E75 , 4917- A - 7 DA¥y Introduction, Spread and Areal Extent of - Saltcedar (Tamarix) in the Western States GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-A sep R uggs "| ~ Science uS Introduction, Spread and Areal Extent of Saltcedar (Famarix) in the Western States By T. W. ROBINSON $T V DLE s O F EVAP O PRAN S P |O N GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-A UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1965 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY Thomas B. Nolan, Director For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 CONTENTS Page Occurrence, spread, and density, etec.-Continued Page pea hes aree euro eles (it en rake aho pt aa e b oo a Al Water consumption: ...;..." cote caclace.s A8 fabroduction =s. == ci aces. 1 Effect on stream regimen... [.ll 10 Entrance into the United States.______________ ___. 3 - Future considerations 10 Time of general awareness of saltcedar in the South- References:._" :~". [Iio one sav act a- m uue 11 west sss poe ale ce noo Zel oa nea n bann o aed 4 Occurrence, spread, and density of saltcedar growth.... 6 320302002 8 ILLUSTRATIONS Prats 1. Map showing occurrence of saltcedar in the 3. Aerial photographs, in 1936 and 1957, of the Page western l} 002 _} In pocket Arkansas River valley at the Otero-Bent Page County line, Colorado______________LL___ A9 Fraur® 1. Dense saltcedar growth along the Gila River A2 2. Increase in the area of saltcedar growth and cover density along the Pecos River in New 7 TABLES Page Page 1. Tamariz specimens collected in the Western 2. Gross areas of saltcedar growth in the West- Slates (1877 to 1920).._._.:..._..._ i.... A5 ern States as of 1961....:.1:.:-.1..0l}.l.. A8 IH STUDIES OF EVAPOTRANSPIRATION INTRODUCTION, SPREAD, AND AREAL EXTENT OF SALTCEDAR (TAMARIX) IN THE WESTERN STATES By T. W. Rostnsox ABSTRACT Saltcedar, the name generally applied to two exotic deciduous species of the genus Tamariz, was introduced into this country more than 100 years ago and has, in the last 30 years, become very much of a nuisance plant in the arid and semiarid regions of the Western States. The species are highly water-consum- ing, salt-tolerant, naturalizing shrubs that have escaped from cultivation and spread rapidly from one stream valley to another. Saltcedar occurs in 15 of the 17 Western States. Areas in- fested range from less than 1,000 acres each in Idaho, Montana, Nebraska, and South Dakota, to about 450,000 acres in Texas. Its dense growth along stream channels presents a barrier to flood flows, and thereby increases flood hazards and sediment deposition. The time of awareness of the plant by residents of the region was generally in the 1920's. The total area of saltcedar growth has increased from an estimated 10,000 acres in 1920 to more than 900,000 acres in 1961. It is possible that by 1970 saltcedar will be growing on 1%, million acres. Not only is the growth increasing in areal extent but also in density of growth. The consumptive waste of ground water by the plant is estimated as 40 to 50 thousand acre-feet in 1920, 3.5 million acre-feet in 1961, and possibly 5.0 million acre-feet by 1970. INTRODUCTION Since about 1930 saltcedar has become a nuisance plant in the arid and semiarid regions of Western States. In these regions saltcedar is the common name by which the deciduous species 7. pendantre Pallas and T. gallica Linnaeus of the genus Tamariz are known. Saltcedar is a highly water-consuming naturalizing shrub that has escaped from cultivation and spread rapidly from one stream valley to another. An ag- gressive plant, it has not only invaded but has entirely replaced the native vegetation in many areas. The dense junglelike growth shown in figure 1 is typical of its occurrence in a well-established stand. Owing to the rapid spread of the plant, its high water consump- tion, and the potential flood hazard engendered by it, saltcedar is of concern to the residents of these regions. This concern becomes greater each year as the demand - for water increases, the need for reducing flood hazards mounts, and, at the same time, the areal extent and growth density of the plants are increasing. The genus TZTamariz, introduced into the Western States from the Mediterranean area, is one of the four genera of the Tamaricaceae family native to Africa, Asia, and Europe. Although many species of the genus have been brought into this country, only two have escaped from cultivation to become important in the saltcedar problem. As a means of evaluating the magnitude of the prob- lem posed by saltcedar, the Phreatophyte Subcommittee of the Pacific Southwest Inter-Agency Committee pro- posed an inventory of the areas of saltcedar growth. In 1958, the author began to assemble information on the location of areas of saltcedar growth. Although saltcedar was known to have spread widely since about 1920 and to be growing on thousands of acres in the Western States, this report represents the first attempt to map its distribution and areal extent. A brief history also was prepared dealing with the time of introduction and subsequent spread of the plants. In addition to field mapping, information was col- lected from all available sources. These include pub- lished and unpublished reports, information from Fed- eral and State agencies, and reports by county agents and consulting engineers and from interested indi- viduals. The offices of the State Engineers of Utah and New Mexico and the Colorado Water Conservation Board provided a large amount of data on phreatophyte growth in these States based on partial inventories made under their direction. In Texas most of the informa- tion was supplied by the Soil Conservation Service and was based on a general reconnaissance of saltcedar in- festation. Vegetative surveys by the Bureau of Recla- mation provided information concerning the Gila River, Ariz., and the Colorado River from Davis Dam to the international boundary. Field offices of the Water Re- sources Division, Geological Survey, located in the Western States, were very helpful in supplying data about their respective States. A1 STUDIES OF EVAPOTRANSPIRATION A2 'se6t 'opung stomog go Artur oy) uf "zury JO 18%0 sou Qp 'oAIX UID 94} Suote YJM018 asu@(G-'T SALTCEDAR IN THE WESTERN STATES The species growing in this country may be divided into the evergreen and deciduous types. The evergreen type generally is represented by the athel tree 7. apAyllae -; Linnaeus. It has been planted extensively as an orna- ' mental, a shade tree, and windbreak largely in the desert areas of the Southwest, but has rarely become naturalized. It is not a problem plant. Likewise, the deciduous 7. tetrandra has been widely used as an orna- mental shrub, but has nowhere become an aggressive plant. Originally thought to be confined to the warm and arid Southwest, saltcedar has spread northward since about 1950 into the Great Basin, the Rocky Mountains, and some of the Plains States. The two species of the deciduous type that have escaped cultivation are 7. pentandra and T. gallica. T. pentandra grows profusely along river bottoms throughout the West, but 7. gallica appears confined to salty soils near the Texas Gulf Coast. ENTRANCE INTO THE UNITED STATES Introduction of the species, which are native to the Old World from western Europe to the Himalaya Moun- tains, into the United States is not of firm record (Bowser, 1958). In discussing its introduction, Bowser notes (p. 13-14) that tamarisk could have been estab- lished in Mexico at an early date by Spanish explorers and conquistadors. Inasmuch as these invaders made expeditions into South-Central United States between 1540 and 1750, it could have been established here also. If tamarisk had been established by the early Spanish explorers, Mexico would have been the center of distri- bution and it would have been found in abundance there. The sparsity of collection of tamarisk specimens from Mexico, however, does not support this assumption. In the National Herbarium of the Smithsonian Institute, Washington, D.C., there is only one collection of tama- risk from Mexico, and this from the border town of Nogales, while there are many from other Latin Amer- ican countries such as Argentina, Venezuela, Chile, and Ecuador. Extensive collections of other plant species from Mexico, some very early, some modern, are filed in the Smithsonian Institute. .. What appears to be support for the assumption that the tamarisk was introduced into this country by the early Spanish explorers may be found in the journal of Father Escalante (Auerbach, 1948). Father Escalante was one of the early Spanish padres who explored the United States as far north as Utah. Ac- cording to the translation of his journal, Father Escalante reported tamarisk at Fort Pierce Wash near the Utah-Arizona border in 1776. Dr. E. L. Little, Jr., Forest Service dendrologist, points out (oral A3 commun., 1963) that there is some question concerning the translation of the word "taray" which Father Escalante used to describe the plant he observed. While "Taray" is defined as tamarisk in the more modern Spanish-English dictionaries published in the United States, it is not so defined in some dictionaries and botanical works published in Mexico. Rami{rez (1902) gives two different species under the heading "Taray." In Vera Cruz the legume Caesalpinic bonducella Roxb. is called Taray, while in "varios lugares" (other places) Eysemhardtia amorphoides H.B.K., also a legume, is known as Taray. Standley (1920), in describing a species of willow Saliz H.B.K., notes that in the State of Durango it is called "Taray" or "Taray de rio" while in Chihuahua it is called "Tarais." Saliz taxifolia occurs along streams and washes throughout Mexico, south to Guatemala, and in the United States from western Texas to Arizona. Inasmuch as this species of willow was common in much of the country traversed by Father Escalante, it seems probable that when referring to the plant in his journal he would use the local name "taray," meaning willow and not tamarisk. According to Christensen (1962, p. 53) , "tamaria was not recorded by the early explorers who traveled on the Green and Colorado Rivers in Utah" in the period from about 1869 to 1875, or by earlier explorers who also crossed these rivers. f The first introduction of tamarisk into the United), States appears to have been by nurserymen in the early _ 1800's. In 1823 according to Horton (1964, p. 2) "tam- arisk was offered for sale in New York City by the Old American Nursery operated by Lawrence & Mills," and in 1828 by Bartram's in Philadelphia. During the 1830's it was listed by several nurseries along the eastern seaboard. R Later the U.S. Department of Agriculture began growing tamarisk and in 1868 their annual report (p. 123) listed six species that had been established in the Department's Arboretum Grounds in Washington, D.C. (Horton, 1964, p. 2).. Between 1871 and 1890 a large number of collections of tamarisk were made from the plants growing in the arboretum. Apparently many plants were growing there before collections were made outside of Washington. The source of the stock is not known; the plants may have been imported or they may have been obtained from local nurseries. The earliest authentic record of Tamarix in the Western States of which the author is aware is found in the catalogs of early-day nurseries. Bowser (1958, p. 14) notes that California firms listed Tamariz, || species unknown, as early as 1856. Dr. H. M. Butter- - field, agriculturist emeritus, Agricultural Extension A4 Service, University of California, reports (written com- mun., 1963) that ZTamariz was available in California from nursery stock of the Highland Nurseries of New York as early as 1854. Dr. Butterfield, who has an extensive collection of early-day nursery catalogs, has the following to say concerning the early listing of T amariz: Some eastern nurseries were listing Tamariz before we had records in California. Most of our early introductions into Cal- ifornia came from nurseries in New York and other States. The old Downing Nursery at Newburg, N.Y., was taken over by the Saul family and in 1854 the Highland Nurseries op- erated by the Saul family listed Tamariz gallica, T. germanica and one called T. libanotis (p. 47 of their 1854-55 nursery catalog). James Saul was sent to California to represent the nursery and was in San Francisco in 1854 and later. * * * A. P. Smith, of the Pomological Garden and Nursery in Sacramento, in his 1856 nursery catalog (p. 14) listed 7. africana and T'. gallica. The Suscol Nursery, operated by the Thompson Brothers at Suscol, Calif., as early as 1856 on, listed two kinds of Zaemariz in 1861-African and German (p. 29 of 1861-2 catalog). (Suscol, now abandoned, was located about 6 miles south of Napa, Calif., and 1 mile west of State Highway 29 between Vallejo and Napa.) James Hutchison, of the Bay Nurseries in Alameda, in his catalog for 1874-75, on page 34 listed Tamariz gallica, while R. D. Fox, of the Santa Clara Valley Nurseries north of San Jose, in the 1884 catalog, page 28 listed Tamariz africana and T. chinensis. In discussing the species listed in these catalogs Dr. Butterfield points out the uncertainty of the proper names or synonyms given in the early listing. He feels that an opinion on synonyms should be based on what was probably grown rather than on present-day name usage. Because of the similarity in appearance, he thinks 7. africana may have been confused with either T. parviflora or T. gallica and that T. parvifiora has often been confused with 7. africana and T. gallica. There has indeed been much confusion of nomencla- ture for the deciduous species. McClintock (1951) has shown that the species common in Arizona and New Mexico is 7. pentandra Pallas rather than 7. gallica Linnaeus as formerly thought (Kearney and Peebles, 1942), and this interpretation was accepted in the later work of Kearney and Peebles (1951). It is apparent that tamarisk stock was available for distribution in the Western States as early as 1854 from nurseries in California, and from the Department of Agriculture Arboretum in Washington, D.C., in the 1870's. Although tamarisk stock was available from the eastern seaboard nurseries as early as 1823, it seems most likely that the role of these nurseries would be as sup- STUDIES OF EVAPOTRANSPIRATION pliers to the western nurseries, rather than as direct distributors. Regardless of the possible avenues by which tamarisk may have come to the Western States, the best evi- dence points to its escape from cultivation in the 1870's. Support for this theory is found in the dates of the early collections. The earliest collection-Z'. gaZ- lica-of which the author is aware was in 1877 at Gal- veston, Texas. 7. tetrandra appeared as a cultivated plant in 1880 at St. George, Utah. 7. pentandra, how- ever, may not have arrived until 1890. Information pertaining to the time and place of speci- men. collection is valuable in dating the introduction of tamarisk in an area. It is also valuable in following the spread of the plant from one area to another. For this purpose a list of tamarisk specimens collected in the Western States has been prepared, showing the dates of collection in chronological order, the collector, and the locality where collected. The present location of the specimen is given when known. (See table 1.) TIME OF GENERAL AWARENESS OF SALTCEDAR IN THE SOUTHWEST After the introduction of saltcedar into the South- west, a considerable period elapsed before residents generally became aware of its presence.. During this period the plants spread and formed stands of such size as to become noticeable. It was in the 1920's when those who lived close to nature-stockmen, farmers, sportsmen-began to realize that a new plant had made its appearance and gained a foothold in the stream valleys of the Southwest. That it was a consumer of ground water was not recognized at that time, nor for nearly 20 years thereafter. No mention is made of saltcedar by Dr. O. E. Meinzer in his classic paper "Plants as Indicators of Ground Water," published in 1927 as U.S. Geological Survey Water-Supply Paper 577, nor in an unpublished paper given before the Geo- - logical Society of Washington in 1922 by Professor ' G. E. P. Smith, of the University of Arizona, who dur- ing the period 1915-25 observed and studied plants that were users of ground water. f The earliest eyewitness report of saltcedar on the Gila River in Arizona comes from Ernest Douglas (written commun., 1962) of the Arizona Farmer-Ranchman, Phoenix. He recalls that it was about 1898 when his father, a cowman, brought home a switch of a new plant he had found growing in a sandbar along the Gila River. The switch was stuck in the moist soil at the edge of a ditch that ran by the house; it grew and in a season or two became a considerable clump of salt- cedar. The ranchhouse was located about 6 miles north of Gila Bend, Ariz., and about 1 mile from the Gila SALTCEDAR IN THE WESTERN STATES Ab TaBLE 1.-Tamarix specimens collected in the Western States (1877 to 1920) [Specimens in the National Herbarium of the Smithsonian Institution at Washington, D.C. (a) were examined and identified by Jerome S. Horton, U.S. Forest Service. Specimens in the Arnold Herbarium of Harvard University (b) were examined and identified by Elizabeth McClintock, California Academy of Sciences. Specimens in the Stanford University Herbarium (c). - Specimens in the University of Arizona Herbarium (d) and Pamona College Herbarium (e) were examined and identified by Miss McClintock, John E. Flood, Botany Department, Arizona State University, and Jerome S. Horton. Specimens in the California Academy of Sciences (f) were ex- amined and identified by Miss McClintock. Specimens without sufficient floral parts for positive identification are indicated with a question mark] Collector Location and Remarks Date Species Apt. 1877 ...-... Sept. 16, 1877... Apr. 14, 1880 T. tetrandra (€). T. tetrandra (a). Galveston Island, Tex. (naturalized). Galveston, Tex. St. George, Utah (cultivated). Texas. Catalina Mts., Ariz May 15, 1892. (10. s 2. 22A. . June 10, 1893 T. pentandra (d).. Brookings, S. Dak. (cultivated). Mar. 30, 1894... T. tetrandra (a)... Harrisburg, Utah. Apr. 5, 10. AREA AUL in as uk ean cams ao Beaver Dam, Ariz. Apr. 9-12, 1894... T. gallica (a, d).. Corpus Christi, Tex. May 5, 1894. .... T. tetrandra (€) .. Silver Reef, Utah. May T. gallica (a, d).. M. Hapeman..._..- cy Galveston, Tex. Apr. 17, 1896.........__.____.__._._... T. gallica (a, b).... A. A. and E. Gertr Cliff House, San Francisco, Calif. Apr. 1898. ._...- T. tetrandra (a) . & .J. M. Milligan. Bonham, Tex. Apr. 11, Deer wee Stanford Arboretum, Calif. Aug. 18, 1901... T. pentandra? (a) Tempe, Ariz. (common in river bottoms). Sept. 22, 1901... T. gallica (a). ..- Galveston Island, Tex. Apr. 15, 1902. T. pentandra (a) Barstow, Tex. Apr. 14, 1905 T. tetrandra (d) - Thatcher, Ariz. % (Oct /AL 1007. . 0200s Wilgus Ranch, Chiricahua Mts., Ariz. July 6, 1909. T. pentandra (a)... Kanab, Utah (cultivated). July 12, 1909 T. pentandra? (a). Winslow, Ariz. (cultivated). : May 23, 1910. T. pentandra (d)... Univ. Ariz. campus, Tucson, Ariz. May 21, 1911.. T. pentandra (a)... Nara Visa, N. Mex. (cultivated). 1911. . /...-L ._. T. tetrandra (a).. ._ El Paso, Tex. (cultivated). Sept. 8, 1912... T. gallica (a)... Galveston, Tex. May 21, 1913. -| T. gallica (d)....-- Univ. Ariz. campus, Tucson, Ariz. May 22, 1913. T. pentandra (d) . .-- Do africana (d) . tetrandra (1) . -| T. tetrandra? (c, 0) - - -| T. tetrandra (f). ...-- z T. pentandra (c, 0) _____ Iss wenee F Marion L. Campbell.. m Munz & Harwood..._.___....__._..