremental reservoir:

Sek =RATIO (SRk). (17)

The value of one or more of the variables, r, a, x, or
DA, was changed slightly for each trial of the
simulation procedure. The optimum values of vari-
ables were naturally those which produced the best
sediment estimates. A listing of the variables and
optimum values are shown in table 22 for the four
periods between capacity surveys and for the 1929—
66 period.

TABLE 22.—0ptimum values of variables from simulation of the
sediment depositional procedure,

 

 

 

 

 

 

Period number Variables

and dates r a x D A
No. 1, 1929-35 ............................ 13.20 0.25 1.30 60,000 ft
No. 2, 1935-37 1.80 .40 .88 40,000 ft
No. 3, 1937-47 2.35 .38 .82 72,000 ft
No. 4, 1947—66 1.35 .55 .82 16,000 ft
No. 5, 1929—66 4.65 .35 1.60 56,000 ft
Symbols:

r is the seasonal ratio of summer to winter sediment concentration.

a is the percent, by weight, of total sediment load deposited when streamflow
reaches the reservoir pool.

x is the exponent of the expression for the distribution of suspended sediment.

DA is the distance along the reservoir centerline from the point of inflow to the point
where no sediment remains to be deposited.

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

INTERPRETATION OF SIMULATION RESULTS

The range of the summer-to-winter concentrations
(r = 135—132) in table 22 is not unexpected for the
time periods and geographical area considered but
may not be meaningful because varying discharge,
velocities, and so forth, were not considered. A
comparison of differences between the summer and
winter suspended sediment concentrations at sev-
eral southern Arizona locations shows that, for short
periods, r can vary more than that shown in table 22.
The period 5 data suggest that the ratio of summer-to-
vvinter concentrations for the 38-year period 1929—66
actually approached the 5:1 value discussed on page
32.

The variable a ranges from 0.25 to 0.55 in table 22
and averages 0.39, indicating that about 40 percent
of the total sediment was deposited near the entry of
the San Carlos and Gila Rivers into the reservoir
pool. This 40 percent probably is the approximate
bed material discharge of the San Carlos and Gila
Rivers.

It was assumed prior to determining an optimum
x, that x would be equal to, or greater than, 1.0; that
is, the rate of suspended sediment deposition would
be at least as great at the point of inflow as in any
other part of the reservoir. The selection of the
optimum x did not confirm this assumption for all
periods. The values of x determine the curvature of
the relations in figure 39. These relations show the
proportional distributions of sediment deposits
along the reservoir using the optimum values of a,
DA, and x determined by the simulation procedure.

DISTANCE DOWNSTREAM FROM POINT OF
INFLOW, lN THOUSANDS OF METERS

0 5 10 15

20 23
100

  
 

Period
Nov. 1928 to 1935

1935 to 1937
1937 to 1947
1947 to 1966
n»—'0 Nov. 1928 1101966

INFLOW DEPOSITED

 

 

20 —
x=0.82 is an exponent of the suspended
10 - sediment distribution, equation 14
0
o

CUMULATIVE PERCENT OF SEDIMENT

10,000 1
20 000 4
30 000 4
40 000

50,000 4
60 000 4
70 000 1
80 000

DISTANCE DOWNSTREAM FROM
POINT OF INFLOW, IN FEET

FIGURE 39.—Simulated sediment deposition, in percent,
downstream from point of inflow.

 

N37

In table 22 DA is an approximation of the average
distance over which sediment is deposited. For
shorter time periods DA can be considerably differ-
ent because the distance through which sediment is
transported is affected by the rate of inflow, the
volume of water in surface storage, constrictions in
reservoir, and so forth. It appears from optimizing
DA that sediment distribution sometimes occurred
throughout the reservoir, even when surface-water
storage was a great amount.

Sediment inflow at San Carlos Reservoir does not
appear to move far into the pooled water under low
streamflow conditions, as illustrated by period 4
results in figure 39. Inaccuracies in the storage
capacity ratings adversely affected all measure-
ments and estimates of sediment deposition. Also,
sediment distribution based on mean cross-sectional
velocities or daily streamflow volume rather than the
mean distance, DA, may improve the sediment
model. These indications emphasize the need for
more development of the sediment distribution
model.

Figure 40A shows a comparison between the vol-
umes of deposits measured and computed during
period 5 for all incremental reservoirs. In figure 40B
a comparison is made between cumulative volumes
of measured and computed sediment deposits for
period 5. The summations were determined by pro-
gressively adding the volume of deposit in an incre-
mental reservoir to the sum of the deposits in all
lower incremental reservoirs.

PROCEDURE TO DEVELOP RATINGS

Reservoir surface-water storage values for the
water budget of water years 1929, 1935, 1937, 1947,
and 1966 were obtained from surface-water capacity
ratings of the five surveys. Ratings for water years
1967 through 1971 were interpolated on the basis of
estimated volumes of sediment deposited from the
simulation procedure by using the parameter values
of period 5. For all other years, surface-water storage
ratings were developed by interpolating storage
changes between capacity surveys.

The development of a rating of surface-water
storage for each water year was begun by inserting
the optimum parameter values of table 22 into the
simulation procedure. Estimates of absolute sedi-
ment volumes, Sek, were obtained for all incremental

reservoirs during the periods between surveys. The
ratio of the measured to estimated volumes,
SMk/Sek, was computed for each incremental reser-

voir of each period. Table 23 lists the S Mk, Sek, and
SMk/Sek values for period 3.

 

 

N38

MEASURED VOLUME OF SEDIMENT
DEPOSITED, IN CUBIC

 

     

HECTOMETERS
0123456789102425
‘21 IIIII‘III I[I|IiIWHIT.‘[I
25
20»
24
19—

)
I

.-
0

co
ESTIMATED VOLUME OF SEDIMENT DEPOSITED,

IN CUBIC HECTOMETERS

Mubmmuooco

_\

 

 

 

   

ESTIMATED VOLUME OF SEDIMENT DEPOSITED
IN THOUSANDS OF ACRE FEET

,, I I I ,ALj‘ \ I 0
0 1 2 3 4 5 6 7 8 I 20 21
MEASURED VOLUME OF SEDIMENT
DEPOSITED, IN THOUSANDS
OF ACRE~FEET

MEASURED CUMULATIVE VOLUME OF
SEDIMENT DEPOSITED, IN CUBIC
HECTOMETERS

QOOO
oV—va

0
Lo
1 00 I_17__I_F_I_.Tm+J7__g_I_r_I_rLfiIT_Li

90

DEPOSITED, IN CUBIC HECTOMETERS

10

DEPOSITED, IN THOUSANDS OF ACRE-FEET
.‘ ..
O1 07
O O
ESTIMATED CUMULATIVE VOLUME OF SEDIMENT

o ._L_J_gI‘LL_ o
0 1O 20 30 4O 50 60 70 80 90 100

MEASURED CUMULATIVE VOLUME OF
SEDIMENT DEPOSITED, IN THOU—
SANDS OF ACRE-FEET

ESTIMATED CUMULATIVE VOLUME OF SEDIMENT

FIGURE 40.—A, Comparison between the volumes of deposits
measured and estimated for each incremental reservoir,
1929—66. B, Comparison between volumes of estimated and
measured deposits cumulative by 5-ft-elevation increments
upward from lowest part of reservoir, 1929—66 water years.

The sediment volume accumulated in each incre~
mental reservoir was also estimated in the simula-
tion from the beginning of a period to each subse-
quent year in the period. These volumes are desig-

 

GILA RIVER PHREATOPHYTE PROJECT

TABLE 23.—Volumes of sediment deposits measured (SMk) and
computed (Sek) for 1937—47, and the SMk/Sek ratios

 

 

 

  

Elevation S M S S / S
increment k ek Mk eh
(ft) (acre-ft) (acre-ft)

below 2380 5,162 5,160 1.00
2380—2385 _ 2,411 718 3.36
2385-2390 . 2,139 780 2.74
2390—2395 , 1,926 2,189 .88
2395-2400 . 1,312 861 1.52
2400-2405 , 381 740 1.19
2405—2410 . 976 695 1.40
2410-2415 , 703 854 .82
2415—2420 ,. 430 428 1.01
2420-2425 ,, 300 ‘ 1,042 .29
2425-2430 .. 128 1,586 .08
2430—2435 ,, 168 661 .25
2435—2440 ., 590 594 .99
2440-2445 ., 795 572 1.39
2445—2450 .. 189 462 .41
2450—2455 .. 809 515 1.57
2455—2460 . 403 714 .56
2460—2465 .. 430 439 .98
2465—2470 .. 469 418 1.12
2470—2475 .. 107 ' 321 .33
2475—2480 304 298 1.02
2480-2485 175 760 .23
Above 2485 0 0 —

 

 

nated Se k, j' Completion of these estimates concluded

the simulation procedure, but much of the interpola-
tive computations remained to be done. The esti-
mates 'were multiplied by the appropriate incre-
mental reservoir ratio, SMk/Sek. This adjustment

was required so that the interpolated storage change
from sediment deposition over a period equaled the
measured sediment deposition. The symbol Z dis-
tinguishes an interpolated volume from the esti—
mated volume, Se- Therefore, the volume for any
incremental reservoir is

Z“ = Sek.j(

The annual Zk, j quantities listed in table 24 for a
part of period 3 are the estimated losses in surface-
water storage through interpolation from the begin-
ning of the period in 1937 to the start of the year
shown. Subtraction of de‘ for an incremental reser-
voir from the surface-water capacity at the start of
the period gives an adjusted capacity. Adjusted
surface-water storage capacities for each year were
then assembled into an elevation-capacity rating for
the year. The elevation and corresponding adjusted
capacity values are shown in table 25 for water years
1938—42.

A comparison was made between a computer-
developed rating and a rating developed by curve
fitting to determine whether computer-developed
ratings were acceptable for the investigation.
Surface-water storage capacities for 1966 from table
1 by 5-foot-elevation increments were used in both

S Mk

Sek . (18)

 

 

HYDROLOGIC HISTORY OF SAN CARLOS RESERVOIR, ARIZONA, 1929—71

TABLE 24.—Estimated change in surface-water storage capacity,
Z, from start ofperiod 3 to 1942, by 5-ft-elevation increments

 

Estimated loss in surface-water storage capacity, Z, in acre-ft

 

 

N39

TABLE 26.—Comparison of segments of surface-water storage
capacity ratings made by curve fitting and by computer

 

Surface-water storage capacity,

 

 

 

 

 

 

 

in acre-ft Difference
Elevation Water year Elevation between ratings,
increment (ft) Developed by Computer acre-ft
(ft) 1938 1939 1940 1941 1942 curve fitting developed
Below 2380 680 1,142 1,663 2,570 3,705 2410.0 33,308 33,308 0
2380—2385 ., 210 360 636 1,008 1,688 .1 33,563 33,589 26
2385—2390 93 173 504 1,291 1,699 .2 33,819 33,869 50
2390—2395 133 363 575 864 1,291 .3 34,076 34,150 74
2395—2400 93 141 327 542 850 .4 34,334 34,431 97
2400—2405 88 124 146 243 547 .5 34,593 34,711 118
2405—2410 118 239 239 369 613 .6 34,854 34,992 138
2410—2415 113 233 233 233 440 .7 35,116 35,273 157
2415—2420 98 184 184 184 312 .8 35,379 35,554 175
2420—2425 47 53 53 53 114 .9 35,643 35,834 191
24252430 20 20 20 20 70 2411.0 35,908 36,115 207
2430—2435 28 28 28 28 112 .1 36,175 36,396 221
2435—2440 171 171 171 171 445 .2 36,443 36,676 233
2440—2445 208 208 208 208 588 ,3 36,712 36,957 245
2445-2450 66 66 66 66 153 .4 36,982 37,238 256
2450-2455 222 222 222 222 555 ,5 37,253 37,518 265
2455-2460 0 0 0 0 216 ,6 37,526 37,799 273
2460—2465 0 0 0 228 .7 37,800 38,080 280
2465—2470 0 0 o 0 329 .8 38,075 38,361 286
2470—2475 0 0 0 0 67 .9 38,351 38,641 290
2475-2480 0 0 0 0 237 2412.0 38,628 38,922 294
2480—2485 0 0 0 0 73 .1 38,906 39,203 297
Above 2485 0 0 0 0 0 ,2 39,185 39,483 298
,3 39,466 39,764 299
. . .4 39,746 40,045 299
TABLE 25.—Computed capaczty ratings of surface-water storage .5 40,028 40,325 297
used for 1938 through 1942 -6 40,310 40,603 296
.7 40,593 40,887 294
Elevation Cumulative surface-water storage capacity in acre-ft by water year .3 .. 1(1):?2; 21:33 :3;
2413.0 .. 41,448 41,729 281
(ft) 1938 1939 1940 1941 1942 '1 __ 41,734 42,010 276
2380 6,759 6,297 5,776 4,869 3,734 -2 -- 42,021 42,290 269
2385 12,253 11,641 10,844 9,565 7,750 -3 -- 421309 42-571 262
2390 19,231 18,539 17,411 15,345 13,122 -4 -- 42,598 42,852 254
2395 27,602 26,680 25,340 22,985 20,335 5 -- 42,888 43,132 244
2400 37,617 36,647 35,121 32,551 29,593 -6 ~ 43,179 43,413 234
2405 49,835 48,829 47,281 44,614 41,352 -7 ~ 43,471 43,694 223
2410 64,168 63,041 61,493 58,696 55,190 -8 »- 43,764 43,975 211
2415 80,289 79,042 77,494 74,697 70,984 -9 ~ 44,058 44,255 197
2420 98,251 96,918 95,370 92,573 88,732 24140 ~ 441353 44536 183
2425 118,813 117,474 115,926 113,129 109,227 1 -- 44,648 44.817 169
2430 142,624 141,285 139,737 136,940 132,988 -2 -- 44,944 45,097 153
2435 169,414 168,075 166,527 163,730 159,694 3 ~ 45,241 45,378 137
2440 199,089 197,750 196,202 193,405 189,095 -4 -- 45,539 45,659 120
2445 231,850 230,511 228,963 226,166 221,476 15 -- 457838 451939 101
2450 267,841 266,502 264,954 262,157 257,380 46 ~ 46137 46,220 83
2455 306,799 305,460 303,912 301,115 296,005 47 ~ 46,437 46501 64
2460 349,359 348,020 346,472 343,675 338,349 ‘8 »- 46,738 46782 44
2465 396,071 394,732 393,184 390,387 384,833 -9 ~ 47,040 41062 22
2470 446,872 445,533 443,985 441,188 435,305 471343 477343 0
2475 501,110 499,771 498,223 495,426 489,476
2480 558,819 557,480 555,932 553,135 546,948
2485 620,285 618,947 617,398 614,601 608,341
2490 685,785 684,446 682,898 680,101 673,341
2495 755,274 754,035 752,487 749,690 743,430 . .
2500 829,288 827,949 826,401 823,604 817,344 dlfferences between ratlngs was 0.23 percent. From
5 907, 95 , 56 04,80 0 , , - - -
321),, 990244 333,305 387,653 333341141) 3333}, these comparisons 1t was concluded that use of either
2511 1,007,672 1,006,333 1,004,785 1,001,988 995,728 '
2515 1,078,079 1,076,740 1,075,192 1,072,395 1,066,135 rating was acceptable.

 

 

ratings. Attention was given to the rate of change in
capacity with respect to 0.1-ft (0.03 m)-elevation
changes in the curve-fitted rating. The computer
rating was developed to 0.1-ft (0.03 m)-elevation
intervals by use of a constant elevation-capacity
ratio over each 5-foot-elevation increment. A sample
of these ratings is given in table 26. The largest
volumetric difference between ratings was 763 acre-
ft (0.94 hm3) at 2,462.5 ft (751 m) elevation, which
is 0.063 percent of the total surface-water storage
capacity. The mean deviation of the 0.1 ft (0.03 m)

 

 

 

 

SUMMARY AND CONCLUSIONS

Existing records of the hydrology of many reser;
voirs can be used to make an inexpensive evaluation
of reservoir performance. The records may provide
sufficient data to estimate available water in bank
storage and to determine water losses by evapora-
tion and transpiration within the reservoir bound-
ary. It is frequently possible to investigate changes
in storage capacity and in the volume of sediment
deposition by use of existing capacity-survey infor—
mation.

At San Carlos Reservoir, the water loss by evapo-

 

 

N40

transpiration (ET) was greatest during and soon
after large water-level recessions when considerable
surface areas of the reservoir were uncovered. Dur-
ing these periods the potential for ET was high
because the exposed soils had a high water content.
ET computed by the reservoir water budget aver-
aged 26,230 acre-ft (32.3 hm3) per year or 1 1.3 percent
of the total reservoir outflow. Mean annual depth of
ET from the exposed area was computed as 1.47 ft
(0.448 m) in the water budget. When precipitation on
the exposed surface of the reservoir was included, ET
depth was 2.62 ft (0.800 m) annually. The changes in
ET caused by the increasing amount of vegetation
during the period of record were not readily apparent
from the existing San Carlos Reservoir data.

Evaporation from the water surface of the lake
averaged 24,611 acre-ft (80.3 hm3) per year and
accounted for 10.5 percent of the total outflow from
the reservoir. During the period 1964—71 when evap-
oration was computed by energy-budget and mass-
transfer methods, the computed pan coefficient was
0.80.

Sediment deposition in the reservoir from 1929 to
1966 totaled 96,719 I‘cre-ft (119 hm3) and averaged
2,553 acre-ft (3.1 hm3) per year. The mean volume of
sediment deposited was 1.2 percent of streamflow for
this 38-year period.

Usable water in bank storage was computed by
water budgets during winter periods when ET was
considered minimal and changes in surface-water
storage were large. Usable bank storage was found
to be about 14 percent of total usable storage at the
elevation of maximum storage capacity. At lower
reservoir elevations, this percentage was even
greater. Ratings were developed relating usable
bank storage to stage.

The simulation of the sediment depositional pro-
cess in a reservoir was developed primarily to aid in
interpolating capacity changes between capacity
surveys. The simulation was successful for the pur-
poses of this study. For applications in which the
primary objectives are to predict location and
amounts of sediment deposited and to estimate com-
paction, the simulation model needs further testing
and modification.

REFERENCES CITED
Anderson, E. R., 1954, Energy-budget studies, in Water-loss inves-

X}U.S. GOVERNMENT PRINTING OFFICE: 1977 — 789-025/74

 

GILA RIVER PHREATOPHYTE PROJECT

tigations—Lake Hefner studies, technical report: U.S. Geol.
Survey Prof. Paper 269, p. 71—119.

Burkham, D. E., 1970, Precipitation, streamflow, and major floods
at selected sites in the Gila River drainage basin above
Coolidge Dam, Arizona: U.S. Geol. Survey Prof. Paper 655—B,
33 p.

—1972, Channel changes of the Gila River in Safford Valley,
Arizona, 1846-1970: U.S. Geol. Survey Prof. Paper 655—G,
24 p.

1976, Flow from small watersheds adjacent to the study
reach of the Gila River Phreatophyte Project: U.S. Geol.
Survey Prof. Paper 655—I, 19 p.

Culler, R. C., and others, 1970, Objectives, methods, and environ-
ment—Gila River Phreatophyte Project, Graham County,
Arizona: U.S. Geol. Survey Prof. Paper 655-A, 25 p.

Davis, A. P., 1897, Irrigation near Phoenix, Arizona: U.S. Geol.
Survey Water-Supply Paper 2, 98 p.

Gottschalk, L. C., 1964, Reservoir sedimentation, in Chow, Ven Te,
Handbook of applied hydrology: New York, McGraw-Hill,
p. 17—1 to 17—34.

Gila Water Commissioner, issued annually, Distribution of waters
of the Gila River.

Hanson, R. L., 1972, Subsurface hydraulics in the area of the Gila
River Phreatophyte Project, Graham County, Arizona: U.S.
Geol. Survey Prof. Paper 655—F, 27 p.

Hanson, R. L., Kipple, F. P., and Culler, R. C., 1972, Changing
the consumptive use on the Gila River flood plain, south-
eastern Arizona, in Age of changing priorities for land and
water: Am. Soc. Civil Engineers, Irrigation and Drainage
Div. Specialty Conf., p. 309—330.

Lane, E. W., and Koelzer, V. A., 1943, A study of methods used in
the measurement and analysis of sediment loads in streams:
U.S. Army Corps Engineers rept. 9.

Marciano, J. J., and Harbeck, G. E., Jr., 1954, Mass-transfer
studies, in Water-loss investigations—Lake Hefner studies,
technical report: U.S. Geol. Survey Prof. Paper 269, p. 46—70.

National Weather Service, issued annually, Climatological data,
Arizona: U. S. Dept. Commerce.

Simons, W. D., and Rorabaugh, M. I., 1971, Hydrology of Hungry
Horse Reservoir, northwestern Montana: U.S. Geol. Survey
Prof. Paper 682, 66 p.

Thorp, E. M., and Brown, C. B., 1951, Sedimentation in San Carlos
Reservoir, Gila River, Arizona: Soil Conserv. Service Tech.
Paper 91, 26 p.

Turner, R. M., 1974, Quantitative and historical evidence of vege-
tative changes along the upper Gila River, Arizona: U.S.
Geol. Survey Prof. Paper 655—H, 19 p.

U.S. Army Corps of Engineers, 1914, San Carlos Irrigation Proj-
ect, Arizona: U.S. 63d Cong., 2d sess., H. Doc. 791, 168 p.

U.S. District Court vs. Gila Valley Irrigation District, et al., 1935,
Globe Equity No. 59: Decree entered June 29, 1935.

U.S. Geological Survey, issued annually, Water resources data for
Arizona—Part 1, Surface water records: U.S. Geol. Survey
open-file reports.

1954, Compilation of records of surface waters of the United

States through September 1950; Part 9—Colorado River basin:

U.S. Geol. Survey Water-Supply Paper 1313, 733 p.

 

 

 

:.1...1..15115511515511.1215... ,L 1 L L: 1 1 L:
.152 L.Lw552,151..LL,,.LL5.51.152...,fimwflfiwmmfi:,L.:.5LL.LLmL,1L.L.L,m..Lm5.;L,L..21L.1.. L1.. :L2,1L: ..1 .:,... L5 , 1,
:,5 1.15.11.51.555 ,
,.. ..512.LL.,5;5.:,1 L1,

1;

1151:3555... .5. .2. ,
5.1::15,,51L,L,...LLL1L,.L,1LL5,,5...LL1L..L,.. 2511.11: 111.,5.LLL..,51,5,5
.2555 :L2,:L..,L,.1 521,. 5151:5155 1

11:11:11.1:5. L11:...2. ,5, 1111:5115 5 ,1

115111 ,1: 5.,:5L5L1...,1, ...:L5LL..L55.,5L,,511:;

1 55551155 1.2111555111511115.1.155521.1111.551.1111.251211.1.11555.
,,:,1.L 5....55.1115..5.. LL53. ,,5L5 2....51L. 1,:. ...5..W..515..125115... 5.1.1.1.2m1215...5.21.1.
.:LL.,,LL,,L .,L111:15::,,1:,1L5,L1LLL5L5.LLL.,L.L..1L,L,L1.1L,..L1L,,LL1LL.LLLL.,L :,1:11: :LL1LLL,1.,L,L:,.:L.1.1L.,L,LL.11L1,,1.LL,LL1L,,..,

,.: :,5 1.5L555.m:,.:,m.L,L.L5L.,5L:,5L,LWL...,L,..L,NL5L,LL 5.1152555 5 2 :1:1 1,55.125.551.521,:.:,2LLL.,HL.1,L1.L,.L,L.L,1.L1,5
LL..W5L1L,L,5LL.1L,L5,L,1,L,1L,2.LL1L2 :::,..:, 1:,.: :5: L ,.: :1,1: 12

 

5

22:51

,:1 ..5.51515 5.151 512.

:,11. 112.25% 1.1,
1: H: :L, :..,.LL. 55:115.... :,5.1L5.1.15L2L. 5
5.2.1:... , ,

:5::
11111-8,
11

L L,L5: ,L1,L,1L,LL., :L1 L,1L ..L .111 L.. L ,.:
11 , ,,.:,5 .....15115L1,..1.:1L,1L.25:.

.

     

1:1
5
if
33F
11

L LL.L,2. 1 :1L1:,2,L21:1,5.1LL.,1LL,1LLL,1
...; . 1 15.:LL..1L..,1L.,,..15L,L,L,..,..LL..:L,

:12: 15

2...

1 51:: 1:1:L.L.15..155,...52L11L5LL:,.L,L,L,11:1,.L
:.1L... 55.151.55.151; 2LL5L1L.,5L1L 1.22.. 5,2

152 1:2,. 1

L,.,,.LL1L5L1,L.1,L5...,.,L..L,L,,L,2.,.151LL,,,:L

2

 

,, L12

1,351 11151.52,
5.15155525.22
.1LL.L,1L,L5..L ,..L, :L:.LL,L:1,..15:.mw5...w1L,L1L.WLLL5.5.1, 1
15:22....5 L
:,5 .LLL,L511,1L1,.LL.,1 :1.:1,LL1L,.,L,L11.L,.L.L5L11LLL:L:.L,NL,.1.L,
..2..,.1.,L,LL,L1,L5L,L5.L:L,L,.1.NL1.L.. L, 5. 5:1,:5
L , :L.. 11,511.11. :1 5.1 2
, 1 51:2
2 L:L,:L,11.L..L1L,.LL21.1..215

2
5

LL11:
11

 

11
,.

~:1
’2

.,.,1L.,1L,,1.,L,.,

1

1.

2: .L,L,.L.5,LL,1L,...L.LL.,,LLLL,.:..mL
., L.,:,.L11: 2 2.:

1 ,.:,HL :2. :,5HL,L,L:L.WL.L,5.,LL.,,1.L,LL,,.1.L
5 2:15:15... ,..LL55:
L,,:L,,1L51,L:.,LL1,LL,.L :L.:.1.112: ,:L..,LL
21.55.1151,...2,:L:11LL..,,..,1,..,

1c

1 21,551.25::,5HLL55L1LL.:,.L5L.L1,L1,LL5.1L.L.L15:..1 L
L 1 L 2 ,.51 512.11.52.11...
:,. .1,:1LLL5L1,1LL2 11:111., 11.1.512.121.1..2151151.15215 .25.1.11..5..155

L :.1:,:1525 L,..:L...L1L,L.L,L1,L.L,LLLL.1,.LL25:5:1,..;LL1 ,

2
1

LL.1:... 5.115.; 2 ...L L: ,:.:,1 1:,; ,:.L. 15.255115 1,,,.L:5552....LL22L11L...L,1L,..LL,5L51, 2131:2512 :L:.1.LL,L.:,1L,L,.L.

J:
‘1

.:.L.

1 :5;

2

,1: LL1:,NL,L.5L,,1LLL.1L,, 5

,.. :11LL,L.;L1LL,1 L

,:.:,

.. ,.: 51:5

1518'

1,L

5... L 25112 ,
121215.212;

11L

 

,:

 

,5, L:

.1... :,1. L:1:5L.1L. :,,
L.::1,.L,.L,L.,5. ...L.L..1.,.2..1L,2w

 

:5.

:,LL
’1

,.,M,5,.»5,:1w..:.:1,L,..L,.15,1:, ,:L.
.W5LL1L,mL,L.L.5www5.Lu,LL,9.2.5:, L
25.1212.2.1.5125.

.., :,,::,L.115LL5:,1,L ,2L51L.:LL.LL ..,:1L525L1,L15L,2,:L,..,1L,1

,
'1
i
’11

'11
1

512 ,

2
L1

5

1
1
,
§
1

1:2

51:2

11L":

.,..1... :LL1L5L1L1, 5.51

 

1

:1... .2525. 25,
.:L,,1L,L.L,L.L..,,L.LL,L,L.L,;,L1;LL,,,LL12,L,L,5,2L.,,LLL,.: 1
,2, L1L,1LL.L1L,LLLL.1L,1LflL5

2

 

2
111

52.11:: :,,..:,1LLL,,1,L,1LL, L1,...:,:.:L,.:,551,121.,LL..:.L,L.L,L.L.,1.LLL1.WLL1

2
L1

,,L

.15.15.125.111251

L5

    

. 11;:

.L L 1,1:,.2L2:1:1

..L
15LC

,.:,..155

L:,.LLL.,.L1L.L L . 1:51.521:L,....L,L.L,LL,,.
..1..11.5..1,1 . 5: 111,: ,:,,:,1 ::,,LL..11,L2 :,:1,:5 .
L,1: 1112,.m,5,L,L.L1LLL1.L,.L.1L.L1,.LL1L, L

5

:L2.LL1L,..L15 LL15 L,.,

1

1.5153111
1.1.5:, 5......211511151
L 5525.112151.

., L1
L51
~11

                

:L:L1,,.L.1L, ...:,.1,...LLL.1,..55L,LL..L,1,LLL,.1..LL,5LL..L,L5,,LL L..::,.:,1LL5L5L.1L,LL:.L..,L,.,,LL, 5. 2511
::,..:115 1 .2112 L 11L,L.LL,LL551551L1LuL1,LL,LL 5.1.221, L
:51, 1.1 215 :,:..,,1: L,L...L,L,15L,1 ,,
, 1,:,L,1,L,L.1..:,5.1L.,,L.L..,5L,1L,L1L,mL1L,.2. L.L,,1L,,2LL,LL.,.L.L.L.11 L225..1,5L,LLL.,.L.1L.L,11L.L,,1L.
..1. ,:,,5LL11.215.1115. L, L
525552.1121 ,1 2.55125 L,:L.,LL5 11,552,

1 .5 .1..15 ,.:.11. ,

 

1
::LL
5’1

:1:

'1
1:5

1'.
1.

L: :5 ,,..:1LL,.L1LL.1,,.
:,,...5 L1,
21.511.

’55::

L21

 

 

LL1

.. ,.L,L.,5,1,5L1L,w.1.1

25
5

,

s11.
1153,13
.,L,
:11;
11L
1'5

:51

:11:.2,.L.L1,11L,1
.:.,1L,L,.,L,L,LL..,.LL.,5LLL,,11..,:. L

1,.2.:L,LL.1LLL. ,.:,LLL11:1L..L.L..

.1
5

211111.. ,5 1,1 11:12:51

 

1:5Z

,,
3L
.

“1,

.2112. 21:;LL,
1,511.51, :.:,:5. 5:,

1

2,:: 1:,:,:L,L,1.LL,,.L.L,L,2,L
::,:,2 :11.. .
5,15,: L,
,,LL.1L11 521.1,,
2 :,.:.2.

,L
,1;
7:???
§
'5
'2

 

11:.
1%

i

1:,1 LL L,L5,,LL . 1:,,L,L,L,L,L,.,L,LL, :1, .1 :.1
L.:LLL,,.L.L,L,L,..L,L1LL,5L,.L,1LL..1.,L :,.:L:,2LLL,L,5LL,,L,L115.L,,.L.L.L,.L, :,L21.. :1:::, ,1.1,L,L,,L,1LL,LL,1L1:,LL,L,L,L.., .
,1 LL.1LL.,,5L.,L,LLL2,LL2L..1L.L,:11LL,L...,LLLL,L,LL.,,L . ,.LL,L,LL,,,1L ,:,..::,:1,,
121.15.. 11.51.52... 25. .2. :1::,:
2121. L. :2LL1L,L1,L. 8,,L,L1LL,L,L,LL12LL,5L,, L:1, L:L:,H1L.L,,LL1L,1L,L:,:.:.L1L1,1.L,5L, ...5. L:..,:

:.1 5
‘1‘ 1
51
,L1

5:
1

:L2:

2

51:,

515,221,.

11

.1.... ,L
,1 L.: , :.L , 1.111.: 11 , 1:: L
L.. 1,L,LLL,,1.LL.LL1L L L.:,LLL,LL,L.LL1..L.L, LL,,.L,LL1,L,LL.5L,L.1LL,.1,L.:,u.LL,LLL,L,1,,L5L,:L.LL,:WL5,L,L11LL,52:,.LL,1,L1,51.5L,.2,L2...,2L,..LL1,2,.LL,
, 1 ,.HLLL1, 2 :1,1:..L5 ....11...1..1...5121

,L1L. .25 :L:..:.L,..L.L,.,

L,1L..,,.,L.L,L,w:,
5:551...

’51;
,L
,1

, 5,::,,:LLL2L1LL.,.LL.1LL,1L,.
L1.. ,:,.L,HL,.....L ,1L: LLL,.1.L.LLL..L,1: :,:.:,.

.:.L , L.,m

.L:.sLL11LL5L,LL...1L,.L11 1

'1

1

2

...:5:1L525.15.511.55...L,,,L,L.L,.LL.:...LL,:,:,..::1: .
L,LL.1L1,..L.L,L,.:.:,NL,L5,5.,12.L,,LLL:1,L..L§.L,1L,1.,11L1,L:5L.L.L,5, :L1,:.
L..::,.: 122.

1,,:,1LL1,5L, :LL11LLL.,..L1L,LLL,LL.,
15:521.: L:,:,L,L1L,,..LL..,L11.

5

1
1
1'

2

.2.1...1.2,5 2,

15,
1;

,.

1,
,5

11

,3. ,

L ,.LL,L,L,1L,LLL1,..,L..L,,,.LL, 11
.1 ,1L,,,L,L..1L.,5L111LLL,L.L, , L.,:

,,:1,
8'?
1'2
2:
5

5~

1

.515 L,.:.L1,LL1 :1:1.5L.1L11

11.11.525.55151555 5.11.1...

121:.

1.11.. ::./...:,.,,::1
.:..L L,: L,.:,L,..LL1,..1LL..,1L2L,.L.L.,5,L,1LLL, :,L.me
.1

1~
2:L

1.5 L1:.
, :1:..LLHL1,,LL,L.1,.L.L,,.,L,L,LL.,,.LL,L.. L1,:1,:1

1'

2

L 2,51 .: ,:L,:,...L

,:.1:L... 111.: L:

  

L L11.L1,L5L :,,:,,::1L,LL,,L .:,,L;,L,LL. ,.:,:,.LM,L,L.L,,L..,LLL,,::1,.::,,1L,
2:,L21L1,,.LL,L,5..L,..LL,2 2,::L:.,1L1L,L.5,:,LL1.,L,,L..,,LL.,..L.L,L,LL 5, :1,.:,:L:1.11L.L,,L,LL1L,L,LL,L5:L,:,1L:,m

.1L
12.55.. ::,:151: .52.. L51:L..:,,1LL.L.L1.L..L2.2.1.:L1,:L.LL,.LLL.1.,.LL.:,LL.LL,N :1:..2L1,,1L,LL :,..:5,

:L::.1 L5 5:,:,,L,51.w..:,2:.1:,:51: 1 L51:,,.:.L1::.:.HLL.,1L1,W.L,...,5,LL25,W1. 521....
, :,.... 512115521 55 :55:

1.21115 51, 21215215. :L:L.1,L12
2.1

L “5:11.:,5.,1L,.LL.1L1L1LLL.,,..L,L1.
,5..LL,.,5:.L1L15..LHL,L,1.,LLL1L,L, :,:.:,:2,

, L , .:1,1:L,L,5 :,1:,,:,:..,.LL111L5L,L,1LLL1...LLLL51:11:11.1:L,1.:,L21.L..LL.L,1.L,.L.NL,51
1:1, 21:11.15 1..1 21.15. 1111111.... 1 25 515151 5.
L: ,1: 5 ,5... L 5. 11.52....,11L1::,:..,LLL15LLL5,LLL.,,.LL,5L5 1212511.». 5.32:2, :.152L1,L .1... ..

       

,;L..,.L1L,5LL .15: 5 , .1
:,,11L,.,L...LL,L,L.LL.L,...LL.,. 2::: ,.:.L,
L1....1L5WL..,.1 ,

5:111”

......LL...L,LLL,1,.L. 1.1L,L:L,:,1,:.LH11LH,.:.1L.L,1L..1.L.,1,: ,,L:
.:1L,L,.w.LL.NuL5,5L1L,m:31.2w,:.::..,.L.,L..L:L..HL,HL.»,LL2.L.LL,.L,

L5

..L. .2251.

11~

L ,::,..::,:,41222 1.55... L:1,LL,..L.L,LLLL..,21,,

:L, :..,. :,1:, 1.

2,:1 2.1.1::1 111,2. 2.1.1:... 121.. L: 1:1,, ..::1:,512 2

:55..5.LL...5L.L.L,1.LL5:11.111.5.2252“.L.5HLL.1,1.LL,.5,LL11,,.L15,2115,,..L..,:,,1.LL,1 1,:,“
L:,,:: ::.555L5L,1L....L,LLL5L,L,.,..1L,L..LL,MLL LLLL..,L..1,..L,NL,5L...L,L.1L1,5L,LL1:..L.:,.L,.LL1,LLL1.,..11.
1 1,1 L:.L,,,LL,,2L,2,L1,.L L..:L,.L.L,L.L,1LLL.1.L1,.LL 5:12LL.L. :L:...L,L,.,L5L1,,5L,.5,..:L,L. ,:..LL,LL

13L
5
11L

11:1

2.512L,:L.,..L1...LL,1L..:2L,mL.1L,W.L.L1.5L,....LL.LLL.1,.:,, 1:L..,,LLL.LL,..1.L,.1..LML1.,5
L 221.1511 5.1

,1...
1,L:L~~,1~

5
5L
'1
1
5.

:2.

L5;

.

1

1251.1 1
. 2 .. . 5.55,..NLL..,.LL.1,1L..,L11.L,
2.55515m555;

1;
1:51
13
2
2
i

5‘,

111151.125:
2.1.1: :,.1LL,L,LL, :1,:,,:1 :5:,11,

:2.

1
5

,;3
1
,L:1
1
2

, 1.11.;2 L..:L:,L..,L,L1LL11L5L,LL,5L,

1
2L
1L

1:88,:
21:58

1.151::L..:,LL11.:LL.,..L.1,2.W.22L.5,L..:11.:,5 L11,:LL.,,LL1,LL,,L,1,.L15LL L L.
111.15.11.511:,.:,,LLL..,LLL1L2..LL.L.11,5.:,:.1L.L,L151,. 1:.... ,:1L.
..::,.:: :5,:LL1L. 5:111:15: L2.L.1:1,
5......55.111.5
L225:...LL1L51L,L,..LmL51L5.L,,L.:.HL....LL.1L..LL,...LL11,
L: :51 L,:L.,1L1L5L1..,L.....:L..:L:,5..15.L1,11LL
.1 1,, 1, L L..LL,..:,1:,m1:,LL,LL.L1LL,5..,,2LL,L.,1L1
5115.51.15.15L,:,LL..LL5,..LL,1,_LL,51 ..1

1

.L,LLL.L. .11....LL,.:..LL..2L,L.1L,1L,L5,,1.L,L..1L.L,,5L.,L.,L..,L,5 .,L:

i’;’
1~1

1

2

L5

:L1, LL, .1LLL..1.

1,

,:L

525
i

1.

1

.2: L.,.LL1,,.LL.,L.1L5, LL ,L :L.

1

1
'5
L,L
..:215

.5 .:.1 1,5.
.: L.:2L151,.,LLL.L..L.1L.L1,, ,: ., 1:1 L.,

5
1,.

5,
'22:, L1
11

2.21115 L,1L,L,.L,.L,L.,LL..L....1L,L,LLL,L11.,.1L.1L,L.L,L1,L.L,.:,L,.L,.LL..,.LL,,.L,L1,L,.1,.L,

L: L ,:.:,.L. 5.:: .1:1
:..:1,:,,. 2. 5 . 1.51..

.L:,LL.L:L,51L,L.L.5.1L,L.LL1,21L5L5,1. L,L..1..1L..
,: 11:5 1:.. :..
1L5.LL1,,LLL,,L.L,2 :.,:L 1 5:: .L
., :..,:1,..1,,2.L.L.51L2,.5.1L11. 1..:... L

5
1

:1:1:

21L

1

.:,1..........HL5H55W...L.,.LL.

3:

L 1 551.11 ,L: .:
,,. :1:,1....25512....
L.:..L.,LL,,L:L,,.L1L,......12. L.

5

L: ,.:,LL.,L ,.: 1. 1 112 1 L.:L..L2..1,L.,L,L

. ,, .:,L,L,L.L1LLL1LL,..L 511

’2
2
,:1L:

                   

:1:::,::, ..1.5
..L: :,.:1:L 521L512... L

L, 5.152: ,51;,: 5
2 1:25.11. 5

1 . :,.:.:11,:,,. :,.::... 8LL,L,,L,,

:.1
15
'2

17

,11,:1,L:,,L, .,,LL1L,.L.,LL.L,LL....LL1,5

L;L1,L.L,L,L,:,1LL,LL.1,1.1,.LL,L,LL,,1
1.15 L 1. 5.12:1.

5
LL55
15
11L
.,L:

5
1::

,1

 

        

55:11., :L....LLL..1.L5,.L.1L,m:,L,L...L1LL15L,HL,5L.L1,.L.L,..L:,1
, ..1 L. :,.:1L.,.LLLL.L, 5 51.11:. 2

.1...,.2.L.2..:L,.. 2,:L11.,,.1L1LL..15, L L: 5
..1 22215.
:,,:5,.L.1..12L,L,L...,:L,5,L,,LL.L,L2LL.,.L,,L2,L,L.1.,1L,1:.1

1 :5:,L111,L.L, 1 :.,:L
L.. 1:L1,,..L, ,:L5,,.LL1L,,L,L,.L
L . :., 2. ..1. .5 :1,: ,..,:1.,.L,2

5,
2

:L,...L11L1,.L.,1L5L,L,L..1,1,1
.1122 11..

      

1.115., ,:.,L... :,:,LL,L1L1,.L,.

:11
11

:L.: ,::.1LL;.L.11L..LL.,., 1: :,::11
L, L: 2:,,L2,..L1LL2L,L,L51.1LL1,..L,5L..,1LL,.L. 1. :,,.: L2

L:1...L,L,L..1LL1L.LL.
1:..L1: L.. ,,

15 L5,:LLL..,1,.L 52. 21 :11

1
’2

22

'5

.:,,:11.1:,L5.LLL..,LL1.,LL,,.L,,L,LLLLLL2.:, L
..

51
1

1:51
L1
11‘
11>
1,
1:

1:.:,5,LL,1..1,.1,.1L.1..L.L15,
,:,.1,5 .
1,1 ,1 1,,:1 .,L

1'1
5
21, 115-.

 

11

1
2::
11:2
:,1, :11
2

1.12.... L L 1 ,: 1....1521

   

L’L‘E

     

L,
5'

  

1.1.1.1...
:.,L.L1,L.. :1:1.1LL,L,,1,LLL.,,..LL,5L, :,5.
::,...2 5:5. :,,:LL1,,.1L,....LL

    

L5 :L:.,L1LL.,,.L.1,LL.L,1,..1L

  

:,, L 1..:1L.L,5.L;LL.1,5.1 L,:L1L,,.L.,LL,1,..LL1LL.1,1
.:,.LL1L121L15: 5:151:22. .:,

118'”

,, L.: :,L.:, 1 L:,,1:,: 1 LL1,5..5.L,. ”W
1,::. :.12111. 1,: L1

.5

L,,

, ,,L :L,,
L HELL:

L15~

  

1
21

 

5

51

15511
1 :'
5 ,

L .52511111. 5
1.2.1552. 12:55 :,.,:11: 1 1,

     

:1 .LL1,LL.LLLL1L,L,.:L1,L.L1L
1:,,::,LL,,LL.L,,LL.L,L.LLL.,L.L..L,LL1L,,1L,L.L.,2L1L.LL

 

'1
1

:2 L,
:1,.:1L1LL, 52:, L
L 111:1 5...:

                                                                                                                                        

1:
,L151L

,....L ,;L ,2 ..

5L2~
1L

1:”:
1°

,:.
«21:15:51-
'1
121

5
22‘
2:
51,
,L
1
,L..,
,,L

1 L . L...LL.,LLL,5LL1:11,.:::,..:.1L,.,,..LL,L5LLL.L,.:,L.LL,L,L,L,1LW.L,:1:L.1L1,.L,,
L:,1 1.:.L.1L..L, :.. :,:1,L,L,.. :.,L..L .. :..,:5 :,.,1
.2 5 . 5 1:,.:: .1 L L. L: 1.51:, . :1,L:1.1L5 2 1,1: L: 112.

.215,5,,L:,,.L.1L,L.1LL,L.,1.5151:5:1: :.. 2,.: ..L, LL1L5L,1.L/,L..,L,LLL ,,L211 5:,5,15,L1,2L.L,2: .,1L,2 .:,:L1,...,, ,L:. L
, .1.222 . :,.52 L,L,LL.,.L2,.L11L.L.,L,L1,L,..L.L.HLLL .,2 5,15 11:5... 5 2
L “L, ,1. .5,1,,:,L..L,L1.,1L11L.1L.. :1:,151LL.

5.:

          

L..:1L.,,LLL1,..5L1,..L21.11.811.15

.1
L: ,L,

25.21.52.125
1:1L: 215.1 ,HL11LL1: ,1
2. 5.5.515: LL.1:,1 55:51.. :,1: :2. 11:5:

’52

11
.11

           

2
5
,5

.51 .1.. ,:,1: 15 5:, ,L
.L:.112 L15.,L,..,L1LL.LL.,L ,..:,L..1,L,,:1L.1,LLL.LL22L,L,.LL.NL,.L,,.L .215.112 1, L: 5 .1,:,,:11: 1::L1...,LLL1L,..1, L
5252.1...25552WML11: 15111.1: ,L 1&5%§§.2.1.21.151.1..,1LL,,1LLL5L,.LL112.LL,L.5H.L,L52:,L.L,.L2L.LL,
LL 1.212 5 212115u.11
5151.2:L,LLL,L,L,L1L ,..:...L, 15.1211 , L .

            

,.L

25-112

.28

  

1::1,:,5:,L.,2L,1: 2515 L:1.5,
, L: 1, :1,1:,,.L11L., L,.,,.L......L,1.,L,L. ,:

'1

21 L.1LL..,1L15L,,5LL.,.:1,.:1LLL..LL51.LL,..,1L,L.L1L...LL ,

,,,:
LL:'2LL
’r

L 2.55.1. . .,:.111 ,..,:151L5,,.L.1L,5.15..,5.1,L, 5125
5.2::,1.1,L..L.2,1L1L2.L1L,..1 .L L. 2111112
5.,;L,L12L1.5.LL.L,2 :5,...1

                                   

, ..
’st
1's
1';

1
1.

1;,
:21

L 1.. 11.
....,5..52.5:.w1 111.12.:,2:,LLL1,WM,52...L2,.LL1,ML.,151LLL, L m
:.., .L:LLL,L51L1,L.LL1.,5L,L,L5.LL1L5:21:11:,:,.L,LLL.,LLL1,.L1LLL1LL,L,L...L11.2
, ,1:L.,:,12112.25.L1L55.2,2,.:1,15L5,:,5L,L1...L 2 :1:
{$22.1wa .,.,MfimeLLlfi 1:11.:

21,
'1

,L.:12i51
5......12

,-
,1

L111

:,.:15 1,L::,:,.1...L,L,.L,5L,%L,ML,L,2.,,.L.1LL,.:.L,2:1,:.1LLLM..L,1LL,..L,2L,HL,LL11:1:

1..
115

1

1.21.2...255115 L
115115 L1
21.1LL..L1.L,L..,LLLL.,.LL,..

1L,,L.1Ln:,L,L.1L,LL5.,.L,LL,ML5L,L,L,..,LL.,L,1L ,L,1.,1.,L,L..:L 2

1: :1 ~’~’

, 1 ; ,::.1 111.1 1. 1 1152
11: 2:5, ,2, :,2:,.. . L 11.1 :,.. 2.5 1:1,.1LL
.. 5:15. .,LL.L . L115...LL.L.L.L1LL1L.L.L.L1.2.L,:,,..:,1LL21L.1L..1..,,LLL.L,;LL,,..::,:..L51L1 L,....
1:51:22. L,..:uLL:5:1,..1.51L.L,5L,LL51.L,LL1,,.L1L,.LL,Lm L
r2121 :L1:1,.

5,: L5

111:1, 1.: 2 L...L 1
L 1::,...11L55:5:LL1,...21LL5,L1L.1LL1,L5LL,.. 1
21.. L..1LL,1,..,.........2512.1

2
2

    

,L

            

255%.. :1,. 121:.

          

,,L
1'8
,
1
1’

1’2
'22
2:32

25:55.51:,.L,L,.L,LL.,L2L,.,L,LL1,LLLML,1 152.55
1.1 ,

,L
'1

,L

51
L5
5
1
x'
1:.,
52
11 , LL
.1-
is?
11:“:
,L.
2 :5-
s

1,: .. :.:1
, .11,.LL1.1LL,5..LLL1.,.;ML1,5L,51LL.,,LL11L L.:L,1L,.L,LL
:,N.L,L.L,L5LL1,.L,L,L,,.L.L,2111.111151...

.5 LL25LLL1,21,.5, L 1:15.152. 2

51:21

 

1.1.2.5.,

11:,:. ..,:1L,,..L1,1.,1L5.1 ,1

.,...:..- :1:: ..L, L.. :,,.:,,... :L:L.,

1
.2
5L

.1

                         

 

 

   

521??”

2

L21:

2,
g,
51:12
1'9?
52
,,,
2
.,L
:8
8,1,
.1;
L11, .L,

1.,

.:,,.,L

1 L L1 .211.25.11...5.12
..25.1..5.52 51.2..
,,. .211155151... ,

15L"

5

1

.1.

1L.
1.,

1
,., L.
‘2

L: 2 L:L5.L,1,5:,.,.,.L,.L,L.1.1

21

                          

1

        

'2:"L°111
:1
11

.L

  

151-

,L
5

5:::

5 11.5....L...1L,.,111
.12

,L.::

:,:,.1 1,151.51. L:,..: ..:,:L

.:, ::,1:,11. :.LL..121LL ,:,::,1L:,:, .
1:, :L:.LLLHL,5LL.HL,,.1.L5,,,.L,1LLL.L,,L2,3..5.1LL5,.LWL..L..L,L1LL5L11LL1 51,
L1 L,.L,..L,5.L111.....LLLL..,.L5L21,L,L.5LLL51 1 L
1 5152.15

L1

2. :,LL,...LL1LL .. :,..:1L., L,1L,.L

2L
1,

1 5

2
L», :1

,.:
'11;

1‘2“

1: ,

2:1
,:12L

5
:1.~,L:1

:.,1L, :,:.1,.L1L,

.:L.. :L:.,.LL., 1 ::,2 1
. 1L.:

1:”:~
5
Y}
1

5
5

.15....11
.,:1... L,1:,..1

1
1
12
2.
:1.

5.5

L,L
Lx'
11

11:2
Lz’s

,, L .:L..,LL.L.1L,..L,LLL,..,.,L .55.5112.5.5.5

5
L1
11~

                

1~~
i:
55

1 1.... L, 1111:,

L,L.,1L111: 55113:.

211:“: 1.

15 1
'8
1

L. . :1::,:,L,.L,, .L:.m,111,.:,:21 L
L1 1, 22.2112
.5.1.1.1.r111L,L11L,5:,L..:...L,5,.L.1
25.1.1.1». 1.,,1LLm,..1,L,1.,,...,.LL,21

2

, 1:2,,

..1. L L L L
L: 1,:.,,..1L,2L11,.1 L,

”.:,1:.2 L:, :L,. , L,
L .:..LL1LL1 L .:,,:L ,5.L1,2L,,2,.L.1LL,L2.LL..,,..L1L.
.L,,1L,L1 2.1.112,..:,L.L,.,LL1,.1L.L,L,LL,,,L.1.,.L 12:25:, :1::,::1, L,.

, L.,.

, L..L,.,1 ::,1: L L , 1,..
L L , 15.

5: L1,

                            

     

5

,L

L5L,..LLL,.,, 5
:1....LLL1L..1..1L..,LL1,.1L1L.LL,1,..L, L1.. L..:L:...
::,1 :1. L. L L, 3.55.1..LLL1111. :11:.L15r15,:1:51 :...1LL11L,..LL ..L: ,.L 5::,,:.,LLL..
22:52.11, :,:1:5 5L,L..u:,.::,:,:1,1:1 ::.,1.

L:LL1

2.1121. , :,1:, 5,: H,
155.1... L,.:.

  

..51
:LL.L1.L,.L,1LLL.L,.L,L.1. 1.1.5.
1 . .:.:L..1,LL.L5L. L,..

.1
1

   
  

,
'1
11
,1
L's

L..,L:,.:1L2L1Lw,.:,.:..,..:,5,5,5::,,.:1L1,L,..L1L LL11...LL,..,,LL1,WL

   

, L. L. .2.2;..1..22 5,2,1,:,1,:,LL1L.L,1L,L:,,:L1LL.WL,,L,.LL,HL,L.L,1.L1L, ,,L., :L.

'1;

               

11

      

1:

      

5

    

 

L2~

     

                                                

         

5.1..

1:1,::,,1LLL , .:,L..,.. L.:.L1L,LL1. 2,
L5, L1,..L,L..,,.L .112
.L: 21112 :5.LLL1L... 2:1,

:13

    

:5,,1L,L,L121LL1L,..L,1L,L:LL,LLL..,LL 2.5:,22,L,LL.,LL,L,L,L2,L,1LL

1L, :,:1, L .111 L2 ,
L 1:5.1,L5L1L,L1L,HL5L1LL,2LLLL,L,L.LL,NL,L.L.LL1L..L,LL11LL11:,1:, :1L
.,,..LLL2LL,.1L,L.L,1,,..,L1LL.L.. :,..5,L5L,.L1,..L1L2L

L11

5

       

15. :.1 52.115.12.251 L2:,,1:5

 
  

51151.11:
5,525 .55.2

1,:

1'2

, , :L,5.1L,., L5,..L,L,2L,,1.,L,2,LLL 1 1:1, L
11:5 115 1,.: ,,,:,,.LL,
1L1,1L1L,.L,.LL..,,LL5L,,.L,L, .

2151525522,... 1:
L2,..L1L5...L,2LL,1LWL1,L L,.mLL.1:m111,L,.L.m.LLL., .,1.,

:"s
1.
.

1

1.21:...

L,:.L,.,L. 1... .5:

5,::

'2:
1

              

5

155,...LL.,,..1L,5L,1L1LL.LL.. L 1:.. L L ,:
.:., 51121212551551 .2.125.,L1
L,L,:,1L.LLL,L.L,W:L.,L.,1L1,LL1L..L..L,L,5L,L,1,.L,:L:LL,.LL1L2L,1L,1L.L,..1L;LLL1,5.L,5.
..1, 2 .:1,::,...LL..:1L1L 5.125,... 5.1.1, :,1,L.

L:1
Q11
~111881::
.311
:11:?
L,.;
11111:
1:17:1
£551.
1, 1.:;
1811,11
8L1,
111:1
158,:
11:
1

, ,

158:1:
5
L'SLL
'83
:1
8‘1

.1....115151 .,:,
1....1111...1 M:., 125512.291:
43 .LL..1L1:,1:..L5L1,2:,L.L.NLL5,HLLL.....55.:121:. ...... L.:.
L .:N..,1L,L.L...L.1L,.LL1,L,L1L...1LL.,,1 51:51:52. L
5151u21.51
5.1.515:L2:,:L,5,,L1L..1LNL,L,1L..,L.1.,.LLL.L..LLL, :11.:
.. 1 51.5
55.21.115.21... ,5L
1.15525 2

L L, 11151151.:112 L1

11
1
2

2
1
1
11L
2
1.
‘1
'5
5
'1

‘53,

                       

18>,
1

   

5
1
a

  

i1:
1:

:5
,5:
1g,

'2
2:

 

21
1;

1.1 L :1 L 1 ..,:L 111:,:11.,:L1:1,,...LLL,..,5LL,1L,;.L,L,L.L1LLL.,.:,.
L 5 L . ,::,.L.,,L1. .:....11
:L:..L,LLL,L :L.:.NL..1L,L,L11L,,2L.1LLLL.1LL L11,, ,L. ,:L.
9,211.22,5L,2:L.:,L,.L,.1,.L1,.,,,LLLL.:

L,:,,L.L,..1LLm:,.LL,.L,....L,.LL,.
:,..1L1,L...L1.L2LL,L,12:1L
LL.1:: ,L11,1L.L, 51 :L2

     

5:
5
x
5
1
.:L
$22
2L

1:L1

      

 

 

LLa

. :L, 2.:,1L,,.., .....5

   

,:

1 .L.1LL5,LL..L,L5:,LL,L:,:,L22L1L..L,L2,,.. 51:1,:11

 

1

:,,::,1: ,1:1,1.,,L1 .
::.1L.Lm:, :,L, .1. LL, :,:...L.L,L..L.: :,5... :1,L.
:,,:,1: 21.15,. 1: , .:,2.51:..LL1:.L, 2

1

,

3L

 

,,L ,,L

  

1
1’
58,
,:

L1.:L15L,L,L,LL.,:L 2:11., .1 . :L.:,
L 22.1255
.55..11111.112
12.5215. . L: ,.:5 . L2,:,LL21
1.5.1.... ,L:L52 L,

, L,:,..:1,L11:..,L,.:,.,L.1L.1L1,L,.LL,L,2.551.::,.::..LL11,.LL1L.L..,..1115...L1,52.,LL,,.LLL,LL,,LL1 L
1, L1,:L:,..1LL::,.,m,1LL,1L,LL.:,L,L.,,.LL.,::,:,:...HLL2L1:
5:, L 21: .1..5115 1. 1,
. 2.1:..,LL,L1,11L1., L..,:L...L,..15L.HLL.,
.. 11111.11211...221.21
1,, ,,LL.,,.LL.,.L.,L1.L,1LL2,1.L1,.LLLL1,.,1L,L.LL.L,1.L1L51,2,1LLL.
L ,::11.2.1,.: 5 , 1
1.:,,LL..1L,1L,.Lm........NLL.L,.L,L..L,,LL5,.LL,:H.L..mL1

1

L1L..1L1,.LLL,L.L,,,.LL,.1..L,.L,,LL,1. 1 :L.,LL1L.L.

1,5,:

12
21:12,
151,125: --

    

s

,LLL,.LLL,,1LL..,.LLLL.,L,L2

:5.

’1

1.11515,

,11

L1,
1

5.21.11..:LL.,.L1,,.L,L1,LLL5.L,,1LL,L,
51.: :.. 51:25:25:wa L1L5 L ,,.:

.:.,LLL, :,:LLL.,,.LL,L,LL5,LL.,LL11L:,,.:,1:,:11L,,1L,55L.,...1L,L.,,LL.1,,.LL1,,.:1:,5

5

 

                
                

         

:5
,.,
21
15
'1:
:1

5
1,~L15

 

  

2 :1,.L12 .:,15 L, L1LL..LL .

2
:5
,L:
32‘,
L11::
,L

.L 1'5:
1 1 1.
sL1,

LLL..,.LLL,5, ...... ..::,.: ..:2 L

'8,
85,

15:2 ,5 L:,.L,...LLL1,5L,11L..LL,L.L,

s

...L.1,.L 21.12.5112 1:1, 5:,:2 ,L:

     

           

:L15L~”

,L:1L.L2 511.5. .1251... :.,.:1 ,:.1L.LLL1,.L ,,2 1

2
’1
1
2,
2
1
51

  

L1.... ,:,:.LL,.

1.

1:11... 1.11. ,1L1:,1:,L: 2: 52112
:1L,.,LLL., .:L,1LL1.L51,15:,1:,:...w,......L..1L,L..L,.LL.,,5.,,5L.L,1LL,.,LL..5,.L1..,2L1LL:,:L.

           

              

 

715-
11L

5
71
5

'2
12.
i1
,2

1......
..2111.,m.LL1mmww.L.HLLLL.,1L
L 11552551 ,:L..:L,5L1,.L.L :,:,..:...1...
L L1LL2,1LL.L,L:L.L,L,..L.,2L.,1LL.L,..LL.,52......L1m5m...

1.

15

5..5. . :L1:1,5:. .5.
:,.:,:,.211.1.2.1.12251m21.211...
51255111.. 1.21112... 5 1.
L,: L11,,.1L.::1L.:L...1HL..L,.1,..1...L,L,..2,L.L ,..:,2LL,L.1. L51:...L1,.LL1L..1LL,.,L51:?

E'§
5
3,

                        

121

1, . :,:...,:1, 1L,:.L
. 555 1.5121 ..,,

1

  
  

 

,:.

15511 :1 L: 15 1121.... 55

                            

1:2,L11,L L.,.LL11,,5.. ..L.51LLL1.L..,,1.,1L1,5L,1,.2,LL.mL1..L.LLfiL,,:,1,L,1L 2,5115 ,..,: L,.:.

1

’1

1
KL

    

 

221.1,, L: L
:11:. ,:L5 L51:: ,,1.

1:51
5

L. , .11 L

:L:.,.:1L5.L . 5 .. L L: 11
12:1. .11.. .:. L. 1,51 5 L ,.:,.L g... 25L,

5’5
51:

       

1,
£8;
’1
5:

, L,:LL.LL, :L:.1.1.L1L,5.,1L,5,...,L5LL1L,L.L,L1LLw:,L11LL,L .,.
.. 25:21:12.1... L,L,2,,L,., :,:,:..1L,L.

1111

H1 5.5,,LL1L1LLL1L1
5 :.1,:,,1 :1,2:2 .

   

,...

                                  

   

2L

1:.,1. :.1,::1L1 2 ...

1

L: ,, L

L: ,::,,.L:, 1152115

:,,1:: ,L.

L. :21L,,LLL.L,L.:

~, 1::

:1,:1

...11 L
1,. LL..1LL1.L ,LQL, LL1 LL11 L.

        

 

5

:L:,:.1.

L5 :5,:L.,,,LL,2L.1L..,,,L: L
11:1 1.55.1 L:
:L,,LL.,L.L...,L,,11L,L,L.L, 511.2111. 2:, ,
L..,,LLLL,L,11L15LL,NL.2.,,,2.,L,.,L1LL...L.,L,.LL,,L,2,LL5:15 :,:,:.LL1L51LW1. :,:...:1...
,1, L 5.2.5.551...5.5.5....:.,L
111....

5121.111

L:1

 

 

5-112

:L,5 5 ,, L,L..L 1.5

L. ..,:L:: 11,. :,1: 11:2“. ,1 L ,,1LL1.,L;L,
1,1 1 1.151155, 12111.. L.:, :,.,L. , L , L
:.,5,,L,.,, L:,L.:m1L51:L...5L»L,L.LLmfi5L,L1LL.5L.,L.L .:,1:1.25 L1,.

1
1'1;
5

,, L .:.:L,LLL,LL,LLL.1.2..,L5 5,:L,L.LL,15LLL115.1,,.LLL1.:L,5,,:12L,515,21

.1:,1,LLL.,L1, :,2LL,L..L,5.1LL..,

5
5

.:LL1 LL5LL,L,..
..:. 5515.2... L,

1
,, :L1;
'11

1151

,1:,5,L.L1,,.,LL.L,LL.,.L.,L,.1,.. 2

 

a

   

,L

.L 1 ,.: 1,:,1LL,L.L, 1.

          

 

  

,:.1:L. L..:L:; 1, :,1:LL,:LL.. L:L,,LL,,,..L,,L.,.L1,LL,.., 1, :L:1 .,,,5. L.,...LL...LL.L.,LL.L,5L,,,2L,1.1

 

3L. 1'
11 :1
2L ‘1

,.: :,5, LL ,,.:,L,1L,. :1,,

5.1,: ..:.1.

                                              

L, :,1 .11. 5,:LL:.,.:,1LL.LL11L,,.LLL.,L.L.L,.LL,.L,.L,LL.LLL.1::,L.:,LL1,.L,,.L1,L1LL1L1 LL :1 ....11. :L1 L1:,

, 12 :1,:L.L, :,.:1L..L,L.L.1,.L,,LLL

L, 52.11.1151 ,1: 1, 1,.:,L,..1LL.LL..L,,L...H.L2L,L.,,1L,5.2,.:. L,1L.,,.L,15,,25:,111:, LL21: 1L,:L.,.LL1,5LL,L...1LL.,,LL,:,,.. LL
.::1..: .2 1,151. $9.1 51:. :,5 :,L. 1:.; ,::1,.L.,2,LL,,..L ,,:12 L: L: , . 11: L ::,1

2
'5

..L 1.21.
1,1:,L5,5L 1: ,..:...LLLL, L 1 :1:,511, L L,
.:,..2,1.L,.,L,L..12..L,2LL...,., 11:15:51.1, L,1:,L.L,L.LLL,,.LL...L,1LL,HL1,..L,L1L,L..L,,L L ,, 11:225.:

. L 1.:,. .1 2 , L L,1:
, , .11 L ,.L1,LLL.L.1,L,,1 L.: 1. :::.1., L1,515LL1..L L51:..LLLL.,
L:..LL,.LLL,L.L,LLL.1LLL,...L :,.:L, ,LL.,.LL1L1.L, L2,:,,LL,,,...L,L.LL,,.L,,.. ,.:,::,.,11,L1:,L1L2,11,5,1:,1LL.1:LL,..L.1LL,L.2,:L1

:.LL 1, L 2 :1:: ,5 :1,; L. ,1: L5: L 1L1L5. L , L11.

'2,

,.: 5 L .L :,,.:,1LL1...H,L.L,L15,1LLL.,. :.L.

.2.: ,..,L,

,,L

..L. :12, ::L. 11...... .1111
2,::1.L,L, L.: L 51.2.55
.,.:.1L,L.1L.L,1L,L..L,.L,L.,,1L..L,5,1.L , 5....151155
.1 2.1.11: 5 L,
L, :1:L5L,,L,L ..::,NLLL..,,LLL.,.L,,.LL5L,L.L5,.LL5:L.,1,..LL,L1,LL11.LL,.. 1,,1,:1 2,:,,..L1
L..,L:.L.,.L.1 :,:.1L,,.LL1L5L,,L2,.LL1,L 5:1

  

15255.11...

             
                                                                                                                                                                                                                                   

:..L,.LLL.,.L;LL.LL,1LL,L..,1LL, ,:1.:, :,1

2:

 
 

.:, :L:.,.:L.L,.

1.15.1... 1.1... 1 ,
1L,:.Lm...:,LL5L1L,L.25LLm:L2L.L,LLL,

2

 

..::,L11L 5.51:1. :,..:,L51L1...L,1L,L5,L,L,2.

.::.L111. L..:: ,:,:5

,,,.5.1L,...1LL.L,,.L 1.11551 L
11.2.5.5...53. 1:15

5
25

,,

:L2 :,,..LLL...L.,LL, L 1:,,::,5L,

L L:,L.,LL.,.LL,,..,1L,.L,,LL.1,,.LL2L:
511,11:1:51:22...
.::,5151 ,1L: 1 ,,..1L

. L,.:,L,L.,,.LL11.LL5,1

  

.:L1L.L1,2.LL.L,1L, 1.
5. LL.,,:L.L,L1

:5:
,

          

’1 E1.

'8, :5'2

1’ 3’

(1'2 ,

5:111“

1 L2: :51, L

L11

LI1'1L7‘1‘1‘
1.5
Lié’21,
1:511
1’11
1 .52
11

LLL,,:L,1.L 3:, 5.11.: :,;L1,L,.L,1..,L,1L

L.:,LL,L,.L1 :L1... 5 5,:,,:, 511, ,:1,. 2,: L,
1 .:,,..L, 5:11, 1.2 151 555 :1, 2 L ,1
:,1L%, ...,: ...,..12L1L.LLLL.,1LL1L.L.,1 2:15,: :1.,LL..1LL1,1L,L,.Lw,..L1L,..1151115.1.. ..15..2m.21...2.m...........2252
.251.5.. LLL5..5,.L..,.L,1L2,,.L.15L1L,.1..:1, :.L,,1LLL1.,L1,.LL1 11 ,1:,LL.L,L.,.,L,,1.15.L2,L,L55L

                            

,..1L,L.E§LL,1L,2..L;1; :,,::,1:.,,LL,.L,1LLL2.11,2L,, .:,L1L,L,.LLL,.,..LLLL,.LLL5.L1L1..L1LL

 

5.1.2.5:L,L..;.LHL..L,L,2L,,L,LL2,.LL.L1 5,151L1L5,LL5L,L..,LLL.:.LLLL..L1,L,,5L1L5.HL,.L,,..LL.,,L.L,.LL:1L..NLLLL.LL2,5.L.1.ML5L111.1.