__ Univ. Ariz. campus, Tucson, Ariz. (5 sheets). Univ. Ariz. campus, Tucson, Ariz. (2 sheets). Salton Sea, Calif. Pecos City, Tex. Agua Caliente, Ariz. Univ. Ariz., Tucson, Ariz. Havilah, Kern County, Calif. Univ. Ariz., Tucson, Ariz. Wilmington, Calif. Ontario, Calif. Hope, N. Mex. Salton Sea, Calif. River. By 1902, when his family left the ranch, an occasional saltcedar could be seen along the Gila River, but the species was not well established. The time of awareness of saltcedar appears to have varied from place to place. Although records (Eakin and Brown, 1939, p. 11-18) indicate the presence of a few seedlings growing on the delta of Lake McMillan on the Pecos River, N. Mex., in 1912, it was not until the late 1920's that the growth commanded attention. Mr. L. E. Foster, Superintendent of the Carlsbad Proj- ect, describes the conditions in the delta area of Lake McMillan under date of July 20, 1928, as follows (Me- Donald and Borland, written commun., 1955, p. 51-52) : "At the present time, the entire upper end of the reser- voir is covered with a dense growth of tamarisk except for a few narrow channels." In the Rio Grande valley, the time of general aware- ness was about 1930. The earliest report of saltcedar growth was in 1910 near Mesilla Park, N. Mex. (Thomp- son, 1958, p. 2). Reports by residents of the valley in- dicate that the plant was uncommon throughout the 1920's and in the early 1980's. Dr. Luna B. Leopold (written commun., 1960) of the U.S. Geological Survey, recalls that his father planted a tamarisk in front of their house in Albuquerque, N. Mex., about 1920, and that the plant was rather un- common. In 1931 his father pointed out a seedling on a road crossing of the Rio Galisteo between Albuquerque and Santa Fe, and remarked that "given time, tamarisk will cover such channels extensively, as it propagates rapidly." Mr. C. C. McDonald (written commun., 1962) of the U.S. Geological Survey, remembers that in 1923, when his family moved from a farm on the left bank of the Rio Grande west of Old Albuquerque, there was some saltcedar growing along an irrigation canal, but none along the river. The recollection of saltcedar stands _ out because of the difference between limbs and twigs of willow and saltcedar when used as fishing poles and for whistles. Willow was very common, having been planted for bank protection, but saltcedar was available at only one spot. Mr. C. L. McGuinness (written commun., 1960), also of the U.S. Geological Survey, reports that his father, a lifetime resident of the Rio Grande Valley, recalls that "as late as October 1931 there were no salt- cedars in the vicinity of the 'lakes' southwest of Las Nutrias, N. Mex." Mr. McGuinness, senior, a sports- man, visited this area frequently as he was part owner and used certain of these lakes for duck hunting. Despite plantings for erosion control in 1926 in J + the tributary streams Rio Salado and Rio Puerco, and A6 perhaps in other localities, the plants did not become important in the plant life of the Rio Grande Valley for nearly 10 years. There was no mention of salt: cedar in a land-classification report by the Middle Rio Grande Conservancy District (McDonald and Bor land, 1955, p. 19) in 1926, although other vegetation was shown. In 1936, the vegetation of the Rio Grande Valley was mapped by the Department of Agricul- ture during the Rio Grande Joint Investigation (Na- tional Resources Committee, 1938). Although saltcedar was present in the valley and was mapped on field sheets, no separate classification was established ; it was included under the heading "Trees-bosque." The author, who was in charge of the ground-water studies in the Colorado portion of the joint investiga- tion, has no recollection of any saltcedar in the San Luis Valley or elsewhere in the headwaters of the Rio Grande in 19836. In the upper Gila River valley in Arizona, salt- cedar may have made its presence known a little before 1920. Mr. Thomas Maddock, Sr. (written commun., 1960), of Safford, Ariz., notes that longtime residents of the Safford Valley date the appearance of salt- cedar from the floods of 1916. The floods caused chan- nel shifts of as much as three-quarters of a mile and denuded large areas of native vegetation. Shortly afterwards, saltcedar was observed growing in the area that had been flooded and denuded. Gillespie Dam on the lower Gila River was completed in 1921. Mr. Thomas Maddock, Jr. (written commun., 1962), reports that by 1929 the lake back of the dam had filled with sediment and a very heavy growth of vege- tation had taken place over most of the reservoir area. Much of this vegetation was saltcedar. Water losses from the Gila River were so great that in 1929 the Gillespie Land and Irrigation Co. began the construc- tion of a drainage ditch on the west side of the river to move water from the end of the Arlington Canal to the face of the dam. In 1931 this drain was extended to a point near the Hassayampa River. This effort is perhaps the first known in Arizona to salvage water normally lost by saltcedar through evaporation and transpiration. Saltcedar appears to have been widely used as hedges prior to 1920. Mr. Thomas Maddock, Jr., recalls a hedge in front of the family home in Williams, Ariz., in 1918. Mr. C. L. McGuinness remembers a well-grown hedge several years old around their house in Albuquer- que in 1921. The species is not known. W. W. Hast- ings of the U.S. Geological Survey recalls a hedge- planting of saltcedar in 1919 at the U.S. Agricultural Experiment Station in Sacaton, Ariz. STUDIES OF EVAPOTRANSPIRATION Mr. Ernest Douglas (written commun., 1962), who left the lower Gila Valley in 1902, returned in 1925 and "for the next four years edited a paper at Mesa, Ariz., but never heard saltcedar mentioned, though it must have been common along the river bottom." In 1929, after becoming editor of the Arizona Farmer-Ranch- man, he noted that saltcedar had become a hard-to- fight nuisance along unlined laterals of the Roosevelt Irrigation District in the Buckeye area and that there were jungles of it along the Gila River. Even then, he . writes, no one recognized it as a consumer of precious water. OCCURRENCE, SPREAD, AND DENSITY OF SALTCEDAR GROWTH Although the information on the occurrence of salt- cedar prior to 1920 is meager, it is sufficient to indicate that the plant did not command much attention in the 43-year period between 1877, when the first specimens were collected in Texas, and 1920. This period of the plant's history contrasts sharply with the next 40 years, when saltcedar was recognized first as a new plant and later as a problem plant. During this latter 40-year period it spread rapidly from one watershed to another and up and down the stream valleys of the Southwest, then northward into the Great Basin and the Rocky Mountains. In 1961, as shown by plate 1, saltcedar was widespread in Arizona, New Mexico, Texas, Oklahoma, Kansas, Colorado, and Utah. Small but well-estab- lished areas of growth occur in California, Nevada, Oregon, Idaho, Montana, Wyoming, South Dakota, and Nebraska. So far as could be ascertained, the plant does not occur naturally in Washington and North Dakota. Data on the rate of spread prior to 1920 is limited to the delta area of Lake McMillan on the Pecos River in New Mexico. There are no records or reports of salt- cedar in this area prior to 1912. - The first report was of a few seedlings in 1912 (Eakin and Brown, 1939, p. 11- 12). By 1915 the plants had spread over an area of about 600 acres of delta land (National Resources Planning Board, 1942, p. 57). In the next 10 years the plants continued to spread over the delta area until by 1925 they covered 12,300 acres. By 1960 the plants covered an estimated 57,000 acres in the 200-mile reach between Alamogordo Dam and the _ New Mexico-Texas State line. Concurrent with the in- crease in total area, there was also an increase in the density and in the cover density (the proportion of an area covered or shaded by the vegetation foliage; usually expressed as a percent). Thus areas of light and medium cover became areas of medium and dense cover,. The increase in the areas is shown graphically in figure 2. SALTCEDAR IN THE WESTERN STATES 60 50 40 Total area / Light, medium, and dense cover 30 7 20 hs adium and // dense cover AREA, IN THOUSANDS OF ACRES 3 #2" 1910 o 1920 \ R 1930 1940 1950 1960 FIGURE 2.-Increase in the area of saltcedar growth and cover density Grande basin of New Mexico has a similar history. along the Pecos River in New Mexico between Alamogordo Dam and the New Mexico-Texas State line. The occurrence and spread of saltcedar in the Rio Al- though it was first reported in the area south of Mesilla Park in 1910 (Thompson, 1958, p. 1502-2), most of the available information covers the 80-mile reach from Bernardo Bridge to San Marcial. In 1918 this reach was included in a topographic and land use survey by the New Mexico State Engineer that covered 150 miles of the valley from Cochiti to San Marcial. No mention is made of saltcedar in the description of the land classi- fication. In describing the "timber" classification the _ survey notes: "The timbered areas are those overgrown _- with timber or brush, usually cottonwoods, willows or _ thorn bushes." A cross-valley profile near San Marcial in 1924 noted the vegetation, but no mention was made of saltcedar. The next survey in point of time was a land-classification survey in 1926 by the Middle Rio Grande Conservancy District. - Here again no saltcedar was reported. According to an unpublished report for the U.S. Bureau of Reclamation (H. R. McDonald and W. M. Borland, written commun. 1955, p. 32-41) dealing with - saltcedar infestation in the middle Rio Grande Valley, _ there was no significant growth of saltcedar prior to - about 1926. About that time, erosion was a problem in _- some tributary streams, and plantings of saltcedar seed- lings were made in 1926 and 1927 at several places in the Rio Salado and Rio Puerco basins and perhaps in other localities as an erosion and silt control measure. The AZT plants spread rapidly after the flood of 1929 and by 1936 had covered about 5,500 acres (sum of the planim- etered areas of parcels of land shown as saltcedar on the maps of vegetative cover. National Resources Com- mittee, 1938), in the Bernardo Bridge-San Marcial reach. Vegetative surveys of the reach were made by the Bureau of Reclamation in 1947 and again in 1955. As part of these two surveys, determinations were made, by species, of the cover density and of the height component of the foliage. Thus it is possible to make comparisons of area of growth, density of growth, and volume of foliage. This area of saltcedar in 1947 was 26,300 acres and in 1955 was 24,800 acres. The decrease from 1947 to 1955 was due to clearing about 10,000 acres of saltcedar for cultivation. These data indicate an increase of about 8,500 acres of saltcedar growth in the uncleared lands. In the 1947 survey, cover densities ranged from 1 to 81 percent and averaged 19.1 percent, while in the 1955 survey, cover densities ranged from 2 to 100 percent and averaged 39.3 percent, an increase of more than 100 percent. At the same time the increase in the volume of foliage was over 75 percent. According to Christensen (1962), there are no rec- ords of Tamarix at Utah Lake, at Great Salt Lake, or on the Colorado and Green Rivers in Utah prior to 1925. He reports that "the period from approximately 1925 to 1960 was one of rapid spread and increase in im- portance of tamarix. The greatest degree of invasion occurred during the twenty-year period from about 1935 to 1955." In the Arkansas River valley of Colorado, Bittinger and Stringham (1963) found a similar history of in- crease in the area and cover density of saltcedar growth. Comparisons of the areas of saltcedar growth by aerial photographs taken in 1936, 1947, and 1957 show quite clearly the progressive invasion of saltcedar. As an example, in the study reach of the valley the area of phreatophytes, largely saltcedar, increased 520 acres in ,, the 11-year period from 1936 to 1947, or at an average | rate of 47 acres a year. In the 10-year period from 1947 to 1957 the average rate of increase was 57 acres a year. Not only was there an increase in area during these periods but also an increase in the cover density. In the aerial photographs in figure 3, taken in 1936 and 1957, the increase in area and cover density in the 21- year period is easily seen. Concerning the spread of saltcedar in Kansas, Mr. P. H. Berg (written commun., 1962), Project Manager, Bureau of Reclamation, states that : * * * Because of experience in other reclamation areas in the southwest with saltcedar, we have kept a close watch in our AS reservoir areas and on irrigation systems for any new infesta- tions of saltcedar that may occur. Within a short period after we complete a dam or reservoir, scattered plants of saltcedar are observed. Each year we observe a few more plants with more mature ones furnishing seed for new infestations. It is logical to assume that were data available for other stream valleys they would show a history of salt- cedar growth and spread similar to that in the Gila, Pecos, Rio Grande, and Arkansas River valleys. AREAL EXTENT The areas of saltcedar growth in the Western States as of 1961 are shown on plate 1. Owing to the small scale of the map, it is not possible to delimit the actual boundaries. Rather the map is diagrammatic, showing the reaches of the streams and included reservoirs, lakes, and playas where saltcedar is known to occur. In com- piling the map, the author attempted to visit and ob- serve areas where saltcedar was unreported, but was suspected to be present. However, it was not possible to examine all of them, so there may be some areas of growth that are not shown. Although these are believed to be few, isolated, and small, they do form a seed source from which the growth may spread. Some observations of saltcedar growth were made at the road crossings of streams, and were limited to the growth seen from the road. The sites of these observa- tions are shown by means of a distinctive symbol on the map. Owing to the aggressive nature of the plant, growth may be suspected for a considerable distance up and down stream from the crossing. Lacking confirma- tion of this, however, the presence of growth at the crossing only is indicated on the map. The largest area of saltcedar, 275,000 acres, occurs in the Pecos River basin of New Mexico and Texas. It was estimated by the Geological Survey ground-water office in Albuquerque that in 1960 some 57,000 acres in the New Mexico portion of the basin was infested. In the Texas portion, on the basis of a general recon- naissance of the basin by the Soil Conservation Service in 1959, there was about 218,000 acres (C. A. Rechen- thin, written commun., 1963). In this portion of the basin according to Mr. Rechenthin, "saltcedar covers most of the bottom lands from the New Mexico line to a point below Sheffield, * * * is found on many tribu- tary streams such as Salt Draw, Toyah Creek, Tornillo Draw and others * * * and is found extensively on 'gyp' soils in the Pecos, Imperial, Fort Stockton, and Girvin areas." The cover density, according to J. S. Horton (writ- ten commun., 1962) is quite variable, ranging from 5 to 100 percent. The area of dense growth, occurring in the flood plain adjacent to the Pecos River channel, was estimated at about 15 percent of the total area of STUDIES OF EVAPOTRANSPIRATION the Texas portion of the Pecos basin. In the remainder of the area the cover density had decreased in the fall | of 1962 to 5 to 15 percent. Horton reports that in these - areas the shrubs are suffering and some have died as the result of either increased salinity in the ground water or a declining ground-water level. A similar condition prevailed shortly after the close of World War II along the Gila River in Arizona for a few miles below Gillespie Dam. Here increased pumpage for irrigation lowered the water table to such an extent that between about 1950 and 1955 much of the saltcedar died. This area of growth is not included on plate 1. No attempt was made to indicate the cover density of the growth on the map. The cover density, however, as indicated above, is known to range from scattered growth of a few percent to 100 percent. Where growth along a stream was not continuous but occurred at in- tervals, it is shown by a broken pattern. On the basis of the assembled information, table 2 was prepared to show the approximate acreage of salt- cedar growth by States. The acreage for each State is not presumed to be exact but is considered sufficiently accurate to indicate the magnitude of the growth area in each State. Neither is the total of 900 thousand acres presumed to be exact, but it is believed to be a realistic indication of the area of saltcedar growth in the Western States at the end of the 1961 growing season. TABLE 2.-Gross areas of saltcedar growth in the Western States as of 1961 Area Arca State (in thousands of acres) State (in thousands of acres) *118 |; North 0 *16' ] 60 Colorado.... __: y 50.) ___ 1 Waho 1" (b) South (b) Kangas: ls 25 | Texas..! i. © 450 (®) s2 000000 : 38 Nebraska;......¢: _.. (P) 0 Neyada.:.:::..-..l..2-. *12 | 1 New Mexico.________. 155 mf tere Total (rounded). 900 a A vegetation survey by the Bureau of Reclamation in 1961 found there was 53,200 acres of saltcedar in the flood plain of the Colorado River from Davis Dam to the international boundary, of which 38,000 acres was in Arizona, 12,700 acres in Cali- fornia, and 2,400 acres in Nevada. These figures included in the State totals. b Less than 1,000 acres. © A brush survey by the Soil Conservation Service in 1963-64 (Smith, H. N. and Rechenthin, C. A., 1964) found that as of June 1964 there was a total of 523,900 acres of saltcedar, of which 273,000 acres were dense stands (more than 20 percent canopy), and 250,900 acres were light to moderate stands (less than 20 percent canopy). WATER CONSUMPTION The term "saltcedar" includes the deciduous species described earlier. Saltcedars are phreatophytes; that is, they depend upon ground water for their water sup- ply. Their occurrence under natural conditions is con- fined to areas where their roots can reach the water FicurE 3.-Aerial photographs, rado. The areas of scattered cottonwood growth o SALTCEDAR IN THE WESTERN STATES taken in 1936 (upper) and in 1957 (lower), of the Arkansas River valley at the Otero-Bent County line, Colo- f 1936 have been filled in and were dominated by a dense growth of tamarisk in 1957. A10 STUDIES OF EVAPOTRANSPIRATION table, such as on the flood plains of stream valleys, on deltas, or along the shoreline of lakes and reservoirs. \ {\, The plants usually grow where the depth to the water table does not exceed 25 feet, and normally where it is \ less than 15 feet. Saltcedars have a wide range of tolerance to saline or alkali soil and water. They have been found growing in Death Valley, Calif., where the ground water contains as much as 5 percent (50,000 parts per million) of dissolved solids. However, it cannot be said they thrive where the concentration of the water approaches the 5 percent limit. Generally they grow best where the ground water is little to mod- erately mineralized. Under optimum conditions the annual rate of con- sumption of ground water by saltcedars is probably the highest of all the phreatophytes The plants have a low economic value, and hence the water used by them is largely wasted. The term "consumptive waste" (that part of consumptive use that is without substantial benefit to man) aptly describes the dispo- sition of this water. The annual rate of use of ground water by saltcedar depends upon several factors such as cover density, size of the plants, depth to the water table, and climatic conditions. Use is greatest where the height and den- sity are at a maximum, the water table lies at shallow depth, and the climate is hot and dry. Studies of the annual rate of use of water by salt- cedar show that under favorable conditions it is more than 9 acre-feet per acre. Experiments with plants grown in tanks in the Safford Valley of the Gila River, Ariz., have shown that at 100-percent volume-density the annual evapotranspiration discharge, not including precipitation, ranged from 9.2 acre-feet per acre when the depth to the water level was 4.0 feet to about 7 acre- feet per acre when the water level was at a depth of 8 feet (Gatewood and others, 1950, p. 137). The aver- age annual use of ground water by saltcedar in the Safford Valley, under natural conditions, was 4 acre- feet per acre. The cover density ranged from scat- tered growth to 100 percent, and averaged 61 percent. At Carlsbad, N. Mex., the annual use of water by salt- cedar grown in tanks was 5.5 acre-feet per acre with a 2-foot water level and 4.7 acre-feet per acre with a 4- foot water level (Blaney and others, 1942, p. 202). The average annual use of ground water by saltcedar in the Pecos River valley, N. Mex., was estimated, on the basis of plants grown in tanks, to be 5.0 acre-feet per acre (National Resources Planning Board, 1942. p. 55). ; EFFECT ON STREAM REGIMEN The regimen of a stream on whose flood plain salt- cedar has become established is usually affected in three ways. There is (1) a depletion of streamflow, (2) an increase in the area inundated by floods, and (3) an | increase in deposition of sediment in the areas of salt- cedar growth. As noted earlier, consumption of ground water by saltcedar is among the highest of all phreatophytes. As a consequence of its draft on the ground-water re- servoir, there is a general lowering of the ground- water level throughout the area of growth. Lowered ground-water levels affect streamflow either by reducing ground-water movement toward a gaining stream or by increasing percolation from a losing stream to the - ground-water reservoir. In either case the result is the same-a reduction in streamflow. In a well-established area of saltcedar along a stream, the plants grow so densely that they choke overflow channels and the flood plain, and so form a partial barrier to flood flows. During periods of flood, this restriction and increased channel roughness cause the water to spread out and inundate areas that normally would not be flooded and thus to endanger lives and damage property. Floodwater is nearly always laden with sediment. The damming or ponding effect of the dense saltcedar growth so reduces the velocity of the water, and thus its power to carry the full sediment load, that much of the sediment is dropped and deposition is accelerated. Substantial deposition of sediment attributed to salt- cedar growth has occurred in the Rio Grande and Pecos Rivers in New Mexico and the Gila River in Arizona. Sediment deposition resulting in part from saltcedar growth is common in the delta areas above reservoirs. FUTURE CONSIDERATIONS The history of the invasion of saltcedar in the West- ern States provides a basis for forecasting its future performance and estimating its areal extent and con- sumptive waste of ground water. Historically, the area and the density of plant growth have increased wherever the species has be- come established. This effect may be expected to con- tinue, wherever area and density of growth have not reached their optimum and wherever new areas become established. At the same time, increased consumptive waste of ground water, increased flood hazards, and continued sediment deposition may be expected. The increase in area and density will not be at the same rate everywhere, but will vary from place to place according to the availability of growing space and to environmental conditions, such as climate, depth to ground water, and degree of alkalinity or salinity. In the stream valleys that now support extensive growth, such as those in Arizona and New Mexico a SALTCEDAR IN THE WESTERN STATES where there is little room for expansion, the areal in- crease will not be large. The density, however, will continue to increase until optimum growth prevails throughout the area. Although no saltcedar was re- ported in the States of North Dakota and Washington in 1961, it is reasonable to expect the plant to make its appearance in the near future. If it is assumed that no control measures will be taken to curb the spread and continued growth of saltcedar, some predictions can be made of the magnitude of the growth area and the consumptive waste of ground water. The area of saltcedar growth at the time of aware- ness of its presence in 1920 must have been small. The meager data indicate that the growth area was about 10,000 acres. Most of this growth was on the delta of Lake McMillan on the Pecos River, N. Mex. By 1961 the area had increased to about 900,000 acres, or at an average rate of about 22,500 acres a year. Projection of this rate over the 9-year period to 1970 indicates an increase of about 200,000 acres. The average rate of increase, however, is much less than the maximum rate. The two areas for which information is available on the average and maximum rates of spread, the Pecos River valley in New Mexico and the Arkansas River valley in Colorado, show that the maximum rate occurred in the decade 1950 to 1960. Thus an increase of 200,000 acres to a total of 1,100,000 acres in the 17 Western States by 1970 would be a minimum. In the Pecos River valley the average rate of in- crease from 1915 through 1960 was about 1,250 acres a year. The rate during the period 1950 to 1960 was double the long-term average, or about 2,500 acres a year. The saltcedar growth in the Pecos River valley dur- ing the 10-year period 1950 to 1960 increased by 25,000 acres, as shown in figure 2. This is 78 percent of the area covered in 1950. In the Arkansas River valley the increase from 1947 to 1957 was about 81 percent of the 1947 area. In the Rio Grande Valley, the increase in the period 1947 to 1955 would have been about 32 per- cent of the 1947 area, had there been no clearing. Ex- pressed as average rates per year, they are 7.8, 8.1, and 4.0 percent respectively. Projection of these average yearly rates over the 9-year period from 1961 to 1970 indicates an increase in the total area of saltcedar of 325,000 to 650,000 acres. On the basis of these estimates of the minimum and the average yearly increase in total area of saltcedar growth by 1970, the increase would range from 200,000 acres to 650,000 acres. The average of these projections indicates a probable increase of 425,000 acres, or a total of about 114 million acres by 1970. All The average annual use of ground water by saltcedar in the Safford Valley, Ariz., was 4 acre-feet per acre, and in the Pecos River valley it was 5 acre-feet. On the basis of an average of 4.0 acre-feet for the entire, . region, the annual draft on the ground-water reservoir / in 1961 was about 3.5 million acre-feet and in 19702: would be about 5 million acre-feet. In 1920 the draft probably did not exceed 50,000 acre-feet. These predictions, as stated earlier, have been made to indicate the magnitude of the saltcedar problem by 1970 and its effect on water supply, under the assump- tion that the spread and growth of the species will not be curbed by control measures. REFERENCES Auerbach, Herbert S., 1943, Father Escalante's Journal, with related documents and maps: Utah Hist. Quart., v. 9, p. 1-142. Bittinger, M. W., and Stringham, G. E., 1963, A study of phre- atophyte growth in the lower Arkansas River valley of Colorado: Colorado Univ. Civil Eng. See., Colorado Agr. Expt. Sta. Blaney, H. F., Ewing, P. A., Morin, K. V., and Criddle, W. D., 1942, Consumptive water use and requirements, Pecos River, New Mexico and Texas, (U.S.) Natl. Resources Plan. Board, Pecos River Joint Investigation-Reports of the participating agencies: Washington, U.S. Govt. Print- ing Office, pt. 3, see. 3, p. 170-230. Bowser, C. W., 1958, Introduction and spread of the undesirable tamarisks in the Pacific southwestern section of the United States and comments concerning the plants' influence upon the indigenous vegetation in Symposium on Phreatophytes, Phreatophyte Subcomm. Pacific Southwest Inter-Agency Comm. Mtg., Am. Geophys. Union, Sacramento, Calif., February 1957 : p. 12-17. Christensen, E. M., 1962, The rate of naturalization of Tamarix in Utah: The Am. Midland Naturalist, v. 68, no. 1, p. 51-57. Eakin, M. E., and Brown, C. B., 1939, Silting of reservoirs: U.S. Dept. of Agriculture Tech. Bull. 524, p. 11-18. Gatewood, J. S., Robinson, T. W., Colby, B. R., Hem, J. D., and Halpenny, L. C., 1950, Use of water by bottom-land vegeta- tion in lower Safford Valley, Arizona: U.S. Geol. Survey Water-Supply Paper 1103, 210 p. Horton, J. S., 1957, Inflorescence development in Tamariz pen- tandra Pallas (Tamaricaceae) : Southwestern Naturalist, v. 2, no. 4, p. 135-139. 1964, Notes on the introduction of deciduous tamarisk : U.S. Forest Service Research Note RM-16 (March 1964). Kearney, T. H., and Peebles, R. H., 1942, Flowering plants and ferns of Arizona: U.S. Dept. Agriculture Misc. Pub. 423, 1069 p. 1951, Flora of Arizona: California Univ. Press, 1032 p. McClintock, Elizabeth, 1951, Studies in California ornamental plants; 3 The Tamarisks: California Hort. Soc. Jour. 12 p., 76-83. f Ramirez, Jose, 1902, Sinonimia vulgory y cientifica de las plantas Mexicanas: Ofecinca Typografica de la Secretaria de Fomento, Mexico. f Robinson, T. W., 1958, Phreatophytes : U.S. Geol. Survey Water- Supply Paper 1423, 84 p. A12 STUDIES OF EVAPOTRANSPIRATION Smith, H. N., and Rechenthin, C. A., 1964, Grassland restoration [U.S.] National Resources Committee, 1938, Regional planning ; the Texas brush problem : U.S.D.A. Soil Conservation Serv- Pt. 6, The Rio Grande joint investigation in the Upper Rio ice, Temple, Texas. Grande Basin in Colorado, New Mexico, and Texas, 1936- Standley, Paul C., 1920, Trees and shrubs of Mexico : Contribu- 37 : Washington, U.S. Govt. Printing Office, v. 1. tions from the U.S. Natl. Herbarium, Smithsonian Inst., v. 23, pt. 1 [U.S.] National Resources Planning Board, 1942, Regional Thompson, C. B., 1958, Importance of phreatophytes in water Pt: 10: Th? Itecos River j.01nt investigation in supply : Am. Soc. Civil Engineers, Jour. Irrig. and Drainage the Pecos RlYer Basin 3 New Meme? and Texas, Sum- Div. Proc., v. 84, no. 1R1, p. 1502-1-150217 (January mary analysis and findings: Washington, U.S. Govt. 1958). Printing Office. U.S. GOVERNMENT PRINTING OFFICE : 1965 0-757-046 UNITED STATES DEPARTMENT OF THE INTERIOR PROFESSIONAL PAPER 491-A GEOLOGICAL SURVEY PLATE 1 128° 127° 126° 125° 124° 123° 122° 121° 120° 119° 118° 117° 16° 115° 114° 113° 112° 111° 10° 109° 108° 107" 106° 105° 104° 103° 102° 101° 100" 99° 98° 97° 47° 50° 46° 48° 44> 43° 38° , a1° 35° 20° maart, beg 1 )' tq? Foros \ * Comin" [engexerr ttt. EXPLANATION Saltcedar Saltcedar intermittent or scattered growth @ Saltcedar growth observed at road crossing of stream 25° TRUE NORTH APPROXIMATE MEAN DECLINATION, 1965 119° 118° 117° 16° 115° 114° 113° 112° 111° 10° 109° 108° 107° 106° 105° 104° 103° Geological Survey Bureau of Land Management U.S. Army-Corps of Engineers Bureau of Reclamation State of Utah 102 101° 100° 99° 98° 97° 96° INTERIOR-GEOLOGICAL SURVEY, WASHINGTON, D. C.-1965-w64207 MAP SHOWING OCCURRENCE OF SALTCEDAR IN THE WESTERN STATES SCALE 1:7 500 000 100 0 100 200 300 MILES C-- --~p------- F- 100 0 100 200 300 KILOMETERS r-- Collings, Myrick-EFFECTS OF JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN-Geological Survey Professional Paper 491=5 7 DAY ffects of Juniper and Pinyon Eradication on Streamflow from Corduroy Creek Basin, Arizona GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-B Prepared in cooperation with the U.S. Bureau of Indian Affairs and the White River Apache Tribe Effects of Juniper and Pinyon Eradication on Streamflow from Corduroy Creek Basin, Arizona By M. R. COLLINGS and R. M. MYRICK STUDIES® OF-EVAPOTRANSPIR AT4ON GEOLOGICAL SURVEY PROFESSIONAL PAPER 1491-B Prepared in cooperation with the U.S. Bureau of Indian Affairs and the White River Apache Tribe UNITED STATES GOVERNMENT - PRINTING OFFICE, WASHINGTON.: 1966 799-428-66 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 20 cents (paper cover) CONTENTS Page Page .:.. .. .s un ebb 2a cha ar rs - awh ees BL | 90002. nan ine onne ss B9 introduction»: 2 _.. _. duc 1 :| Records AVYAlADIGEE L LXII sua ace- 9 Description of es 1 anno she aon 9 Corduroy Creek 3 Procipitation.2 « ~. sus nell al ass 9 Carrizo Creek basin.. ... _. ... _... col nl ual ile 9 Of Ab .. » 2 - 22 cue oo Ln ene nnn abn anld o 10 SireamIlOW.t -c.... 3 SfreatmfloOW.. cL... =~ «ahs . sonne s m mEk bh Liks - aie 10 ModificatiOAIDrOSTAM . . _.. ...- 3 PrecipitatIOn. _. _ . L 2C Ln ne sel nace ah = aca ca v aga 10 Mydrologio ANAIYSIE . .. . . L _ en sl enn enn de doc! G | Selected 10 an tme= sale se 8 ILLUSTRATIONS Page Figurng 1. Index map of Carrizo Creek-Corduroy Creek B2 2. Map showing areas cleared and seeded in the Corduroy Creek basin. 4 8. Graph showing relation between Carrizo Creek and Corduroy Creek runoff before and after treatment-______.. 5 4. Graph showing relation between precipitation and 0 6 5. Double-mass curves of precipitation data, Corduroy Creek 11 6. Double-mass curves of precipitation data, Carrizo Creek 12 TABLES Page ABLE 1. Runoff of Corduroy and CaTTIz2O CIEOLk§_ -.... ans senses bose BT C aAnalyses.:Of COVATIAHCEEL- _- 2-22. .= nam t+ mss aa asan aln 7. O. November through JUuNC precipHAtION.'-.s .- se sss o con eek 10 it: Annual 2 ah nA nse aln a mle 10 10 5. Results of Thiessen method for weighting precipitatiGn IH STUDIES OF EVAPOTRANSPIRATION EFFECTS OF JUNIPER AND PINYON ERADICATION ON STREAMFLOW FROM CORDUROY CREEK BASIN, ARIZONA By M. R. and R. M. MyxricK ABSTRACT An investigation to determine the effect of juniper and pinyon removal and of controlled burning on runoff was made on the adjacent Carrizo Creek and Corduroy Creek basins, Fort Apache , Indian Reservation, Ariz. The watersheds encompass areas of / 237 and 213 square miles, respectively. The study was begun in ~' 1957 with 5 years of streamflow records already existing. Thirty-eight percent of Corduroy basin was modified; Carrizo basin was left undisturbed. There were 7 years of premodifica- tion data (1952-1958) and 5 years of postmodification data (1959-1963). Comparisons were made on the runoff relations from adjacent basins and precipitation-runoff relations over each basin for water-year periods, summer storm periods, and winter storm periods. No statistically significant difference in runoff relations could be detected ; however, a significant differ- ence between precipitation-runoff relations was indicated for the winter storm period on both the modified basin and the control basin. A test of precipitation relations of the control versus the treated basin for the before- and after-modification periods indicated no detectable difference in precipitation be- tween basins. A test of precipitation for the period before versus the period after modification over each basin showed a statistically significant change in both basins; therefore, the change in the precipitation-runoff relations for the before- and \|{after-modification periods was concluded to be the effect of a \'climatic change. The statistically significant change in the precipitation-runoff relations for the control basin was no dif- ferent than would be expected by chance than the change in the precipitation-runoff relation for the modified basin. If a change does exist because of vegetation modification, the change is masked by the variance of the data. Prior to this study the theory had been advanced that if un- desirable species of vegetation were eradicated from a basin, such as the one studied in this investigation, runoff would be ' increased and measurable and additional discrete quantities of water would be made available for appropriation. From the results of this study, however, it cannot be demonstrated that the partial clearing of Corduroy Creek basin resulted in either an increase or a decrease in water yield. INTRODUCTION Juniper and pinyon plant communities occur between altitudes of 4,000 and 6,500 feet above sea level and oc- cupy extensive areas in the Southwest. These trees cover about 74 million acres if the juniper types that extend northward in the Rocky Mountains into Canada are excluded. The trees have increased in number in the last 30 years, possibly because of increased livestock grazing-seeds germinate more readily after passing through the alimentary tract of animals (Arnold and Schroeder, 1955)-and because of the reduced number of forest fires. In 1955 Arnold and Schroeder stated that the en- croachment of juniper is believed to have reduced graz- ing capacities, increased erosion, increased livestock- handling costs, and possibly decreased water yields. The next year Barr and others (1956) estimated the probable increase in water yield that would result from the removal of pinyon and juniper if the land were reseeded with grasses. In 1957 the Carrizo Creek and Corduroy Creek water- sheds on the Fort Apache Indian Reservation were selected as investigation sites for a study of the hydro- logic effects and the probable water-yield change produced by the Bureau of Indian Affairs vegetation- modification program. The criteria used to select the basins were that they be adjacent, about the same size, relatively unmodified by man, and representative of the juniper and pinyon woodlands. It was possible to start basin modification early in the investigation because streamflow and pre- cipitation stations having 5 years of record already existed. This report is based on work done by the U.S. Geo- logical Survey in cooperation with the U.S. Bureau of Indian Affairs and the White River Apache Tribe, Fort Apache Indian Reservation, Ariz. The description of vegetation was aided by suggestions from R. M. Turner, botanist, U. S. Geological Survey. DESCRIPTION OF AREAS The Carrizo Creek-Corduroy Creek area is entirely within the Fort Apache Indian Reservation (fig. 1). The northern boundary of these watersheds is the Sit- B1 B2 STUDIES OF EVAPOTRANSPIRATION ; N e & R Q Pvp ? ¢ Heber / /* i / /1 > e er/ SITG/REAVES Area of this I /_,‘ / Pij 6 ) : t report X .n |/ i. [ ay f < 4 / NATIIbNAL ) Pinedale ? { / c-" t Z/N j T1 . :'FOR}€:/S\’I‘/' ef . 110° v as sl. ~Ty" a &l "f+ Am \ Cleans! : x. 1p f > AZ X - «"*_ Show Low \/. ja / QM f \ p ' j s "137" Zo var n ~ asst ~ ? f _n y'" a png / SF g. Z Cp [=e _> \-. \xf>‘$- h J, ) | « \ {_ _fP & ¥ &* * Forestdale , A« * *i o T ® y 3 ( Tojfeg (M a af j i <6§ & [ k $k Co yal e airy % ) J l“ AE %’\ ang McNary < {*.* 9° ef pes A- | /s. wan EXPLANATION \, /f/ J al gle. & A mn le F om" Precipitation gage « ° ) Bieta NAYAIO {COUNj Eli) e k omy... . Oita C Al ©> precipitaion gage 3 | $ pl rise! AB) pasa AJ. ony Surfac: water 4 a# \/ als t gaging station a wf ) Jz :; cme e 6 mms e ce J L.-" Watershed boundary >/ # «-/ K ( i Pr-O. =p. / \ \Y 110° FiGURE 1.-Index map of Carrizo Creek-Corduroy Creek area, showing location of instruments. greaves National Forest, which borders the Mogollon Rim, and the southern boundary is just north of the mouth of Corduroy Creek on the Navajo-Gila County line. Both drainage areas are in Navajo County and are adjacent-Corduroy Creek lies to the east and Car- rizo Creek to the west. In summer, almost all the precipitation is produced by windborne moisture from the Gulf of Mexico. Winter storms generally originate along the Pacific coast and are carried to Arizona by frontal action. The mean annual precipitation is about 20 inches, although local topographic features may cause significant vari- ations from this average. Summer precipitation occurs during thunderstorms in the late afternoon ; it is fre- quently intense and of short duration and is very local- ized. Winter precipitation occurs as rain and snow; it is usually of low intensity and long duration and is widespread. Drought conditions are common from late April through early July. A distinctive plant-community pattern is recogniz- able in the area. Chaparral grows at 5,200 feet-the lowest altitude in the basins. Above the chaparral, the lower reaches of the watersheds are sufficiently elevated to support the stands of juniper and pinyon pine spe- cies which characterize the next vegetation zone. - Open low forests dominated by these plants occur where soil or topography compensate for the relative aridity to produce habitats with adequate moisture for the growth of these plants. The juniper-pinyon woodlands occur as high as 6,500 feet on the southern slopes of the Mogollon Rim. Utah juniper (Juniperus osteo- sperma), alligator bark juniper (J. deppeana), and pin- yon pine (Pinus edulis) are the dominant conifers; Arizona white oak (Quercus arizonica) and emory oak (Q. emoryi) are the most important live oaks in this vegetation zone. The third and highest (above 6,000 ft) vegetation zone in the watersheds is dominated by ponderosa pine (Pinus ponderosa)-a tall long-leafed pine that con- trasts sharply with the low short-leafed pinyon pine of the zone below. JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN Corduroy Creek drains an area of 213 square miles. The main stream is 30 miles long, and the average chan- nel slope is 68 feet per mile. The slope of the major tributaries to the stream is 101 feet per mile. Stream density, the length of channel per unit area, in the basin is 1.25 miles per square mile. The drainage area is somewhat triangular in shape, the hypotenuse being the southeast side. Altitudes range from 5,350 to 7,300 feet. Basalt crops out in the eastern part of the basin along the main stem, and lava- capped cuestas and sedimentary rocks crop out in the rest of the area. The soil is derived from the Supai Formation, Kaibab Limestone, Cretaceous units, Coco- nino Sandstone, rim gravels, and Quaternary lava. CARRIZO CREEK BASIN Carrizo Creek basin drains an area of 237 square miles. The main stream flows in a southeasterly direc- tion and is 35 miles long. The drainage density is 1.66 miles per square mile. The slope of the major tribu- taries of the main stem is 155 feet per mile. The basin is triangular in shape, the hypotenuse being the southwest side. Altitudes range from 5,200 to 7,400 feet. The bedrock geology and soils are similar to those of Corduroy Creek, except that the Quaternary lava is present only near the mouth of Corduroy Creek. STREAMFLOW The streamflow pattern in the basins shows the inte- grated effect of the climatological and physical char- acteristics of the area. Except for the transbasin diversion from Show Low Lake into the Corduroy basin during the spring and midsummer, the streams rise dur- ing the winter, reflect the heavy precipitation of sum- mer storms, and are very low to dry from early summer through fall. The Carrizo Creek flood plain is 14-14 mile wide in the lower reach and is covered with riparian vegeta- tion which includes many phreatophytes. Perennial flow from the upper reach is lost as ground-water re- charge upstream from the Carrizo Creek gaging sta- tion. The Corduroy Creek flood plain is very narrow and has little or no riparian vegetation. Corduroy Creek is a perennial stream. MODIFICATION PROGRAM To modify the pinyon and juniper vegetation on the Corduroy watershed, the following criteria were used in selecting the areas to be treated: The slope of the land was not to exceed 20 percent; undesirable vegeta- tion was removed only where treatment would cause CORDUROY CREEK BASIN, ARIZONA B3 no damage to the commercial timber stand ; excessively rocky or inaccessible areas were not included in order to keep the eradication program as economically fea- sible as possible. These criteria are in agreement with those recommended by Wilm (1956, p. 210-212). Data on the modification program was compiled from the Corduroy watershed progress report, Bureau of In- dian Affairs (1959). Clearing of the watershed was started in 1957 and completed in 1959. Bulldozers and crawler tractors dragging chains were used to clear most of the area, but some hand cutting, girdling, and grubbing was done. Areas having less than 10 percent native sod were reseeded from the air. Prescribed burning, which is controlled burning of dense under- brush and duff in areas of ponderosa pine, a common practice on the Fort Apache Indian Reservation, was restricted to the Corduroy basin. Prescribed burning was discontinued on the study areas after 1959. Vegetation modification in the Corduroy Creek basin (fig. 2) consists of the steps shown in the unnumbered table. Corduroy Creek basin modification Percent Eradication of juniper, pinyon, Acres of basin and manzanita....; __.... 34, 500 25 Prescribed burning.___._._... 18, 000 13 52, 500 38 Reseeded to grass__________._. 9, 400 T HYDROLOGIC ANALYSIS All analytical methods used herein are subject to the basic assumptions that the data suitably describe the variables and that the relations among the variables is described properly. The validity of the interpreta- tions and conclusions derived from any analysis are related directly to the reliability of these assumptions. , Correlations between monthly runoff from adjacent areas and between precipitation and monthly runoff are | poor. In the following analysis of runoff and precipi- tation, it was necessary to use periods greater than one month. Monthly data cannot be considered as inde- pendent events because of serial correlations. In the analysis, two periods based on storm type were defined. The first period, November through June, is dominated by widespread frontal-type storms; during the second period, July through October, localized thunderstorms prevail. The water year, beginning October 1, was also tested in the analysis so that all chronological periods could be examined and compared. The data (1952 through 1963) were divided into two periods: before vegetation modification (1952-58) and after basin modification (1959-63). The area was modified in 1957, 1958, and 1959. The total area of B4 STUDIES OF EVAPOTRANSPIRATION EXPLANATION Area cleared Area cleared and seeded yses. sets # A * sepa ? MILES ¥ FicUrm 2.