,L

1

‘5;
L”11
1,
5:

:1
1:11
:1

118111:

1:11

11811:
LL11 1 L1

1 LL1,L.L.HL1L,,.:....12..,1L,L11.L1.LL..1L,L.1L,,2L,1,L .1.L
.21....1
1..5 1...... 1,:,
::1,1:1...L12,,m..L1L,N:L,LL.,,5.LL5L,L,L,.L,.1L ,:.1:,: L:
15.11.11.1..51.5..11.15. 1
L 2: .125...15....5
1 1.5.12.2. 12155.2
L L: .:,.,.L.,L.L,,1...5LL11LL5.,L,5515. :,:,25L.,.L,L,.L5,1L,5LLL5
2151.115 :..,.L,L..:,:,5L1 1,5511...L.,:L11,L.L,,L,L.1L
.111...1..

:,:1 :L,,.LL1,21.2,.1,s111.115152 1,.L.L1,.LL2,.LL1,,.LLL,. L$21,125,L1,,1LL,,....L.12,,,12,..21111:
L 5: L 1 5551555111 L. :,:,825 1152325 2:...

25
1'

,

22

5
LL11

,:,.L:,..,. 5:15,.
.,...515 2..

5

1:. 1: ,.:,L,:.:,,LL.,,.LL,L,1L,,LLL11.L,.L,.,.LL

 

2
i5
:1
L:1?
E
'1

1:,, L :1:.

i

5.. 51:
512..
L 2.,.,,LL.51LL.,,2,L1,.LL.L,,

L ... LLL..LL,L5,.L
.::1L.. :..,:1, ..:L.,

    

:i’1.~'
11

’11
51:1
111

21LL,.w:,.LL1L.,L1,1, L11: 11:.LL1,.L.1LL1,
,..:, .11152:

, :.:..1L1LL,,,LH,LLL.L,L,.LLHL,L,.L.,1LLL..,LL.2,.M....,L.L.,.1,L,NL,LL..L,,L,
5111.511 5.

 

:.LL:1.L:1,1L,,L,:1 ,5 L..:L 2 1512..

  

.1 :52
27,12
'2

11:5
'2
1

i

'1

 

1:

  

    

%
5

.:1LL.1.,.L1 .

11

      

,,1

.. :,L..:,:,:2L1....L.1LL,1LL,L,L1L,,.LL..L,L.L,LL,.

... :L.. :2

..1:2 :,:...:1

2

L 5 L: ,5: LL1

L L 51:11LL,1LL..1L,5..LL1,51,.,L
25LL,1LLL...L...L5,.L.15L,5. L2,12.,LL.1L,11,.,,1.L,.LLML,LL.1
.1.:1L,L,:.,1.,21.., 121,5..1: , 11.12.125.255:
1151.5112111151555 ,.LL.51 52:15.1:
1 ,..LL..,L,.L11,L.L 2212.: 5:511: 555115 1

.:,...LL:,.L,L L,:L..,.L ..:,:1, L1

5

,, LLL ,5 L:

         

5

N 11.11551.2
1:25,...15NLL51L5, L:1L..LL, L..:L 115.5,

2515...... 2,512..
11:,:L.,L,..L.,1L,L.L,.L.,,1.L.L,L.1L ...L:,L,L,L,L.L,L.NLL.,5

:,,:,2, :L1

1::L1:
1; 3
,L L

'1

,:.2:,
5

1,5L,1,LL,,..LL.NL12

2

 

 

                                 
   

  

’5
'1

L2,,L,.L,,,L,L.L,,.,L5..LL5..LL,LL51:.151

  

'2
,1:
1...

5152511111. 2L5,1L,2L.L...L1L1,1LL1LL5,
.,:1, 11 1,1:,: 3,5,1 L .

5:
L,L

5
,L

1 1
,,L

,....2211
,, :,1,L,..L: :.1.,.LL11L.,.,1L.,L,L1.L,1...L,LL.LL.,, LLL.,,.1LL
LL ,:LL,,1,,LL1, L,.. , . LLL5,,.LLL :,:LL.LL,,
1:1L... ,L:,,.::.,.::31:15.21,LLLL1,1M2L,L12.L.1.5L1
5:15,:,,..,LL2,,LLL1,.L,1,L

L ,1 1555.11.51.11. 1.,
1 L 1:.LL,L,L.L,LL1, L: ..;.5.LLL,.1.L1LL.,,:1 1L: L5.1.LL.,L,L1,L.1LL,LLL.51,1L,LL ,..,LL1L,LL,.,11. 2.
L,5.:,:LL,.,..,.,L,L.LL..,LL..,,21:15:12.5: ,5 5 5: .55 2:125

1, LL ..512.12.5.121
:L:....LL1,, 22.. L: 552,151,...“ 1 ..1:.1,:

L,LL5

1 L..,1L,L.L,.LL.,5LLL,,5:L1,:.LL..,.,L,1LLL.L,,LL.,1LL,...1LL,5
112. 1: 21 ::.,L. ., :.., .:....12..LLL..1LL,L.:,L,LLL11L,1L,.1L5L.L
L .211...52..1,.1.
2111: L 1.2. 1.2 ,:..L,1L,L..

1' 1
L :5
1

 

L2

'5;

:,.1.ML,1L,51252 L5 :.:1

L1 :,:.1 5:115,

:L,.:,:,5L, LLL.,,L,,.LLL,1LLL,L.1L 1.515. :,:LLL1,, L
2: 15.1

5

1LHLL,wL:,§,JL,wL,, ,,
2:,.1 5,:LL.,,.LL 2,:1,,:51 :1,:LL1.5LL,L :1L: ,1.. L L
....15.1 ., L251: 5

1'2
1

   

a

5,1: L,...L,L.,..LLL..5.,N.1.L

   

.,:
$7
,,1
1

5,111

 

2
5

L2:,

  

12
L22
:1
L1
11
5
1
22:
1

 

. 111: 1:1,: L,
:,L :..,:111.151..5.1211552.51...
,,.. :,LL :5,:1115 ::...L1LL,,51L 21.15.L.L.,5..LL,..L L.,1LL,L1

11
11'

   

2

L, 15.5.15522

:1
2 11
2,: ‘5

2 :L,5L1L,.. L.1 . :11LL,LL
:L1,:5.1L15.,,1LLL 1:,:,:L,1,1,L,..,L2,LL.11 LL.L,L,L.. ,
L L,: m5,L,L..1L.L1L,L..L,L1LL5L5LL.,..LUL, 1

’s"

                                     

11
5,

 

5

.L,,;. :,L..:,1L,_L1L15,,2.,1:5,

1~
,1,:
1

.,,L
'2

:..,:5 L ,
,,..L 8: .,LLL1L.L,LLL,.
.L: LL:.,.L.L1L,5.

,L
’5
i
Li1'
,,L
LL15

1

1‘53?

.51..

1

L 1L5 11 L, .: 211,115.11... 1 ,L
1 L 2.2.1.1.:2.31.52121151124215 ,5....HL1,.LL1,.L,L1,LL ,.. 11.5.51 :L.:.,.LLL.L,.2115:LL1,.LLL1..5.wL,m,..L...,1L,1.,M5C21 5,,:11: 5.51111

,,L

. , .. :1.:1,.,,LLL..,..L,:,5.,L,1L,L1.,.LL1.L,1,
.1 1155:, 1::,,..,1:,5,
LL..L..11L,1. 1,:1L,.L,1:1 5.:: L,.:, 1

1;
L:1,

        

 

2

,L

1 1.:1,:L1.1,2:11.L1:11. 552.15
15,: ..11151. ., ,
.2. ...55 .1L:1,LL.L. ..1 ..:,...LL,.,,LL,1 1,125,251.. L 1:1.L.L..

1,1,1

  

L11~
112
,,1

2
:1’
12
'5

L L 25%.. :..:, ,L:,fimm.,512L.LnL,1L,L.L52,5...L,5MWL,,1,L.LL,L,LL.
11.. L 1 . .1 51,2..,.:,1LL..1.LL1L.L.,1LL.L,.5,LL12:15.5...L,.,L,.LLL.,,.L ,.
51L,51.125.11,:L.:..L,15,1,..LL.,1.:1.,:L...L,5L,.L,L.L1LL,.L1L..LL.1.LL,1L, ,

52.15.22 51511

22.. :.,.:1: 11.1.2512. 31152531122211 .L .,
1.1.5u.....2.51 1

:5
:1
’3

11: 1.2.1 .25225

'1
E
5
1:'
11
2,1

1.2

.....551..1.255115
.L: 2 .15 2.5:, :,:,.L1L,L.LL1,5.,L.L1L, L,L..L.L,LL1L..,..LL,NL,..1L,u..1,LL.LL,1,L..1LLL1. 5 1
:,:..,.L L.:1,LLL.L5LLLL..1,L,L.,,.LLL,L11LL, .1 L.:..L1, :L..

:1:.1...LLL.:,1L

1:1
5.1,
1*
2
2:2

L.:,:,:...,1L...LL.,,.L1,5,:,,1.,,.LLL..,5HL,L.LL,L,...1LL5L,1LL,L.LL,,5....» :L:.1.L, 1.5.5.... L
L.: 11:,:.1.111,..:L,...,L5, 21515.1...L1L115 L211:
:L.LL,L.1L,.L,L,LLLL:.L,:,,:,,L5L,1LL,5.LL 5

....1 1:L.
52.1.52... 1:515:12.

11,:

,1 115
2

,....,1,5LL,..,LLL,,.LL,LL$5,512,. 51,151.12 :1,2: L:,, .
5....L2L.L.1..L1L,L.5..,,5,.:,,LL..:u:,L1L, .:.:L.L,.LL.,..2M5.,,11LL,L,1, L,. ; :,..:mLLLL5LLmL. 15.2.1.1:M2L15L,
25 1352:1111 2.1:“. 111.52.221.25
, ..111.5..25

51L,

1

11:2:
21:15:21
11
,,,:
1L

'§

1
'5
11.2

251215151 ,:L..,L.1L,L.L1L.L.L.

5
5

'1
1

2L

5:
L 255551521211

.5,

1,551,111.15 ,
.1,:1L,1L,L..LL..L,.L,L1..,.LL.,,1L1
.1551. .:1.,; 2 ,
L,,1L5...LL.,,1.1L.LL.5.L,L.LL,2:,..L1L5,LL, :1L11L,1L ,:
1:., L: ..:LL..: :,:5 ,1 L,:L5:,,.. 551
L.: ,5L.,...: ..1. , 55.551511. 1L
,: 2......12111..55112
55555.12

.,.
2'2

1

L: 11.11551...
1:.. LL..:L1L.,L 2252

 

52.253... ... 511.5531
..,.L: ,L: 1
.:, L:,... :LL

1
1.

5

15.1551 :11:
L 5......1525...
1::,..515515. :L.L,:

           

2

LsLL:

1

2.
E2
1

  

5,512.,

,::L
111,51
:,L 111,
L211:
8111
,L
1
:31,

”55

5

'2

5

:: i :1
11:11,
1:2,
51
~12

 

1.

  

  

L
1'1

12

215
5
5
5

LL5LLL.L,5LL1L2L,NLLL.12,1L1,51L,,.:11

21L:
,5
’2

           

  

.1,.1:L12,1L ,L.:,
11522;

  

51:11,
'52
21:2
:5
5
211:1»:

.:L“...

1
111

1:,.:. 1:1L...L51:..,..L,.,L.2,L1.LL L

11

L,1:,.L1,.W.L.,..1/1..L1LL ,,:,12.11LL.L,1211L,.L,L,..L..
2.5 ,2, 51,L1,..L,51LL.L,.L,L,,::1: .,:1

5.1.1... L .11,.
551.2 12
1::,, . ,.1 2 511.151.151.15

   

,

  

 

1

    

.1.L; 155211.:,...L....LL151511,5.,LL5,..

.,L.

..L. ,L

5

5.51.5125...

.151 51152:.
:LHHL.W52LL5:,,L,L .. 2

5 L 151115 ,2:,
1.,:L2,,1,.. 51:51:55.5}:

1

51.211151
,,:5 L :5, ,::1,.1; L11LL.1.1., L2.:1,1L1L, 15.11.: 1:51.,

,

 

52.1251 ,21, .2
1 L 1:,”... 15552251....

1.. ,5 1,.. 1122.....2
1.21. :,,:.15,L11L,1L 5:25: 1..., ,

 

5

.11. 15.22. ,:,.L,L1L,

15

       

       

L2

51
11‘

L2
38,:
5:2
1'?

 

.,L

2
L11

1 ,:.,L.., .,,.L :L., 5,115.32. ..

L, .1:,1/L.

 

5151‘:
5L? :'

2
5
i5:

2
151

1 L,1L
.:,LLL,,.LL,,.L

11
‘9‘,
’1:
:8
2;
ii

§
‘83; 1
L5
5

    

1
1
1'.
,1
5;

 

,..5L,5L,,1L,L :,.:L...LL...,L,.1...,L..1.. L,L...,L.L.,1L.. :,:.:,“..L,.1:,L.LL, ,.:...221..L,LL,11:,L1, 1

                 

5
L2 :1
:,1:121

51......52...

L5. 5LL..LL,
, L1:L,1L,..LL1 5
LL,L,L1LL11LL.L.LL.1LLL1L,1 5.2.11 2 L
5.,.LL..L5L,L..L.LL.,1L1L,5LLL5L.5.L5,1511 L ,,.L,LL,1:L1L, :,:1: ..L L2:,L,1,

L 1 :,1L...
L.L,,12,L.2, L. ,.:.121 1

LL2‘
'5

L.:..LLL2...LL,.L,.

1
5

1..:: L2:,LL,

’21

111

  

15.;LHLWLHMNH5MHJH ,L
1 L1,1L11.
151.115,:

215,52, 11.2.

,12

,,.;
:,L

11: 12
5:

1.

WW1. 1. :L..:.

'2

:1L ,.:. 121:5... ,:...1: 1
12255155125

 

,.
L1
1

11.15 .
L...,L,LL..L ...515 .151
15515:..125515. .
..1211..15515515

15~
3,”?
1:4
'1:

:.1
:1»,
,

5L

5;
2

L,L

L, L, ..555...

53
'11
L::.;
1'

51:8,
LL15
12:3

:,.:1LLL...LL11.1,1L,LLL..L, L: L21: :..1LLL1: L1
:,.:L5n::.....LL.L.L.L1LL L5: 1.25.1.25152 1
5.5121152.

          

15

 

.:1,L,..LL,,1L,L1:.:,L11LNL,LL5L,1.,.L1L.H,LL1LL,.L..L1.1.,LL.L,
...:,1,.5LLL.L.,.L.,1L ,:8, 5,15,

::,,5 :,.,1LL2L.:, ,1,
2:,:11L1L..L.L,LL,.,..LLL,:5..512:1,1LL..,L. L,1:,LL..5L,L1L,.L,L.LLL1L,L,L,L,.,,2L.152:...me,.:..LL.LL..LL,L,.WL1,11L1.LLL

L L ,:::.1,L.:

.15.. 51
1512.55.21.11: ,5.

LL: 52.513.12.53
L ,..:.L5,LL,51L,..L,,,L,L5 L:1L,.

L:1"

:.L, L1L,.

1,.

L550 L,
21:11:52.1.

L 1.. L 5£12515.2152 L5 :L., :,L..: ,,.L,,L,L.,.L1L.L,L1L. L:1L,:L.L,,L51L1LL,.LL,L.L,L.,, L:,5L,LL2L1LL2..L1,1.LLLL.11LL1, L :5.

,..
5

11
,L
18,
,,L

                                                         

.515, 15 L1 25.11.11
.:.:51, 5511121 L211: 215 L, :1,
L,,L: .,1L..1.1 1:,.1,:L....LL,555121..L;,.L2L.H,..m21.:1:,11.,1,LL2 ,:
2.2111.5..51

15321.12... ....LLL..,5.11L.5,..L..1,LL.,1

1: 11.25:», :L1L.5..LLL..L1
:,:111: ,:1,:L .1151. L 11..1..L,5.
1.111111, ,5 2211....511112. .5112151
:L., :L:.L..L,.L.LLHL,L,1L1LL.,2.L,L.NL,L.L,,..,
1::,1L,L.L..,L,L1,,.L.1LL.,,,L,.L :.1.

1..1;
.,L

5

2

1 51

2

             

          

, .:., 55
515. LL
5 5511.

,.:11: ,..,,L1,..,1L,L5,..1L.1L.

15.1.1152 2221.12,: ,
:LL..LL..LL1.L.1LLL...L,1.LL..1,1L5.,LL.L5L 25:22:,
1211111,:1L.L.LL5,1.LL1,L5L.1L5.1, ... .11:

5

11:
1111
:L1
L1:

15 L: 1, 11:11. 2 5 11:,:, ,..:,

       

 

L1 1:,1:. L .,:L52.LLL5..5111.55111, 5:,:1,
L 11 515.111.555111221.111...1.511.525....5.22.1.n.
..L ,.L.L..L1L,1L, ,:1:,L:1L...L1,5..5:LL,.....1,1LLL L..:L..LL,,...L..,2L,H 2.21,, ...55.2,L1,....51,1:L,1Ln5:..1,:1 1,.. L L,:.11,5LL...L....,,.LL,5.,21,L..,.L.L,,.LLL1L5,1.L, ,,.:: :L:1...5,
..1.....1. 1.55255 1. 2.511.1125151 15.55.12.521

11
5
2’1

52::

1.. L51: :1.

1.5.

11112:, , :1, 1.1.1.1112: . .51. ,1:...»

        

LL11...,,.5.L11L,. mu..5au1155512m..m......2;.1115

 

,1
1
,
1'.
11
.,L:

:.,..L1L5.15,L1L ,, :11:,:,:..,.L.151LHLL11L,.1L1L.L.L1,2,,..
5... ,L.L,.L,.1..,

:2 L: ,5. L,1:L..,.1 5.

2

  
  

  

 

,1:LL,5.,. ,.:...LLW5,LL5LLL
L. L 5:,.LL15.L,L5.1L.5,.1L,2.L
:,L.: 5 L 1:L...L1L, :,,LL.1.2L,1L,,LL,...L..,L..Lu,LLL.1L,2L5L..,2L,L,2.LL1L,,5..LL2
..1 ”22522151215522 2.5.5

1 :11:. ,..:,L,L...,:L1, 1,1, 11.,...
,5 , L L,., ;5L1LLL . L,.: 1

,,::,5LL,L,.5L,1LL:2,

1:11.515 115 11,2:111
5.51:5: :55: 51,... L.,1,L.,..LL,1L,..L L: 5

1.:,;

 

is

   

:11
1’1
“11
12

:,,.: .:,15111 L ..1 L
.::151LLL,LL.L 15...; :1L:,L:L1 1L5 ,,.. ,L:,L..L,:1, 1.1.11.5. 11,51,111
22.... ,:.LL...L1L..L1.,..L 55:15:51.: L 5 L:,.LLLL:1C,. .

1:,.:, 51.21.55, L

11

.1.511.211;51.L,:
25.5.2151 L

1L :.1L5,L

,.:,:LLLL2. L

L1L,..LLLL

:.: LL53. ,L.:.12L,111
L L L1L.,,5.2121

5

11:5,..1115 2,521.,

:,:,:1,: 51:21: ,: 5:1: 2.5152
1,,L,L,.,L...L .:1125125
.,LL1,5L,LL,L,5LL,,..1,L.L.L1,1115..55..11..1.5121512...,.,:551 1 .:
:.:1,L,1,.L1525525,:215,:L:,51L5.,L5.55,525,11L1.1L.5,2,L11: 2L 252.11.155. ,1:..1:2.2...151,1..5

’?$LL
1.5111
1:4
51‘
1
1
12
5

..L, 51515 L,:L..,.L1L2

5
5
:15

1.
11

5 ,,L: 1.1:,L,.L,

5155,2512125: 1L:,.:,:2.,,L111,1,1,L,.L,1L1;L,L,.,,LLL1,,LL1,15 5:51:11

2

  

15
$1
'1
'3
E
E

:L.L,,LL11LL15L.11LL
5.1:,1LL1,LL,.L, L 11.21, .511
52: 5325152....5551522. L Mam»... ”2,51,,1LL

,,:1,. L21,:L,2.LL111
1 L12... ,LL,.LL,L1.. L:L5.1.,.L5L,LL,5.1L.2..1 L
L , 5111151,..1111551 ,:,,.111. L:LL11L,: LL.
W5,::,,L.1L..,LL,1L, L,:,1LL.,..1. 5225:1512: 5.15.1351... ..:L: ,5.. :.1,:,.

 

1:11

                    

      

, L :,:,:5: 2122,... L:1.L2L5L,LL5,LL.L,..
L.:;,,..,,...., L. 1

:L,,LL1.L, 1...... L152 , 1L1 L :L1, L11,

555.11.55.155151 1L

9“,,
12
SL
:"s
15:
is”:
5

2

L1:,5.12,L.,.L.L,1.,L,LL.,L .L5 2..151 L1 1::: L1, .1 ,
2.1L...L,1LL,LL2LL.LLL,.L..LLLL5mL,.H1L1L. ::,:L5L. 151.151.. :1:,mL11L§LL.L.51L.L1..,W.,
..LL.L.,2;1,L,5.,,15L,L:,5W5L,.15.152L L ,
1.,1L,L.L5,...L,1LL.,,L,...LL52...L,L.L,1L1L,
.1 .1.... 1...
:,,LL1.L,LLL2,1 :5

12

,,L ,L
'5

11.22 :1.. L,..

  

    

1

5
,:1’5»:

      

51
51

 

51
§
'3
25

:.1.LL.LLL,.L,.L.1.L1L1L1L.,L1 ,, .:1.
115151.252 ,1:..LLLL5 ,
.1....525551525...21 12.155.55.15
:,,..:L, ,, 12122.512 L
1,1,. 5: 11. L1.: 1, LL.1.1LW.1L.L,L..5 2. :L.L,..,..L.1,.11...2 L .11.:1:.
11 :55. 25 :,5: 1,1:,1: LL12L1L111:
L5. ... :,5: 2 ,L,: ,.....L1
11:12:11:1..1L.1.,.L..,L21,L.

5

                                                     

5

,LL... 1,1L 1
:,5,21,L,...5L11L2..:

 

'5
5;

1: 112,215., 1,: 155555.121;

11
12
12

'2

5: L581:

11.

12

:.1... 11115252111... L1,..LL,15,1L51..,55155551115,.
1511.21 :,11
11.55.151.512..,.15..1:2....1,
1. 1:.:,2L1,.LL1L..L51.12:,LL111L5.,,L.:,L L $115.41.: 121112 5115
..L: 2222.12.52; .., :.....5155513.:,1,N515..,552,.L5L5.L,11L5,1L.,5 L: 21.15221251
112....155115 1.5.5:, LL11,22515.12...25.2.1151: 1L.:1L5L1..L11:,,LLLL1LL,L1LL1 :1L:,1:,
1L,:.LL.:.55215.1 1 55115.15 L551L..1,L,..L.L,..LL5LLLMML11L15L.1L5L,
.1,2.L,LL..1LL..1,2LL,L,2.L15L,LLL5L, L, L,: 252.535.”...
..1,L,1L ,11 L, .

.:,1., :..,L

,81:
2:
12
13‘,
11.2
L.
18,2

2

,L

11:,: :1,1:,L2.:,...1LL15
51me2...

52

1

12

12

51:22:.

:15L

1:

1
12

:,5:
2:,:

                    

.,L
151
5
1

1.:,:111:

:1 ~
:1,: ..L

1;:
,1
1
5
'21
2
11
'5
2
5

L1

11,:
5

15,1:

§,

 

5.:
2
1°:

:2

,.

5

,,L

2

.,1. :,:,51: :,LL.: :L2.1L,5 L
L .L:L5,1LL1,2L15,L
n11 L.:,1LL...L,LL5L5L,.5LL.L,1L,L,.,5

1211
.1
,

L5
L3,, LL

1
,.1

5:15.115 ,...LL,L..L1L, L, L

12-
,L11
L1
3%
:11

 

 

, ,. 555.111 1
515.....1... ..:,LLL...LLL1L2.LL2
L ,;L,LL.2,LL.;.L5.1L.L5§.;L

 

11
2

   

 

'5

:,L.: 15:11.;
1 L11..1L, ,. :L:..:1, L
12.5.5152.:L:..L.rL.L,L21.LL5L2L..L,L11..,1L,5 L.

1
8,
6
~15
11»
2;:

:1
,L

>111;

.... , ,
,..:, LL2L15.1LL.,

 

,LL
1‘
5
1

L. .:1,:12
51.15.

 

154,

.52

,... L11:,5L..,,LL1L,111,5 .111,..L1

L21.LL,..:,L,L.,LLL..LL1..L5 1 1:

1.5..L..,5.L.,.LLLL.L.L.,LL. :,.:L .... L. L L

551.15....5515 1,.... L 5.15:1.
5:11, 2 L5.LLL..,,LL,5511,115.:,....511:

121
1

L:1LL5, 511.151.11.11.
:,L..:L :L,.,,1L.L,5..,...L1
.1....1:1:,L,,,L..L.LL.,,...L.1.W.,5,L,

5

,:2
’51

2

.,:5, 25,515 L:,5.1,L.L, L

5

:5L:,12 11.1:2, 2,125.11: 5...... L: ...,..L1:2.:.1.1L11,..H
L ...:L.L,,....L,L.L5L,NLL.L.L2,1:1L..L5L1L,L..1, L ,

., ,:
..,:L1LL..L,,5L,H5,

51.15.15: L

11212.5

1'
:5 ~.

LL11: 121251
L,,1:L.L.L,L5LLL.,..LL..LL,.LL,,1L,L.1.,.L.Nfi11,,1
1111......L...NLLL.1...L;5:1L11L ,,:,5:
21:51: ...

'1: 1..

152.112.215.51

..,.21
115.51.11.12.
5.51

.2
'5

21

L: XL

5,1:1 15:11 . :L..
L, ,L .1155. L .:1LL,1LL,1,.11
1 ,,L1L12,,L,L;.,.LL,11,2.,.1,,LL,,1LL,,,.LL,L;L,1LL 1,1:,:21, ,
15:.2525,L.:1LL1,2.,11.2.L,L,.1, . 55.5 L155:

151.255.5515....1 L :5LL,,L..2,L :5...

 

1151
1,

,L,
12
’2
,:

L:1,5L,,L.L.L,L,5.L.1.,

1 1~15
L5”

L11

~1
1
2
5
’5
.1.
,,L,

L. .1LL.1.1L,.LL1,2L.L,LLLLLL15
5,, 51,...1,L,WL,LL..5L_L1L.21L1L,L.L,2LLL1, 2:1. 5
L 5115.. L
152512211112..5552
81,115,55511251

2. L,L:2,LL.,,

21 5. 5 5.
,5LL251L...515.5.L..5

.WWLL.W5L1,L.L,,L,L.L1L,1 L5

   

,:

1.5115
.,...1LL5,1,L..LL1.L,L,L5L.1,L5. :1 L, L L

. 22,...2,..L.1L..LL.L,1L,5L1L.L.,5,11L1L:”.1121. E11w2smeLéwaw 1::,:L11.
51.21 L L, 1555.11.15.15

        

  

..1 ,:,..L...L2 ..L... 1.1551.
m1551515....12........1

5..125 5551525....

.,.
1
1:
1:8

11%,
1

:,,Ls 1:1,.L.L,21L L ..1 1 ..L: ::1, .1 .:1....11511551 .,:LL1L.
115: 32121555..53.155.535.125...512.....5.1
.n122515555515255.21.1 L L,
L... LL1L51 1:15.12,L:,.L:..L,L,L51.1LL.L...1L,5.LML,L.L,L,..L1LLm:L5L1, 5,1515%... ...,L,

,:,1 L5 1.:,, :L., LL25

51
'2
1», L1
15
18.
L,:
12::
1'
’1
'1
5
:2

      

2%
1

L 2511..::1L,..L,LLL1L1L,..5251,5151

51
.1,: L1

8557
11
21
1
'2
5
1
1'

.,,LL

522...,L..,11LL5,,.1
L ,,..meL. L,LmeuL , L:, w.
1,:.L,L,.L..,5LLL.L,.L,L1L,L.L.H
L,,L.2LL..1L1.,,L,25.,5H52..L.1.,5.L1L.LL,L1L,..WLLLw.21.1

1

1

1.1.1.1...152111.
L 3.5255:

.1
.51:

:3

51
:51
221

:5

51:1
13
5
La:‘:,
.11
5,

L. , 1 1
..1.
...2215.J552,5.1L1L1L2m

5.

   

2
5.
‘5

.:,1L,L.5L1L,L1...1,L5..LL1.55, L ,
5 :L,.,1LL..5.,,..L.L5L,51LLL.1LL,.

L 51
1L
,5,
‘81
~é
,,L
1
11

1:

9r

1:

1

L111

8,:

15
L1:

511

,L

...,...L..1LL1 5152.1,

,L51‘
L15L'
is

5
:12

4:

:12:

mm.
.

21

           

,L
2
11:

5:: ,,5L,,:.1, :,5;.,1., .:LL2 51.512

1'

,2: 251
2.,1LL..LLL..1.1,L .::11.L1L.L,.L5
2525255512215L522555
L....L5..LL..5...L,.L.HL5L,L,L 151152555,

1

,.::L..L1LLL,, 2
..L :.:1:, L

 

12,

g
,
3'8

1
1L
11
«,::12

,1:,1.....1L,.1.LL1,L5,1L,N5.

1

94,:
g:
1

  

,1L...
1.1, L , ...
. :1,L.LL.L, 1.251,,2..L,L.,..L.L.,L11. L,:L5
L ,... :,:,..,.LL,L.L,L,LLL.,,1LL,,..L
2.1,222.,L5,1.L1..21.L1,,5
.:,.L .:..1
, 55.25.
:,1, 25:51:15,... 1,
,,:,5.,L..LLL1,L.,1L.5,L 1.2.5.. 1.5.
L 1, LIM5.L1L2,LLL..W,1.5LLL5.NLLL.LL
,: 21.L,L,2LLL1L1.
51:1. :,:...:1

1:11....211,L1.L,52L,,51LL1.L.

:1,L.L,,LL55L,..LL,1L,L25LLL..,5..L1/..,LL, :.,L,1,. LL,L5..LL.L,1L..LW.LL,,L1,.L,LLL.:.L: 12.21 ::,..LL . L
.1:,1,,1.,,1.LL5 ,1,:1,1,. :52: 5. 21:1,. 1.11... ,5: 11:1, ,1. 5 ,:1.: L

15

:,LL..,.

:5L

L1.
111
LL51
.1

15..
1,:::,1.: 5:,:,: 11 ,,..::,: 11:5: LL2L,LLLL,L,1

'1
1
3

                                     

12
‘2

5
11521
12:
:5
,
'11
12
.2
~81
’2:
1
1
1:
,,1

.5 ,5.:1,L.. L 1
.5 . .1.LL.L.L.M2.L.5,L

51,
>5 1»,
13L
8,,"

2

1
.5
L28

, :..:: :L,...:.11L,.LL1L1,1,:,.::.L1LL.1,,1.L.1 ,:,,.:1
:1 5112151.... ,.L . 112.151..
, 15,22,151:,1L:,LL...,L55L..5,1LLL..,,.L, 15:11.55 L,

1L:...LLL1L1: L...:5LLLL,
::,1 L.,,;1.,.1L

11
,1,.:,
1,;:
L2
51

521225.1
LL.1:,1LL,HL,L.1.,.L....L,L,1LL,1L1. 1..

 

5
5:
5
5
vi
::2

 

1

.25,...L,1,.., , .:L,1:1.2, 1
552:, L1.L,LLL.L.,: ....1....1251....52.15.111
1.5.4.1355:L255,2L1LL5L,L,LLL..,5L.L1L,..L1,.:.1L5LL,1.mLLL1L,LL..L....,..,.5L,.,

1: 1,:, L .51 :,L21,L
L 1.2.2,;1LL1,1L,...5L511

15
,
:5

24115

          

L 51.5.:

~L1
~,1
:5
39;, L

L5

,2. L: 1

.1

5.15.522.

25

W,LL 1:: L ,1 L: :L:,:: L 1:.:,11LL,W,2.
2151,555.12...1L:,,,LL2,55....,...5..1L5 L
L: ,.L5LL251L5...W.51L5L,.LL.2

, L,L..,L111...L ..L: 5,1,5L.LL,,5L;1,L....L.L1, L... 11..:
21515....15515555.LL.L,L,12L5L,.L,L...L1L
.L:, 552,515.15 125...

L,. 15.551.115.51
::...LL,L,L5L..L,L.12 L1,.L1L..,5L,, .2551: w :,:,:,1,.L151,.LL5,
15.155155 :L.. “:.:1: 1.155551...

111.11.,

E:
1.
:3

552,525.51. 211.55. :1L1:.L 22, 5.
2.515. 555.51.151.21.

5L
:1
18:

2.. .1.L...LL1,5L,.,LLL55,L1,H,,1LL.5.L,.1L.L..1.L,:,L1L 5 L,
:,:,:, .L5..,.L1L5L1L1L5L,L1L,5,:.,L5LL.L..L,L,1,.LLL5LL1,L5L:LL.;.L.
1:,.,:,1, 12121,,5.12,L...L,LL25.:,15..5.L., 15112.15: :2 1
1 L 12.1.51:,,L:,,.LL2,5.,LL1,L5.L:,L.21
512.1255... ,..1:,1.L,,L2. 512111,... :...LLL2LL5

11...: 5:1, 5.155., 5 5.15, L,1...:1,L.5..L,L.L.,.1L1,

11
1'1

  

,L..:§1.L,L
,5.:1;

555:,1L..,..LL,L,1LL,,...L,LLL..5
L1 .5 551 .2.52555e1 52155.2...
51:22.52. , 11::L,1.L.LL..LL,L,..L,1.LL..L..,.L.L,1L,L1:1 5.1 .252115555252 L
5 11.5.: ...L1.,:1,.:,...L...1,L.,.55LL55...,..L,1L1.1,..:..11:1.:1:.1,,:.1...L,LL1,25,51,5L,.LL.,5L1,L,5L1LL.1..11.155....1,,5L1,5 ,
,: ,:,.L.LL 1 1555.5. 12.11.. .:.:L... . L5.55....L...,.L11L,L.2LL...LL,,1L5LL1L,L1L,.L.L1L.:,.1L1L,L,2.21.151.2LL1.L.L5.L.w51L.L5
L L 5155:1155:,,:.11.LLL.L,L:,L,L.1.L1LL...1:,1...:,L.LL1.L.,...,LL.LL,.1L,L/L1,.L,.L.L.L,L...L,. L,.,L..,L1L.L,,LL5L2L512.51. 11.15L1L1L.1.,LL.,15L,L,LLLL1L1 5551511,:
5555.15.52.511525212255.51.52.112 1 2.112. 55:11:, . L. .
. ,:.,:L:1...L11,.L.,51LL..15...5.5.11;:,L, 5522 .1 LL51...1.L,5,LLL.L,L,

:,5.L,51.1L:L,151LL51...1L,:1,L
15.1521155555555.225115.2,
..:11LL.: 5521511551525 . ::1L.1L,.L,..,..LL,L..L1:..5.1L..LL,L,L...L15.1,.L.L1,L.,,..LL
2.2.. 5151.21.11. 12......555.

:,:1,:L. L,.:, ... 2 L. ,1.: :,L.:.

 

..:1LL.L,....,..1.,LL.,5L5L,1.L

L1.,

              

5.21.515 2

       

       

5
1

:5

   

X,
5

     

1%
1:12,

                               

2. ..:,2,: :.L, ,1:LL5.:1.1.L1,.115L L.
..,. 15555.1..525552. 55551.5, .:.,1.
11.1.5112, L: 1 51.21551
L L 1.5.5 .1151. 51,2: 2.55.1:LL55.L,2,1L.,5.5.L
15:. :L.. 1 L, :..,, L.

,, 2.51 1515212121,: L
L ,:1.., L11,, 5.155.155.1251 1L11,.L1.:
LL .,:1, L 11221522552255.1212...5 :.:2,,15..1515155 2:15.15 .1.,

    

13,
11,1

             
                                    

  

L1

  

251
1,
'81

,:,..::,1LL.1,L.L1L.,5 52111.5 21.2.11...» , ,
515.115 L: .5 51 15.11 ..5 15155555215
2521152 ....2351.5 5.225

1
::1,1
4:8 1
5
21L
5L,
: 1:1:

111.1151: 1......L.1L,..L..L11.L , 12: :,L.:,5.L LL...L1L.,,,..L,.L..LLL1....51:55:11;
5 .1:;LL1,.1LLL,.L1L 5.1.5.5 5:15....815 1,:,,.. 2 ...: 5:..1,L5.L.1:
1.455.155: .11555....522.2..A2.5. 5111155522, 1.. 51:,L1L1,, L L,1..LL.L.,,LL1,

..,:L ,,:L. L1,, ,.. 5....
51,1:,LL.1.2.,5L,LL2,,.L..1,2. .:5525.L1,:L1, 1,5,1 1
,.....L...LL..LL.L.1,2L55551

  

5
12

 

 

.,.
'1
‘2

,:1

”1’1

11:

18
1.

1.. 1..:... 1:5
1 1.1.... :5L:
1 11L1:. L L

                    

      

 

1 L:1L :,1:1, :15: L ,:,11: 5.: ,:L.., ,:L11L1551: 15.311.512.151...
51.. :1 511,512 :51 51 5:, 55%.... 25:51:51:,.:.1LL.L,5..,L,.,LLL..5L,12w.,..wa.,:.1:,:.w:.1.1:5.1,:,5,.21LL1L11L.L5m5...522:,u2:.1,:5. L
, . ..L :..,1LL,.L1...,LLL, .fifi... 1515.11 L,1L.511115.5111.1251. . L :,:1 1
.115..,1L,LLL.w.5;.1L55L $55....” , 5.512? 5.15:1.52.25125551215512115512.1591 .1LL.,
.5.5..511..n2.1..
,5....25..5nz.1
2215mm...
1,::..

                               

 

L 5.5.15.1, L81:,2.,:, L. 25.2 .525:L5,L,L..:L21L.,L5.1LL..,:.1L..L,L:..,L..LL..L.L,LLL5L,1L.2,1LL5L15152125... :.1,:... :5,
51,51, L 5315.211... 1.2254255:5.5.2512...515,515.L5HLW51L5W1155M25.1L5,

551:... 2253 L, :,:,..:LL2 21.. 1:,..L1 25125111212

::,.:.... L
.111, L 5: 1:1.
1::15,1L,...,11,,1.L,L,5,.L.1.. 1:,

L .LLL15L,1L ,,..,L1.L..1LL,...115LLL,5LLL,5L,L,
..:L11LLL..2.....121L111L...1,

2L

   

,1 5.1.11, L
,.1. 211.11: L1 11115115155 ,:L2.,L5

.,:.2 .1:;

xi,

L1..2...1,..L,1L1LL5,L,5.LL, L,L...1L5,L5,1.LLLL1L15L1L1 .,,5125 ,:,..L 51.22
.5: ,2 5.111, ,:,,,1,:1L,

..1 5 55 1 5 5 2 5 :15L.5.1L.,..LL,NL,5LL,
.: .:L1LLL.L, . :,L . 21:1 2. .:L8 5.1:81
..LL 1:21.151: .51er 5:,: ,, :L,

,
:1
L2
3
8

 

5
'2

   
 

512 1 1:11.512... :,:,:..5...: .1 5
1222.11. 1, 5 251:5.:...L25.LL15L1,.L.ML.L..LL51.L2L 21111512515155 ;.5..11 ..2 2.5.1255. :LL1..1L.11L5.,L15.HL1..LL,1L L5.
1.1.21. 55 115521.11L.,1L,L..,LL;.5.5L,. L 112.1551 ,L .1L...L11..555 55.5.11...L..LL..W,5.1L, ,L......1L2L L
,11,L.,, 5222.5552232w21525211.51.211.5.555.1.511.5115,.. 1..1 . 211.1..1.151..5..N.11.555. .5
5 . .55w2..2.m..5.51......5...m...2.212.215..
LL25.L,2.,LLL.1L.1,;L...,2.,L;1.1L 1.15,: 255.5%,5...:N2..5..1L.1....LLL.W51,,.L.HL..L55.L2:, ,::1,.1, :L1:,1:..1L1 .1
1 .1 1 2:,: ,L11.,1L 15551 .,
L,L. 15:21.1551... 5.5.1.221“..125...1..1 55...
1 ,L,. , 1 :5:,52me..ww..w.21Lm..
1 5.. L1,8.L,,.:.,.«11LL5,1LLL..,1L1,5.,2:,..:, ,:
.. L 11,5... .111LL1L5:L.,,LLLL5L1LL51 ,,.::151:5.1.5.111, . L: ,
:LL 1115.: ,: 55.2.1.5: 15.5.15 5311,1511: 115...: , L: :L:.1,,:21,5L1L1L..L,5: ,: ::L:....1,1L,2LL..L L L1: L , :L,
1,1 :1 .1.L,LL1L..,.L,H 21,5 :,.. L1,521:1,.5L15L1.,L55..Lw 5.555.511.5515m5255fi 21555221551w. 1555.5.25m555222222514115125d1155252
L . L 1;. ,L11, .122 L...us,.:.1L,L,L.HL,5.L 1 53.5.. :215111111: :1. 1:11, :15}, LL .
:,5 5,: 5.15:5..12 1.1251525 .51..1L...,...L.51,L5.1.. L :1, 11.,: L 1 11, ,
1. .512 5121.15.51.51 .215 51.12.25.511, 11:15.: L 11.. ,1 :,1,LL11., 1.2.... L. :L1 1::: ,::1,:,
L,.. ::,:LLL1.L.,...1,5LL.1,5.55. 252115.515. :,..:..m L5151L1,5.,L1,5L 51. 52.515282 L:,L51L5L2522LL32L5W11L2 1 55.155355 15%....

1

   

LL;.L,L::.LL1 515 “:,1:1,5. 25113215551255 1.52.5.1... ..1
..15L:L5..L5L1,.5...M...L5,,Lw55555 125.2555. 1.152....5.L5,L.55,1L5.5.w51fi: 555.5% 5,155,. 5 5.251.555.1555, 55 .
L 8LL5.,5..,1.LL,LL.,1555.511.5L.L 12.155.51LL252552...5:1:5.5.11.2,155212 .2, :L:1:522:51:52.21.15. L
. 1.155.512.5255 55551515525515m25511.515255$552....1525r155fi51255
L ..22L5..,15.5.5L5.5L1.5LL5551L.2.,...1121215555511,1.:..fl,..1,5..:11, 2:22,: L

     

5
if“
L?
L

,L.

115
1%?
'x
2
:1
i
”If,
5
' ‘35
J's'
11

L1.

25. ..,,1L

16,

L L..:L:.2..LL.1,5512,1:11.5.L5, 1,
2.55521 1....5..5.....W55 11151.25... 2 5:5
.122522551.55..
, .22 552522.22
1112...... 5
1 L,....LL,1, 5.11.115
55...... 253212.
.5, 21.5.1: 11,115 ,
1, .11,. .. 21.12 L:
12: 51L:51.5,::L:..15.1L15L1,L..1.L 211

31,251, 5... 1.:L2LL11L,.5L..1L,5,:1 15215:,

1
52L
11,
11
.88,
18,
18.
:2
2,
:1
5
,:1:
5:
E91
51L
4,1
.3:

15.1

1
1
1,

  

1:51,
L8}?

111’s.

151212552511 L.
521155151525. W552... 2 L L,

21
31,
,L
‘3
:5
'5
EL“?
§
.18,

11L.

,3:

’81.
1:
11
,

1:12

18,
.-
1'

..1L,.., 51.5: 11:1,:11:
L.:..LL1.....225
....1

 

1L5. . 1,..
, 1. 1.. L .. L
2111.;1111. 5.1.1., 1

82
12 1‘11
1,1 1C"
2 8.8,;
5 51;:

                

,L
2:8
1’5.
2
:1
a

                  

515.12,...LL5,
121.522
, :,2LL1.1,1L.L1L,LL1L5,.L1L1L1,5L15.1 L,::1LL1L
25... 2252115512... L .
:1 ,..:C ..151521.5.:,....L1,LLL,1LLLLL,1LL1,.L5L

   

2:
5
1
,:5.‘
1221 y

15

,.

225.11 1: L. :,:L5. LL LL... :1

H.

:5

    

L ,::1:,:12 5115...... .L5,11L.2L,,L1:1, L ..L
,..,L,2.LL,,5. 1.5..» 52.5.
L ::,,1 .5... 1 1..:, L

          

    

L,.8'5L8:
..L

52:11.1.

:,,
.,:
1’2
111,13
2

5.... 51: 1
1. 55.5.11: 2.1 11.1.21: ,, 1
52.5.5.2? L5 1151155555....M21 5,.
.. 1: 52555515522115.1121 L.,2LL5.LW.111LL2
55E.:L2...L.,L.L.L,L1LL.L.&5LL15.L,LL52M,$LL1,2L1%1M2L .:1.:L;.L,,5....LLL,2.L,
122.51.115.55 5211... 51.51 15552.2 5... L
1 ,..1L5n21.25.15. :L:1.LL5.11.;22L11L5,1L,L11.1LLL2252525.5 L :1 LwL

    

12,”

L 1:.. .311 L .5 L L :1, LL 11:113.: :1
555%....555 L1 L5.15512LLL....521L115..,5,2.1
L 1. .1 .5.1.21..1112...2.1
:5,. 11.111115 ,:
2...... .L: ......5...5.......
,11....5.LLL2L,;.L.L..5:,5L,, L, :1 N.:, ::.:11L11.L,..L5L,1L
5.1.55. , L

5

              

1

112
E3
”12%
:13,

E?”
LL57
1L1
.1
‘5’:
£1
“,3

’1

  

1
,L
41%
1 ,
:,,L

11

1'5

1
1:1.
5L1.

81

25'2“

 

12

 

1;
i,
. £17,:
8.8L
_.
1
.,L‘L
.
Li
1'54
i’x

,L121
,.

L,...L1LL,L..L1LL..L, ,
:5:...255. .L:

5

                           

1
:,1::
22
L'2
§1
E
,1

  

, ,:.,1:,,1L L
1:1L.1,.., 1,.:,1:1,. :1,2.
,:L....M22L..1n,.:...,..:.15:55:11.”

L,:,1L,L.1.,,..,1L ,5 E L.L

  

             

L,:

12
4,1
g
5:
3"
£323

5:12., 22..

.:.... 1::,: L5

2;
E

      

 

2%:
1

1.515, ,:L... .
.:..LL1L211.1,21
5.2555235

L5
.,11

L..,:L:1,..1L,.,..;L1, 51,...1
555252151:

.: M5.5.25 15:5,:
1,2 L ,.....

  

21:1:

2.5.55
L5: L £21.15: 1.5
.1....15m..5..2

1
'18?
Ffi

5

15:11.5
$5
LL. 5:,.
“53.:
.3111
:5.
L1.-
8.:
11.
'15,:
$375
:2
115,
135155

21.1.1LL5L 1..,
, L 5.1: MNL1LL,11,5.11L..L1,L5.5.5.1221. ......
.,,...uwfiwmflL . .,.2.LL2LL.LL,,1L,2L.5ML1,.LL,LL21,L1L,,L..1
1555152.. L 1 15.12.15... .:,...

.1.... L .1111 ,L51L51L,..1

  

    

5812

12 1
855.,
8:11:81
11:52
.53;
.18
58:1
15
2122
131:
112:3:

1, L.
1115

 

1‘3”
2

     

’2,

,1 L 1:13.15.” LL1L§1L1WILL

1, L2, :::.2;15,1

...,.

                 

5

   

   
   

5,1,1;

1 :51. 5,..

5.15.1: L211?

15‘
1~1

515

:41:
15

21:

, L5 115,511., ..,,.
5L 1:: .152 2.

5

xi?

11.111.51.21...m.%11.55.2..2. L..:L:... L
12; 5:22.11 ,L,L:L2L2L11.1L...L,51L15..L,2::L.115
.1,, L.1,L:: ,1LL11L,.LL L,.: ,L,.:5LL1L:L:L,5....,,..LL,5.1.
L 5.11..52225515551559.5225... 122.51.15.55
1 “511.51.. 112,. :11:...115525

g,
15

             

 

“35 L
x 51%
18‘
51L
,:5
1 ., L15

.,. 11.1.: ,L
1 1525152111..
.2

,11.»...:L:.1L,.,L,1L...L,.,mL,LLL.LL 5:,

12

5.,,5,:1,.:,,LL..,LL1,1,1.,....L 5.1 L L

:,.:L L..1,..L,LL ,5. . , 55 L ,2

5 LL ..1 215115:,..: 211.515.351.51 152215.52 L

1:512, :1,L... 5.1252511. L , 15.11 N

1552 25215255212 ,,..» 15221 L ,

L L ,:1 , L , ,.:L.,..L.L2,.LL.11L»:,,.L..,.L22L1L1.5251512, ...,.L111.5.1.1,5.L1,2 2,1::151.1,525,1L11,u.:.:2«.. ,:L,:,u2.LL,,.LL .,:1,,1

12,11,512 1:21.. .. 25:11,L5L1,L.,1L,L..1.,2.51.51.12.225:151.511.1515. . 1.1215 :L. ,5. 12 55. L: ..122 ,2. . ,,L.:

L: 1:52: L L,:L..,.LLLL1,.L..,.LL.15L,5....1LL5.L,5,1.L,L21..1,,.mL L,...L1L5HL...L1,5.,LL..L,,,LLL,1.1152 51.1.5212.

, , 52.51555... .qu1 51:...515551525512152....... 5255.25.12.25555521
:L..:1LL1:L..,L,:1 L 5,11 ,.15L..,L..L,L,L5..LL.L,5,1 1 1 5.1155525 .:.m, L, 1,:,1L, .1,1L11L1L.m:252:,

                   

15.,
1:33.
:_ L g, 1

, 51152.“; 1,5; 1255M2w55w5w1551552221555m5552w115ww1m521.55521152321515.5 525.51.25.15 L.:
:,:...:LL5W5L1L51111L51 22.15 55.55.13.212... L......2m,515,1LL1L5.L1.1L.: 5.2.511... . L1,: 5 2.15.25.15.51:
. 2.5 5.:11.,1L5....,.L.1LL.1.L,,.,L.,.,,L,1,.L. 511,121.51... LL
51 H .5

1..
1 15,15,221,

1.11:2: .5H3LL1,L:11:,L..L,L..,1.L,LL21LL1
:..L.,5.L5.L211LL511,.LL,5.L 2: 1:5,:

”2:

   

1

5L
52
1
§

5

.553}
L511 51.1: 15 L :55... .1552... :,L,
,....112552515J51551
.:,,1.LLL...55L5..L.12: 1.::L..,LL1,115,L5,.. 2125.22

21
&

.... 2.5

1,L
'5

                

1.1.2.1511: .
251.5. 5:15:11 152.51.515.12
LL, 52.55.55”.
,:155 :1,1::,...21. 31L51.LL1,,L1,L1.:,..1.,L2 L

, 555.511 ,L ,
25155152515521. :,..:

2’8:
'11,.
21;,
28
SL:

'1

, :11:,LL., ,11.“ 255.2
, 11:11..5.;.251.151. 1.,
...,LLL1,1LL,5, 1m21521n15....212m15m«5.252:1,. “1,111,:5
2 :1,:L1,LL1,,.L1,,1LL ,.:L L11:...1LL,....L1L ,L:1....LL111L1:51L2L.LL,.1LL.L,, 25:1,”:
,L, , ,5.... 2525551125
L L 121251.511.:.....L,15L,LL5,

1L5 :,:.1L :1,. L L 1::L,.L,.LL...L 1,..,L,
L. 1,;, 511,2: L1 2..L L: .55.... .1..2..11
1.5.515 15252221151251:,1.1L,...LL1,,5N5L1LL..,.LL,,5LLX,2L.NLL

21.

      

 

:L2.LL1L1L, :.::1,.5LL.L.2,.5LL1, L: 15,1:1,:1L L:
2:11:15. 2.2 :,.:.:255 L

L,
, 5911

31:

       

 

x 1
521 1,112

..1.1..1...1..211..11,11,2.L,,5 55.21

.2

    

  

   

1: ,::L :,1
11 22.55.?
....1.

5,
58:1
5~

.,‘2’:
5

 

11.2.5511? 11.2121. .

:1

...mf5 ,:.:1,..L,,.L,L;,,LL,

125.551.251.15.

, .5.125.5.512515 .1155...5...5...1.115 . 2.. 1 551 .1..

111212...LLL..,1L2L.1L..,1L5..m,5m.21:..,152xfi,wmm§:L2 1L: LL11

151,21... 12:51.1 52.15.22.55212215515512.2122..5555 1.: 55155
15L L2. :18:5 L5:

.1.: . 555 1 5X51

:,1:..1, 2.215221 L 1, L: ,§L.,LL.,,L1.L
5.5251,, 1,.. ...,1L.

1'15
,.Léé,
,.,
2;.
.51
11
£1
5

,

L1L15.
2‘;

1

                          
           
      

   

:5
2.8,

  

.521 ...15

 

‘5,
L38
2

5:

    

2. 21.111. 5531:: .1, ,5.: 5.11:1. :1, :,1:,.L,

,
1
2,
L.
1:
22
.1
1.
$31 ,

  

21
.512

,L

H..:,.:...:L:.::,WLL.LL,.LL
.1:,11L52LLL,.L.LL1L,1L.L,,L,.,

12
$1
1
2
51
1.
1:

158

5

2

L L1: 12.2
5:15,. L1 L1:

:51;

      

                  

, 1 L15. 351.2151 :15,
, 25151.21... 1:,..L1, 52151
1 :,:.:...1 1:15,: 1, , 122 51515
.1, .2 2..2255552.25%.. 51.515552 55225511151555.5131
L 131112515, 5555255122155... 111. 15.5,. 511:5 .

1., L: 5.1.11,

1
S5
15
15
,5,

              

   

., :,..:,HLLL..LL 2215212151515
1 155.21 LL25L5LL...,51.1...LL2.15L

11,1
5
1:
5
1L
2

.2, 1,.... :.1L...

  

151.15 2.5.5.1.?:,5..L.L,5L.L,,1L,L....LL5..1,::L

L,L

1,.

52.25.111.551 55:21.....,1LL..L1,.5L2, 12,
11.1.. 52521.52 1111511? 151,52 ,. 1
11:15.1 55.51.1115,:1112555 ...:L.1L5L1L....,. 11L. .25 5..5L..1LL5.21L2.LL1L,
5:15....L121m55 m :,:15.L1,..L 5.5.1155... 5:155 512.15,. 1,5,,:,L,1:

1
.1
11
1,:.

               
   

,1

 

115521. 5.1
:5, 1111.51.15
L 1.5122212:

                

    

51
,LL 15L

    

 

              

      

              

 

1
,1:
12

.,L

5.11: , :1L, 1, 1,5111 ,1,...L .1 :.:1: .L .1211
..1 L.:...12L .11.... . 1.555.555, . 2L,.L.LLL5.L5. .2 15ww5n5.5125..5.581:2m21§L5:LL

1
1

 

    
 

 

   

   

      

       

L1’s

         

  

 

    

 
 
      

     

 

.. 2 L1LL..LL2 :,,5L,,L1. L1L,.L,,.:,.L, 55. 51112 L
L L 12.... ::,.:.1: 1:2 1:1LLL,L,L
,L 1 ,L . , .:L..,::, .,L 1: 1.,: ....25551.22M52552155552w555m1212.
L5.L........LL.W.5L,.11L12L...L2..L.52.25...”55555255252255... 21: 2515515525851 .L.L5.._...,...: 1 15.25.515.5255152H151 5.15,...

1...,

.5,

 

 

    

 

   

 

 

 

    
 
          

   

     
 

      

    

         

:.1:L

    

    

    

 

   

     

  
    

 

 

 

    

     

 
   

 
        

      

   

        

          

  

        

  

        

   

      

 

     

 

15,:: .1:1:1,.2LL

    

     

   

 

     

   

   

    

 
      
 
 

Calculation of Evapotranspiration

Using Color-Infrared Photography

By jOHN EDWIN JONES

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—0

 

Work done in cooperation with the National Aeronautics
and Space Administration

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1977

 

 

UNITED STATES DEPARTMENT OF THE INTERIOR

CECIL D. ANDRUS, Secretary

GEOLOGICAL SURVEY

V. E. McKelvey, Dirertor

Library of Congress Cataloging in Publication Data

Jones, John Edwin, 1940-
Calculation of evapotranspiration using color-infrared photography.

(Gila River phreatophyte project) (Geological Survey Professional Paper 655-0)

Bibliography: p. 044-045.

Supt. of Docs. N0.: I 19.162655—0

l. Evap0Iranspiration——Measurement. 2. Evapotranspiration—-Remote sensing. 3. Photography, Infrared. 4. Aerial
photography in botany. I. United States. National Aeronautics and Space Administration. II. Title. III. Series.
IV. Series: United States Geological Survey Professional Paper 655—0.

QE75.P9 no. 655-0 [QK873] 557.3’08s [551.5’72’028] 76-608355

 

For sale by the Superintendent of Documents, US. Government Printing Office
Washington, DC. 20402
Stock Number 024-001-02985-5

CONTENTS

 

 

Page Page
Symbols __________________________________________________ IV Identification and measurement of vegetation parameters—
Glossary of terms __________________________________________ V Continued
Abstract __________________________________________________ 01 Vegetation—Continued
Introduction ______-___,1___,_________,____,__' _____________ 1 Signature discrimination—Continued
Purpose and scope of the investigation __________________ 1 Spatial computer analysis ______________________ 026
Location and extent of the study areas __________________ 2 Volume of canopy foliage __________________________ 28
Acknowledgments ____________________________________ 2 Foliar cover ______________________________________ 31
Aerial photography ____________________________________ 2 Techniques for analysis of color—infrared photography ________ 33
Vegetation ____________________________________________ 3 Cost of photographic analysis __________________________ 35
Color-infrared photography as a tool for vegetation analysis __ 4 Basic photographic concepts ____________________________ 35
Color-infrared film ____________________________________ 4 Spectral sensitivity ________________________________ 35
Leaf and canopy reflectance ___________________________ 5 Film density versus exposure ______________________ 35
Visual analysis ________________________________________ 6 Base plus fog density __________________________ 35
Data collection ____________________________________________ 7 Gamma ______________________________________ 35
Photographic data collection ____________________________ 7 Data standardization and sources of variability __________ 35
Hydrologic data collection ______________________________ 10 Analytic optical density ____________________________ 36
Vegetation data collection ______________________________ 1O Film-type correction ______________________________ 36
Identification and measurement of vegetation parameters ____ 1O Flight-altitude correction __________________________ 38
Evapotranspiration ____________________________________ 10 Filter correction __________________________________ 38
Depth to water-soil moisture __________________________ 18 Solar angle correction ______________________________ 40
Vegetation ____________________________________________ 19 Standard correction ________________________________ 40
Signature discrimination __________________________ 19 Statistical analysis of variables ____________________ 42
Time—independent signature ____________________ 21 Summary ________________________________________________ 42
Time-dependent signature ______________________ 24 References cited __________________________________________ 44
ILLUSTRATIONS
Page
FIGURE 1 Map of study area __________________________________________________________________________________________ O3
2 Graph showing spectral sensitivity versus wavelength for the three dye layers of color—infrared film ______________ 4
3 Diagram showing color formation on color-infrared film and its relation to plant reflectance ______________________ 5
4. Graph showing analytic optical density versus log exposure for three dye layers of color-infrared film ____________ 6
5 Graph showing average leaf reflectance and absorption as a function of wavelength ______________________________ 7
6. Photographs showing vegetation in the Gila River flood plain __________________________________________________ 8
7—23. Graphs showing:
7. Transmittance Versus wavelength for the densitometer filters used in film analysis ______________________ 9
8. General percentages of irradiance for different earth-surface characteristics ______________________________ 11
94 Remote sensing and water-budget values of evapotranspiration versus time, Gila River Phreatophyte Project
area ____________________________________________________________________________________________ 13
10. Regression relations between measured remote sensing and water-budget values of evapotranspiration,
reaches 1 and 2, 1968, Gila River Phreatophyte Project ____________________________________________ 14
11. Monthly consumptive—use coefficients for areas of indicated percent of foliar cover of phreatophytes, and
average monthly consumptive-use factor __________________________________________________________ 15
12. Monthly consumptive-use coefficient versus relative near-infrared irradiance ____________________________ 16
13. Consumptive use of water and modified relative near-infrared irradiance versus time for grain sorghum
during 1968 growing season ______________________________________________________________________ 17
14. Regression relation between consumptive use of water and modified relative near-infrared irradiance for grain
sorghum during the 1968 growing season __________________________________________________________ 17
15. Depth to water versus relative near—infrared irradiance, reach 2 0f the Gila River Phreatophyte Project-___ 18
16. Relative near-infrared irradiance, relative red irradiance, evaporation, and water in the capillary zone for a
dense 40-acre (16—ha) saltcedar site, 1968 __________________________________________________________ 19
17. Relative near-infrared irradiance, relative red irradiance, cumulative soil moisture, and depth to water, 1968
and 1969 ________________________________________________________________________________________ 20
18. Regression relations between relative near—infrared irradiance and relative red irradiance among vegetation
types in and adjacent to the Gila River Phreatophyte Project, 1968 and 1969 ________________________ 22

III

 

 

IV CONTENTS
FIGURES 7—23. Graphs showing—Continued Page
19. Regression relations between relative near-infrared irradiance and relative red irradiance among vegetation
types in and adjacent to the Cibecue Ridge area, summer 1971 ______________________________________ 023
20. Relative near-infrared irradiance versus relative red irradiance for five vegetation types in and adjacent to the
Gila River Phreatophyte Project, summer 1968 and 1969 ____________________________________________ 24
21. Relative near-infrared irradiance, relative red irradiance, and model curves, reach 2 of the Gila River
Phreatophyte Project, 1968 and 1969 ______________________________________________________________ 25
22. Regression relation between relative near-infrared irradiance determined from film and from a time-
dependent model equation, reach 2 of the Gila River Phreatophyte Project, 1968—69 __________________ 26
23. Relative near-infrared irradiance and relative red irradiance versus time for five vegetation, types in and
adjacent to the Gila River Phreatophyte Project area, 1968 and 1969 ________________________________ 27
24. Diagram showing computer printout codes used in spatial analysis of relative near-infrared, relative ‘red, and relative
green irradiance ___-_______________________1___-___-___-___-____-__._ ____________________________________ 28
25. Color-infrared mosiac showing spatial computer analysis by spectral signature for 3.67-acre (1.49-ha) plots in reach 2,
Gila River Phreatophyte Project, June 27, 1968 __________________________________________________________ 29
26—33. Graphs showing:
26. Volume of canopy foliage versus relative near-infrared irradiance for photographic flights during 1968, reach
2, Gila River Phreatophyte Project ________________________________________________________________ 30
27. Relative near-infrared irradiance versus time for three saltcedar volumes of canopy foliage classes, reach 2,
Gila River Phreatophyte Project __________________________________________________________________ 31
28. Relative near-infrared irradiance versus time for three mesquite volumes of canopy foliage classes, reach 2,
Gila River Phreatophyte Project __________________________________________________________________ 32
29. Foliar cover versus relative near-infrared irradiance for 50 3.67-acre (1.49-ha) plots, subreach 2b, Gila River
Phreatophyte Project _____________________________________________________________________________ 33
30. Adjusted and unadjusted relative near-infrared irradiance, subreach 2b of the Gila River Phreatophyte Proj-
ect, 1967—71 ______________________________________________________________________________________ 34
31. Major versus the two minor optical densities for the three dye layers of film type 8443 ____________________ 37
32. Filter transmittance versus wavelength for the standard filter pack used in aerial photography in the study
areas ____________________________________________________________________________________________ 39
33. Relative near-infrared irradiance and relative red irradiance versus time of day for bare ground and dense
saltcedar ________________________________________________________________________________________ 41
TABLES
TABLE 1. Time-independent analysis using color-infrared film for vegetation in the Gila River Phreatophyte Project and adjoin-
ing areas ______________________________________________________________________________________________ 021
2. Time-independent analysis using color-infrared film for vegetation in the Cibecue Ridge area and adjacent areas __ 23
3. Atmospheric transmittance at different flight altitudes for the peak wavelengths of dye-layer sensitivity of color-
infrared film ____________________________________________________________________________________________ 38
4. Atmospheric transmittance for the peak sensitivity of dye layers of color-infrared film at summer and winter solstice 40
5. Statistical index of variables in equations used in this report __________________________________________________ 43
SYMBOLS
A Altitude correction factor D Calendar-year day for the end of dormancy, determined by the
(A) Matrix A low relative near-infrared irradiance
Q4") The inversion of matrix A D’ Calendar-year day for the end of dormancy, determined by the
AOD Analytic optical density high relative red irradiance
Subscript—data from an evergreen species of saltcedar E Exposure, defined as E =(I) (t)
(Tamarix aphylla) E Evaporation from bare ground
at Subscript—analytic transmittance E(-y) Energy in ergs/cmz.
B Integral optical density—blue color END Equivalent neutral density
3 Relative green irradiance, in percent ET Evapotranspiration, measurement based on water or energy
C Analytic optical density—~cyan dye layer ,, A budget data
C' Calendar-year day, January 1, C'=1; December 31, 0:365 or ET Evapotranspiration, estimated by relative near-infrared ir-
366 (leap year) radiance
c Subscript—cyan dye layer exp Exponential function of, as in exp (x) for ex
CI Foliar cover, in percent F Filter correction factor
DU Base plus fog density F F test (statistical test)

 

CONTENTS V

f Monthly consumptive-use factor, defined as f = %(f;—)

G Integral optical density—green color

G Relative red irradiance, in percent

0’ Value of the model curve for relative red irradiance

H Amplitude of the sine waves, for relative near-infrared ir-
radiance

H’ Amplitude of the sine wave, for relative red irradiance

 

 

Ah Difference in altitude
A hl Camera altitude
A hz Ground surface elevation
I Illumination
l Modified relative near-infrared irradiance, in percent
10D Integral optical density
J Bridge correction factor
k Monthly empirical consumptive-use coefficient
12’ Monthly empirical consumptive-use coefficient, defined as
being independent of soil evaporation
L Duration, in days for the period of plant vigor, determined by
the change in relative near-infrared irradiance
L’ Duration, in days, for the period of plant vigor, determined by
the change in relative red irradiance
ln Natural logarithm
M Analytic optical density—magenta dye layer
m Subscript—magenta dye layer
fit Subscript—month
. . 4 180°
N Function of time, defined as N = EC —D) ( L )]
. . 4 180°
N’ Function of tlme, defined as N ’= EC —D’) (. L )]
n Number of samples
CD Optical density
0d Subscript—optical density
P Mean relative near-infrared irradiance during dormancy
P' Mean relative red irradiance during dormancy
P Percent of daylight hours in the month
q Subscript—bridge reading for any photographic mission
R Integral optical density—red color
R Relative near—infrared irradiance, in percent
R’ Value of the model curve for relative near-infrared irradiance
r Coefficient of correlation
r Subscript—reach
SD Standard deviation

 

s Subscript—standard bridge reading (June 27, 1968, photog-

raphy) '

sec Secant (of)

sin Sine (of)

Sy.x Standard error of estimate; the standard deviation of residuals
from a regression line

SQ) Spectral sensitivity, defined as S()t)=1/E(}t)

T Transmittance

T’ Atmospheric transmittance of radiation at a particular
wavelength, through the total atmosphere or between the
ground surface elevation and the camera altitude

T Transpiration

t Mean monthly temperature, in degrees Fahrenheit

f Time

t’ Student t test

u Monthly evapotranspiration, in inches

11 Subscript—This subscript indicates that the relative ir-
radiance data (R, G, or B) was from a nonstandard flight
altitude or filter combination and that the data are corrected
to standard values by the equation

Depth to water (table) level, in feet

Analytic optical density—yellow dye layer

Subscript—yellow dye layer

Subscript—~year

The Fisher Z transformation

The maximum and minimum values for Z,’ or Z p as deter-
mined by the t’ distribution

The Fisher Z transformation of a function

The inverse of the Fisher Z transformation of a function

Micrometers

Value of the null hypothesis

Summation

Extinction optical thickness

Solar angle

Function (of)

Matrix notation

Gamma

Subscript—reach 1

Subscript—reach 2

Subscript—subreach 2a

2b Subscript—subreach 2b

8443 Film type 8443

2443 Film type 2443

NW“ as

um

3:
B

gmwazkise-an

GLOSSARY OF TERMS

Aerosol scattering coefficient. A coefficient of particulate scattering
in the atmosphere; particle size may approximate or exceed the
mean wavelength of light.

Analytic optical density. The optical density of the individual film
layers of a multidye layer film.

Base plus fog density. Film density resulting from film base optical
density and chemical fog; the optical density resulting from any
change in exposure is greater than gross fog density.

Densitometer. A device used to measure the optical density of the
dye deposit in a photographic image.

Equivalent neutral density. The visual density of dye layer that a
multidye layer film would have if it were converted to a neutral
gray by superimposing the minimum necessary amounts of dye in
the other dye layers.