-Areas cleared and seeded in the Corduroy Creek basin. modification on Corduroy basin amounted to 38 per- cent; 6.2 percent was treated in 1957, 11.8 percent in 1958, and 20 percent in 1959. The assumption is made that prescribed burning would have the same effect as juniper and pinyon removal. In 1957 and 1958, 47.4 percent of the total 38 percent had been modified or less than half the total modification had been accomplished. During 1959 the remainder, or 52.6 percent, of the modification was accomplished. On this basis the 1957 through 1958 period was considered premodification and the 1959 period postmodification. The first step in the analysis was to relate the runoff from Carrizo basin to the runoff from Corduroy basin by means of a regression analysis. The November through June, the July through October, and the water- year periods for before and after Corduroy basin modification did not vary more than would be expected by chance. Therefore, the runoff relations are not significantly different (tested at the 1 percent level of significance). All the plotted runoff observations (fig. 3) may be best fitted by one line. If one curve were fitted to all observations, 1952 through 1963, its error band (90 percent confidence limits) would overlap both curves (1) and (2) in figure 3. Thus curves (1) and (2) are not farther apart than might occur by chance, considering the variance of the data, even though the 1959 through 1963 line (curve 2) visually indicates an increase in the runoff relations. An increase in water yield on Corduroy basin after removal of vegetation is not indicated by the runoff data. In their report, Barr and others (1956, p. 218) esti- mated that the probable increase in water yield due to juniper and pinyon removal would be 0.46 inch (weighted average per unit of area treated), the initial yield being 1.2 inches. Because 38 percent of Corduroy Creek basin was modified, the increase in water yield CORDUROY CREEK DISCHARGE (R,,), IN INCHES. DRAINAGE AREA 213 SQUARE MILES JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN, ARIZONA B5 19" I I 1 fof ~] I 1 I (3) (1) 1.0 0.5 3s | | | us _; | | | 0.1 0.5 1.0 5.0 CARRIZO CREEK DISCHARGE (R,,), IN INCHES. DRAINAGE AREA 237 SQUARE MILES Curve (1), 1952 through 1958. @, before treatment of Corduroy Creek basin. R..=1.15 R..*. Standard error=0.186 log units, {+37 percent, -27 percent. Curve (2), 1959 through 1963. X, after treatment of Corduroy Creek basin. R..=1.32 R..". Standard error=0.147 log units, +41 percent, -29 percent. Curve (3), Barr and others (1956). Dashed line indicates suggested probable water-yield increase. FIGURE 3.-Relation between Carrizo Creek and Corduroy Creek runoff before and after treatment. B6 STUDIES OF EVAPOTRANSPIRATION - py, Fy] | 10 )/ Y. 1.0 o. 10 & (= (e] 2 ¢ 1 < €" . |= € 3 [- $. a # 2 > a fa ws co frr I L LJ te (a 5 9 > O O S [+3 & a 3 fel '% €] y 6 5 i / 5 2 61 & 0.10 p . 0.10 Ix T g os- * -. $ 2 2 o 5 §. ? $ l- $ 8 & & 0.02 0.02 L JAS I £1 10d | 5 10 20 5 10 20 PRECIPITATION (P,,), IN INCHES PRECIPITATION (F,,), IN INCHES Carrizo Creek basin, November through June data : Curve (1), @ 1952 through 1958. Log R..=-3.234-3.00 log P... Standard error=0.232 log units, +37 percent, -27 percent. Curve (2), X 1959 through 1963. Log log P. Standard error=0.2561 log units, +64 percent, -39 percent. Corduroy Creek basin, November through June data : Curve (1), @ 1952 through 1958. Log Re=-5.47-4+5.03 log Pco. Curve (2). X 1959 through 1963. Log R..=-4.124+4.27 log Ps. Standard error=0.075 log units, +11 percent, - 10 percent. Standard error=0.224 log units, 454 percent, -35 percent. FiGuRB 4.-Relation between precipitation and mnoff. JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN, ARIZONA B7 Tasts: 11-Runoff, in inches, of Corduroy and Carrizo Creeks | data are not significantly different; however, Corduroy Creek basin (the modified basin) and Carrizo Creek Annual November-June « + Hast P Water year basin (the control basin) show precipitation-runoff re- Carrizo Corduroy Carrizo Corduroy lations that are significantly different (at the 90 percent confidence level) during November through June (fig. 1092. 2. 918 4. 043 2. 780 3. 984 The ; ipitation. 5 3 . 585 1 . 489 . 420 1 . 432 4) * Char-lge in prQCIPIta'tlon runoff 12313119115 Of | {I . 526 1 . 698 . 418 . 558 | Corduroy basin from before to after modification, is | i822 ----------- 8 i’é‘; o i253 s iii (2 $3? not different (statistically not significant) from the |/ gong .t" _._. . 402 1. 425 . 362 .330 | change between these relations in the Carrizo basin . 501 1. 624 . 432 . 593 i r Toso . 224 153 A27 ' oss | before and ‘i‘f?er 1th mOdlficatmn' f $960......:...... 2. 233 1 2. 848 2. 185 2. 736 The precipitation relations over the basins was ana- ewok . 198 . 166 175 . 101 F F ' 1822 ___________ tray | oid 447 * 794 1.417 | lyzed by comparing the Corduroy Creek basin precipi- joo... _. . 297 . 296 . 180 . 169 | tation for before- and after-treatment periods with the Carrizo basin precipitation for the same periods. No change in precipitation relations between basins could be detected, but the precipitation over each individual basin for before and after periods was tested and found to vary significantly on both the control and the modi- fied basins. The precipitation was substantially less during the period following treatment. If the precipi- tation is the independent variable (fig. 4), the signifi- cant change in the precipitation-runoff relations for s f both basins could be accounted for by the difference and modified basins were compared. The water-year (decrease) in precipitation over the same periods on period, the period July through October, and the period | each basin. In summary, the change indicated by pre- November through June were tested. The water-year | cipitation-runoff relations over the control and the \[ and the July through October precipitation-runoff re- | modified basins probably resulted from climatic change. | lations show no significant difference in means for | If a change exists because of vegetation modification, - - either basin-that is, the before- and after-modification | it is masked by the variance of the data. (See table 2.) _ 1 Corduroy Creek minus Forestdale Creek diversion. would be 38 percent of 0.46 inches, or 0.18 inches. In figure 3, curve (3) is the increase in water yield sug- gested by Barr and others. This projected increase in water yield, if it exists, could not be detected from the records examined in this report. Curve (3) assumes no significant difference in slopes for the three lines. Next, the precipitation-runoff relations of the control TaBuE 2.-Analyses of covariance [F ratio: * denotes a significant difference at the 10 percent level; ** denotes a significant difference at the 1 percent level. Logarithmic transformations were used in parts 1-5 to obtain linearity and normalization of the data (Cramer, 1946)] DISCHARGE CARRIZO CREEK BASIN (R..) VERSUS CORDUROY CREEK BASIN (Reo) Sum of squares F ratio ) (No significant Source of variance Degrees About Test difference at 90 of freedom regression percent level Rc. ReaReo Reso unless otherwise indicated) Water-year period [See fig. 3] Within each group: o- o ch nven ne unseal seals awe 6 0. 80583 1. 00184 1. 33728 0.09176 || Total means. ..._..._._.... 0.45 1959-63........._. 4 . 78274 . 99490 1. 32050 . 06494 || Regression._..._.__________. 20 f SIODOSeIe_ LEI ..... r .01 Within groups...... f 10 1. 58857 1. 99674 2. 66678 «15700 AMORE MACBNS. . . ...- caned eb crncere cane 1 . 00796 . 00206 . 00053 . 00001 DOLAL. - :+: cu. DUCK AE cua th ae be rope ae 11 1. 59653 1. 99880 2. 66731 . 16489 November through June period Within each group: ol cos be c one bh ods bean aar ao 6 0. 92231 1.17929 1. 74274 0.23492 || Total means._...__..___.... 2. 59 --. - 9~ - » a 4 1. 09971 1.42443 1.92798 08290 || Regression.......____.._.... 1.15 Within groups...... 10 2. 02202 2.60372 3. 67067 . 31796 Among means.. 1 . 01909 . 01453 . 09733 . 08628 c: L 13% LEC ... . SLS LJ CCL 11 2. 04111 2. 61825 3. 76800 . 40048 B8 STUDIES OF EVAPOTRANSPIRATION TABLE 2.-Analyses of covariance-Continued PRECIPITATION-DISCHARGE RELATIONS CORDUROY CREEK BASIN PRECIPITATION (P.;.) VERSUS RUNOFF (Re) Sum of squares Source of variance Degrees About Test F ratio of freedom. regression Pco PeoReo Pco Water-year period Within each group: edie cos » abc wan on niks dee 6 0. 06618 0. 26291 1.33728 0.20284 || Total means....._..._._.... 0. 09 1090-00 eur c s- onn ee d 4 . 00726 . 08996 1. 32949 . 21478 WAbRIN STOUDSL H GL... 10 . 07344 . 35287 2. 66677 . 97128 AMIONE THEANS L 22. 2 o- «cul ors son Prasanna ean 1 . 00028 -. 00039 . 00054 . 00001 . BC cut cach o- one boe aie es arai 11 . 07372 . 35248 2. 66731 . 98199 November through June period [See fig. 4] Within each group: 10924001 ono 20+ 2 2 cite me - I seein a 9 B cain aie cfc aaa 6 0. 06768 0. 34065 1. 74274 0.02817 || Total means_...._____.._... **35. 31 1950-00. -. L2 .OO ASLE Fanaa n Bn oe 4 . 09681 41371 1.92793 +19098 11} Reo MEANS. . 03 Regression **8. 65 Within groups. 10 . 16449 . 75436 3. 67067 AMI4 JI Pro means. *4. 06 Among means. 1 . 06666 02715 . 01105 £10001 41 Slopes.z. ... . 2. . 98 eerie ce- - aie Pe deus be 1 . 23115 «78115 3. 68172 1. 03947 ! CARRIZO CREEK BASIN PRECIPITATION (P..) VERSUS RUNOFF (Ree) [See fig. 4] Sum of squares Source of variance Degrees About Test F ratio of freedom. regression Pca PeaFee Pc. Within each group: Acer - 1 o 2 ugh ai aa ee Se nr she a+ a Bow bean die ceed 6 0. 07257 0.21754 0.92230 0.27019 || Total means...._._.._.___.. *7.10 4 . 08212 27220 1. 09882 . 19657 i 23 10 . 15469 . 48974 2.02112 . 47064 Fo *5. 06 1 . 07842 \ . 03881 . 01920 +0000L I! BIODESL 20 C rana ane es . 07 11 . 23311 . 52855 2. 04032 . 84190 CARRIZO CREEK BASIN PRECIPITATION (P..) VERSUS CORDUROY CREEK BASIN PRECIPITATION (Pco) [Logarithm transformations were not used] Sum of squares Source of variance Degrees About Test F ratio of freedom. regression PeaPco Pco Within each group: 1002 oduct es ch n=n snus on en ae 6 34. 48620 39. 81910 46. 91840 0.91505 || Total means........_....... 1. 68 4 18. 92048 23. 26434 29. 61092 1.01004 !i . 91 10 53. 30508 63. 08344 76. 52932 2. 00044 AMORE MORNSI SC- - oes LLL ol cas. 20.60 1 28. 22742 23. 91881 24. 63078 . 00001 cbc ones sabes 1 76. 62310 87. 00225 101. 16010 2.37277 DISCUSSION The curve relations of runoff from Carrizo basin versus runoff from Corduroy basin (fig. 3) indicate an increase in runoff after basin treatment; however, the scatter of the data, or the poorness of the relations, is such that it would be correct, 90 times in 100, to draw one line through all the data (1952-63). It is con- cluded that an increase cannot be detected from the runoff relations. If the variance of the before- and after-treatment runoff data were constant, that is, if everything were to remain static, the visually suggested change in water yield shown over Corduroy basin would be statistically significant after about 38 years of post- modification data. A statistically significant change in the precipitation- runoff relations is indicated over both basins for the before- and after-modification periods for the winter months. The tests show that there is 90 percent con- fidence in this change; however, the change over Cor- duroy basin is not different from that over Carrizo JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN, ARIZONA basin. If precipitation were 10 inches and if one stand- ard error were used for the variance of the data, Car- rizo basin would have a change in runoff of 1.17 inches which could range from 0.25 to 2.49 inches, and Cor- duroy basin, a change of 1.08 inches which could range from 0.48 to 1.92 inches. Thus, a change on Corduroy (the treated basin), if it exists, is completely masked by the change on the control basin. The precipitation was shown to vary significantly over each basin between the before- and after-modifica- tion periods. The precipitation data over the basins was weighted by the Thiessen method. A multiple- regression method (Linsley, Kohler, and Paulhus, 1949, p. 436) was also used to weight the precipitation data, and the results of the analysis were not changed except that the scatter of the data was slightly increased. During the winter period (November-June) trans- basin diversions were made into Corduroy basin in the 1953 and 1956 water years. These diversions were sub- tracted from the data in this report. The total diver- sion amounted to about 0.047 inches of runoff. No diversions were made in the 1959 through 1963 winter periods. If the diversions were appropriately added to the Corduroy basin runoff, none of the results of the analyses in this paper would be changed. The question arose as to whether the years of clear- ing (1957-59) should be included in the analyses. The total analyses were made using data from 1952-56 for premodification and 1960-63 as the postmodification. The conclusions are not different from those already made in the report. CONCLUSIONS The removal of pinyon and juniper from approxi- mately 38 percent of the drainage basin of Corduroy Creek produced no significant change in runoff. If clearing had been complete, a significant increase might have resulted ; however, as much of the basin was cleared as was considered economically practicable, and this re- striction presumably would also prevent complete clearing of other basins. This data is not to be interpreted to mean that no '\ increase in runoff can result from the eradication of | undesirable vegetation. In the two basins studied here, however, the relation between rainfall and runoff is poorly defined, as is common in arid and semiarid regions, and this natural variability masks any small- scale effects of man's endeavors. In other words, an increase in runoff may result from vegetation modifica- tion, but its magnitude is small and is so masked by other factors as to be indistinguishable. The most im- portant fact is that in the Corduroy Creek basin neither B9 an increase nor a decrease in flow could be proved from the available data and, therefore, no discrete or measur- able quantity of water was made available for appropriation. RECORDS AVAILABLE STREAMFLOW Stream-gaging stations (fig. 1) were established ion Carrizo Creek in June 1951; on Corduroy Creek near its mouth in September 1951 ; on Forestdale Creek diver- sion from Show Low Creek in May 1953; and on Car- rizo Creek above Corduroy Creek in October 1953. Carrizo Creek gaging station was discontinued in June 1961. Data-collecting facilities on Carrizo Creek above Corduroy Creek and on Corduroy Creek near its mouth consist of continuous water-stage recorders and concrete controls; on Carrizo Creek a continuous water-stage re- corder and a natural rock control; and on Forestdale Creek diversion from Show Low Creek a continuous water-stage recorder and a V-notch sharp-crested weir. Records of discharge and runoff have been published as part 9 of the annual water-supply papers of the U.S. Geological Survey. PRECIPITATION The U.S. Weather Bureau operates six precipitation stations in or near the study area. In 1958, the U.S. Geological Survey project personnel established five recording gages at less accessible interior sites to sup- plement the Weather Bureau data. All the precipita- tion stations are shown in figure 1. Data from the five project stations were not used in the analysis, as no record is available for the pretreatment period. The supplemental record shows that a vast network of sta- tions would be required to accurately measure the mean precipitation from convective-type storms. In the Southwest, convective-type storms usually do not produce rain of a general nature or greatly influ- ence peak discharges from larger watersheds. The storms are, however, of utmost importance in the pro- duction of maximum runoff from watersheds of 10 square miles or less or for parts of large watersheds (Dorroh, 1946). Precipitation measurements in the area were based on records from Cibecue, 5,300 feet above sea level; Heber, 6,400 feet; Pinedale, 6,500 feet; Lakeside, 6,800 feet; McNary, 7,250 feet; and Forestdale, 6,200 feet (tables 3, 4). Data for the stations are published by the U.S. Weather Bureau as part of the annual climato- logical data summary. B10 Tarts 3.-November through June precipitation, in inches Year Heber Pinedale | Lakeside | McNary | Cibecue | Forestdale 14. 34 12. 31 19.12 24. 03 14. 46 15.78 8.13 8. 45 10. 68 14. 95 9.88 10.14 8. 47 9. 94 11.73 15. 20 10. 22 10.48 4. 61 5.28 7.290 9. 59 4, 50 7. 58 7.23 7.18 8. 28 13. 38 5.30 8. 43 9.75 7. 57 11. 04 17.12 9.82 9. 66 8. 22 10. 60 11. 58 16. 63 10. 78 10.10 5. 60 4.77 6.11 6.77 3. 90 5. 64 8. 64 8. 04 12.98 14. 72 12. 07 9. 84 4. 88 3. 88 6. 97 8. 36 6.72 4. 41 7.33 7. 00 12. 44 12. 61 12. 92 10. 81 4.30 4.02 8. 22 9. 95 8.34 6.35 TABLE 4.-Annual precipitation, in inches Year Heber Pinedale Lakeside | McNary | Cibecue | Forestdale 22. 87 23. 14 30. 46 35. 45 21.33 26.17 15.75 18. 08 15.37 22. 30 16. 37 14.25 16.12 16. 51 22.39 27. 47 17.25 19. 72 11. 91 13. 25 17.49 21. 07 10. 98 14.47 13. 05 10. 31 18. 05 20. 58 10. 82 13.18 15. 01 18. 27 20. 50 26. 92 15. 91 20. 00 19. O1 20. 87 24. 18 29. 98 22, 00 21. 09 16. 03 15. 63 15. 44 18. 05 17. 09 16. 69 18. 09 18. 89 22. 81 24.18 21.71 20. 48 17.17 15. 81 19.74 21. 14 17. 09 17.70 16.42 14. 57 22. 68 25. 83 20. 25 20. 80 18. 88 12. 57 20. 46 23. 00 16.92 19. 30 RELIABILITY OF DATA STREAMFLOW Daily discharge was computed for the six stream- gaging stations using stage-discharge relations estab- lished for each station. Because the stage-discharge relations were stable, the records are considered ac- curate streamflow measurements, except for brief periods of recorder malfunction. Estimates of flow during these periods of lost record are generally less than 3 percent of the annual streamflow and, therefore, represent a negligible amount of error. The transbasin diversion from Show Low Creek, which was started in May 1953, is pumped from Lake Show Low in the Little Colorado River basin into headwaters of Forestdale Creek in the Salt River basin (Corduroy Creek basin). The volume of water and periods of diversion depend on the storage in Lake Show Low. Records on Corduroy Creek near its mouth were adjusted for the transbasin diversion by subtracting the measured inflow into Forestdale Creek. PRECIPITATION The precipitation data were thoroughly analyzed for consistency as an index of the actual basin precipitation. Each station record was analyzed for any trend or abrupt change in the measured precipitation. The method used to detect these changes was the double- mass-curve technique. The double-mass curves were constructed by plotting the cumulative precipitation STUDIES OF EVAPOTRANSPIRATION for each gage against the average for a group of gages for the same period of record. An abrupt change in the slope of the double-mass relation indicates some inconsistency in the record of one of the gages. The precipitation stations have given consistent records (figs. 5, 6). The average precipitation in the basins was de- termined by constructing a Thiessen network, which is weighted in respect to the areal distribution of the measuring stations. The results, as the percentage of precipitation given to each gage, are shown in table 5. TaBt® 5.-Results of Thiessen method for weighting precipita- tion records Gage Altitude (feet Weighted above sea level) percentage Carrizo Creek Basin Forestdale: ..2.l { 6, 200 7. 92 pimedfile.| 2:24 ed l on ruma ne -_ 6, 500 30. 20 Hebets. o 2 con oul. crece eine e wee 6, 400 26. 28 Cibeoue. 2 HIV _-C 5, 300 35. 60 Corduroy Creek Basin Pinedale. .. 2... 3.00384. 6, 500 8. 6 MeCNary 2.2.2 nude aetna nah a Ras 7, 250 8. 7 Lakeside. _cous occur .od 6, 800 12. 5 Porestdale . 2 : . .. =% ree ae c= ar anand 6, 200 75.2 SELECTED REFERENCES Arnold, J. F., and Schroeder, W. L., 1955, Juniper control in- creases forage production on the Fort Apache Indian Res- ervation: Rocky Mountain Forest and Range Expt. Station, Paper no. 18, p. 1-85. Barr, G. W., and others, 1956, Recovering rainfall: Report of Arizona Watershed Program, Univ. Arizona, Tucson, pt. 1, 83 p., pt. 2, 218 p. Bureau of Indian Affairs, 1959, Corduroy watershed progress report: Whiteriver, Fort Apache Indian Agency, 24 p. Cramer, Harald, 1946, Mathematical methods of statistics: Princeton, N.J., Princeton Univ. Press, 575 p. Dixon, W. J. and Massey, F. J., Jr., 1957, Introduction to sta- tistical analysis [24 ed.]: New York, McGraw-Hill Book Co., p. 112-138. Dorroh, J. H., Jr., 1946, Certain hydrologic and climatic char- acteristics of the Southwest: Univ. New Mexico Press, 64 p. Ezekial, Mordecai, 1941, Methods of correlation analysis: New York, John Wiley & Sons, 535 p. Linsley, R. K., Jr., Kohler, M. A., and Paulhus, J. L. H., 1949, Applied hydrology: New York, McGraw-Hill Book Co., 689 p. Schneider, W. J., and Ayer, G. R., 1961, Effect of reforestation on streamflow in central New York: U.S. Geol. Survey Water-Supply Paper 1602, 61 p. I Sellers, W. D., ed, 1960, Arizona climate: Univ. Arizona Press, 60 p. 280 240 200 160 120 CUMULATIVE PRECIPITATION, IN INCHES 80 40 Stoyanow, A. A., 1936, Correlation of Arizona Paleozoic forma- tions: Geol. Soc. America Bull., v. 47, no. 4, p. 533-536. Wilm, H. G., 1949, How long should experimental watersheds be calibrated?: Am. Geophys. Union Trans., v. 80, no. 2, p. 272-278. 