Evapotranspiration. Water withdrawn from soil by evaporation and
plant transpiration.

Exposure. The product of the illumination of a unit area of sensitive

 

material and the duration of time through which this illumination
acts, in ergs/cmz.

Extinction optical thickness. The mean value of the sum of the
Rayleigh scattering coefficient, aerosol scattering coefficient, and
the ozone absorption coefficient.

Foliar cover. The amount of ground covered or shaded by a vertical
projection of plant foliage; usually expressed in percent.

Gamma. Slope of the linear portion of the characteristic curve.

Ground scene. Earth surface characteristics perceived both spa-
tially and spectrally.

Ground-truth data. Data obtained independent of a remote sensing
system that describes the object or phenomenon sensed by the sys-
tem.

Integral optical density. The total optical density of a multidye layer
film.

Irradiance. A measure of radiant power per unit area that flows
across or into a surface. The specific definition of the term as used

VI CONTENTS

in this paper is defined as the energy recorded on the film. Addi-
tional clarification of this term is presented in the section "Purpose
and Scope of the Investigation.”

Light. The visible portion of the spectrum, located approximately
between 0.40 and 0.70 micrometers.

Optical density. The negative logarithm to the base 10 of the trans-
mittance. Optical density in a multidye layer film can be described
in two ways, analytic optical density and integral optical density.

Ozone absorption coefficient. The coefficient of ozone absorption in
the atmosphere.

Phreatophytes. A plant that habitually obtains its water supply
from the zone of saturation (water table).

Radiance. The radiant flux per unit solid angle per unit of projected
area of the source.

Rayleigh scattering coefficient. A coefficient of gaseous scattering in
the atmosphere; the gaseous molecules (primarily nitrogen and
oxygen) are smaller than the mean wavelength of light.

 

Reflectance. The ratio of the reflected to the incident radiation.

Sensitivity. The reciprocal of exposure (ergs/cmz) to produce an
equivalent neutral density, usually of 1.0.

Spectral signature. The spectral signature is traditionally defined as
the reflectance characteristics (curve of percent reflectance) that
have been determined for an object. But, because of the nature of
this study, the spectral signature will be defined in this report as
the relationships between the relative irradiances (R , G, and)? ) for
a particular object. These relationsips are variable with time and
changing conditions.

Step tablet. A strip of film or a glass plate whose optical density
diminishes in incremental steps of density used to calibrate film
density.

Transmittance. The ratio of the transmitted to the incident radia-
tion.

GILA RIVER PHREATOPHYTE PROJECT

CALCULATION OF EVAPOTRANSPIRATION USING
COLOR-INFRARED PHOTOGRAPHY

By JOHN EDWIN JONES

ABSTRACT

In the 5-year period 1967—71, 38 color-infrared photographic
missions were flown over the Gila River Phreatophyte Project in
southeastern Arizona. Data from these missions were analyzed to
determine the possibility of identifying and measuring vegetative
parameters and their associated hydrologic variables by spectral
analysis of the photography. During the summer of 1971, additional
data from six color-infrared photographic missions flown over the
Cibecue Ridge Watershed Study in central Arizona were used to test
the validity of some of the techniques developed in this study.

The photographic missions were flown at altitudes between 1,500
and 60,000 feet (460 and 18,000 meters) above land surface during
many different climatic conditions using a variety of cameras, films,
and filter combinations.

A transmittance densitometer was used to obtain density readings
in each of three primary colors of the positive transparencies. The
irradiance (defined in this report as energy recorded on the film)
sensed from the dye concentration in each of these primary colors
was determined and related to the total energy sensed by the film,
achieving three parameters of relative irradiance (near-infrared,
red, and green) which were functionally related to spectral regions
indicative of plant activity. These parameters were then corrected to
a “standard photo flight” by adjusting the densitometric data for
flight altitude, filter combination, film type, and a standard correc-
tion based on the spectral signature of a bridge located on the project.

Calculations of evapotranspiration based on remote sensing from
13 photographic missions flown during 1968 were related to water—
budget measurements for a 1,700-acre (690-hectare) area cleared of
vegetation and a 2,200-acre (890-hectare) phreatophyte-covered area
of the Gila River flood plain.

The coefficients of correlation between the water-budget mea-
surements and the remote sensing calculations were 0.88 for the
cleared area and 0.86 for the phreatophyte-covered area. Evapo-
transpiration calculated from seven photographic missions flown
during 1968 over a spatially homogeneous grain sorghum field gave
a coefficient of correlation of 0.93 when related to evapotranspiration
computed by the Blaney-Criddle equation.

Spatial variability of vegetation on the flood plain was defined by
the discrimination between eight different plant communities in-
cluding both deciduous and perennial species. A spectral evaluation
of near-infrared versus red irradiance of these plant communities
showed that for each community the standard error of the irradiance
terms was less than 10 percent. Determining variations in the depth
to ground water spectrally was not practical, although estimates of
depths to water of less than 30 feet (9 meters) had a relatively small
standard error of 5.4 feet (1.6 meters). Moisture content of the soil
could not be determined spectrally either, but plants under stress
from moisture deficiencies were detectable.

 

Temporal changes in vegetation were evaluated by computer maps
showing the distribution of spectral signatures of the Gila River flood
plain for each photographic mission flown. A mathematical model of
the change of spectral signature with change in season for the flood
plain gave a coefficient of correlation of 0.98 between modeled and
observed densitometric data. When an increase in volume of canopy
for both mesquite and saltcedar was noted in the field, there was a
corresponding increase of near-infrared irradiance, but the mea-
surement errors of both irradiance and canopy volume were very
large. The pooled standard deviations of irradiance for all canopy
classes of mesquite and saltcedar were 7 percent and 14 percent,
respectively. The pooled standard deviations of the 1968 irradiance
data for the largest canopy classes were 13 percent for mesquite and
12 percent for saltcedar. Foliar cover versus irradiance had a high
coefficient of correlation, 0.85.

During this study it was determined that a color-infrared photo-
graphic mission and a computer analysis of the photographic data for
the Gila River Phreatophyte Project area cost about a tenth of the
amount of conventional species classification and canopy-
measurement techniques.

A short discussion of the derived spectral equations and a table of
24 statistical parameters describing the spectral and hydrologic var-
iables is included.

INTRODUCTION

This report is the result of the study of color-infrared
photography as a quantitative vegetative and hy-
drologic monitor in the Gila River Phreatophyte Proj-
ect area and the Cibecue Ridge area. The study is a
part of the comprehensive Gila River Phreatophyte
Project, which was undertaken by the US. Geological
Survey to evaluate evapotranspiration from an
analysis of all the significant components that compose
the hydrologic system. The work was done under the
general supervision of R. C. Culler, project chief.

PURPOSE AND SCOPE OF THE INVESTIGATION

The purpose of this investigation was to develop and
evaluate color-infrared photographic methods and
techniques as a remote sensing tool for the identifica-
tion and spacial measurement of vegetation cover, its
condition, and temporal variability, and to functionally
relate these measurements to quantitative estimates of
hydrologic parameters, such as evapotranspiration,

01

02 GILA RIVER PHREATOPHYTE PROJECT

depth to ground water, and soil moisture. Near-
infrared photographic sensing involves the detection of
the spectral radiance that is reflected from the photo-
graphed surface and is particularly well adapted to hy-
drologic analysis. The sensitivity of two of the three
dye layers of the film corresponds to ranges in
wavelengths that can be used to measure plant status,
and the photographic format lends itself to a visual
evaluation of ground-truth data.

At this point clarification concerning the terminol-
ogy used in the paper is in order. The photographic
data used was in general standardized to a specific
flight altitude (8,500 feet or 2600 meters) above land
surface using a ”standard filter pack.” The analysis in
the paper was, therefore, based on the effect that ir-
radiance had on the film as a result of radiance from a
ground scene and atmospheric “stray light” being col-
lected and focused through the filters and lenses upon
the film. This "at a distance” and “filtered” approach
required the unconventional use of the word irradiance
as a descriptor for the ground scene throughout the
report to distinguish this data from data which are
standardized to actual "ground-scene radiance” at sur-
face elevation. The term "irradiance” as used in this
report is, therefore, defined as the energy recorded on
the film. It includes the reflected solar energy, modified
by the atmosphere, and stray light that reaches the
film, exposes the film, and is processed as the energy
recorded on the film. A more general definition of the
term is given in the section “Glossary.”

The author feels that the nature of the data in this
report is such that if presented as "radiance” the reader
might try to compare the results in this paper directly
with other data which are actual “radiance,” and this
would yield spurious relationships.

LOCATION AND EXTENT OF THE STUDY AREAS

The Gila River Phreatophyte Project area covers a
6,000—acre (2,430-ha) 15-mi (24-km) span of the Gila
River valley immediately above the San Carlos Reser—
voir in south-central Arizona, and the Cibecue Ridge
Watershed Study area includes two small 60—acre (24-
ha) watersheds in the Fort Apache Indian Reservation
in east-central Arizona (fig. 1). For this study, the
15-mi (24—km) span of the Gila River valley was di-
vided into two reaches—1 and 2; reach 2 was further
divided into two subreaches—Za and 2b (fig. 1). The
photographic— and hydrologic-measurement sites in
the reaches of the Gila River Phreatophyte Project area
analyzed in this report are located by a three-number
coordinate system. The reaches are divided by 2,000-ft
(610 m)-square grids, which are numbered 1—28 down-
river and 1—5 cross river, and the grid unit is divided
into 25 plots, each of which has an area of 3.67 acres

 

(1.49 ha). Therefore, each 3.67-acre (1.49 ha) plot is
located by a three-number coordinate system (for
example, 14—2-25), in which the first number of the
coordinate indicates the downriver grid, the second the
cross-river grid, and the third the plot. The location of
the photographic- and hydrologic-measurement sites
in the two small watersheds in the Cibecue Ridge area
are described only by dominant vegetation type.

ACKNOWLEDGMENTS

The interest shown in this study by many people in
the fields of photogrammetry and the optical sciences is
gratefully acknowledged by the writer. Of special help
and value were the suggestions and contributions of
Dr. P. N. Slater, Associate director of the Optical Sci-
ences Center, University of Arizona; Mr. N. L. Fritz,
research associate of the Photographic Research Divi-
sion of Eastman Kodak Co.; and Mr. W. E. Evans,
senior research engineer of the Electronics and Radio
Sciences Division of Stanford Research Institute. Dr.
Ran Gerson of the Department of Geography of the
Hebrew University in Jerusalem, Israel, reviewed the
manuscript and offered many helpful suggestions. The
writer also gratefully acknowledges the generous con-
tributions to this study by the following personnel of
the US. Geological Survey: R. M. Myrick furnished
data for the Cibecue Ridge area and aided greatly in
the computer analyses; R. M. Turner furnished the de-
scriptions of the vegetation and made many helpful
suggestions during the study; D. R. Dawdy aided in the
statistical analyses; and L. B. Leopold made many
helpful suggestions for the improvement of the paper.
R. C. Culler conceived, initiated, and guided the study.

This study would not have been possible without the
data from the aerial photographic missions flown by
the US. Geological Survey and the National Aeronau-
tics and Space Administration. The Survey missions
were flown under the direction of H. E. Skibitzke; the
National Aeronautics and Space Administration
(Johnson Space Center) furnished the high-altitude
photography.

AERIAL PHOTOGRAPHY

Data from 36 color-infrared photographic missions
flown by the US. Geological Survey over the Gila
River Phreatophyte Project area were analyzed for this
study. In addition, data from two NASA (National
Aeronautics and Space Administration) photographic
missions flown over the area were analyzed. The US.
Geological Survey missions over the project area were
flown at altitudes of 1,500, 3,600, and 8,500 ft (460,
1,100, and 2,600 m). Trial missions were flown at al-
titudes of 500 and 1,000 ft (150 and 300 In) on April 5,
1969, and at an altitude of 11,000 fl: (3,300 m) on

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

EXPLANATION

5

Boundary of study area for reaches

Q
Location of study area

Peridot

  
 
 
  
  

 
  

Dense R sat/:27?”
Saltcedar snte site 13L2 (-

  

SAN CARLOS

RESERVOIR 24-21254

(Grid units in area)

 
 

01234

 
  
 

Mesquite\ st\\
well Site Evergreen K I
c 26 9L3 Q

03

 

ARIZONA

\

\

l

‘3 Cibecue
\ Ridge area
0

 

 

I
\ ‘320
\\
100 MILES\\\ L
\\______ __

0 50 100150 KILOMETERS

../
__/

' Whitethorn site

 

 

 

INDEX MAP

Gila River Phreatophyte
Project and adjacent areas

  
  

Creosote
bush

       
  
 
  
 
 
  
 
    

Bare ground site

Grain sorghum field

     

Cultivated
grass

saltcedar site

Reach 1

3 4 5 MILES

5 6 7 8KILOMETERS

FIGURE 1.—Map of study area.

March 22, 1968; hourly missions from 9:00 a.m. to 3:00
pm. were flown at an altitude of 3,600 ft (1,100 m) on
May 31, 1968. The NASA missions were flown at an
altitude of 60,000 it (18,000 m) on September 30,1969,
and September 11, 1970.

From July 21 to September 28, 1971, six missions
were flown over the Cibecue Ridge area by the US.
Geological Survey. The missions were flown at al-
titudes of 1,500 and 4,000 ft (460 and 1,200 In). An
analysis of the data obtained during these flights is
included in this report.

Kodak Ektachrome Infrared Aero Film 84431 was
used in 1967—69; after 1969, Kodak Aerochrome In-
frared Film 2443 was used.

VEGETATION

In 1967 the Gila River flood plain was covered by a
dense growth of phreatophytes—mainly saltcedar
(Tamarix pentandra). The vegetation in reach 1 was
cleared before April 1967 as part of the comprehensive

1The use of name products in this report is for identification only and does not imply
endorsement by the US. Geological Survey.

 

plan as outlined by Culler and others (1970, p. 3). Veg-
etation clearing in reach 2a was started in January
1969 and completed in December 1969, and clearing in
reach 2b was started in November 1970 and completed
in March 1971.

The terraces adjacent to the flood plain support small
open forests of mesquite (Prosopis juliflora) and plant
communities of xerophytes, such as creosote bush
(Larrea tridentata), whitethorn (Acacia constricta), and
various cacti. Interspersed within these plant com—
munities are small ephemeral plants that grow in re-
sponse to winter and summer rains; the species that
respond to winter rains are different than the species
that respond to summer rains (Shreve, 1964, p. 127—
142). Cultivated crops, such as tall wheat grass and
grain sorghum, are grown adjacent to the project area.
A small community of the evergreen species of
saltcedar (Tamarix aphylla) is located in an adjacent
area to reach 1 of the project area. An extensive de-
scription of the vegetation of this area was prepared by
Turner (1973).

The main types of vegetation indigenous to the
Cibecue Ridge area are juniper (Juniperus osteos-

O4

perma), pinyon pine (Pinus edulis), ponderosa pine
(Pinus ponderosa), and native grasses. In 1967, one
watershed was cleared of native trees and seeded with
natural and exotic grasses.

COLOR-INFRARED PHOTOGRAPHY AS A
TOOL FOR VEGETATION ANALYSIS

An economic method of obtaining frequent descrip-
tions of ground-surface conditions and changes, par-
ticularly that of vegetation, is needed for watershed
management and hydrologic analyses, especially in
large areas. Traditionally, black and white and color
aerial photography have been used to inventory vege—
tation, but recent studies show that color-infrared pho-
tography generally is superior for species identifica-
tion, plant-vigor measurements, and vegetation map-
ping (Hunter and Bird, 1970, table 1).

COLOR-INFRARED FILM

The advantage of color—infrared film for vegetation
analysis is due to the high reflectivity of active vegeta-
tion in the 0.70—0.90 Mm (micrometer) near-infrared
wavelength range. The tripack color infrared film has
three dye layers, cyan, magenta, and yellow, that
mainly are sensitive to near-infrared (0.70—0.90 ,um),
red (0.60—0.70 pm), and green (0.50—0.60 ,um)
wavelength bands, respectively. In order to eliminate

 

GILA RIVER PHREATOPHYTE PROJECT

the sensitivities of the three dye layers in wavelengths
of less than 0.50 ,um, a yellow Wratten 12 or 15 filter is
used (fig. 2). After film processing, the combination of
the three dye layers exposed to their sensitivity ranges
yields false colors—near-infrared radiance generally
appears red, red may appear green, and green may
appear blue—in the multilayered film (fig. 3). The term
"false color photography” originates from this "incor-
rect to the eye” rendition.

Figure 4 shows the generalized characteristic curves
for color-infrared photography. A knowledge of the re-
lationship between optical density and log exposure is
essential for any quantitative photographic analysis.
Analytic optical density, which is shown as the scale on
the ordinate of figure 4, is defined as the optical density
within the individual dye layer of a multidye layer
film.

The two parameters, optical density (OD) and expo-
sure (E) are defined below:

OD = filog (T) (1)
E = (I) (2) (2)
where
T = transmittance,
I — illumination, and
2 = time.

Any characteristic curve such as figure 4 has three
basic parts: the shoulder, the linear portion, and the

 

 

   
    
 
  
  
  
 

 

 

  

 

 

100 . . _ 100
C I I I l l :

o '- _

I f' _

m _

u: )— _.

E Yellow‘forming '—

g — /1\\dve layer \ — 5:

Lu ’- \ / \ U

" \ E

“2‘ 1o :— ‘ / \ —: 10 a.

F : | / 3 E

E " l LMagenta-forming _ uI

U ' l / dye layer — U

u — l — Z

a: _ <

< — l2

I) _

o _ ‘ E

m ‘2

Z a:

>- 1 -— —__- 1 ._
— dye layer

'2 )— : a:

2 p _ E

t _ - g

3 ~ — t

Lu —

m .—

2 Wratten 12

:1 F filter —

LL

0.1 I 0.1
0.40 0.50 0.60 0.70 0.80 0.90

WAVELENGTH, IN MICROMETERS

FIGURE 2.—Spectral sensitivity versus wavelength for the three dye layers of color-infrared film.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

05

RESULTING VISUAL COLORS ON COLOR-INFRARED
FILM, IN MICROMETERS

 

 

 

 

I Blue I Green ' Red I
0.40 0.50 0.60 0.70
PRINCIPLE SENSITIVITY OF DYES BLEACHED
FILM LAYER BY EXPOSURE
Red
M enta
0.6—0.7 ‘ 39
Green
Y II
0.5—0.6 ° °""

 

 

Near infrared
0.7-0.9

 

 

 

 

 

‘ Cyan
l

SOURCE OF RADIANCE, IN MICROMETERS

 

Green I Red

0.60

r Blue I
0.40 0.50

I Near infrared —I
0.70 0.90

 

VisuaI sensitivity range

 

 

 

 

 

Sensitivity range of color-infrared
film with yellow filter

 

 

 

Spectral range of
low pIant pigment
reflectance

 

Spectral range of
high plant mesophyll
reflectance

 

 

 

FIGURE 3.——Color formation on color-infrared film and its relation to plant reflectance.

toe. The shoulder is the part where increasing exposure
produces an increasing rate of density change, on the
straight—line part the rate of change of density with log
exposure is constant, and on the toe the rate is decreas-
ing. It must be noted that figure 4 is for a positive film;
the characteristic curves would be reversed for a nega—
tive film.
LEAF AND CANOPY REFLECTANCE

When radiation comes in contact with a leaf or a
plant canopy, the energy either is reflected, absorbed,
or transmitted (fig. 5). In the near-infrared spectral
(0.70—0.90 ,um)-sensing range of the cyan dye layer,
individual leaf reflectance is relatively high—as great
as 50 percent; this relatively high reflectance of near-
infrared radiation occurs because of the difference in

 

the refractive index when the radiance that enters the
inner part of the leaf is refracted and reflected at the
interface between the water film covering the
mesophyll cell wall and the adjacent air cavity (Knip-
ling, 1969, p. 19). In the red-color (0.60—0.70 mm) sens-
ing range of the magenta dye layer, individual leaf
reflectance is very low—about 10 percent—because the
amount of absorption by leaf pigments (generally
chlorophylls) is high. Leaf reflectance is these two
bands is influenced by two distinct plant systems, one
involving pigment chemistry, the other mesophyll
anatomy. Both of these systems are indicators of the
volume of foliage and can be remotely sensed by color-
infrared photography.

The spectral reflectance of a typical vegetation

06

GILA RIVER PHREATOPHYTE PROJECT

 

3.8 \ I

Shoulder

0 Linear
m .4 .
\: a, portion
\ 02 a of
Z \ 9
Q 2' "fig curve
a) m a
__ a ' -< __
2.0 ‘w .2 g
c: a
“‘\
6

1 .0 —
Characteristic
curves

ANALYTIC OPTlCAL DENSITY, IN NEGATIVE LOG OF PERCENT TRANSMITTANCE

 

 

 

0
—2.0 —1.0

0.0 1.0

LOG EXPOSURE, IN METER—CANDLE—SECONDS

FIGURE 4.—Ana1ytic optical density versus log exposure for three dye layers of

color—infrared film.

canopy is considerably different from that of an indi—
vidual leaf, because of angular relationships, shadows,
and background surfaces. The visible and near-
infrared reflectances of the total canopy are of the order
of 5 and 35 percent of incident radiation, respectively.
The disproportionate reduction of spectral ranges for
an individual leaf versus a canopy, which enhances the
near—infrared radiance range, is due to greater reflec-
tion of near—infrared illuminance by multiple leaf
layers within the canopy (Knipling, 1970, p. 157).

The difference between the spectral reflectances of
these two radiance ranges is compensated somewhat
by the decreased incident solar energy in the near in-
frared versus the red. The energy received at the earth
surface in the near-infrared irradiance range is ap-
proximately 0.1 langley per minute, which is about 75
percent of that received in the red region (Reifsnyder
and Lull, 1965, fig. 12).

 

VISUAL ANALYSIS

Visual color-infrared photographic interpretation is
useful for the identification of the shape and color of
objects and is of value for studies of a reconnaissance
nature. The interpretation of subtle temporal changes,
however, requires a rigorous classification of color. In
such a classification, the amount of spectral irradiance
absorbed by individual dye layers of the film may be
measured and then related to the total irradiance,
sensed by the film, which gives a set of parameters of
relative irradiance that may be functionally related to
plant activity. Figure 6 shows two color-infrared aerial
photographs taken over the Gila River flood plain at
times of general vegetation dormancy (fig. 6A) and of
Vigorous plant activity (fig. GB). Photograph A is Visu—
ally interpreted as showing leafless saltcedar and mes-
quite and some spring grass on the south-facing ter-

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

O7

 

100

| l
90
80
70 —

60-—

NOTE:

plus absorption

40—

30—-

LEAF REFLECTANCE (MESOF‘HYLL RELATED) IN PERCENT
LEAF ABSORPTION (LEAF PIGMENTS RELATED), IN PERCENT
U1
0

 

        

The transmitted radiance is equal to
100 percent minus the sum of reflectance

Leaf absorption

 

 

20 —
% Leaf reflectance /
/ \ /
1o — / /
/ \\ /
_./
//
/
o | I |
0.40 0.50 0.60 0.70 0.80 0.90

WAVELENGTH, IN MICROMETERS

FIGURE 5.—Average leaf reflectance and absorption as a function of wavelength.

races. PhotographB is visually interpreted as showing
increased foliation and vigor of the saltcedar and mes-
quite, particularly by the saltcedar adjacent to the
river, and dormant spring grass; this is a major vegeta-
tion change that occurred in less than a month—March
22, 1968, to April 19, 1968. Healthy growing vegeta-
tion is shown as a red tone on figure 6B, which is prin-
cipally due to high values of near-infrared irradiance.
Differences in soil types and the amount of soil mois-
ture probably account for the braided patterns that are
shown in figure 6.

DATA COLLECTION

The evaluation of color-infrared photography as a
means of evaluating hydrologic variables required the
collection and tabulation of photographic data, which,
in turn, was related to hydrologic data.

 

PHOTOGRAPHIC DATA COLLECTION

Optical density data were obtained from the positive
transparencies using a Macbeth Model No. TD 402
transmittance densitometer with a 3-millimeter apera-
ture using Wratten filters numbers 106, 92, 93, and 94.
These filters transmit, respectively, the entire light
(0.40—0.70 ,um) spectrum (corrected to the visual re-
sponse), the red (0.60—0.70 um), green (0.50—0.60 ,um),
and blue (0.40—0.50 ,u.m) light (fig. 7). The resulting
data from each filter is expressed in terms of the optical
density of the multilayered film in each respective
wavelength range.

The instrumentation and calibration limitations and
the nature of the study prohibited photogrammetric
data from being analyzed as a precise spectral radio-
metric pheonomenon. The need for quantitative data
for computer analysis did, however, require an em-

08

GILA RIVER PHREATOPHYTE PROJECT

Saltcedar

 

FIGURE 6.—Vegetation in the Gila River flood plain. A, General dormancy;B, Vigorous plant growth.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

 
   

09

 

(“—1 3'0 _ Filter
92
(green) (red)
UJ
0
Z
<
.—
'2
E
m
z
<
m
I—
I—
,_ 1.0—— 2 2.0 —-
z LIJ
LIJ o
o u:
n: u:
Lu 0.
D- u.
2 0
.:~ 8
(i) J
Lu
<
t 2
I: l-
2 3
0’ Lu
2 z
E z
’- 1o.o— ;. 1.0 ——
'2
m
2
Lu
0
.1
<
9
'-
D.
0
100.0 -— 0.0

 

 

 

0.40 0,50

0.60 0.70

Data from Kodak (1970).

WAVELENGTH, IN MICROMETERS

FIGURE 7.—Transmittance versus wavelength for the densitometer filters used in film analysis.

pirical approach, which could be used to monitor the
ground-scene changes and would allow the necessary
corrections. It was decided to view each dye layer of the
film as a radiometer, which was sensitive to a particu-
lar wavelength range, and then to relate the irradiance
seeeennsed in each dye layer to the sum of the ir-
radiance sensed by the three layers (fig. 3). This proce-
dure allowed the film to be utilized as a relative
radiometer.

Using such an approach required converting the in-
tegral optical density, R, G, and B of the film to the
analytic optical density of each dye layer of the film.
This technique is explained in the section “Analytic
Optical Density”; these analytic optical densities C, M,
and Y, respectively, were then converted to analytic
trasmittance (Cm, Mat, Yat) for each dye layer.

T=10 40D) (3)
Where

 

= integral optical density—red color,

integral optical density—green color,

= integral optical density—blue color,

analytic optical density—cyan dye layer,

= analytic optical density—magenta dye layer,
and

analytic optical density—yellow dye layer.

a gamma
II H

II

The analytic transmittances for each dye layer were
then divided by the sum of the dye layer transmit-
tances, thus achieving a set of parameters of relative
irradiance which could be used to monitor the vegeta-
tive and hydrologic variables. The three equations for
converting to relative irradiance are shown below:

R = relative near-infrared irradiance, in percent:
[Cat/ (Ca2+Mat+Yat)]x 100, (4)

G = relative red irradiance, in percent:
Mda/(Czr+AL.+12»]x 100, (5)

010

B = relative green irradiance, in percent:

[Yat/ (Ca,+Ma,+Yat)]>< 100. (6)
Thus, by definition .
R +G +3 = 100 percent. (7)

Two characteristics of these trichromatic parameters
(Wright, 1969, p. 83) are of particular importance to
the researcher. R and G are generally inversely pro-
portional and linearly related, because of the three
following effects.

1. The spurious correlation due to the “closed sys-
tem” (Chayes, 1971) interrelationship between the var-
iables as shown in equation 7. For example, an in-
crease in R dictates an equal decrease in the sum 0+3 .
This effect is discussed in the section “Statistical
Analysis of Variables.”

2. The high reflectance in the near infrared and high
absorption in the red by active vegetation as shown in
figure 5.

3. The radiance from vegetation affecting the yellow
dye layer (0.50—0.60 ,um) is not significantly related to
plant vigor.

Ideally, photographic step tablets and spectral sen-
sitivity curves should be generated for each roll of film
to calibrate precisely the energy received by the film
and to compensate for processing and aging effects.
This was not done because of the limitations of the
study, which have already been discussed. The actual
area sampled by the 3-millimeter aperature on the
8,500-foot (2,600—m) photography is 0.50 acre (0.20 ha)
per plot. This was a 14 percent area sample and was
considered representative owing to the large sampling
populations.

Densitometric data which were not on the linear
parts of the characteristic curves were rejected. The
difficulty of handling data from the toe or shoulder of a
characteristic curve is obvious from figure 4.

HYDROLOGIC DATA COLLECTION

Hydrologic data sampling on the Gila River Phre-
atophyte Project was done using the plot system
already described. The hydrologic variables were mon-
itored by a network of ground-water observation wells,
soil-moisture access tubes, rain gages, and river and
tributary gaging stations. The sampling techniques
and methods of evaluation have been discussed by
Hanson, Kipple, and Culler (1972, p. 315).

VEGETATION DATA COLLECTION

The plant type and spatial distribution of vegetation
on the project area was described by a combination of
photographic reconnaissance and field checking.

Black and white aerial photographs with a scale of
1:7,200 were viewed through a dissecting binocular

 

GILA RIVER PHREATOPHYTE PROJECT

microscope, which had been fitted with an appropriate
reticle, and were used to measure plant-crown cover-
age (Culler and others, 1972; Turner, 1971). Tentative
boundaries defining areas of homogeneous vegetation
were first drawn on the photographs. A field check was
then made to confirm these boundaries and determine
the dominant species within each area. Finally, the
photographs were examined under magnification. The
squares on the reticle, when viewed through the micro-
scope, were considered to be plots projected onto the
photographs. Plant coverage within each plot was es-
timated according to a size-class system. Averaged
data, taken from an examination of each plot, provided
numerical values of percentage of crown coverage.
Comparison of the photogrammetric method with one
of the standard field measures of crown coverage was
made (Turner, 1973), and agreement between the
methods was close.

IDENTIFICATION AND MEASUREMENT OF
VEGETATION PARAMETERS

Conventional hydrologic instrumentation is usually
limited to point samples, but photographic remote
sensing offers a method of obtaining a record of both
the spatial and temporal variability of many vegeta-
tive parameters.

Definitive radiometric identification of gross earth-
surface characteristics, such as bare ground, water,
and vegetation type by photographic remote sensing, is
prerequisite to the quantitative monitoring of many
hydrologic variables. Figure 8 illustrates the relative
irradiance characteristics (R, G, and B) of eight
ground—scene conditions viewed with color-infrared
photography. Automatic computer identification of
earth-surface characteristics based on the spectral ir—
radiance appears possible and should be a productive
research field, particularly with the advent of satellite
imagery.

EVAPOTRANSPIRATION

The feasibility of estimating evapotranspiration by
photographic remote sensing has been studied by the
personnel of the Gila River Phreatophyte Project for
several years. A primary objective of this project is the
study and measurement of evapotranspiration.

Evapotranspiration is water withdrawn from a land
area by evaporation from water surfaces, from moist
soil, and by plant transpiration (Langbein and Iseri,
1960, p. 9). The vaporization process has three basic
requirements: (1) heat for the phase change from liquid
to vapor; (2) unsaturated air to remove the vapor; and
(3) liquid water available at an air-water interface.
Any of these factors can control the rate of vaporiza-
tion, and the quantity of water vaporized can be meas-

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

1001

(I)
O
l

01
O
l

011

EXPLANATION

EARTH —SUR FACE CHARACTE R ISTIC
O—-O
Water in reservoir
Water more than 10 feet (3 meters) deep;
small amount of suspended sediment
o— ——————————— 0
Water in river
Water less than 2 feet (0.6 meters) deep;
large amount of suspended sediment

o— __________ _o
Deciduous forest in summer
(Saltcedar)
o—-————o
Deciduous forest in autumn
(Saltcedar)

o— ——————— <3
Conifer forest in summer
(Pine)

 

N
O
l

O

 

o— ---------- —o
Conifer forest in autumn
(Pine)

....0
Bare ground
(Desert)

0

o— ———————— -o
Asphalt paved road

 

Relative red
irradiance (0.60—0.70):
sensed by
magenta dye layer

Relative green
irradiance (0.50—0.60);
sensed by
yellow dye layer

IRRADIANCE SENSED BY INDIVIDUAL FILM DYE LAYERS, IN PERCENT
b
O
I

 

Relative near» infrared
irradiance (0.70—0.90);

sensed by
cyan dye layer

WAVELENGTH RANGE, IN MICROMETERS, AS A PERCENT OF THE
SUM OF IRRADIANCES SENSED BY THE FILM (0.50—0.90 MICROMETERS)

FIGURE 8.——General percentages of irradiance for different eart h-surface characteristics Photographed using color-infrared film
2443 and filters numbered W12, CC20B, and CC30M.

ured by a rigorous budget of any of the three factors.
The definition of the spatial variability of these factors
is one of the major problems in estimating evapotran-
spiration. Heat and unsaturated air can be measured
by thermal and other meteorological instruments, al-
though conventional instrumentation limits these ob-
servations to point samples. The availability of liquid
water is dependent on the type of surface and, for most
types of land area, the condition of the surface. As an
example, the area of air-water interface for a lake can
be defined simply as the surface of the lake. For bare
soils or for vegetated areas, however, the air-water
interface may be in the pores of unsaturated soils or
within the mesophyll tissue of leaves. Thus, for a re-
mote sensing technique to be practical, these gross
surface types must be observed and determined. The
feasibility of discriminating earth-surface conditions
by the relative radiometer technique is demonstrated
in figure 8 and has been discussed in the section
"Identification and Measurement of Vegetation Pa-
rameters.”

Remote sensing offers a method of obtaining a record
of spatial variability of these surface conditions within
the detection capabilities of the sensing equipment.

 

This method of observation can be applied to the
estimation of evapotranspiration by detecting and
monitoring surface types or conditions which are func-
tions of, or are functionally related to, the factors
controlling evapotranspiration. Spatial mapping of
spectral signatures will be discussed in the section
"Spatial Computer Analysis.”

Temporal variability also presents a problem in es-
timating evapotranspiration. The seasonal change in
deciduous vegetation, ranging from bare limbs in
winter to complete foliation in summer, is an example.
Disease and moisture stress can also affect transpira-
tion. Shallow soil moisture resulting from rains affects
both evaporation from bare soils and the distribution of
annual vegetation. In addition to these natural tem-
poral changes, the activities of man can dramatically
alter surface conditions in a relatively short time.
Thus, spatial variability is not uniform in time. Repeti-
tive remote sensing is an efficient means of monitoring
and analyzing this variability and is discussed at
length in the section "Time-Dependent Signature.”

The usual method of estimating evapotranspiration
is to use one of a number of empirical equations, which
express the relations between measured evapotranspi-

012

ration and climatic conditions as described by Veih-
meyer (1964, table 11—2).

Remote sensing in the form of color-infrared photog-
raphy has been applied to the widely used equation
developed by Blaney and Criddle (1962). Studies by
Cruff and Thompson (1967, p. 22) indicate that this
equation is the most reliable for estimating potential
evapotranspiration in arid and subhumid climates
yielding variability ranging from —44 to +22 percent
of adjusted evaporation data from Weather Bureau
Class A pans. The equation is

u=kfi ( 8)
where

u =calculated monthly evapotranspiration, in
inches, derived from the Blaney-Griddle
equation;

k = empirical consumptive-use coefficient, which
can be defined for a vegetative area by re-
mote sensing (the coefficient is dependent on
the species, composition, and quantity of
vegetation as well as the background signa-
ture which may be soil or water), and

pt
d fi ed ———,
e n as E 100
summed for the number of months of anal-
ysis, in which
p = monthly percentage of annual daytime hours

f = consumptive-use factor,

and
t = mean monthly temperature, in degrees
Fahrenheit.
In metric units,
. +
u = kp (————45 71t00813 ) =monthly consumptive

use, in millimeters and
t = mean monthly temperature, in degrees centi-
grade.

The k value is assumed to represent a gross measure of
the air—water interface. It therefore represents the spa-
tially variable factor defined by the surface conditions,
including such botanical parameters as vegetation
type and aerial extent, plant density, and physiologic
conditions. For the purposes of this study, k was de-
fined in two ways. The vegetation description of the
project area was used to define a k which was used in
the water-budget computations of evapotranspiration,
and a second k was defined by remote sensing which
was dependent on relative near-infrared irradiance.
Values of p are tabulated in Blaney and Criddle (1962),
andt is generally available in published US. National
Weather Service (issued annually) summaries. It is as-

 

GILA RIVER PHREATOPHYTE PROJECT

sumed that the factor f describes the thermal and vapor
conveyance requirements of evapotranspiration.

The assdmption that the relative near-infrared ir—
radiance can be used as a measure of the Blaney-
Criddle k for determining evapotranspiration was
tested on two large reaches of the Gila River
Phreatophyte Project area using the photographic data
for 1968. The two test sites were reach 1, which is a
1,732-acre (701-ha) area cleared of phreatophytes but
partially covered by grasses, and reach 2, a 2,268—acre
(918-ha) area covered by phreatophytes. The evapo-
transpiration was computed as a residual in the water
budget. Median monthly values of evapotranspiration
were calculated for reach 1 after clearing and for reach
2 prior to clearing (Hanson and others, 1972, figure 5).
These median monthly values were then compared to
calculated evapotranspiration based on relative near-
infrared irradiance of 13 separate photographic flights
of both reach 1 and 2 of the project area during 19682.
The mean deviation between the evapotranspiration
computed from the water budget and the calculation
derived from the photography was 32 percent. Figure 9
shows the evapotranspiration values derived by both
methods for both reach 1 and 2; also shown is the per-
cent deviation of the spectral calculation from the
evapotranspiration computed from the water budget.

The equation used to calculate evapotranspiration
by relative near-infrared irradiance for both reaches 1
and 2 is in a form comparable to the Blaney-Griddle
equation (8), u=kf. The general remote sensing form is
shown below:

E‘T=Lf(R)]W)L (9)

where
ET = evapotranspiration, calculated by relative
near-infrared irradiance,

f(R) = O.37+8.25 [2(R/100)2/n]2-45 (this term is con-
sidered to be equivalent to k for this test
site),

n = number of samples,
R = relative near-infrared irradiance, and
flf) = f( % )1
The k term was developed by optimization procedures
using the water-budget data and the remote sensing
data. For convenience, equation 9 has been expanded
and is shown below:

Wm 13 of the 15 flights flown during 1968 were used because of the large
departure from the expected signature exhibited on the May 3, 1968, and August 28, 1968,
photography. The May 3, 1968. photography was affected by a frost occurring on April 20,
which defoliated all vegetation with 8 R (2.4 m) of the ground. The August 28, 1968,
photography demonstrates the 0.89 inches (23 mm) of rain which occurred that day on
reach 2. This water on the ground appreciably decreased the near-infrared albedo, there~
fore decreasing the R and proportionally increasing the G and 3. Because of these two
events, the May 3 and August 28, 1968, photographic missions appear as anomalies on
figures 13, 16, 17, 21, 23, 27, 28, and 30 in this report, and the data from these missions
were not used in the computations presented in figures 9, 10, 21, and 22.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

EXPLANATION

 

 

 

 

 

 

 

 

 

 

 

 

013

 

 

 

7 — 0 Reach 2, 2,268 acres (918 hectares),
c: ‘_ before clearing phreatophytes
. .— . .
6 _ _ fl I Mean absolute deVIatIon, 25 percent _ 150
Calculated evapotranspiration: w ‘9 V
Calculated using relative near-infrared {-2 .— ‘— _ u:
5 __ irradiance.Calculatad values are from + ' ‘ >.
U) 13 photographic missions flown in1968. V V :9 <1:
>- Number, +20. is percent of deviation . D
< from measured evapotranspiration Q 8
g 4 — D — 100 m
n Median monthly evapotranspiration. E
a: Computed using water- budget methodm m m
E 3 _ v r:
I 01 u
U) V I-
u.I o ' LlJ
I
<2) 2 — ‘I' g — 50 E
.I
_ i 3 _I
E I E
i 1 - a
9 .
F S
g o I I I l I l l I I l I I I l I I T l I I I I l I I I l I I I I I I I I I I I | I I I l I I I I I I I I I I I I I I I o I:
E <
(n 4 __ v ._ Reach 1, 1,732 acres (710 hectares), _ 100 E
<2: i: m m N after clearing phreatophytes _ n.
N + o m U)
E + ‘2 + Mean absolute deviation, 39 percent <2):
0 3— g "‘ II
0' + K, I-
< + O
> m :
u.I
2 .— (f a) —_ 50 >
o
l N LIJ
+
1 — E] _.
0 I I I I I I I | I l I I | I l T I I I I T I I I I I l I I I l I | I I I l I I I I I I I I' I l I I I I I I I I I I I I I 0
Jan. Feb. I Mar. Apr May June July Aug. I Sept. Oct. Nov. I Dec.

 

 

 

FIGURE 9.—Remote sensing and water-budget values of evapotranspiration versus time, Gila River Phreatophyte Project area.

ET = {O.37+8.25[2(R/100)2/n]2'45}[(f)(103/12)].(10)

Several items will be noted by scrutiny of figure 9.

1. The evapotranspiration calculated spectrally is
consistently high for reach 1 and consistently low for
reach 2. To be an effective tool over large areas of land-
scape, the remote sensing estimate of evapotranspira-
tion should be independent of any knowledge of ground
condition, and for this reason the same spectral equa-
tion (equation 10) was used for both the phreatophyte-
covered reach 2 and reach 1, which was cleared.

2. The nature of the different covers in the two
reaches shown in figure 9 must also be considered to
gain an understanding of the departure noted in item
1. Reach 1 was cleared of phreatophytes prior to the
1968 photographic flights, and the bare areas produced
by the clearing have been reoccupied to varying de-
grees by perennial and ephemeral herbaceous plants
(Culler and others, 1972). The primary source of water
for this replacement vegetation is from shallow soil
moisture. The water availability for these plants is
considerably less than for the phreatophytes on reach
2, which predominately obtain their water from the
water table. The relative near-infrared irradiance of

 

these replacement grasses is probably disproportion-
ately higher than phreatophytes for a given amount of
water use because they are short, yet dense, and reflect
greatly in the near infrared. Therefore, evapotranspi-
ration calculated by relative near-infrared irradiance
on reach 1 is assumed too high. This would also account
for the underestimation of evapotranspiration on reach
2 because of the same equation being used simultane-
ously for both sets of data.

3. Figure 9 indicates that the relation between k and
relative near-infrared irradiance is closer for dense
vegetation than for sparse cover. The evapotranspira-
tion calculated by irradiance has a closer correlation on
the phreatophyte-covered reach 2 (25 percent versus 39
percent). This assumption is also shown in later
examples.

4. An error source, which should be considered in
analyzing figure 9, is the inherent deviation of the
evapotranspiration derived from the water budget.
These median evapotranspiration values have a devia-
tion of as much as 40 percent (Hanson and others,
1972, p. 326). This fact may also explain a portion of
the deviation between the water-budget evapotranspi-
ration and the remote sensing approach.

014 GILA RIVER PHREATOPHYTE PROJECT

The data represented in figure 9 may also be viewed spiration for both reach 1 (fig. 10B) and reach 2 (fig.
as two discrete tests for evaluating evapotranspiration; 10A). The regression equations and coefficients of cor-
this is done in figure 10. Figure 10 shows the functional relation are given below:
relationships for remote sensing versus water-budget
calculations (derived from equation 10) of evapotran— ET1 = 0.64( ET)+20.5; (11)

CALCULATED EVAPOTRANSPIRATION, IN MILLIMETERS PER 30 DAYS

 

  

 

   
  
 

 

 

 

0 50 100 150
l l l
7 I
U)
2
3 o
< 5 — —150 o
D m
0 [I
"’ u:
[I n.
LU U)
m D:
— LL!
3 5 l;
I
o E
Z 3
' 1'
E 2
Z. 4 i— _—100 z
9 2-
'3 9
D_: 2 EXPLANATION
n.
3 3 — — E ET
A . .
<1 ET = 0.88 (ET) + 108.2 % Computed median monthly evapotransplratlon:
E V = 0.86 E Computed using water-budget methods
0 Sy-x = 71.00 a: A
a p. ET
< 2 _ _ 50 0 Calculated evapotranspiration:
E — % Calculated using relative near-infrared irra—
0 > diance. Calculated values are from 13 photo-
‘-'-' L” graphic missions flown in 1968
I D
3 Lu S
(I) D: .
< 1 — _ 3 y x .
L“ (1) Standard error of estImates
E <
u.I
E r
Coefficient of correlation
o I | I l l o
O 1 2 3 4 5 6

CALCULATED EVAPOTRANSPIRATION, IN INCHES PER 30 DAYS
A. Reach 2, 2268 acres (917 hectares), before clearing phreatophytes

CALCULATED EVAPOTRANSPIRATION, IN MILLIMETERS PER 30 DAYS
0 50 100 150
l l l

 

 

 

 

i 2‘
o 3 I l I I I o
F :2
<0: S.
9 >1“ :3
113: E0
30 W"
((02— ——50<ZKE
E5 on
Os 5::
:0) A n-I.u
“" ET=0.64(ET)+20.5 <I—
>1 >LU
wo1_ r:0.88 _ m2
03 Sy-x 224.73 0:
Lu LIJ_,
[12 D:—
D" 32
m a)
E I I I I I (E
o o w
2 o 1 2 3 4 5 6 E

CALCULATED EVAPOTRANSPIRATION, IN INCHES PER 30 DAYS
B. Reach 1,1732 acres (701 hectares), after clearing phreatophytes

FIGURE 10.—Regression relations between remote sensing and water-budget values of evapotranspiration, Gila River Phreatophyte
Project.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

the coefficient of correlation for this equation is 0.88;
and

ET2 = 0.88( ET)+108.2; (12)

the coefficient of correlation for the equation is 0.86;
where
ET1 = evapotranspiration, reach 1, calculated as a
residual by the water-budget method and
ET? = evapotranspiration, reach 2, calculated as a
residual by the water-budget method.
Figure 10 is a refinement of the data presented in
figure 9, and although figure 10 should be viewed with
caution (for the reasons presented in the discussion of
fig. 9), the correlation within each reach is good, and
more research of this type would appear justified.
Hanson, Kipple, and Culler (1972) described an
analysis in which the measured evapotranspiration for
each month of record on each reach of the Gila River
Phreatophyte Project is related to foliar cover for the
respective reach by solving the following transforma-
tion of equation 8:

“my
frfij/
where subscripts f“, r71, and 5/ represent the reach,
month, and year, respectively. The months of available
evapotranspiration data were used to compute the
monthly k coefficients for various values of foliar cover
seasonally as shown on figure 11.

The foregoing discussion has been directed toward
an analysis of evapotranspiration based on the varia-
bility of near-infrared irradiance. The image of wet

kmy =

(13)

 

1.8

015

bare soil on color—infrared film is dark green; thus, the
dominant irradiance is in the green and red (fig. 3). The
interpretation of this condition will require multiband
analysis, which is not possible with the available pho-
tography. Most logically the comparison should be be-
tween near-infrared irradiance and that component of
evapotranspiration directly related to plant volume—
that is, transpiration. Because evapotranspiration in-
cludes both transpiration from plants and evaporation
from the soil, transpiration can be calculated by sub-
tracting evaporation from total evapotranspiration. A
comparison has been made, although direct soil-
evaporation data are not presently available for the
Gila River site. Data for this area can be estimated by
using published values from evapotranspirometers op-
erated near Yuma, Ariz. (McDonald and Hughes, 1968,
table 6).

In order to define the relationship between transpi—
ration (T) and k, it was necessary to subtract the evap-
oration (E) calculated for the Yuma site from the
evapotranspiration computed from the water budget on
the Gila River:

T=ET—E. (14)

A consumptive-use coefficient (k’) independent of
evaporation may then be defined by modification of
equation 8:

T

7‘
Figure 12 shows the relationship between k’ and
relative near-infrared irradiance for the Gila River

k’=

(15)

 

ET: kf

.1
'm

_.
1:.

2
i0

_.
'o

9
00

O

07

.0
A

0 percent

    

#200

——150

—100

IN INCHES PER 30 DAYS

—50

CONSUMPTIVE-USE FACTOR,f,

MILLIMETERS PER 30 DAYS

 

CONSUMPTIVE-USE COEFFICIENT, k

.0
N

O

 

CONSUMPTIVE-USE FACTOR, f, IN

 

 

 

From Hanson, Kipple, and Culler
(1972, fig. 7).

Jan. [Feb.iMar. IApr. I May 'June ’ July ’Aug. [Sept] Oct. [New] Dec.

 

FIGURE 11.—Monthly consumptive-use coefficients for areas of indicated percent of foliar cover
of phreatophytes, and average monthly consumptive-use factor.

016

GILA RIVER PHREATOPHYTE PROJECT

EXPLANATION

Relationship between k’ and data for un~

 

2.0
l l

    

1.5 __ k’2a = 0.046 1% — 1.35

r = .80
Sy-x =0.33
1.0 —- 2
r = .77
Sy-x = 0.32

MONTHLY CONSUMPTIVE —USE COEFFICIENT

 

o I / I |

= 0.04113 -1.14_

cleared subreach (2b)

Relation between k’ and pooled data
for cleared (2a) and uncleared (2b)
subreaches

Data from energyabudget station, represent-

—-— ing especially dense vegetation, is in,
cluded in both regression analyses

0
Subreach 2a (cleared of phreatophytes)

A
Subreach 2 b (uncleared of phreatophytes)

0
Energy budget station

A

R
Relative near - infrared irradiance

k’
-— Monthly consumptive—use coefficient
V
Coefficient of correlation
Sy.X
Standard error of estimates

 

 

0 20 40 60 80

2

100 Subscript, reach 2

RELATIVE NEAR-INFRARED IRRADIANCEJN PERCENT 2a

Subscript, subreach 23

FIGURE 12.—Monthly consumptive-use coefficient versus relative near-infrared irradiance.

Phreatophyte Project. The equations shown on figure
12 are shown below:

kg, = 0.046(R)—1.35 (16)

(the coefficient of correlation is 0.80, and the standard
error is 0.33) and

k’2 = 0.041(R)—1.14 (17)

(the coefficient of correlation is 0.77, and the standard
error is 0.32)
where

k’2a = monthly empirical consumptive-use coeffi-
cient (independent of soil evaporation) for
subreach 2a;

k’2 2 monthly empirical consumptive-use coeffi-
cient (independent of soil evaporation) of
all of reach 2; and

R = relative near-infrared irradiance.

The data (k’ versus R) for subreaches 2b (cleared of
phreatophytes with a partial cover of annual vegeta-
tion) and 2a (phreatophytes undisturbed) and one data
point computed from an energy-budget station located
on the project are plotted on figure 12. Equation 16 is
based on data from reach 2a and the energy-budget
station. Equation 17 is based on data from subreaches
2a and 2b and the energy-budget data; this was de-
veloped to test the effect on the regression of the
sparsely vegetated subreach 2a. The standard error is
less for equation 17 than for equation 16 because of the
greater number of data points. However, the coefficient
of correlation is higher for equation 16, indicating that

 

the relation between k’ and irradiance is closer for
dense vegetation than for sparse cover.

It must be emphasized that the statistical signifi-
cance of equations 16 and 17 are not great because of
the small number of data points. This illustration is
not meant to show the precise relationship between k’
and R, but rather to show the development of an ap-
proach to evaluating evapotranspiration.

A different approach to correlating spectral signa-
ture t0 evapotranspiration was made using a Blaney-
Criddle consumptive—use curve modified from Erie,
French, and Harris (1965, fig. 10). The consumptive-
use factors (f) were evaluated for the project area. The
tests were performed using a grain sorghum crop adja-
cent to the Gila River Phreatophyte Project during its
growing season. The regression equation describing
the relationship between consumptive use and R is

u = 0.0075 (R)—0.29 (18)

(the coefficient of correlation is 0.84, and the standard
error is 0.09).

A refinement using a modified R parameter was
made of the sorghum crop photography (figs. 13, 14).
The spectral signatures was modified by the equation

R—Rq
“W’

a q

(19)

where

f = modified relative near—infrared irradiance,
q = subscript, bridge reading for any photographic
mission,

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

017

 

        
  

  

 

 

 

 

 

0.5
— 12:
>.
< ’\ o
D
D 67/ \ “J (I
504— {5/ Y \ #10035: _10$
3 Q\/ lrradiance \\ E I— g
m Q‘V/l affected by \\ ZE LU
5 rain IT 0 m 8 E
II
E 0.3 — — 75 <31 E
m _I
Z 22 -I
. LIJ _ _
m > u; — 6 E
3 I: U E
0.2 —— g — 50 < z .
“J _J <12 m
> LU — u)
— O — 4 3
i— [C <
n. / uJ
/ DII
E um: 2
3 01 — — 25 -- I-
U) ' u. m
z Irradiance prior 5 — 2 2
O to plant emergence O D
U 2 g
0.0 o — o 8
JUNE I JULY AUGUST I SEPTEMBER J OCTOBER
1968 Consumptive-use data modified

from Erie, French, and Harris
(1965, fig. 10)

FIGURE 13,—Consumptive use of water and modified relative near-infrared irradiance versus time for grain sorghum during
1968 growing season.

 

 

 
 
   
    
    
   

 

 

 

°'5 I I | I I

EXPLANATION — ‘2:

> A
< 1 0
0 Modified relative near-infrared irradiance o:
E 0.4 _ u - — 10 3’
a. Consumptive use . g;
3 r S w
I Coefficient of correlation y-x— ’—
o — 8 “J
z 0.3 I— — E
2 Sy~x Sy~x=0.06 _l
_ Standard error of estimate 1'
u; _ 6 E
(D Z
3 0.2 — _ —~
“-1 w
> (I)
I: — 4 3
n. m
5 2
a 0.1 — — I-
z — 2 E
O 3
U m
5
0 v I I — Q

0 20 4O 60 80 100 120

MODIFIED RELATIVE NEAR—INFRARED IRRADIANCE IN PERCENT

FIGURE 14.—Regression relation between consumptive use of water and modfied
relative near-infrared irradiance for grain sorghum during the 1969 growing

season

a = subscript, data from a sampled plot of an ever-
green species of saltcedar (Tamarix aphylla)
located on the project area.

i was then related to the consumptive use; the regres-

sion equation is
u=0.0049 (f)—0.14 (20)

(the coefficient of correlation is 0.93, and the standard
error is 0.06).

 

l

The parameter 1 presents the amount of radiance of
the grain sorghum crop as a function of two presumed
constant radiating surfaces, and although this assump-
tion is not exactly true, particularly with regard to the
saltcedar, this empirical ratio does show promise. The
increase in the coefficient of correlation and the de-
crease of the standard error of estimate between equa-
tions 18 and 20 seem to verify this. Equation 19 repre-
sents a method of calibrating for the change of chroma

018

(and to a lesser extent the hue) as sensed by the film.
Ideally this should be done by using a series of large
panels with precise reflectance characteristics, which
could be photographed and used to calibrate the spec-
tral data.

DEPTH TO WATER-SOIL MOISTURE

Vegetation is dependent upon water for growth. This
water may be obtained from the water table, from soil
moisture, or from both of these sources. Photographic
remote sensing is a means of approximating a canopy
measurement, but this technique does not allow one to
determine the source of the available water.

One of the principal objectives of this study was to
determine if a photographic remote sensing technique
could be developed for measuring depth to ground wa-
ter. This was found impractical. Figure 15 shows the
depth to water versus relative near-infrared irradiance
for the 15-acre (6-ha) vegetation areas surrounding 14

 

GILA RIVER PHREATOPHYTE PROJECT

observation wells located in reach 2 of the Gila River
Phreatophyte Project. The regression equation for this
relationship is
W=—0.64 (R)+45.4 (21)
where
W=depth to water (table level, in feet).
The coefficient of correlation is 0.67. The coefficient of
correlation associated with equation 21 is not high, but
an inverse relationship between depth to water and
irradiance is apparent for depths of less than 30 it (9 m).
The relationship between soil—moisture depletion,
which results in plant stress, and the resulting change
in spectral signature is implied by figure 16, which
shows the R and 0 Signature curves for 1968 for the
dense saltcedar growth located in grid 24—2 (see fig. 1).
Below the R curve is a plot of a quantity of water in the
capillary zone (defined for this analysis as the quantity
of water, in inches, from the water table to 3 ft (0.9 m)

 

 
  
  

   

 

 

 

30.0 LU
w I I I I I I I T I 0
0 <
< LL
LL -—8 a:
03: 25.0— W=-0.64R +454 _ 3,
"’ Sy-x r = 0.67 D
0 S . : 2
z \ yx 5.42 <
< 4
4 s
g 20.0 —— —_ 5 o
0 _J
-‘ In
L” in
CD (I)
D:
E 15.0 — _ L”
m E
U.
— 4
2 5.
’~ 3
I ~— — .
M 10.0 a:
'— LLI
E E
o —2 g
I- 5.0 »— a o
I l-
}; I
g I:
LU
o I I I l l- I | J | 0 Q
o 10 20 30 40 . 50 60 70 so 90 100

RELATIVE NEAR—INFRARED IRRADIANCE, IN PERCENT

Data for June 27,1968

EXPLANATION

W

Depth to water, in feet

A

R

Relative near- infrared irradiance

r

Coefficient of correlation

SM

Standard error of estimate

Observation well
Relative near~infrared irradiance was measured
in a 15—acre area around each weII

FIGURE 15.-——Depth to water versus relative near-infrared irradiance, reach 2 of the Gila River Phreatophyte
Project.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

019

 

  
   

  

 

 

     
 
 
   
 
    
    
     
 
 

 

 

 

 

1°° I I | I I I I I I I g j
2 n.
90 — < fix — 13 3 ui _
“- 0.5 — U, z
< < 350 g
a E U) _ 12 d >. (I)
80 .. <1: q UJ 13.0 v II n: _.
a: j I - 2 < 1‘3
" 3% Water in/‘\_‘ /\. — 11 o jm
> _ 12.0 . JV . _ 2
70 _ < < capIIIary zone I: 300 m _ _
_ 0 Z \ Lu < J
‘- 0 0.4 2 _ _ 10 a. U _'
z I - uj11.0 u: 2 g
1.11 E u: z I: _
U 60 — w E O Evaporation __ 9 E I Z ‘
E g < N 10.0 m 250 f3
0. E
o 3 8 3 g
2 _ '_ _
__ 50 E 0.3 — EXPLANATION \ :1
‘0 E 1% 2
o ‘ Relative near-infrared irradiance \ \ _ 7 Z
Z 40 _ Z l \ \ - —
< 9 I 6 \\ z~
Q: :1 I Relative red irradiance \ _ 6 9
a: \ l-
30 — < —
2 0.2L I \ \\ L 5 co:
<
> ' \\ ‘2
20 _ LLI ll \ > —‘
\ NJ
I \
I \ _
10 — I Standard \
deviation \
\
0 Jan. I Feb. I Mar. I Apr. I May I June I July I Aug. I Sept. I Oct. I Nov. I Dec. I ‘-
1968

FIGURE 16.—Relative near-infrared irradiance, relative red irradiance, evaporation, and water in the capillary zone for a dense 40-acre
(16-ha) saltcedar site, 1968.

above the water table). The effect of soil-moisture de-
pletion is very evident on both the August 13, 1968,
and August 28, 1968, flights. Another interesting fea-
ture of figure 16 is that the May 3, 1968, flight was not
significantly affected by the April 20, 1968, frost. The
dense saltcedar at this site has a mean height of 13 ft
(4 m), whereas the frost only affected vegetation less
than 8 ft (2.4 111). Also shown in figure 16 is the pan
evaporation (Class A, Weather Bureau), in inches per
day, for part of the growing season during 1968. The
various effects (temperature, radiation, and so forth)
which contribute to pan evaporation are probably also
contributing factors for the anomalous August signa-
ture. The confusion as to the relative effects of different
hydrologic parameters on spectral signature em—
phasizes the need for more research, particularly re-
search of a quantitative nature.

Figure 17 shows the respective R and G signatures
for two 15—acre (6-ha) sites of different types of vegeta-
tion during 1968 and 1969. The lower diagram shows a
mesquite site which has an approximate 50 percent
foliar cover, the corresponding accumulated soil mois-
ture above the water table, and the water-table eleva-
tion. The upper diagram depicts some data for a dense
(nearly 100 percent foliar cover) saltcedar area.

 

The respective signature for the two sites show that
during the summer months the R is greater than the G
at the saltcedar site, while the R is generally less than
the G at the mesquite site. A correlation between soil
moisture or water table is not immediately apparent
from the R and G signatures.

VEGETAT ION

The two spectral parameters of leaf reflectance, rela-
tive near-infrared irradiance (R) and relative red ir-
radiance (G), can be used as a tool for the time-
independent discrimination of plant species and as a
means of monitoring the change in vegetation status
with time.

SIGNATURE DISCRIMINATION

The spectral signature is defined in this report as the
relationships between the relative irradiances (R, G,
and B) for a particular object. These relationships are
variable with time and changing conditions. The more
traditional definition is given in the “Glossary.” Dis-
crimination of vegetation by spectral signature can be
achieved using the two spectral bands, R and G, which
are functionally dependent on vegetation status.

020 GILA RIVER PHREATOPHYTE PROJECT

1

 

 

 

 

 

 

 
   

   
 
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n: 9 of
E 0 IE .0
$38 “’— - “33w
§m4< 11— — EMA:
OI-L'gu. 053m
I—LLon: 12— — F205:
24: I JD
I— m 13— “— Z
I- Lu 4|- m0?
3 “3 14 $—'”
0 I I I I I I I I I I I I I I I I I I I I I I I D
. 16 400 w
IJJ LLI
Egg) IBI— _ Egg
pkg 20— ——5°°:PE
<90 22— _ <95
302 So:
22; 24a _—soo§§_.
gi— 26- _ 8&2
8 28 —700 8;
I | I I I I I I I I I I I I I I I I I I I I
90
'2 80— ——
E,” 70— _
E 60— —
L
2 50— —‘
_ _ \ _
<6 40
0 3°- / I 1% \ —
z 20 _ ‘ R I t' e —' f d l \ —
< / II eaIv near In rare / \
W 10 _ I ~\‘ ‘ irradiance / \\ _
o ‘\‘-_______ K
JIFIMIAIMIJIJIAISIOINID'JIFIMjAIMIJIJIAISIOINID
1968 1969
A. Well 13L2 in saItcedar area
05 43 ”135.
g o Fmgm
(1:38 44— _ §$<o
§m4< 45— — F-‘E
0L“ LL 0mg“:
l-LLEI 46— ——14 "203
129.3 _ _ ,5sz
III—Em 47 n._g
Lu
3 48 I I I I I I I I I I I I I I I I I I n
. ‘UI
UJ
LU 50 LL! fl:
fig”, 52 —-13003§;{I
— uI — I-
I-I-I _ — <27];
<90 54 J—
302 56 ——————— —14ooDc2)3
332 w 34;
05 I I I I I I I I I I I I I I I I I I 062
m w-
I. so
E 70— —
g 60— —
E 50— _
Z 40— _
(O 30 — I \ f \ \V \ 1% /r'—”’— \\ _.
O __ [I \I ‘ ’ \
Z 20 , \/ \V"\\‘/Relative near-infrared ’,,/ \\ ‘—
< 10 _ ’ \\ irradiance ,/ \\ —
(m 0 ‘\\‘\\””~‘_J \~‘
J I F I MIA I MI J I J IA I 5 I0 I N I D| J I F IMI AIM] J I J I A I sIo I NI D
1968 1969

B. Well 9L3 in mesquite area

FIGURE 17.—Relative near-infrared irradiance, relative red irradiance, cumulative soil moisture, and depth to water 1968 and 1969.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

TIME-INDEPENDENT SIGNATURE

A series of time-independent linear relations for R
versus 0 was found for all of reach 2 of the Gila River
Phreatophyte Project and five vegetation types located
in the adjacent to reach 2 for all photographic missions
flown in 1968 and 1969. The regression equations, co-
efficients of correlation, and standard errors of esti-
mate for each vegetation type are shown in table 1. The
relationships between R and G are graphically illus-
trated in figure 18.

The same type of analysis was performed for the veg—
etation in and adjacent to the Cibecue Ridge area and
is presented in table 2 and figure 19. Several conclu-
sions can be postulated from the linear relations shown
on figures 18 and 19.

1. The sparse upland vegetation—creosote bush and
whitethorn—differs from the other vegetation shown
in figure 18. The slope of the regression equations and
the coefficients of correlation (table 1) distinguish this
type of ground cover. The nonuniformity of spectral
signature between the upland- vegetation and the other
vegetation types shown in figure 18 mainly is because
the creosote bush and whitethorn have a foliar cover of
10—25 percent and a low moisture availability. The
predominant signature that the film records is actually
a combination of bare ground, cacti, litter, and short-
lived annuals that form the understory.

2. The dense saltcedar, cultivated grass, and all
saltcedar and mesquite in reach 2 show a general trend
of diminishing R spectral signatures in that order. The
trend from dense saltcedar to mesquite is the expected
tendency; the more open mesquite community has a

 

O21

lower R signature than the saltcedar community. The
regression equation for the vegetation in reach 2,
which is a combination of mesquite and saltcedar, is
between the regression lines of the two tree types. The
cultivated grass has a higher R than any vegetation
type shown except that for dense saltcedar; the grass
has a foliar cover of about 50 percent and is subject to
little or no stress owing to irrigation. The dense
saltcedar community is adjacent to the river, where the
average depth to water is 10 feet (3.0 m) below the land
surface, and has the highestR signature because of the
high water availability.

3. The slope of the regression lines for the denser
vegetation types is roughly parallel, indicating that
the rate of R versus G change for vegetation with high
moisture availability is fairly constant. Low moisture
availability, as shown by the creosote bush and
whitethorn signature, tends to decrease the slope.

4. The vegetation types of the Cibecue Ridge area
(fig. 19) show the same trends as the vegetation in and
adjacent to the Gila River Phreatophyte Project, with
an increase of R signature from bare ground to
nonhomogeneous grass sites to the juniper-pinyon
stands and, finally, to the denser ponderosa pine.

5. The very low coefficient of correlation of the grass
sites in figure 19 is due to the nonhomogeneity of sam-
pling; the six grass sites used have a foliar cover of
between 5 and 20 percent. An added signature error is
encountered because of the large percentage of mulch,
which in many cases obscures the grass. This is also
true to a lesser extent of the bare-ground signature.

6. The ponderosa pine and juniper—pinyon regres-

TABLE 1.-—Time independent analysis using color~infrared film for vegetation in the Gila River Phreatophyte Project and adjoining areas
[Data used in analysis are from flights made in 1968—69, except as indicated; for a more complete statistical analysis see table 5}

 

Standard

 

 

 

 

 

 

 

Number Coefficient error
Area 0 Equation Regression of of
Vegetation (in acres) flights number equation correlation estimate
Reach 2 of the Gila River Phreatophyte Project area
All saltcedar (Tamarix pentandra) ......... (1) 25 22 R 24.91 (G) + 73 9 —0.92 7.9
All mesquite (Prosopisjuliflora) ........... (1) 25 23 R=— .91 (G) + 69 9 — .89 7.6
otal ,,,,,,,,,,,,,,,,,,,,,,,,,,,,, ‘2.600 25 24 R=— 89 (G) + 71 0 a .90 7.9
Adjacent area
Creosote bush (Larrea tridentata) 66 24 25 R : 10 81(0) + 63.9 —0.86 6.8
Whitethorn (Acacia constricta) ...... 92 24 26 13:, 73 (G) + 60.7 — .81 7.7
Grass (Agropyron elongatum) ,,,,,,,, 22 15 27 13:71 01 (G) + 82.3 — .91 9.4
Dense saltcedar (> 75 percent foliar cover) ,,,,,,,,,,,,,,,, 37 215 28 12:71 02 (G) + 96.2 — .99 2.4
1Owing to clearing of reach 2 during 1969, the sample area changes. The sample areas and their corresponding flights are given below:
2Only 1968 data used.
Dates Number Acreage of Acreage of
of o saltcedar mesquite
flights flights sampled sampled

Mar. 22, 1968—Mar. 6, 1969 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 16 1,380 960

Mar. 6—June 17, 1969 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4 1,020 380

June 17vNov. 20, 1969 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 5 800 190

 

022

GILA RIVER PHREATOPHYTE PROJECT

 

 

 

 

 

 

100
T I I | I I
EXPLANATION
90 \‘\ Dense—saltcedar
.._ \‘ ___ —
\_\ Cultivated grass
'\-\ All saltcedar in reach 2
E 80 — \\ Average radiance.in reach 2 ""
LLI '\ __._
g ‘\ All mesquite in reach 2
LU .\ -¢—1—u—1-4-I
l 70 _ '\ Whitethorn _
E ‘\ ______
‘ ‘\ Creosote bush
LLI .
o \.
Z
S 50 — _
D
<
D:
II
_ \\\
Q 50 —— ‘—
LI.I
II \ \
< \
II
LL
3 4o — ——
l
I
<
Lu
2
u.I 30 -— —
2
f.
<
_I
LLI
c: 20 — _
1O — —
o I I I I I l
0 10 20 30 4O 5O 60 7O 80 90

RELATIVE RED IRRADIANCE, IN PERCENT

100

Data from Table 1

FIGURE 18.——Regression relations between relative near-infrared irradiance and relative red irradiance
among vegetation types in and adjacent to the Gila River Phreatophyte Project, 1968 and 1969.

sion equations (table 2) indicate that the spectral sig-
nature for ponderosa pine is more nearly constant and
has less variation than juniper-pinyon.