1956, Treatment of vegetation on the Salt River water- JUNIPER AND PINYON ERADICATION, CORDUROY CREEK BASIN, ARIZONA B11 YEARS O 0 < u w m co o («] ea o & & F & & & & & & & t r- «- - r- «- «- - +- «- « «- | I I / McNary / L / Lakeside 2 ® Forestdale ¢Pmdale o A a> (3 y ® -I | | | I I 40 80 120 160 200 240 CUMULATIVE PRECIPITATION, IN INCHES, FROM AVERAGE OF FOUR INDEX STATIONS, 1952-62 Fieur® 5.-Double-mass curves of precipitation data, Corduroy Creek basin. shed to increase water yield, in Recovering rainfall : Report of the Arizona watershed program, Univ. Arizona, Tucson, ' pt. 1, 33 p., pt. 2, 218 p. Wilson, E. D., Moore, R. T., and O'Haire, R. T., 1960, Geologic map of Navajo and Apache Counties, Arizona ; Arizona Bur. Mines, Tucson, Ariz. B12 IN INCHES CUMULAT'VE PRECIPITATION, 240 200 160 120 80 40 STUDIES OF EVAPOTRANSPIRATION YEARS a m s w w m co a o va o & a * &. a & &a & & & & <- <- v- «- - «- Lal r- v- «- «- I | I I I I I | Forestdale #: | Cibecue Hleber Pinedale 2. / 7 # @ V @ | , 40 80 120 160 200 CUMULATIVE PRECIPITATION, IN INCHES, FROM AVERAGE OF FOUR INDEX STATIONS, 1952-62 FIGURE 6.-Double-mass curves of precipitation data, Carrizo Creek basin, 240 7 DAY __Composition of Saline Residues on Leaves and Stems of Saltcedar (Tamarix pentandra Pallas) GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-C Composition of Saline Residues on Leaves and Stems of Saltcedar (Tamarix pentandra Pallas) By JOHN D. HEM $ TU DIES OF EVAP OT RANS PIR AT IO N GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-C Analyses of saline deposits leached or washed from saltcedar plants UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1967 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 20 cents (paper cover) CONTENTS Page *.. aes siles - C1 | Interpretation of results-Continued Uptake and circulation of inorganic ions by 1 Relation of leached solutes to ground-water quality . Collection Of tals" 2 Effects of other Interpretation of results. 6:1 Conclugi0n8-. ... -. Nature and amount of solutes......._..___.__.__._.. 6:1 Literature cited._...2..2.2! :r club ILLUSTRATION Figur® 1. Graph showing average sulfate to chloride ratio in solutes washed from saltcedar leaves, Pecos River basin, MN. Ll.... 2... ILL.. ul anna nac as ne aln f an aln tin a's al tms 6 t Wile a mmc in aw nne w met ss ws mn TABLES TaBLs 1. Composition of solids leached from saltcedar leaves; 16 to 24 hours contact time_________________-___---------- 2. Composition of solids washed from saltcedar leaves, few minutes contact time, and of solids dissolved in nearby - .. 2 . 2 o ce o o le la a e aa a a Le ia tee ae n hs oe wn i a ane we f ht Ne ht i on mal a as e an he e nn mee ae nn ae boe e e e me At a in he on me og He ol e hee i me in o as he t ae in im e he in 3. Miscellaneous analyses of water associated with III 236-478-66 Page w t \[ o Page C8 Page C3 O i> STUDIES OF EVAPOTRANSPIRATION COMPOSITION OF SALINE RESIDUES ON LEAVES AND STEMS OF SALTCEDAR (TAMARIX PENTANDRA PALLAS) By Jounxn D. Hex ABSTRACT Saltcedar leaves and stems collected along the Gila River in Safford Valley, Ariz., and at various locations in the Rio Grande and Pecos basins of New Mexico, were analyzed for inorganic ions. The total content of cal- cium, magnesium, sulfate, and chloride in the leaves generally ranged between 5 and 15 percent of their dry weight. From a few tenths of a percent to more than 3 percent of the dry weight of the leaves consisted of inorganic ions which could be washed off by rainfall. In most samples, the predominant inorganic anion was sulfate. Leaves highest in sodium and chloride content were obtained from trees growing where ground water had a high salinity. The amounts and composition of inorganic solutes which are present in saltcedar leaves are influenced by numerous factors including composition of water avail- able to the plant, the time of year, the rates of growth and transpiration before sampling, and the amount and frequency of antecedent rainfall. UPTAKE AND CIRCULATION OF INORGANIC IONS BY PLANTS The processes whereby plans absorb inorganic ions from soil solutions and the means which certain species of plants have for coping with excessive salinity in water and in soil have been studied for a long time. The published literature on these subjects, however, indicates that although some understanding has been gained, the details of the processes involved are far from being com- pletely understood. Some of the major constituents dissolved in natural water, in particular calcium, magnesium, sulfur, and carbon, are essential for plant growth. Most of the carbon required for carbohydrate synthesis is taken from atmospheric carbon dioxide, however, and the oxy- gen and hydrogen required are supplied from the water itself. Chlorine is required for plant growth only in small quantities and sodium apparently is essential for some but probably not all plants. Other essential ele- ments such as iron are commonly present in smaller amounts in natural water (Sutcliffe, 1962, p. 6). It is to be expected that plant parts where metabolic proc- esses are actively taking place, especially the leaves, would contain all the elements needed by the plants. Other elements extracted with the nutrients from the medium in which the plant is growing also might be present. Some plant species are noted for their ten- dency to accumulate elements present in minor amounts in the soil. Certain species of Astragalus, for example, accumulate selenium in large amounts (Anderson and others, 1961, p. 35). Sutcliffe (1962, p. 5) presented data he had obtained from the literature showing that specimens of stonewort Chara ceratophyllae, contained similar amounts of cal- cium, magnesium, and potassium regardless of the com- position of the medium in which they were grown but that amounts of sodium and chloride taken up did de- pend to some extent on the concentration of these ele- ments in the available water. He commented, "In par- ticular, sodium is present at a low concentration in the above-ground parts of many plants, irrespective of the amount of sodium in the soil * * *." Evidently some plants reject sodium, and others may take it up only incidentally in obtaining other nutrient ions. Saltcedar is the common name applied to two decid- uous species of the genus Tamariz (T. pentandra Pal- las and 7. gallica Linneaus) which have become nui- sance plants along streams in parts of the Western United States (Robinson, 1965). One of the distinctive traits of the saltcedar is its ability to thrive in areas where the ground water, from which its water require- ments are met, has a rather high dissolved-solids con- centration. The plant does not appear to be a true halophyte in the sense that it requires a saline environ- ment, but it can thrive in places where the growth of C1 C2 | | most other vegetation is strongly inhibited and thus shares some of the properties of the halophytes. The assimilation of water by plant roots involves the passage of water molecules through root membranes by osmosis, movement of water being toward the solution of higher solute concentration. Some solute ions pass with the water through the root membranes and gen- erally supply the nutrient needs of the plant. To obtain water from the soil, plants must overcome two forces. The first of these, soil-moisture tension, causes water films to adhere to soil particles. The other force is the osmotic pressure required to separate solvent molecules from a solution, and this force increases as the concentration of dissolved solids in the solution rises. Thus, as the salinity of available water around a plant's root system increases, the osmotic pressure required to extract water molecules becomes greater and greater. If the salinity of the soil water is too great, the plant will succumb from lack of water because it is unable to attain a high enough osmotic pressure. Halophytic plants in general can exert a high osmotic pressure. According to Sutcliffe (1962, p. 156), this is mainly because of high sodium and chloride content of the cell sap. In some plants of this class, the salt concentration of the cell sap rises throughout the grow- ing season, but in others the concentration is maintained at a fairly constant level by certain regulatory mecha- nisms. The presence of large amounts of water in the tissues of succulent plants is one mechanism for prevent- ing excessively high dissolved-salt concentrations with- in the plant. Many halophytes are succulent plants. More freely transpiring halophytic plants, however, require other mechanisms. Certain plants, including saltcedar (Decker, 1961), have organic structures termed "salt glands" on their leaves which can excrete fairly concentrated solutions. The salt in this solution is left behind on the leaf surface when the water is evaporated and may be washed from the plant by rain or blown off hy wind. | The manner in which salt glands remove or concen- trate salt from the internal solutions of the plant is not fully understood, although Sutcliffe (1962, p. 161) thought it probable that the glands perform osmotic work. If so, the plant actually expends energy in the process of eliminating salt. A fairly detailed anatomi- cal description of the salt glands of saltcedar has been given by Campbell and Strong (1964). Another process by which a plant may rid itself of saline material is the process of guttation. When tran- spiration ceases, at the end of the day, the solutions in the plant circulatory system may continue moving up- ward and outward to leaf surfaces, and drops of mois- ture may thus appear on the leaves. This moisture, STUDIES OF EVAPOTRANSPIRATION although it has the appearance of dew condensed from the atmosphere, can be highly charged with salt derived from residues left in the leaf structures from the day's transpired solution, as well as from residues on the leaves from the activities of the salt glands and the evaporation of prior guttation. The green leaves of saltcedar plants commonly carry small cubic crystals which are readily seen with the aid of a hand lens. Although the appearance of the crystals suggests they are nearly pure sodium chloride, the salt brought to the leaf surfaces also includes considerable amounts of other ions. The salt is transferred to the ground by rainfall or wind. Actual quantities of salt involved in such transfer mechanisms in areas of dense saltcedar growth have never been determined. Qualita- tive evidence includes the absence near saltcedar of vegetation having a low salt tolerance and the high cor- rosion rate of metals exposed in saltcedar thickets. Decker (1961) noted the salt crystals on the foliage of saltcedar and the formation of whiskers of salt at the sites of the salt glands on the leaf surfaces. In his ex- periments, these salt deposits proved to be deliquescent and went into solution when the twigs were exposed to humid air. Drying caused cubic salt crystals to form on the leaves. Decker believed the liquid observed on saltcedar leaves in the field, which had been ascribed to guttation (Gatewood and others, 1950), was actually the result of the condensation of atmospheric moisture. COLLECTION OF DATA During investigations of water quality in Arizona and New Mexico in which the writer participated in the 1940's, the presence of salt crystals on saltcedar foilage was noted on many occasions, and from time to time several types of experiments were performed on samples of foilage to determine the quantity and nature of the saline material. The results of these experiments were not highly consistent or simply interpretable, and most of them were never published. Owing to increased interest in the processes whereby salt-tolerant plants are able to grow and in the details of the transpiration proc- esses of the saltcedar, the data have been reexamined and are presented here as an aid in understanding some features of the solute circulation of the saltcedar and to aid in planning further studies of the process. The samples usually consisted of green leaves and associated stems. Some collections were made after the end of the growing season when the leaves were no longer green, but most specimens were actively growing when collected. Generally, a sample of about 25 grams was obtained in the field by clipping leaves and stems from a single plant. In the studies of 1940 and 1944, the samples were brought to the laboratory and allowed COMPOSITION OF SALINE RESIDUES ON LEAVES AND STEMS OF SALTCEDAR to dry in the open air. About 2 g of this material was then oven dried at 110° C for 1 or 2 hours. The weight loss from oven drying, generally between 10 and 15 percent, was ascribed to moisture not removed by air drying. The plant material lost most of its green color during oven drying, however, so some breakdown of the organic material probably also took place. About 5.0 g of the air-dried material was placed in a beaker, 100 milliliters of distilled water was added, and the solution stood with occasional stirring for 16 to 24 hours. In some of the experiments, the leaching was done at room temperature; in others, the solutions were kept warm on a water bath. The temperature at which the leaching was performed did not affect the results significantly. C3 After the leaching, the solution was filtered and quan- titatively transferred to a 250-ml volumetric flask. The flask was then filled to the mark with distilled water, and the solution was analyzed for calcium, magnesium, sulfate, chloride, and bicarbonate content and for total dissolved material. Results were calculated as percent- ages of the air-dried sample weight. A large part of the material leached from the leaves and stems was organic, and the final solutions generally were strongly colored. The inorganic ions which were leached probably represented not only the saline de- posits on leaf surfaces but also any ions that were pres- ent anywhere in the leaf tissues in combination with organic ligands. The analytical results are given in table 1. TaBL® 1. Composition of solids leached from saltcedar leaves, 16 to 24 hours contact time Percent of air-dried sample Sample] Residue after 1 g 3 4 (812g; 21311151233; Description evaporation | Galcjum | Magne- | Sulfate | Chloride (heated for 1 stam hr at 180°C) A.-Vegetation along Pecos River between Artesia and Carlsbad, N. Mex., sampled March 4-7, 1940 1 4.9 AT 1.8 | New growth on tree at bridge on New Mexico Highway 83, east of Artesia, N. Mex. 2 3.1 2.2 1.4 | New growth on tree on river bank near Carlsbad Spring, Carlsbad, N. Mex. 3 6. 0 2.6 2.3 | Dead leaves from previous season on tree at bridge on New Mexico Highway 83, east of Artesia, N. Mex. - Average of two samples. B.-Vegetation in Safford Valley, Ariz., sampled November 1944 4 28.5 1.7 0.7 5.2 0.8 8. 4 6.5 | Saltcedar growing in tank 16, Glenbar experiment station, Arizona. Water table held at 6.0 ft below surface. 5 26.5 2.4 .5 74 1.0 11. 0 7.1 | Saltcedar growing in tank 21, Glenbar experiment station, Arizona. Water table held at 6.0 ft below surface. 6 26.8 1. 4 .6 4.7 .9 7.6 5.2 | Saltcedar growing naturally near tank 16, Glenbar experiment station, Arizona. 7 26. 8 1.3 .5 5. 4 4.1 11.3 1.3 | Saltcedar growing beside drain ditch from Knowles flowing well, Geronimo, Ariz. 8 23.2 -F. 1.2 4.3 1.2 7.4 3.6 | New growth on young saltcedar near well 18-79, SWINE see. 34, T. 4 S., R. 23 E. Graham County, Ariz. 9 24. 2 1.1 1.0 4.8 1.0 7.9 4.8 Méturetgroxtgh on saltcedar 16 ft tall, sec. 34, T. 4 S., R. 23 E. Graham ounty, Ariz. 10 24. 8 .6 .4 2.4 4.8 7.9 .4 | New grovéth in area cleared 2 months previously, near Fort Thomas, Ariz. 11 26. 8 .9 .8 3.2 4.2 9.1 .8 | Mature growth just outside cleared area near Fort Thomas, Ariz. 12 26. 0 1.4 * 4.7 1.8 8.6 2.6 | Mature growth on Gila River near Geronimo, Ariz. C.-Vegetation in Pecos River basin, N. Mex., sampled November 1944 13 28.9 2.8 0.9 9. 4 1.6 14.7 5.9 | Saltcedar on Pecos River near Frazier, N. Mex., bridge U.S. Highway 70. 14 31.0 2.2 1.9 8.7 4.3 17.1 2.0 | Saltcedar on Pecos River near Artesia, N. Mex., New Mexico Highway 83 bridge. 15 20. 9 2.1 1.1 9.3 2.8 15. 4 4.0 | Saltcedar on Pecos River near Major Johnson Springs below McMillan Dam, N. Mex. 16 28. 2 1.5 .9 6.6 28 11.3 2.9 | Saltcedar in Malaga Bend area, New Mexico, Pecos River. 17 23.9 .8 t 4.1 3. 0 8.2 1.4 | Saltcedar in Malaga Bend area, New Mexico, 150 yds downstream from No. 16. 18 26. 3 2.3 14 8.7 1.0 13.1 8.7 | Saltcedar on Delaware River at U.S. Highway 285 bridge, south of Malaga, N. Mex. D.-Vegetation in Rio Grande basin, N. Mex., sampled December 1944 19 9. 4 1.7 0.5 4.3 1.4 7.9 3.1 | Saltcedar beside interior drain, San Acacia, N. Mex. Leaves brown from recent frost. 20 13.5 .8 .6 3.6 1.3 6.3 2.8 | Saltcedar beside interior drain, San Acacia, N. Mex. Leaves still green. 21 22.8 .9 +4 5.3 2.0 8.9 2.6 | Saltcedar on old U.S. Highway 85 in Bosque del Apache Wildlife Range near San Antonio, N. Mex. 22 18. 6 .8 +2 2.1 1.3 5.0 2.1 | Saltcedar in arroyo 114 miles south of Truth or Consequences, N. Mex. 28 16. 3 1.1 . 3 2.9 1. 4 5.7 2.1 | Saltcedar by road 20 yd below No. 22, 114 miles south of Truth or Consequences. E.-Vegetation growing where permanent water table is very deep, sampled December 1944 24 11.8 0.6 0.6 22 0.5 3.9 4. 4 | Lone saltcedar in highway borrow pit, 5 miles west of Hatch, N. Mex. 25 15.1 .8 .4 2.6 ...... _ 2 ue cles . 43 . 44 Specific conductance, in micromhos at 25°C... 55,500 3, 260 1. Water accumulated on saltcedar foliage near Glenbar, Ariz., during night preceding Oct. 18, 1944. 2. Water from well 11-14 near Glenbar, Ariz., Aug. 30, 1943. INTERPRETATION OF RESULTS The analytical data should be capable of providing in- formation bearing on the following three questions : 1. How much and what kinds of solutes are present in and on saltcedar leaves ? 2. What relation does the amount and nature of solutes in the leaves have to the chemical quality of ground water in the vicinity of the plant ? 3. What effects do other environmental or internal fac- tors have on the composition of solutes in the leaves? NATURE AND AMOUNT OF SOLUTES The amount and composition of material leached from the saltcedar leaves obviously are functions both of the amounts present on the leaf surfaces or within the plant tissues and of the effectiveness of the method used to bring them into solution. The rather rigorous leach- ing used for samples listed in table 1 probably explains the fact that this procedure gave soluble contents from 9.4 percent to more than 30 percent of the dry-foliage weight. - The simple washing of leaves used for samples listed in table 2 gave soluble contents ranging from a few tenths of a percent to a little over 3 percent, or about a tenth as much as the leaching treatment. The anal- ysis in table 3 of the moisture naturally appearing on the leaves is obviously influenced by many extraneous factors and cannot be related to leaf weight or area. The solutes in this water are the most transient part, being subject to rapid removal by wind action. The methods used to obtain the data in table 2 give results that are indicative of the amounts of soluble salt which might be washed from saltcedars by rainfall dur- ing the growing season. The more rigorous leaching procedure used for samples in table 1, however, probably approximates the total amount of the determined ions which can be extracted from leaves which fall to the soil surface and decompose. The sodium content of the leaf extracts in table 1 was STUDIES OF EVAPOTRANSPIRATION not determined ; and because of the large amount of or- ganic matter in the solutions, a meaningful value cannot be calculated from the ion balance. Therefore, ion per- centages cannot be calculated for these data. The ratio of sulfate to chloride reported for the analysis, however, does have some usefulness in comparing data in table 1 with that in table 2. The sum of inorganic ions reported in table 1 is gen- erally no more than half the total soluble residue. Al- though sodium was not determined, it could account for only a small part of this difference. No significant con- centrations of bicarbonate were found in those extracts which were analyzed for this constituent. For most samples, the four ions determined comprised from 5 to 15 percent of the dry weight of the leaves. Although one might casually expect the inorganic solutes extracted from saltcedar leaves to be mostly sodium and chloride, the results of the leaching and washing experiments show that sulfate is generally the predominant anion. Surficial moisture from the leaves, however, may contain principally sodium and chloride as indicated by the single analysis of this material in table 3. It seems probable that sulfate is held to a greater extent within the leaf structure, and chloride to a greater extent is transported to the outer surface of the leaf from which it can be periodically removed by wind and rain. This inference is supported by the gen- erally higher sulfate to chloride ratio for the more in- tensively leached samples (table 1) than for those sim- ply washed with distilled water (table 2). RELATION OF LEACHED SOLUTES TO GROUND- WATER QUALITY The kinds and amounts of inorganic solutes on the sur- faces of saltcedar leaves, as well as the amount present within the leaf structures and available for leaching, are influenced by numerous variables. In a general way, one would expect the composition of the leachings to be related to that of the water available to the plant roots, modified by whatever selective processes the plant might be able to exercise to absorb or reject certain solute ions. The composition observed would then be further modi- fied by such factors in the sampled part of the plant as the rate at which water has been transpired and the rate of metabolic processes. More external but still impor- tant factors affecting the composition of leachings might include the length of time which has passed since the last rain