The spectral signature of the total ground cover may
well be used as a remotely sensed index of hydrologic
activity. The signature is an integrated index of the
actual plant morphology and pigment chemistry,
which is influenced by many direct and indirect hydro-
logic parameters such as radiation, temperature,
ground-water level, and available soil moisture.

During the period of Vigorous plant activity on the
Gila River Phreatophyte Project (May through August
for most vegetation, April through September for dense
saltcedar), the time-independent signature of the vege-
tation types were related (fig. 20). This illustration
demonstrates that discrimination of the primary vege-
tation types (dense saltcedar, saltcedar, and mesquite)
on the Gila River Phreatophyte Project is possible with

 

these parameters during this seasonal period. The
sparse upland vegetation cannot be distinguished be-
cause of the reasons given in conclusion 1.

The signature equations expressed for saltcedar and
mesquite in table 1 do not show a large difference in
the R versus 0 signature between the two plant vari-
eties. However, these signature equations represent
very large areas where foliar cover (the amount of
ground covered or shaded by the vegetation foliage)
varies from 0 to 100 percent and the underlying cover
and soil type vary considerably. These variabilities
tend to mask the actual R versus G signature of the
specific plants. The plant signature must be isolated
from the contribution of underlying vegetation and
ground irradiance. Horizontal photography at a dis-
tance of approximately 50 feet (15 In) from a densely
foliated saltcedar and a similar mesquite was used to
determine the R versus G signature for the two types of

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

023

TABLE 2.—Time—independent analysis using color—infrared film for vegetation in the Cibecue Ridge area
[Data used in analysis are from flights made in 1971; for a more complete statistical analysis see table 5]

 

 

 

 

 

 

 

 

Number Number of Coefficient Standard
0 sample points Equation Regression of error of
Vegetation flights per flight number equation correlation estimate
Ponderosa pine (Pinus ponderosa) ,,,,,,,,,,,,,,,,,,,,,,,, 6 5 29 R :—0.85(G) + 88.3 #086 2.9
Juniper-pinyon pine (Juniperus osteasperma, Pinus edulis) m, 6 10 30 R = 7 64(0) + 701 7 ‘94 3.2
Grass sites (Blue grama, Boueteloua gracilis; R : 7 .42(G) + 55.4 7 .68 55
weeping love grass Eragrostis curuula) ________________ 6 30 31
Bare ground ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 6 15 32 R=— 138(0) + 494 — ,92 212
1 00 -l l I I
90 — —-
E X P LA NAT I O N
Ponderosa pine
E 80 "‘ Juhi—p-e: fp-ihvon "-
8 ._ _ __
I: G rass sutes
:1 ———,
Z 70 _ Bare-ground Sites __
u;
0
Z
S 60 —— —
D
<
CC
CC
0 50 — -—
Lu
{I
<
[C
u.
g 40 — —
I
<
Lu
2
Lu _ __
2 30
'—
<
J
Lu
m 20 _ _.
1O — —'
o I | | I l J l | l
O 10 20 30 40 50 60 70 80 90 100

RELATIVE RED IRRADIANCE, IN PERCENT

Data from Table 2

FIGURE 19,—Regression relations between relative near-infrared irradiance and relative red irradiance
among vegetation types in and adjacent to the Cibecue Ridge area, summer 1971.

vegetation. The regression equation for saltcedar is \
R=—2.30(G)+109.0, (33)
and the coefficient of correlation is —0.99. The regres-
sion equation for mesquite is
R —0.92 (G)+90.9, (34)
and the coefficient of correlation is —0.94. Discrimina-

tion of vegetation type is possible with ground-level
photography, because the rate of change of R versus G

with ground-level photography for saltcedar is approx—
imately twice that of mesquite. The given slope of the
R versus G equations for the two vegetation types i1-
lustrated cannot be directly equated to aerial photog-
raphy, because of the increased contrast ratio which is
used to compensate for haze in aerial photography.
This effect will be discussed in the section “Gamma.”
The signature equations (table 1) for these two types of
vegetation are integrated readings of the total ir-

024

GILA RIVER PHREATOPHYTE PROJECT

 

100

l
90
80 —

70—-

50—
40—-

30—

RELATIVE NEAR~|NFRARED IRRADIANCE, IN PERCENT

20..—

 

o | J

 

EXPLANATION

I
Dense saltcedar

El
Saltcedar

A
Mesquite

.
Whitethorn

O
Creosote bush

 

Vegetation zone

 

I v

 

0 10 20 30 40

50 60 70 80 90

100

RELATIVE RED IRRADIANCE, IN PERCENT

FIGURE 20.——Relative near-infrared irradiance versus relative red irradiance for five vegetation types in
and adjacent to the Gila River Phreatophyte Project, summer 1968 and 1969.

radiance from the photographed surface. For precise
discrimination of plant type by aerial photogaphy, the
relative contributions of each ground parameter
(short-lived ephemeral plants, soil, water, and others)
should be evaluated.

TIME-DEPENDENT SIGNATURE

The spectral signature change with time is used to
monitor a dynamic phenomenon, and this precludes its
use as a predictive tool; however, a model curve can be
used to determine the departure of the spectral signa-
ture from the expected value, thus enabling the re-
searcher to see trends and anomalies which should be
investigated. The variability of the R and G spectral
signature change with time is shown in figure 21; R, G,
and the modeling curves are given for all of reach 2
during 1968 and 1969.

The general model curve used in this study has four
distinct segments. First, is a linear segment for the

 

mean spectral signature during the period of dormancy
of the deciduous vegetation, from late fall to early
spring (4 —1’ on the R’ curve). Two nonlinear segments,
1—2 and 3—4, on the R’ curve (fig. 21) represent the
transition from winter dormancy to high summer plant
activity and from high summer plant activity to winter
dormancy, respectively; these segments are described
by equation 35. The last is a linear segment (2—3 of the
R’ curve) derived by assuming linearity between the
two peaks of the bimodal curve from equation 35.
These peaks occur on May 19 and August 11. This
segment is assumed to represent the average summer
maximum plant activity value.

The generalized form for the equations describing

the model curves are given below:
R’=[H(sin N+1/5 sin 3N) +P]>< 100, (35)

where
R’ = value of the model curve for relative near-

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

025

 

T
O) \l 00
O O O

01
O

N (.0
O O

RELATIVE IRRADIANCE, IN PERCEN
8 S

 

I I I I I I I I I I I

 

O

 

 

 

EXPLANATION

0
Relative near »infrared irrad iance

1
_—

Model curve for relative near-in»
frared irradiance,‘number
indicates a time of change in
the spectral signature of
vegetation

I
Relative red irradiance

Model curve for
relative red irradiance

FIGURE 21.—Re1ative near-infrared irradiance, relative red irradiance, and model curves, reach 2 of the Gila River Phreatophyte
Project, 1968 and 1969.

infrared irradiance, in percent,

H = the amplitude of the sine wave for relative
near-infrared irradiance,

P = the mean relative near-infrared irradiance
during dormancy, and

N = a function of time, expressed as

N=[<C—D>(1—§f£)l

where
C = the calendar-year day on which the photo-
graphic mission was flown [January 1, C = 1;
December 31, 0:365 or 366 (leap year)],
D = calendar-year day for the end of dormancy de-
termined by the low relative near-infrared
irradiance, and
L = duration, in days, for the period of plant vigor,
determined by the change in relative near-
infrared irradiance,
and '
G’={H’[sin (180°+N’)+
1/5 sin(180°+3N’)]+P’} X 100, (36)
where
G’ = value of the model curve for relative red ir-
radiance, in percent,

 

H ’ = the amplitude of the sine wave, for relative

red irradiance,

 

P’ = the mean relative red irradiance during
dormancy, and

N’ = a function of time, expressed as

180°
N:[(C —D,) ( r )J’
L
where
D’ = calendar-year day for the end of dormancy

determined by the high relative red ir-
radiance and
L’ = duration, in days, for the period of plant
vigor, determined by the change in relative
red irradiance.
The actual model equations used for figure 21 are given
below:

 

R’=[0.47(sinN+1/5 sin 3N)+0.03]><100, (37)
where
180°
N—[(C'—74)( 214 )]and
G’={0.48[sin(180°+N’)+1/5 sin
(180°+3N')]+0.76}><100, (38)

where

O26

180°
245

 

N’=[(C'—60)(

)l

The accuracy of fit of the model equations to spectral
data was checked by relating the data from the 233
data points presumed to be representative of the
change of signature with time to the predicted values
from the R’ model equation 37 represented in figure 21.
The standard error of estimate (Sy'x) from the line of
equality of 0.04 is shown in figure 22.

The time-dependent plant signature (R and G) of
various types of vegetation are shown on figure 23, and
curves similar to those of figure 21 could be developed.

Several observations can be made with the aid of the
model curve.

1. The time-dependent signature of large areas can
be used to recognize and map vegetation types by using
the sensing bands of satellite imagery in a similar
manner to that proposed here for aerial photography.
The 9- or 18-day cycle of LANDSAT is ideal for
monitoring these changes.

2. Probably the most useful application of the model
curve for aerial photography is to determine a schedule
of flights to define the periods of growth at a particular
study site. The minimum number of flights during a
year to define the growth periods reliably for the Gila
River Phreatophyte Project, as indicated by figure 21,
is five: one taken during the period of dormancy
(November through February), one during the spring
transition period (April 1), two taken while the plants
are vigorously active (June 1, July 30), and one during
the fall transition period (September 30). For a long
term investigation where only one flight per year is
feasible, it would appear that sometime in June is the
best choice for this particular area.

3. The transition period, from dormancy to Vigorous
activity as sensed by the G sensitivity range, was not
the same in 1968 as it was in 1969. The transition
period started approximately 1 month later and ended
1 month earlier in 1968 than in 1969. The reflectance
in the visual spectral range, of which G (red color) is
functionally related, which responds to photosynthetic
activity, is well documented (Knipling, 1970, p. 158;
Gates, 1970, p. 226). This time lag in activity between
1968 and 1969 is probably a function of many interre-
lated contributing variables such as radiation, temper—
ature, rainfall, and others.

4. These modeling techniques were not successful on
the Cibecue Ridge area, primarily because the pre-
dominant vegetation is perennial and the small signa-

3Only data from 23 ofthe 25 photographic missions flown during 1968 and 1969 were used
in this computation because of the reasons given in footnote 2.

 

GILA RIVER PHREATOPHYTE PROJECT

 

50 I l | I

A

R: 1.00 R’+ 0.00
r = 0.98
Sy-x : 0.04

40

0.)
O

N
O

10

RELATIVE NEAR—INFRARED IRHADIANCE,
IN PERCENT FROM FILM

 

 

 

1
o I I | I
o 10 20 30 4o 50
RELATIVE NEAR—lNFRARED IRRADIANCE, IN PERCENT
FROM TIMEEDEPENDENT MODEL EQUATION

EXPLANATION

A

R
Relative neareinfrared irradiance sensed by the film
R’
Relative neareinfrared irradiance predicted
by the model equation

 

r
Coefficient of correlation

Sy.x
Standard error of estimate

1968
A
1969

FIGURE 22.—Regression relation between relative near-infrared ir-
radiance determined from film and from a time-dependent model
equation, reach 2 of the Gila River Phreatophyte Project, 1968 and
1969.

ture change with time is largely over-shadowed by the
inherent error in the photography.

SPATIAL COMPUTER ANALYSIS

The R versus G signatures can also be used as a
means of spatial evaluation. Figure 24 shows the tri-
chromatic coordinates subdivided into 10 percent sub-
groups. The first digit indicates the percent of R, and
the second digit the percent of G; for example, the
number 72 means 70 percent relative near-infrared ir-
radiance and 20 percent relative red irradiance. The
remaining parts out of a hundred, in this case 10 per-
cent, signify the relative green irradiance. Figure 25
shows a computer printout of the R and G trichromatic

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

90
80
70
60
50
40
30
20
1O

0

A

R AND C, IN PERCENT

A

90
8O
70
60
50
4O
30
2O
10

A

R AND C, IN PERCENT

A

80
70
60
50
40
30
20
10

A
R AND C, IN PERCENT

A

80
7O
60
50
40
3O
20
10

A
R AND C, IN PERCENT

A

90
80
70
60
50
40
30
20
10

A

A

R AND C, IN PERCENT

 

NOTE: Irregularity of —
spectral signature
mainly is due to
irrigation

   

 

| I I I I I I I I | I I I | | I I I I I I I

CU LTIVATED GRASS STUDY SITE

 

    

‘~‘ [z
‘~/

 

| I I I I I I I I | I | I I | I | I | I I

WHITETHORN STUDY SITE

 

 

 

 

CREOSOTE BUSH STUDY SITE

 

 

 

 

ALL MESOUITE IN REACH 2

 

 

\
‘ ’— _// \\
.- ——

 

 

 

\ \
F'M'A'MJ JAs'o'N'DJ‘F'M'A'M'J'J‘A'S'O‘ND
1968 1969
ALL SALTCEDAR IN REACH 2

EXPLANATION

’ Relative red irradiance
= Relative near-infrared irradiance

W>Q>
I

FIGURE 23.—-Relative near—infrared irradiance and relative red irradiance versus time for five vegetation

types in and adjacent to the Gila River Phreatophyte Project area, 1968 and 1969.

()2?

028

100 PERCENT

 

 

 

 

GILA RIVER PHREATOPHYTE PROJECT

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G
10} EXPLANATION
§ fi = Relative near~infrared irradiance, in percent
a = Relative red irradiance, in percent

1— 90 091 = Relative green irradiance, in percent
2 Multiply first digit by 10 to obtain percent R, 30
8 80 08 18 28 percent;Arnultiply second digit by 10 to oinAtain
E percent G, 60 percent; percent B = 100 — R — é
a.
Z 70 O7 17 27 37
Lu.
2 60 06 16 26 36 45
S
9t 50 05 15 25 35 45 55
EC
0:
a 40 04 14 24 34 44 54 64
Ll.l
fl:
u.I 30 03 13 23 33 43 53 63 73
2
l-
2: 20 02 12 22 32 42 52 62 72 82
NJ
[I

10 01 11 21 31 41 51 61 71 81 91

O 00 1O 20 30 4O 50 60 70 80 90

O 10 20 30 40

100 PEfiCENT
B

50 100 ;
RELATIVE NEAR—INFRARED IRRADIANCE, IN PERCENT

60 70 80 90

100 PERCENT
13

FIGURE 24,—Computer printout codes used in spatial analysis of relative
near—infrared, relative red, and relative green irradiance.

coordinates for each 3.67-acre (1.49—ha) plot in the
heavily vegetated reach 2, sampled photographically
on June 27, 1968. Areas of near-equal signature can be
contoured on the printout and compared with ground
truth. Above the printout is a reduction of a composite
print mosiac copied from the positive transparencies
analyzed. The connecting lines link nine pairs of tri—
chromatic coordinates (R and G) with their correspond-
ing position on the photograph.

Five photographs were combined for the photomosaic
(fig. 25). An important feature of the photomosaic is the
inconsistent spectral signatures of the five sequential
photographs. The mosaic demonstrates the necessity of
quantitative adjustment of the signature data by some
technique if a numerical approach to photographic
evaluation is required.

Computer printouts of the type shown on figure 25
were run for all reaches in each flight. This aided
greatly in the spatial and temporal evaluation of the
vegetation. The potential for computer interpretation
of large areas by this method appears to merit consid-
erable research.

VOLUME OF CANOPY FOLIAGE

Volume of canopy foliage is a difficult and expensive
hydrologic parameter to measure by conventional
means; it is conventionally defined in cubic feet per
square foot of area. The technique used to determine
this parameter on the Gila River Phreatophyte Project
has already been outlined in the section “Vegetation
Data Collection.”

 

Thomas, Wiegand, and Myers (1967, p. 553) found
that an exponential increase in reflectance was
achieved by stacking individual leaves on one another.
This enhancement of reflectance would imply that
canopy volume could be estimated by photographic re-
mote sensing, but the condition of stacked leaves might
result in quite different reflectance relations than
those of a natural plant canopy.

Figure 26 shows the relationship between nine
canopy volume classes of saltcedar and three canopy
classes of mesquite versus relative near-infrared ir-
radiance from four photographic missions.

The photographic missions chosen represent the
spring period of new leaf production (April 5), the sta-
ble summer period of dense foliage (May 31 and June
27), and the fall transition period leading to dormancy
(October 10). These periods are demonstrated on figure
21 and discussed in the section “Time-Dependent Sig-
nature.”

The data from the saltcedar classes (B) indicate that
the relationship between canopy volume and R is not
linearly related. An important point, which must be
considered while viewing figure 26, is that the vegeta-
tion description of canopy classes was done in 1965,
while the photography used in the figure was taken in

 

FIGURE 25.——Color-infrared mosiac showing spatial computer
analysis by spectral signature for 3.67-acre (1.49-ha) plots in reach
2, Gila River Phreatophyte Project, June 27, 1968.

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

  
   
 
   
 
 
  
 
 
    
 
 
 
 
  

mu RIVEG DHREATDDNVVE noun
InVLflPHEYATKON or AERIAL INFRAiED EKYACM‘OIE Motown—v me»- : 1w: or ass 3

rue-n on on 6-27-08 ELEvnlou of rum" 1; nsan FEEY

_ "5.. ..... "5..

   

5‘ 3A 3t 3‘ ‘1 3A 3‘
u ‘1 3‘ tt 3A 3‘ 3‘
ts tt 3‘ IN tt ‘3 ll
Cleared area .. .. .3 .3 u at
53 3A 3‘ 5‘ ‘l

‘3 at u u u

 

:n as u u u

3‘ as u 53 u

 

 

 

as u u u u 52

Open mesquite forest “ .1 u .. .3 u

 

 

 

  

Dense saltcedar forest .. s: .2 7 ,2 n

53 5: 3A 72 n

 

 

 

u 11 u 71 71

35 u 53 6| u 12

u u 53 In 11 n
u at 5: 72 u 53
u 02 3t 3‘ 12 I] 12
u 53 a u 11 12 63

:5 as 53 5: 7a 52 a: 12 53

u u 7: 12 5: u 15 u 1: 63
n :3 12 a: ‘3 u 25 u 53 a:
u u 72 25 25 551510 25 1:
‘3 53 25 3A 15 25 15 u oz 52
53 u 25 15 u 15 at 25 72 t:
25 a7 25 n 15 u u. 25 3t 53
01 25 M u )5 lo M In on 53

 

n7 2! M 05 at n :5 u M 5:

Inundated area— water ,. u ,. ,,, .. ,5 ,5 5,

n M u an on as at 15 u ‘3

 

 

 

u 53 u u n 15 u ‘15 62 u
53 u u n on no on us A: u
u 3‘ u 53 at no at 25 52 u

o 1M|LE

O 1 KILOMETER

 

u
5:
53
a:
5:
u

u

"1..

 

1n
n
20
an
to
an
an
ll
21
n
21
an
22
22
22
z:
a:
23
2:
23
z:
a:
h
u
a.
2‘
as
25
25

25
25
a
u
:5
20
u
21
27
21
21
21
n
z.
u
u
an

029

030

GILA RIVER PHREATOPHYTE PROJECT

 

 

 

 

 

 

 

 

 

 

 

    
    

 

 

 

 

 

 

 

 

 

 

 

| I I l | | |
EXPLANATION
4e 5—68 Partial foliation U
------- 5—31 ~68 Fullfoliation a
——— 6-28—68 Full foliation 3
Average Range —————— 10—10—68 Unfoliated 0 cc
number Range of of 1 Standard deviation E E
of _canopy, height, H—' M";
sampled In percent 1n feet Pooled standard deviation of irra» _ 2 211.1
plots of area (meters) diance for the three mesquite — a:
canopy classes;4 45768 data is 3<
7 percent LL 8
5.0 - -
,_ > w
173 50775 7+ 111 a- 1:1:
(2.1) > u.1 I 9 0 Lu
:1. LL I- 40 _ 1 / _ Z 1:1.
0 O ' <1
0 I I / (1)
Z — O I . 0
<( 1:: IL I I ,/ — 1 1.1. E
7+ 0 3 Lu 3-0 — ' l ‘ o 1—
45 25—50 1.1. 0 1:1: I u:
(2.1) o z < 1 L” 2
Lu ' D 2 o — 1' _ '2'
2 “id ' .' 3
D 0 w I o
5 o 2 7+ 5‘: E 1'0 _ 1 _ >
1 ' 5 (21) >311 A. MESOUITE
”- 1 1 1 1 1 1 1 0
0 1O 2O 30 40 50 60 7O
RELATIVE NEARilNFRARED IRRADIANCE, IN PERCENT
Average Range of Range of I I ‘ I I I I a:
number canopy, weight, i— 14'0 _ ' 1.1.1
of same in percent in feet 0 E
pled plots of area (meters) E 130 _ _— 4 2
”J Lu
1: I
g 12.0 — ‘ g
_ 13+ 0 o
51 75 100 (4) m 1 U,
a: 11.0 — I‘ _ 0:
E 1 E
I m
E 10.0 — \‘ _— 3 5
1.1.1
t 1 »
13+ U 9.0 — I - g
31 50775 (4) 5 xx)
166 757100 6.5413 3 T ' 9
(2—4) 0 8-0 — \ ‘ g
z \ 0
_~ ‘
u; 7.0 — \ - E
0 1 _2 .
_ 1 111
19 50775 6-5 13 3 r o
(2-4) —I 6.0 — ,‘ / ‘ <
O / //‘ . . . —
13+ 1.1. , ///_/ Pooled standard devuatlon of Irra- _|
30 2550 (4) > 50 _ 1‘ diance for the nine saitcedar 0
CL ‘ I canopy classes;4~5—68 data is LL
2 III 14 percent 3-.
_ 4.0 — — O
34 2550 6g A? 8 4 \ \: z
0 V LL \\ \\ -\\ _ 1 <
23 757100 ‘6-5 o 3 _ \-\ _ 0
(0—2) '0 \\ \ \ 1.1.
0 g \\ \\ ’\, o
9 50775 '6'5 3 ' /) a 5" N
(042) 2.0 — // /// " E
0 65 6' // //// 3
6 25750 ‘ > ’ ./- .1
(0—2) 1.0 _ _ o
B. SALTCEDAR >
1 1 1 1 1 1 1 0
0 1O 20 30 40 50 60 70

RELATIVE NEARJNFRARED IRRADIANCE, IN PERCENT

FIGURE 26.—Volume of canopy foliage versus relative near-infrared irradiance for photographic flights during 1968, reach 2, Gila
River Phreatophyte Project.

1968. The effect which this time lag had on the data is
unknown. Several conditions and relationships are i1-
lustrated by figure 26. Mesquite volume (A) also seems
to indicate a nonlinear relationship, although it is dif-
ficult to postulate any conclusions based on these data

alone, because of the limited number of canopy classes.
It is apparent that although the canopy class versus
irradiance curves are nonlinear, they all consistently
exhibit the same general shape.

The transition period leading to dormancy, as shown

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

by the October 10 flight, shows that the irradiance is
almost constant and apparently not related to canopy
volume. But there is a general trend for greater foliar
cover to produce higher irradiance, particularly in the
larger volume classes.

A representative standard deviation for each data
point for April 5 is graphically shown in figure 26. The
pooled standard deviations (Dixon and Massey, 1957,
p. 109) of the irradiance for saltcedar and mesquite are
14 and 7 percent, respectively, which indicates that the
error is too large for any precise conclusions.

Figure 27 illustrates the relative near-infrared ir-
radiance of three saltcedar foliage classes versus time.
This illustration shows that an increase in irradiance

 

031

accompanies an increase in volume of foliage. The
pooled standard deviation of the largest class is 12 per-
cent. Figure 28 illustrates a similar relationship for
mesquite; the pooled standard deviation is 13 percent.
This large error factor indicates that for the data used,
it is not practical to use this technique to measure
canopy foliage.

FOLIAR COVER

Foliar cover is defined as the ground covered or
shaded by a vertical projection of plant foliage. Two
techniques of photographic analysis may be used to
determine the percent of canopy intercept in an
area—Visual examination and spectral signature.

 

 

 

 

 

       

 

 

 

‘00 I I I I I I I I I
90 — _
F.
E 80*— ? 1Standard deviation '—
0
CE
m
a. / \
z 70 _ / \ _
‘ / \
LIJ‘ / \
/ \
U / / \
z \\ // \ // \
<51: 60 _ I \‘K \ I \ _
I
< I \\ I\\
I I \ I \
[C I, \ I \\
“ \ I \
D 50 — / I I \ “
Lu / \ I ,/ / \ \ I \
I: / \ I /' \ \ I
< / \ I / \ I,
I: \ I I L \ , \
u. \ I l \
z 40 — \ I \ —‘
—I \ ‘
[I I, A \ \\
3 I, / \
Z 30 -— I / Pooled standard deviation of irra» \ \ —
g / diance for11.4 ft3/ft2 canopy \
’: class during 1968 is 12 percent \
3
u, 20— —
CC
10 — —
0llllllllllllllllllllllllilllllllllllllllllllllllllllllllillllllllllllll
Jan. Feb. Mar. Ap-r. May June July Aug. Sept. Oct. Nov. Dec.
1968
EXPLANATION
VOLUME OF CANOPY
FOLIAGE, lN CUBIC
FEET PERSOUARE NUMBER OF SAMPLED TOTAL ACREAGE
SYMBOL FOOT PLOTS SAMPLED
3.6 (1.1 m3/m2) 34 125( 51 he)
—— 3.5 (2.6 m3/m2) 166 608 (246 he)
————— 11.4 (3.5 m3/m2) 51 187 ( 76 ha)

FIGURE 27.—Relative near-infrared irradiance versus time for three saltcedar volumes of canopy foliage classes, reach 2, Gila River

Phreatophyte Project.

O32

GILA RIVER PHREATOPHYTE PROJECT

 

 

 

 

 

 

  
 

  
  
 

 

100 I i I | l l
EXPLANATION
VOLUME OF CANOPY
90__ FOLIAGE,|N CUBIC ___‘
FEET PER SQUARE NUMBER OF SAMPLED TOTAL ACREAGE
SYMBOL FOOT PLOTS SAMPLED

#- _ — 0.9 (0.27 m3/m2) 173 635 (257 ha)
2 so 3 2 _.
uJ —— 2.7 (0.82 m /m ) 45 165 ( 67 ha)
3 ————— 4.4(1.34 m3/m2) 51 187( 76 ha)
LU
D.
z 70 — __
ui
0
Z
1‘ 60 —- _
D
<
n:
E
o 50 — —
LIJ " \
0:
<
n:
LL
2 40 — 1 Standard deviation -‘
c'c
<
u]
2 30-— _
L; Pooled standard deviation of irra-
_ diance for 4.4 ft3 /ft2 canopy
2 class during 1968 is 13 percent
—l _
Lu 20 _
a:

10 — "‘

0 I|||lllllilllllilllllllllllli llllllllllllllllllllllllllllllllllllllli J

 

 

 

 

Jan. Feb. Mar. Apr. May June

July Aug. Sept. Oct. Nov. Dec.

1968

FIGURE 28.—Relative near—infrared irradiance versus time for three mesquite volumes of canopy foliage classes, reach 2, Gila River
Phreatophyte Project.

Visual examination generally requires mapping of
canopy-intercept classes from aerial photography and
field reconnaissance, planimetering, and computing
the mean percent of canopy intercept in the area
(Griffith and Howe, 1960). Spatial computer evaluation
of imagery can aid greatly in the use of this technique,
as shown in the recent studies by Denny, Morrison,
Worthman, and Lucht (1971).

Photographic analysis using the spectral signature
technique can probably be best accomplished by near-
infrared photographic sensing. Figure 29 illustrates
the relationship between foliar cover and the R signa-
ture for the 50 plots within grids 2—24 and 2—25 of the
Gila River Phreatophyte Project. The regression equa-
tion describing the relationship is

CI=1.54(R)—29.9, (39)
where
CI=foliar cover, in percent.
The coefficient of correlation is 0.85. The variations

 

within the individual grids as shown in figure 29 are
due primarily to two reasons.

1. Each plot has a different percentage of water
showing through the canopy, which greatly affects the
spectral signature registered on the film.

2. The actual distribution of vegetation types
(saltcedar, mesquite, and grasses) is different in each
plot, therefore yielding variability in the signatures.

Interception of precipitation by vegetal cover is a
difficult hydrologic parameter to measure, but photo-
graphic remote sensing can be used to estimate both
the spatial and temporal variability of interception. In-
terception is defined as the amount of rain or snow
stored on leaves and branches that eventually evapo-
rates back into the atmosphere. It is equal to the pre-
cipitation on the vegetation minus stemflow and
throughfall (Langbein and Iseri, 1960, p. 12) and is
basically a storage characteristic of the hydrologic
cycle; it also is a function of other hydrologic, climato-
logic, and botanical variables, such as vegetation type,

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

1°°IIII|I

 

‘0
o
l

=1.54fi—29.9

= 17.7
=0.85

C]
S y.x
r

on
o
l

   

01 m \I
o o o
| l l

CANOPY INTEHCEPT, IN PERCENT
&
O
I

 

0 1O 20 3O 40 50 6O

   

033

EXPLANATION
o
Plots located in grid 24—2
_ (fig. 1)

o
Plots located in grid 25-2
(fig. 1)

A

R
Relative near-infrared irra-
diance, in percent

— Cl

Foliar cover, in percent
Sy.x

Standard error of estimate

r
Coefficient of correlation

 

 

I
80

70 90 100

RELATIVE NEAR—INFRARED IRRADIANCE, IN PERCENT

All data from photography
of May 31, 1968; altitude
3,600 feet

FIGURE 29.—Foliar cover versus relative near-infrared irradiance for 50 3.67 -acre (1.49—ha)
plots, subreach 2b, Gila River Phreatophyte Project.

leaf area, Wind velocity, and rainfall intensity and du-
ration. As precipitation strikes a vegetative canopy,
some is retained and distributed over the exposed sur-
face area. If there is adequate precipitation, a quasi-
equilibrium is reached, whereby interception storage
becomes stable and stemflow plus throughfall equals
precipitation. The intercepted precipitation is then re-
tained (stored) on the leaf and stem surface until after
the storm or until the vapor-pressure gradient is high
enough to initiate evaporation. For 100 percent canopy
cover, interception of precipitation from an individual
storm may be estimated using the curves of Kittridge
(1948, fig. 12) or from the interception data of M01-
chanov (1960). Photographic remote sensing then can
be used to approximate the percent of can-
opy cover—the macroleaf count—on a watershed,
which is then used to determine the interception
storage.

TECHNIQUES FOR ANALYSIS OF
COLOR-INFRARED PHOTOGRAPHY

In the 5-year period 1967—71, 38 color-infrared pho-
tographic missions were flown over the Gila River
Phreatophyte Project. During the summer of 1971, six
color-infrared photographic missions were also flown
over the Cibecue Ridge Watershed Study. The photo—
graphic missions were flown at altitudes between 1,500

 

and 60,000 feet (460 and 18,000 m) above land surface
under many different climatic conditions using a va-
riety of cameras, films, and filter combinations. This
is shown in figure 30. The quantitative utilization of
these diverse data required the development of both
deterministic and empirical techniques of film stand-
ardization.

Figure 30 shows both the adjusted and unadjusted R
for subreach 2b from 1967 through 1971. The unad-
justed R is the parameter described in the section
"Data Standardization and Sources of Variability.”
The adjusted R has been adjusted for the variability
due to the inconsistency in flying altitude, filter combi-
nation, processing techniques, or other contributing
factors. The effect of adjustment is illustrated by the
signature difference in unadjusted R between the
summer of 1968 and 1969. The difference between un-
adjusted signatures is not related to the vegetation
condition because the phreatophytes were equally
dense during both periods. The adjusted values for this
same period are nearly equal, as would be expected if a
correlation exists with ground conditions. Most of the
photographic flights during the 5-year period (1967—
71) were flown at an altitude of 8,500 ft (2,600 m) above
land surface; however, six flights were flown at 3,600 ft
(1,100 m), and two at 60,000 ft (18,000 m); therefore,
the adjustments described above are necessary.

GILA RIVER PHREATOPHYTE PROJECT

034

.Elbme £8.55 wismfimogsh 33m «:0 23 mo am £3895 ducwmwmgi vwaméfiimw: 9,532 $55.95;: 38 wwumdwfiw‘dm 559m

 

 

 

 

 

 

Fhmr Ohms 000—. moor nwmw
o_z_o_m_<_q_ 131.2?— _. 1107??? T1127; 3 o z_o_m_<_ a? 7312?: o_z_o_w_<_1_n_2_<_s__m T 1111111312 0
Ma
3
—I
V] b v
< u
U an A
/ a
.l N
l D 4 / mu ow 3
, / IW
/ N _|
O 0 MR u N
mu
4 av
MB
2..
I Q l ow o
\ m
H
w
1 m
N
m
L l ow .
DmFUmr—KOU DWFUMKIOU DmFUWIKOU DWFUNIEOU DWFUwEIOU OmFUmEIOU DwFUwEKOU DwFUwEIOU Humvtmv DI<DZ<FW
DWPUWIEOUZD DmPme—EOU DWFUMIIOUZD DwFUmmw—OUZD DmFUwr—EOU QNFUWEEOUZD DmFUwEw—OUZD ONFUNEIOQ mmmFl:l
DmFUWKIOU DmFUwr—KOU DkamIIOU DMFUWEIOU QwFUmIEOQZD DWFUMKKOUZD DWFUNKEOUZD DWFUNEIOUZD ml)... _2l__n_
QmFUmImOUZD DNFUWIKOU DWFUmmm—OUZD DmPUWEEOO DmFOmIm—OU DmFUmEKOUZD DNFUNEEOUZD DwPUwIEOU mODF_._.J_<
WFZWEFWDfiD<
20000 EOMUU .MONUU mF>> EONUU
EOMUU .mONUU .mF>> m_.>> .MONUU .N—LS .N:S Eme—Z ENFWSZJJZZ .mONUO ‘Nr>> 20000 .mONUU .NF>> NF >> [mg-12¢
IUZ_ m‘MVVN IUZ_ mwmvvw IUZ_ m\m<vfl I_l_|_:>_ ONIMVQN mMImvfiw IUZ_ GIMva IUZ. m>mfl¢m I02. m‘MQVm F<§EOEIEIZ¢
_._.r._. mm_wN hpr. D<Jm4mmw<I ZOv:Z hrtV. NFIM hrlv. (Em—240
meu oom‘w men. OOOdm mem 00m.m Emma comm Fwwl acqow meu. 00mm meu 00mm Fwwu Doom wDDF_._.I_<
mw4m<2m<>
Emu 320 3a....

Emu “553.695 . O
Hymn nwumlum. .

">a>5m $030.80 .m .3

033.52: .0 gaming: . D
$qu 33mins K. .5me DoumEUm .-

UmumiUmc: q
Hum—u uufilnu “

mam—u naming: .q
“nun—u Duamamvm ‘ ‘

"notmbflfiE—ulx

 

”:o_umtm_c_rcu<

 

mownm ucn u>m>5m n>m>5w
3:22.92 2360.80 .m .3 $33.80 ‘m .3,
3:2me

oumnm Ucm
motzmcohmd‘ 3:2me

320 3333:: .O
3qu “533:3 ..

”>w>.:m
$23.80 .w .3

mama Damp—DEL: . O
“Emu Swami—nu . .
u>w>5m _mu_mo_owmu .w .3

 

mama uuumiumcz . O
.Vmumu vagina . O
u>m>hzm 3030250 .w .3

FIG—IE 023:1). >02m0<

 

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

COST OF PHOTOGRAPHIC ANALYSIS

The cost in 1968 of using conventional techniques to
assess the vegetation on the Gila River Phreatophyte
Project (the technique is outlined in the section “Vege-
tation Data Collection”) was approximately $6,600,
while the cost of a color—infrared photographic mission
and a complete computer analysis of the photographic
data cost approximately $650.

BASIC PHOTOGRAPHIC CONCEPTS

The analysis of photographic remote sensing data
requires a basic knowledge of the fundamentals of pho-
tographic science. Spectral sensitivity, density versus
log exposure, and analytic optical density are photo-
graphic topics necessary for the type of analysis
utilized in this report.

SPECTRAL SENSITIVITY

Spectral sensitivity, S()\), for color-infrared film is
defined as follows:

S()\)= 1/E(}\), (40)

where E (A) is the energy in ergs/cm2 of monochromatic
radiation required to produce in the individual dye
layer an equivalent neutral density, usually 1.0 (N.L.
Fritz, 1969, written commun.). Equivalent neutral
density (END) is defined as the visual density a dye
layer would have if it were converted to a neutral gray
by superimposing the minimum necessary amounts of
dye in the other two dye layers (Todd and Zakia, 1969,
p. 215—219).

The relationship between wavelength sensitivity
and color formulation of color-infrared film was discus-
sed in the section “Color-Infrared Film” and was dem-
onstrated in figures 2 and 3.

Multigeneration photography such as received from
NASA during the study refers to the fact that the
color-infrared photography has been duplicated on
color film for data analysis. A quantitative analysis of
the data must include consideration of the sensitivity
of the duplicated and duplicating films.

FILM DENSITY VERSUS EXPOSURE

The relationship between film density and exposure
was discussed in the section "Color-Infrared Film” and
shown in figure 4.

BASE PLUS FOG DENSITY

Base plus fog density (D0) may be referred to as the
fog level of the film. It is the optical density obtained
after processing film which has received no exposure
and is the result of chemical reduction by the developer
on unexposed grains. For example, the base plus fog

 

035

level for the cyan dye layer in figure 4 is an optical
density of approximately 3.6. The actual base plus fog
densities of the dye layers of film type 8443 analyzed
with our equipment are shown below:

Cyan D0 = 2.76,

Magenta D0 = 1.38, and

Yellow D0 = 2.22.
GAMMA

Gamma (y) is the slope of the linear portion of the
characteristic curve:
2 AOD .
Alog E

 

v (41)

A perfect tone reproduction of a scene is obtained
when the gamma is equal to — 1 (for positive film), and
a gamma less than —1 means that the scene is repro-
duced at a higher contrast. (This discussion applies to
positive film only; a negative film would have a posi-
tive gamma.) Aerial films, such as color-infrared film,
which are used for high-altitude photography usually
have gammas of approximately —2.0. The reason for
the low gamma of aerial films is because the original
scene contrast is greatly reduced by atmospheric at-
tenuation, and the lower gamma compensates for this.

The gammas for the dye layers of film type 8443 as
analyzed by our equipment were:

Cyan V: —2.00,
Magenta y: -O.81, and
Yellow 31: —2.82.

DATA STANDARDIZATION AND SOURCES OF VARIABILITY

The use of color-infrared photography as a tool for
the quantitative monitoring of ground-scene changes
with time requires the standardization of the data
taken from the film. Certain sources of variability from
one photographic flight to another and within a roll of
film can be controlled by consistent practices and care-
ful planning. Controllable variables are camera, film,
filter combinations, flying altitude, and certain proce-
dures in the film processing. Variability in the data
from these sources can to some degree be controlled by
consistency, or if necessary, by mathematical adjust—
ments. Other sources of variability, such as atmos-
pheric changes, film—emulsion variation, and film-
processing changes and inconsistency, are less deter—
ministic in nature, and the investigator has very little
control over these sources of error. Variability in the
data from these sources are very difficult to correct
without elaborate instrumentation and calibration
and, to a large degree, must be adjusted by an empiri-
cal approach if precise calibration is not available.

Theoretically, the standardizations of photographic

036

data for such variables as atmospheric attenuation,
solar altitude, and filter transmittance should be made
as exposure adjustments for the film gamma. This was
not done for two reasons: (1) the high cost involved in
precisely measuring the energy radiated from the
ground scene and subsequently sensed by the film and
(2) the high error involved in measuring hydrologic
phenomenon such as evapotranspiration.

As pointed out previously (in the section "Photo-
graphic Data Collection”), each dye layer of the film
was viewed as a relative radiometer, and the irradi-
ance sensed in each dye layer was related to the sum of
the irradiance sensed by the three dye layers. For
clarification the reader should refer back to equations
3, 4, 5, 6, and 7.

ANALYTIC OPTICAL DENSITY

Analytic optical density is defined as the optical den-
sity within the individual dye layers of a multidye
layer film. The necessity for isolating the individual
dye-layer spectral contribution is illustrated in figure 2
by the spectral sensitivity curves for color-infrared
film. It should be borne in mind that these curves (fig.
2) are only for the film and do not show the effect of the
instrumentation (filters, lenses, densitometer, and so
forth) used for analysis. The color density measured by
the densitometer in each selected wavelength band
(fig. 7) is the sum of the radiation absorbed by all three
of the dye layers.

Several techniques are available for isolating the
exposure in each of the three dye layers. The process
used in the following analysis involves selective expo-
sure yielding a step tablet—in this case a film strip
graduated in transparency from one end to the other,
for each dye layer of the film. The relative contribution
of each dye layer in the three sensing bands was then
isolated algebraically through the use of a matrix in-
version (Evans and others, 1952, p. 441—447).

The step tablets were analyzed using the transmit-
tance densitometer (Macbeth TD 402) by recording the
three optical densities (Rod, God, Bod) transmitted at
each exposure step of each step tablet. These were then
plotted graphically with the major optical density (RC
for the cyan dye layer) versus the two minor optical
densities (G0 and BC for the cyan dye layer) for each
step tablet (the slope of the major optical density of the
dye layer taken as 1, fig. 31). The slopes of the major
versus minor optical densities for each dye layer were
then entered into a 3 X3 matrix (A). The inverse (A'l)
of the matrix (A) was then taken. (A71) was then used
for the three equations to convert the three integral
optical densities (R, G, B) to their respective analytic
optical density (C, M, Y). Matrices (A) and (A_1) and
the resulting analytic optical density equations for film

 

GILA RIVER PHREATOPHYTE PROJECT

type 8443 analyzed with our equipment are given be-
low:

RcRmRy 1.000 0.065 0.015
04),,“l3 = GchGy = 0.184 1.000 0.106 (42)
BcBmBy 0.046 0.192 1.000
and 8443 8443
C R
M =(A‘1)8443x G , (43)
Y B
8443 8443
where
1.012 —0.063 —0.008
(A‘1)8443= —0.185 1.032 —0.107 (44)
—0.011 —0.195 1021
Therefore 8443
08443 =1.012(R)—0.063(G)—0.008(B), (45)

M8443 = —O.185(R)+1.032(G)—0.107(B), and (46)
173443 = —0.011(R)—0.195(G)+1.021(B). (47)

FILM TYPE CORRECTION

During the period of study for this report, two color-
infrared films were used. Film type 8443 was used in
1967—69; after 1969, film type 2443 was used. The 2443
film has three advantages over its predecessor: (1) bet-
ter infrared color balance; (2) increased overall speed;
and (3) improved consistency from emulsion to emul-
sion. The quantitative use of the photographic data re-
quired that the data from film type 2443 be corrected to
the same spectral response as film type 8443. The
technique for accomplishing this was to premultiply
the inverse type 8443 [(A‘1)3443] by a conversion ma-
trix, [(A)g443], which was supplied by Kodak (N. L.
Fritz, written commun., 1972):

(A —1)2443 = [(14 )gfigl X [(A 71)8443], (48)
where
1.094 —0.024 —0.009
(A)::::= 0.000 0.997 0.003 (49)
0.011 —0.029 1.043

It must be emphasized that the technique illustrated
by equations 48 and 49 is not precise but is a usable
approximation.

The solution for (A ‘1)2443 is given below:

1.112 —0.092 -0.015
—0.184 1.028 —0.104
0.005 —0.240 1.065

(A ‘1)24432 (50)

2443

The analytic optical densities (C, M, Y) for film type
2443 were derived as shown below:

C R
M =(A’1)2443 X G ;
Y B

(51)

2443 2443

therefore,

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY O37

 

 

 

 

 

 

 

 
 

 

 

 

>' .60
:°IIIIIIIIIII»II
"’ z
E LU Yellow dye layer A Am
03:1 0.40 — _
-’ 0
So
; Z
“(020— Slo e=0015 - GD—
33 M8 o—O-f-CP ‘0? gr" p ‘ “ +4)
0 I 0559
Z
3 o I I I I I I I I I
0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80
MAJOR OPTICAL DENSITY, BLUE

>- 0.60
l- I
7)
E L" Magenta dye layer
0 D

.J 0.40 — __
<in:

D
9.2.
ED 020 _ Dunn” W _
a: g 0000
O O
2
g 0 I I I I I I I I I I

0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80
MAJOR OPTICAL DENSITY, GREEN
0.80

 

Cyan dye layer

0
in
o
I

 

MINOR OPTICAL DENSITY
GREEN AND BLUE
0
b
O
I

 

 

0IIII I I

0 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80

 

MAJOR OPTICAL DENSITY, RED

EXPLANATION
El
Blue minor optical density
Green minor optical density

Red minor optical density

FIGURE 31.—Major versus minor optical densities for the three dye layers of film type 8443.

C2443 = 1.112(R)—0.092(G)—0.015(B), (52)

M 2443 = ‘0'184(R)+1-028(G)—0-104(B)’ and (53) in the section “Photographic Data Collection” and by
Y2443 = 0-005(R)—0.240(G)+1.065(B)- (54) use of the equations shown in that section—equations

These analytic optical densities were then converted to 3’ 4’ 5’ 6’ and 7-

relative irradiance by the technique already discussed

038

FLIGHT‘ALTITUDE CORRECTION

The atmospheric attenuation of radiation varies as a
function of wavelength and altitude. An atmospheric-
attenuation model based on the US. Standard Atmos-
phere, 1962 (Elterman and Toolin, 1965), can be used
to determine the transmittance of the atmosphere as a
function of both wavelength and altitude. With vertical
photographs an attenuation factor may be described by
the following equation (Elterman and Toolin, 1965):
TA}; =expl—(Th1 —Th2)l, (55)
where

TM = atmospheric transmittance of a particular
wavelength, between the ground-surface
elevation and the camera altitude,

77,1 = extinction optical thickness at the camera

altitude, and

extinction optical thickness at the ground-
surface elevation; extinction optical
thickness is defined as the mean value of
the sum of the Rayleigh scattering coeffi-
cient, and the ozone absorption coefficient
at a particular altitude.

Applying this equation to the peaks of the dye-layer
sensitivity curves (fig. 2) at 0.55 ,um, at 0.65 nm, and
0.72 am, the atmospheric transmittances at 3,600,
8,500, and 60,000 ft (1,100, 2,600, and 18,000 m) for
these peaks of the sensing ranges are computed. These
are shown on table 3. The peaks were used because
they are the wavelengths where the greatest exposure
results for a given unit of radiant energy. The altitude
correction needed to convert 3,600- and 60,000-ft (1,100
and 18,000 m) flights to the standard 8,500-ft (2,600 m)
flight is then tabulated in table 3.

The values obtained from equations 6, 7, and 8 are
therefore corrected (for the photographic missions
which were flown at altitudes other than 8,500 ft
(2,600 m)) by their respective atmospheric correction
factors as follows:

=RvAc/(RVAc + GVAm + BvAy),

=GVAm/(Ii’VAC + GVAm + BVAy), and
=BvAy/(RVAc + GVAm + BvAy),

H

Th2

(56)
(57)
(58)

CU>CD>W>

where
v = subscript indicating that the relative ir-
radiance data (R, G, or B) was from a
nonstandard flight altitude or filter com-
bination and that the data are corrected to
standard values by the equation,

Ay = altitude correction factor for yellow dye
layer,

Am = altitude correction factor for magenta dye

layer, and
Ac = altitude correction factor for cyan dye layer.

 

GILA RIVER PHREATOPHYTE PROJECT

TABLE 3.—Atmospheric transmittance at different flight altitudes for
the peak wavelengths of dyeJayer sensitivity of color~infrared film

 

Correction factors for
conversion of non-
standard flight altitudes

Peak wavelength Percent of atmospheric

 

 

Dye sensitivity transmittance at flight to standard flight
layer (in micrometers) altitude altitude of 8.500 ft
3,600 it 8,500 ft 60,000 ft 3.600 ft 60,000 ft
(1,100 m) (2600 m) (18,000 m) (1.100 ml (18,000 m)
Yellow ,,,,,,,, 0.55 88.5 84.3 77.7 0.953 1.085
Magenta ,,,,,, .65 90.1 86.7 82.7 .962 l .048
Cyan" ,,,,,,, .72 90,9 87.8 84.9 .966 1.034

 

This correction as well as the corrections in the fol-
lowing sections could be improved by vectorially mul-
tiplying the atmospheric transmittance by the spectral
sensitivity at small wavelength increments (0.01 ,um)
to achieve a spectral sensitivity at a specific altitude
(US. Air Force, 1968, p. 91—96). However, this proce-
dure does not compensate for the effect of increased sky
luminance with height on aerial photography. Sky
luminance affects both the highlights and shadows
equally, thus in effect causing a lower contrast ratio
and a narrower density range (Smith and Anson, 1968,
p. 312). An expansion of the density scale (fig. 4) to
compensate for sky luminance was not necessary be-
cause the relationship between the illuminance sensed
by each dye layer was not affected significantly.

FILTER CORRECTION

The filters chosen as a standard for vegetation differ—
entiation with color-infrared film were the Wratten
numbers 12, CC20B, and CC30M (Culler, 1970). Filter
combination was chosen as a result of testing proce-
dures suggested by Fritz (1967). The testing procedure
required a series of photographs with various combina-
tions of filters and exposures. A cliff looking over a
collection of plants typical of the Gila River flood plain
was used as a platform to simulate aerial photographic
conditions. A Graflex XL camera with a 100 mm Zeiss
Tessar lens, aperature to f/3.5, shutter speed to 1/500,
and 70 mm film type 8443 was used for this photog-
raphy. Photographic testing equipment was not readily
available, and comparisons were made by Visual in-
spection. A filter combination of Wratten numbers 12,
CCZOB, and CC30M provided the best visual color reso-
lution for identifying flood-plain vegetation. This
"standard” filter pack was adapted beginning with the
1968 photography. Figure 32 shows the transmittance
curves for the W12, CCZOB, and CC30M filters and the
product curve for this combination. Optical filters obey
the Bouguer—Lambert Law, which states that the
transmittance of a series of filters is the product of the
transmittance of each filter.

The need for color correction filters to enhance the
spectral signature on color-infrared photography was
explored by Pease and Bowen (1969). Figure 2 shows

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

O39

 

 

100 I l I
W12 Filter
90 // \________./_.: “fig—“*—
G} / ./
so 5/ / -
// ' Prod uct filtration
/ /—-—--———-.—-_-_-
7o / /,/ _
CC /- I
./‘- 2o .
60 X \5 1‘]th /./ I/ _‘
so ___________/’ —
/

4O

30

FILTER TRANSMITTANCE, lN PERCENT

20

 

 

 

 

I
10 — / Principal sensing Principal sensing PrinCipal sensing .-
renge of the <__ range of the , ‘ range of the .._>
/ yellow dye layer .1 magenta dye layer 1 cyanldye layer
0
0.50 0.60 0.70 0.80 0.90

WAVELENGTH, IN MICROMETERS

Data from Kodak (1970) and N. L.
Fritz (written commun., 1971)

FIGURE 32.—Filter transmittance versus wavelength for the standard filter pack used in aerial photography in the study areas. .

lower sensitivity relative to the other layers of the cyan
dye layer of color-infrared film. The purpose of the color
correction filters used as a standard in this study is to
decrease the sensitivity of the magenta and yellow dye
layers and therefore proportionally increase the cyan
sensitivity, which has the lowest sensitivity as shown
in figure 2. In this study the two spectral ranges of
interest (because of the physiological and plant chemi-
cal relationship)——-these sensed by the cyan and the
magenta dye layer—must be accentuated to yield a
signature more useful for the determination of plant
condition. Figure 32 shows that the filter pack trans-
mits approximately 65 percent in the cyan sensing
range, 45 percent in the magenta, and 30 percent in the
sensing region of the yellow dye layer (all percentages
refer to the percent radiation transmitted which is in-
cident to the filter pack in each of the three different
spectral ranges); the filters thereby compensate for the
lower sensitivity of the cyan and increase the magenta
dye layer sensitivity with respect to the other two dye
layers. This filter pack accentuates the two dye layers
used as indices of plant vigor. This can be Visualized by
noting figure 4; the differential shift resulting from

this filter pack would diminish the yellow dye layer
image with respect to the other two.

During 1967 only a number 12 filter was used. The
two NASA flights (September 30, 1969, and September
11, 1970) included in this analysis used a number 15
only. Filter corrections were therefore necessary to
convert to the standard filter pack used during most of
the study.

The transmittances of the two color correction filters
(CCZOB, CC30M) were multiplied at 0.01 ,u.m intervals,
and these products were summed over the high sen-
sitivity wavelength range of the particular dye layer.
The mean transmittance of the filters for each dye
layer was then calculated. These filter corrections were
then used to adjust the nonstandard photographic
flights by the same method used for atmospheric at-
tenuation. No compensations were made for transmit-
tance differences between the filter numbers W12
and W15 because of similar transmittance properties
(fig. 2).

The correction equations used for the filters is illus—

trated below for the near-infrared irradiance:
R =Rch/(Rch+Gva+Bva),

 

(59)

040

where
Fy = filter correction factor for the yellow dye
layer,
Fm = filter correction factor for the magenta dye
layer, and
F c = filter correction factor for the cyan dye layer.

The actual values for the correction factors for the
specific filters cited are given below:

Fe = 0.73,
Fm = 0.50, and
Fy = 0.35.

This filter correction could be improved by the vectorial
multiplication of the filter correction and the spectral
sensitivity at small increments of wavelength as pro-
posed for atmospheric attenuation, but it was not done
because of the limitations of the study.

SOLAR ANGLE CORRECTION

The effect of solar angle ((1)) should also be considered
when interpreting color-infrared photography taken at
different times of year or day. The solar angle, the an-
gular departure from the local vertical, can be calcu-
lated at solar noon as

<I>=latitude—solar declination. (60)
The effect of solar angle on the amount of radiation
reaching a reflecting source can be computed as (El-
terman and Toolin, 1965)
T’=exp(—Th2 sec 43), (61)
where
T’ = atmospheric transmittance of radiation at a
particular wave length through the total
atmosphere and
Th2 = extinction optical thickness at a particular
wavelength through the total atmosphere
to the ground-surface elevation.

The relationships between object reflectance and
solar angle, incidence look angle, and azimuth look
angle are treated in Egbert and Ulaby (1972), while
solar angle versus exposure is explored in Sprecht,
Fritz, and Sorem (1966). The Gila River Phreatophyte
Project is at latitude 33°10’, and the noon solar angle
that occurs during summer and winter solstice (June
21, December 22) are 988° and 55.62°, respectively.

A computation of the transmittances at the peak of
each dye-layer sensing range for June 21 and De-
cember 22 are given in table 4. Table 4 shows that the
maximum possible solar angle effect on the three
wavelength regions of interest would be significant if
all other effects were constant. During the period of
plant activity (April through September) this effect is
not significant at this latitude when the photographic
data is analyzed as a ratio. Further evidence of this is
shown in figure 33, which shows that the spectral sig—

 

GILA RIVER PHREATOPHYTE PROJECT

TABLE 4.—Atmospheric transmittance for the peak sensitivity of dye
layers of colorinfrared film at summer and winter solstice

 

Percent of
atmospheric transmittance Ratio of

Peak wavelength winter solstice

 

sensitivity Winter Summer to summer
Dye layer (in micrometers) solstice solstice solstice
Yellow ,,,,,,,,,,,,,,,,,, 0.55 65.9 78.6 0.84
Magenta ,,,,,,,,,,,,,,,, .65 74.0 84.1 .88
yan ,,,,,,,,,,,,,,,,,,,, .72 78.9 87.2 .91

 

nature parameters of plant activity (R, G) are largely
independent of solar angle. Figure 33A is the bare
ground site shown in figure 1, while 33B is a part of the
dense saltcedar site in 25—2 also shown in figure 1.

The decrease of R and increase of G in figure 333 at
1200 hours results primarily from the high sun angle
which allows irradiance from the ground between the
trees to register on the photographs. The increase in R
at 1400 hours in figure 33A is more difficult to explain,
but this is perhaps due to an increase in temperature of
the soil which may cause a depletion of the surface soil
moisture and a change in signature.

STANDARD CORRECTION

Many photographic variables, such as atmospheric
changes, processing inconsistencies, and shifts in film
color balance due to aging or improper storage, are
very difficult to correct. Elaborate ground instrumen-
tation data and the use of complicated sensitometric
methods are necessary to derive such corrections.
However, a simple empirical correction for this type of
variability was attained by using the spectral signa-
ture of a highway bridge as a standard calibration. The
necessity of a standard correction is indicated by the
uncorrected signature of reach 2b during 1968 versus
1969 (fig. 30).

The basic theoretical assumptions inherent in the
bridge correction procedure are given below.

1. The bridge is assumed to be a gray surface, such
that the film (using the “standard” filter pack) regis-
ters equal quantities of energy in the three wavelength
regions sensed by the individual dye layers; the photo—
graphic mission flown on June 27, 1968, was used as
the standard photographic flight for this reason.

2. The characteristic curves (fig. 4) are superim-
posed at the analytic densities of the bridge readings,
achieving an equivalent neutral density.

The accuracy of this empirical method is considered
adequate for this study, and although it is recognized
that neither of these conditions is precisely met, the
effect of error is overshadowed by the need for a usable
inexpensive correction technique.

Densitometric readings of the highway bridge lo—
cated on the Gila River were taken on each photo-
graphic flight, conver’oedto R, G, and]? by the technique

CALCULATION OF EVAPOTRANSPIRATION USING COLOR-INFRARED PHOTOGRAPHY

 

O41

 

 

 

 

 

 

 

 

 

40 _ A. Bare ground (all standard deviation less than 0.01). —
I- I?
z —
u.| 30 --
U
I A
33 G
2 20 —— _
<6
0
z .—
< 10 —
‘3:
0
I I I I
0900 1000 1100 1200 1300 1400 1500
TIME OF DAY, IN HOURS
EXPLANATION
1%
Relative neariinfrared irradiance
6;
Relative red irradiance
100 —- B. Dense saItcedar é —
Standard deviation
I E _
’_ 90 -— I
2
Lu
0
II
Lu
n. 80 I I
E
<6
G 20 — __
Z
<
(m a
10 — ,—
» 1 i; ii
0
I I I I I I I
0900 1000 1100 1200 1300 1400 1500

TIME OF DAY, IN HOURS

FIGURE 33.—Re1ative near-infrared irradiance and relative red irradiance versus time of day for bare ground and dense saltcedar.

illustrated in equations 4, 5, and 6, and corrected for
film type, altitude, and filter combination when neces-
sary. All bridge readings were then corrected relative
to the bridge reading flown June 27, 1968, which con-
formed to the assumption (1) given above. The June 27,
1968, irradiance was
165:0, =Bs=0.33, (62)

where

s=subscript—standard bridge reading June 27,

1968, photography.

The standard correction factors for each photograph
run derived from the bridge reflectance are

Jc=Rq/Rs, (63)
JmZGq/Gs, (64)
and
J, =Bq/Bs, (65)

 

where
q =subscript—bridge reading for any photographic
mission.
The data for a specific flight was then corrected as
demonstrated by the following equation for relative
near-infrared irradiance:

R =(RUJc)/(RUJC+GUJm +BUJy). (66)

It should be realized that in this specific instance
equations 63, 64, and 65 may be simplified (J(.=3Ii’q,
Jm=3Gq, and Jy=33q), but the technique is presented
here for use if the condition of equation 62 is not met.

This technique is only usable during the summer
months (May through August) because the correction
factors tend to overcompensate the cyan layer by as
much as a factor of ten during the winter. The cause of
this high overcompensation is that the skylight con-

O42

tribution to the cyan dye layer is much lower in the
winter than in the summer, the vegetation adjacent to
the bridge becomes dormant during the winter, and the
effect of the scattered near-infrared radiance from this
source is greatly reduced, therefore reducing the quan-
tity of near-infrared irradiance received by the film
sensing the bridge.

STATISTICAL ANALYSIS OF VARIABLES

Extensive statistical analyses have been performed
on the equations presented in this report. For the use of
those interested in these evaluations, the statistical
parameters are presented in table 5.

To attain significance in a quantitative study of
spectral data requires a rigorous statistical approach.
This is of particular importance in the study of indi-
vidual plant signatures, where the two primary
parameters used (R , G) are elements of a closed system,
as shown by equation 7 in the text, R +G+B =100 per-
cent. The interrelationship between variables in a
closed system such as expressed in equation 7 yields a
spurious correlation. For example, an increase in R
dictates an equal decrease in the sum of 0 +3 . In addi-
tion, the fact that the sum of the parameters is a con-
stant destroys the independence of both the variance
and covariance and induces a bias toward a negative
correlation. Therefore, the correlation coefficient of a
closed system, such as that used in this report, may not
be related to the conventional null hypothesis, p=0.
Chayes (1960, 1971) proposed that the expected value
for a null hypothesis with this “closure restraint” may
be approximated by p=(1 —n)_1, where n is the number
of interrelated variables in the closed system. This
value can then be utilized in the Fisher Z transforma-
tion (Fisher, 1950, paper 3, p. 125) to determine the
distribution of the coefficient of correlation. A pictorial
representation of the distribution of this transforma-
tion is shown by Natrella (1963, 20—1e).

The normal sampling distribution at a particular
confidence level can be approximated by the equation
for Z given below; for ease of calculation a table of Z
values has been published in Dixon and Massey (1957,
table A—30b). The equation used to determine the con-
fidence levels of the coefficient of correlation is

Z,r ZZr: (t'XSD),
where
Z’r = the maximum and minimum values on for
the coefficient of correlation as determined
by the t’ distribution,

_ _ .117:
z, —Z[f(r)]—O.5[ln 1_r],

SD = 1/Vn—3, where n=number of samples,

 

GILA RIVER PHREATOPHYTE PROJECT

t’ = "student” If value of a particular significance
level, and
r = coefficient of correlation.
This procedure was used in this report to solve for the
rmlx and rmin by the inverse transformation,

r=Z‘1(Z’,).

Similarly, the corresponding confidence levels for the
null hypothesis were computed as

Z’p=Zpi(t ') (SD),

where
Z ’p = the maximum and minimum values of Z for
the null hypothesis as determined by the t’
distribution,

 

1
Zp=zmp>1=05 [1n< 1:; )1,

p = value of the null hypothesis,
and the values for pmax and pmin were derived by the
transformation,
P =5 ‘1(Z'p).

This was done for both the null hypothesis and the
coefficient of correlation at the 95 percent confidence
level and is presented along with the value for p and r
in table 5.

The percent confidence level of the computed r when
related to the null hypothesis was computed by

Zp=Z,——(t’)(SD),

which is then solved for t’,

t,_ zp—z,
‘ SD

 

The confidence level for the t’ for each computed corre-
lation coefficient as related to the null hypothesis is
also given in table 5.

The statistical parameters shown in table 5 which
have not been discussed were evaluated from the defi-
nitions given in Ezekiel and Fox (1963).

SUMMARY

In the 5—year period 1967—71, 38 color-infrared
photographic missions were flown over the Gila River
Phreatophyte Project in southeastern Arizona. Data
from these missions were analyzed to determine the
possibility of identifying and measuring vegetative
parameters and their associated hydrologic variables
by spectral analysis of the photography. During the
summer of 1971, additional data from six color-
infrared photographic missions flown over the Cibecue
Ridge Watershed Study in central Arizona were used
to test the validity of some of the techniques developed
in this study.

A transmittance densitometer was used to obtain a

043

.Nm 3 Fm 89¢ flaw—HE? vm ”555:5 8:. mmofi 23? ea 8 men Sod emu—8:0 mm :ofimaew Sm 3:on coEEmm me 525:: 2: .mwmbwnm 98 v2.8a m5 wcidu mEumm—u 3 mats?

 

 

2 ON. «a. S. E. 3. mm. 2. i E. E E E. S cm. mm. mm om. om. ““““““““““““““ SE
@5985 23 mo 3:0 mime—Saw
v.5 mA 9N Wm m.m Wm mm «m vm >6 m6 m6 es ms «1m me. me Q: mew ““““““““““““““ 8:36qu
vamESmm as She unavnmum
mm. wa mm \ whf :f mmf mmf ewf emf mhf mbf mhf mhf mbf mm: mm mm. mm mm. ......... Ema—399E 25¢
.wa Se. .9. .ve. .me .eH. .mm. «5.! .He. .mfil .NHf .mfif .me .2f .3 1 .3 f .mnf .mm 1 .mmf mo £32 853:8 €523 mm
ee. emf emf emf emf emf emf emf em I emf emf emf emf emf ee. ee. ee. ee. ee. l 83:8an :5: wfi no 2:5»
mm. mm. mm. mm. em. mm. em. mm. mm. mm. mm. mm. mm. mm. em. mm. mm. mm. mm. “““““““ mwmwficnms :a: 23
0a mumwemmwh EMF—3 :Omudmvhhoo
mo acmMuEwg («a ~w>£ gamut—EU
Hm wmf mmf bmf mwf wmf wmf mmf emf Hmf mmf mmf mmf emf emf mm. hm. mm. mm. 11 5:38:83 EwCEuS
.mb .vwf .mmf .ewf .vvf .ewf .emf Hmf .whf .eef .wef .mbf .whf .wa .mwf .mm. .mm. .mm. .ve. :8 £33 85238 H.533 mm
mm? odw 33 9% EN 3; mi 333 gm Ev man 35 gm wdfi 3m 33 NE mom Em ffffffffffffffffffffff no:
.Ewtou we $85525 we amen. k
Nb. 1-1 yr; ‘1‘ 5“ f--- 11 I? 1-: ‘1‘ \l :ff 1;. 1} mv. mm. E «R. Em I‘mflomaaEEkumveo unmmoEwoO
mm. vmf mmf Nmf wef vmf emf mmf Sf Bf emf emf mmf Nmf he.) mm. we. mm. mm. 11111 nosflwbou mo «56530
mml mom msmfi Hm b6 mefi w.» «NE e42 fim 0: QB 93 PS m.m eHef :ef vN mm. 11 38.33:“. 33 L unwedum
Tm. wA m. «mm NH We 0: mp. fie be 5m 5% e.m me 9: m4: Q: mwv «Am l £5885? 55¢va Ewe—Ea
m.mNI mem emefl «.mv Ymm Heb m.ww Nem me bee mew Q: mmm m.m> «.3 if mmf ”we” mew 11 3.: :Bmmvnmfi mo 38»an
fl: m.w\ vefil 1m! Wml emf Tm! @va >61 Sew NBS! Em! Hm! 12‘ fiml mm m.m w.m mm ffffffff mac? mo 23 _“ anmcfim
3. ea. 3. ve. we. be. mm. Ne. ma. 3. 2. me. 3. me. E. weee. Nee. Hm. mu. 1-- 33m :8 5.325 Ewanfim
«m4 Nmf emNI wmf Nvf $f mmf Neg! HeAI mbf wa mwf Sf Hmf vef mwee. whee. we. we. yyyyyyyy 95 56mm 3 mo mac—m
meow «.mfif meal m.$f QB! fiZHI m.mmf hmwwf ememl w.«m~l mhif eAbmf ESNI «Maw w.mm| 9v «N Neewfifi NdFfi ................ wonmigoo
mdm Nd mefi mm ms m.w Sm emm «Am e mg 5 NH v.5 mgfi m.mH es 3. 3. e62 v 3 ................. final:
Emmimnmv .«o nodumgmu flaw—Bum
bdweé ”3 mm: New whm New hem wane eemv h mm: m 5H eAem @me N 5mm 93 me. me. :3de fimvmd léfiwfig Ewe—heave «o magma;
Qwe v.3 m.F mmm mime. mAm men wimp m.mm m mm m mm e.wm «hm hem Q: mm. mm. 5 bmm mmmfi -flemig engage mo :32
NE way :u «.3 em” NS Wm mew fima .9: v m: v.5 WE new ms new Q: wmmfi Nme .............. Emmi? ”Emu
«flammvflm .wo £2551,me UhNfl—nflum
wmmm WE @me we? méE ed: wsm «meme Even m mHN v m: Nmem fiewm 12¢ fimm eAmm ms?” eANeHH fivav ffffffffffffffffffffff 033
35> uqmunwnmefi no mun—«tax»
w.me mg: mg“: 1mm edm dew wév fimm new wAm «.mv fiwv e.mv esv fimm emu mme véww veew 123$“: Eocnamui «o :82
em 9 ofi ma em 2 m m w mm wH Eh— mrm. hwh Va m m bub th fffffffffffffffffffff ENE
.59 $53 mEEmm .«o 53552
H 11 ‘1‘ e e m e E 2 Va mm mm mm mm H N. n ma 3 .................... REES
mEME uEAEMEOaE .«o HEESZ
mm vm mm mm 5 on mm mm hm mm mm vm mm mm Hm em 3 Na 3 EwquEam
:oflusvm :oEmScW dosaamm nofiaacm :23st 53¢:UH nofiasmum aofimnvm :Eamzem :ofianmm :23er

Elana—am HESS—Um nofiwdum

:Eumflam cogs—Gm Egan—Um :ofiwzrm nomawflvm

 

tom? «3“ S New»: «$3333 5 $33.23: Kc 8&5.» BREESWIh and?