or windstorm, the season of the year, and the man- ner in which the leaves for sampling are selected. It probably is not surprising that very little can be deducted from the data in tables 1 to 3 to indicate the effect of the composition of the water supply. In many places, the composition of the water available to the COMPOSITION OF SALINE RESIDUES ON LEAVES AND STEMS OF SALTCEDAR plants, from which leaf samples were taken, is very im- perfectly known. In table 2 an analysis of ground water or of low-flow water from the adjacent stream is included for each sample location. - Somewhat similar information is available (Hem, 1950) for some of the sampling sites in Safford Valley. The actual composi- tion of moisture available to plant roots may differ somewhat from the ground-water composition in the vicinity; however, until more is known about root-zone behavior of solutes, it is useless to speculate about this question. Gross differences in ground-water composition do seem to be reflected to some extent in the composition of the extracts in table 1. Sample 7 was obtained from a saltcedar growing along a drainage ditch which had ap- parently been used for some years to convey the unused salty water from a small flowing well to a nearby wash channel. The chloride concentration of this water was 6,800 ppm (parts per million), and its dissolved solids were 14,000 ppm. The percentage of chloride in the leaf sample was higher than for most others in table 1, which represented for the most part trees having a less saline water supply. Samples 10 and 11, also having high chloride percentages, came from trees growing near Fort Thomas, Ariz., in an area where the ground water probably had between 3,000 and 4,000 ppm of chloride. The remaining two samples (Nos. 28 and 30) which showed very high chloride contents were from trees growing under a bridge, where the leaves were protected from rain, and therefore they are not really comparable to the others. The total percentages of inorganic ions reported in table 1, although not including sodium, do seem to have a systematic trend. The highest totals generally are found in group C, which represents trees growing in the Pecos River basin. The dissolved-solids concentration of ground water available to the plants probably was considerably higher in that area than in the Rio Grande basin. The lower total ion contents of the leaves in group D are in accord with this interpretation. Table 2 also shows some effects related to composition of the water supply. The tree growing near well 5 in the Malaga Bend area where the shallow ground water is very saline gave consistently higher chloride per- centages in the leaves than did trees at other locations. It also had a sulfate to chloride ratio below 1.0 con- sistently through the growing season. The sulfate to chloride ratio of the surficial moisture from saltcedar leaves (table 3) is nearly the same as that of ground water in the vicinity. Although this may be fortuitous, the ratio for leaf samples from nearby trees (table 1, sample 6) is much higher and suggests the sulfate is mostly present within the leaves. CT EFFECTS OF OTHER FACTORS Several analyses in table 1 give some indication that the younger, more vigorously growing leaves contain somewhat larger proportions of chloride than do mature leaves. Pairs of analyses showing this effect are those for samples 8 and 9 and samples 10 and 11. The first in each pair represents new growth and has a significantly lower sulfate to chloride ratio than the second, which represents mature growth. The successive samples in table 2 give some general indications of effects of rainfall and passage of time during the growing season. A major factor affecting the results is the transfer of loose salt deposits from leaf surfaces to air or ground by wind movement, by rain, or by the drip of moisture accumulated on the plant leaves in other ways. Surficial deposits of sodium chloride, being loosely attached and readily soluble, probably are rather easily removed by wind and by dripping of condensed moisture or guttation. Salt exuded from salt glands in the form of so-called whiskers (Decker, 1961) would readily become air- borne. Even in the absence of heavy rain, therefore, it might be expected that total content of soluble ions and the sulfate to chloride ratio would tend to increase with time. The effect of rainfall is evident in some of the anal- yses. A moderate to heavy shower should remove all readily soluble material from the leaf surfaces ; if some of this intercepted rainfall actually enters the plant di- rectly, as may happen, the solution within the leaf tisues may be diluted somewhat. In any event, several days might have to pass after the rain before a normal deposit of saline material would again build up on the leaf surfaces. Leaves collected soon after a rain should contain a lower percentage of solute ions than leaves which had not been rained upon for some weeks. Samples of leaves obtained July 10 and 11, 1945, from the locations at and downstream from Artesia, N. Mex. (table 2), were influenced by a general rainstorm which had occurred in that area a few days before. For all the sites where rain had occurred, the percentage of solute ions in the leaves obtained in July was lower than at the previous sampling. Very little precipitation oc- curred in the Pecos basin during the months of April, May, and June of 1945. Figure 1 shows the average ratio of sulfate to chloride for each of the four sets of samples obtained during the growing season of 1945 (table 2). The general trend in the ratio was upward until after the July sampling. In common with other properties related to the plant- growth pattern, this ratio and the soluble percentage of the dry weight of the leaves would both be expected to increase most rapidly during the period of highest C8 STUDIES OF EVAPOTRANSPIRATION 5.0 (£1944) 4.0 on fs erf - [xe irie - oar Si / \\\\\ ae 3.0 % is \\ & | sc 9 |o / \\ 2.0 \ \ \ | e | // | | 1.0 f 6 0 Mar. Apr. May June July Aug. Sept. Oct. Nov. 1945 Fieurm 1.-Average sulfate to chloride ratio in solutes washed from saltcedar leaves, Pecos River basin, N. Mex. Heavy rainfall preceding October 1945 sampling probably caused an abnormally low sulfate :chloride ratio. growth rate and to level off later in the growing season when growth rates are lower. The last sampling, how- ever, which was made in late October, showed a very pronounced decrease in the ratio, to a little under 1.0. Precipitation in the Carlsbad-Artesia area in October 1945 was unusually heavy and probably accounted for the ratio decrease. From this rather fragmentary in- formation, one cannot be certain that the decrease was the result of rainfall, but the average ratio of sulfate to chloride for samples obtained in November 1944 in the Pecos River basin (table 1) was a little over 4.0. The data probably are not completely comparable, but the general seasonal trend and the effect of rainfall would be interesting subjects for further study. In table 1, there are two samples of leaves obtained from trees growing under the county highway bridge across the Gila River at Safford, Ariz. The trees in this location were protected from rain, although they were probably still losing salt from their leaves in other ways. Both samples 28 and 30 contained at least 5.0 percent of chloride and thus had the highest concen- trations observed in any of the samples represented in the table. CONCLUSIONS The total inorganic solute content of saltcedar leaves and deciduous stems may exceed 15 percent of the air-dry weight. The predominant anion in the analyzed sam- ples was generally sulfate, although chloride was also abundant and was the predominant anion in a few. The composition of total solute content of leaves was not closely related to composition of the ground water in the area where the plants were growing. It seems likely, however, that chloride must be strongly predominant in the water available to a plant to maintain a value below 1.0 for the sulfate to chloride ratio in the leaves. The solutes which can be washed from the leaves by fairly brief contact with water, as during a rainstorm, constitute as much as 3.5 percent of the dry weight of foliage of plants growing where the water supply is high in dissolved solids. Where the available water is lower in dissolved solids, the solute content of the leaves is lower. More rigorous leaching may bring into solution as much as 30 percent of the dry weight of foliage, but a considerable part of the dissolved material is organic. As the growing season advances, inorganic solutes tend to accumulate in and on the leaves. This trend may be interrupted or reversed by rainfall that is sufficiently heavy or prolonged to wash away some of the accumu- lated material. The available data suggest that the ratio of sulfate to chloride in the leaves usually tends to in- crease during the growing season. Except where the ground water in the vicinity was very high in chloride, COMPOSITION OF SALINE RESIDUES ON LEAVES AND STEMS OF SALTCEDAR the solute washed from the leaves contained more sul- fate than chloride. Moisture which collected on saltcedar leaves at night along the Gila River near Glenbarr, Ariz., was strongly saline and contained mostly sodium and chloride. The cubic erystals, apparently sodium chloride, which occur on saltcedar leaves, are probably deposited when such saline solutions evaporate. The chloride ion content of the solutes present in salt- cedar leaves is probably affected more than the sulfate content by various processes that may remove salty fluids and dry salt crystals from the leaves. Samples were obtained from trees growing in widely separated areas, and the composition of ground water available to them had a wide range. The range of com- position of the solutes leached from the leaves, however, was narrower than the range of associated ground-water composition. One may thus conclude that selective processes within the plant strongly influence the com- position of solutes which can be leached from the leaves. Obviously the data presented here do no more than suggest some general tendencies in the movement of inorganic solutes from the ground water and soil mois- C9 ture through the saltcedar plant. More detailed studies in which the environmental variables are more closely observed and can be controlled are needed before more firm conclusions can be reached. LITERATURE CITED Anderson, M. S., Lakin, H. W., Beeson, K. C., Smith, F. F., and Thacker, Edward, 1961, Selenium in agriculture: U.S. Dept. of Agriculture, Agr. Research Service, Agriculture Handb. 200, 65 p. Campbell, C. J., and Strong, J. E., 1964, Salt gland anatomy in Tamariz pentandre (Tamaricaceae): The Southwestern Naturalist, v. 9, p. 232-2838. Decker, J. P., 1961, Salt secretion by Tamariz pentandra Pall. : Forest Sci., v. 7, p. 214-2117. Gatewood, J. S., Robinson, T. W., Colby, B. R., Hem, J. D., and Halpenny, L. C., 1950, Use of water by bottom-land vegeta- tion in lower Safford Valley, Ariz. : U.S. Geol. Survey Water- Supply Paper 1103, 210 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. Robinson, T. W., 1965, Introduction, spread, and areal extent of saltcedar (Tamarix) in the Western United States: U.S. Geol. Survey Prof. Paper 491-A, p. A1-A12. Sutcliffe, J. F., 1962, Mineral salts absorption in plants: London, Pergamon Press, 194 p. U.S. GOVERNMENT PRINTING OFFICE: 1967 £2 75 et 2 7 PAY . 491-D -Evapotranspiration by Woody Phreatophytes in The Humboldt River Valley Near Winnemucca, Nevada GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-D & ayn GA oct 14 1979 X); i/ Evapotranspiration by Woody Phreatophytes in The Humboldt River Valley Near Winnemucca, Nevada By T. W. ROBINSON W ith a section on SOIL-MOISTURE DETERMINATIONS By A. 0. WAANANEN STU DIES O F- -EV APO TR ANS PIK AT IO N GEOLOGICAL SURVEY PROFESSIONAL PAPER 1491-D Quantitative studies of water use by greasewood, rabbitbrush, willow, and wildrose UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON :- 1970 UNITED STATES DEPARTMENT OF THE INTERIOR WALTER J. HICKEL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 55 cents (paper cover) CONTENTS Page Page ADSUTACLE ELETE EL. nlc: oll rina nea ane tac, aon D1 | Problems-Continued frc cle 1 Salt content of the D17 Parpose and V.. 2 Boron toxicity... 17 L 2 Source and accumulation of boron in the top soil. 19 Climate}: "129 AE EH wanes oa ans ag a 2 Leaching as a corrective measure. _________________ 22 Species of woody phreatophytes_.________________. 3 | Factors influencing evapotranspiration rates-__________. 24 Winnemucca fest ls, 8 Expression of evapotranspiration losses and extrap- Weather statiOn-L L 2 020 cru 0 sel L Dua dra uwe aula hel 3 polation to growth areas.: 26 Evapotranspiration 8 Areal method.... s cell cks sis. 27 Definition of 3 Volume-of-foliage 27 Evapotranspiration 4 Results of evapotranspiration studies.____________. 27 Construction of evapotranspiration tanks... .._. 4 27 Water-supply system... _c 5 _ 29 Transplanting, survival, and growth of woody Willow.. ...s. rel loll l ul aul ad aoa da doan o 30 T 5.2 anus me ae blau Aisles a 30 Operation of the 8 Bare soil-: .-- esl. 31 Water _c. r 0 - Summarys. .e aa r bee tac ak ao a o 31 Sources of water for 9 | Soil-moisture determinations, by A. O. Waananen-____ _. 32 Rainfall 10 Soil-moisture observations at the Winnemucca test Water supplied to the tanks.._______________. 10 fie 32 Soil 0000000 nls 10 ar agar taf a anat pone ates Com utati n f ev. tr n i tion _______________ 11 EqUIpment and procedure _______________ 32 p ons of evapotranspira Plant growth and development.-__________________ 11 Water contents ob.ser.ved_._ ““““““““““ . FRolage }.. cls. t' 12 Water-content variations in 1963-__._________. 35 le _ lin ci tl. 16 Water-content changes in shallow flood-plain Damage by webbing 16 .L - -o eels nere enne n e Paik tie a s ales § ala ao 35 Damage by lege. 1? | Aeforences. . ... }. . 22 nil ely a s ave oad auc e e ae 41 ILLUSTRATIONS Page Fraur® 1. Map showing location of Humboldt River Research Project test site_ D2 2. Plan sketch of Humboldt River Research Project evapotranspiration test site 4 3. Photograph showing installation of a plastic membrane liner in a 30- by 30-foot evapotranspiration tank: 1X12-inch boards on edge outline the sides of the tank.... null lll lesen ai 5 4. Photograph of the bottom of the tank in figure 3 showing the water-distribution system in place...... 5 5. Plan and section of evapotranspiration tank showing water-distribution system_________________________ 6 6-11. Photographs showing: 6. A 30- by 30-foot tank about half completed, with the corners of the membrane fastened at the land surface. The photograph shows the method of 4 7. Installation of the water-distribution system in a 10- by 10-foot evapotranspiration tank.. 7T 8. The 5-year-old greasewood plants in tank 2 and the two 350-gallon elevated reservoirs that supply water to greasewood-tanks 1 and 2-12 22.20.00. osc l [nol nlc LLIN IESE eds T 0. Greasewood plants in tank 2 on September 30, 1960......_.... 8 10. Growth of plants in willow tanks on October 1, 1960, from cuttings planted on April 14, 1960... 8 11. Scientist measuring cover density and plant height on a transect across greasewood tank 2_______. 12 12. Graphs showing seasonal foliage volumes and operating water levels for the four species of phreatophytes grown in cvapotranspiration tanks.... 200-200. celled abee ende ceeded 14 13. Graph showing soluble boron content of the soil in greasewood tanks 1 and 2 before leaching and in tank 1 afterdeaghinig -t scc en cel eu e 20 o t os pu TET IC Joe. uR cn a aun idle ann eties ceara gees 19 14. Plan sketches showing location of points and lines used in sampling undisturbed soil, in relation to clumps of streasewooldl ... .= carens be ene ie uel, c uk aus beak bls aes ce Leas aanenetos tans end cae 20 III IV CONTENTS Figurss 15-19. Graphs showing: TABLE 20. 21. 22. 23. 24. 25. 26. co 10. 11. 15. Results of boron, sodium, and electrical-conductance determinations of soil samples collected from sampling points shown in figure tace oes sou a alk mas 16. Decrease in boron and specific conductance of the effluent from greasewood tanks 1 and 2 during leaching.... . . .. 22 lll cl oe Eu PAL w WD wen nt e wr ve e ne m Hy rs n he hme ne fue Te e a n an ie e on mean sna Salee nt mnt Snee e had H a we me me e on 17. Comparison of specific conductance and boron of the effluent from rabbitbrush tanks 2 and 3 for the leachings of 1063 and 1966: .L ie uns ee ses us mak 18. Relation of monthly draft on ground water to temperature April through October 1966 for three species of te n ae onl ule me ana bn he to i in ha on i ame ho 19. Evapotranspiration by four species of phreatophytes during the growing seasons 1963 through 1967 for indicated depth of water d ensbe tire ae aise o aly a an u tak Photograph of neutron-meter scaler, probe, and shield, with soil augers and typical access tubes-_______. Generalized section of evapotranspiration tank showing neutron-meter depth probe in access tube for soil- moisture observations, and typical soil-moisture profile _ LLL Photograph of neutron meter set up for use in gréeasowood tank 1.2 : .o oul bul aL ILL t ean ae aio ea ae a Photograph showing rabbitbrush tanks and access tubes installed for soil-moisture observations-___.__.. Soil-moisture profiles at the beginning and end of the growing seasons 1962-67 in evapotranspiration tanks and the flood plain at the Winnemucen test site.. /l. abs agus e Graphs showing variations in water content of soils in evapotranspiration tanks at the Winnemucca test site during the 1963 Graph showing water content at end of each season and maximum content observed in June 1962, in inches, as observed at three sites in Humboldt River flood plain near Winnemucca, Nev----__________________ TABLES . Growth and development of woody phreatophytes grown in evapotranspiration tanks at the Winnemucca est aThe . 22 22 sel ae eee e e a ue a i o a o lee a ed he m e ie ine tn ra an n wt t o In oe Neca T h nd t t e hn i an a n ented t an at e be in it an oe in hea n he on n an eve ae ar ie . Comparative data on greasewood growth in the field under different growth conditions and the average in the two evapotranspiration . Chemical analyses of water from the saturated soil in the evapotranspiration tanks of the Winnemucca test Site EXL g ITL a a a ee e 2 Els a ae at h wade e he ie cal ed aa a ae te Hale ta ae od i e ali at i ie he a elan a tole n an in to he i n e ol s e as t mre at wea ie Results of analysis of samples of undisturbed soil profile. . Results of analysis of surface soil sampled to 3-inch depth at 1-foot intervals from central stem of a grease- wood plant...... L Dak oie aoe ha a in beg a oa a mie al m ale an e ane in be as in in eline he e calle we ul al Len tn he e oa e ale . Quantity, rate of inflow to tank, and chemical analyses of samples of the effluent from greasewood tank 1 and rabbitbrush tank 1, October cenas ane nen . Comparative warmth for the period, April 1 to October 31, 1962-67, by months, at the Winnemucca test site, in degree days above a base of 32° . Climatological data for the April 1-October 31 period, 1962-67, at the Winnemucca test site___________-_.- . Evapotranspiration by four species of phreatophytes grown in tanks at the Winnemucca test site and evaporation from bare soil during the growing seasons Soil-moisture changes in evapotranspiration tanks at the Winnemucca test site during summer and winter periods 1902-07. .... . 