044

14 percent sample of the study area. Optical density
readings in each of three primary colors of the positive
transparencies were taken. The irradiance (defined in
this report as energy recorded on the film) sensed from
the dye concentration in each of these primary colors
was determined and related to the total energy sensed
by the film, achieving three parameters of relative ir-
radiance (near-infrared, red, and green) which were
functionally related to spectral regions indicative of
plant activity. These parameters were then corrected
to a "standard photo flight” by adjusting the den-
sitometric data for flight altitude, filter combination,
film type, and a standard correction based on the
spectral signature of a highway bridge located in the
project area.

Remote sensing calculations of evapotranspiration
from 13 photographic missions flown during 1968 were
related to water-budget measurements for a 1,700—acre
(689-ha) area cleared of vegetation and a 2,200-acre
(891-ha) phreatophyte-covered area of the Gila River
flood plain. The coefficients of correlation between the
water-budget measurements and the remote sensing
calculations were 0.88 for the cleared area and 0.86 for
the phreatophyte-covered area. Evapotranspiration
calculated from seven photographic missions flown
during 1968 over a spatially homogeneous grain sor-
ghum field gave a coefficient of correlation of 0.93 when
related to evapotranspiration computed by the
Blaney-Criddle equation.

Determining depth to the ground-water level spec—
trally was not practical, although depths to water of
less than 30 ft (9 m) had a standard error of 5.4 ft
(1.6 m). Moisture content of the soil could not be de-
termined spectrally, but plant stress due to moisture
content deficiencies was detectable. Plant-type dis—
crimination was possible between eight different plant
communities, which included both deciduous and pe—
rennial species. A spectral evaluation of near-infrared
versus red irradiance of these plant communities
showed that each had a standard error of less than 10
percent. Computer maps of the spectral signature dis—
tribution for each photographic mission flown of the
Gila River flood plain were found useful for spatial and
temporal evaluation of vegetation. The spectral signa-
tures for different types of vegetation were harmoni-
cally modeled and used as a means of recognizing sea-
sonal trends and anomalies in the vegetation. When an
increase in volume of canopy for both mesquite and
saltcedar was noted, there was a corresponding in-
crease of near-infrared irradiance, but the measure-
ment errors of both irradiance and canopy volume were
very large. However, foliar cover versus irradiance had
a high coefficient of correlation, 0.85.

During this study it was determined that a color-

 

GILA RIVER PHREATOPHYTE PROJECT

infrared photographic mission and a computer analysis
of the photographic data for the Gila River
Phreatophyte Project area cost about a tenth of the
amount of conventional species classification and
canopy-measurement techniques ($650 versus $6,600).
A short discussion of the derived spectral equations
and a table of 24 statistical parameters describing the
spectral and hydrologic variables is included to aid the
reader in evaluating the significance of the study.

REFERENCES CITED

Blaney, H. F., and Criddle, W. D., 1962, Determining consumptive
use and irrigation water requirements: U.S. Agr. Research Serv-
ice Tech. Bull. 1275, 59 p.

Chayes, Felix, 1960, On correlation between variables of constant
sum: Jour. Geophys. Research, v. 65, no. 12, p. 4185—4193.
1971, Ratio correlation—a manual for students of petrology
and geochemistry: Chicago, 111., Univ. Chicago Press, 99 p.
Cruff, R. W., and Thompson, T. H., 1967, A comparison of methods of
estimating potential evapotranspiration from climatological
data in arid and subhumid environments: U.S. Survey Water-

Supply Paper 1839—M, 28 p.

Culler, R. C., 1970, Application of infrared color photography to the
description of flood plain vegetation, in Workshop on aerial color
photography in the plant sciences: Gainesville, Florida Dept.
Agriculture, Div. Plant Industry, p. 159—164.

Culler, R. C., Jones, J. E., and Turner, R. M., 1972, Quantitative
relationship between reflectance and transpiration of
phreatophytes—Gila River Test Site: Fourth Ann. Earth Re-
sources Program Review, Natl. Aeronautics and Space Adm.,
v. 3, chap. 83, p. 1—9.

Culler, R. C., and others, 1970, Objectives, methods, and
environment—Gila River Phreatophyte Project, Graham
County, Arizona: US. Geol. Survey Prof. Paper 655—A, 25 p.

Denny, C. H., Morrison, E. L., Jr., Worthman, C. D., and Lucht,
D. D., 1971, Automated processing of forest imagery, in Applied
remote sensing of earth resources in Arizona: Univ. Arizona,
Proc. Arizona Regional Ecological Test Site Symposium, 2d,
p. 111—125.

Dixon, W. J., and Massey, F. J., Jr., 1957, Introduction to statistical
analysis: New York, McGraw-Hill Book Co., 488 p.

Egbert, D. D., and Ulaby, F. T., 1972, Effect of angles on reflectivity:
American Soc. Photogrammetry Jour., v. 38, no. 6, p. 556—564.

Elterman, Louis, and Toolin, R. B., 1965, Atmospheric optics, in Val-
ley, S. L., ed. Handbook of geophysics and space environment:
Cambridge, Mass, US. Air Force, Office Aerospace Research,
Cambridge Research Labs, chap. 7, p. 1—32.

Erie, L. J., French, 0. F., and Harris, Karl, 1965, Consumptive use of
water by crops in Arizona: Univ. Arizona, Agr. Expt. Sta. Tech.
Bull. 169, 44 p.

Evans, R. M., Hanson, W. T., Jr., and Brewer, W. L., 1952, Principles
of color photography: New York, John Wiley and Sons, Inc.,

‘ 709 p.

Ezekiel, Mordecai, and Fox, K. A., 1963, Methods of correlation and
regression analysis: New York, John Wiley and Sons, Inc., 548 p.

Fisher, R. A., 1950, Contributions to mathematical statistics: New
York, John Wiley and Sons, Inc, 43 papers.

Fritz, N. L., 1967, Optimum methods for using infrared-sensitive
color film, in Workshop infrared color photography in the plant
sciences: Winter Haven, Florida Dept. Agriculture, Div. Plant
Industry, pt. 3, p, 1—19.

Gates, D. M., 1970, Physical and physiological properties of plants, in

 

CALCULATION OF EVAPOTRANSPIRATION USING COLOR—INFRARED PHOTOGRAPHY

Remote sensing with special reference to agriculture and fores-
try, by Committee on Remote Sensing for Aricultural Purposes:
National Acad. Sci., p. 224—252.

Griffith, S. V., and Howe, R. H. L., 1960, Photo interpretation in
hydrology and watershed management, in Colwell, R. N., ed.:
Manual of photographic interpretation: American Soc. Photo—
grammetry, p. 539—560.

Hanson, R. L., Kipple, F. P., and Culler, R. C., 1972, Changing the
consumptive use of a flood plain, in Age of changing priorities for
land and water: Am. Soc. Civil Engineers, Irrigation and Drain-
age Div. Specialty Conf., p. 309—330.

' Hunter, G. T., and Bird, S. J. G., 1970, Critical terrain analysis: Am.

Soc. Photogrammetry Jour., v. 36, no. 9, p. 939—952.

Kittredge, Joseph, 1948, Forest influences: New York, McGraw-Hill
Book Co., 394 p.

Knipling, E. B., 1969, Leaf reflectance and image formation on color
infrared film, in Johnson P. L., ed. Remote sensing in ecology,:
Athens, Univ. Georgia Press, p. 17—29.

1970, Physical and physiological basis for the reflectance of
visible and near‘infrared radiation from vegetation: Remote
Sensing Environment Jour., v. 1, no. 3, p. 155—159.

Kodak, 1970, Kodak filters for scientific and technical uses: Roches-
ter, New York, Eastman Kodak Company Pub. B3, 88 p.

Langbein, W. B., and Iseri, K. T., 1960, General introduction and
hydrologic definitions: U.S. Geol. Survey Water-Supply Paper
1541—A, 29 p.

McDonald, C. C., and Hughes, G. H., 1968, Studies of consumptive
use of water by phreatophytes and hydrophytes near Yuma,
Arizona: U.S. Geol. Survey Prof. Paper 486—F, 24 p.

Molchanov, A. A., 1960, Gidrologicheskaya rol’esa, Izdatel’stvo
Akademii Nauk, SSSR, Moskva, translated from the Russian by
Prof. A. Gourevitch (1963): Washington, D. C., U.S. Dept. Com-
merce, Office Tech. Services, 407 p.

Natrella, M. G., 1963, Experimental statistics: U.S. National Bur.
Standards Handb. 91, 504 p.

 

 

O45

Pease, R. W., and Bowen, L. W., 1969, Making color infrared film a
more effective high-altitude remote sensor: Remote Sensin En-
vironment Jour., V. 1, no. 1, p. 23—30.

Reifsynder, W. E., and Lull, H. W., 1965, Radiant energy in relation
to forests: U.S. Dept. Agriculture Tech. Bull. 1344, 111 p.
Shreve, Forrest, 1964, Vegetation of the Sonoran Desert, in Forrest
Shreve and Ira Wiggins, Vegetation and flora of the Sonoran

Desert: Stanford Univ. Press, Calif. 2 v.

Smith, J. T., and Anson, Abraham, eds., 1968, Manual of color aerial
photography: Am. Soc. Photogrammetry, 550 p.

Sprecht, M. R., Fritz, N. L., and Sorem, A. L., 1966, The change of
aerial camera exposure with solar altitude: Photographic Sci.
and Eng, v. 10, no. 3, p. 150—155.

Thomas, J. R., Weigand, C. L., Myers, V. 1., 1967, Reflectance of cot-
ton leaves and its relation to yield: Jour. Agronomy, v.59,
p. 551—554.

Todd, H. N., and Zakia, R. D., 1969, Photographic sensitometry—the
study of tone, reproduction: New York, Morgan and Morgan,
Inc., 312 p.

Turner, R. M., 1971, Measurement of spatial and temporal changes
in vegetation from color-IR film: Am. Soc. Photogrammetry
Proc., San Francisco, ASP Fall Convention, p. 426—441.

1973, Quantitative and historical evidence of vegetation
changes along the upper Gila River, Arizona: U.S. Geol. Survey
Prof. Paper 655—H, 20 p.

Veihmeyer, F. J., 1964, Evapotranspiration, in Chow, V. T., ed.,
Handbook of applied hydrology: New York, McGraw-Hill Book
Co., chap. 11, p. 1—38.

Wright, W. D., 1969, The measurement of colour: New York, Van
Nostrand Reinhold Company, 340 p.

U.S. Air Force Avionics Laboratory, 1968, Color tone
reproduction—Part 1, Theory Manual: U.S. Air Force Avionics
Laboratory Tech. Rept. AFAL—TR—67—164, 65 p.

[U.S.] National Weather Service, issued annually, Climatological
data, Arizona: U.S. Dept. Commerce.

 

flu. s. GPO: 1977—791- 786

 

 

 

 

 

 

 

 

 

Evapotranspiration Before and
After Clearing Phreatophytes,
Gila River Flood Plain,

Graham County, Arizona

By R. C. CULLER, R. L. HANSON, R. l‘vl. lVIYRICK, R. M. TURNER, and F. P. KIPPLE

GILA RIVER PHREATOPHYTE PROJECT

 

GEOLOGICAL SURVEY PROFESSIONAL PAPER 655—P

 

 

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1982

 

 

IJNITEI)STHXTESl)EPAJ(TthbWT()F'T}U£IbYFERJCHK
JAMES G. WATT, Secretary

GECHJ)GJCAJ.SLHIVEY
Dallas L. Peck, Director

Library of Congress Cataloging in Publication Data

Evapotranspiration before and after clearing phreatophytes,
Gila River Flood Plain, Graham COunty, Arizona

(Geological Survey Professional Paper 655—?)

Bibliography: p. PSO-PSI.

Supt. of Docs. no.: I 19.16z655-P

1. Evapotranspiration——Arizona——Gila River watershed——
Measurement. 2. Phreatophytes——Arizona——Gila River
watershed. 3. Water balance (Hydrology)--Arizona——
Gila River watershed. I. Culler, R. C. (Richard Carlton),

1912— . II. United States. Geological Survey.

[11. Series.

QC915.7.U5E9 551.57’2 81-607801
AACR2

 

For sale by the Distribution Branch, US. Geological Survey,
604 South Pickett Street, Alexandria, VA 22304

 

PREFACE

This report is the last of the US. Geological Survey
Professional Paper 655 Series describing the hydro-
logic and environmental studies associated with the
Gila River Phreatophyte Project. The following is a
list of the publications originating from this project:

SURVEY PUBLICATIONS

Culler, R. C., and others, 1970, Objectives, methods, and environ-
ment~Gila River Phreatophyte Project, Graham County,
Arizona: US. Geological Survey Professional Paper 655—A,
p. A1—A25.

Burkham, D. E., 1970, Precipitation, streamflow, and major floods
at selected sites in the Gila River drainage basin above
Coolidge Dam, Arizona: US. Geological Survey Professional
Paper 655—B, p. B1—B33.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis of
streamflow data for an alluvial stream: US. Geological Survey
Professional Paper 655-C, p. C1—C13.

Weist, W. G., Jr., 1971, Geology and ground-water system, Gila
River Phreatophyte Project: US. Geological Survey Profes-
sional Paper 655—D, p. D1—D22.

McQueen, I. S., and Miller, R. F., 1972, Soil-moisture and energy
relationships associated with riparian vegetation near San
Carlos, Arizona: US. Geological Survey Professional Paper
655—E, p. E1—E51.

Hanson, R. L., 1972, Subsurface hydraulics in the area of the Gila
River Phreatophyte Project: US. Geological Survey Profes-
sional Paper 655—F, p. F1—F27.

Burkham, D. E., 1972, Channel changes of the Gila River in
Safford Valley, Arizona, 1846—1970: US. Geological Survey
Professional Paper 655—G, p. G1—G24.

Turner, R. M., 1974, Quantitative and historical evidence of vegeta-
tion changes along the upper Gila River, Arizona: US. Geologi-
cal Survey Professional Paper 655—H, p. H1—H20.

Burkham, D. E., 1976, Flow from small watersheds adjacent to the
study reach of the Gila River Phreatophyte Project, Arizona:
US. Geological Survey Professional Paper 655—I, p. 11—119.

1976, Hydraulic effects of changes in bottomland vegetation

on three major floods, Gila River in southeastern Arizona: US.

Geological Survey Professional Paper 655-J, p. J1—J14.

1976, Effects of changes in an alluvial channel on the
timing, magnitude, and transformation of flood waves, south-
eastern Arizona: U.S. Geological Survey Professional Paper
655—K, p. K1—K25.

Hanson, R. L., and Dawdy, D. R., 1976, Accuracy of evapotranspir-
ation rates determined by the water-budget method, Gila River
flood plain, southeastern Arizona: US. Geological Survey
Professional Paper 655—L, p. L1—L35.

Laney, R. L., 1977, Effects of phreatophyte eradication on quality
of water: US. Geological Survey Professional Paper 655—M,
p. M1—M23.

 

 

Kipple, F. P., 1977, The hydrologic history of the San Carlos I
Reservoir, Arizona, 1929-1971: US. Geological Survey Profes- :

sional Paper 655—N, N1—N40.

 

Jones, J. E., 1977, Calculation of evapotranspiration using color-
infrared photography: US. Geological Survey Professional
Paper 655—O, p. 01—045.

Leppanen, O. E., 1980, Barren area evapotranspiration estimates
generated from energy budget measurements in the Gila River
Valley of Arizona: US. Geological Survey Open-File Report
80-1003, 31 p.

1981, Evapotranspiration from rapidly growing young salt-

cedar in the Gila River Valley of Arizona: US. Geological

Survey Open-File Report 81—485, 26 p.

1981, Evapotranspiration from forage grass replacing native
vegetation in the Gila River Valley, Arizona: US. Geological
Survey Open-File Report 81—1018, 38 p.

Park, D. M., Culler, R. C., and Turner, R. M., 1978, Management of
flood-plain vegetation, 1967 to 1972, San Carlos Indian Reser-
vation, Arizona: US. Geological Survey Open-file Report
78-412, 21 p.

Culler, R. C., Hanson, R. L., Myrick, R. M., Turner, R. M., and
Kipple, F. P., 1982, Evapotranspiration before and after clear-
ing phreatophytes, Gila River flood plain, Graham County,
Arizona: US Geological Survey Professional Paper 655—P
(this paper).

 

 

BASIC DATA COMPILATION

Myrick, R. M., Kipple, F. P., Culler, R. C., and Hanson, R. L., 1975,
Basic hydrologic data for the Gila River Phreatophyte Project,
Graham County, Arizona: US. Geological Survey Basic-data
compilation, 13 p. and 400 p. of tables. Available for inspection
at the Water Resources Division, US Geological Survey,
Tucson, Arizona. Data are also available on magnetic tape in
WATSTORE, US. Geological Survey, Reston, Va.

JOURNAL ARTICLES

Culler, R. C., 1965, The Gila River Phreatophyte Project: Arizona
State Land Department Ninth Annual Arizona Watershed
Symposium Proceedings, p. 33—38.

1970, Water conservation by removal of phreatophytes:
EOS Trans. American Geophysical Union, v. 51, no. 10, p.
684—689.

Culler, R. C., Hanson, R. L., and Jones, J. E., 1976, Relation of
consumptive use coefficient to the description of vegetation:
Water Resources Research, v. 12, no. 1, p. 40—46.

Hanson, R. L., Kipple, F. P., and Culler, R. C., 1972, Changing the
consumptive use on the Gila River flood plain, southeastern
Arizona, in Age of Changing Priorities for Land and Water:
Proceedings American Society Civil Engineers, Irrigation and
Drainage Division Specialties Conference, September 1972,
p. 309—330.

Hanson, R. L., 1973, Evaluating the reliability of specific-yield
determinations: US. Geological Survey Journal Research,
v. 1, no. 3, p. 371—376.

Warren, D. K., and Turner, R. M., 1975, Saltcedar (Tamarix
chinensis) seed production, seedling establishment, and re—
sponse to inundation: Arizona Academy Science Journal.
v. 10, no. 3, p. 135—144.

 

III

CONTENTS

  
 
 

  
 
 

 

 

   

 

 

  

 
  

    

Page Page
Conversion factors ............................................................................... VI Collection and conversion of basic data—Continued
Scientific names and common equivalents .................................... VII Basin-fill inflow ______________________________________________________________________________ P16
Equations ________________________________________________________ . VII Ground-water movement downvalley ______________________________________ 17
Symbols _________________________________________________________________ . VIII Ground-water movement crossvalley ______________________________________ 18
Abstract _________________________________________________________________ P1 . Evaluating evapotranspiration ________________________________________________________ 18
Introduction ............................................................................................ 2 Relative significance of the water-budget components ______ 18
Personnel and acknowledgments ............................................ 3 Measurement errors in evapotranspiration ________________ _ 24
Description of the study area ............................................................ 3 Criteria for rejecting measured ET values ___________________________ 26
Geology ............................................................................................ 5 Use of ET data for defining a prediction equation ______________ 27
Climate 6 Evaluation of the derived evapotranspiration
Vegetation ............................................................... 7 equations ...................................................................................... 32
Surface water ......................................................... 8 A comparison of Blaney-Criddle method
Ground water .................................................................................. 9 with other methods .................................... . 37
Man’s influence .............................................................................. 9 Comparison of results with other studies 41
Method of analysis ................................................................................ 10 Effects of phreatophyte clearing on ground-water
Collection and conversion of basic data ........................................ 11 levels and seepage measurements ........................................ 48
Gila River streamflow ....................... 11 Ground-water levels ..................................................................... 48
Change in channel storage ........................................................ 11 Seepage measurements ................................................................ 48
Tributary inflow ............................................................................ 12 Conclusions ................................ 49
Precipitation .................................................................................... 14 > References cited ____________________________________________________________________________________ 50
Soil moisture .................................................................................... 14
ILLUSTRATIONS
Page
PLATE 1. Vegetation map of the Gila River Phreatophyte Project ............................................................................................................. In pocket
FIGURE 1. Map of project area ............................................................................................................................................................................................ P4
2. Geologic section across the study area .................................................................................................................... 5
3. Map showing grid network and the system for numbering quadrangles used in describing vegetation cover ._ 8
4. Photograph showing plugging of channel on section 23 ...................................................................................................................... 12
5. Graph showing monthly discharge of the Gila River at Calva ............................................................................................................ 13
6. Map showing location of ground-water observation wells .................................................................................................................... 15
7-26. Graphs showing:
7. Measured evapotranspiration for each budget period during 1963—71 and monthly pan evaporation at San
Carlos Reservoir .............................................................................................................................................................................. 19
8. Sources of water available for E Tin reach 2 during each budget period of 1965 ............................................................ 20
9. Average contribution to ET per month of precipitation, surface-water inflow, soil moisture in storage, and
ground-water inflow for the preclearing period and the postclearing period of 1965 ................................................ 22
10. Average monthly ET before and after clearing based on selected ET data from all reaches ...................................... 23
11. San Carlos pan evaporation, evapotranspiration per budget period, and measurement error in evapotranspira-
tion for reach 2 during calendar years 1965 and 1970 ........................................................................................................ 25
12. Boundaries within which the adjusted ET values (ET’) were accepted as reliable estimates ...................................... 27
13. Relation between monthly and average monthly values of kp ............................................................................................ 31
14. Mean and standard deviation of monthly measured ET’ and climatic factor f .............................................................. 33
15. Monthly variability in climatic factors ........................................................................................................................................ 38
16. Seasonal variability in average monthly values of the coefficient kp ______________________ 39
17‘ Field estimates of seasonal variability of foliation .......................................................................................................... 39
18. Seasonal variation of adjusted red transmittances obtained from Ektachrome—IR images of a saltcedar forest,
Gila River, Ariz. .............................................................................................................................................................................. 40
19. Seasonal variability of f kp, the difference in U between an area having a 100 percent areal density of
phreatophytes and that from an area of no phreatophytes .............................................................................................. 41
20. Evapotranspiration from Cottonwood Wash ............................................................................................................................ 43
‘21. Values of k for Cottonwood Wash ................................................................................ 43
22. Relation of measured to computed evapotranspiration for reaches in Safford Valley 1943—44 ......................... 44
23. Relation of measured to computed evapotranspiration for evapotranspirometers at Buckeye, Aria, and
Bernardo, N. Mex. , ........................................................................................................................................................................... 45
24. Comparison of water-table elevations before and after clearing phreatophytes ............................................................. 47
25. Results of simultaneous discharge measurements of the Gila River channel .................................................................. 48

V

VI

TABLE 1.
. Number of 1-acre plots in Reach 3 characterized by combinations of species (mesquite and saltcedar), canopy-cover

b3

mommy»:

12.

13.

14.

15.

16.
17.

18.
19.

20.
21.

. Monthly percentage of daytime hours of the year for latitudes 24° to 500 north of equator._
. Summary of vegetation survey ......................................................................................................................................................................
10.
11.

CONTENTS
TABLES

Climatological data for San Carlos Reservoir ..........................................................................................................................................

 

class, and height class ............................................................................................
Area from which vegetation was cleared on project reaches during 7 years ....................................................................................

. Annual and extreme discharges of the Gila River, at Calva, cross-section 9, during the study period ....................................
. Precipitation for reaches 1 and 2 by water year ......................................................................................... ; ............................................
. Water-budget components, resulting ET, and total measurement error for each budget period during water years

1963—71, reaches 1, 2, 2a, 3 ......................................................................................................................................................................

. Number of ET values measured in each reach and the number rejected as outliers or because of a large measurement

error ................................................................................................................................................................................................................

 

Application of vegetation description to empirical equations ..............................................................................................................
Number of accepted budget period E T’ data for each month as related to the status of clearing on each reach and the
numerical vegetation descriptors ..........................................................................................................................................................
Summary of monthly and annual k0 and kp coefficients in equation 13 derived from equations 12 and 14 where f =
pt/lOO and x = 0.75 ........................................................................................................................... : ........................................................
Average monthly and average annual U rates for each reach before and after clearing phreatophytes, computed from
equation 12 using the monthly f and average monthly kp values given in table 12 ..............................................................
Variability of measured E T’, climatic factors, coefficients. and computed U for the month of June for preclearing and
partial clearing ..........................................................................................................................................................................................
Average monthly measured (E T’) and computed (U) evapotranspiration for all reaches and the standard deviation (SA)
of the differences E T’ - U ........................................................................................................................................................................
Variability in U due to fitting coefficients to data from different periods and areas .....
Summary of monthly and annual 120 and kp coefficients in equation 13 derived from three expressions for the climatic
factor f in equation 12 ..............................................................................................................................................................................
Annual evapotranspiration computed for each reach by the Blaney~Criddle method (BC), solar radiation (R), Jensen
and Haise (JH), and potential evapotranspiration (PE) methods ..............................................................................................
Derivation of consumptiveuse coefficients for Cottonwood Wash during growing season, March through October ________
Application of vegetation descriptions from Gatewood and others (1950, tables 7 and 8) to empirical equations ................
Comparison of evapotranspiration computed by empirical equations with the measured draft on ground water
presented by Gatewood and others (1950, table 58) ........................................................................................................................

 

CONVERSION FACTORS

 

Multiply English unit b—y To obtain metric (SI) unit
Inches (in.) 25.4 Millimeters (mm)
Feet (ft) .3048 Meters (m)
Miles (mi) 1.609 Kilometers (km)
Square feet (ftz) .0929 Square meters (m2)
Acres .4047 Hectares (ha)
Square miles (miz) 2.590 Square kilometers (kmz)
Acrefeet (acre-ft) 1233 Cubic meters (m3)
Acrefeet (acreft) 1.233 X 10'3 Cubic hectometers (ha3)
Cubic feet per second .02832 Cubic meters per second

(ft3/s) (ml/s)

Acrefeet per square 4.761 X 10" Cubic hectometers per

mile (acre-ft/miz)

square kilometer
(ha3 /m2)

Page
P6

7

8
12
14
52
27
28
29
30
30
30
32
35

35
36

38
40
42
42

44

CONTENTS VII
SCIENTIFIC NAMES AND COMMON EQUIVALENTS

Common name Scientific name
arrowweed Pluchea sericea (N utt.) Coville
baccharis Baccharis species
Bermuda grass Cynodon dactylon (L.) Pers.
catclaw Acacia greggii Gray
cottonwood Populus fremontii Wats.
creosotebush Larrea tridentata (DC.) Coville
mesquite Prosopis juliflora (Swartz) DC.
red willow Salix laevigata Bebb.
Russian olive Elaeagnus angustifolia L.
saltcedar Tamarix chinensis Lour.
saltgrass Distichlis stricta (Torn) Rydb.
seepweed Sueda torreyana Wats.
seepwillow Baccharis glutinosa Pers.
whitethorn Acacia constricta Benth.
willow Salix species
EQUATIONS

Equation No.

(1) ET:QI;Q0+QT+AC+F+_AMS+AMI (13) k:ka+ka
+AMC+ GB+ G,- Go+AMn~

 

4
_ n n (14) V: X Al [Cl/100 + (Cl/100V ]/ 2
(2) P = E AJPI‘ 2 Aj U21
j=l j=1
321
_ n n <15) Am”, = 2 [(ET,- U,)] / 321
(3) AMzt : 2 (AsztAj) / 2' A] t:
j=1 1:1
ET’ - k”
(4) AMZUZszII'll_M2/l (16) kp: 7‘;—
(5) A111: — A71 S’A
. N N 2 [/2
6 :
U G ‘TWD 2 A12 (2 A) /N
v ‘ 2 1 z z _ 2 _ i=1 i=1
(7) 6m; = [613, + 6-0“ + 5 QT+6 Ar + 6 17+ 6 AMs + 6 AM, (17) gA : N 1
+ 6331+ slim] 1/2
(8) (:I-f’l'h : [620“ + 620', + 620“] 1/3 12 1/2
(18) §A = 2 8.x";
(9) em = [6251; + 551%] 1/3 "1:1
(10) PET 2 (0.014t - 0.37)]? (19) k : k0 + kchcw

(11) ET’ = ET — 13 — Am
(12) U = f k <20) k = kn + kp V + lei-“Am

 

 

VIII

{bib}.

ran

“1

BC

ET
ET’

AMI

AMC

CONTENTS,

SYMBOLS

 

Area of reach or surface area assigned to a sample point.

Area assigned to rain gagej or to soil-moisture hole 1'.

Fraction of total area having a canopy cover falling in
density class 1).

Intercept of the relation defined in unnumbered equation
page P38.

Slope of the relation defined in unnumbered equation
page P38.

Subscript denoting “basin fill.”

Subscript for f, k, k0, kp, and U denoting that the Blaney-
Criddle PET equation was used to determine 1' in
equation 12.

Bias error.

Subscript denoting “capillary zone of soil profile.”

Change in Gila River channel storage.

Average cover density in class 11.

Subscript for consumptive use coefficient for cottonwood
and willow, in equations 19 and 20.

Total number of days in a budget period.

Difference between measured and computed ET for a
given budget period, equation 17.

Evapotranspiration measured by the water budget.

Evapotranspiration measured by the water budget with
P and 11713 removed.

Climatic factor.

D-ownvalle-y ground-water flow through the alluvium.

Inflow of ground water from “basin fill.”

Downvalley ground-water inflow.

Downvalley ground-water outflow.

Average ground-water level change.

Downvalley ground-water slope (ft/ft).

Subscript denoting “inflow” or “intermediate zone of soil
profile.”

Subscript forf, k, k0, kp, and U denoting that the Jensen-
Haise PET equation was used to determine f in equa-
tion 12.

Sample point—ie, precipitation gage or soil-moisture
access hole.

Consumptive use coefficient which is dependent on the
kind and quantity of vegetation.

Consumptive use coefficient for no phreatophyte cover.

Increase in the consumptive use coefficient for phreato-
phyte cover, dominated by saltcedar.

Subscript denoting month in equation 18.

Moisture content of soil.

Average change in moisture content in the unsaturated
soil zone immediately below the land surface.

Ave-rage change in moisture content in the unsaturated
intermediate zone located between the overlying soil
zone and the underlying capillary zone.

Average change in moisture content in the capillary zone

"3":

 

szt
AMzJ-t

AM”

Q 23

39"“)

‘ PET

PAN

[2

(gmme

a <=qe”

[>me

E.
3

located below the intermediate zone and within the
zone of water-table fluctuations.

Lateral ground-water movement through the capillary
zone between the flood plain and adjacent terrace area.

Soil-moisture content in zone 2 of hole j at the end of
budget period t.

Change in soil-moisture content in zone 2 of holej at the
end of budget period t.

Average weighted change in moisture content in zone 2
of the reach during period t.

Total number of sample points in a reach.

Total number of budget—period evaluations of ET or ET’
for a month.

Subscript for consumptive use coefficient k denoting no
phreatophytes.

Subscript denoting “outflow.”

Average precipitation on the area.

The accumulated precipitation at gage j for the budget
period, in inches.

Precipitation.

Monthly percentage of daytime hours of the year used in
equation 12 to determine f BC.

Potential evapotranspiration.

Subscript for f, k, k0, kp, and U denoting that pan
evaporation was used to determine f in equation 12.

Subscript for consumptive use coefficient k denoting
phreatophytes dominated by saltcedar.

Gila River inflow.

Gila River outflow.

Tributary inflow.

Subscript for k, k0, kp, and U denoting that the solar
radiation equation was used to determine f in equation
12.

Subscript denoting a sampling type of error.

Standard deviation.

Standard deviation of differences.

Subscript denoting “soil zone of soil profile.”

Apparent specific yield.

Temperature or subscript denoting a given budget period.

Transmissivity or subscript denoting “tributary.”

Computed evapotranspiration.

Subscript denoting a given cover density class.

Numerical descriptor of phreatophyte cover on a given
area.

Width of saturated alluvium.

Exponent in equation 15.

Subscript denoting a given soil—moisture zone.

Measurement error in water-budget component.

Minimum average difference between measured and
computed evapotranspiration for all accepted budget
periods.

GILA RIVER PHREATO'I‘YPE PROJECT

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING
PHREATOPHYTES, GILA RIVER FLOOD PLAIN,
GRAHAM COUNTY, ARIZONA

By R. C. CI'LI.ER, R. L. HANSON, R.

M. lVI‘i'RICK,

R. M. TI'RNER,and F. P. KII’PLE

ABS'I‘RAUI‘

The conveyance of ground water to or from a river channel and
down its valley is an important hydraulic function of the alluvium
underlying that river’s flood plain. In the arid southwestern
States, evapotranspiration from a flood plain can result in a
significant reduction in the quantity of water conveyed to down-
stream users. A large part of this evapotranspiration is transpira-
tion from deep rooted plants, called phreatophytes, which obtain
most of their water from the saturated zone and capillary fringe.
Phreatophyte control, consisting of the removal of the phreato-
phytes and substitution of plants having a lower consumptive use
and higher economic value, has been proposed for and applied to
large areas of flood plain in an attempt to reduce the conveyance
losses. The relatively high consumptive use by phreatophytes has
been documented by numerous studies, but the actual reduction in
evapotranspiration resulting from the application of phreatophyte
control on the flood plain of a major river has never been measured.

The US. Geological Survey initiated the Gila River Phreato-
phyte Project in 1962 with the following objectives: (1) develop
methods of analyzing the hydrology of a flood plain; (2) determine
the evapotranspiration and the change in evapotranspiration
resulting from the application of phreatophyte control on a flood
plain typical of areas of existing or proposed application; (3)
develop methods of extrapolating results to other areas; and (4)
evaluate the reliability of the results. The project site consisted of
15 miles (mi) or 24 kilometers (km) of the Gila River flood plain in
southeastern Arizona, subdivided into four contiguous reaches.
The areas of the reaches ranged from 1,400 to 2,300 acres or 570 to
930 hectares (ha). In 1962, the vegetation consisted mainly of
saltcedar and mesquite of variable heights and densitites ofcover.
Removal of the phreatophytes was done in stages beginning in
1967 and completed in 1971. Postclearing attempts to establish
grass were unsuccessful because of heavy grazing and adverse
weather conditions, but annual plants did provide temporary cover
when shallow soil moisture was available during the growing
season.

Evapotranspiration was evaluated for each reach as the residual
in a water—budget equation consisting of twelve components
measuring all inflow and outflow of water through each reach, for
budget periods of two or three weeks, during the study period 1963
through 1971. Evaluations were made for 414 budget periods.
Measurement errors in the water budget are important because the
accuracy of the evapotranspiration data is dependent on the
quantity of water measured as inflow and outflow rather than on

 

the magnitude of the evapotranspiration. The errors in each
component and in the total budget were evaluated and the maxi-
mum potential evapotranspiration before and after clearing was
computed. Acceptance criteria based on measurement errors and
potential evapotranspiration were used to establish acceptable
maximum and minimum evapotranspiration values and maxi-
mum errors in these values. Applying these tests to the water
budget evaluations provided 321 acceptable evapotranspiration
values.

The accepted evapotranspiration data were fitted to four pre-
viously developed and widely used empirical evapotranspiration
equations by use of an optimization program. Optimum fitting was
achieved when the average difference between measured (accepted)
and computed evapotranspiration for each accepted budget period
was minimized. An analysis of variablity between measured and
computed values indicated a possible error in the annual values
computed by empirical equations of 15 percent before clearing and
25 percent after clearing.

Annual evapotranspiration on the project area averaged 43
inches (in.) or 1,090 millimeters (mm) before clearing, and ranged
from 56 in. (1,420 mm) for dense stands of phreatophytes to 25 in.
(630 mm) on areas of no phreatophytes. The removal of phrea-
tophytes resulted in a reduction in evapotranspiration averaging
19 in. (480 mm) per year and ranged from 14 in. (360 mm) on reach
1 to 26 in. (660 mm) on reach 3 because of the difference in the
density of phreatophytes. This reduction is temporary and would
not apply after permanent replacement vegetation became estab
lished. A flood plain without phreatophytes is in an artificial
condition, and the water requirements for maintaining this condi-
tion will depend on the land-management practices applied.

A logical replacement of phreatophytes would be a cover of
forage grasses. For this reason the consumptive use of water for
various grasses was computed with empirical equations using
previously published parameters derived for optimum production
of grasses under irrigation near Mesa, Ariz. The computations
indicated a consumptive use greater than the evapotranspiration
from the Gila River flood plain before removing the phreatophytes.
Assuming that these grasses could be established, it can be
postulated that the consumptive use would be less than under
irrigation, production would be less than optimum, and some
water would be salvaged. Data to confirm or disprove this postula-
tion must await further studies.

P1

P2 GILA RIVER PHREATOPHYTE PROJECT

INTRODUCTION

The principal source of water in the southwestern
United States is the relatively high precipitation
falling on headwater areas. Use of this water is
largely confined to cities and irrigated farms in the
arid low lands. Water is transported from source to
user through a conveyance system consisting of
channels and flood plains of rivers and their tribu»
taries. The efficiency of this conveyance system is
reduced by phreatophytes, which are deep-rooted
plants growing on the flood plains and drawing their
moisture from ground water. The flood plain serves
two functions in the hydraulic system of a drainage
basin: (1) it conveys surface flows that exceed the
capacity of the river channel and (2) it conveys
subsurface flows to or from the river channel and
down the valley through the underlying alluvium.
Because of the abundant water supply and fertile
soil, the flood plain is an ideal environment for the
production of plants. Dense thickets of phreatophytes
now cover many of the flood plains in the south-
western United States, retarding the movement of
flood water over the surface of the flood plain and
causing greater flood damage. The high consumptive
use of water by the phreatophytes constitutes a with-
drawal from the subsurface flow and results in a
reduction in the quantity of water available down-
stream. The scarcity of water in the southwestern
United States has prompted a search for additional
approaches to water management.

Phreatophyte control, consisting of the removal of
the phreatophytes and their replacement with other
types of vegetation, has been proposed and applied at
numerous sites. The intended benefits from phreato-
phyte control are: reduced flood damage, reduced
evapotranspiration, and greater economic return
from the site by the production of more valuable
vegetation. Nonbiologic problems caused by the
removal of phreatophytes include an increase in
flood-plain erosion and in downstream silt load
should the replacement vegetation not become well
established. Also, phreatophytes provide a wildlife
habitat and a greenbelt of luxuriant vegetation in
otherwise sparsely vegetated areas, and loss of these
features must be considered. A comprehensive discus—
sion of the problems of managing the phreatophyte
habitat has been presented by Horton and Campbell
(1974). Quantitative data showing benefits and detri—
ments are necessary to determine the desirability of
applying phreatophyte control to any particular site.
The prediction of the quantity of water diverted from
phreatophyte use is of primary importance in plan-

 

ning phreatophyte control; the quantity thus saved is
equivalent to the amount by which evapotranspira-
tion is changed following vegetation modification.
In 1962, before initiating this study, the Geological
Survey examined the available data on evapotran-
spiration. The high consumptive use of water by
various species of phreatophytes had been measured
at several locations. Blaney and others (1942) re-
ported an annual use of 4.68 feet (ft) or 1.43 meters
(m) by saltcedar planted in tanks at Carlsbad, New
Mexico, where the average depth to water was 4 ft
(1.2 In). Gatewood and others (1950) measured evapo-
transpiration from 9,303 acres or 3,765 hectares (ha)
of the Gila River flood plain near Safford, Arizona. A
total of 28,000 acre-feet (acre-ft) or 34.5 cubic hecto-
meters (hmi) was used during the 12-month period
ending September 30, 1944. During this period esti-
mates were obtained of the total water use for various
species of phreatophytes growing in tanks (Gatewood
and others, 1950) and evapotranspiration was eval-
uated at several ground-water well sites by means of
the transpiration-well method developed by White
(1982). These studies showed that the observed
annual use of water for 100-percent volume density
was 7.2 ft (2.2 m) for saltcedar, 4.7 ft (1.4 m) for
baccharis, 6.0 ft (1.8 m) for cottonwood, and 3.3 ft (1.0
m) for mesquite. Methods of extrapolating these data
to the heterogeneous vegetation on a typical flood
plain were neither developed nor tested. Phreatophyte
control was not performed on any of these sites and
changes in evapotranspiration were not measured.
The determination of the quantity of water that
could be saved requires an estimate of evapotran-
spiration both before and after the application of
phreatophyte control. The postclearing condition of
the flood plain is only temporary unless some form of
regular maintenance is performed to prevent re—
invasion by phreatophytes. Large areas of privately
owned flood plain have been cleared to provide agricul-
tural land. In this case, the postclearing evapotran-
spiration depends on the crop planted. Phreatophyte
control projects are generally planned to convert the
phreatophyte areas to grass in order to minimize the
cost of maintenance and to provide an economic
benefit in the form of forage. Ideally, the replacement
vegetation should minimize consumptive use, maxi-
mize forage production, and resist invasion by phreato-
phytes. The types of vegetation which will completely
satisfy these criteria in the flood—plain environment
have not been identified and the density of grass
species that can be established is not predictable.
Thus, an estimation of evapotranspiration for the
postclearing conditions with a beneficial replacement

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P3

vegetation on the flood plain cannot reliably be
made.

An evaluation of available evapotranspiration data
indicates that several deficiencies exist when apply-
ing these data to present water-management prob-
lems. In an attempt to correct these deficiencies the
following objectives were established for the Gila
River Phreatophyte Project: (1) develop methods of
analyzing the hydrology of a flood plain; (2) deter-
mine the evapotranspiration and the change in
evapotranspiration resulting from phreatophyte con-
trol on a flood plain; (3) develop methods of extrap-
olating results to other areas; and (4) evaluate the
reliability of the results. The water budget was select-
ed as the primary method of measuring evapotrans-
piration from flood-plain areas.

The criteria for selecting a study site were set by
the objectives and the methods of measurement. The
site requirements included the following: (1) hydrau-
lic characteristics suitable for evapotranspiration
measured by the water-budget method; (2) a large
area of dense phreatophytes; (3) authorization to
apply phreatophyte control; and (4) uniform land
management. A reach of the Gila River flood plain
within the San Carlos Indian Reservation was select-
ed as the best site available. Continuous records of
the flow of the Gila River through the reach were
available for the preceding 33 years (Burkham, 1970);
changes in the channel and flood plain had been
observed since 1929 (Burkham, 1972), and the vege-
tation had been repetitively mapped since 1914
(Turner, 1974). Phreatophyte control on this reach, as
proposed by the Corps of Engineers, was authorized
by Congress in Public Law 85—500 (US. Congress,
1958) and approved on July 3, 1958 as part of the
project entitled, “Gila River channel improvements
between Camelsback Reservoir site and Salt River,
Arizona.” A formal agreement was made on May 28,
1962 between the US. Geological Survey and the San
Carlos Apache Indian Tribe, with the approval of the
Bureau of Indian Affairs, for the use of reservation
lands for the project.

Evapotranspiration was measured by the water-
budget method during the period March 1963 through
September 1971. The phreatophytes were not dis-
turbed until December 1964. Removal of the phreato-
phytes was done in stages and was completed in
March 1971. The volume of water lost to evapotran-
spiration was measured for 414 two- or three—week
budget periods on four contiguous reaches of the
study area. Evapotranspiration was related, by use of
empirical equations, to potential evapotranspiration
and to the type of vegetation. Error analyses of the

 

components of the water budget, prepared by Burk—
ham and Dawdy (1970) and Hanson and Dawdy
(1976), were used to evaluate the reliability of the
results.

PERSONNEL AND ACKN'O‘VLEDGMENTS

The project was planned and activated under the
general supervision of Thomas Maddock, Jr. and F.
E. Clark, former chiefs of the General Hydrology
Branch and G. E. Harbeck, Jr., research hydrologist.
From 1967 to 1972 general supervision was provided
by R. W. Stallman, research coordinator, Rocky
Mountain Region, and since 1972 by P. C. Benedict,
research coordinator, Western Region. The study was
under the direct supervision of R. C. Culler, project
chief, from 1962 to 1973. R. L. Hanson became project
chief in 1973. Collection of records and analysis of
data on surface water, sedimentation, and chemical
quality of water were under the supervision of H. M.
Babcock, district chief, Water Resources Division in
Arizona. The authors of this report were responsible
for processing and interpretation of data used in this
analysis.

Extensive cooperation and assistance were pro-
vided by the San Carlos Agency, C. J. Rieves and T.
B. White, superintendents; by the San Carlos Irriga-
tion Project, M. D. Young, general engineer, Bureau
of Indian Affairs; and by the San Carlos Apache
Tribe, Marvin Mull, Tribal Council President.

DESCRIPTION ()F THE STUDY AREA

The study area is located on the Gila River flood
plain within the San Carlos Indian Reservation,
Graham County, in southeastern Arizona. The area
extends from the US. Highway 70 bridge near Bylas,
15 miles (mi) or 24 kilometers (km) downstream,to the
mouth of Hackberry Draw, which is within the San
Carlos Reservoir and is located 11 mi (18 km) up-
stream from Coolidge Dam. The flood plain ranges
from 3,500 to 5,500 ft (1,100 to 1,700 m) in width and
has an average downvalley slope of 0.0016. The Gila
River is normally confined to a meandering channel
about 110 ft (34 m) wide and 7 ft (2 m) deep. Terraces
up to 25 ft (8 m) in height border the flood plain
throughout most of the reach. The channels of numer-
ous tributary streams are graded to the flood plain
and fanshaped deltas are formed at the mouth of
each stream. The flood plain ranges in elevation
above mean sea level from 2,460 to 2,585 ft (750 to 788
m). The Santa Teresa Mountains to the south and the
Gila Mountains to the north reach elevations of 8,200
and 5,000 ft (2,500 and 1,500 m), respectively.

The upstream extension of the study area was

GILA RIVER PHREATOPHYTE PROJECT

P4

A8582 33::me 38 3.8 e835 33m «$0 MERE? mmEIA $5ch

muss. v m N F 0 295mm $05 22E be 4:55
_ _ _ _ _ L

_

o

 

                  
    

ummm =033§ow£
1 _ _ _ _ _ _ _ lav

wmmhmfiogg w .528». 5:: m¢0

a? 9.2353 c 95. O
= a : «Ir—IQ:

 
     

 

 

 

 

O U _
Sam mm: .b
/ O
V Ema
.5qu EEmEEéom IQ O
F .3 0&9 333 :05.“ D
\\l

\\ Boa .mEboZ

$0> Immmt
m0 qt «0 >36

 

559$
1/?

 

 

 

 

 

.585...

r!) «.65 >36
_

_ \I E V

_ ~ ovm 835n—

\ ,

_ ‘

_ /

_

— <ZONE< , Mu

_ m m. 33m 25 sf =0 53% 9.350 695:. 95 E503 $80

a

_ . |||I®.|||

— ? r own ‘

1 i /x t)\

_ ’03 Em a o 0 PE :5

_ . _ we av v on

llllllllll
FLI: LIEIL .\/\!\:}-:(:
co: 0N: ow: ZO—P<Z<An—Xm—

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P5

limited by irrigated farm land on the flood plain
above the Bylas bridge on US. Highway 70. Down-
stream, the area was limited by possible inundation
by the San Carlos Reservoir. Although no part of the
study area had been flooded by the reservoir since
1944, the project was designed to preclude total inter-
ruption of the study by high water levels in the
reservoir. The study area included 5,500 acres (2,200
ha) and was initially subdivided into three reaches
(fig. 1). The upper reach, designated reach 1, is above
the maximum water level which can occur in the San
Carlos Reservoir. This reach extends from the US.
Highway 70 bridge downstream to the mouth of Bone
Spring Canyon. Reach 2 extends from the mouth of
Bone Spring Canyon to the mouth of Salt Creek. In
1966, the lower part of reach 2 was inundated by
reservoir water resulting in the designation of the
upper half of reach 2 as an additional reach (2a).
Reach 3 includes the area between the mouths of Salt
Creek and Hackberry Draw. This reach was flooded
by the reservoir in January 1966 which terminated
the collection of data there because the Gila River
channel was plugged with sediment.

Topographic maps of the project area include plane-
table surveys by the Soil Conservation Service in
1914—15 to a scale of 1:12,000 with 5 ft (1.5 m)
contours. All of the project area except 1 mi (1.6 km)
of reach 1 at the upstream end was included in a 1947
map of the San Carlos Reservoir by the Corps of
Engineers at a scale of 1:7,200 with 5 ft (1.5 m)
contours. Topographic maps to a scale of 1:62,500
with 80 ft (24.4 m) contours were published by US.
Geological Survey (USGS) as the Bylas Quadrangle
in 1960 and the San Carlos Reservoir Quadrangle in
1962. Provisional unedited USGS 71/2 minute quad-
rangle maps to a scale of 124,000 with 40 ft (12.2 m)
contours were also available. Vertical control for
project surveys were tied to US. Coast and Geodetic
Survey (C&GS) bench mark elevations above mean

      
 

 

 

  

sea level (supplementary adjustment of 1937) by third
order leveling. Horizontal control was established by
third order triangulation from C&GS monuments to
relate project surveys and the Arizona Zone East
Grid System.

Prior to the beginning of this project aerial photo-
graphy was available for 1935, 1942, 1947, and 1954.
Black-and-white, color, or color-infrared aerial photo-
graphy ranging in scale from 1:900 to 1:120,000 was
obtained on 76 dates between March 1962 and
November 1972. Vegetation, channel changes, inun-
dation by the reservoir, and clearing progress were
observed by ground surveys and the use of aerial
photographs.

Access to the flood plain was provided by roads
crossing the flood plain along 13 cross sections at
about 1 mi (1.6 km) intervals as shown in figure 1. All
instrumentation was located along the cross sections
to minimize the area of phreatophytes disturbed by
the installation and servicing of equipment. Repeti-
tive surveys were made along the cross sections to
measure erosion or deposition on the flood plain and
changes in the river channel.

GEOLOGY

The Gila River flood plain at the project site is
within the Safford basin, a typical basin-and-range
downfaulted sediment-filled trough between uplifted
ranges as described by Davidson (1961). Two water-
bearing sedimentary units, identified as basin-fill
and alluvial deposits, are included in the valley fill as
described in detail by Weist (1971).

The basin fill, divided into a silt and sand facies
and a limestone facies, consists mainly of a very fine
sand and silt that is partly cemented. The beds dip
gently from north to south into a poorly defined axial
area slightly south of the Gila River. The basin-fill
deposits extend to a depth of about 1,000 ft (300 m)

  

 

 

   

 

 

s
3000' — N_
900
b
Floodplain and g
terrace alluvium a
= 800
2500' — o _ g
’—
LU
..... 2
-_ _ . ._. 700
1" ' Basin fill 5.;
2000' ‘—
o 500 1000 METERS Geology by E. 3. Davidson, 1964 600
o 2000 FEET

VERTICAL EXAGGERATION ><4
NATIONAL GEODETIC VERTICAL DATUM OF 1929

FIGURE 2,—Geologic section (diagrammatic) across the study area.

P6 GILA RIVER PHREATOPHYTE PROJECT

and reach a width of 10 mi (16 km). Water enters the
basin fill along the outer boundaries of the formation
and along tributary stream channels and flows
toward the center of the valley where it is discharged
into the overlying alluvial deposits. The basin fill
contains water generally under artesian pressure, but
because it is fine grained, it yields only a few gallons
per minute to wells.

Alluvial deposits occur in the channels cut into the
basin fill by the Gila River and its tributaries. The
deposits are divided into terrace alluvium, which
underlies the entire Gila River flood plain, and the
flood-plain alluvium, which overlies the terrace allu—
Vium and occupies the central part of the flood plain.
These two units consist of poorly sorted lenticular
deposits of sand, gravel, and some silt and form a
single aquifer. Figure 2 shows a typical geologic cross
section of the study area.

The soils on the project are a heterogeneous
complex of alluvium ranging from clay to cobbles
and are constantly relocated by erosion and deposi-
tion. All tributaries to the project area are graded to
the flood plain and not to the river channel. Flows in
these tributaries are infrequent and consist primarily
of high peak discharges carrying large quantities of
sediment which are deposited in alluvial fans on the
flood plain. Flood flows in the Gila River redistribute
these deposits on the flood plain creating a hetero—
geneous and frequently changing soil surface. Soil
texture profiles at a number of sites in the project
area have been described by McQueen and Miller
(1972).

(ILIMATE

The project area is classified as semiarid by Thorn-
thwaite (1948, pl. 1A). Mean annual precipitation is
about 12 in. (305 mm). Longterm climatological data
are available and have been summarized by Sellers
and Hill (1974) for National Weather Service stations
at San Carlos Reservoir, 20 mi (32 km) west; San
Carlos, 20 mi (32 km) northwest; and Safford, 35 mi
(56 km) east of the project. The mean annual precipi-
tation for the 30-year period 1941—70 ranges from 8.43
in. (214 mm) at Safford to 14.15 in. (359 mm) at San
Carlos Reservoir. The seasonal distribution of precipi-
tation and temperature is similar at all stations and
is shown in table 1 as mean monthly totals at San
Carlos Reservoir for the period of study and the long-
term record 1941—70.

Burkham (1970) has described in detail the types of
storms which produce runoff. The seasonal distribu-
tion of precipitation can be divided into two periods
which have distinctly different types of storms. About
40 percent of the annual precipitation occurs as late

 

afternoon thunderstorms during the four-month
period July through October. That rainfall is of high
intensity, short duration, and covers small areas.
Moist tropical air, usually from the Gulf of Mexico,
enters east-central Arizona during this period and
the storms are triggered by orographic uplift and
high surface temperatures. Winter precipitation is
very erratic from year to year although it is generally
less violent and of longer duration than the summer
rains. Cold season precipitation is normally associat-
ed with cyclonic storms that develop in the North
Pacific Ocean and move eastward over the continent.
These storms usually remain too far north to bring
more than strong winds and cloudy conditions to the
area. However, when they follow a more southerly
track and intensify off the coast of southern Cali-
fornia, significant quantities of precipitation can
occur. The maximum (monthly precipitation for
December at San Carlos Reservoir was 8.53 in. (217
mm) in 1965; the December mean of 1.77 in. (45 mm)
has been surpassed 16 times in the 42 year period
1931 through 1972. Drought conditions are most
prevalent in May and June when the average month-
ly precipitation at San Carlos Reservoir is less than
0.25 in. (6.4 mm); the total precipitation for both
months has been zero in nine years during the period
1981—72.

Temperature extremes range from 10° Fahrenheit
(F) or -12° Celsius (C) to 115°F (46°C). Each of the
seven months from April to October has experienced
maxima exceeding 100°F (38°C) and minimum tem-
peratures below freezing have been observed in all
months except June through September. Mean daily
temperatures range from 32°F (0°C) to 60°F (16°C) in
winter and 65°F (18°C) to 100°F (38°C) in summer. The
average diurnal temperature variation exceeds 29°F
(16°C) in both winter and summer as shown in table
1. The estimated mean monthly relative humidity

TABLE 1,—Climatological data for San Carlos Reservoir

 

Temperature DF
means

Precipitation (inches)
means

 

 

 

 

March 1963 March 1963-
1941—1970 June 1973 1941—1970 June 1973
Month Daily Daily Monthly Monthly Monthly Monthly
Maximum Minimum
January 58.4 32.6 45.5 45.2 1.58 0.90
February . . 653.9 35.6 49.8 49.2 1.01 107
March 68.8 40.0 54.4 54.3 1.46 1.22
April . 78.6 47.9 63.3 61.6 .48 .38
May 88.0 56.8 72.4 72.0 .22 .20
June 97.1 65.9 81.5 80.6 .24 .10
July 99.8 73.2 86.5 86.8 1.81 177
August . 97.1 71 ‘5 84 2 83.9 2.32 2 83
September 93.7 64.8 79.3 78.0 1.28 l 60
October . 83.1 52.6 67.9 67.7 1.08 .54
November 69.2 39.9 54.6 55.9 90 1 45
December 59.7 33 4 46.6 45.8 1 77 2.71
Annual

average 79.8 51.2 65.5 65.1 1.18 1.23

 

Annual total

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P7

ranges from 23 to 64 percent with the minima
occurring at 1800 hours in either May or June and the
maxima occurring at 0600 hours in August. Annual
pan evaporation averaged 97 in. (2,460 mm). Average
total monthly wind movement ranged from 650 mi
(1,050 km) in December to 1,250 mi (2,010 km) in July
at the San Carlos Reservoir station.

The preceding climatic data are based upon long-
term records from National Weather Service stations
in the vicinity of the project area. The climate near
the ground on the project area differs to some extent
from these data because of the moderating effect
induced by the high phreatophyte transpiration. In
dense thickets of phreatophytes during the growing
season, the diurnal variation in temperature is re-
duced and the relative humidity is maintained at a
high level. Another local influence, cold air draining
from the adjacent mountains onto the valley floor,
can produce extremely low minimum temperatures in
winter.

VEGETATION

At the time the project began in 1962, the moist
conditions prevailing along the flood plain had pro-
moted a dense growth of phreatophytes. This vege-
tation comprised mainly saltcedar and mesquite.
Cottonwood, seepwillow, seepweed, and arrowweed
were also present (Turner, 1974). On the uplands
above the flood plain grew low, open stands of

 

creosote-bush, mesquite, catclaw, and whitethorn.

The vegetation of the study area was mapped on
aerial photographs at a scale of 127,100 (plate 1). Two
major vegetation types, saltcedar and mesquite, were
recognized. Within these two types, irregular parcels
of apparently homogeneous vegetation were outlined
on the photographs. Canopy cover estimates were
made photogrammetrically (Turner, 1974). Average
plant heights were determined for each parcel from
field observations. Canopy cover values were grouped
into four cover classes: 1—25 percent, 26—50 percent,
51—75 percent, and 76—100 percent. The following
height classes were recognized: for saltcedar, 0—6.5 ft
(0—2.0 m), 6.6—13 ft (2.0—4.0 m), and greater than 13 ft
(4.0 m); for mesquite 0—7 ft (0—2.1 m) and greater than
7 ft (2.1 m).

The method devised for describing the hydrologic
parameters of the study area utilized a grid system of
quadrangles each 2,000 ft (610 m) on a side (fig. 3).
The quadrangles were further subdivided into one
hundred square plots, each 200 ft (61 m) on a side
with an area of 0.918 acre (0.372 ha). The plots were
assigned vegetative descriptors based upon the can-
opy coverage and height classes of the parcel into
which each quadrangle fell. Table 2 gives the number
of plots in reach 8 that fell within the vegetative
classes noted above. Where the vegetation comprised
only ephemeral plants, the parcel was regarded as
bare ground. In a few instances the plots fell within
parcels of upland vegetation. The upland vegetation

TABLE 2.—Number of l-acre plots in Reach 3 characterized by combinations of species (mesquite and saltcedar), canopy-cover class, and
height class. Values apply only to the area of flood-plain alluvium.

 

 

 

 

SALTCEDAR MESQUITE
CANOPY
COVER, A B c D A B c D
HEIGHT... . 1 2 3 1 2 3 1 2 3 1 2 3 1 2 1 2 1 2 1 2 TOTAL
~ 3 - - - » 9 3 20
- 9 1 - 13 - 19 - 33 16 4 2 3 100
. 1 - » - 3 - 8 - — 12
« 1 - » - - 1 1o 12
. 2 - 6 - . 75 6 1 2 8 100
- 3 » 20 - - 65 - - 88
. . » — . 4 4 ~ 4
- - - » - 46 34 30
. 1 - 13 - » 85 1 100
- - - - - 56 - 56
- 1 » - - - 3 3 12
- 2 — 18 - - 72 8 100
- - - » - 80 — 80
. 1 - - - 10 1 12
- - ~ - 3 97 . 100
» - » 1 31 . 32
- - 1 » 23 ~ 24
- - - - 31 69 100
- - 43 50 7 100
1 - - 15 — 15
- - 2o . 2o
- 42 58 100
3 2 1 17 10 31 12 76
- 3 55 29 84
6 1 62 7 76
» - 20 - 20
15 27 42

 

 

Canopy coverage, in percent—Class A, 1—25 percent; Class B, 26—50 percent; Class C, 51—75 percent; Class D, 76—100 percent.

Height, in feet: Saltcedar——C1ass 1, 043.5 ft; Class 2, 65430 ft; Class 3, 13.0+ ft.
Mesquite—Class 1, 0—7 ft; Class 2, 7+ ft.
' See figure 3

P8

TABLE 3.—Area from which vegetation was cleared on project
reaches during 7 years

 

 

 

Calendar A C R E S C l. E A R E 1)

year Reach 1 Reach 2 Reach Ila Reach 3 TOTAL

1964 ,,,,,,,,,,,,, 1160 360

. 268 268

1,040 1,040

1,440‘ 1,440‘

1,374 1,374

114 114

819 819

 

 

5,415

TOTAL , .......

1,668 933 1,374 1,440

 

“ Vegetation killed by inundation.

was not classified further. Tables similar to table 2
were prepared for each reach and the data for all
reaches are summarized in table 9 (see “Use of ET
data for defining a prediction equation”).

Removal of all flood-plain vegetation was a treat-
ment condition incorporated into the experimental
design of this research program. The vegetation
removal or “clearing” is described in detail by Park,
Culler, and Turner (1978) and will be discussed here
in general terms only. Root plows were used to cut the
roots of the phreatophytes below the crown. The
debris was collected and piled into windrows for
burning. Cleared areas were left fallow for one year to
locate areas of re-establishment by phreatophytes.
These areas were cleared a second time. Most of the
vegetation was removed during the period from 1966

110°25’

 

GILA RIVER PHREATOPHYTE PROJECT

through 1971. The acreage cleared, the location, and
associated clearing dates are given in table 3. As
vegetation removal progressed, the vegetative descrip-
tors for each plot were changed to reflect the new
unvegetated condition. The vegetative condition of
the project area will be defined as: (1) preclearing—
conditions on the flood plain before clearing began;
(2) partial clearing—conditions during the period of
clearing; (3) postclearing—conditions after the phreato-
phytes were removed.

Postclearing attempts to establish grass were
unsuccessful because of heavy grazing and adverse
weather conditions. Bermuda grass established itself
in some areas and various annual plants appeared
for a few weeks each year.

SURFACE WATER

Discharge in the Gila River, as inflow at the up-
stream end of a reach and as outflow at the down-
stream end, is the largest and most variable com-
ponent in the water budget. Long-term records of the
flow of the Gila River are available in the annual
reports of the US. Geological Survey for the Calva
gaging station at cross-section 9 in the project area.
The drainage area above this station is 11,470 square
miles (miz) or 29,707 square kilometers (kmz) and the
average annual discharge for the period 1929—72 was

 

 

   
 

 

33015, 110 15
\
5 flood plain 5
a .
33°10’ — <9
3 \
\ 0 1
32
Bylas o
THOUSANDS OF METERS
0 1 2 3 4 5 6 7 8
I l l l I | l | l
I l I | I l
0 5 10 15 20 25

THOUSANDS OF FEET

FIGURE 3.—Map of study area showing grid network and the system for numbering quadrangles used in describing vegetation cover.

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P9

181,100 acre-ft (223 hm3). Flow in the Gila River is
highly variable, as indicated by the annual discharges
for the period of record which ranged from 20,870
acre-ft (25.7 hm3) in 1956 to 804,100 acre-ft (991 hm3)
in 1941. Instantaneous flow ranged from zero, which
occurred many times, to 40,000 cubic feet per second
(ft3/s) or 1,133 cubic meters per second (m3/s) on
August 13, 1967. The post-1914 peak flow occurred on
January 20, 1916, and was estimated to exceed 100,000
ft3/s (2,800 m3/s).

Burkham (1970) in a comprehensive analysis of
runoff in the Gila River above Coolidge Dam shows
that major flows are confined to two distinct periods
because of the seasonal distribution of precipitation.
Runoff from July through October is produced by
convective storms and the discharge is highly vari-
able with numerous flood peaks of short duration.
High sustained flows occur in winter from snowmelt
and frontal storms. May and June are ordinarily the
months of extremely low flows although discharge
can be very low in any month. A particularly low
runoff year occurred in 1956 when the total monthly
discharge was zero at the Calva gaging station in
June, and from September through December.

Diversions of the Gila River above the project area
are used for metallurgical treatment of ores, for
municiple use, and for irrigation of about 69,000 acres
(28,000 ha).

The tributary streams to the Gila River in the study
area drain about 260 mi2 (673 kmz) and individually
range in size from 0.1 to 39 mi2 (0.3 to 100 kmz).
Burkham (1976) gives a complete description of these
tributaries including the seasonal runoff for the per-
iod of 1963—71. The basins are long and narrow and
drain the north slopes of Mount Turnbull and the
south slopes of the Gila Mountains. The channel
slopes range from 2 percent near the river to more
than 40 percent at the higher elevations of Mount
Turnbull. All tributaries are ephemeral and most of
the flow is the result of summer thunderstorms. High
rates of discharge occur for short periods, generally a
few hours or less, and many tributaries do not flow
every year. The channels are graded to deltas on the
flood plain and many flows seep into the alluvium
without reaching the Gila River channel.

GROI'ND WATER

The 1,000 ft (300 m) thick underlying basin-fill unit
described previously is recharged on the high moun-
tain slopes adjoining the flood plain and because of
its low permeability is an artesian aquifer under the
flood plain. Weist (1971, fig. 8) mapped contours of
the potentiometric surface in the basin fill and found

 

this surface to be subparallel to the land surface on
the steep valley slopes. The depth to water in wells 2.5
mi (4 km) from the flood plain was 360 ft (100 m).
Ground-water movement in the basin fill is down-
slope toward the Gila River and slightly downstream
to the west. The water in the basin fill is under
sufficient artesian pressure to rise above the water
table in the overlying alluvium and water movement
is always from the basin fill t0 the alluvium.

The alluvial deposits have a relatively high perme-
ability. Water in the alluvium is unconfined and is
recharged from the basin fill, downvalley movement
of ground water in the alluvium, overbank flooding
from the Gila River, surface flow from the channels
of tributary streams which spread over the flood
plain, and precipitation falling on the flood plain.
Discharge from the alluvial aquifer is by downvalley
underflow, transpiration by phreatophytes and other
vegetation, and evaporation from the soil. The Gila
River channel is hydraulically connected to the allu-
vial aquifer and water can move either to or from the
aquifer.

The depth to ground water on the flood plain
ranges from 5 ft (1.5 m) near the river to 20 ft (6.1 m)
near the outer boundaries of the flood plain. On the
adjacent terraces the depth to water is from 20 to 40 ft
( 6 to 12 m). During high reservoir-water levels in the
San Carlos Reservoir the ground-water level rises to
the ground surface near the river in the downstream
part of the study area. Aquifer tests (Hanson, 1972)
show that the average storage coefficient is 0.15 and ,
the average transmissivity is 28,000 square feet (ftz)
or 2,600 square meters (ml) for the alluvium. The
average downvalley slope of the ground-water table
is 0.0016 and the resulting downvalley subsurface
flow averages 5.1 acre-ft (0.0073 hm3) per day.

MAN‘S INFLUENCE

The observed influence of man on the project area
began as early as 800 years ago with the construction
of pueblos by members of the Salado culture (J. E.
Ayres, Arizona State Museum, oral commun., 1974).
Remains of these houses are in evidence on the
terrace south of the Gila River in the vicinity of Calva
and at Dewey Flat. These sites were surveyed in 1959
and the more important ones were excavated. In
1966, prior to clearing, the area was resurveyed for
archeological sites under the direction of J. E. Ayres,
Assistant Archeologist, Arizona State Museum, Uni-
versity of Arizona. These prehistoric residents may
have raised crops on the flood plain but any lasting
effect was quite insignificant. Cultivation of the proj-
ect area in historic times has not been extensive.

P10

About 170 acres (69 ha) were farmed on Dewey Flat
in about 1870 (Bureau of Indian Affairs records, San
Carlos Agency, unpub. data). At Calva, 65 acres (26
ha) were cultivated in the 1930’s (BIA, San Carlos
Agency), and an area of 360 acres (146 ha) near the
Bylas highway bridge was cleared and leveled for
flood irrigation by diversion from the river in the fall
of 1964. Only 65 acres (26 ha) of this area were
continuously farmed during the period 1965—71.