20.0.2 el a tink co mick a ane bell a aimee alm mio ue Sie man h a ak ale sieve a a aan of a ae in moons Water content of soils in evapotranspiration tanks at the Winnemucca test site, as observed during the 1963 SCAROMELL CLZ L Lene ede ee Lieu ire 2a nak ank aas o e a aik ale a ace ole aun an a i ain a Amare a me hae io a ules a toe Page D21 28 28 26 38 40 41 Page D13 15 18 20 21 24 25 26 28 36 40 STUDIES OF EVAPOTRANSPIRATION EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY, NEAR WINNEMUCCA, NEVADA By T. W. Rosinsor ABSTRACT This report presents the results of cooperative studies of evapotranspiration by phreatophytes in the Winnemucca reach of the Humboldt River valley. Water that is wasted by evapo- transpiration from areas of low beneficial phreatophytes is one of the largest unknowns in the water budget of the reach. In order to obtain information with which to evaluate the consumptive waste, studies of the water use of four woody phreatophytes-greasewood, rabbitbrush, willow, and wild- rose-were undertaken in evapotranspiration tanks at the Win- nemucea test site. Twelve tanks ranging in size from 30 feet square and 10.5 feet deep to 10 feet square and 7 feet deep were constructed. Seed- lings of greasewood were planted in 2 tanks and rabbitbrush in 3 tanks; cuttings of willow were planted in 3 tanks and wild- rose in 3 tanks. The twelfth was left bare. The tanks were constructed in place by lining excavated pits with watertight plastic membranes, providing a water-distribution system on the bottom and backfilling with the excavated material. The tanks were operated during the growing season April 1 to October 20 from 1961-67, inclusive. Water metered into elevated reservoirs was supplied by gravity to the evapotranspiration tanks. Evapotranspiration by the plants was computed as the sum of rainfall, soil-moisture depletion and water supplied to the tanks during the growing season. Plant growth and development were recorded by photographs and transects. Foliage volumes were computed from the transect data as the product of the average height, the cover density of the plants, and the area of the tank. Foliage volumes and water use were affected by difference in depths to the water level, damage to the plants by rabbits and by insects, and, for the greasewood and rabbitbrush, by accumula- tion of boron in toxic concentrations in the root zone. The dam- ages were alleviated by removing rabbits from the test site, by spraying the insect-infested plants, and by reducing the boron content in the root zone by backwash leaching. Boron content of undisturbed soil, adjacent to the grease- wood tanks, ranged from 13 to 32 milligrams per kilogram in the top 1.5 feet, and from 5 to 0.9 milligrams per kilogram between 5 and 9 feet. Green greasewood leaves had a boron content that ranged from 196 to 233 milligrams per kilogram. Evapotranspiration by a given species of phreatophytes is affected by climatic conditions, of which temperature is the most important. Water use by greasewood and rabbitbrush in April was only about 2 percent of the yearly total, whereas dur- ing the months of peak use it was 28 percent. More than two- thirds of the annual use occurred during June, July, and August. Evapotranspiration, expressed on an areal basis as depth over a unit area, gives no indication of growth conditions for which the information was obtained and may result in serious error when transposed to areas of dissimilar growth conditions. Some of the difficulties and uncertainties of the areal method may be avoided by expressing evapotranspiration on a volume-of-foliage basis, as a quantity of water per unit of foliage volume. The method presumes that transpiration by a species is proportional to the total transpiring leaf area, and so proportional to the foliage volume. In the results of the studies, evapotranspiration is expressed in both quantities. The annual use of water ranged rather widely over the study period, as the plants responded to the effect of plant damage, boron toxicity, depth to the water level, and warmth and length of the growing seasons. Draft from the water table, equivalent to the water supplied to the tanks, varied with the rainfall. It was greatest when rainfall was scant, and least when the rains were copious. INTRODUCTION The Humboldt River Research Project is a Federal- State cooperative project concerned with developing data and techniques by which to evaluate the water resources of the Winnemucca reach of the Humboldt River. The agencies cooperating in the study were the U.S. Geological Survey, the U.S. Bureau of Reclama- tion, and the Department of Conservation and Natural Resources of the State of Nevada. The Winnemucca reach extends from the Comus gag- ing station downstream to the Rose Creek gaging sta- tion. The Comus gage is about 23 miles east, and the Rose Creek gage is about 15 miles southwest of the city of Winnemucca. The distance between the stations along the meanders of the river in the flood plain is about 92 miles, whereas the distance along the meandering flood plain is about 45 miles. The flood plain ranges in width from 0.2 to 5 miles; the average width is 0.8 mile. The altitude at the Comus gage is slightly less than 4,400 feet above mean sea level, and that at the Rose Creek gage is about 4,200 feet (fig. 1). D1 D2 STUDIES 120° 118° 116° 114° 42° j j 42° Winnemucca ~ Zs Rose Creek Ne & Comus 5 Elko [ | | (T y Ely'! 40° 40° 38° 1 18} \ EXPLANATION \\ X \ Test site \\ | A Las Vegas Gaging station 36°- Ro ° 36° 116° * 114° XK ~ % 150 MILES 0) ard WWW] Figur 1.-Location of Humboldt River Research Project test site. One of the largest unknowns in the water budget for this reach of the Humboldt River is the evapotranspira- tion loss. This loss includes the losses from water and bare-soil surfaces by evaporation and losses from soil moisture and ground water through transpiration by phreatophytes. Phreatophytes are plants that habitual- ly send their roots to the water table, and obtain their water supply primarily from ground water. Several species of woody phreatophytes thrive along the streambanks, in the flood plain, and on the lower parts of the adjacent alluvial fans. Water use by such plants is by far the largest of the several evapotranspiration losses. The purpose of this study, one of several in the inter- agency Humboldt River Research Project, was to de- termine the water use by woody phreatophytes, notably several species of low beneficial use. The water used by these plants constitutes a consumptive waste of water, for it is discharged into the atmosphere with but little benefit to man. This is one of nature's preemptive taxes, and results in depletion of the water resources of a region and reduction in the quantity of water available to man. With proper management, man can reduce this draft and benefit from the reduced draft to the extent of the salvageable portion. Possible modifications might OF EVAPOTRANSPIRATION be to replace the phreatophytes of low economic value with plants of higher economic value or to salvage the water otherwise -consumptively wasted and use it bene- ficially. To assess the economic feasibility of such opera- tions, quantitative evaluations of the consumptive waste are needed. PURPOSE AND SCOPE The purpose of the evapotranspiration studies was to obtain data for evaluation of the consumptive use of water by four woody phreatophytes widespread in the Winnemucca reach. These shrubs occur generally throughout the Humboldt River basin and in other areas in Nevada. The studies involved growing the plants in large evapotranspiration tanks and determin- ing their seasonal water use for different depths to the water level and for different cover densities. The first tanks were constructed at the Winnemucca test site (see page D4) in the fall of 1959 and were planted in the spring of 1960. Later plantings were made in 1961 and 1962 as additional tanks were constructed . ACKNOWLEDGMENTS Funds for construction of the tanks and for part of the operation and maintenance costs were provided by the U.S. Bureau of Reclamation. The evapotranspira- tion studies were under the technical direction of the Water Resources Division, U.S. Geological Survey. The cooperation and assistance provided by personnel of the Nevada Department of Conservation and Natural Resources are acknowledged. The author is especially grateful to George Hardman, Assistant Director, for his unflagging interest, his counsel and helpful sugges- tions. Vernon Laca, assisted at times by James Starr, provided 7-day-a-week operation of the tanks during the growing season. The wholehearted cooperation of Mr. and Mrs. H. T. Harrer, on whose property the Win- nemucca test site and access road are located, is appreciated and acknowledged. CLIMATE The Winnemucca reach of the Humboldt River lies largely in the 5- to 8-inch rainfall zone. The climate is characterized by few cloudy days and moderate wind movement and may be classified as arid to semiarid. About two thirds of the annual precipitation falls as rain or snow during the winter period, December to May. During the growing season, April to October, the precipitation falls largely as scattered summer showers that occasionally exceed one-half inch. On rare occasions as much as 1 inch of rain has been recorded during a single storm. The prevailing winds are wester- lies, and these lose much of their moisture as they pass over the Sierra Nevada about 150 miles to the west. EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY The summers are marked by warm days and cool nights. Temperatures tend to rise sharply with the sunrise and remain comparatively high during the day- light hours, then drop rapidly about sundown. A daily temperature variation of 50°F is not uncommon. According to the U.S. Weather Bureau, the average monthly temperature during the growing season for 92 years of record at and near Winnemucca has ranged from 47.1°F in April to T1.T°F in July. Extremes of temperature have ranged from 36°F below zero in Janu- ary to a high of 108°F in July. Temperatures rise to 100°F or more on the average of 2 days a year. The growing season during the 92-year period of record has varied considerably, the shortest on record being only 63 days and the longest 184 days; the average growing season is about 140 days. Measurements of pan evaporation for the months of April through October were made at the Winnemucca test site from 1962 through 1967. Evaporation during this 6-year period ranged from 52.7 inches in 1965 to 66.8 inches in 1966 and averaged 58.0 inches. SPECIES OF WOODY PHREATOPHYTES The common woody phreatophytes that are native to the Winnemucca reach of the Humboldt River include greasewood (Sarcobatus vermiculatus), rabbit- brush (COArysothammus), and willow (Saliz). Wildrose (Rosa) and buffaloberry (Shepherdia) occupy less ex- tensive areas. Saltcedar (Z'amariz), an exotic plant from the Mediterranean area introduced into this country about the turn of the century, is invading the lower part of the Winnemucca reach. Saltcedar is also grown as an ornamental shrub in several places in the city of Winnemucca. In addition, hydrophytes, such as cattails and bullrushes, grow in several small areas adjacent to the river. Greasewood is the most extensive phreatophyte in the area, and willow and rabbitbrush are the next most common species. The open spaces on the bottom lands of the flood plain are covered with a variety of bene- ficial phreatophytic grasses. The two common species are bluejoint or creeping wildrye (W/ymus triticoides) and saltgrass (Distichlis stricta). Great Basin wildrye (Elymus cinereous) was once common in the flood plain, but is now found only in areas protected from heavy livestock grazing. WINNEMUCCA TEST SITE The Winnemucca test site is a parcel of land 300 by 600 feet on the Harrer ranch, in the NEY sec. 2, T.35, R.37 E., Mount Diablo base line and meridian, on the south side of the Humboldt River. The site is 344 miles southwest of Winnemucca, and three-fourths mile D3 west of U.S. Highway 40. About half the site lies on the present flood plain of the Humboldt River, and the re- mainder on a terrace about 4 feet higher. Both parts are nearly flat, with some slope northward toward the river. The terrace is at an altitude of about 4,260 feet. To prevent inundation of the part of the test site on the flood plain during periods of high water in the Humboldt River or when the adjacent flood plain is ponded for flood irrigation, an earthen dike, approxi- mately 214-3 feet high, was constructed. The site also was fenced with a combination of barbed wire and wire mesh, which was ostensibly rabbit tight and stock proof. The layout of the test site, showing the evapotrans- piration tanks, water-supply reservoirs, pipelines, sup- ply well, and weather station, is shown in figure 2. WEATHER STATION A class "A" weather station, installed at the test site in March 1962, was operated each season during the pe- riod April 1 to November 1. Instrumentation consisted of a 4-foot evaporation pan, a totalizing anemometer, maximum and minimum thermometers, hygrothermo- graph, and an 8-inch rain gage. The temperature and hu- midity instruments were housed in a cotton-region type shelter. A pyrheliograph for measuring incoming radia- tion was installed in the low-lying meadow a short dis- tance from the weather station. In addition, a thermograph was installed at grease- wood tank 2 to measure the surface temperature of the soil in the tank. The sensing element was covered with about one-fourth inch of soil, enough for shielding from the direct rays of the sun. All the nonrecording instru- ments were read daily, and the charts on the recording instruments were changed weekly. The station is about 30 feet lower in altitude than the Weather Bureau station at the Winnemucca Airport, 3 miles to the south. EVAPOTRANSPIRATION STUDIES DEFINITION OF EVAPOTRANSPIRATION The first use of the term "evapo-transpiration" was by Sondregger (1929) in May 1929; he used it as a synonym of evaporation-transpiration losses, a term coined by Lee (in Lee and others, 1926) to describe water lost to the atmosphere. Later in the 1930's as the term came into common usage the hyphen was omitted. The term is generally considered to be synonymous with the term "consumptive use" and is defined in Manual No. 43, "Nomenclature for Hydraulics," of the Ameri- can Society of Civil Engineers (1962, p. 156) as "water withdrawn from soil by evaporation and plant transpiration." D4: STUDIES OF EVAPOTRANSPIRATION 5 600 FEET & err rr c __ 0, | TTT_rTT—IRWater-supply barrels +_\ K p- op Meadow Graveled drive dike x 2¢---q1 -I- j | \ | G A\ et" \p Al II e ® \ 2 ¢ m_ (1 g Weather \ I ® [232mm i Rabbitbrush tanks Tool houses 6-inch well Elevated water reservoirs (¥. | | I| #2? Elevated water reservoirs Pressure tank sae -f | j | x FIGURE 2.-Humboldt River Research Project evapotranspiration test site. EVAPOTRANSPIRATION TANKS Twelve evapotranspiration tanks were installed at the Winnemucca test site. Eleven tanks were used to meas- ure the evapotranspiration from four species of woody phreatophytes-greasewood, rabbitbrush, willow, and wildrose-and the twelfth was used to measure the evaporation from bare soil. The plant species, tank sizes, and planting dates are given in the following tabulation : Evapotransptiration tanks Tanks Plant species Date planted Number Size (ft) Greasewood.__.... 2 30X30X10. 5 Apr. 183, 1960. Wwillow..."....... 3 10X10X 7.5 Apr. 14, 1960. Wildrose:.__.____ 3 10X10%X 7.0 May 20, 1961. Rabbitbrush. ._.. 3 20X20X10.0 Apr. 10 and 11, 1962. bare €10X10y% 7.0 The greasewood and rabbitbrush tanks are on the terrace at the test site, and the willow, wildrose, and bare-soil tanks are on the low-lying flood plain. The terrace, according to Cohen (1965, p. 28, pl. 1), con- sists of deposits of Quaternary age, whereas the flood plain consists of deposits of Holocene (Recent) age. The soils and vegetation in the two parts are quite dif- ferent. The terrace deposits consist of poorly sorted gravel, sand, silt, and clay, whereas the flood-plain de- posits consist of clay loam with considerable organic matter and occasional balls of volcanic ash. The terrace deposits support a vigorous and well-established growth of greasewood that averages 2.5 feet in height; individual plants are as much as 5 feet high. Saltgrass grows in open spaces between the clumps of grease- wood. About three fourths of the flood-plain part of the test-site meadowland is covered with several kinds of grasses, and about one fourth, in the northwest cor- ner of the site, has a dense growth of willows and wild- rose. The willow and wildrose tanks are in this north- west corner. The bare-soil tank is in the meadow-grass area, at the edge of the willow growth. Saltgrass occurs among the willows and also in the meadow. The pre- dominant meadow grass, however, is bluejoint or creep- ing wildrye. As the result of the fencing of the site for protection from grazing, a few clumps of Great Basin wildrye have made a healthy growth in the meadow. CONSTRUCTION OF EVAPOTRANSPIRATION TANKS The tanks were constructed in place by installing watertight plastic-membrane linings in pits excavated into the terrace and flood-plain deposits. The membranes for the 30- by 30-foot tanks and the 10- by 10-foot tanks were fabricated in sheets 57 and 30 feet square, respec- tively, from black polyvinyl plastic sheets 20 mils (0.02 inch) thick. The plastic sheets for the 20- by 20-foot tanks were 45 feet square and were fabricated from aqua polyvinyl material 22 mils thick. The pits for the tanks were excavated to the water table and were 2-3 feet larger than the finished size. Excavation was done by a dragline equipped with a 1- yard bucket. When the excavation was completed, the bottom of the pit was leveled manually, raked clean of EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY stones and sharp objects, and shaped to size. Next, the dimensions of the finished tank were outlined on the bottom of the pit by using 1 X 12-inch boards placed on edge, and the plastic membrane was positioned in the bottom of the pit as shown in figure 3. Marks on the center of the membrane served as guides to fix the liner position in the pit. After the membrane was positioned, the part forming the bottom of the tank as outlined by the 1 X 12-inch boards was covered with a 4-5-inch layer of pit-run medium to coarse sand. A water-distribution system was then placed on the sand layer as shown in figure 4. The system consisted of a horizontal E-shaped component made of 2-inch rigid plastic pipe perforated with holes on 14-inch centers on the underside, and a 4-inch riser pipe for adding or withdrawing water from the system. The riser pipe extended 0.5 foot above the ground surface and was also used to monitor the water level in the tank. A plan and section view of the tank, including the water-distribution system, is shown in figure 5. The 2-inch pipes were covered with sand throughout their perforated sections in order to facili- tate movement of water into the sand layer which has been described. With the water-distribution system in place, the pit was backfilled with the material previously excavated, a clamshell bucket being substituted in place of the drag- line bucket for this purpose. The excavated material was replaced only approximately in the reverse of the order in which it had been removed. The positioning of the membrane for the sides of the tank was accomplished by alternately draping the membrane on the outside of the 1 X 12-inch boards and placing fill on the inside, and then draping on the inside and placing fill on the out- side. As the level of the fill on both sides became even with the top of the boards, the boards were raised 8-10 inches and the backfilling process repeated. Thus, as the backfilling progressed, the membrane that was origi- FIGURE 3.-Installation of a plastic membrane liner in a 30- by 30-foot evapotranspiration tank: 1 X 12-inch boards on edge outline the sides of the tank. 312-487 O-10--2 D5 nally draped over the 1 X 12-inch boards in the bottom of the pit became the sides of the tank. Experience demon- strated the advantages of keeping the backfill higher along the sides than in the center of the tank, and the desirability of fastening the membrane corners on the land surface at the corners of the pit as shown in figure 6. The 30- by 30-foot and the 20- by 20-foot tanks were constructed in the fall of the year. Upon completion, the tanks were tested for watertightness by being filled with water to within about 4 feet of the surface and allowed to stand in this condition during the winter months. No leaks developed. A 30- by 30-foot tank was constructed in about 4 days; the excavation required about 114 days; the installation of the membrane, water-distribution system, and the backfilling about 214 days. The smaller 20- by 20-foot tanks were constructed in 214 days each, with the excavation usually completed in about 7 hours; the 10- by 10-foot tanks required 1-114 days each. The method of construction and the design of the water-distribution system in the shallower 10- by 10- foot tanks was different from that for the larger tanks. For the smaller tanks, the pit was excavated to the finished size, and the membrane suspended loosely from the top edges of the pit. During backfilling, care was taken to keep the membrane on the sides of the pit loose and free from tension. The distribution system resting on a 4- to 5-inch layer of sand, as shown in figure 7, was an 8-foot circle of 2-inch flexible plastic pipe that was cross connected on one diameter, perforated on the bot- tom, and connected to a 6-inch riser pipe. WATER-SUPPLY SYSTEM Water for use in the test site was obtained from a drilled well, 25 feet deep, located inside and near the entrance to the site (fig. 2). The well was equipped with an electrically operated jet pump that delivered water under a working pressure of 30 pounds per square inch FIGURE 4.-The bottom of the tank in figure 3 showing the water- distribution system in place. D6 STUDIES OF EVAPOTRANSPIRATION o | 30'0 | 5-- B-» 10' & 6 o 2 g £ 0 o 3 £ $% gE I € w 2 Ee O 2 bo 5m C ~ ~. E7 2+ C wat C E46 _E a 3 0 a € £E% 30 9 o 4 50 £28 o a a. 30'0" 4-inch water-supply and dewatering pipe Backfilled with excavated Land surface material PLAN r4-inch water-supply and dewatering pipe Tank surface Fine sandy silt Sand and small gravel 4-5-inch layer of pit-run sand CG Excavation line goeth P fe oo or aa frites yo's: SECTION FiGurE 5.-Plan and section of evapotranspiration tank showing water-distribution system. to a 500-gallon buried pressure tank. Water from the pressure tank was delivered through a pipeline to ele- vated metal reservoirs adjacent to the evapotranspi- ration tanks (fig. 8). The pipeline consisted of about 500 feet of %-inch iron pipe and 400 feet of 1-inch plastic tubing. The volume of water delivered to the reservoir was measured by industrial cold-water meters that have horizontal nonsetback-reading registers and a 10-gallon dial circle calibrated to one-fourth gallon. Each meter was tested for accuracy before installation; for flow rates greater than 3 gallons a minute, the measurement error was found to be less than 1 percent. Water from the reservoir was delivered by gravity through a %-inch pipeline to a float-controlled valve at the 4-inch supply pipe (fig. 5) of the evapotranspira- tion tank. Water levels in the evapotranspiration tanks were controlled by means of a float in the EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY FIGURE 6.-A 30- by 30-foot tank about half completed, with the corners of the membrane fastened at the land surface. The photograph shows the method of backfilling. FicurE 7.-Installation of the water-distribution system in a 10- by 10-foot evapotranspiration tank. FiGurE 8.