The historical changes in natural flood-plain vege-
tation have been described by Turner (1974). Man has
caused changes in both prehistoric and historic times
by the deliberate burning of the vegetation to improve
grazing or to flush animals from the dense thickets.
The frequency of burning cannot be determined nor
can its long-term effect on the vegetation be defined.
Man is responsible for triggering a significant change
in the dominant species of the flood-plain vegetation
by the introduction of saltcedar. This exotic and
prolific plant was not found on the project area in
1914, but has since invaded and dominated the area.
Removal of the riparian vegetation by the phreato-
phyte control project, as described by Park, Culler,
and Turner (1978), produced a dramatic change in the
vegetation which in turn produced a significant
change in the evapotranspiration of the area.

The construction of the San Carlos Reservoir has
altered the flood-plain topography and the river chan—
nel. The flood plain in reach 1 is wholly above the
effect of reservoir backwater and is gently sloping
toward the entrenched river channel with terraces
along the outer boundaries. The flood plain in reach
3, however, contains large deposits of reservoir sedi-
ment, eliminating the cross-valley slope toward the
river, obscuring the adjacent terraces, and reducing
the channel conveyance so that natural levees have
developed. The reduction in channel conveyance has
in turn caused complete plugging of the channel by
debris (Kipple, 1977).

The construction of levees and bridges has pro-
duced local irregularities on the flood plain. In 1907, a
2,000 ft (610 m) fill was built across the Gila River at
the railroad bridge 1 mi (1.6 km) above the mouth of
the San Carlos River. The bridge was abandoned in
1928. This fill constricts the river channel and has
increased the sediment deposition on the upstream
side of the fill (Kipple, 1977). A replacement railroad
bridge upstream near Calva (cross-section 9) confines
the low water channel to the south side of the flood
plain and the Bylas highway bridge at the upstream
end of the project area near section 1 confines the
channel to the north side of the flood plain. A 2,000 ft
(610 m) levee was constructed in 1964 on the north
bank of the river below the highway bridge to protect

 

GILA RIVER PHREATOPHYTE PROJECT

the farm land in the project area. The floods of
December 22, 1965 and August 13, 1967 overflowed
the levee and caused some erosion and sediment
deposition.

METHOD OF ANALYSIS

The determination of evapotranspiration (ET) of
phreatophytes from a flood plain by the water-budget
method requires that all significant movement of
liquid water into and out of the flood plain be
measured. Twelve liquid-water components have been
defined as significant in the water budget of the Gila
River flood plain. An equation expressing these
components is _

ET: QILQO'F QT+AC+P+AMS+AMI

+AMC+GB+GI'GO+AMTC (1)
where

ET = evapotranspiration from the area,

Q 1 = surface inflow of the Gila River,

Q0 = surface outflow of the Gila River,

QT = surface inflow from tributaries bordering
the area,

AC = change in Gila River channel storage,

P = average precipitation on the area,

AMS = average change in moisture content in
the unsaturated soil zone located immedi-
ately below the land surface,

AM, = average change in moisture content in
the unsaturated intermediate zone located
between the overlying soil zone and the
underlying capillary zone,

AMC = average change in moisture content in
the capillary zone located below the inter-
mediate zone and within the zone of
water-table fluctuations,

G B = ground-water inflow vertically upward
into the alluvium from the underlying
basin fill,

G1 = ground-water inflow downvalley through
the saturated alluvium,

G0 = ground-water outflow downvalley

_ through the saturated alluvium, and
AMTC = lateral ground-water movement through
the capillary zone between the flood plain
and adjacent terrace area.
ET determined with equation 1 will henceforth be
referred to as “measured ET.”

One factor not considered in the water-budget
equation which may be significant during some bud-
get periods is surface depression storage. Reliable
field measurements of depression storage were not
possible. However, the only time this factor appears
to be significant is during periods of high flow in the

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

Gila River when the error in the measurement of the
QT and Q0 components and the resultant ET are
large.

The water-budget equation can be solved for any
given length of period, if that period is long enough
for measurable changes to occur in the components.
However, the budget period must also be short enough
to permit detection of seasonal variations in ET.
Other factors considered in the selection of the
budget-period length (for this study) included the
time and cost in data collection and the frequency
with which data could be measured without redun-
dancy of information. Because of these considera-
tions, budget periods of 14 and 21 days were used—
the length depending on the frequency with which
soil-moisture and precipitation measurements could
be obtained. Except for basin-fill inflow, which was
assumed constant, all other components of the water
budget were recorded continuously and could, there-
fore, be evaluated for any length of budget period.

COLLECTION AND CONVERSION
OF BASIC DATA

The project was instrumented for the independent
measurement of each component in the water budget.
The collection and conversion of data for these com-
ponents are described in the following pages. The
errors involved in computing the individual com-
ponents and in evaluating ET as the residual in the
budget equation were analyzed and described by
Burkham and Dawdy (1970) and Hanson and Dawdy
(1976).

GILA RIVER STREA MFLOW

Measurements of discharge were required at the
upstream and downstream ends of each reach. A
gaging station, Gila River at Calva, has been in
operation since 1929 at cross-section 9. Additional
gaging stations were installed at cross-sections 1, 17,
and 23 in 1963. A gaging station was also established
at cross-section 13 in June 1966 following inundation
of the stations at cross-section 17 and 23 from back-
water in San Carlos Reservoir. A continuous record
of river stage was obtained at each gaging station
and a stage-discharge relation was established and
maintained by repetitive current-meter discharge
measurements.

The gain orloss of flow within a reach of the river
is the difference between the measured inflow and
outflow. Generally, this difference is a small fraction
of the flow passing through the reach and accurate
discharge records are essential to obtain reliable
measures of gain or loss. Burkham and Dawdy (1970)

 

P11

analyzed the accuracy of the measured discharge and
developed criteria to predict the error in flow volumes.
The magnitude of this error is a function of both the
quantity of flow and the accuracy of the stage
discharge relation. The Gila River channel is subject
to considerable scour and fill; thus, good definition of
the stage-discharge relation requires frequent dis—
charge measurements. The frequency of measure-
ments ranged from one week in the winter to three
per week in the summer with additional measure-
ments during floods.

Records of daily discharge are complete for the
study period at cross-sections 1 and 9. The gaging
station at cross-section 13 was in operation after
June 1966 but the discharge record is incomplete
during floods and during the first six months of 1968
when backwater affected the stage-discharge rela-
tion. No streamflow data in excess of 3,500 fth (99
m3/s) were obtained at cross-section 17 and the data
are incomplete during periods in 1966, 1968, and 1969
because of backwater from the reservoir. A stage
discharge relation could not be established for flows
exceeding about 500 fth (14 m3/s) at cross-section
23. The channel at this station was completely
plugged in 1964 as shown in figure 4 and the station
was relocated 1,000 ft (305 In) downstream on an
excavated channel. The station records were not used
after July 1965 because of renewed inundation from
the reservoir and plugging of the excavated channel.
The channel plugging process is described by Kipple
(1977).

The mean annual discharge, the maximum and
minimum discharge, and the total annual discharge
for the gaging station at cross-section 9 for each
water year (October 1 to September 30) during the
period of study are given in table 4. The variability in
monthly discharge at this station during the study
period is shown in figure 5.

CHANGE IN CHANNEL STORAGE

The water stored in the channel within a reach is
computed as the product of the average cross-sec-
tional area of the stream and the channel length.
Discharge measurement notes provided the data
necessary to compute the cross-sectional area at the
time of measurement. These data were used to develop
an area-stage relation for the cross sections at the
ends of the reaches. The recorded stages provided the
data for determining the cross-sectional area at the
beginning and end of each budget period. Average
cross-sectional area for the reach is the average of the
areas for each end of the reach. Change in channel
storage (AC in equation 1) during a budget period is

P12

TABLE 4.——Annual and extreme discharges of the .Gila River
at Calua, cross-section 9, during the study period

 

Water Mean annual Extremes (ft‘/s) Total annual

 

 

year (ft1 /s) Maximum Minimum (acreft)
1963 ................ 242 3,240 0 175,100
1964 . 130 3,060 0 94,390
1965 126 4,700 0 90,960
1966 737 39,000 11 533,400
1967 205 40,000 9.1 148,400
1968 798 8.960 21 579,300
1969 83.6 1,160 37 60,560
1970 43.1 982 2.6 31,220
1971 82.6 7,470 1.2 59,770
1972 243 7,160 4.0 176,400

 

 

the difference between the channel storage at the
beginning and end of the budget period.

TRIBI'TARY INFLOW

The discharge from the numerous tributaries to the
Gila River within the project area is an input to the
water budget. The method of measuring the flow from
these tributaries was designed on the following
assumptions and criteria: ( 1) the total discharge during
a water-budget period from all tributaries to a reach
would be required; (2) although the smallest water-
budget period would be two weeks, longer water-budget
periods might be used; (3) significant amounts of
runoff occur only during the months of July through
October; (4) tributaries contribute surface flow to the
study reach for only a small part of the time and this
flow is assumed to be a small part of the total volume
of water included in a water budget for most periods;
(5) accurate records of flow for each tributary would
not be required; (6) the water-budget computations
could be omitted for periods of excessive discharge
from tributaries; and ('7) periods of no flow in the
tributaries would be accurately defined.

 

GILA RIVER PHREATOPHYTE PROJECT

The location of tributary basins and gaging instal-
lations are shown in the report describing the
measurement of tributary streamflow (Burkham,
1976). Gages were installed on streams draining 235
mi2 (608 kmz) of the total 260 mi2 (673 km?) of area
tributary to the project area. Continuous recording
gages were established on streams draining the 10
largest basins, which included 54 percent of the
tributary area. Six other recording gages were located
at sites selected on the basis of the physiography and
orientation of the basins. Crest-stage gages were
established on 47 other tributary streams. Stage
recorders were operated during the July through
October runoff season and the crest-stage gages were
regularly inspected during this 4-month period.
Tributary runoff data were obtained from 1963—7 2 for
reaches 1 and 2. The discharge of the Gila River at
cross-section 23 could not be measured during periods
of tributary discharge, therefore, flows from the
tributaries to reach 3 were not used.

A stage-discharge relation was computed for each
tributary gaging station using the standard-step
method as described by Burkham (1976). The initial
intent was to verify these relations by using current-
meter and slope-area measurements of discharge. This
plan was later discarded when it became apparent
that this refinement of the stagedischarge relation
could not be justified because of the limited signif-
icance of tributary runoff in the water budget and
because of the difficulties involved in making verifica—
tion measurements. Thus, only a few current-meter
and slope-area measurements of floodflow were
obtained.

Runoff from tributaries having recording gages
was computed by applying the stagedischarge rela-

 

 

FIGURE 4.—Views looking upstream at the gaging station and channel at cross-section 23 on July 15, 1964 (upper photo) and on
September 6, 1964 (lower photo) after deposition from channel plugging had filled the channel.

tions to the recorded gage heights. Discharge data
from watersheds near the project area having drain-

MONTHLY DISCHARGE, IN ACRE-FEET

170

160

150

140

130

120

110

100

90

80

70

50

40

30

20

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l

210

—- 200

— 190

— 180

-—— 170

— 150

-— 140

—- 130

— 120

_— 110

—— 100

"90

-—80

—-70

-—60

—50

—4o

—30

—20

—10

 

 

1963 l 1964 I 1965 l 1966 I 1967 I 1968 I 1969 l 1970 i 1971 l 1972

WATER YEARS

FIGURE 5.—Monthly discharge of the Gila River at Calva, Arizona.

 

MONTHLY DISCHARGE, IN CUBIC HECTOMETERS

P13

age areas of less than 100 mi2 (259 km?) were used to
develop a relation between peak discharge and

P14

volume of storm runoff. The volume of runoff from
each storm event was then obtained by applying the
discharge-volume relation using the peak discharge
computed for those tributaries with crest-stage gages.

Average seasonal runoff from tributaries to reaches
1 and 2 for the period 1963—71 was 1,370 acre-ft (1.69
hm”) or 9 acre-ft per mi2 (0.0043 hm” per kmz). The
seasonal runoff to reach 1 ranged from a minimum of
40 acre-ft (0.049 hm“) in 1970 to a maximum of 1,620
acre-ft (2.00 hm“) in 1967. For reach 2, the seasonal
runoff ranged from 90 acre-ft (0.11 hmil) in 1970 to
2,220 acre-ft (2.74 hm3) in 1971. The largest runoff
from an individual storm occurred on August 5 and 6,
1967 and totaled 280 acre-ft (0.35 hm“) in reach 1 and
690 acre-ft (0.85 hm“) in reach 2. The highest peak
discharge on an individual tributary was estimated
as 8,000 ft“/s (227 m3/s) during the storm. On a unit
area basis, the maximum peak discharge was 2,800
ft3/s per mi2 (25 m3/s per kmz) on July 16—17, 1967.
There was no flow in any of the tributary streams
during 96 percent of the days in a year.

PRECIPITATION

Precipitation falling on the project area is an inflow
component of the water budget. The limited areal
extent of the summer storms required a relatively
dense network of rain gages to provide an adequate
sample for the computation of the volume of precipi-
tation. Three types of gages were used: float—actuated
digital recorders, weighing recorders, and non—
recording wedges. Recording gages were installed at
the ends of each reach and wedge gages were located
at the ends of each cross section as shown in figure 1.
The distance between gages is about 1 mi (1.6 km).
Hourly data were recorded by the digital gages and
the weighing gage charts were interpreted for one-
hour intervals. Wedge gages were read every 2 to 3
weeks when soil-moisture measurements were made,
to obtain the precipitation accumulated during a
budget period.

Precipitation records for wedge gages are complete
for reach 1 from September 1963 through September
1971, for reach 2 from October 1963 through Sep-
tember 1971, and for reach 3 from June 1964 through
February 1966. The recording gages were operated
from January 1964 to September 1971, except for
periods of instrument malfunction.

Data from the wedge gages were used to compute
the volume of precipitation for a budget period.
Occasionally the precipitation at a gage was not
obtained. In such instances, the precipitation was
estimated using observed data from nearby wedge
gages or from the recording gages. Representative
portions of the project area were assigned to each

 

GILA RIVER PHREATOPHYTE PROJECT

wedge gage by a modification of the Thiessen Method
(Thiessen, 1911). The total accumulated precipitation
for a budget period was computed as an average
weighted value from

n n
(.EAij>/2Aj (2)
1:1 j21

P = the average weighted precipitation
for the budget period, in inches,

P- : the accumulated precipitation at
gage j for the budget period, in
inches,

A - = the area assigned to gage j, in acres,
and

n : total number of gages in a reach.

where

Hanson and Dawdy (1976) analyzed the measure-
ment errors associated with the precipitation data.

Mean annual precipitation for the 8-year period
1964—71 was 11.15 in. (283 mm) on the project area.
Precipitation increased downstream from a mean
annual value of 10.35 in. (263 mm) at cross-section 1
to 12.18 in. (309 mm) at cross-section 17; an 18 percent
increase. A comparison of the mean annual precipi-
tation between reaches shows that precipitation on
reach 2 was 1 in. (25 mm) more than on reach 1. Also,
the precipitation on the south bank averaged 11.31
in. (287 mm) compared to an average of 10.98 in. (279
mm) north of the river. This spatial variability in
precipitation is attributed to the orographic features
of the area. The maximum daily precipitation re-
corded was 2.39 in. (61 mm) at the north end of cross-
section 17 on December 27, 1968. The total annual
precipitation for 1964—71 in table 5 is the average
from the wedge-gage data for reaches 1 and 2, and
ranges from 7.50 in. (191 mm) in 1971 to 14.6 in. (371
mm) in 1966.

SOIL MOISTURE

The flood-plain alluvium constitutes a water-

TABLE 5.—Precipitation for reaches 1 and 2 by water year

 

Total precipitation
(inches)

Water year

 

1964 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8.36
1965

1966
1967
1968
1969
1970
1971

 

 

Average ....................................................................................

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

P15

 

 

    
  

 

  

“0°25 460,000 FT 110°15'
I
114° 112° 1_1|0° (/ l (l ' (
.—____ -_‘- . ) F
. . o /
El: '36 / \/ 1762 / f
H ARIZONA l / 35/ 7 f /
/ grl’hoenix Salt Rj®134° lé C? / f /
as 610W is / / g,
1 4"/Project area ‘3 ‘EN / é /_ _/
\L“ I o) 2 °J{ (y
33°15’ — \‘*\__.4_J’32 \\ Basin fill ( K / / _
9 ) / V / /
w o
)i «9/ v/
E Q‘ xx”?
0 j 9/ o /
8 ..... / / ‘
.9 { $7
in 3 s
______ \ \ Qe/
\ (a a e ‘e® } < Flood-plain
Flood-plain )r”%\~n \ \Og/l‘ alluVIum

 

  

alluvium

\

SAN CARL OS
RESERVOIR

33°10’

I

EXPLANATION

Contact
Between basin-fill deposits and
sunlrated alluvium

(D ‘
Cross section and number “i
Wells

Well numbers (0517) given only for
wells discussed in this report

aioweoAS

./

ust

W

G) 0720

Observation well and soil-
moisture access pipe
in cross section

0

Observation well at
miscellaneous site

 

l l

 
    
    

 

Basin fill / / é\
'°/
\4 /
<52
/ Bylas '3
0 2 4 KILOMETERS
0 1 2 MILES

 

L

 

 

Geology by E. S. Davidson, 1970; hydrology by R. L. Hanson, 1970

FIGURE 6.—Location of ground-water observation wells and soil-moisture access pipes.

storage reservoir with a capacity of about 25,000 acre-
ft (31 hm3) within the study area. Approximately 30
percent of the average annual ET is supplied from
changes in this storage. The soil-moisture content
was measured within two areas of each reach: (1) the
flood plain which corresponds to the area for which
ET is evaluated and (2) the adjacent terrace area
which extends out from the flood plain to the contact
of the saturated terrace alluvium with the basin fill.
Measurements were made with neutron soil-moisture
meters at 2- to 3-week intervals, thus defining the
water-budget periods. The difference between the
moisture content measured at the beginning and end
of the period defines the change in moisture content
for the budget period. Three access holes for measur-

 

ing the moisture content were installed on each side
of the Gila River at each cross section as shown in
figure 6. Each hole was classified as one of the
following three types: (1) river hole located adjacent
to the river, (2) flood-plain hole located between the
river and the terrace, or (3) terrace hole located on the
terrace adjacent to the flood plain. The river and
flood—plain holes were used to obtain the change in
moisture content in the unsaturated zone of the flood-
plain alluvium and the terrace holes were used to
obtain the change in moisture content in the unsatu-
rated zone of the adjacent terrace alluvium. The
installation of access holes and the calibration of soil-
moisture meters are described by Myrick in Culler
and others (1970).

P16

Moisture content, expressed as percent by volume,
was measured with the soil-moisture meter probe at
0.5 ft (0.15 m) and 1 ft (0.30 m) below ground surface
and at 1 ft (0.30 m) intervals through the remaining
depth of each access hole. The observed moisture
content ranged from 8 percent at the soil surface to
over 40 percent below the water table. The change in
moisture content was determined for three zones in
the profile: (1) the soil zone extending from the land
surface to 2.5 ft (0.76 m) below land surface in the
flood plain and to 5 ft (1.52 m) below land surface in
the terrace, (2) the intermediate zone extending from
the bottom of the soil zone to about 3 ft (0.9 m) above
the maximum observed ground-water level, and (3)
the capillary zone extending from the bottom of the
intermediate zone to the bottom of the hole in the
flood plain and to about 3 ft (0.9 m) below the
minimum ground-water level in the terrace. No inter-
mediate zone was defined for the flood plain of reach
1 because of the relatively shallow ground-water table
in the reach. The change in moisture content in each
of these three zones within the ET area of the flood
plain corresponds to the water-budget components
AMS, and AM], and AMC, respectively, in equation 1.

The average change in moisture content in a given
zone of the reach for a budget period was computed

from
_ n n
AME: : 2 (AsztAj) / .2 Aj (3)
i=1 1:1
where _
AMZ, = average weighted change in mois-
ture content in zone 2 of the reach
during period t.
AMztj = sz(t-1)_ Maj (4)
where
M2,)“, and M2,, = measured moisture content
in zone 2 of hole j at the beginning
(t-l) and end (t) of the budget
period,
A : surface area assigned to hole j, and
n = total number of access holes.

The surface area A ,- assigned to each hole was
determined using the same modification of the Thies-
sen Method that was applied in assigning areas to
the precipitation gages. When moisture-content data
were missing for an access hole, the change in mois-
ture content for the hole was approximated using the
average unweighted change computed from the
measured access holes of the same type (e.g., river,
flood plain, or terrace) in the reach as the unmeasured

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

hole. A negative change in moisture content (—AM)
indicates an increase of the moisture in the profile
(negative ET component) during the budget period,
whereas a positive change (+AM) indicates a loss of
moisture in the profile (positive ET component).
During some high-flow periods, inundation over
the flood plain prevented access to many of the soil-
moisture access holes to obtain moisture-content
measurements. However, water levels in the ground-
water wells adjacent to each access hole were record-
ed continuously, thus providing a complete set of
water-level data for each reach. It was found that
when data from over one-half of the access holes were
missing, a better estimate of moisture change in the
capillary zone could be obtained from the more com-
plete set of water-level data. The estimate of moisture
change in this case was determined by the relation

AMC= - A71 SA (5)

where AMC (or AM“) is the average moisture change
in the capillary zone of the flood plain (or terrace) in
acre-ft, A72 is the average change in the ground-water
levels in the flood plain (or terrace) of the reach in feet
(positive for a rise and negative for a drop in water
levels), S ’ is the apparent specific yield of the aquifer
in the zone of water-level change (dimensionless),
and A is the area of the flood plain (or terrace) in
acres. An average value of S ’ was determined for both
the flood plain and terrace areas of each reach by
relating A72 to the corresponding All—lg (or A1171“) using
budget periods containing a complete set of water-
level and moisture-content data (Hanson and Dawdy,
1976).

BASIN-FILL INFLOW

Ground water is conveyed from the steep valley
slopes through the saturated zone of the basin fill to
the overlying alluvium on the valley floor. This water
is under sufficient artesian pressure to rise above the
water table in the alluvium. Water-table elevations in
20 existing stock-water wells on the adjacent valley
slopes were observed at about monthly intervals. The
depth to water in these wells ranges from 22 ft (6.7 m)
on the terrace near the flood plain to 360 ft (110 m) 2.5
mi (4 km) south of the river. The slope of the
potentiometric surface toward the Gila River, as
defined by Weist (1971, fig. 8) is 0.017 north of the
river and 0.029 south of the river. The variation of the
water levels in the observation wells that were not
pumped was insufficient to produce any significant
change in the artesian pressure under the flood plain.

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

The discharge from the basin fill was therefore
assumed to be constant.

The hydraulic characteristics of the basin fill were
investigated by aquifer tests at two wells, an analysis
of water-level recessions, and by analysis of geo—
thermal gradients as described by Hanson (1972).
These investigations indicate that the average stor—
age coefficient of the basin fill is 0.0005 and the
average transmissivity is 15 ft2 (1.4 m2) per day. An
analysis of moisture movement in the capillary zone
of the deep terrace wells of reach 1 during the selected
winter periods indicates that the basin-fill inflow is
about 0.3 ft (0.09 m) per year (Hanson and Dawdy,
1976). The ground-water contribution from the area
tributary to the flood plain in the project area origi-
nates in the basin fill and is therefore assumed to be
included in the estimate of artesian discharge from
the basin fill. This value was also tested and con-
firmed in an optimization analysis discussed in a
subsequent part of this report.

GROUND-WATER MOVEMENT DOWNVALLEY

A network of 78 recording wells was established to
measure depth to and movement of ground water in
the project area. Three wells were installed on each
side of the river at each cross section as shown in
figure 6: one near the river, one between the river and
terrace, and one on the terrace. The installation of
wells is described by Myrick in Culler and others
(1970). Water-table elevations were recorded by digital
stage recorders with hourly punch intervals. Records
are available for reach 1 from April 1963 to September
1971, for reach 2 from October 1963 to September
1971, and for reach 3 from May 1964 to September
1965. An exception to the preceding periods of record
are five wells on cross-section 12 which were not
installed until 1966. Several of the wells near the river
were relocated at different times during the study
when channel changes destroyed the original wells
and extensions were required on some wells because
deposition of silt during overbank flooding elevated
the ground surface.

Thirty-eight additional wells located near the outer
boundary of the flood plain between cross sections
were drilled, primarily to define the contact between
the alluvium and basin fill. Water-table elevations in
these miscellaneous wells were observed at about
monthly intervals. These data were used to supple-
ment ground-water elevation data from the network
of recording wells at the cross sections.

Data from all wells indicate a consistent decline of
water-table levels from April through June of each
year. The magnitude of this decline depends on the
flow of the Gila River during the preceding winter.

 

P17

The lowest level occurs in midsummer either before or
after the summer storms, depending on the magni-
tude of the summer flow. Water—table levels near the
river are lower than the river bed during periods of
low or zero flow. Highest water-table levels occur
either in midwinter or in late summer in response to
the flow in the river. The annual variability in water-
table levels in reach 1 ranged from 1 ft (0.3 m) in most
wells during 1970 to 8 ft (2.4 m) at wells near the river
at cross-section 3 in 1968. Prior to 1966, the annual
range in water-table levels increased progressively
downvalley with the maximum range occurring at
cross-section 23. After January 1966, backwater from
San Carlos Reservoir raised the water table in
reaches 2 and 3. The maximum change recorded-
during the study was a 25 ft (7.6 m) rise at well No.
2161 between October 1965 and April 1966.

Ground-water movement downvalley through the
upstream and downstream ends of each reach was
calculated from

G=iTWD (6)

where

G = the downvalley ground-water flow
through the alluvium in acre-feet
per budget period,

1' = average downvalley gradient in
feet per foot of the groundwater
surface during the budget period
through the upstream or down-
stream end of the reach,

T = transmissivity of the alluvium in
acre-feet per day per foot,
W : width of the saturated alluvium at

the upstream or downstream end
of the reach in feet, and
D = number of days in the budget period.

The transmissivity, T, of the alluvium was deter-
mined by Hanson (1972, p. F27) to be 0.644 acre-ft per
day per ft (2,600 m3 per day per m) and was assumed
to be constant throughout the study area. The width
of the saturated alluvium, W, is the distance between
the points of contact of the alluvium with the basin
fill at the water table on each side of the flood plain.
The downvalley slope, i, was computed from the
average ground-water levels for the budget period
measured at the river wells and flood—plain wells at
the cross sections on or adjacent to the ends of the
reach. As an example, the slope through the upstream
end of reach 1 (cross-section 1) was computed from
the average water levels at cross-sections 1 and 3.
Similarly, the slope through the downstream end of

P18

reach 1 (cross-section 9) was computed from the
average water levels at adjacent cross-sections 7 and
11. Average downvalley slopes during a budget period
between cross—sections 1 and 3 ranged from 0.00122
in May 1971 to 0.00183 in October 1966 with the latter
slope being the maximum average slope observed in
the project area during the study. Before 1966, the
minimum average slope observed was 0.000305 be-
tween cross-sections 21 and 23 in August 1964. After
1966, the slopes between these cross sections
approached zero because of the high water levels in
San Carlos Reservoir.

(;R()['ND-\\'ATER MOVEMENT CROSSVALLEY

The previously described network of recording
observation wells was designed to define also the
gradient of the water table perpendicular to down-
valley ground-water movement. However, an analysis
of the ground-water level data indicated that the
crossvalley water-table slopes were too variable to
adequately define either the direction or magnitude
of the crossvalley ground-water flow component.

Neutron-log measurements of changes in storage
in the capillary zone under the terraces indicate,
however, that water does move vertically into and out
of this zone. Any significant loss of water in the zone
by ET from the terrace vegetation is not likely be-
cause the depth of the water table under the terraces
is too deep, 20 to 30 ft (6 to 9 m), to be readily
extracted by the overlying vegetation. Therefore, all
significant movement of water out of the terrace
capillary zone is assumed to be lateral and in the
direction of the flood plain in response to an overall
drop in ground-water levels with a general water-
level gradient towards the Gila River. All significant
movement of water into the terrace capillary zone is
also assumed to be lateral but originating from the
flood plain in response to an overall increase in
ground-water levels with a general water-level grad-
ient away from the river. The crossvalley component
(AM TC in equation 1) was then determined by use of
equation 3, which computes the changes of moisture
in storage during the budget period.

EVALUATING EVAPOTRANSPIRATION

ET was evaluated by equation 1 for each budget
period containing a complete set of water—budget
data. The four reaches provided 530 budget periods
during the 1963—71 study period. The length of each
budget period was 14 or 21 days for most periods, but
did range up to 63 days. The duration of a period
depended on the frequency of soil-moisture measure-
ments. Because of significant missing data in 116

 

GILA RIVER PHREATOPHYTE PROJECT

periods, only 414 periods were actually evaluated to
obtain estimates of ET. Of these a further 93 were
rejected on the basis of criteria described in a fol-
lowing section of this report. ET values and their
corresponding water-budget components are given in
table 6 (at end of report) by water years for each
reach. Figure 7 shows a plot of all ET values, both
accepted and rejected values, expressed in inches per
30 days and monthly pan evaporation at San Carlos
Reservoir.

Each ET value in table 6 represents the rate from
essentially the entire flood plain within the reach as
shown by the boundaries in figure 1. Periods of
missing data and periods when one or more of the
water-budget components were not observed are indi-
cated by blanks in the table. Periods for which no ET
values were computed generally coincide with high
flows in the Gila River that inundated the streamflow
gaging stations.

RELATIVE SIGNIFICANCE
OF THE WATER-BI'DGET COMPONENTS

The quantity of water removed from the project
area of the Gila River flood plain by ET is small
compared to the quantities measured by the water
budget. The relative significance of the various
sources in providing water for ET can be illustrated
by grouping the components of the water budget in
equation 1 as follows:

1. Surface-water inflow ........... Q, + Q,

2. Soil-moisture storage ........... MS +M,+MC+M,F
3. Ground-water inflow ........... GB + G,

4. Precipitation .................... P.

As an example, the inflow from surface water, precip—
itation, and ground water, and the volume of water in
the observed soil profile for each budget period of
1965 in reach 2, are shown by the bar graphs of figure
8. Surface water and ground water are not only
sources of inflow but also components of outflow, and
soil-moisture storage can be either depleted or re-
charged during any budget period. The losses (posi-
tive as contributions to the reach) and gains (negative
as contributions to surface and ground-water outflow
and as recharge to soil moisture) are computed as
follows:

1. Surface water Q, + Q, +AC - Q0

2. Change in soil-moisture storage All—45 + AM,

+ A117,. + AM",

3. Ground water GB + G, - GO.
The resulting gains or losses are also shown in figure
8.

The budget does not imply the disposition of water
from each source in the reach but, rather defines the
residual of all water moving through the system. For

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

SAVCI DE Hid
SHEHINITIIW NI '(lEI) NOIIVHIdSNVHLOdVAEI

<1-
I I I I I
I I I

§8§§§

— —I[IJ

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

SAVG 02 Had SHEEWITIIW NI '(13) NOIIVHIdSNVHIOdV/G Z I
8 8 I 2 t8
§ § .— § 32 § § § 5 a T : : 2
I I l I I I I | I | _ ._
._ I _ _ | I I _ .. _
I I g I I g I E
I I '— I I F I _ <r
m
' ‘ ' - ‘ ‘ 1’
: ,9 : : : :
: .9 5 : : : g _ ‘— T
3 E E I E I I E SAVCI 09 Had 3mm
2 g g I F I I F NI '(.LEI}N0|1VHldSNVH10dV/\3
_ I. _ _ _
_ 8 _ _ _
_ E a; _ _ _
—3 fi
_ g g : I :
: E 3 g : : :
an E t
— .5 a 3 — a: — — a, sum as Had
'— 23! — cc: — — no
— g g 5 — 1’ — — 2 SHEEWIIIIW NI '(13)NOIIVHIdSNVH.LOdV/\3
: 5 E E : : : 8 §
_ CI: 5 _ _ _ “7 I
—E£‘£'fi I
- a _ r _ _
I E g 9 Z : I — _
_ ‘D E _ w _ _ w I I
- 2 E “L - E» - - g — — a
_ o o : : Z : I '—
: o : : : : :
4%
I I _ Z I N I Z
: : E E : : 52:? Z 3 g
m m — _ —
_ _ a: _ _ I _ _
: : : .'- : : :
_ _ _ o - _ _
_ _ _ o _ _ _
2 ; E E ' 2 = I
_ o _ _ _ : :
3 ° 2 E 3 2 2
_ _ _ o _ _ _
I I I o I I I
I o I v I I Q I I N
: : E : : E : - 8
1 I I ' ° 2 : :
._ _ _ 0° _ _ _
_ _ _ o- _ _
_ _ g _ ., ° : g : : g
“ I l 1 ‘ ‘ I I I I I — I —
In 0 Ln 0 L” D Ln L!) D I.” 6 LO 0 LO Q I!) I!)
N N \— ‘— , N N .— .— | N P. |

SAVCI 09 Had SEIHONI NI '(.LEI} NOIIVHIdSNVHIOdV/G

REACH 3

REACH ZA

FIGURE 7.—Measured evapotranspiration for each budget period during 1963—71 and monthly pan evaporation at San Carlos Reservoir.

P19

P20

INCHES, PER 30 DAYS

30

20

10

110

100

90

80

70

60

50

40

30

20

GILA RIVER PHREATOPHYTE PROJECT

 

I I I I I I I

Precipitation

 

I WA-OGI. "I .0 ..01I IJVAHW/A OIi-‘IOIOV/lf/A

— 500

 

 

Surface water (G[+GT)

EXPLANATION

 

Outflow Inflow

 

 

 

 

 

 

 

Apr. May June July Aug. Sept.
1965

FIGURE 8.——Sources of water available for ET in reach 2 during each budget period of 1965.

2500

2000

1 500

500

MILLIMETERS, PER 30 DAYS

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

example, the loss of water from the surface-water
source during January and February of 1965 pro-
bably went into subsurface storage, to be removed by
ET during the following growing season.

Figure 8 shows that some sources include much
larger volumes of water than other sources. The total
annual volume of water supplied to the project area
by precipitation is relatively small, but it does pro-
vide a significant part of the total ET during the late

P21

summer rainy season when the potential ET is high.
Of the four sources in figure 8, surface-water inflow is
both the largest and the most variable with time-—
reaching maximum inflow rates during the winter
and late summer storm periods. However, in June
and early July, surfacewater inflow is a relatively
insignificant source. In fact, during this period there
was at times no surfacewater inflow at all for several
days. The loss or gain through the reach per budget

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reich 2
m 20 1965 "’
>.
g I I I I I I I I I I I a
s — 4°° a
E 10 — Ground water (GI+GB) — fi
0’ tr.
§ I I I I I I I I I l E
Z O 171/ 4/ WAY I; ,I' /V/y 11/ A’ ,r/ A/ m/ ,r/A/ /,I' 171/ M /V A771 0 E
3
2
90
l l l I I l I l
EXPLANATION - .
Change in soil-moisture storage SOII mo'Sture
3" "' (Ms +MI+MC +M'rc) "— 2000
WA
Gain Loss
n — N.
N a I
o
5 E
g 60 —_ _— 1500 g
z
0 2
In
DJ 50 __ _.
a E
g :I'
g i
E 40 _— Soil-moisture storage _—1000 E
L” (Ms+M1+MC+MTc) E
o
w 8
g 30 -—— —— m
I; 3
I:
< s“
20 —— —_ 500 <
10 —- ~—
0 I I I I I I I I I I
Jan. I Feb. I Mar. l April I May June l Ju|y I Aug. l Sept. l Oct. l Nov. I Dec.
1965

FIGURE 8.—Continued.

P22

GILA RIVER PHREATOPHYTE PROJECT

period from this source is also highly variable and
depends on the potential ET rate, the amount of
precipitation occurring during the budget period, and
the amount of moisture available in the soil profile.
Figure 8 shows a loss in flow through reach 2 during
all months of 1965 except March and April when a
gain in flow occurred.

Ground-water inflow is a relatively small and insig-
nificant source and remains nearly constant through-
out the year.

ALL REACHES BEFORE CLEARING

NUMBER OF BUDGET PERIODS USED TO DEFINE THE AVERAGE VALUES

12 14 17 20 15 23

18 14 5 20 24 IS

 

3 I I I I I

2 > PreclpitatIon (15)

0—l.lIl.T.T.T

 

. EL .1. .l‘

E l l

 

 

 

 

 

 

 

 

 

75
50

25

The alluvium underlying the flood plain distributes
the water from the various sources to the plants. It
also serves as a regulatory reservoir retaining water
from periods of large supply and low demand for use
during periods of high demand and low supply. The
bar graph of soil-moisture storage in figure 8 shows
the total moisture measured in the soil profile which
extends from the land surface down through the
unsaturated zone and several feet into the saturated
zone below the ground-water table. The total moisture

8 B 13 13 13 18

ALL BEACHES AFTER CLEARING
NUMBER OF BUDGET PERIODS USED TO DEFINE THE AVERAGE VALUES

15 7

B 5

7

11

 

I I 71 I l I
PreCIpiIation (P)

BIL .i.T.T.

 

‘l'l

 

—75

‘— 50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

U)
Q 9 3:
g s a g
g 5 I I l I I I I I I I I 725 a 8 4 I I I I I I I I I I — 100 2
2 Surface water @1er +AC+QTI E E Surface water IQI-Qo mc +er E
a4 4~ —rlUDw. w. 3— ——75 g
23 E E E
5 3 — -— 75 g z 2 — '- 50 E
Z 3 =‘
Z ' —— 50 S I —T I I I I -_ 25 2
I- I -m .. IITI 0
I I l I I I
O i i J. I L O 71 — —— -Z5
-1 _ I 725 _2 I I I I I I I I l I _50
Jan, Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.
-2 I I | | | I | I | | l ._ ——50
Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct, Nov. Dec.
5 l I I I I I I I l I I — 150 5 l I I I I I I I I I I — 125
5 fl 7 Soil moIstuIe 7 1 125 _ Soil moisture _ 100
(A Ms +A MI+AMC 4 (AMS+A M1+AMc+A Mrcl
4 e m MTCI —— 100 3 - “ 75
3 S
g — “k 75 I: 2 _ —_ 50
3 a
RZ‘ “WE 1’ l l l_I ”E
D: EL
H: _ w” T ‘0
~ I —— 25 n: 0 - I 0 >
ll 1 s I I I I I I 5‘
5 0 J. a 0 2 E‘ —I — _, _25 8
-1 _ __ —25 E #2 e ~— —50 E
#2 — _._ A50 E _3 | | | I | l | I | I | —75 E
L, I
Z i
#3 I I I I I I I I l I I _ —75 2
3 I I I I I I I I I I I — 75
2: Ground wate
E 2 I I I I I I I I I I I 50 E (G *G +Gr “— 50
Z): S 1 Ground water IGFGOI'GB) E E I 0 Bi
_ ~— 25 m a 1 '—
E g I I I I I I I I I 3: I I 2 8 I I 3: I I I I I I I I I 25
_ [I I I I I I I I I I I I g I] I I I I I I I I I I I
Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov, Dec. 2 Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.

FIGURE 9.—Average contribution to ET per month of precipitation, surface-water inflow, soil moisture in storage, and ground-water
inflow for the preclearing period (left) and the postclearing period (right).

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

measured in the saturated and unsaturated zones of
the soil profile of reach 2 averaged about 72 in. (1,800
mm) during 1965. The figure indicates that soil mois-
ture is recharged when streamflow is high. Most of
the loss of soil moisture occurs during the spring and
early summer months (April-July) when precipitation
is negligible and the potential ET is high.

The soil-moisture graph also indicates that the
gain in soil moisture during 1965 nearly equals the
loss. For this reason, soil moisture is commonly
ignored when evaluating ET rates for periods of one
or more years.

The losses and gains in water from each source, as
shown in figure 8, were computed on a monthly basis
for each reach to define the average monthly contri-
bution each source makes to ET—both before and
after clearing phreatophytes from the flood plain. A
plot of these average values is shown in figure 9.
Only ET values satisfying one or more acceptance
criteria (to be described later) are included in the
figure. The number of budget periods used in com-
puting each average monthly value is shown at the
top of the figure and the extent of the vertical lines
through each value defines the standard deviations.
Positive values indicate a loss from the system
(addition to ET) and negative values indicate a gain
(recharge) to the system (subtraction from ET). As in
figure 8, the algebraic summation of the average
gains and losses in all sources for a given month
gives the average ET for that month in inches per 80
days.

ET values from budget periods with excessively
high streamflow were not used in this analysis,
resulting in a biased estimate of some of the average
monthly values shown in figure 9. This bias applies
primarily to surfacewater and soil-moisture values
for the winter. Omission of these ET values gives an
underestimate of both the average loss in surface
water and the average gain in soil moisture for the
winter months. However, because these underesti-
mates are opposite in sign and about equal in magni-
tude, they essentially cancel when computing the
average monthly ET for the winter periods. These
underestimates also explain why the soil-moisture
graph of figure 9 shows less gain than loss during
the year—when, in reality, the total annual loss
should approximate the total annual gain as in figure
8.

Definite seasonal trends are apparent in the graphs
of figure 9, particularly for soil moisture which shows
a buildup (recharge) during the winter months fol-
lowed by a loss due to ET during the summer months.
Obviously, of the four hydrologic sources, soil mois-
ture is the predominant contributor to ET during the
period of a high potential ET from May-July. Surface

 

P23

 

7l||||lll||l

Uncleared and partly cleared

  
  

  

/)_ ._ —L ‘ ‘
ClearedA\

  

EVAPOTRANSPIRATION, IN INCHES PER 30 DAYS

 

 

Jan. Feb. Mar, Apr. May

June July Aug. Sept. Oct. Nov. Dec.

FIGURE 10.—Average monthly ET before and after clearing based
on selected ET data from all reaches.

water and precipitation become the predominant con-
tributors during the winter months and during the
late summer thunderstorm period. Variability in gain
or loss in a given source for a particular month—as
indicated by the vertical lines through each average
value—are due primarily to: (1) year to year differ-
ences in the volume of moisture available from the
source to meet the ET demand, (2) random errors in
the measurement of the source, and (3) differences in
vegetation between reaches and changes in vegeta-
tion due to partial clearing. A discussion of these
errors is presented below.

A comparison between the before and after clearing
graphs show similar seasonal trends but substantial
differences in the amount of water each source con-
tributes to ET. As expected, all sources, other than
precipitation, show a reduction in the amount of
water contributed to ET after clearing. Of particular
interest is the surface-water graph which indicates
that the Gila River changed, on the average, from a
losing stream during the pre- and partial—clearing
period to a gaining stream during some months
during the postclearing period.

The gains and losses from each source in figure 9
were summed to define the average monthly ET for
both the pre- and postclearing periods. Graphs of
these monthly values are given in figure 10. The
average annual pre— and postclearing ET rates from
these graphs are 39.3 in. (998 mm) and 24.5 in. (622
mm), respectively—thus, defining an average annual
reduction in ET of 14.8 in. (376 mm). It should be
emphasized that these annual ET rates represent an
average of reaches 1, 2, 2a, and 3 and include precip-
itation which averaged 11.2 in. (284 mm) per year.

P24

The quantity of phreatophytes varied from reach to
reach before clearing and the data in figure 10 repre-
sent the average of the observed data not adjusted for
the variation in vegetative cover. Empirical equations
with coefficients related to the description of vegeta-
tive cover will be used in a later section of this report
to adjust for this variation.

MEASI'REMENT ERRORS IN EVAPOTRANSI’IRATION

Evapotranspiration represents a comparatively
small loss from a large volume of water as illustrated
in figure 8, and the error in the measurement of this
loss becomes highly significant as the volume of
water measured in the study area increases. This is
particularly true of the surfacewater components in
which the volume of inflow can range from essen-
tially zero to several thousand cubic feet per second.
The fact that some ET estimates are better than
others is apparent in figure 7—i.e., some ET values
(indicated by 0) follow the expected seasonal trends
while other values (indicated by X) are unrealistically
high or low and, in some instances, even negative.
These outliers are obviously in error and should be
discarded, or at least, given little weight when com<
puting average seasonal or annual ET rates.

To establish guidelines for selecting the most re
liable ET data, an evaluation was made ofthe relative
measurement errors associated with each of the 12
measured water-budget components and the corres—
ponding ET values expressed in equation 1 (Hanson
and Dawdy, 1976).

The total measurement error of each component
consists primarily ofa sampling error which is depen—
dent on the number of observation points used to
measure the component. This sampling error is time
variant—reflecting both the variability in repetitive
measurements and the error due to missing data.
Nine of the 12 water-budget components were found
to contain significant sampling errors because the
measurement of each component is obtained from an
independent observation. The estimate of the sampl-
ing error in ET may be computed from

657's : [EZQI + EZQO + EZQT+62A(' + 62F+ EzAfiS + 62AM,

where EMS is the sampling error in ET and the error
terms on the right side of the equation are the
sampling errors of the components indicated by their
respective subscripts.

Included in the total measurement error is a bias
error which gives a constant over- and under-estimate

 

GILA RIVER PHREATOPHYTE PROJECT

of the component. Only the three ground-water com-
ponents (GB, GI, and Go) were found to contain a
measurable bias error. This total bias error was
computed from

5E7}, Z [6203 + 62a, + 5260] 1/2 , (8)
where ”1Tb is the expected bias error in ET and the
error terms on the right side of the equation are the
bias errors for the components indicated by their
respective subscripts. Equations 7 and 8 assume that
the sampling error and bias error are independent
and unknown as to direction. The total measurement
error, em, of each ET value, thus, becomes

GET : [62E]; + 621577,] V2 - (9)

Because of independence between the components,
no covariance term had to be included in the computa-
tion of this total measurement error.

Hanson and Dawdy (1976) indicate that the assump-
tions and criteria used to obtain the total measure-
ment error produce an over—estimate of the true
measurement variability in ET. This error is shown
to be significantly greater than the expected standard
deviation of the computed ET value and is, therefore,
considered to be only an indicator of the relative
significance of each ET value.

As an example, the total measurement errors com-
puted for each ET value are included in table 6, in
column e”. A detailed description of the methods
used to derive the sampling and bias error associated
with each water-budget component is given by Han-
son and Dawdy (1976).

Figure 11 shows the ET values computed for each
budget period in reach 2 for calendar year 1965—a
year prior to clearing when streamflow was moder-
ately high (140 percent of the average annual flow)—
and for calendar year 1970—a year when reach 2 was
partly cleared and streamflow was relatively low (16
percent of the average annual flow). The extent of the
vertical lines through the ET values define the total
measurement error (6”) in ET as defined by equation
9. The vertical bars indicate the error in ET attributed
to the streamflow components Q, and QO and to the
error attributed to the soil-moisture change compo-
nents AM» AM ,, All—lo and AM Tp. The increase in the
error in soil-moisture measurement between 1965 and
1970 was caused by the partial clearing in 1970. A
comparison of the values of Q, in table 6 with the
streamflow errors in figure 1 1A shows that discharge
is directly related to the magnitude of the streamflow
errors—with the largest errors occurring during
periods of highest discharge.

The minimum errors in ET frequently coincide
with the period of maximum ET during May, June,

P25

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

SAVE 08 Had SHEIEWITNW NI ‘SHOHHS UNV NOINHJdSNVHlOdVAJ

6an EB $3 3.8% Rune—mo
mctsw N 5&2 .8m 8.5m “5.8853”: van axiom $3.53 Ea dozfifimcmboawz .5330an can ioimmom moimo QNMIAH $505

8:3 32.3 § km 3358
6&2 4 E; .02 < .E 4 “m2 4. “52358 e=:_oe.__8 5 55 .i
50 E; R3 £2328 Became: 5 ES .1

km E 5% 3 EB we:

:2. .UB 32 do .amm .m:< 23. 22, >22 .5< .52 gm;
8N1
29.5225

Em—
.=m_. .80 .52 So 3% 9% 22. 22. £65. .a< 52 .23 :3. So

 

9:1

V]

_H____fi_J__

 

Q: J

SN]

Smll

o3

   

m2: 85.25%; :8
.__o>._wwax @250 :mm

EB 55855 :8
:95me mozmu :mm

SAVE] 08 33d SEHUNI N1 'SHOHHEI GNV NUIlVHIdSNVUiOdVAE!
SAVE 08 83d SUHBWITIIW Nl ‘SHOHHE UNV NOIIVHIdSNVHiOdVAE

1.1 .m .I 3.

__‘_,_ _ _:_ _;_ _ J _ F‘s

SEED C255“. 8. >j<_E<n_
N 55:

 

 

 

..\

szEd mmofim
N 55m

:2.

.umo

 

SAVCI DE 83d SEHDNI NI 'SHUHHE! GNV NDIlVHidSNVHLOdV/E

P26

and early July when streamflow is low and precipita-
tion is negligible. Generally, for those years in which
streamflow was below average the computed ET
values define a definite seasonal trend with relatively
little scatter in the data as shown by the 1970 data in
figure 113.

An important point to be realized from figures 8, 9,
and 11 is that the total measurement error in ET is
dependent on the volume of water measured in the
reach and not on the magnitude of ET. This is
emphasized in figures 11A and 113 by the nearly
constant error in soil-moisture change for each budget
period, reflecting not the large variation in soil-
moisture change shown in figure 9, but rather the
total volume of soil moisture measured in the reach
which fluctuates relatively little with time as shown
in figure 8.

The seasonal trends in ET data for 1965 and 1970
in figure 11 are indicated by the pan evaporation
curves for these years obtained immediately down-
stream from the project area at San Carlos Reservoir.
These curves can be assumed to approximate the
upper limit of ET throughout the year.

Some negative ET values occur in the water budget,
but their measurement error is generally large and
the error extends into the positive ET range. In a few
instances the measurement error does not explain a
large negative ET value or unrealistically high ET
value. These outliers generally occur during periods
of high streamflow and are assumed to reflect large
unmeasured changes in the stage—discharge relation
which was not fully accounted for in the streamflow
error analysis (Burkham and Dawdy, 1970). These
outliers may also be attributed to unknown quantities
of surface water moving into or out of depression
storage as mentioned under “Methods of Analysis.”

CRITERIA FOR REJECTING MEASURED ET VALUES

As indicated previously, all measured ET values
that were obvious outliers (fig. 7) or contained large
measurement errors, a”, were omitted in the compu-
tation of average seasonal and annual ET rates.
These values were also omitted in the development of
an empirical equation describing the relation between
density of plant cover and ET.

Reaches 1, 2, 2a, and 3 provided a total of 414
budget periods generally 14 to 21 days in length from
which ET could be measured. However, ET values
from 93 of these periods were not considered reliable
because of their extreme magnitude or large measure-
ment error.

The criteria used for rejection of the ET values were
arbitrary, but were based on a consistent set of rules.

 

GILA RIVER PHREATOPHYTE PROJECT

First, all ET values containing a measurement error
which exceeds the maximum potential evapotran-
spiration (PET) for the study area (about 4.8 in. or 122
mm per budget period) were discarded.

The PET used in this test was computed from the
relation

PET : (0.014t - 0.37)R (10)

where t is the mean air temperature in CT and R is the
short-wave solar and sky-radiation flux expressed as
inches per day evaporation equivalent (Jensen and
Haise, 1963). Two adjustments were made. Precipita-
tion, an erratic and generally inconsequential source
of water to the flood plain, is stored in the upper soil
zone and returned to the atmosphere as evaporation
from bare soil or is lost by transpiration from either
shallow-rooted plants or phreatophytes within hours,
or at most, a few days after falling. The PET in an
arid region is sufficiently high to remove this inciden-
tal moisture without significantly changing the
average rate of depletion from ground water during a
budget period. Precipitation (P) and the change in
moisture content in the upper soil zone (AMS) were
therefore removed from the water-budget evaluation
of ET before applying the ET data to empirical
equations. The adjusted ET values (E T’) are thus
defined as
ET’=ET—P-AMS. (11)
Upper and lower boundaries were then established
within which the adjusted values (E T’) could logi-
cally be expected to occur, and all adjusted values
falling outside these boundaries were rejected. The
following is a summary of the rejection criteria used.

1. Reject if eET > 4.8 in. (122 mm) per budget

period (maximum acceptable error).

2. Reject ifET’ < -0.5 in. (-13 mm) per budget period

(minimum acceptable negative ET’).

3. Reject if the preclearing and partial clearing
ET’ > PET, provided that ET’ > 1.8 in. (46 mm)
per budget period (maximum acceptable pre-
clearing and partial clearing ET’ in excess of
PET).

. Reject if preclearing and partial clearing ET’ <
0.25 in. (6.4 mm) per budget period from May
through October (minimum acceptable pre-
clearing E T’ during summer months).

. Reject if postclearing ET’ > 1/2PET, provided ET’
> 1.2 in. (30 mm) per budget period (maximum
acceptable postclearing ET’ in excess of 1/2PE T).

4:.

O1

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

 

U1

4;

N

T’m, IN INCHES PER BUDGET PERIOD
(.0

 

E
l
'4:-
I
I\
I
I
I
I
|
I
I
I
I

 

 

 

 

A. Before clearing and partly cleared

 

N b.) A
I
|

 

C:

ET'm, IN INCHES PER BUDGET PERIOD

 

 

 

 

Oct. Nov. Dec. Jan. Feb, Mar, Apr.

May June July Aug. Sept.

8. After cIearing

FIGURE 12,—Boundaries within which the adjusted ET values
(ET’) were accepted as reliable estimates. Number within the
graphs refer to rejection criteria described in the text.

Figure 12 shows the boundaries within which the
ET’ values (with measurement errors 6,” < 4.8 in. (122
mm)) per budget period were accepted for the pre- and
postclearing periods.

A summary of the total number of ET values
measured, both before and after clearing in each

 

P27

reach, and the number rejected by the above-described
criteria is given in table 7.

The water-budget data indicate a seasonal varia-
tion in ET and a significant change in ET after
clearing. Each reach had a different quantity of
phreatophyte cover before clearing as shown by the
vegetation surveys. Therefore, a difference in the rate
of ET between reaches can be expected before clear-
ing but the rate should be similar between reaches
after clearing. This hypothesis is tested in the next
section of this report by use of the project data in
empirical equations.

USE 01" ET DATA FOR DEFINING
A PREDICTION EQI'ATION

Coefficients for several equations defining the
evapotranspiration process were derived in this study
by use of the ET’ data. These coefficients were derived
for the following purposes: (1) to compare the meas-
ured ET’ from different reaches with the project
area having different quantities of phreatophyte
cover; (2) to compare the measured ET’ for different
seasons on the same reach; and (3) to develop methods
for estimating ET from flood-plain sites in the arid
and semiarid Southwestern States. Evapotranspir-
ation is a complex process which is dependent on any
of three factors: heat, vapor transport, and water
availability. Heat and vapor transport are climatic
factors which define the potential evapotranspiration
(PET) 0f the site. Water available for vaporization on
a flood plain is, to a large extent, dependent on the
phreatophytes which extract water from subsurface
storage and convey it to the evaporative surfaces of
the leaves. The preceding considerations indicate
that the desired equation should include a PET para-
meter, based on climatic data, and a coefficient relat~
ed to a quantitative description of the vegetation.

Many empirical equations for predicting ET have
been developed from field measurements of E T. How-
ever, equations such as those presented by van Bavel
(1966) and Ritchie (1972), requiring intensive data
describing the climate near the ground, are consider-
ed inappropriate for this study because of the wide
range in spatial variability of these data on the flood

TABLE 7. —Number of ET values measured in each reach and the number rejected as outliers or because of a large measurement error

 

 

 

 

 

REACH SUII— REACH SUB- Tngf‘L
1 2 2a 3 TOTAL 1 2 2a 3 TOTAL REAcHEs
PRE- AND PARTIAL POST—
CLEARING CLEARING
Number
measured 75 11a 50 17 260 103 11 4o 0 154 414
Number
rejected ..... 24 29 10 o 63 26 0 4 o 30 93
Number
accepted 51 89 40 17 197 77 11 36 0 124 321

 

P28

plain. A single species of phreatophyte such as salt-
cedar or mesquite may be dominant on all or part of a
flood plain but the aerial density and height of the
canopy is not uniform. Also, irregularly shaped open
spaces occur within the canopy where gravel bars or
other surface conditions prevent phreatophyte estab-
lishment. The heterogeneity of the vegetation produc-
es spatial variability in transpiration resulting in a
wide range of surface temperatures and significant
differences in the humidity of the air near the ground.

The methods and equations based on macroclimat-
ic data developed and presented by Blaney and
Criddle (1962), Jensen and Haise (1963), and Chris-
tiansen (1968) are considered more appropriate to
this study. Studies by Cruff and Thompson (1967)
indicate that the Blaney-Griddle equation (1962) is
the most practical and widely used equation for
estimating potential evapotranspiration in arid envi-
ronments such as exist in Arizona. Rantz (1968)
expanded on the Blaney-Griddle equation by relating
the consumptive use coefficient for various species of
phreatophytes to depth to ground water and to den—
sity of stand. Thus, only the BlaneyvCriddle (BC)
method is discussed below. The other methods are
described and compared to BC in a subsequent section
of this report entitled “A comparison of the Blaney-
Criddle method with other methods.”

The Blaney-Griddle empirical equation (Blaney and
Criddle, 1962, p. 1) applied in this study has been
reduced to the basic form

Uka (12)

 

GILA RIVER PHREATOPHYTE PROJECT

capabilities (PET) of the atmos-

phere in inches per 30 days.
Blaney and Criddle applied 30-day measurements of
ET for various crops to equation 12 and derived
monthly consumptive-use coefficient k defined as

k = ET/f
where
ET A measured evapotranspiration in
inches per 30 days,

pt/lOO,

f

where

monthly percentage of daytime
hours of the year listed in table 8,
and

mean temperature in degrees
Fahrenheit.

p _

A parameter describing moisture availability was
not included in the equations developed in this study
because the ground-water table underlying the flood
plain provides a relatively constant supply of mois-
ture for transpiration by phreatophytes.

Before applying the budget-period adjusted evapo-
transpiration values (ET’) to a prediction equation,
each value was expressed in inches per 30 days.
Months were used as the unit of temporal distribution
within the year and each budget-period value was
assigned to the month within which the midpoint of
the budget period occurred.

Hanson, Kipple, and Culler (1972) present prelimi-
nary work on developing a consumptive-use coeffi-
cient to describe the seasonal variability in ET’ and

 

 

 

 

Where the difference in E T’ due to varying phreatophyte
U : computed evapotranspiration in covers. An expression relating k, in equation 12, to
inches per 30 days, the phreatophyte cover as described by the vegeta-
k = consumptive-use coefficient which tion surveys is defined as
is dependent on the kind and
quantity of vegetation, and k I [8,, + kp V (13)
f = climatic (consumptive-use) factor where
which is a measure of the avail« k, : consumptive-use coefficient for no
able heat and vapor transport phreatophyte cover,
TABLE 8.—Monthly percentage of daytime hours of the year for latitudes 24° to 50° north of equator
[From Blaney and (lriddle, 1962, p. 43]
Month Latitude, in degrees north of equator
24 26 28 10 32 34 36 38 40 42 44 46 48 50
January . , 7.58 7.49 7.40 7.30 7.20 7.10 6.99 6.87 h 71 6.60 6.45 6.30 6.13 5.98
February 7.17 7.12 7.07 7.03 6.97 6.91 6.86 6.79 6.73 6.66 6.59 6.50 6.42 6.32
March . 8.40 8.40 8.39 8.38 8.37 8.36 8.35 8.34 8.30 8.28 8.25 8.24 8.22 8.25
April 8.60 8.64 8.68 8.72 8.75 8.80 8.85 8.90 8.92 8.97 9.04 9.09 9.15 9.25
May , 9.30 9.37 9.46 9.53 9.63 9.72 9.81 9.92 9.99 10.10 10.22 10.37 10.50 1069
June .. 9.19 9.30 9.38 9.49 9.60 9.70 9.83 9.95 10.08 10.21 10.38 10.54 10.72 1093
July 941 9.49 9.58 9.67 9.77 9.88 9.99 10.10 10.24 10.37 10.50 10.66 10.83 10.99
August 9.05 9.10 9.16 9.22 9.28 9.33 9.40 9.47 9.56 9.64 9.73 9.82 9.92 10.00
September 8.31 8.32 8.32 8.34 8.34 8.36 8.36 8.38 8.41 8.42 8.43 8.44 8.45 8.44
October 8.10 8.06 8.02 7.99 7.93 7.90 7.85 7.80 7.78 7.73 7.67 7.61 7.56 7.43
November 7.43 7.36 7.27 7.19 7.11 7.02 6.92 6.82 6.73 6.63 6.51 6.38 6.24 6.07
December 7.46 7.35 7.27 7.14 7.05 6.92 6.79 6.66 6.53 6.39 6.23 6.05 5.86 5.65
Total ............ 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

k = the increase in the consumptive-use
coefficient for phreatophyte cover,
numerical descriptor of the phreato-

phytes on a reach defined as

p

V

4
V=2A, (14)

11:1

[ C,,/100 + (CU/100‘ ]/ 2

where
A, = fraction of the total area in a given

reach having a canopy cover
falling in density class 1),

average percent of cover for one of
the classes of canopy coverage
listed in table 2 as: A, B, C, and D,
withA= 13,B=38,C=63,D=88
percent, where U = 1, . . . 4, respec—
tively, and

exponent accounting for the non-
linearity in the relation between

kp and CU.

Equation 13 shows that k = k0 when the cover
density of phreatophytes is zero (V: 0) and k = k, + kp
when the cover density is 100 percent of the entire
area (V = 1 in equation 14). As defined in this
analysis, k0 applies to surface conditions on areas of

CU

P29

the flood plain with no phreatophytes both before
and after clearing. These conditions can include
seasonal grasses and small areas of exposed surface
water in the Gila River channel. The CU and corre-
sponding AU values were obtained for each reach
from field measurements and from aerial photo-
graphy as described in the section on “Vegetation.”
Table 9 summarizes the data from the vegetation
survey for each reach. Depth to ground water, al-
though seasonally variable, was estimated for each
plot in each reach (see example in table 2 for reach 3)
and the averages are shown in table 9. Coefficients
for the two dominant species of phreatophytes
(mesquite and saltcedar) and depth to ground water
were introduced as variables in preliminary attempts
to define an expression for k other than that shown
by equation 13. Differences between the two species
could not be defined because mesquite is relatively
insignificant in relation to the total phreatophyte
coverage. Thus, no distinction was made between the
two species in computing area of canopy cover. Also,
the differences in the average depth to ground water
for the various reaches are relatively insignificant
and were disregarded in all subsequent analyses.
Volume of canopy has sometimes been assumed to be

 

TABLE 9.—Summary of vegetation survey

 

 

 

 

 

 

 

 

 

REACH... ,, , ,, , 1 2 23 3
TOTAL AREA (acres)..,.,,,, ,, 1,723 2,307 1,374 1,440
AVERAGE DEPTH
TO GROUND WATER (feet) 8.5 12.5 11.0 11.5
CANOPY HEIGHT CANOPY OF PHREATOPHYTE OVERSTORY
COVER CLASS _
CLASS AREA VOLUME AREA VOLUME AREA VOLUME AREA VOLUME
(acres) (acre (acres) (acre- (acres) (acre (acres) (acre-
feet) feet) feet) feet)
SALTCEDAR
1 — — 9 4 9 4 4 2
A 2 184 233 40 51 39 49 23 29
3 21 35 37 63 34 57 7 12
1 - — 14 17 10 12 1 1
B 2 93 345 166 615 95 352 80 296
3 17 84 102 504 33 163 — -
1 5 10 35 72 33 68 1 2
C 2 290 1,781 67 412 62 381 35 215
3 136 1,114 98 803 84 688 52 426
1 25 71 69 197 34 97 — —
D 2 206 1,767 613 5,260 148 1,270 965 8,280
3 55 629 176 2,013 72 824 235 2,688
SUBTOTAL 1,032 6,069 1,426 10,011 653 3,965 1,403 11,951
MESQUITE
A 1 3 l 6 3 6 3 — —
2 549 500 469 427 398 362 11 10
B 2 45 120 189 503 153 407 13 35
C 2 44 194 164 723 129 569 —— -
D 2 - — 3 18 3 18 11 68
SUBTOTAL ,, 641 815 831 1,674 689 1,359 35 113
TOTAL ....................................... 1,673 6,884 2,257 11,685 1,342 5,324 1,438 12,064

 

Canopy cover in percent—class A:1~25 percent, class B226—50 percent, class 0:51—75 percent, and class Dz76-r100 percent. Height of canopy in feetnfor saltcedar—class 120—65 feet,
class 2=6.5—13.00 feet, class 3:13.0+ feet; for mesquite—class }:0-7 feet, and class 217+ feet.

Volume : Average Cover x Average Height

Note: Some of the area in each reach contained no phreatophytes therefore the area of phreatophytes is less than total area.

P30

GILA RIVER PHREATOPHYTE PROJECT

TABLE 10,—Application of vegetation description to empirical equations

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

STATUS OF AREA CANOPY COVERAGE V
REACH CLEARING (acres) PERIOD
CLASS FRACTION OF FROM EQ. 14
. Cu TOTAL AREA WITH
(m percent) (AU) x : 0.75

1 1,723 3/63-4/65 13 0.439 0.076

38 .090 .039

63 .276 .185

88 .166 .148

TOTAL .971 .448

1 Partial ______________________ 1,723 5/652/67 13 .327 .057

38 .074 .032

63 .241 .161

88 .120 .107

TOTAL .762 .357
1 Post __________________________ 1,723 3/67-7/71 0 0

2 Pre ____________________________ 2,307 7/63-12/69 13 .243 .042

38 .204 .088

63 .158 .106

88 .373 .334

TOTAL .978 .570

2 Partial ______________________ 2,307 1/70—2/71 13 .039 .007

38 .076 .033

63 .022 .015

88 .267 .239

TOTAL .404 .294
2 2.307 3/71-7/71 0 0

2a 1,374 6/66-11/69 13 .354 .061

38 .212 .092

63 .224 .150

88 .187 .167

TOTAL .977 .470
2a Post __________________________ 1,374 12/69/71 0 0

3 Pre ............................ 1,440 1/64-6/65 13 .031 .005

38 .065 .028

63 .061 .041

88 .841 .752

TOTAL .998 .826

 

TABLE 11.——Number of accepted budget period ET data (see fig. 7) for each month as related to the status of clearing on each reach and
the numerical vegetation descriptors

 

 

 

 

REACH 1 REACH 2 REACH 2a REACH 3 TOTAL
MONTH PRE- PARTIAL POST- PRE- PARTIAL POS’ - PRE- POST- PRE- PRE- AND POST-
CLEARING CLEARING CLEARING CLEARING CLEARING CLEARING CLEARING CLEARING CLEARING CPARTIIIVL CLEARING
LEARI G

3 2 5 3 3 3 1 12 8

2 1 4 5 3 1 2 2 14 6

5 7 4 2 2 4 4 2 17 13

5 7 5 2 2 5 4 3 20 13

2 2 7 5 2 2 2 4 2 15 13

4 2 9 6 3 3 5 6 3 23 18

3 3 10 5 2 2 4 3 1 18 15

1 2 5 6 2 3 2 14 7

1 5 1 2 l 3 5 8

3 3 5 7 1 6 20 5

3 2 6 9 2 6 1 2 24 7

1 1 7 6 3 3 4 1 15 11

32 19 77 62 27 11 40 36 17 197 124

0.448 0.357 0 0.570 0.294 O 0.470 0 0.826 - -

 

 

TABLE 12.——Summary of monthly and annual kg and kp coefficients in7equati0n 13 derived from equations 12 and 14 where f: pt/100 and
x = 0. 5

 

 

 

 
 

 

VALUESOFf,ININCHESJepANDkU ArlNgI‘LAA‘L .
A
JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. (inches) 011311;)

Monthly f ..... 3.12 3.58 4.41 5.40 6.71 7.83 8.27 7.68 6.57 5.11 3.91 3.09 65.68
Monthly k0 .21 .21 .21 .21 .21 .21 .21 .21 .21 .21 .21 .21
Midmonthly hp ...... -.04 .16 -.04 .17 .52 .76 .69 .69 .73 .68 .10 .17 0.95]
Average

monthly kp .01 .11 .01 .19 .51 .72 .70 .70 .72 .61 .18 .14
f kp __________________ 0.031 0.394 0.044 1.026 3.422 5.638 5.789 5.376 4.73 3.117 0.704 0.433 30.70

 

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

a more definitive quantitative descriptor of vegetation
than area of canopy. However, the use of volume of
canopy as a variable gave a greater error of fitting in
the derived expression for k than did area of canopy.
Thus, canopy cover as defined by equation 14 is
considered to provide the best parameter to use in
equation 13.

Table 10 presents the vegetation data found to be
significant in defining a relation for the consumptive
use coefficient. The only measurable change in phreato—
phyte cover was that produced by the clearing opera-
tions. Clearing was done on the project area during
winter months when ET was low and the phreato—
phytes were defoliated. Values of V, the numerical
descriptors of vegetation, are shown for each reach
and period used in the analysis. These descriptors
were adjusted only after the winter clearing was
completed on all or part of a reach. The fraction of
total area (A L.) of phreatophytes is shown for each of
the four canopy cover classes (CU) in table 10 with the
classes representing average cover densities of 13, 38,
63, and 88 percent. The derivation of the value of “x”
as 0.75 will be described later.

The number of accepted budget period E T’ data for
each month as related to the status of clearing and
value of V on each reach is presented in table 11. The
seasonal distribution of accepted data is fairly uni-
form except for September, when the ET’ were fre-
quently rejected because of the variability in the flow
of the Gila River. Data for all 321 periods were used to
define k0 but the data from only 197 periods (data
representing pre- or partial-clearing conditions) were
available to define kp and x.

The first attempt to define the factors ko, kp, and x
was made using Blaney-Criddle’s expression for the
climatic factor, f. Equations 13 and 14 were substi-
tuted in equation 12 and, for each budget period in
which an acceptable E T’ had been measured, repeti-
tive computations of evapotranspiration (U) were
made by varying k0, kp, and x simultaneously within
preset limits until the computed U agreed closely
with the measured ET’. Included in each computation
was the known climatic factor (f = pt/100) for the
budget period, and observed CL, and AU values corre-
sponding to the reach at the time of year in which
E T’ was measured. The repetitive computations were
performed with a digital computer for a total of 321
budget periods using a trial and error technique
developed by Rosenbrock (1960) and applied to hydro-
logic studies by Dawdy, Lichty, and Bergmann (1972).
Numerous preliminary runs of the optimization pro-
gram were made to determine the seasonal variability
and logical limits for k0, kp, and x. For the final
determination of monthly values, each of the factors

 

P31

 

-7 l I l l l

 

 

 

 

Oct. Nov. Dec. Jan. Feb. Mar
EXPLANATION
o Mid-monthly value (table 12)

l—'—i Duration and plotting position for
a measured budget-period value

:1 Average of the daily values
within a given month

FIGURE 13.—Relation between midmonthly and average monthly
values of k p.

in k, km and kp in equation 13 were optimized to
satisfy the following conditions:

1. Define a linear variation in [2,, between the
midpoints of each month with the value of kp
at the midpoint for any given month lying
within the limits -0.1 < kp < 2.0.

2. Define one x for the year within the limits
0.4 < x < 1.0.

The best estimates of k0, kp, and x were defined
when the accumulated sum of the absolute differ-
ences between U and the corresponding ET’ for all
321 periods reached a minimum value (Amm) defined
as

321
Am,” = 21 [(ET,- U.)] /321. (15)
t:

The “best fit” k0 and kp values obtained from this
computation for each monthly f values are shown in
table 12. The total annual f, f kp, and the minimum
fitting error (Amm) as defined by equation 15 are also
included in the table.

Values of k p were computed for each budget period

from the combination and transposition of equations
12 and 13, or

: ET’-fk0

(16)
P f V

k

P32

GILA RIVER PHREATOPHYTE PROJECT

TABLE 13.—-A verage monthly and average annual U rates for each reach before and after clearing phreatophytes, computed from
equation 12 using the monthly f and average monthly kp values given in table 12. All values exclude precipitation and are in inches

 

 

 

per month.
‘JAN. FEB, MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. ANNUAL
P R E C L E A R I N G
Reach 1
(V = 0.448)
h ________________________ 0.21 0.26 0.23 0.30 0.44 0.53 0.52 0 52 0.53 0.48 0.29 0.27
U ...................... .66 .93 1.01 1.62 2.95 4.15 4.30 3 99 3.48 245 1.13 .83 27.50
Reach 2
(V : 0.570)
________________________ .22 .28 .25 .32 .50 .62 .61 .61 .62 .56 .31 .29
U ...................... .69 1.00 1.10 1.73 3.36 4.85 5.04 4.68 4.07 2.86 1,21 .90 31.49
Reach 3
.22 .31 .27 .38 .63 .80 .79 .79 .80 .71 .36 .33
.69 1.11 1.19 2.05 4.23 6.26 6.53 6.07 5.26 3.63 1.41 1.02 39.45

 

 

Average annual preclearing U for reaches 1, 2, and 3 weighted by area is:
(1,723 X 27:50 + 2,307 X 31.49 + 1,440 X 39.45) / (1,723 + 2,307 + 1,440) = 32.32

 

POSTCLEARING

 

All Reaches
(V = 2)

U ......................

.21
1.63

.21
.75

.21
.93

.21
1.13

.21
1.41

where U in equation 12 is replaced by ET’ and f
represents the average climatic factor for the budget

period. Measured ET’ values for each budget period

represent an average rate of ET’ for the duration of
the period. Figure 13 shows the k,D values for a few

selected budget periods during the fall and winter

months (October-March) when the values are typi-

cally low and erratic. Included in the figure are the

optimized midmonth kp values which were computed

assuming a linear variation in lap between midpoints

of adjoining months. The line connecting midmonth

points defines the variability of k p within the month.

Midmonth values of hp for all 12 months are listed in

table 12 and were used to determine Am,” in equation

15. The average monthly k p values are also listed and

are the best estimates to be used with average-
monthly values of f.

Table 13 shows the average monthly and annual
rates of U computed for each reach using f and the
derived coefficient [2,, of table 12. The average pre-
clearing U from all three reaches was 32.32 in. (832
mm). After clearing, the average U was 13.79 in. (350
mm). The water salvaged, computed as the difference
between the preclearing average U and the post-
clearing U is 18.53 in. (471 mm) or 8,447 acre—ft (10.43
hm3) on the 5,470 acres (2,214 ha).

Assuming that all precipitation is evaporated, the
average annual evapotranspiration for the uncleared
project area can be estimated as U + P or 32.32 + 11.15
= 43.47 in. (1,104 mm). The maximum annual U
which represents areas of 100 percent phreatophyte
cover (v = 1), is computed as the summation of f (k0 +
19,- 1) + P for 12 months and is 56 in. (1,420 mm). The
minimum annual U for areas of no phreatophytes is

 

.21
.81

.21
.65

.21
1.74

21

. 21
1.61

. 21
1.38

107 13.79

computed as the summation off 120 + P for 12 months
and is 25 in. (630 mm). '
The seasonal variability of the f and k, values
listed in table 12 and the fitting error (AW-n) defined by
equation 15 can be attributed to three sources: (1)
measurement errors in the water-budget data, (2)
variability of factors affecting ET which are not
defined by equations 12, 13, and 14, and (3) invalid
application of the optimization procedure in assign-
ing limits for the variables k0, hp, and x, and in
defining the optimizing criteria stated by equation
15. The following study was made to determine the
source and magnitude of these errors and differences.

EVALUATION OF THE DERIVED
EVAPOTRANSPIRATION EQUATIONS

Considerable variability exists in the water-budget
E T’ data and the climatic factor (f) used to define the
coefficients kc and kp, in equation 13. Figure 14 shows
this variability in f and in ET’ for the combined pre-
and partial-clearing periods and for the postclearing
period. The upper dashed curve in each graph shows
the average monthly climatic factor f. These f values
are computed from budget-period data which are
randomly distributed within the months. The vertical
lines define the standard deviation of f for each
month. A comparison of the two f curves shows that
the average monthly f was nearly the same for both
periods and the standard deviation of monthly values
averages less than 0.4 in. (10 mm). This variability is
not a measurement error but rather a real variability
resulting from year-to-year and within-monthly dif-
ferences in temperature.

EVAPUTHANSPIHATlON, IN INCHES PER 30 DAYS

EVAPOTRANSPIRATION, IN INCHES PER 30 DAYS

w

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

 

l l

Uncleared and
pamy cteared

 

l llll

—-200

I I l
//% $\\

/ 23

\1

  

 

 

 

 

 

Cleared

 

 

Jan.

reb‘

Mar.

 

Apr.

Dec.

Nov.

May June July Aug. Sept. Oct.

EVAPUTHANSPIHATVDN, IN MlLUMETERS PER 30 DAYS

EVAPOTRANSPIRATION, IN MILLIMETERS PER 30 DAYS

EXPLANATION

Average climatic
factor 1'80
0
Standard deviation of
climatic factor

12—Number of observatinns

Standard devianon of
measured evapotranspiration

Average measured
evapotranspiratiun, E T’

FIGURE 14.—-Mean and standard deviation of monthly measured ET’ and climatic factor f.

P33

P34

The lower (solid) curve in each graph shows the
average monthly water-budget E T’. These ET’ values
represent the average for all reaches with the vertical
lines defining the standard deviation of the water-
budget ET’ data. The number of observations of ET’
used to define the average values are included for
each month in the graph. Some of the variability in
these ET’ data is real, reflecting actual climatic
differences and differences in the cover density be-
tween reaches. Part of this variability, however,
includes measurement errors in the ET’ data as
illustrated in figure 11. The climatic differences are
explained by f in equation 12 and the1 differences in
phreatophyte cover are explained by AD and CU in
equation 14.

The deviation in the ET’ values in the upper graph
reflect the differences in both the climatic factor and
the phreatophyte cover; the deviations in the ET’
values in the lower graph reflect differences in the
climatic factor only.

The coefficient kg was assumed to be seasonally
constant and preliminary runs of the optimization
program in which 1?.0 was varied indicated some
variability from month to month, but no seasonal
trend was apparent. Evapotranspiration from precip—
itation and from shallow soil moisture (see equation
11) have not been included in the development of the
empirical equations. Therefore, k0 describes the ET
maintained by the upward movement of water from
the subsurface source exclusive of phreatophytes,
and there is no reason to expect a seasonal trend.
Applying the k, derived from this study to other
flood-plain areas in the arid and semiarid regions
may give erroneous results because the value of this
coefficient is a function of soil type and depth to
ground water. The seasonal variability of kp will be
discussed in the comparison of the Blaney—Criddle
method with other methods.

The value of k for a given month will change from
year to year only if the density of canopy cover
changes as indicated by equation 13. The possible
ranges of k for July (month of maximum ET’) is
illustrated in the following table.

Values of k for different ranges and averages of CU

 

C, k
Range Average —

1 0 0.21
1-25 13 0.33
26-50 38 0.51
51-75 63 0.68
76-100 88 0.84
>100 100 0.91

 

GILA RIVER PHREATOPHYTE PROJECT

These k values were determined by applying equa—
tions 13 and 14 to the full range of possible cover
densities (C) using k0 = 0.21 and kp for July = 0.70
(table 12).

In order to show the variability of the consumptive-
use coefficient for only that portion of an area having
canopy cover, assume that k0 applies only to that part
of the area having no phreatophytes (1-C u/ 100). If the
coefficient for the area with canopy cover(C ,./ 100) is
expressed as k1,: k0 + kp then k = k0 (1—CU/100) + k’p
(CU/100). As an example, an area having a range of
canopy cover of 1—25 percent (CU = 13 percent) has a k
value cf 0.33 and k’,, : 0.33 — 0.21 (1013) / 0.13 = 1.13.
The following table shows the k’p values and the
percent of total area for which le’p is valid for each of
six classes of canopy cover. These values apply only
to July.

Values ofk’p and percentage of area for which k’,, and
k0 are valid

 

CL. Percent of area for which
Range 12", HP is valid 12“ is valid

<0 0.00 0 100

1725 1.13 13 87
26—50 1.03 38 62
51—75 0.96 63 37
76—100 0.95 88 12
>100 0.91 100 0

This table shows that, for instance, an area having
a canopy cover falling in C,. class 1—25 percent has an
average of only 13 percent of the area under phreato-
phyte cover. The coefficient k’p for this part of the
area is 1.13 or 124 percent of 0.91, the value of the
coefficient 02,) for an area of complete (100 percent)
canopy cover, The relatively high value of the coeffi-
cient for the space under canopy in areas of in-
complete cover can be explained by the “oasis effect”
as defined by Tanner (1957).

It should be noted that the relative value of k’ for
different percentages of canopy is controlled by the
value of the exponent “x” in equation 14. As pre-
viously mentioned, the value of “x” was determined
as a variable between the arbitrary limits of 0.4 and
1.0 by the optimization procedure. A value of 0.75
provided the minimum value of Am," in equation 15.
However, changes in this value did not produce
significant changes in the value of Am. The reason
for this lack of sensitivity can be explained by an
examination of the data in table 10. The value of “V”
in equation 14 will have the greatest variation for low
values of CU in response to changes in “x.” In table
10, the value of “V” for classes of low CU (13 and 38)
is a small part of the total “V” for the reach. An
exact value of “x” could not be determined because
the optimized fitting of the variables in equations 13

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

and 14 are based on the relation of ET’ to the total
“V.” Although seasonal variability in the value of
“x” resulting from changes in foliation might be
expected, such a trend could not be defined by the
available data and a constant value of 0.75 was used
for all computations.

The computation of U for areas of phreatophytes
involves f and kp (which have within-month vari-
ability), the vegetation descriptor (which varies be-
tween reaches but is constant for the month), and k0
(which is constant for the year). Fitting the monthly
computed U to the monthly measured ET’ by use of
the optimization program averages the ET’, thereby
reducing the scatter in ET’ data caused by errors in
the measurements but retaining the differences in V
and the within-month variability of f and kp. As an
example of the results from the fitting process, table
14 shows the average and standard deviation of
measured ET’, climatic factor f, coefficients kp, com-
puted U, and difference between ET’ and U for the
month of June for pre— and partial clearing.

An estimate of the possible error in monthly values
of U can be obtained by analyzing the difference
between ET’ and U for the budget periods used in the
optimization program. The standard deviation of the
differences is defined as

 

 

P35
N N 2 1’2
2 All 2 A, / N
t=1 t=1
- = (17)
SA N - l
where
At = E T’t ' ’t ’
t = a given budget period, and
N = total number of budget periods in a

month.
Table 15 lists the average monthly ET’ and U for all
reaches, whether cleared, uncleared, or partly cleared,
and the standard deviation (SA).
The average standard deviation of the difference
(ET’ — U) for an annual estimate of evapotranspira-
tion is defined as

12 !/2
§A= 2 5mm?)
m=1

where sA is the standard deviation of the difference,
ET’ - U, for month m. Applying the monthly sA
values in table 15 to equation 18 defines average
before- and after-clearing values of 4.6 in. (117 mm)
per year and 3.2 in. (81 mm) per year, respectively.
Thus, annual computed U values obtained from the

(18)

TABLE 14.—Variability of measured ET’, coefficients, and computed Ufor the month of June for preclearing and partial clearing

 

 

 

 

 

 

 

 

 

STATUS MEASURED COMPUTED DIFFERENCE
REACH OF ET’ f kp U ET’- U
CLEARING
AVERAGE 3 AVERAGE 8 AVERAGE s AVERAGE 3 AVERAGE s
Preclearing 4 3.88 1.10 7.62 0.46 0.524 0.015 4.00 0133 -0.12 1.31
1
Partial clearing 2 4.49 ‘99 7‘66 .30 .466 .006 3157 .85 ‘92 .90
Preclearing 6 5101 1.10 7.71 .40 1618 .008 4.77 .28 .24 .97
2
Partial clearing 3 4.71 1.57 7192 .54 .416 .014 3.30 .29 1142 1.31
23 Preclearing 5 4.04 1.62 7.89 .42 .546 .007 4.31 .27 - .28 1‘55
3 Preclearing 3 6150 .57 7.48 .37 .794 .028 5.94 .50 .55 .73
Total all reaches ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 23 4.71 1.38 7173 ‘40 .570 .111 4.39 .84 .32 1.20

 

N : number of budget period data
s : standard deviation

TABLE 15.—Auerage monthly measured (ET’) and computed (U) euapotranspiration for all reaches and the standard deviation (sA) of the
difference ET’ - U. All values are in inches per 30 days

 

 

 

 

 

 

 

 

 

 

JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT‘ OCT. NOV. DEC. égrlNAJLAL
PRE- AND PARTIAL CLEARING
12 14 17 20 15 23 18 14 5 20 24 15 197
0.060 0.69 0180 1.79 3.48 4.71 4164 3.82 2,98 2.60 1.43 0.99 28.53
.68 .93 .94 1.74 3.37 4.39 4.50 4,10 3.25 2,59 1.20 .86 28.55
1.13 .83 1.17 .90 1110 1.20 1.54 2.12 1.76 1.29 1.15 .97 4.55
POST CLEARING
8 6 13 13 13 18 15 7 8 5 7 11 124
0.30 1.12 0.35 0.83 1110 2.12 1.74 0.91 2.22 1.69 0.55 0.26 13119
.66 .75 .93 1.13 1.41 1.64 1.74 1.61 1.38 1.07 .82 .65 13.79
.98 .70 .68 1.07 .69 .69 1.13 1.49 1.01 .51 .62 .88 3.15

 

 

P36

prediction equation provide estimates of the measured
ET’ that are accurate within about 15 percent before
clearing and within about 25 percent after clearing.

These average standard deviations indicate that
monthly winter ET rates cannot be predicted within
less than about 150 percent of the measured pre- and
postclearing E T’ rates. In contrast, monthly summer
rates can be predicted within about 30 percent of the
measured values for preclearing conditions and with-
in about 55 percent of the measured values for post-
clearing conditions.

The expected error in the estimate of the average
annual water salvage of 18.53 in. (471 mm) (table 13)
as a result of clearing the phreatophytes from the
flood plain is computed as

S—AU: [(4.6)2 + (3.2)2] ”2 = 5.6 in. = 142 mm

per year or about 30 percent of the average salvage.
The validity of the basin-fill discharge, G B = 0.3 ft
(0.09 m) per year was evaluated by a specially design-
ed application of the optimization program. As pre-
viously described, the quantity of artesian discharge
from the basin fill into the alluvium (G 8) could not be
accurately determined and was assumed to be con-
stant. G B was introduced as a variable component in
equation 1 which thus alters the value of E T’ for all
budget periods. The previously optimized values of
the coefficients, k, and kp shown in table 12, were
then held constant and used to recompute U. It was
assumed that if the estimated value of G 3 was signifi-

 

GILA RIVER PHREATOPHYTE PROJECT

cantly in error, then the optimized value would differ
from 0.3 ft (0.09 in). The optimized value of G B was
0.306 ft (0.093 m) and Am,” in equation 15 was not
changed, indicating that no improvement could be
made in the prediction equation by changing
GB.

The preceding analysis was based on the deriva-
tion of coefficients using all of the measured data and
is therefore an examination of the fitting process.
Dawdy, Lichty, and Bergman (1972, p. B10) describe
the difference between the error of fitting and the
error of prediction and indicate the desirability of
using split-sample testing to define the error of predic-
tion. In this study, the split-sample test was applied
by using part of the E T’ data to derive coefficients for
estimating U for budget periods not used in the
derivation of coefficients. The test was applied to the
variability in both time and space.

The optimization program was used in fitting to
derive values of 12,, and kg in equation 13 within the
previously described limits. The value of x in equa-
tion 14 was not optimized but was retained at 0.75.
The program was also used to compute the value of U
for each budget period for comparison with the
measured E T’. The number of budget period data was
inadequate to fit coefficients to any single year or to
any individual reach. For temporal variability, data
from odd numbered years were used to derive coeffi-
cients to predict the values of U for budget periods
occurring in even numbered years. The process was

TABLE 16.—-Variability in U due to fitting coefficients to data from different periods and areas

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n Average annual totals §A Average annual totals
Determined from budget periods Determined from average monthly values
11m." Reaches Reaches Reaches U : f lap U : f k0
(1n) Uncleared Cleared Uncleared Cleared Uncleared Cleared ET’ due to ET’ for no
and (in) and (m.) and (in.) phreato phreato
partially partially partially phytes phytes
cleared cleared cleared
(in) (m) _U Per- 'U Per-
ET’ U ET’ U (m.) cent (m.) cent
(in) (in.) (in.) (in) 1
Column number ,,,,,,,,,,,,,, 1 2 3 4 5 6 7 8 9 10 11 12 13
All data ,,,,,,,,,,,,,,,,,,,,,,,,,, 0.95 197 124 28.53 28.55 13.19 13.77 4.55 3.15 30.70 100 13.79 100
Fitted w .84 104 74 25.431 25.81[ 11,78 12,33 3,761 2.53 33.16 108 12.35 90
even years
Applied ‘0 1.11 93 50 28.85 27.30 13.99 12.38 4.92 3.48
odd years
Fitted to 1.1 I 93 50 28.85 28.28 14.73 15.74 4.81 3.43 26.04 85 16.42 119
odd years
Applied to .86 104 74 25.431 26.47 10.79 17.38 31671 2.53
even years
Fitted tu
reaches 2 and .96 129 47 28.22 29.45 14.54J 14.79 4.59 2.323 26.36 86 16.42 119
2a
Applied to
madches l .97 68 77 27.97 28.72 11.75 16.46 4.78 2.91
an 3

 

Note: Am is the criterion for optimizing (equation 15); n is the number ofdata (budget periods);§.5 is the average standard deviation
(equation 18); percent is the relation, in percent, of U computed from coeffiments fitted to data from different periods and areas
to U computed from coefficients fitted to all data; ‘ no data for September; 2 no data for October.

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

then reversed and coefficients were fitted to data
from even numbered years to predict values of U for
odd numbered years. A test of spatial variability was
made using data from reaches 2 and 2a to predict the
values of U for budget periods measured on reaches 1
and 3. The results were compiled in a manner similar
to that shown in table 15 and are summarized in
table 16. Annual totals from table 15 for fitting
coefficients to all data are shown on the first line of
table 16 for comparison. The variability in budget
period ET’ data together with the inadequacies of the
equations for computing U produced some surprising
values of Am," (equation 15). This value was less for
fitting to even years than for all data and the value
for the application to even years was less than for
coefficients fitted to odd years. The variability in
differences (E T’ - U) for postclearing as indicated by
§A was also less for even years and reaches 1 and 3
than for the fitting based on all data. The lack of data
for September during even numbered years and dur—
ing October for reaches 2 and 2a produced false
values of SA for these tests. The data in columns 4—7
compare ET’ with U for each fitting or application
and are summations of the budget periods, grouped
by months. The values are variable because the
distribution of budget periods within the year are
neither uniform nor complete when the number of
data is small. Data in columns 10 and 11 were
computed using average values of monthly f from
table 12 and average monthly values of kp derived
from each fitting. These values provide a comparison
with the average annual total U using all data.
Values of annual U for no phreatophytes in columns
12 and 13 varied up to 19 percent from the values
determined from all data. The product fiep varied up
to 15 percent from values computed from all data.
The sums of columns 10 and 12, f (k, + kp), vary only 5
percent from values computed from all data indicat-
ing that high values of k, are compensated for by low
values of k p and vice-versa.

This limited application of the split-sample test
indicates that there are no unique characteristics in
the data from different groups of reaches nor from
different periods of years that produce bias in the
fitting process. The percent difference shown in
columns 11 and 13 and the average of these columns
is less than the previously estimated errors of fitting.
It is therefore concluded that the errors of fitting are a
reasonable estimate of the error of prediction.

A COMPARISON OF THE BLANE\'-(IRII)I)LE METHOD
WITH OTHER METHODS

As indicated in the previous section, several empiri-
cal methods other than the Blaney-Criddle method

 

P37

are considered appropriate for expressing evapotran-
spiration in arid environments typical of the Gila
River study area.

In this section three commonly used expressions
for the f and k coefficients are described and com-
pared with the Blaney-Criddle f and k coefficients.

Jensen and Haise (1963) used solar radiation (R) as
the climatic factor for computing ET. They applied
the ratio E T/ R to approximately 1,000 measurements
of ET for individual sampling periods for various
crops. ET/R is equivalent to k in equation 12 since
ET is the actual measured rate of ET and R is the
observed solar radiation expressed in in./day evapor-
ation equivalent, assuming that 1 gram of water
occupies 1 cm3 and requires 590 calories to evaporate.
The determination of leg from equation 12 is then

kg : ET’/fR

where ET’ = adjusted ET as defined by equation 11,
and fR = solar radiation, R.

Jensen and Haise (1963, equation 8, p. 34) also
developed an equation for potential evapotranspira-
tion (PET) which was used previously in this report
(equation 10) as a criterion for rejecting measured ET
values. The application of the Jensen-Haise PET to
equation 12 in defining a consumptive-use coefficient
is

kJH : ETl/fJH
where E T’ is as defined previously and

fJH = PET = (0.014t — 0.37)R. (10)

Solar radiation data for defining fR and fJH were
obtained during 1964—71 at an installation 350 ft (107
m) north of the National Weather Service station at
San Carlos Reservoir. The radiation data were not
continuous and the calibration of the pyrheliometer
was incorrect after 1967 due to the degeneration of
the thermopile coating. Thus, extrapolation of data
from the Phoenix and Tucson National Weather
Service stations was necessary to obtain a continuous
record of solar radiation for the project site. The
Phoenix station was used as the primary source of
data and the Tucson station was used for periods of
missing record at Phoenix.

A linear regression was used to define the relation
between the San Carlos Reservoir radiation and the
Phoenix and Tucson radiation. Monthly averages for

P38

all months of continuous records in 1965 and 1966
were used in the analysis. The relation was defined
by an equation of the form

Rsc I at) + (11R

where

= San Carlos Reservoir radiation
data,

a, and a1: constants, and

R = Phoenix or Tucson radiation data.

RSC

Thirteen months of Phoenix radiation data define

R... : -47 + 1.113,,

with a correlation coefficient : 0.991.
Twelve months of Tucson radiation data define

R“ = -72 + 1.09Rt

with a correlation coefficient = 0.998.

Christiansen (1968), expanding on earlier studies,
confirmed the use of pan evaporation as a climatic
factor in conjunction with a coefficient related to
measured ET for various types of crops. Pan evapora-
tion data for the project were available during the
9-year study period (1968—71) from the National
Weather Service station at San Carlos Reservoir. The
application of pan evaporation to derive a consump-
tive-use factor using equation 12 is

kl’AN : ETl/fPAN
where ET’ is as defined previously and

fPAN I measured pan evaporation.

GILA RIVER PHREATOPHYTE

 

PROJECT

 

f, INCHES PER 30 DAYS
f. MILLIMETERS PER 30 DAYS

 

 

 

 

Mar. Apr. May June July Aug. Sept. Del. Nov. Dec.

EXPLANATION

(,ch =pt/1UO

ofR = solar radiation
AfJH =(0.014t —0.37)R
AfPAN : pan evaporation

FIGURE 15,—Monthly variability in climatic factors.

Monthly values of the climatic factors for each of
the methods of computing f are shown in figure 15.
The monthly range in climatic factors when express-
ed as the ratio of highest to lowest for a particular
method is greatest for the J ensen-Haise PET equation
(fyH) and least for the Blaney-Criddle equation (ch)

TABLE 17.—Summary of monthly and kg and [en coefficients in equation 13, derived from three expressions for the climatic factor f in
equation 12 (x = 0.75 for all computations of k)

 

 

 

 

 

 

VALUES OFf,IN INCHES,kpANDkU ANNUAL
JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. TOTAL . 3...
(inches) (Inches)
Monthly fR ____________ 5.51 7.48 9.84 12.65 14.13 13.69 13.06 11.40 10.40 8.35 5.94 4.63 117.08
Monthly kuR ......... 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11 0.11
Midmunthly
pg .................... -.03 .04 -.04 0 .27 .43 .50 .51 .63 .40 .14 .08 0993
Average
monthly kpR... -.01 .02 -.02 .03 .26 .42 .49 52 .59 .40 .16 .07 32.78
Monthly my ..... .. 1.44 2.28 3.92 6.22 8.98 10.48 11.07 9.42 7.53 4.77 2.44 1.26 69.81
Monthly 120.111 ........ 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16
Midmonthly
pJH .................. .04 .54 0 .18 .47 .50 .62 .54 .78 .81 .68 1.03 0977
Average
monthly kaH ...... .23 .41 .09 .19 .44 .51 .60 .58 .75 .79 .74 .86 36.49
Monthly prN ........ 2.50 4.05 5.96 8.77 11.70 13.76 13.67 11.04 9.10 6.56 3.46 2.24 92.81
Monthly kpAN ........ 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13 0.13
Midmonthly
kaAN ................ 7.04 .24 «.04 .13 .34 .38 .42 .43 .46 .61 .40 .60 0.984
Average
monthly pPAN .08 .17 .02 .14 .32 .38 .42 .43 .48 .56 .45 .50 32.42

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

Figure 15 shows that the total annual factor is
greatest for solar radiation (f R) and least for the
Blaney-Criddle equation (ch)- The seasonal distri-
bution of all factors is similar and the location of the
apex is dependent on the relative significance of
temperature and solar radiation in the determination
of the climatic factor. The apex occurs the earliest
(May) for solar radiation and the subsequent decrease
in fR is caused by an increase in cloud cover begin-
ning in mid or late June. The factors reach a maxi-
mum in June and July for pan evaporation (f M N), and
in July for Blaney-Criddle (2‘30) and J ensen-Haise
(fJH) because average monthly temperatures are high-
est during this period.

The optimizing procedure described previously to
define the factors f, and kp in the Blaney-Criddle
expression were similarly applied to obtain best esti-
mates of these factors with f expressed as fR, fJH, and
f M N, respectively. For this analysis the exponent x in
equation 14 was held constant at 0.75. Table 17
summarizes the results of these computations.

The coefficient kp relates the numerical vegetation
descriptor for a reach, determined by Al, and CU, to
the climatic factor f. Since the vegetation descriptor
did not change seasonally, the derived kp must define
any seasonal variability in this relation. The phreato-
phyte cover described by AU and Cl. is deciduous and
a distinct seasonal trend exists due to spring foliation
and fall defoliation. Leaves are the primary evapora-
tive surfaces of a plant and the leaf area is directly

 

AVERAGE MONTHLY k p

 

 

 

_041 I | | I I I I I | l |

 

Jan Feb Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.
EXPLANATION
o kpBC A kaH
O kpRS A kaAN

FIGURE 16.—Seasonal variability in average monthly values of
the coefficient k p.

 

P39

related to the area of air-water interface provided that
moisture is available to the leaves. The seasonal
changes in the physiological condition of the plants
are also involved, not only in the production and
maintenance of the leaves, but also in the process of
extracting moisture from the soil and conveying it to
the leaves. The availability of moisture to the roots of
the plants is not a significant factor for the phreato-
phytes in this study because of the relatively constant
level of the water table under the flood plain. The
plotting of average monthly values of k p, from tables
12 and 17, in figure 16 illustrates the seasonal vari-
ability. The gradual development of foliation begin-
ning in April and continuing through May and June
is indicated by the increase in kp. During June
through September kp is relatively constant for all
coefficients except kaH. The fall dormancy and ulti-
mate defoliation is defined by the reduction in kch
and kpR near the end of the year. The symmetrical

shape of the graphs of 12ch and k file which corresponds

to the seasonal location and duration of the growing

season on the project area, indicates that fBC and fR

provide a better measure of the seasonal variations in

climate than do fJH and fPAN which produce un-

explained high values of kaH and kaAN during

October, November, and December. The preceding

rationalization can be supported by comparing the

shape of the Ian graphs to the seasonal variability of

foliation as obtained by field inspection and by inter-
pretation of aerial infrared color photography.

 

IIIIIIIIII__I IIIIIIIII

IIIIHIII
mo

     
  
  

‘1

  

I

80—

IILLTIJ‘L

60—

40—

1_,_1__1_1_1__<_,_____¢—

  

PERCENT OF ANNUAL TOTAL

20—-

 

 

‘_,4_1

 

i .'
IIIIIIIIIII

I
I
I v
0— d d
IIIIIIIII IIIIIIIIIII
JFMAMJJASONDJFMAMJJASONDJFMAMJJASOND
1963 1954

EXPLANATION

Saltcedar

1965

—o——

..-___ Mesquite

FIGURE 17.—Field estimates of seasonal variability of foliation.

P40

 

oo
o
I
l

RELATIVE NEAR-INFRARED
IRRADIANCE, IN PERCENT
m
o
I

 

 

 

 

 

 

EXPLANATION
AVERAGE
HEIGHT (FT) n ACRES
3.6 —— —— 34 - —— 125
8.5 166 608
11_4 ...... 51 ....... 137

FIGURE 18.—Seasonal variation of adjusted red transmittances
obtained from Ektachrome—IR images of saltcedar forest, Gila
River, Arizona. Data points are mean density values for forest
plots representing three different foliage volume classes. n =
sample size. (After Turner (1971, figure 7)).

Figure 17 shows the estimated foliation as percent
of total annual for 1963, 1964, and 1965 on the project
area as obtained by field inspection. The develop-
ment of leaves begins at least one month earlier on
saltcedar than on mesquite. Defoliation of saltcedar
occurs after the first frost while mesquite retains its
leaves for another month even though the foliage
may be dormant and ineffective as an evaporative
surface. The graphs in the figure indicate a variation
in the seasonal distribution of foliation from year to
year. The period of significant foliation for saltcedar
typically extends from April through October which
corresponds to the period of relatively high values of
kp.

Color-infrared photographs of vegetated areas can
be used to derive relative measures of foliation
(Turner, 1971). Beginning in 1967 aerial photographs
using color infrared film were taken of the project
area at frequent intervals. Figure 18 shows the 1968

area. Figure 18 was described by Turner (1971) as
follows: “The increase in red transmittance from
March 22 to April 5 was in response to spring branch-
let growth. The sharp reduction in values between
April 19 and May 3 reflects a frost on April 20 which
caused partial defoliation. New growth soon restored
this loss and the transmittance values increased

 

GILA RIVER PHREA’I‘OPHY’I‘E PROJECT

abruptly in response. The values slowly declined
after the maximum of late August as the slow
autumnal defoliation typical of the species took
place.” The shape of the graphs and indicated dura-
tion of foliation in figure 18 correspond to the shape
and seasonal extent of relatively high values of kp.
Jones (1977) provides additional confirmation for the
relation between transmittance on color—infrared photo-
graphy and ET by relating the 1968 photographic
data for reach 1 and 2 to the Blaney-Criddle consump-
tive-use coefficient “k.”

Both the climatic factor f and the consumptive-use
coefficient for phreatophyte cover, kp, have a wide
range of seasonal variability as shown in figures 15
and 16. The value of f kp is the difference between an
area having a 100 percent areal density of phreato-
phytes and that from an area of no phreatophytes.
The seasonal variability of the product f k p is shown
in figure 19 for each of the four methods. The dif-
ferences in the response of f to seasonal changes
when fitted to the variable budget-period ET’ data by
optimizing the consumptive-use coefficients, kp, pro-
duce the differences in the product f kp. The most
significant differences in f k p occur during the grow-
ing season in August and September, the months
with the fewest and most erratic ET’ data.

Computations similar to those shown in table 13
were made using climatic factorsz, f H, and f M N with
appropriate coefficients, and the average annual
values of U are shown in table 18 and compared with
annual values from table 13. The average U from all
three reaches was 32.38 in. (822 mm) before clearing
and the greatest departure from the average was
minus 8 percent for U M N. After clearing, the average
U was 12.48 in. (317 mm) and the greatest departure
from the average was plus 10 percent for UBC. The
salvage of water, computed as the difference between
the average U before clearing minus the average U
after clearing, is 19.90 in. (505 mm).

By varying the monthly values of kp and annual
values of k“, the optimization program will fit U to
E T’ with equal success for all of the climatic factors.
However, the monthly values of km”, and kaAN for

. . . . . . 1 TABLE 18.—~Annual eva otrans iration com uted or each reach
seasonal variability in denSItometrlc data from photo- . p p p f

graphs of selected areas of saltcedar on the project .

by the BlaneyvCriddle (BC), Solar Radiation (R), Jensen and
Haise (JH), and Pan evaporation (PAN)

[All values exclude precipitation and are in inches per year]

 

Preclearing Postclearing

 

Method 1 2 3 l, 2, 3

 

39.45
39.97
41.26
38.88

13.79
12.90
11.16
12.07

 

 

39.89 12.48

 

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P41

 

 

7 I I I I I I I I I I I
n
— 150
——- I40
— 120
U)
m >-
3 "— 100 E
a a
CC
E — a“
m E
E — 80 E
Z — E
g 2
. E
s _ Z
r— d.
50 i
_— 40
— 20
—— 0
I I I I I I I I | I I

 

 

 

Jan Feb. Mar. Aprv May June July Aug. Sept Oct, Nov, Dec.
EXPLANATION
° fBCkpBC AfJHkaH
’ f R kpR ‘ fPANkp PAN

FIGURE 19.——Seasonal variability offkp, the difference in U between an area having a 100 percent areal density of
phreatophytes and an area having no phreatophytes.

October through December appear unreasonable, as
previously mentioned. Obviously, nine years of ET
data at one location do not provide an adequate test The results of the Gila River Phreatophyte Project
of the empirical equations used in this study. can be compared with data from other studies by use

COMPARISON OF RESI'LTS WITH OTHER STUDIES

P42

GILA RIVER PHREATOPHYTE PROJECT

TABLE 19.——Deriuation of consumptive-use coefficients for Cottonwood Wash during growing season March through October

 

 

 

 

MAR. APR. MAY JUNE JULY AUG. SEPT. OCT. Average for
growmg season

Upper reach )3

Average 1959—63 . .39 .38 .73 .96 1.10 1.23 1.15 .61 .32

Lower reach [it

Average 1961563 ..10 .83 .56 .44 .46 .50 .55 .43 .48

KwAw. .. .. . .29 1.45 .17 .52 .64 .72 .60 .13 .34

k . 1.47 12.29 .86 2.64 3.26 3.66 3.05 .92 1.7

(‘u'

 

of the derived empirical equations. Temperature and
solar-radiation data for evaluating climatic factors
can be obtained for the various sites used in the
comparison. The coefficients he in equation 13 must
be selected on the basis of soil characteristics and
depth to ground water and k p on the basis of species
and quantity of vegetation. Differences in phreato-
phyte species and in methods of quantifying vegeta-
tion data require modifications in the methods for
applying the coefficients. These modifications will be
illustrated in the following comparisons with two
other previously conducted flood-plain studies.
Water use by riparian vegetation on the flood plain
of Cottonwood Wash in northwestern Arizona was
reported by Bowie and Kam (1968). A 4.1-mi (6.6-km)
reach of the stream channel was divided into a 2.6-mi
(4.2km) upper reach and a 1.5-mi (2.4-km) lower
reach with flood-plain areas of 29 acres (11.7 ha) and
22 acres (8.9 ha), respectively. ET from these reaches
was measured by the water-budget method during
the growing season for the period 1959—63. The flood-

 

plain vegetation, as described by Branson and Aro
(Bowie and Kam 1968), consisted primarily of mature
cottonwood trees (average height 27 ft [8.2 m]) and
red willow trees (average height 19 ft [5.8 m]) dis-
tributed as individuals or clumps over the flood plain.
Depth to water table on the flood plain ranged from
2.5 to 3.0 ft (0.8 to 0.9 m). The quantitative measure-
ments listed by Bowie and Kam (1968) give a total net
canopy cover of 5.7 acres (2.3 ha) on the upper reach
and 5.9 acres (2.4 ha) on the lower reach. This
measure is described as the equivalent part of the
flood plain actually covered by vegetation. The vege-
tation in the lower reach was defoliated in June 1960
and eradicated in February 1961. No change was
made on the upper reach.

The results of the monthly water-budget measure-
ments of ET presented by Bowie and Kam (1968,
table 7) have been reduced to inches per month on the
flood-plain area and plotted in figure 20. The monthly
values ofET for the two reaches were similar in 1959.
A moderate reduction for the lower reach in 1960 is

TABLE 20.—Application of vegetation description from Gatewood and others (1950, tables 7 and 8) to empirical equations
[Area in acres]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thatcher Glenbar Fort Thomas Black Paint Thatcher
to to to o o
Glenbar Fort Thomas Black Paint Calva Calva
Total gross area ........................ 2,159 2,011 1,818 3,315 9,303
Saltcedar
Gross area. 1,302 1,426 852 1,002 4,582
0.603 0.709 0.469 0.302 0.492
72.2 63.7 55.0 54.3 62.4'
0.454 0.479 0.279 0.178 0.326
Baccharis
279 266 202 764 1,511
0.129 0.132 0.111 0.230 0.162
46.2 26.3 38.8 27.9 32.41
0.056 0.042 0.049 0.076 0.061
Mesquite
54 43 263 624 984
0.023 0.021 0.145 0.188 0.106
50.3 57.6 61.1 40.9 47.6'
0.014 0.013 0.094 0.087 0.056
Total of saltcedar, baccharis, and mesquite
V ...................................... 0.534 0.534 0.422 0.341 0.443
Cottonwood and willow
Area at 100 percent
volume density ........................ 31 16 60 73 280
(“V .................................. 0.061 0.008 0.033 0.022 0.030

 

AU : fraction of total gross area covered by the species

Cl, : areal density in percent

V= AUICU/IOO + (CU/100)“? /2.

AM- : fraction of total gross area covered by Cottonwood and willow.
' weighted average for the four individual reaches.

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

 

1

N

IIIIIIIIIH IIIIIIHIII IIIIIIIIIIIIIIHIIHHI HIIIII 300
Defolietion on lower reach June 1960

reach February 1961

Clearing on lower
/ T— 250

  

_— 200

‘— 150

‘— 100

  

'—— 50

INCHES PER MONTH ON TOTAL FLOOD-PLAIN AREA

   

.5
lllllllllll I'Illlllll llfllllllllll
1961 1962 1963

EXPLANATION

0 Upper reach

 

 

 

 

 

MILLIMETERS PER MONTH ON TOTAL FLOOD-PLAIN AREA

 

 

 

6 Lower reach

FIGURE 20.—Evapotranspiration from Cottonwood Wash.

apparent following defoliation and a drastic reduc-
tion following the eradication of vegetation on the
lower reach is reflected in the ET data for 1961-63.

Monthly values of the Blaney-Criddle expression
for climatic factor “1‘” were computed using temper-
atures for the site as given by Bowie and Kam (1968,
table 8). Temperatures for periods of missing record
at the site were estimated by linear correlation from
temperatures recorded at Kingman, Ariz., 30 mi (48
km) northwest of the site (National Weather Service).
Monthly coefficients were computed as k = E T/f and
plotted in figure 21. The differences between the two
reaches, resulting from defoliation and eradication
on the lower reach were expanded and a basis for
extrapolation of data was developed. The flush of
spring transpiration from replacement vegetation on
the lower reach is assumed to have produced the
relatively high values of k for April and May in 1962
and 1963.

Depth to ground water on the Cottonwood Wash
flood plain as recorded at observation wells was less
than 3 ft (0.9 m) on both reaches. The soil character-
istics were also similar on both reaches. The consump-
tive-use coefficient for no phreatophytes can therefore
be assumed to be equal for both reaches. Equation 13
as applied to Cottonwood Wash can be stated as

k : k0 + kchcw (19)

where
k = consumptive-use coefficient for a

reach,

 

 

 

 

 

 

 

     

 

P43
1-5 lllllllllll HHIHHH HIIHIHH lllllllllll Hlllllll.
Defcliation June 19'60 lower reach
1'4 T / (Clearing February 1961 lower reach
1.2 1
I“
ll
51
I .‘l o
1 11 ’
1'0— 1 :, /
1 1
r | \
: : u
I 1
. 1
4: _3~ i 1
l 1
a J
1 l
.6— I
E
l
l
‘ l
l
l
l
l
l
l
l
l

 

 

1 if
n 1 1 l
O lllllhllllllll llllllllllll‘llll lllbllllllll

1959 1960 1961 1962 1963

EXPLANATION

0 Upper reach

k=%uwhere f=1%t0

o Lower reach

FIGURE 21.—Values ofk from Cottonwood Wash.

k0 = consumptive-use coefficient for no
phreatophytes,

km. : the increase in the consumptive-use
coefficient for the area under a
canopy of cottonwood or willow, and

Am. = the fraction of the total area under

cottonwood or willow canopy.
Average values of k for the lower reach during
1961—63, after the phreatophytes were eradicated are
equal to k0 for both reaches because Am was zero for
the lower reach. The value of km. is computed by the
transformation of equation 19 from the values of k for
the upper reach as

k _ k - k0
cw _ ACLU
where
_ canopy cover of the upper reach _ 57 acres _

 

0.197.

cw T

total area of the upper reach 29 acres

Monthly and average monthly values of k, kg, and km
for the growing season are listed in table 19. The
seasonal variation of km is similar to the seasonal

P44

variation ofkp except for the high value in March and
the large negative value in April resulting apparently
from unusual transpiration requirements by the re—
placement vegetation. The erratic values for the
spring months do not have a significant effect on the
average monthly value for the growing season.

The average value of 1.7 for k,.,,, as shown in table
19 is 3.2 times the Gila River project area k1» value of
0.52 (see monthly [2,, values in table 12) during the
period March to October. Mature cottonwood and
willow in open stands, where each tree is an indivi-
dual oasis, provide ideal conditions for transpiration.
Rantz (1968, fig. 2) indicates a value of 1.6 for km. in
the Blaney-Criddle equation for cottonwood and
willow with a depth to water table similar to the
Cottonwood Wash flood plain. The coefficients deriv-
ed from Cottonwood Wash data are used later in this
report for comparing kp from the Gila River project
with data from a previous study of the Safford
Valley, also dominated by saltcedar but containing
some cottonwood and willow.

The water use by bottom-land vegetation in the
lower Safford Valley, Ariz., was reported by Gatewood
and others (1950). The study reach extended from
Thatcher to Calva and included reach 1 of the Gila
River Phreatophyte Project. The draft on ground
water (identical to E T’ in equation 11) was measured
during the period October 1, 1943 to September 30,
1944 by six different methods described as tank,
transpirationvwell, seepagerun, inflow-outflow, chloride
increase, and slope-seepage for four reaches and the
sum of the four individual reaches. These data are
compared with data from the Gila River Phreato-
phyte Project by the following method.

The description of vegetation by Gatewood and
others (1950, table 7) includes an average areal den-
sity for each species of phreatophyte. These data are
used to evaluate the vegetation description “V” as
shown in table 20. Saltcedar, seepwillow, and
mesquite are assumed to be equivalent to the phreato-
phytes whose transpiration is defined by the coeffi-
cient hp in equation 13. The method used in comput-
ing average areal density in the Safford Valley pro-
duces a different value of V from that of the summa-
TABLE 21.—-Comparison of euapotranspiration computed by

empirical equations with the measured draft on ground water
presented by Gatewood and others (1950, table 58)

Area ET’ U

Difference
(acres) ET’ , U

 

 

 

Reach (acreft) (inches) (inches) (inches) percent
(1 ' 1“" ’
_ fir.—
Thatchor to
(ilcnbar 2,159 7,420 41.24 315.41 5.821 ‘ 14
(ilenbar to
Fort Thomas 2,011 5,810 114.67 29.13 5.24 ~15
I‘ori Thomas in
Black Paint 1,818 4,700 31.02 29.06 1.96 . 6
Black l’ninl to
(‘lava 21,315 5,030 18.21 25.45 7.24 “10
'1'halt'her to
(‘alva 9,303 22,960 29.62

29.31 .511 v 1

 

GILA RIVER PHREATOPHYTE PROJECT

tion for each density class as used in equation 14,
although the difference is relatively insignificant.
The estimation of transpiration by cottonwood and
willow is determined by application of the coefficient
km. from the previously described Cottonwood Wash
study. The areas listed as having 100 percent volume
density are assumed to be equivalent to that part of
the flood plain actually covered by vegetation used in
Bowie and Kam (1968), and are used to determine
ACW.

The only available monthly values of E T' for the
Safford Valley reaches were measured by the inflow-
outflow method; these values and the mean daily
maximum and mean daily minimum temperatures
for computing average monthly temperatures are
listed by Gatewood and others (1950, tables 4, 47, 48,
and 49). Monthly values of U were computed by the
Blaney-Criddle method using equation 12 with the
coefficient k evaluated as

k : k() + kl} V + k(‘lL‘A('l(' (20)

where
k,, = 0.21 from table 12 (soil type and
depth to water table for the Lower
Safford Valley are assumed to be
similar to those on the Gila River

Phreatophyte Project),

 

I I | 1 1 I | l I l 1
Glenbar to Ft. Thomas

— 200

    

— 150

 

 

INCHES. PER 30 DAYS
MILLIMETEHS, PER 30 DAYS

 

 

-100

 

 

 

 

1
Mar. Apr, May June July Aug. Septl Oct. Nov. Dec.
1944 1943

EXPLANATION

—°— Measured

Jan. Feb.

- “+ ‘ ‘ Computed

FIGURE 22.——Relation of measured to computed evapotranspira-
tion for reaches in Safford Valley, 1943—44.

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

k p = monthly values in table 12,
= numerical vegetation description from
table 20, and
km = monthly values from table 19, and
Ac“, from table 20.

Measured ET’ and computed U are shown in figure
22 for October-December 1943 and January-
September 1944. Data for 1944 are plotted on the left
of 1943 data to show the monthly variability during a
calendar year. Computed monthly values for the
Glenbar to Ft. Thomas reach are lower for the period
March—August and higher during September—
December, than the measured values. The values for
total annual ET’ and total annual U are 29.2 in. (741
mm) and 29.4 in. (747 mm), respectively. ET’ is higher
than U for the entire growing season on the Ft.
Thomas to Black Point reach; annual totals are 37.3
in. and 28.5 in. (947 mm and 723 mm), respectively.
On the Black Point to Calva reach, the relationship is
reversed with U being higher than ET’ for the entire
year. Annual totals are 14.7 in. and 25.1 in. (373 mm
and 638 mm) for ET’ and U, respectively. Differences
between monthly ET’ and U range up to 126 percent
of ET’ for September on the Glenbar to Ft. Thomas
reach, which indicates the possible error in monthly
data for an individual reach measured by the inflow-
outflow method when the net ground inflow was not
computed for individual months.

The average values of annual ET’ determined by

 

12 I I I I I I I I I I I

Bernardo—average for tank 5
dunng 197I and I972 /'\

 

 

 

 

INCHES, PER 30 DAYS
:3

 

MILLIMETERS, PER 30 DAYS

Buckeye—average for tanks 2 and 6
3 ‘ during I962 and 1983A -_ 200

‘- 150

‘* 100

    

“— 50

 

 

[J r‘I"I I I I I 1 1 I l |
Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.

EXPLANATION

- - +-- Measured

 

-—°—' Computed

FIGURE 23.—Relation of measured to computed evapotranspiration
for evapotranspirometers at Buckeye, Ariz., and Bernardo, N.
Mex.

 

P45

the six methods listed by Gatewood and others (1950,
table 58) are shown in table 21. Annual computed
evapotranspiration U was determined by the applica-
tion of equations 12 and 19 with the coefficients
evaluated as

k, 0.21,
annual totalfkp = 30.70 = 0.467
annual total f 65.68
(average annual value from table 12),
km. = 1.70 (table 19) + 0.11 (average kp for
November, December, January, and
February from table 12), and
Aw. from table 20.
The differences between ET’ and U are listed in table
21 and show only a +1 percent difference from ET’ for
the combined reaches but a -40 percent difference for
the reach between Black Point and Calva.

Phreatophytes have been grown in evapotranspiro-
meters at various sites throughout the United States
to provide accurate data on the consumptive use of
water (evapotranspiration) by the vegetation. Data
from two sites, one near Buckeye, Ariz., and the other
near Bernardo, N. Mex., have been selected for com-
parison with the results of the Gila River Phreato-
phyte Project because some of the evapotranspir-
ometers at these sites had the combined features of
relatively large area, dense saltcedar, and depths to
water table approximating the depth to ground water
on the Gila River flood plain.

Six evapotranspirometers, 900 ft2 (84 m2) in surface
area and 14 ft (4.25 m) deep, were installed and
planted with saltcedar at Buckeye in 1959 as describ-
ed by van Hylckama (1974). Monthly water use,
exclusive of rainfall, (corresponding to ET’ on the
Gila River project) for tank Nos. 2 and 6 during 1962
and 1963 was used for comparison in this study. The
depth to water table in these tanks was maintained
at 8.9 ft (2.7 m) and the areal cover density of canopy
was 80 percent in tank No. 2 and 75 percent in tank
No. 6.

Nine evapotranspirometers, 12 ft (3.7 m) deep with
a surface area of 1,000 ft2 (93 m2) were installed by the
US. Bureau of Reclamation on the Rio Grande flood
plain near Bernardo in 1962. Saltcedar, Russian
olive, and saltgrass were planted in the tanks and the
water table was maintained at various levels. Water-
use data for tank No. 5 during 1971 and 1972 were
selected for comparison. Depth to water table was
maintained at 9.0 ft (2.7 m) and the saltcedar in this
tank had an areal density of 92 percent in 1971 and
97 percent in 1972. The monthly water-use data for
comparison with U and Buckeye consumptive use
were obtained by subtracting the precipitation listed
in US. Bureau of Reclamation (1973, table 1) from
the water use for tank No. 5 listed in US. Bureau of

 

P46

Reclamation (1978, table 15).

Equations 12, 13, and 14, using coefficients kp and
k0 from table 12, were applied to the Buckeye and
Bernardo evapotranspirometer sites to provide values
of computed U for comparison with the measured
evapotranspiration exclusive of precipitation. The
results are shown in figure 23. The graph of measured
ET’ for Buckeye shows greater values during April
through June than the computed U. The reduction in
ET’ during July and August was observed at all
tanks at Buckeye and is attributed to extremely high
temperatures, up to 115°F (460C), and to excessive
convected heat from the surrounding desert, creating
moisture stress in the plants and reducing tran-
spiration (van Hylckama, oral commun., 1976). The
difference between ET’ and U is 102 percent of U in
July and the annual totals are 36.09 in. (917 mm) for
ET’ and 41.67 in. (1,058 mm) for U, a difference of 15
percent.

The measured ET’ for the Bernardo evapotran-
spirometer, as shown in figure 23, is primarily con-
fined to the period May through September with June
through August averaging 27 percent greater than
the U values. The annual total is 33.58 in. (853 mm)
for ET’ and 36.94 in. (938 mm) for U, a difference of 10
percent. Minimum temperatures at Bernardo were
freezing or below during November through April,
which caused the low ET’ values for these months.

The graphs in figure 23 illustrate the limitations of
equation 12, and its application to this study, with
regard to describing evapotranspiration for wide
ranges in climate. Neither the effect of high temper-
atures nor below freezing conditions are adequately
defined to provide monthly averages.

EFFECTS OF PHREATOPHYTE CLEARING
ON GROUND-WATER LEVELS AND
SEEPAGE MEASUREMENTS

GROUND-“HATER LEVELS

An ' increase in ground-water elevations can be
expected as a result of eliminating water withdrawal
by phreatophytes. Ground-water levels measured in
the observation wells on the Gila River flood plain
are primarily controlled by the stage and discharge
in the Gila River channel. Annual and seasonal
variability in the flow of the river obscures the effects
of water use by phreatophytes on ground-water eleva-
tion. Therefore, periods of similar river discharge
before and after clearing were selected to illustrate
the differences in ground-water levels. Discharge
during the period February through July of 1964
before clearing and 1969 after clearing were reason-
ably similar as shown in figure 5. Water-table eleva-

 

GILA RIVER PHREATOPHYTE PROJECT

tions during these periods for flood-plain wells 0517
and 0720 (see figure 6) are shown in figure 24.
Elevations at well 0517 were higher in 1969 than in
1964. The rate of recession was similar for both years
until the middle of May when water use by phreato-
phytes produced an increased rate of recession in
1964. Increased discharge in the Gila River from
summer storms after July 15 of both years terminated
the ground-water recession. Water-table elevations
were higher at well 0720 in February of 1964 than in
1969. There was no flow in the Gila River channel in
reach 1 from June 28 to July 14, 1964, whereas, the
minimum inflow during this period of 1969 was 2.7
ft3/s (0.08 m3/s) and outflow was 4.0 fth (0.11 m3/s).
The graphs in figure 24 indicate that the removal of
phreatophytes reduced the rate of recession in ground-
water elevations but the maximum difference in
elevations before and after clearing was less than 1 ft
(0.3 m).

SEEPAGE MEASI'REMENTS

Discharge measurements of the flow in the Gila
River channel were made at about six-week intervals
to observe the interchange of surface and ground-
water flow as described by Burkham in Culler and
others (1970, p. 14). Measurements were made at each
cross section on the same day during periods of
uniform flow by two or three stream gagers. These
essentially simultaneous measurements were made
on 53 dates during the term of the project. The results
of measurements taken on seven dates were selected
to represent the range and variability in discharge
for before clearing (fig. 25A) and after clearing (fig.
253).

The flow in the river channel at any cross section is
affected by the subsurface conveyance; that is, the
depth, width, and transmissivity of the alluvium and
the slope of the water table. Differences in channel
flow between cross-sections reflect not only differ-
ences in subsurface flow but also contributions to or
depletions of water from the reach of flood plain
between cross sections. Figure 8, based on complete
water-budget data, indicates that the Gila River chan-
nel was a losing stream before clearing and a gaining
stream after clearing. The graphs in figure 25 tend to
confirm this characteristic of the channel flow, al-
though the data for certain dates, such as May 2,
1965, before clearing and May 18, 1971, after clearing,
are contradictory. Changes in subsurface storage
undoubtedly account for the variability in the gain or
loss characteristics of the river. The only information
provided by these measurements is that the relation
between surface and subsurface flow is reasonably
constant from cross-section 1 to cross-section 17 in

P47

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

GZSL :IO OADN EIAOEIV SHEIJBW NI 'NOILVAS'IEI EI'IEVL'HELVM

.mwtmaaofiwwwaa MEEEQ Sim UGM whomon wfiofim>3m $33.33? mo cemimnEoO|.wN $565

 

 

 

>43w mzaw ><§ amm< Iom<2 >m<3mmmu wwmw
_ _ _ _
0.055 I
. ‘ogi’k’8:o)o
N ORR Iva. coco-o amp
I coon-Woo: II thN
coco- CNN :w>>
cocooooOooooKuSoooooSooooooooooooooo on tool:
82\ 9332302535
V.Ohfi I Icoomoaooooo ooao
Into-nonchoo: oooooadaooooIloooo
0.0.0. o oaollooooocccooacc
II I. conga-ooooocooouuo-ooo!on: to cocoaeeoeoooonn mNmN
no . o‘nocooioho-nflv (M I
0.955 I :- 0
man oRSSMuwmuuoom
wdhh I
_ _ _ _ mNmN
mmmN
_ _ _ _
odhh I
Néhh [I .1 OVmN

Yet. I

9th |

 

wénn

:m __w>>

 

3mm

 

,NVmN

 

GZGL :IO OADN EIAOHV J.E|3:I NI 'SNOIlVAEHH EI'ISVL'HEILVM

GILA RIVER PHREATOPHYTE PROJECT

P48

'EIEJHVHSSICI

UNOCES 83d SIHEW OIBIIU NI

mo I

Amv Marga? Eda 98 A3 Manda? 98me 3:550 33%

 

   

 

    

 

 

 

 

m
E; B 82585 2_ 22: 8o: 2,52 5255
mm a om 8 an 8 S o
_ _ _ _ _ fi
955% $qu
t 2 2 : m N m m _
I IIIfiIIIa IIIIIII fiIIIIfiIIIIqIIIIJ IIIII fiII_I
I RE .2 Eamamm H
I o/O'IIO/O|\\I\Z\\O|III\\’/|
E: £52
I E: I; 22 I
I. %I
I :2 .: 5.22 I
I II \OI’I I\0I
/\\II III.\ 1
$2 .: .8588
FIILIIIIF IIIIIII HIIIIrI -IILIIIILIII—ILIIL
_ * _ _ . _
_ _ fl _ _ _ _ _ _
a e 2 E 2 S m w v N a

mEEE mo moz<mpoi z_ 2:); mac: @202 “oz/‘55

.mwazsnofiwwza

«:0 m5 5 3:052:33: mMRAomE waomnmfizfimm mo mfismmMIhN ”.5505

am

om

co
8

Dow

'3EJHVH38I0

ONOOBS EH 1331 DIEII'II] N|

GNOIJEIS 83d SHEIEIW JIBIIC) NI 'EIEJHVHCJSICI

<
EM: “5 $2585 2_ 22‘: good wzo._< mqumE

 

 

           

  

            

     

 

 

     

 

 

mm 8 B S 8 a S o
_ _ _ _ _ _
wzgzmm W85
2 m: 2 : m N m m P
I JIJI
S I n _
I _ _I
_ _
I _ _I
_
I _ _I
I _ _|
I _ T
I _ |
_ u
_
_ N
N 2 .2 >22
3 I n
I _ I
n _ £39 225 _
_
_ _ _ _ _ _ _ _
I _ _ £2 .2 52 I
_
F | _
I _ M ma: .9 $532 _ _ I
w
I , _ _ _ I
~ _ _ _ _ _ _ _ _
I . ~ _ _ _ _ _ _ *1
_
I _ _ ._ _ .II
I u _ $27: 52552 _ n _ A _ . I
I _ . _ l
u _ _ _ _ n . _ _
_ _ _ _ . _
_ w . _ —
_ . am: i E2
IIIILIIIIL IIIIIII L|||lI_. IIIII .IIIIIFIIIILIIL
m I .
a _ _ _ _ _
_ . _ a _ _ . _ _ _
ON a 2 3 a E m w v N a

mEEE mo 32330:.— z_ 233 good @202 $2456

 

E

ow

am

3
om

09

SN

UNODBS Hid 133i UIBIIIJ NI 'EIEIHVHOSICI

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

the project reach. Continuous records of the complete
water budget are necessary to accurately evaluate the
relation of one component to another.

CONCLUSIONS

Annual evapotranspiration (ET), including precipi-
tation, on the project area averaged 43 in. (1,090 mm)
before clearing. Annual ET ranged from 56 in. (1,920
mm) in dense stands of phreatophytes (100 percent
areal coverage) to 25 in. (630 mm) on areas of no
phreatophytes. The removal of phreatophytes result-
ed in a reduction in ET averaging 19 in. (480 mm) per
year. The reduction ranged from 14 in. (360 mm) on
reach 1 to 26 in. (660 mm) on reach 3, a difference
attributed to the difference in density of the phreato-
phytes. This reduction is temporary because replace-
ment vegetation was not established. ET after clearing
consisted of evaporation from bare ground and tran-
spiration from annual vegetation. An estimate of the
permanent reduction can be obtained by comparing
the ET before clearing with the consumptive use of
possible replacement vegetation. Erie, French, and
Harris (1965) measured the consumptive-use require-
ments for optimum crop production of various irrigat-
ed grasses near Tempe and Mesa, Ariz., and computed
semi-monthly values of the coefficient “k” in the
Blaney-Criddle equation. The application of these
coefficients to the values ofthe Blaney-Criddle clima-
tic factor for the Gila River flood plain provided
annual estimates of 69 in. (1,750 mm) for alfalfa, 49
in. (1,240 mm) for blue panic grass, and 42 in. (1,070
mm) for a Bermuda grass lawn. The consumptive use
for alfalfa exceeds the maximum observed ET; that
for blue panic grass use exceeds the average ET; and
for Bermuda grass, use is only 1 in. (25 mm) less than
the average ET. According to these estimates, there
would be no significant salvage of water if any of the
grasses were established on the entire area, if they
maintained optimum production, and if their roots
extend to the capillary fringe of the water table.
Selective clearing of areas of dense phreatophytes
converted to blue panic or Bermuda grass would
provide a salvage of 7 in. (178 mm) and 14 in. (360
mm), respectively, from these areas. Because the
average depth to ground water exceeds 8 ft (2.4 m) on
the project area, it can be postulated that the con-
sumptive use of the grasses would be less than under
irrigation, crop production would be less than opti-
mum, and more water would be salvaged. No data
are available from this study to prove or disprove this
postulation. A flood plain without phreatophytes is
in an artificial condition, and the water requirements
for maintaining this condition are dependent on the
land-management practices applied. The maximum

 

P49

possible salvage for sites similar to the Gila River
flood plain, as observed in this study, is 31 in. (790
mm) for areas of 100 percent area coverage of phreato-
phytes converted to no permanent vegetation.

The preceding data were obtained by computing
ET as the residual in a water-budget equation, involv-
ing twelve measured components, consisting of all
inflow and outflow of water. Four contiguous reaches
of the flood plain were studied and measurements of
ET were obtained for budget periods of two or three
weeks between 1963 and 1971. The accuracy of the
ET data is dependent on the quantity of water
measured as inflow and outflow; the average annual
ET for an individual reach was only three percent of
the average annual quantity of water moving through
the reach before clearing and one percent after clear-
ing. Thus, errors in the water budget can completely
obscure the ET values. Fortunately, however, maxi—
mum rates of ET do not generally coincide with
maximum rates of flow and ET is a significant
component of the water budget for many budget
periods. Arbitrary criteria based on consistent and
unbiased rules were established for rejecting all
obviously erroneous data. The errors in each compo-
nent and in the total budget were evaluated and the
maximum potential evapotranspiration for before
and after clearing was computed. Acceptance criteria
based on the measurement errors and potential evapo-
transpiration were used to establish acceptable maxi-
mum ET values and maximum errors in these values.
Minimum acceptable negative ET values were also
established. Applying these tests to the water-budget
evaluations provided 321 acceptable ET data.

Accepted data were too few and their distribution
too irregular to define ET accurately for any individ-
ual reach during a particular year. The ET data were
also spatially variable before clearing because of
differences in the density of phreatophytes on the
various reaches and temporally variable because of
seasonal and annual differences in available energy
and atmospheric conditions. In order to combine data
from all reaches and to compensate for this spatial
and temporal variability, four previously developed
and widely used empirical ET equations were fitted to
the accepted ET data. The equations provide a clima-
tic factor that compensates for differences in solar
radiation and temperature. This factor was used to
derive monthly coefficients for each equation related
to the areal density of phreatophytes. The following
equations or data were used to define the climatic
factors: (1) the Blaney-Criddle equation based on the
monthly percentage of total daytime hours in the
year and mean temperature; (2) solar radiation; (3)
the J ensen-Haise equation based on solar radiation

P50

and mean temperature; and (4) pan evaporation.
Coefficients for no phreatophytes and for varying
densities of phreatophytes were derived by fitting the
climatic factor to the data by use of an optimization
program. Optimum fitting was achieved when the
average difference between measured and computed
ET for all accepted budget periods was minimized.

The average standard deviations of the annual
computed ET from the measured ET indicate a vari-
ability of i4.6 in. (117 mm), or i15 percent, before
clearing, and i3.2 in. (81 mm) or :25 percent, after
clearing. The deviations indicate an error in the
computation of monthly rates of ET ranging from a
low of 30 percent in summer before clearing to 150
percent in winter for both before and after clearing.
These statistical tests of accuracy for fitting mea~
sured ET t0 the various equations for the climatic
factor indicate no significant difference in the accu-
racy of prediction among the equations. However,
seasonal variation of the coefficients for both the
Blaney-Criddle equation and for the solar-radiation
equation is similar to the seasonal variation of folia-
tion based on field estimates and on repetitive
infrared-color aerial photography. In contrast, the
variation of coefficients for the Jensen-Haise and
pan evaporation equations differ considerably from
the observed variation of foliation. Thus, if it is
assumed that the seasonal variations in the monthly
coefficients for phreatophyte cover are due to seasonal
variation in leaf area, the Blaney-Criddle and solar-
radiation equations must be considered to be superior
to the other two.

The empirical equations with coefficients derived
from this study can be used to estimate ET and water
salvage for other areas. Annual coefficients for no
phreatophytes and monthly coefficients for varying
densities of phreatophytes for each of the four climatic
factors are listed in tables 12 and 17. The value of the
coefficient for no phreatophyte describes evapotran-
spiration maintained by the upward movement of
water from subsurface sources and is related to the
soil type and depth of the ground-water table. These
features should be considered in projections to other
areas. The coefficient for varying densities of phreato-
phytes is primarily related to the quantity and con-
dition of foliation, which in turn is related to the length
of growing season. Coefficients for the transition
months, such as May and October, should be in-
creased for growing seasons longer than this season
on the Gila River flood plain, or reduced if the
growing season is shorter. Average values of the
coefficient for the year, or for the growing season,
may provide adequate estimates for many purposes.
The coefficients for both no phreatophytes and for

 

GILA RIVER PHREATOPHYTE PROJECT

phreatophyte cover were derived from data which
excluded precipitation, therefore, the local precipita-
tion should be added to obtain estimates of total ET.
The application of the coefficients from this study to
areas other than the Southwestern States may pro-
vide erroneous estimates.

Usable methods were developed by this study for
comparing ET from reaches having different quanti-
ties of vegetation. However, methods of obtaining the
quantitative description of vegetation as related to
transpiration, in particular, should be improved. A
rational interpretation of the empirical equations
used indicates that the climatic factor is an index of
potential evaporation and, therefore, the vegetation
description should be an index of the area of evapora-
tive surfaces, which for deciduous trees is the seasonal-
ly variable foliation. The vegetative measures used in
this study were based upon the canopy, which is a
function of the species present and the habitat. As
such, the canopy is an integration of the growth
characteristics during the life of the vegetation and
does not vary seasonally. The resulting measure
reflects long-term conditions. Seasonal trends in the
consumptive use coefficient were calculated from
measured ET, but were not defined by seasonal
trends in the vegetation description. Transfer of the
ET value determined by this method should be
restricted to areas having seasonal, climatic, and
environmental trends similar to the project site.

The problem of obtaining an adequate description
of the vegetation was recognized at the beginning of
the study and various methods were investigated.
Repetitive infrared-color aerial photographs were avail-
able beginning in 1967 and the development of a
technique for relating photographic spectral response
to ET was described by Jones (1977). The results are
encouraging because the photogrammetric data are a
measure of the contemporary transpiration charac-
teristics of the vegetation. This method was not
available until the latter part of this project and
therefore could not be thoroughly developed and
tested for use in this report.

REFERENCES CITED

Blaney, H. F., Ewing, P. A., Morin, K. V., and Criddle, W. P., 1942,
Consumptive water use and requirements: Pecos River Joint
Investigation Reports of Partcipating Agencies, National
Resources Planning Board, p. 170—200.

Blaney, H. F., and Criddle, W. D., 1962, Determining consumptive
use and irrigation water requirements: Us Department of
Agriculture Technical Bulletin 1275, 59 p.

Bowie, J. E., and Kam, William, 1968, Use of water by riparian
vegetation, Cottonwash Wash, Arizona: US. Geological Sur—
vey Water-Supply Paper 1858, 62 p.

Burkham, D. E., 1970, Precipitation, streamflow, and major floods

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

at selected sites in the Gila River drainage basin above

Coolidge Dam, Arizona: U.S. Geological Survey Professional

Paper 655—B, p. B1—B33.

1972, Channel changes of the Gila River in Safford Valley,
Arizona 1846—1970: U.S. Geological Survey Professional
Paper 655—G, p. G1—G24.

_ 1976, Flow from small watersheds adjacent to the study
reach of the Gila River Phreatophyte Project, Arizona: U.S.
Geological Survey Professional Paper 655—1, p. 11—119.

Burkham, D. E., and Dawdy, D. R., 1970, Error analysis of
streamflow data for an alluvial stream: U.S. Geological Survey
Professional Paper 655-C, p. C1—C13.

Christiansen, J. E., 1968, Pan evaporation and evapotranspiration
from climatic data: Journal American Society Civil Engineers,
Irrigation and Drainage Division, v. 94, no. 1R2, p. 243—265.

Cruff, R. W., and Thompson, T. H., 1967, A comparison of methods
of estimating potential evapotranspiration from climatologi-
cal data in arid and subhumid evnironments: U.S. Geological
Survey Water-Supply Paper 1839—M, p. M1—M28.

Culler, R. C., and others, 1970, Objectives, methods, and environ-
ment—Gila River Phreatophyte Project, Graham County,
Arizona: U.S. Geological Survey Professional Paper 655—A,
p. A1—A25.

Davidson, E. S., 1961, Facies distribution and hydrology of
intermontane basin fill, Safford basin, Ariz., in Geological
Survey Research, 1961: U.S. Geological Survey Professional
Paper 424—C, p. C151-C153.

Dawdy, D. R., Lichty, R. W., and Bergman, J. M., 1972, A rainfall-
runoff simulation model for estimation of flood peaks for small
drainage basins: U.S. Geological Survey Professional Paper
506—B, p. Bl—B28.

Erie, L. J., French, 0. F., and Harris, Karl, 1965, Consumptive use
of water by crops in Arizona: Agricultural Experiment Station,
University of Arizona Technical Bulletin 169, 44 p.

Gatewood, J. 8., Robinson, T. W., Colby, B. R., Hem, J. D., and
Halpenny, L. C., 1950, Use of water by bottom-land vegetation
in lower Safford Valley, Arizona: U.S. Geological Survey
Water-Supply Paper 1103, 210 p.

Hanson, R. L., 1972, Subsurface hydraulics in the area of the Gila
River Phreatophyte Project: U.S. Geological Survey Profes-
sional Paper 655—F, p. F1—F27.

Hanson, R. L., Kipple, F. P., and Culler, R. C., 1972, Changing the
consumptive use on the Gila River flood plain, southeastern
Arizona, in Age of Changing Priorities for Land and Water:
Proceedings American Society Civil Engineers, Irrigation and
Drainage Division Specialties Conference, September 1972, p.
309—330.

Hanson, R. L., and Dawdy, D. R., 1976, Accuracy of evapotran-
spiration rates determined by the water-budget method, Gila
River flood plain, southeastern Arizona: U.S. Geological Sur-
vey Professional Paper 655—L, p. L1—L35.

Horton, J. S., and Campbell, C. J., 1974, Management of phreato-
phyte and riparian vegetation for maximum multiple use
values: U.S. Department of Agriculture, Forest Service Re—
search Paper, RM 117, 23 p.

Jensen, M. E., and Haise, H. R., 1963, Estimating evapotran-
spiration from solar radiation: Journal American Society Civil
Engineers, Irrigation and Drainage Division, v. 89, no. 1R4,
p. 15—41.

Jones, J. E., 1977, Evapotranspiration calculated using color-
infrared photography: U.S. Geological Survey Professional
Paper 655—0, p. 01—045.

 

 

P51

Kipple, F. P., 1977, The hydrologic history of the San Carlos
Reservoir, Arizona, 1929-71, with particular reference to tran-
spiration and sedimentation: U.S. Geological Survey Profes-
sional Paper 655—N, p. N1—N40.

McQueen, I. S., and Miller, R. F., 1972, Soil moisture and energy
relationships associated with riparian vegetation near San
Carlos, Arizona: US Geological Survey Professional Paper
655—E, p. E1—E51.

National Weather Service, issued annually, Climatogical Data,
Arizona: U.S. Department of Commerce.

Park, D. M., Culler, R. C., and Turner, R. M., 1978, Management of
flood-plain vegetation, 1967 to 1972, San Carlos Indian Reser-
vation, Arizona: U.S. Geological Survey Open-File Report
78—412, 21 p.

Rantz, S. E., 1968, A suggested method for estimating evapotran-
spiration by native phreatophytes: U.S. Geological Survey
Professional Paper 600-D, p. D10—D12.

Rosenbrock, H. H., 1960, An automatic method of finding the
greatest or least squares value of a function: Computer J our-
nal, v. 3, p. 175—184.

Richie, J. T., 1972, Model for predicting evaporation from a row
crop with incomplete cover: Water Resources Research, v. 8,
no. 5, p. 1204—1213.

Sellers, W. D., and Hill, R. H., 1974, Arizona climate, 1931—1972:
Tucson, Arizona, University Press, 616 p.

Tanner, C. B., 1957, Factors affecting evaporation from plants and
soils: Journal Soil Water Conservation, v. 12, p. 221—227.

Thiessen, A. H., 1911, Precipitation from large areas: Monthly
Weather Review, v. 39, p. 1082-1084.

Thornthwaite, C. W., 1948, An approach towards a rational classi-
fication of climate: Geographical Review, v. 38, no. 1, p. 55—94.

Turner, R. M., 1971, Measurement of spatial and temporal changes
in vegetation from color IR film: Proceedings, International
Workshop on Earth Resources Survey Systems, Ann Arbor,
Mich., May 3—15, 1971, p. 513—525 and Proceedings, American
Society of Photogrammetry Fall Convention, September 1971,
16 p.

1974, Quantitive and historical evidence of vegetation
changes along the upper Gila River, Arizona: U.S. Geological
Survey Professional Paper 655—H, p. H1—H20.

U.S. Bureau of Reclamation, 1973, Progress report, phreatophyte
investigations, Bernardo evapotranspirometers: U.S. Bureau
of Reclamation, Middle Rio Grande Project Office, v. 2, 50 p.

U.S. Congress, 1958, Public Law 85—500: 85th U.S. Congress.

U.S. Geological Survey, Surface-water supply of the United States-
Part 9, Colorado River basin: U.S. Geological Survey water-
supply papers (published annually through 1960; published
periodically, 1961 to present).

van Bavel, C. H. M., 1966, Potential evaporation: the combination
concept and its experimental verification: Water Resources
Research v. 2, no. 3, p. 455—467.

van Hylckama, T. E. A., 1974, Water use by saltcedar as measured
by the water-budget method: US. Geological Survey Profes-
sional Paper 491—E, p. E1—E30.

Weist, W. G., Jr., 1971, Geology and ground-water system, Gila
River Phreatophyte Project: U.S. Geological Survey Profes-
sional Paper 655—D, p. D1—D22.

White, W. N., 1932, A method of estimating ground-water supplies
based on discharge by plants and evaporation from soil,
results of investigations in Escalante Valley, Utah: U.S. -
Geological Survey Water-Supply Paper 659—A, p. A1—A105.

 

 

 

 

 

 

 

TABLE 6

 

 

P54

GILA RIVER PHREATOPHYTE PROJECT

TABLE 6.—Water-budget components, resulting ET, and total
measurement error for each budget period during water years
1963—71, reaches 1, 2, 2a, and 3

[All values are in acre-feet per budget period]

 

REACH 1
Budget Proj—
per1od ectl/ _ _ _ _ 2/
endlng day— Days ET QI Q0 AC QT P AMS AM Mc GB GI Go AMTC SET—
3—19—63 170 14 521 4641 ~4601 68 99 62 91 41 86 ~77 111* 478
4— 2—63 184 14 30 1358 ~1529 8 1 6O 97 41 86 ~77 ~15 249
4—16—63 198 14 129 1344 —1477 1 0 105 ~7 41 86 ~77 113 248
4~30~63 212 14 228 1261 ~1255 0 0 69 29 152 41 86 ~77 ~78 215
5—14—63 226 14 ~171 847 ~904 11 0 0 14 —8 41 87 ~77 182 194
5—28—63 240 14 312 505 ~521 6 0 0 88 67* 41 87 ~77 116 193
6—11—63 254 14 389 256 ~231 4 0 0 34 171 41 88 ~77 103 178
6—25—63 268 14 220 153 ~111 5 0 0 41 90 41 89 ~77 ~11 158
7— 9~63 282 14 306 26 —9 2 0 33 34 86* 41 91 ~77 79* 159
7~23~63 296 14 254 7 0 0 0 0 18 96* 41 91 ~78 79* 165
8— 6—63 310 14 446 3858 ~3574 ~128 67 148 ~34 6 41 88 ~78 52 598
8—20—63 324 14 851 4220 -3613 64 18 175 20 ~12 41 87 ~77 ~72 490
9— 3~63 338 14 4122 30698 ~26253 ~84 10 309 ~21 —377 41 89 ~77 213 1997
9—17—63 352 14 2556 19267 ~16780 33 0 49 34 ~16 41 88 ~77 ~83 1336
10— 1—63 366 14 728 8215 ~7761 64 0 15 21 69 41 87 ~77 54 727
10—15—63 380 14 293 983 ~960 28 1 0 7 137 41 87 ~77 46 210
10—29—63 394 14 582 11960 ~11288 ~42 21 96 ~24 —128 41 87 ~77 ~64 1106
11~12~63 408 14 137 5036 ~5022 ~13 36 26 30 41 85 ~77 ~5 496
11-26-63 422 14 209 4107 -3873 17 50 ~23 ~70 41 83 ~77 ~46 419
12-10-63 436 14 525 3491 ~3131 7 24 7 71 41 83 -77 9 364
12—24—63 450 14 494 2691 -2316 18 0 10 ~17 41 82 ~76 61 315
1~ 7-64 464 14 230 2527 ~2269 ~29 0 0 ~18 41 83 ~76 ~29 311
l—21—64 478 14 —640 4203 ~4474 ~5 0 ~18 -166 41 82 ~76 191 438
2~ 4—64 492 14 ~40 4322 ~4401 7 29 ~2 6 41 82 ~76 ~48 445
2~18~64 506 14 97 2185 —2241 33 0 9 17 41 83 ~76 46 296
3— 3—64 520 14 9 1015 ~1108 3 0 48 ~20 17 41 84 ~76 5 200
3-17—64 534 14 195 958 ~898 5 0 35 6 21 41 85 ~76 18 192
3—31-64 548 14 196 977 ~928 0 0 26 ~2 7 41 85 ~76 66 197
4—14—64 562 14 112 987 ~890 0 0 75 ~17 -51 41 85 ~76 -42 195
5— 4—64 582 20 269 1166 -999 4 0 4 16 27 59 121 —107 ~22 200
5—25—64 603 21 410 688 ~600 8 0 0 23 127 62 128 ~113 87 184
6—15—64 624 21 372 173 —144 7 0 0 17 162 62 131 -113 77 172
7— 6—64 645 21 424 8 ~3 0 0 5 15 145 62 132 -114 174 161
7—27—64 666 21 1056 7614 ~6839 0 14 93 6 36 62 130 ~115 55 819
8-17-64 687 21 1903 18502 ~17068 98 176 272 ~61 ~86 62 128 -113 -7 1427
9- 7-64 708 21 513 1051 -1107 23 252 174 8 82 62 124 ~112 -44 341
9—28—64 729 21 2799 24456 ~21821 -165 310 289 ~73 ~272 62 123 —113 3 1732
10-19-64 750 21 229 3080 ~3254 148 0 75 27 89 62 127 —ll4 ~11 409
11- 9~64 771 21 132 696 ~674 11 O 0 25 1 62 125 -114 0 177

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

P55

TABLE 6,—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,

reaches 1, 2, 2a, and 3-C0ntinued

 

REACH 1
Budget Proj—
per1od ect / _ _ _ _ 2/
endlng day— Days ET Q1 Q0 AC QT P AMS AM Mc GB GI G0 AMTC EET—
11—30—64 792 21 132 2223 ~2094 ~27 103 ~33 ~69 62 122 ~115 ~40 262
12—21—64 813 21 113 2519 -2499 10 88 ~18 ~24 62 121 —115 ~31 275
1—11—65 834 21 42 3907 ~3857 ~71 78 —8 ~72 62 121 —115 —3 408
2— 1-65 855 21 ~193 10502 —10619 26 96 ~17 -113 62 119 —115 —134 776
2—22—65 876 21 465 10649 —10155 ~15 172 ~63 —109 62 118 ~115 ~79 757
3—15-65 897 21 247 8414 ~8082 ~15 96 ~25 ~65 62 118 -114 ~142 655
4— 5—65 918 21 33 6736 —6785 23 42 29 ~35 62 119 ~114 ~44 545
4—26—65 939 21 378 5631 ~5460 11 29 58 ~22 62 120 —114 63 470
5-17—65 960 21 82 3098 —3231 46 0 5 66 ~11 62 121 —114 40 335
6- 7—65 981 21 407 902 ~934 13 0 0 16 235 62 124 —113 102 187
6—28—65 1002 21 448 355 -287 6 O 40 27 146 62 123 —113 89 159
7—19—65 1023 21 469 316 —203 2 0 41 26 115 62 123 —113 100 164
8— 9~65 1044 21 1979 16734 —15225 ~26 333 320 ~51 —159 62 120 —113 ~16 1294
8~30~65 1065 21 265 3078 -3651 ~87 456 225 ~10 70 62 118 ~112 116 600
9—20—65 1086 21 1224 15221 —14261 80 51 124 17 ~14 62 114 ~113‘ ~57 1247
10—11-65 1107 21 436 1434 —1223 22 0 O 8 92 62 113 —112 40 206
11— 1~65 1128 21 302 1180 ~1079 ~15 0 0 24 59 62 114 ~113 70 192
11—22—65 1149 21 265 1767 ~1590 0 0 11 0 5 62 114 ~113 9 221
1—24—66 1212 63 -287274 968 -628 ~1061* 187 352 —336 ~1957
2—14—66 1233 21 —228 24032 ~24629 —2 150 ~53 110* 62 135 —120 87 1467
3— 7—66 1254 21 1702 25095 ~23774 26 15 105 119* 62 129 —121 46 1464
3—28—66 1275 21 ~104514 7 —192 ~637* 62 119 ~118 ~430
4—18—66 1296 21 —1333 40110 -42587 114 0 254 496* 62 129 —116 205 2214
5— 9—66 1317 21 999 15433 —15584 71 0 0 161 403 62 136 ~115 432 1044
5—30—66 1338 21 683 4800 —5399 52 0 2 217 453 62 138 -115 473 456
6—20—66 1359 21 671 1614 —1882 13 0 38 111 340 62 136 —115 354 228
7—11—66 1380 21 738 928 ~987 9 28 120 20 234 62 132 ~115 307 188
8— 1—66 1401 21 405 758 —1075 —3 382 347 -179 28 62 130 -115 70 428
8—22—66 1422 21 911 3885 ~3639 ~107 233 69 132 161 62 132 -115 98 501
9—12—66 1443 21 564 6402 —6373 93 0 330 ~26 78 62 131 ~113 ~20 633
10— 3—66 1464 21 366 13093 ~12924 ~14 42 52 104 —135 62 134 —112 64 1096
10—24-66 1485 21 109 2217 —2338 21 0 9 33 71 62 129 ~113 18 264
11-14—66 1506 21 350 2233 ~2235 ~19 77 7 86 62 128 —113 124 258
12— 5—66 1527 21 263 3346 —3241 13 0 l7 2 62 126 —113 51 325
12—19—66 1541 14 110 1382 —1416 0 95 ~29 17 41 83 ~75 12 215
1—16—67 1569 28 44 2917 ~2921 ~13 41 —5 ~37 83 167 ~151 ~37 286
2~ 6—67 1590 21 144 3901 —3778 16 36 3 ~35 62 124 ~113 ~72 363
2-27—67 1611 21 105 1428 —1420 1 40 —5 ~38 62 125 0113 25 206
3—20-67 1632 21 —273 1336 —1477 4 15 1 ~98 62 124 —113 ~l27 206

P56 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6,—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3—Continued

 

REACH 1
Budget Proj—
5331:: 3:;l/ Days ET Q Q AC Q 5 AM AM M G G G AM a 3/

I 0 T s I c B I 0 TC 8T
4—10—67 1653 21 199 1138 ~1108 0 22 17 14 62 124 —114 44 190
5— 1-67 1674 21 75 910 ~936 5 10 11 —9 62 124 —114 12 181
5—22—67 1695 21 221 632 ~666 3 0 0 57 103 62 124 ~114 20 169
6—12—67 1716 21 327 409 ~459 1 9 45 44 105 62 123 —115 103 162
7~ 3—67 1737 21 200 440 ~541 2 0 31 14 71 62 120 ~115 116 166
7—24—67 1758 21 413 5567 ~6148 ~21 746 280 ~55 ~59 62 116 —115 40 1048

8—14—67 1779 21 ~71778 520 263 —746 -l675 62 116 —114 ~400

9~ 4—67 1800 21 —34631 313 82 326 860 62 135 ~98 ~224
9—25—67 1821 21 738 8035 —7660 —248 33 145 52 163 62 142 ~110 124 810
10—16—67 1842 21 446 9467 —9554 250 0 106 30 125 62 137 —111 ~66 805
11- 6—67 1863 21 159 2947 —3116 ~10 0 37 86 62 137 —110 126 306
11-27—67 1884 21 212 2890 —3054 3 47 30 44 62 136 —109 163 302
12—18—67 1905 21 397 5000 ~4885 ~64 582 ~179 ~48* 62 133 ~109 ~95 464
1— 8—68 1926 21 —11978 57007 ~68105 ~95 48 ~119 ~410* 62 129 ~113 ~382 4448
1—29—68 1947 21 1767 44924 ~42768 ~261 49 87 ~234* 62 123 ~113 ~102 2181

2—19-68 1968 21 —125535 209 ~205 ~499* 62 104 ~115 ~869
3-11—68 1989 21 —13600 107485 ~120537 ~89 130 16 —134* 62 102 —111 ~524 4452
4— 1~68 2010 21 1228 78645 —77971 137 12 150 142 62 112 ~1o9 48 3589
4—22—68 2031 21 —1641 59584 —61349 110 48 52 ~128 62 114 —109 ~25 2927
5—13—68 2052 21 —999 27556 ~29324 76 0 1 243 213 62 125 —110 159 1649
6— 3—68 2073 21 752 12212 ~12541 85 0 o 256 450 62 131 ~112 209 882
6~24~68 2094 21 403 2878 ~3239 32 0 11 108 310 62 137 —113 217 314
7- 8—68 2108 14 351 1071 ~1106 2 0 66 22 ~50* 41 93 ~76 288 211
7—22-68 2122 14 25 690 ~819 10 0 0 10 7 41 92 ~76 70 173
8— 5—68 2136 14 254 2035 ~2304 ~65 296 244 ~39 —74 41 90 ~77 107 426
8—19—68 2150 14 ~55 12024 ~11764 25 70 190 ~59 -213 41 80 —77 —372 998
9— 2-68 2164 14 55 6739 ~6716 ~47 2 66 -5 ~47 41 84 ~76 14 632
9~16~68 2178 14 348 2275 ~2602 79 0 1 63 217 41 87 —76 263 317
9—30—68 2192 14 90 657 —774 10 0 1 15 101 41 90 ~76 25 178
10—14—68 2206 14 —42 1225 ~1469 5 0 47 0 23 41 95 —81 72 188
10-28—68 2220 14 161 813 ~809 -6 0 0 4 57 51 86 ~76 51 175
11—11—68 2234 14 97 1584 ~1505 ~15 8 2 ~36 41 84 ~75 9 225
11—25—68 2248 14 83 3342 ~3358 —28 136 ~27 ~44 41 81 ~75 15 384
12— 9—68 2262 14 ~31 4080 —4109 7 14 3 ~33 41 80 ~75 ~39 426
12—23—68 2276 14 244 2832 ~2782 11 132 0 ~3 41 82 -75 6* 327
1— 6-69 2290 14 ~235 4952 ~5099 ~24 160 —144 ~80 41 81 ~75 ~47 501
1—20-69 2304 14 317 5779 —5569 —16 56 2 —9 41 81 —75 27 535
2— 3—69 2318 14 78 7624 ~7408 ~10 60 ~10 —124 41 80 ~75 —100 663

2-17-69 2332 14 199 6172 —6135 33 38 11 49 41 81 —75 —16 573

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P57

TABLE 6.—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71.
reaches 1, 2, 2a, and 3—Continued

 

REACH 1
Budget Proj—

per1od ect / _ _ _ _ 2/
endlng day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC SET—
3— 3—69 2346 14 105 2893 —3030 34 30 17 73 41 85 —75 37 350
3—17—69 2360 14 99 1600 —1711 2 85 -5 25 41 86 —75 51 251
3—31—69 2374 14 0 1178 —1336 2 0 46 53 41 86 -75 5 225
4—14—69 2388 14 56 1350 —1449 0 13 60 27 41 86 -75 3 234
4—28—69 2402 14 96 1041 —1104 6 0 60 14 41 86 —75 27 213
5—12—69 2416 14 98 989 —1069 —2 120 —18 20 41 85 —75 7 212
5—26—69 2430 14 200 624 —710 9 0 3 60 74 41 85 —75 89 191
6— 9—69 2444 14 138 430 —474 5 0 0 3 85 41 85 —75 38 184
6—23—69 2458 14 192 228 —267 1 0 0 5 78 41 85 -75 96 176
7- 7—69 2472 14 228 187 —195 4 0 10 3 79 41 84 —75 90 176
7—21—69 2486 14 183 442 —747 —8 168 159 —29 71 41 82 -75 79 261
8- 4—69 2500 14 89 397 —437 0 0 13 27 29 41 80 —75 14 183
8—18—69 2514 14 154 508 -782 —5 92 195 -11 68 41 79 —75 44 223
9— 1—69 2528 14 —17 657 —654 -27 6 41 8 ~61 41 78 —75 —31 193
9—15—69 2542 14 —233 7144 —7374 -59 46 154 —28 —132 41 74 —75 —24 702
9~29—69 2556 14 183 1608 —1662 91 O 0 35 85 41 76 -75 -16 279
10—13—69 2570 14 208 357 —374 2 0 0 2 62 41 79 —74 113 169
10—27-69 2584 14 207 952 —858 —10 15 73 —12 —22 41 77 -74 25 221
11—10—69 2598 14 -67 1158 —1209 —3 66 —3 —16 41 75 —74 -102 221
11—24—69 2612 14 112 2118 —2054 —29 225 —114 —81 41 75 —74 5 292
12— 8—69 2626 14 86 3014 —2971 —9 110 —18 —25 41 75 —74 -57 346
12—22—69 2640 14 —87 3183 —3286 27 5 7 —33 41 76 —74 -33* 360
1— 5—70 2654 14 106 1467 —1505 1 93 —16 12 41 78 —74 9* 227
1—19—70 2668 14 117 1805 —1703 O 7 17 -37 41 80 —74 -19 247
2— 2-70 2682 14 98 1227 —1188 11 0 20 -l 41 81 -74 —19 215
2-16v70 2696 14 40 1108 —1124 —2 25 4 —10 41 81 -74 -9 208
3— 9—70 2717 21 127 3645 —3746 —20 389 —143 ~60 62 121 -112 -9 435
3—23-70 2731 14 81 1596 —1672 17 31 72 -27 41 78 —74 19 242
4— 6—70 2745 14 135 1184 -1253 3 22 72 40 41 81 —74 19 215
4-20e70 2759 14 18 1051 ~1146 —1 34 42 5 41 75 —74 —9 207
5— 4-70 2773 14 19 1023 —1063 1 O 30 —6 41 76 —74 —9 204
5—18v70 2787 14 72 722 —743 7 0 0 1 21 41 80 -75 9 186
6— 1—70 2801 14 23 515 —579 3 0 8 12 7 41 82 —75 9 177
6-15—70 2815 14 136 334 —382 5 0 0 7 50 41 82 —75 74 170
6-29-70 2829 14 269 220 e236 1 O 44 12 107 41 81 —75 74 166
7—13—70 2843 14 154 144 —171 3 2 14 9 42 41 80 —75 65 164
7—27—70 2857 14 209 357 —369 —2 8 114 —10 1 41 79 —75 65 176
8—10—70 2871 14 153 1441 —1370 —23 2 63 —3 —37 41 77 —75 37 282

8—24-70 2885 14 183 1164 —1259 19 0 90 26 65 41 75 -75 37 235

P58 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6.—Water—budget components, resulting ET, and total measurement error for each budget period during water, years 1963—71,
reaches 1. 2, 2a, and 3—Continued

 

 

REACH l
\l
Budget Proj—
per1od ectl/ - _ _ _ 2/
endlng day- Days ET QI Q0 AC QT P AMS AMI MC GB GI GO AMTC EET—
9— 7~70 2899 14 262 213 ~235 —1 7 155 ~29 73 41 76 ~75 37 168
9—21—70 2913 14 193 641 —587 0 16 52 0 ~8 41 76 ~75 37 196
10— 5—70 2927 14 568 5577 —4727 —154 0 68 —9 -145 41 74 ~74 ~83 750
10—19—70 2941 14 113 1864 ~1922 170 0 1 11 33 41 72 ~74 ~83 306
11— 2—70 2955 14 97 573 ~523 ~2 0 0 17 9 41 74 ~73 ~19 177
11—16—70 2969 14 41 727 —731 2 0 8 12 41 74 ~73 ~19 187
11—30—70 2983 14 14 585 -597 —1 7 ~6 —7 41 74 ~73 -9 179
12—14—70 2997 14 55 678 ~720 O 6 14 44 41 74 ~73 —9 184
12—28—70 3011 14 5 668 —735 —2 53 ~10 ~2 41 74 ~73 —9 184
1—11—71 3025 14 126 2729 ~2598 ~35 29 —3 —28 41 73 ~73 ~9 328
1—25—71 3039 14 ~40 3498 —3483 1 0 4 —54 41 72 ~73 ~46 376
2— 8—71 3053 14 91 3211 -3064 4 0 3 ~57 41 72 —73 ~46 351
2—22—71 3067 14 356 2929 —2648 2 95 —26 7 41 66 ~73 ~37 326
3— 8—71 3081 14 —8 2096 —2096 7 11 —1 —17 41 61 —73 -37 274
3—22—71 3095 14 ~88 1090 ~1208 20 5 1 —10 41 74 ~73 ~28 212
4— 5-71 3109 14 27 739 —739 3 0 5 4 41 75 —73 ~28 186
4—19—71 3123 14 74 698 —668 O 54 -8 10 41 77 -74 ~56 184
5— 3—71 3137 14 ~20 597 ~618 1 0 8 12 41 69 —74 ~56 180
5-17—71 3151 14 41 434 ~444 5 O 0 7 -10 41 63 ~74 19 173
5~31~71 3156 14 82 333 —337 1 0 0 1 33 41 65 ~74 19 169
6—14—71 3179 14 93 160 ~218 4 0 0 12 46 41 76 ~74 46 165
6~28~71 3193 14 106 92 —118 2 0 0 7 33 41 77 ~74 46 163
7—12—71 3207 14 203 160 ~136 1 O O 12 66 41 77 —74 56 165
7—26—71 3221 14 192 257 ~321 —8 42 142 ~43 25 41 75 ~74 56 177
8— 9—71 3235 14 —728 2701 —3666 ~43 199 248 ~71 ~45 41 74 ~73 ~93 462
8—23—71 3249 14 ‘ ~14671 318 176 27 —282 41 71 ~72 O
9— 6-71 3236 14 518 6032 ~5803 197 21 5 61 56 41 72 ~71 ~93 626
9-20—71 3277 14 ~183 2939 —3122 —128 O 7 30 46 41 75 ~71 O 463
REACH 2
7— 9—63 282 14 427 9 0 2 50 0 312* 42 77 ~65 0 154
7—23—63 296 14 573 O 0 0 29 29 461 42 77 ~65 O 130
8- 6-63 310 14 —195 3574 ~3560 0 ~200 —14 -49 42 77 ~65 0 566
8—20—63 324 14 —455 3613 ~4153 66 32 —7 ~60 42 77 ~65 O 468
9— 3-63 338 14 329 26253 —24712 ~75 ~33 0-3077 42 77 —65 ~81* 1832
9—17—63 352 14 ~755 16780 —17387 32 65 ~17 —246 42 77 —65 -36* 1275
10~ 8—63 373 21 —293 8372 —9296 92 113 18 310 64 115 ~97 16* 749
10—22—63 387 14 111 7906 ~7680 —123 85 ~43 8 ~116 42 77 ~65 20* 967

11— 5—63 401 14 535 5918 ~5535 68 43 36 ~13 —4 42 77 ~65 ‘32* 543

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA P59

TABLE 6.—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3—Continued

 

REACH 2

Budget Proj—

per1od ectl/ _ _ _ _ 2

endlng day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC SET
11—19—63 415 14 437 4911 —4522 19 31 4 13 —41 42 77 —65 —32* 457
12— 3—63 429 14 —53 3491 —3483 —5 59 —20 —18 —107 42 77 —65 —24* 360
12—17—63 443 14 —246 2870 —3120 9 40 —6 -7 —61 42 76 —65 —24* 322
12—31—63 457 14 97 1670 —1662 23 0 23 12 —12 42 76 —65 -10* 219
1—14-64 471 14 27 3516 —3395 -41 0 -6 —9 —64 42 76 —65 ~27* 363
1—28—64 485 14 167 4857 —4605 0 41 —13 —7 —134 42 76 —65 -25* 453
2-11—64 499 14 92 3403 —3342 25 0 —11 O 7 42 76 —65 ~43 368
2-25—64 513 14 104 1412 —1394 17 0 0 —9 33 42 76 —65 -8 222
3—10—64 527 14 302 1003 —946 5 108 -13 3 51 42 76 —65 38 191
3—24—64 541 14 103 *878 —890 0 48 1 —8 —2 42 76 —65 23 185
4— 7—64 555 14 246 954 —835 1 61 —20 11 15 42 76 —65 6* 159
4—27—64 575 20 143 1100 —1049 3 18 29 —6 —31 61 108 —94 4* 168
5—18—64 596 21 387 749 —646 7 0 0 53 7 101 64 113 —98 37* 153
6— 8—64 617 21 728 280 —166 9 0 0 49 9 372 64 113 —96 94* 135
6—29—64 638 21 789 18 0 0 0 14 50 12 532 64 114 —95 80 141
7—20—64 659 21 1058 3945 —3644 0 3 17 47 1 464 64 115 —94 140 647
8—10—64 680 21 1068 15463 —14642 —9 138 208 -36 3 -142 64 113 —90 —2 1257
8—31—64 701 21 537 5523 —5730 10 356 391 —134 1 6 64 113 —90 27 763
9—21—64 722 21 439 7923 —7941 —43 118 318 —16 -6 ~33 64 113 -93 35 791
10—12—64 743 21 —409 16722 —17155 35 0 61 68 0 —178 64 113 —93 —46 1495
11— 2—64 764 21 363 1001 —826 4 0 93 3 —1 77 64 114 —94 —72 177
11—23—64 785 21 85 1388 -1305 —32 141 —43 0 —93 64 115 —95 —55 202
12—14—64 806 21 145 2776 —2568 9 6 1 —7 —83 64 115 —96 —72 269
1— 4—65 827 21 197 2045 —1910 —21 197 —110 1 -91 64 115 -97 4 227
1—25—65 848 21 283 9625 —9040 ~45 167 —60 —6 —322 64 115 —94 —121 712
2—15—65 869 21 293 10334 ~9643 0 336 —252 —13 —419 64 115 —90 -139 728
3— 8—65 890 21 328 9270 —9021 46 8 102 —7 —147 64 114 —87 ~14 680
3-29—65 911 21 —391 6839 —7033 —8 126 2 3 —324 64 114 —88 -86 564
4—19—65 932 21 -65 5914 —6127 1 127 39 4 —89 64 114 —91 -21 491
5—10—65 953 21 793 4224 —3911 28 0 0 201 4 110 64 114 —90 49 368
5—31—65 974 21 616 1287 -1152 13 0 1 74 3 239 64 113 -91 65 171
6—21-65 995 21 563 433 —317 14 0 35 28 2 213 64 113 —90 68 128
7—12—65 1016 21 832 204 —101 —32 0 91 64 8 516 64 113 —91 -4 129
8— 2—65 1037 21 1195 9876 —9236 -210 254 499 -223 —9 97 64 113 —90 60 946
8—23—65 1058 21 842 7569 -7424 219 216 150 65 6 34 64 113 —86 —84 821
9—13—65 1079 21 1188 13715 —12815 —143 188 369 —131 0 -72 64 113 —87 —13 1184
10— 4—65 1100 21 763 3003 —2901 149 72 0 203 —16 107 64 112 -87 57 346
10—25—65 1121 21 562 926 —662 O O 14 46 —4 92 64 113 —88 61 148

11-15~65 1142 21 250 1414 -1213 —4 O 69 10 —96 64 113 ~89 -18 181

P60 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6,—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3—Continued

 

REACH 2
Budget Proj—
2:333 3:;l/ Days ET Q _ _ ’“ -— 2
I Q0 AC QT P AMS AMI MC GB GI GO AMTC SET—/
12— 6—65 1163 21 253 2943 —2648 —22 251 —211 —1 -101 64 113 —89 —46 290
2— 7—66 1226 60 299841 1371 —807 —166-1991* 192 341 —269 —524
2—28—66 1247 21 —838 25726 —26632 —26 28 137 —15 —78* 64 121 —90 —73 1549
3—21—66 1268 21 82666 0 23 12 —238 64 120 -77 —119
4—11—66 1289 21 62847 0 239 8 139 64 116 —43 -125
5— 2—66 1310 21 21203 0 233 27 424 64 115 —30 113
5—23-66 1331 21 8013 0 0 213 35 456 64 115 —31 84
6—13—66 1352 21 2515 0 0 141 47 799 64 115 ~38 115
7— 4—66 1373 21 1211 29 38 132 56 556 64 115 —59 121
7-25—66 1394 21 759 2 155 37 21 375 64 115 —57 151
8—15—66 1415 21 604 1628 —1557 —15 28 264 —50 13 109 64 115 —65 70 221
9— 5—66 1436 21 460 8132 —8455 —1 120 228 80 1 129 64 114 —74 122 738
9—26—66 1457 21 —98 12617 —12884 —32 120 424 —166 —7 —199 64 112 -78 —69 1117
10—17—66 1478 21 795 2947 —2711 30 0 56 8o 10 241 64 112 —80 46 284
11— 7—66 1499 21 323 1781 —1622 -2 0 45 29 —12 8 64 113 —82 1 206
11—28—66 1520 21 64 3387 —3298 —15 43 7 5 —101 64 113 —83 —58 314
12—12—66 1534 14 132 1590 —1572 15 148 —151 —1 40 42 75 —56 2 210
1— 9—67 1562 28 250 2646 —2640 —3 50 18 0 33 85 151 —112 22 253
1—30—67 1583 21 —48 3937 —3915 —16 57 1 —8 —186 64 113 —84 ~11 355
2—20-67 1604 21 54 1682 —1908 27 4 23 3 55 64 113 -84 75 214
3—13—67 1625 21 225 1549 —1449 1 81 0 0 30* 64 113 —84 —80* 218
3-27—67 1639 14 17 793 -807 2 24 0 0 24* 42 75 —56 -80* 184
4—17—67 1660 21 285 1106 —1090 2 60 35 -20 39 64 114 —84 59 163
5— 8—67 1681 21 271 813 —805 4 0 74 22 95 64 114 —84 -26 146
5—29—67 1702 21 720 601 -522 4 1 75 51 13 268 64 114 —85 136 135
6—19—67 1723 21 739 422 —286 0 0 26 37 —8 310 64 115 —84 143 131
7—10—67 1744 21 974 973 —694 ~18 2 87 o 8 347 64 115 -84 174 170
7—31-67 1765 21 531 9288 —9822 —48 433 625 -253 16 114 64 115 —80 79 991
8—21—67 1786 21 97483 1105 502 —497 —77 2428 64 107 —76 -385
9—11—67 1807 21 464 7061 —8088 88 76 115 374 35 744 64 101 —84 —21 664
10— 2—67 1828 21 —212 11784 —12751 11 4 234 137 4 305 64 112 —85 —31 1070
10—23—67 1849 21 219 4454 —4948 38 0 69 108 10 301 64 110 -85 98 421
11—13‘67 1870 21 408 3161 —2961 -2 0 0 52 9 69 64 110 —85 —9 291
12— 4—67 1891 21 318 3556 —3447 —16 179 -78 4 36 64 109 —85 —4 323
12—25—67 1912 21 675 37789 —36169 —132 891 -516 —43-1088 64 110 —84 -147
1—15—68 1933 21 —3712 46413 449488 —26 62 —79 —35 -498 64 114 —79 -160 2467
2- 5-68 1954 21 71931 111 —l38 —3 -425 64 115 —78 —298
2—26—68 1975 21 125158 275 —143 —19 —870* 64 114 —73 ~313
3—18—68 1996 21 115935 237 85 -190 —612* 64 110 -39 -263

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

P61

TABLE 6.—Water‘budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3—Continued

 

REACH 2

Budget Proj-

per1od ectl/ _ _ _ _ 2/
ending day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC ET—
4— 8-68 2017 21 66744 37 288 —187 —360* 64 109 —16 —73

4—29-68 2038 21 46889 65 75 —189 320 64 109 —12 -12

5—20—68 2059 21 24010 0 9 174 26 134 64 111 —10 84

6—10—68 2080 21 8174 0 21 132 6 521 64 113 —10 190

7— 1—68 2101 21 2275 0 O 202 118 378 64 114 —12 40

7-15—68 2115 14 1112 5 169 115 42 220 42 76 —10 196

7-29—68 2129 14 1003 32 240 —8 96 67 42 77 —12 65

8—12—68 2143 14 9806 189 221 ~37 17 —528* 42 77 —18 —49

8—26—68 2157 14 6987 0 170 23 42 —188 42 76 —21 -137

9- 9—68 2171 14 5272 0 25 185 32 167 42 76 —23 76

9—23—68 2185 14 1088 O 0 80 0 142 42 76 —25 155
10— 7—68 2199 14 646 0 96 —8 6 99 42 76 —28 55
10—21—68 2213 14 733 O O —3 15 24 42 75 —30 43
11— 4-68 2227 14 1055 0 4 12 —4 —41 42 76 ~33 —36
11—18—68 2241 14 2162 241 —73 -18 —111 42 75 —36 -17
12— 2—68 2255 14 4322 20 6 —16 —13 42 75 —37 -31
12—16—68 2269 14 3411 0 —9 —1 —59 42 75 —38 -47
12—30—68 2283 14 3917 489 —232 —13 —35 42 75 —39 -51

1—13-69 2297 14 5022 44 —3 —3 —114 42 75 —39 -88

1—27-69 2311 14 6204 131 —36 —16 —74 42 75 —38 —34

2—10—69 2325 14 7608 46 2 —5 —182 42 75 —37 —16

2—24—69 2339 14 4601 66 4 0 —45 42 75 ~36 27

3—10—69 2353 14 2043 81 16 8 —22 42 75 -36 -35

3~24-69 2367 14 1515 38 30 7 82 42 75 —36 8

4- 7—69 2381 14 1346 0 53 13 —25 42 75 —37 39

4—21—69 2395 14 1328 12 58 20 12 42 75 —38 169

5— 5—69 2409 14 991 173 32 34 57 42 75 —39 4

5—19—69 2423 14 991 0 43 13 6 —61 42 75 -41 —5

6— 2—69 2437 14 289 585 ~442 6 O 1 39 8 3O 42 75 —43 -12 139
6—16—69 2451 14 527 347 —186 5 0 0 29 23 182 42 75 —45 55 128
6—30—69 2465 14 360 231 —51 2 0 3 16 6 37 42 75 —47 46 119
7—14—69 2479 14 510 222 —25 0 0 3 48 28 105 42 75 -49 61 125
7-28—69 2493 14 1026 802 —88 —11 104 153 —15 8 1 42 75 —50 5 183
8-11—69 2507 14 721 388 —190 8 3 130 —13 17 251 42 75 —54 64 125
8—25-69 2521 14 576 801 —495 —5 29 173 -43 l 48 42 75 —48 -2 155
9- 8—69 2535 14 3345 0 55 3 4 —117 42 75 —48 —55

9—22—69 2549 14 5879 72 124 —36 3 27 42 75 —51 —58
10— 6—69 2563 14 194 495 —355 13 0 O —14 1 47 42 74 —55 -54 128
10—20—69 2577 14 419 353 —207 0 19 192 —21 —5 5 42 74 —57 24 122

P62 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6.—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3-Continued

 

REACH 2
Budget Proj—
per1od ect / - _ _ _ 2/
endlng day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC SET—
11— 3—69 2591 14 194 1283 —1096 —18 0 0 11 5 —55 42 74 —58 6 189
11—17—69 2605 14 344 1245 —1128 —19 366 —106 —10 —57 42 74 —59 —4 183
12— 1—69 2619 14 24 2957 —2804 —7 11 —9 —12 —134 42 74 —58 —36 312
12—15—69 2633 14 474 3492 —3139 —10 119 —27 —1 16 42 74 —58 —34 352
12—29—69 2646 14 220 1949 —1872 26 134 —17 —9 —16 42 74 —57 —34 243
1—12—70 2661 14 1 1797 —1779 —6 6 1 1 —45 42 74 —56 -34 228
1—26—70 2675 14 28 1368 —1424 17 4 17 —1 —13 42 74 —56 0 198
2— 9—70 2689 14 —62 1090 —1213 0 0 23 O —22 42 74 —56 0 179
2—23-70 2703 14 179 1073 —1027 1 69 21 3 6 42 74 —56 —27 174
3—16—70 2724 21 306 4127 —4147 —11 534 —151 —26 —85 64 111 —83 —27 436
3—30—70 2738 14 165 1422 —1465 6 20 86 7 8 42 74 —55 20 202
4—13—70 2752 14 270 1146 —1136 3 4 95 5 73 42 74 —56 20 179
4—27—70 2766 14 143 1134 ~1150 1 51 44 3 —27 42 74 —56 27 179
5—11-70 2780 14 257 926 —892 4 O 78 —4 57 42 75 —56 27 162
5—25—70 2794 14 272 660 —636 5 0 0 56 16 77 42 75 —57 34 145
6— 8—70 2808 14 314 482 —378 6 0 15 36 —1 59 42 75 —56 34 133
6—22-70 2822 14 520 297 -102 6 0 0 38 9 156 42 75 —56 55 124
7— 6-70 2836 14 583 204 —8 —1 2 21 37 4 209 42 75 -57 55 122
7—20—70 2850 14 587 120 —8 1 5 109 —11 18 271 42 75 ~55 20 125
8— 3—70 2864 14 479 420 —137 —1 O 69 6 3 39 42 75 —57 20 136
8-17—70 2878 14 558 2126 —1647 —55 0 40 29 —6 —14 42 75 —59 27 280
8—31—70 2892 14 505 494 —420 56 43 147 —2 0 103 42 75 -60 27 154
9—14-70 2906 14 507 285 —219 -1 34 286 -58 6 98 42 75 —61 20 137
9—28—70 2920 14 275 571 —443 0 O 0 12 O 60 42 75 —62 20 156
10—12—70 2934 14 6146 0 68 —43 —30 —424 42 74 —63 —61
10—26—70 2948 14 187 646 —567 12 0 0 33 17 55 42 73 —63 —61 143
11— 9—70 2962 14 71 690 —637 —6 0 0 8 —2 —4O 42 73 —64 7 147
11—23—70 2976 14 177 591 ~509 2 0 10 3 22 42 73 —64 7 140
12— 7—70 2990 14 122 672 —634 —4 37 —4 —6 9 42 73 —63 0 146
12—21—70 3004 14 110 694 —640 4 82 —17 1 —66 42 73 —63 O 148
1— 4—71 3018 14 177 1297 —1162 —37 91 —13 25 —42 42 73 —63 —34 199
1—18—71 3032 14 —91 3403 —3350 —4 2 0 —7 —154 42 73 —62 —34 353
2- 1—71 3046 14 31 3387 —3334 3 O l 0 —68 42 73 -59 -14 351
2—15—71 3060 14 —134 2846 —2999 6 0 3 —4 —29 42 73 —58 —14 318
3— 1-71 3074 14 260 2348 -2281 9 152 —28 0 9 42 73 —57 -7 270
3—15—71 3088 14 53 1767 —1745 10 10 —0 —3 —38 42 73 —56 —7 226
2—39—71 3102 14 8 868 —908 9 O —2 0 —38 42 73 —56 20 161
4—12—71 3116 14 170 660 —678 0 0 17 0 92 42 73 —56 20 146

4—26—71 3130 14 268 638 -583 3 73 9 6 55 42 74 —56 7 143

TABLE 6.—-Water-budget components, resulting E

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

,

reaches 1. 2, 2a, and 3—Continued

 

 

 

REACH 2
Budget Proj—
per1od ect / _ _ _ _ 2/
endlng day — Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC ET—
5-10—71 3144 14 125 547 -519 3 0 0 24 O 3 42 74 —56 7 143
5—24—71 3158 14 127 374 —353 2 0 0 23 —4 ~2 42 74 —56 27 135
6— 7—71 3172 14 205 285 —238 4 O 0 10 6 51 42 74 —56 27 125
6—21—71 3186 14 188 172 —106 1 0 0 12 5 18 42 74 -57 27 122
7- 5—71 3200 14 144 109 —89 —3 0 0 1 —6 46 42 74 —57 27 122
7—19—71 3214 14 245 168 —184 -18 83 86 3 11 30 42 74 ~57 7 154
8— 2—71 3228 14 340 1168 —1432 —49 354 533 —181 —7 —111 42 73 —57 7 429
8—16—71 3242 14 8249 928 360 —10l l —194 42 73 —53 ‘41
8—30—71 3256 14 12702 218 137 1 -18 -273 42 72 —54 -41
9—13—71 3270 14 2999 0 7 71 4 179 42 71 —55 ‘7
9—27—71 3284 14 5504 O 6 82 7 3O 42 71 —56 ‘7
REACH 2a
10— 4—65 1100 21 230 19 0 139 —11 98 41 112 -149 47
10—25—65 1121 21 926 0 6 31 0 74 41 113 —149 46
11—15—65 1142 21 1414 0 36 2 —25 41 113 —147 —54
12— 6—65 1163 21 2943 144 —133 0 —26 41 113 —147 —19
2— 7—66 1226 63 299841 817 —593 —6 —494* 124 340 —401 ~334
2—28—66 1247 21 25726 21 109 —13 0* 41 121 ~13O —63
3—21—66 1268 21 82666 0 —11 —3 —521* 41 120 —135 -59
4—11—66 1289 21 62847 0 139 —12 796* 41 116 —133 —95
5— 2—66 1310 21 21203 0 146 15 282 41 115 —118 87
5—23—66 1331 21 8013 O O 135 8 290 41 115 —114 73
6—13—66 1352 21 2515 0 0 98 16 498 41 115 —114 133*
7— 4—66 1373 21 656 1211 —1106 7 0 23 91 O 301 41 115 —116 89 178
7-25—66 1394 21 663 759 —595 -1 0 55 37 7 229 41 115 -118 134 154
8—15—66 1415 21 203 1628 —1668 —4 4 162 —57 2 41 41 115 —119 58 223
9— 5—66 1436 21 8132 47 131 38 1 89 41 114 —114 71
9—26—66 1457 21 12617 11 242 —117 —1 —113 41 112 —112 —41
10—17—66 1478 21 —621 2947 —3881 33 O 37 48 7 109 41 112 —118 44 349
11— 7—66 1499 21 228 1781 —1594 —1 0 14 22 —4 —8 41 113 —120 —16 205
11—28—66 1520 21 180 3387 —3175 -9 39 0 0 —52 41 113 -120 —44 312
12—12—66 1534 14 198 1590 —1402 9 91 —120 —2 8 27 75 —80 2 209
1— 9—67 1562 28 92 2646 —2657 —2 33 19 0 20 55 151 —161 -12 255
1—30—67 1583 21 -125 3937 —4008 —9 37 2 2 —119 41 113 —121 0 361
2-20—67 1604 21 —78 1682 —1904 15 4 11 4 31 41 113 —121 46 219
3—13—67 1625 21 102 1549 —1467 0 49 0 0 14* 41 113 —121 -76 212
3—27—67 1639 14 28 793 —763 1 15 O 0 10* 27 75 —80 ~50* 177
4—17-67 1660 21 195 1106 —1077 0 34 26 —4 22 41 114 —122 55 167

P63

T and total measurement error for each budget period during water years 1963—71,

P64 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6.—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,
reaches 1, 2, 2a, and 3—Continued

 

REACH 2 a

Budget Proj—

per1od ect / _ _ _ _ 2/

ending day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC SET—

5— 8-67 1681 21 86 813 ~843 2 0 36 >15 41 41 114 ~122 ~11 149
5—29-67 1702 21 388 601 ~570 2 1 45 29 3 144 41 114 ~122 100 137
6—19—67 1723 21 343 422 —414 ~l 0 22 16 ~13 188 41 115 ~122 89 139
7—10-67 1744 21 323 973 ~1024 —6 0 44 9 ~7 178 41 115 ~123 123 183
7—31—67 1765 21 9288 204 370 -180 13 ~11 41 115 ~123 51

8—21—67 1786 21 97483 242 279 ~269 ~10 ~611* 41 107 ~128 ~306

9-11—67 1807 21 7061 39 61 237 12 614* 41 101 ~128 ~12
10- 2~67 1828 21 11784 4 127 73 ~14 205 41 112 ~129 ~38
10—23—67 1849 21 4454 0 61 29 8 170 41 110 ~130 89
11~13~67 1870 21 314 3161 ~2915 ~1 O 0 50 ~1 18 41 110 ~130 ~19 290
12~ 4—67 1891 21 188 3556 ~3481 ~10 96 ~37 5 41 41 109 ~130 -2 325
12-25—67 1912 21 37789 526 ~296 ~14 ~651 41 110 ~128 ~128

1—15—68 1933 21 46413 41 ~76 ~25 ~263 41 114 ~l35 ~109

2~ 5—68 1954 21 71931 53 ~100 5 ~228 41 115 ~136 ~208

2~26—68 1975 21 125158 173 35 ~12 ~185* 41 114 ~138 ~219

3—18—68 1996 21 115935 137 52 ~10 ~3* 41 110 ~122 ~157

4~ 8—68 2017 21 66744 22 170 ~24 113* 41 109 ~70 ~19

4—29-68 2038 21 46889 54 41 10 212 41 109 ~41 8

5—20—68 2059 21 24010 O 7 113 6 94 41 111 ~29 62

6~10~68 2080 21 8174 0 16 85 18 306 41 113 ~27 129

7~ 1—68 2101 21 2275 0 O 123 4 255 41 114 ~40 36

7—15—68 2115 14 1112 O 71 46 0 152 27 76 ~38 189

7~29~68 2129 14 151 1003 ~1088 ~2 15 113 ~13 -5 34 27 77 ~46 36 177

8—12-68 2143 14 9806 93 156 ~44 0 ~137* 27 77 ~56 ~23

8—26—68 2157 14 6987 0 113 18 ~6 ~129* 27 76 ~57 ~107

9— 9-68 2171 14 5272 O 13 113 5 116 27 76 ~56 64

9—23-68 2185 14 328 1088 —1069 6 0 0 47 6 116 27 76 ~56 87 173
10— 7-68 2199 14 379 646 ~501 6 0 58 0 ~10 91 27 76 ~57 43 139
10—21—68 2213 14 232 733 ~601 ~1 0 O —1 3 22 27 75 ~58 33 144
11— 4~68 2227 14 27 1055 ~1043 ~7 O 0 9 0 ~21 27 76 ~58 ~11 170
11~18~68 2241 14 106 2162 ~2088 ~16 136 ~40 ~2 ~80 27 75 ~58 ~10 255
12— 2-68 2255 14 230 4322 —4101 —4 14 —1 2 ~16 27 75 ~57 ~31 416
12-16—68 2269 14 ~31 3411 —3439 11 0 ~4 1 ~33 27 75 ~57 ~23 357
12—30-68 2283 14 42 3917 ~3969 ~21 281 ~156 ~6 ~35 27 75 ~56 ~15* 418
1-13~69 2297 14 ~325 5022 ~5268 6 24 —1 0 ~85 27 75 ~56 ~69 487
1—27—69 2311 14 ~111 6204 ~6327 ~11 81 ~17 ~4 ~51 27 75 ~57 ~31 567
2~10~69 2325 14 ~706 7608 ~8263 3 29 8 —2 ~118 27 75 ~57 ~16 683
2~24~69 2339 14 47 4601 ~4667 17 38 5 2 ~19 27 75 ~56 24 452

3-10—69 2353 14 55 2043 ~2039 13 39 11 —3 ~29 27 75 ~55 ~27 251

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

TABLE 6.—Water—budget components, resulting ET, and total measurement error for each budget period during water 3!

Budget Proj—

reaches 1, 2, 2a, and 3—Continued

 

peréod ectl/ _ _ _ _ 2/
endlng day— Days ET QI QO AC T P AMS AMI MC B GI GO AMTC, ET—
3—24-69 2367 14 186 1515 ~1503 3 28 22 1 68 27 75 ~56 6 207
4— 7—69 2381 14 102 1346 —1315 -2 0 33 0 ~25 27 75 ~56 19 193
4—21-69 2395 14 141 1328 ~1285 4 8 25 ~1 15 27 75 ~57 2* 193
5 —5—69 2409 14 223 991 —967 1 101 14 7 18 27 75 ~57 13* 169
5—19—69 2423 14 45 991 ~964 3 0 36 0 —5 ~50 27 75 ~58 ~10 169
6— 2—69 2437 14 68 585 ~567 3 0 1 16 ~3 7 27 75 ~60 ~16 142
6—16—69 2451 14 218 347 —329 3 O 0 15 3 99 27 75 ~62 40 130
6—30—69 2465 14 146 231 ~198 1 0 l 14 1 31 27 75 ~65 28 128
7—14—69 2479 14 187 222 ~l70 0 0 1 26 ~1 28 27 75 ~69 48 126
7—28—69 2493 14 302 802 ~578 —6 7 84 ~11 —2 ~20 27 75 ~72 —4 165
8~11~69 2507 14 301 388 —349 6 0 76 2 4 111 27 75 ~70 31 132
8—25—69 2521 14 350 801 ~593 —1 14 99 ~33 0 24 27 75 ~73 10 174
9— 8—69 2535 14 3345 0 27 ~3 —2 ~106 27 75 ~78 ~45
9—22—69 2549 14 5879 36 68 ~26 0 —5 27 75 ~73 ~54*
10— 6-69 2563 14 34 495 ~533 6 0 0 ~10 1 48 27 74 ~73 —1 153
10—20-69 2577 14 249 353 ~324 O 14 101 4 0 54 27 74 ~73 19 128
11— 3—69 2591 14 45 1283 ~1239 ~10 0 0 —2 3 ~27 27 74 ~73 9 198
11-17—69 2605 14 269 1245 ~1180 ~10 217 ~40 0 3 27 74 ~73 6 186
12— 1-69 2619 14 ~114 2957 ~2993 ~5 6 ~12 -5 ~62 27 74 ~73 ~28 321
12—15—69 2633 14 262 3492 ~3278 —5 66 ~13 —1 0 27 74 ~73 ~27* 356
12—29-69 2647 14 35 1949 ~2027 16 78 —1 ~3 —3 27 74 ~73 ~2* 248
1—12—70 2661 14 —9 1797 —1787 —3 3 —3 —7 ~39 27 74 ~69 —2* 228
1—26—70 2675 14 ~36 1368 —1467 9 1 15 4 7 27 74 ~69 —5 199
2- 9—70 2689 14 —131 1090 ~1221 0 O 13 —5 ~34 27 74 ~69 ~6 178
2—23-70 2703 14 132 1073 ~1039 1 45 15 5 15 27 74 ~70 ~14 171
3~16~70 2724 21 327 4127 -3998 ~6 315 ~89 —5 ~51 41 111 -104 ~14 425
3—30—70 2738 14 125 1422 ~1396 3 11 50 3 ~14 27 74 ~69 14 198
4—13—70 2752 14 202 1146 ~1130 1 4 78 —4 60 27 74 ~69 15 177
4—27—70 2766 14 160 1134 ~1053 1 31 26 0 ~29 27 74 ~69 18 174
5~11~70 2780 14 161 926 ~914 2 0 52 —2 47 27 75 ~69 17 162
5—25—70 2794 14 66 660 ~686 2 0 0 24 5 13 27 75 ~69 15 145
6- 8—70 2808 14 111 482 ~462 2 0 5 21 —6 22 27 75 ~70 15 134
6—22-70 2822 14 153 297 ~292 2 0 O 9 1 72 27 75 ~72 34 125
7~ 6—70 2836 14 229 204 ~164 1 0 12 16 1 97 27 75 ~73 33 122
7—20—70 2850 14 180 120 ~63 1 0 33 ~2 4 52 27 75 ~80 13 120
8- 3~70 2864 14 151 420 ~345 ~2 0 42 1 0 O 27 75 ~79 12 135
8-17—70 2878 14 169 2126 ~1975 ~35 O 27 4 ~3 ~19 27 75 ~80 22 290
8-31-70 2892 14 232 494 ~484 36 38 76 ~1 0 31 27 75 ~81 21 149
9—14-70 2906 14 262 285 —264 —2 0 159 ~17 4 65 27 75 ~82 12 130

P65

ears 1963—71,

P66 GILA RIVER PHREATOPHYTE PROJECT

TABLE 6.—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963-71,
reaches 1, 2, 2a, and 3—Continued

 

 

 

REACH 2a
Budget Proj—
per1od ect / _ _ _ _ 2/
endlng day— Days ET QI Q0 AC QT P AMS AMI MC GB GI G0 AMTC eET—
9-28—70 2920 14 40 571 —579 O O 0 —5 0 20 27 75 —81 12 159
10—12—70 2934 14 6146 0 43 —25 —12 —225 27 74 —81 —50
10—26—70 2948 14 646 0 0 12 9 35 27 73 —77 —50
11— 9—70 2962 14 690 0 0 0 0 —3 27 73 —74 12
11—23—70 2976 14 55 591 —607 2 0 9 2 18 27 73 —73 13 141
12— 7—70 2990 14 73 672 —676 —3 l6 0 —3 36 27 73 —72 3 145
12—21—70 3004 14 32 694 —710 3 40 —3 2 —23 27 73 —71 0 147
1— 4—71 3018 14 138 1297 —1200 —37 65 —8 14 8 27 73 —70 —31 199
1-18—71 3032 14 —11 3403 ~3338 —3 1 0 —2 —70 27 73 -72 ~30 351
2— 1—71 3046 14 216 3387 —3155 2 0 -2 0 —36 27 73 —73 —7 344
2—15—71 3060 14 —237 2846 —3092 6 0 0 —2 —15 27 73 —73 —7 321
3— 1—71 3074 14 123 2348 —2348 8 94 —14 O 18 27 73 —73 —10 272
3—15—71 3088 14 10 1767 —1805 13 7 3 O 8 27 73 —73 -10 228
3—29—71 3102 14 —7 868 —900 7 O —3 —2 —21 27 73 —72 16 159
4—12—71 3116 14 95 660 —682 0 0 11 —3 66 27 73 —72 15 145
4—26—71 3130 14 121 638 —634 2 42 2 5 38 27 74 —72 —1 143
5—10—71 3144 14 77 547 —513 4 0 0 11 0 0 27 74 —72 —1 137
5—24—71 3158 14 112 374 -357 l 0 O 21 0 18 27 74 -72 26 128
6— 7—71 3172 14 92 285 —277 4 0 0 l 4 19 27 74 —72 27 125
6—21—71 3186 14 109 172 —146 2 O 0 6 3 13 27 74 —72 30 121
7— 5-71 3200 14 77 109 —115 —2 0 0 —1 —1 28 27 74 —72 30 120
7—19—71 3214 14 155 168 —140 —11 42 27 8 4 22 27 74 —72 6 131
8— 2—71 3228 14 1168 131 300 —147 —2 —31 27 73 —72 6
8~16—71 3242 10 8249 507 179 —66 —1 —136 27 75 —70 —33
8—30—71 3256 14 12702 167 85 5 —4 —170 27 72 —73 —34
9—13—71 3270 14 219 2999 —3080 142 O 4 55 2 76 27 71 —73 —4 365
9-27—71 3284 14 5504 0 5 58 2 18 27 71 —72 —8
REACH 3 —— 1964 Water Year
10— 8—63 373 21 9296 35 10 —0
10—22—63 387 14 7680 8 3 —30 23 7 —0 7*
11— 5—63 401 14 5535 16 0 3O 23 7 '—0 —24*
11-19—63 415 14 4522 —18 0 70 23 7 —0 —10*
12— 3—63 429 14 3483 34 0 7 23 7 —0 -5*
12—17—63 443 14 3120 —39 0 —56 23 7 —0 —6*
12-31—63 457 14 1662 l 4 —7 23 7 —0 1*
1—14—64 471 14 3395 —25 —2 —23 23 65 —32 —9*
1—28—64 485 14 25 4605 —4589 0 7 0 —45 23 65 —32 -9* 446
2—11—64 499 14 36 3342 —3322 20 —19 —3 —34 23 65 —32 -4* 361

EVAPOTRANSPIRATION BEFORE AND AFTER CLEARING PHREATOPHYTES, ARIZONA

P67

TABLE 6,—Water-budget components, resulting ET, and total measurement error for each budget period during water years 1963—71,

reaches 1, 2, 2a, and 3—Continued

 

REACH 3

Budget Proj—

per1od ect / _ _ _ _ 2/
endlng day— Days ET QI 00 AC QT P AMS AMI MC B GI G0 AMTC SET—
2«25—64 513 14 —27 1394 -1495 14 —4 —1 6 23 65 —32 3* 209
3—10—64 527 14 47 946 —1035 2 6 1 70 23 65 —32 1* 168
3—24—64 541 14 —34 890 —934 1 —5 —2 —40 23 65 —32 0* 170
4— 7—64 555 14 70 835 —821 3 —8 2 —4 23 65 —32 7* 165
5—11—64 589 34 246 1529 —1455 6 25 3 —16 57 160 —78 15* 190
6- 1—64 610 21 433 328 —218 8 33 —1 168 35 97 —47 30* 95
6—22—64 631 21 541 5 O 0 47 2 388 35 95 —50 19 94
7—13—64 652 21 580 0 0 0 23 42 2 444 35 94 —54 —6 90
8— 3—64 673 21 16210 146 —26 1 -206 35 91 —36 13

8—24—64 694 21 7175 90 —3 1 -23 35 90 —17 13

9—14—64 715 21 4529 250 -46 —3 21 35 93 —26 -16

10— 5—64 736 21 21548 54 —47 —1 —268 35 93 —24 0

10—26—64 757 21 999 81 —7 2 57 35 93 —29 —9

11—16—64 778 21 239 582 —432 -14 51 15 0 —57 35 95 —33 —3 114
12— 7—64 799 21 46 2598 —2509 —12 44 —14 0 —119 35 96 —32 —41 246
12—28—64 820 21 108 1914 —1880 14 73 —19 -3 —86* 35 96 -29 «7 209
1—18—65 841 21 6198 98 —56 -7 —201* 35 95 —26 —82

2- 8—65 862 21 9378 273 —204 0 —186 35 92 —19 —42

3— 1—65 883 21 10687 6 161 —2 067 35 86 —19 —28

3-22—65 904 21 7392 80 —142 —3 —238 35 88 —24 —12

4~12—65 925 21 6299 78 48 —1 ~68 35 90 —28 18

5— 3—65 946 21 400 4776 —4490 13 0 88 4 —73 35 90 —32 —11 390
5—24—65 967 21 374 1763 —1682 23 0 62 -6 110 35 90 —38 17 191
6—14—65 988 21 661 558 —408 8 1 103 9 269 35 90 —45 41 100
7— 5—65 1009 21 702 40 —5 0 31 84 O 447 35 90 »44 24 83
7—26—65 1030 21 3648 136 —13 2 76 35 91 —39 37

8—16—65 1051 21 12174 286 ~217 0 —288 35 87 —42 19

9— 6-65 1072 21 9471 67 81 0 —173 35 87 —46 18

9—27-65 1093 21 6857 134 —30 0 216 35 86 «70 -25

 

1/ Project day 1 is October 1,

2/ Measurement error (see text).

1962

* Soil moisture change estimated from change in ground—water levels

.LDEII’OHd ELLAHdOLVElHHd HEIAIH V119 Ell-LL :IO dVW NOILVLEIDEIA

node; slul ;o BaJV

ISLOES

 

AZZODL L

"08

 

N\N V sopeg ues pue EN 5 soueg ues

sluud eoueApe paupaun Aa/uns magfimoeg 'S'fl tum; aseg

 

Jepaones
JEPBOJIBS
HOWEAOO

 

 

 

133} SI 01 9'9
waned 09—93
Juaoxad QZ—I

Juaaxad gL—[g

_,\,

133} 81 ueq; 3.10m

73% x
/u

aunbsaw
" mus
NOILVNV'ldXZ-l

pu9[ paggssepun JO
‘pamnmno ‘uaneq pue [auueqo lama

 
 

waoxad 001‘9L
Juaoxad QL‘IS
Juaoxad 09—93
Juaoxad 93—1

133} L ueq) 310m

 

 

 

«EC—m ZODAI

 

LQGK ‘NOIJNNI'IOECI
NVBW BLVWIXOHddV

 

 

BZSl 50 NEW] WOHHEA $130039 TVNOIiVN
lEBfl UV 1VAHHLN| HflOlNOC)
000 1783 BWVOS

 

 

HELENO'HM I

 

 

 

v
\ \ H2 ./ .

((1 I

J » Wfl“
\f//’\ I
/’\\ ‘ 7
s o '1 WV av J

 

 

99908M—-l85l"V/\ 'NOlSEH 'ASAHHS ‘IVIJIEJOWDBEJfiOIUElNI

.,3>/

n}.\/>\./.u

/.//

r

M11 ,1 it ,

$9059

‘ X f
\J

/

Defl“

//,

 

 

”081 ZZOU l l

AHAHOS 'IVOIDO'IOEID

HOIHEILNI Ell-LL :IO .LNEIWLHVdC-ICI SELLVJS CIELLIND

 

 

 

 

 

 

 

 

Q/
S»

30 A $35

/

73??

 

‘\/

\%

 

d‘999 HEIdVd 'IVNOISSEHOHd

I '3le

 

 

“OE/[000 l L

«108110088

   

”id
SL
329