-The 5-year-old greasewood plants in tank 2 and the two 350-gallon elevated reservoirs that supply water to greasewood tanks 1 and 2. D7 supply pipe that was directly connected by a %%-inch rod to the float-valve mechanism. The float valve was activated by the float; as the water level in the evapo- transpiration tank fell, the valve opened to admit water ; when the water level rose, the valve closed. The inter- val between the high and low position of the water levels was less than 0.1 foot; so a nearly constant water level was maintained in the evapotranspiration tank. TRANSPLANTING, SURVIVAL, AND GROWTH OF WOODY PHREATOPHYTES Transplanting of greasewood and rabbitbrush to the tanks was undertaken with some trepidation, for it was the first time transplanting on a large scale for experi- mental purpose had been attempted. Seedling plants were used for both species. Some consideration was given to transplanting mature plants, but it was feared that damage to their deep-root systems would be so great that many of the plants would not survive. During excavation for the tanks, greasewood roots were ob- served and photographed at depths of 7 and 8 feet; judging from the root size at these depths, the roots probably extended to a depth of at least 10 feet. The seedling greasewood plants were obtained in the general vicinity of the tanks. In the transplanting pro- cedure, the seedlings were dug from the ground with care taken to keep as much soil as possible around the roots, placed in a wheelbarrow, covered with wet bur- lap, and transported to the tank for planting. Each seedling was examined carefully before planting, and seedlings with obvious root damage were discarded. Prior to planting, the average spacing of the plants in the area was ascertained by random transects. Because some loss of the transplanted shrubs was anticipated, the seedling spacing was closer in the tanks than in the sur- rounding growth, for which the average spacing was 3- 31/, feet. Eighty-five plants were planted in greasewood tank 1 and 105 in tank 2 in mid-April 1960. At the end of September, 71 plants were growing in tank 1 and 89 in tank 2, a survival rate of 84 and 85 percent, respectively, for the two tanks. Figure 9 shows the growth in grease- wood tank 2, 5% months after planting. Some addi- tional transplanting was done in the spring of 1961 to fill the spaces where two or more adjacent plants had died. Rabbitbrush plants were not available in the im- mediate vicinity of the test site; however, an exten- sive growth about one-half mile distant provided an ample supply of seedlings. The young plants were D8 FIGURE 9.-Greasewood plants in tank 2 on September 30, 1960. Of 105 seedlings planted on April 13, 1960, 89 rooted and thrived. easily dug from sandy loam with a minimum of dam- age to the roots. The seedlings were planted in irregular rows, and at a closer spacing than for the greasewood, to simulate observed field conditions. The plantings on April 10 and 11, 1962, were as follows: 75 plants in tank 1, 86 in tank 2, and 90 in tank 3. A count of the living plants on July 9, 1962, showed 63 surviving in tank 1, 75 in tank 2, and 86 in tank 3. An item of interest was the location of the dead plants. Plant mortality in each of the three tanks was greatest in the center of the tank and least along the edges. Replanting in the spring of 1963 did little to alter this pattern. The willow tanks were planted on April 14, 1960, using cuttings from the thicket in which the willow tanks were located. The cuttings, about 10 inches long, were stuck into the wet soil surface of the tanks to a depth of about 6 inches at approximately 1-foot inter- vals. An inspection on September 30 showed that about 99 percent of the cuttings had taken root and were growing. The growth by October 1, as shown in figure 10, ranged in height from 11/4 to 314 feet and averaged about 244 feet. The wildrose tanks were also planted with cuttings from the surrounding area. The plantings on May 20 took root, thrived, and were well established by the end of the 1960 growing season. Survival rate was high, averaging about 95 percent. OPERATION OF THE TANKS With the exception of the year of planting, studies of evapotranspiration in the tanks were started each year on April 1 and ended on October 20. The emphasis following planting was on the establishment of a healthy and vigorous growth. The period April 1 to October 20 was selected because it spanned the dates of earliest and latest beginning and end of the growing season. Generally water could not be supplied to the tanks for STUDIES OF EVAPOTRANSPIRATION the entire period each season. Water systems were dis- rupted in some years near the beginning of the season and in other years near the end of the season when below freezing temperatures at night caused water lines and water meters to freeze and burst. Samples of the soil and of the water in the tanks were collected once or twice a year, following the first full growing season, for chemical analysis. The analyses provided data on the concentration of salts in the soil and water, and warned of any salt buildup that could be detrimental to the plants. Once the operating water level for the growing season was established and the float mechanism for controlling the water level adjusted, the day to day operation was fairly simple. At the start of the season, the reservoirs were filled to their operating capacity, indicated by a reference mark on a manometer tube attached to the side. The reservoirs supplying the grease- wood and rabbitbrush tanks had a capacity of 350 gal- lons, and those for the willow, wildrose, and bare soil tanks 100 gallons. With the outlet valve of the reservoirs open, water delivered to the evapotranspiration tanks was controlled at the tanks by the float-valve mecha- nisms. The quantity of water supplied to the evapo- transpiration tank in any period replaced water used by the plants. It is a measure of the draft from the ground-water reservoir. The quantities of water used were determined by refilling the supply tanks to the reference marks at the end of each period with water measured through the water meter. It was not feasible to meter water directly into the evapotranspiration tanks because of the low demand-rate of flow. The rate varied from a trickle to about 1 gallon a minute during the course of a day. These rates were too low for proper actuation and opera- tion of the meter. The interval between refills varied, depending on the FicURE 10.-Growth of plants in willow tanks on October 1, 1960, from cuttings planted on April 14, 1960. Average height, 2% feet, density comparable to that in surrounding thicket. EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY date within the growing season. At the beginning and end of the season, when water use by plants was low; the reservoirs were refilled at weekly or 10-day inter- vals. During the height of the season, in July and August, it was not uncommon to fill them every day or every second day. All reservoirs, regardless of the need, were refilled on the first day of each month. In this way, the water use by months was obtained. Re- filling was usually done in the morning, generally between 9 and 10 o'clock. At each filling, the operator compared the water use of the current period with that of the previous period, and thus was alert to any unusual condition. Some difficulty was experienced with sediment clogging the float-valve outlets, reducing and even stopping delivery of water to the evapotranspira- tion tanks. On a few occasions, the submergence of a float that developed a leak kept the float valve open and thereby allowed excess water to enter the tank. On other occasions, the horizontal-distribution system in the bottom of the greasewood tanks became partly clogged and restricted inflow. Unusual conditions such as these were quickly noted and easily corrected. WATER LEVELS With the exception of the interval following plant- ing, while the plants were getting established, the water levels in the evapotranspiration tanks were controlled at predetermined depths below the surface of the tank. During and following planting of the seedlings or cut- tings, the depth to water was maintained from 1 to 2 feet below the surface so that the soil at the root level might be kept moist. As the plants became established and the roots developed, the water level was allowed to decline slowly, until the approximate operating level was reached. All the evapotranspiration tanks were operated for the first 3 of 4 years with a water level 5 feet below the surface of the tank. When adequate data for that depth had been obtained, the water level was raised or lowered, and water-use data was obtained for different depths to water. The operating water levels for the different spe- cies and the different tanks, for the period of study, are given in the following tabulation : DQ Operating water levels for the evapotranspiration tanks Depth to Growing season. Tank wate; §evel (ft Remarks Greasewood LEAL e eaves ad Pre aas there e 1,2 fig Aug. 1 to Oct. 21. 6.0 7.5 6.0 Apr. 1 to Aug. 4. Membrane accidentally perforated on Aug. 4. 7.6 6.2 7.8 July 13 to Oct. 21. 5.0 5.0 g; Tank 3 discontinued. 5.8 4.2 5.4 4.1 100e 64-65 .l. .is. ce 1,23 5.0 966 ts reel CIOL 12,3 5.3 f cpr 12% 6.2 Tank 3 discontinued. t> & > on gs on on mH to 01000 P5 pbo e oweone sOURCES OF WATER FOR EVAPOTRANSPIRATION Evapotranspiration rates, as determined from the tank studies, are for the period of the growing season, April 1 to October 20. The data are indicative but may not be applicable to the full year, as precipitation was not measured during the nongrowing season. In other similar tank studies in the southwestern United States where the winters are mild, evapotranspiration rates are determined for the full year. D10 The water that made up the evapotranspiration dis- charge was supplied from three sources, namely, (1) rainfall, (2) water supplied to the tanks, and (3) soil moisture. RAINFALL As noted earlier, rainfall was measured by a standard 8-inch rain gage at the weather station. The catch of the rain gage was considered the water supplied by rain. During the period of study, rainfall ranged widely not only in amount for the growing season but also in the intensity and frequency of individual storm periods. Rainfall for storm periods ranged from a low of 0.01 inch in July and September 1966 to a high of 1.96 inches during a 4-day period in June 1964. During July 1962 and 1963 and October 1964 and 1966, there was no meas- urable rainfall. Rainfall, in inches, from April 1 to October 20, is given in the following tabulation : Rainfall, in inches, from April 1962 to October 1967 Year Apr. - May June July Aug. Sept. Oct. Period 0.23 0.35 0.25 0.00 0.13 0.05 - 0.35 1.36 1.20 2.07 . 00 .25 .25 . 49 6. 51 . 87 97. 2.10 .35 . 08 46 00 4. 83 . 81 T7 1.06 . 34 42 09 14 3. 63 38 33 48 . O1 .02 42 00 1. 64 98 21 | 14. 51 . 36 19 22 05 3. 52 WATER SUPPLIED TO THE TANKS The water supplied to the evapotranspiration tanks was the largest of the three sources of water for evapo- transpiration. This water represented the draft on the ground water by the plants. Draft on the ground water will vary from season to season depending on the rain- fall, the climatic conditions, and the portion of the winter rainfall available to the plants. The quantities, expressed as acre-feet per acre, supplied to each tank during the growing seasons from 1962 thru 1967 are given in the following tabulation : Water, in acre-feet per acre, supplied to the evapotranspiration tanks during the growing seasons from 1962 to 1967 Species Tank 1962 1963 1964 1965 1966 1967 Greasewood................ 1 0.90 _ 0.56 0.42 0.83 _ 0.56 2 88 .51 () . 98 . 62 Rabbitbrush............... ELECT 1.39 - 1.10 '68 <_ 1.18 - 1.99 2 1.42 . 98 . 53 . 85 72 3 - 1.76 .- 1.20 T5. £16 91 Willowe c 2. ulin lil. 1 2.89 1.76 148 230 1.55 2 297 180 213 3.19 18 3 3. 82 RST L(t) 0 Leta a an aan Wildrose}. 1. 04 . 86 . 83 1. 23 1.08 Pe 69 ~80 : 1.26% "A046 Divs eca se 58 MTO CNM Lu tite i eels B are Soll. .cc db Don eden ison at's "47 30 . 63 23 18 () ! Membrane perforated August 1965. 2 Discontinued in 1965. * Unmeasured water entered tank 1967. NotE.-1 foot of depth of water over the tanks is equivalent to: 6,733 gallons for greasewood 1; 6.695 gallons for greasewood 2; 2,992 gallons for rabbitbrush 1, 2, and 3; 748 gallons for willow 1, 2, and 3, wildrose 1, 2, and 3, and bare soil. STUDIES OF EVAPOTRANSPIRATION SOIL MOISTURE The plants in the evapotranspiration tanks obtained part of their seasonal water supply from soil moisture in the unsaturated zone above the water table. The differences in the water content of the soils in the tanks at the beginning and end of the growing season pro- vided a measure of the quantity of water obtained by the plants from this source. Determinations of the volumes of water in the soils in the evapotranspiration tanks and of seasonal varia- tions in the water content were started in 1962. Neutron- meter observations of soil moisture were made in access tubes installed in each of the tanks at the beginning, middle, and end of the growing season except in 1962 and 1963 when additional observations were made. The tubes were of sufficient length to permit sampling to nearly the full depth of the tanks, and the bottoms were sealed in order to permit observations at depths below the water levels in the tanks. The quantities of water, in acre-feet per acre, ob- tained by the plants from soil moisture during the growing seasons 1962 through 1967, and the observed changes in water content of the soil in the bare-soil tank are shown in the following tabulation : These data represent net changes in water content for the full pro- files sampled, although the principal depletions of soil moisture occurred in the unsaturated zone. Soil-moisture depletion, in acre-feet per acre, in evapotranspiration tanks at the Winnemucca test site during the growing season [Parentheses indicate géin in soil moisture, generally as result of adjustment of water level in tank after start of seasonal operation] Species Tank - 1962 1963 1964 1965 1966 1967 1:>:0.97, 0.46 - 0.200 0.42) 24 0. 43 2 . 38 . 58 M89 . 25 . 64 1 . 09 . 22 .19 . 21 . 65 . 40 2 . 09 . 28 . 09 14 .28 .37 3 .10 .23 12 . 09 . 36 . 52 Willow. cores nue erie ns ons ay 1 . 34 27 .10 . 29 . 28 . 25 2 . 51 . 28 . 23 15 . 22 . 24 3 . 38 . 25 (BO .. Wildrose. . 2.52 22sec yal 1 (. 01) 14 .10 .10 .19 .21 2 . 03 21 14 18 . 24 .16 3 15 . 27 (aU ves ogee racer ey Bate lures. ave cns, 14 .03 _ (.03) (.01) .04 08 The water provided to the plants from soil moisturt: each season constituted a significant part of the seasona ' supply for each species. This source supplied 27 percent of the total use in the greasewood tanks during the study period, 17 percent in the rabbitbrush, 10 percent in the wildrose, and 8 percent in the willow tanks. The contri- butions varied widely from year to year owing to dif- ferences in plant growth, water level, and the avail- ability of soil moisture. The changes in water content in the bare-soil tank, however, were very small, and the water lost by evaporation was derived in approximately EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY equal shares from precipitation and water supplied to the tank. Infiltration from winter precipitation was the princi- pal source for replenishment of moisture in most of the soil above the capillary fringe. The water available to the plants from soil moisture varied annually as a result of variations in winter precipitation and opportunity for infiltration. The changes in soil-moisture volumes between the end of one growing season and the begin- ing of the next were a measure of the portion of the winter precipitation that was stored during the non- growing season. The total precipitation during this pe- riod actually was greater than the measured quantity, as part was lost by evaporation from the soil and plant surfaces, perhaps some by evapotranspiration, and some by sublimation of snow and ice. A description of the instruments and procedure used in the measurement of soil moisture in the evapotrans- piration tanks, illustrations of composite soil-moisture profiles in the tanks for the species studied, and some selected results are given in the section "Soil-moisture determinations" in this report. The section also includes a brief discussion of seasonal soil-moisture changes in shallow flood-plain deposits as observed in the vicinity of the Winnemucca test site. These latter findings demonstrate (1) the characteristics of soil-moisture availability to phreatophytes growing on the flood plain and (2) the relative volumes of water stored in the unsaturated flood-plain sediments and car- ried over after wet seasons to the following season or discharged by evapotranspiration or by ground-water outflow in dry seasons. The volumes of water in storage in these sediments also may be indicative of the capaci- ties available for inflow and storage of floodflows and the corresponding effects on streamflow downstream. Observations during a flood period in June 1962, for example, indicated the water content of the sediments to be more than 1 acre-foot per acre greater than in Sep- tember 1961, which was near the end of a 3-year dry period. By October 1962, the water content had declined only about a half of this amount, with resultant carry- over of about one-half acre-foot per acre to the follow- ing season. The water content at the end of the follow- ing seasons, 1963-65, increased slightly, but by the end of the 1966 season, which was dry, the water content dropped to the September 1961 level. The large changes in storage in 1962, 1966, and 1967 thus would be rep- resentative of significant differences between the annual precipitation, streamflow, and water available to plants on the flood plain. D11 COMPUTATIONS OF EVAPOTRANSPIRATION Evapotranspiration was computed as the sum of the water supplied to the tank, precipitation during the growing season, and the difference in soil moisture at the beginning and end of the growing season. Because the water from the three sources was measured in different units, each was converted separately into equivalent depth of water in feet for the area of the tank; that is, acre-feet per acre. Water supplied to the tank during the season was considered to represent the draft on ground water under the conditions of growth in the tank. Water that may have been added or withdrawn prior to the beginning of the growing season in order to estab- lish the desired operating water level for the season was not included in the computations of evapotranspiration. PLANT GROWTH AND DEVELOPMENT The growth and development of the plants were ob- served and recorded by photographs and measured by means of transects across the tanks. Photographs in color (35 mm transparencies) and in black and white were taken at 4-6-week intervals during the growing season. The color transparencies provided a chronological rec- ord of changes in the color of the foliage. Subtle changes in color, indicative of distress in the plants resulting from an unfavorable environment, could be detected by comparing the transparencies with transparencies taken at a prior time or with the actual growth in the tanks. For example, during the summer of 1962 there was a slow progressive change in the color of the foliage of the greasewood plants in both tanks from dark green to yel- lowish green, the result of a deleterious salt buildup in the soil. After the plants had become well established, meas- urements of the vegetative growth were made, beginning with the second full growing season. From a series of measurements made in July, August, and September of 1963 and 1964, it was found that the foliage was at a maximum and that most of the plant growth for the season had occurred by about the first of August. When the early part of the growing season was warm, maxi- mum plant development occurred in late July ; when the season was cool, the maximum occurred in early August. Some variation in foliage development among the four species was observed also. Willows were the earliest to reach their peak in growth and foliage development, and rabbitbrush was the latest. In a given season the differ- ence in time was not great, generally only 2 to 3 weeks. D12 FOLIAGE VOLUME Foliage volume, computed from measurements of the height and crown intercept of the plants, is the product of cover density (expressed in percent), vertical thick- ness of the foliage, and surface area of the tank. In gen- eral, the foliage of the four species extended from the crown of the plants to the surface of the tanks, and the height of the plants thus was equal to the thickness of foliage. The cover density and thickness of foilage were obtained by transects across the tanks. The number of transects for a tank depended on the size of the tank. Four transects having a total length of about 140 feet were measured across each greasewood tank-t wo on the diagonal and two at the midsections. Diagonal transects were used for the smaller rabbitbrush tanks in which the two transects measured totaled 57 feet, and the wil- low and wildrose tanks in which the two transects totaled 28.5 feet. Cover density is considered synonymous with crown cover and is defined by Horton, Robinson, and McDon- ald (1964, p. 9, 36) as "the amount of ground covered or shaded by the vegetation foliage." Cover density is measured by vertically projecting the transect intercept of the crown of the plant onto a tape stretched hori- zontally across the tank and by noting the length of the intercept. (See fig. 11.) The summation of the vertical projections of the crown intercepts, expressed in per- centage of the transect length, is the measure of the crown cover for the transect. FIGURE 11.-Scientist measuring cover density and plant height on a transect across greasewood tank 2. The results of one diagonal transect across grease- wood tank 2, the computations for obtaining the aver- age weighted height or thickness of foliage, and the cover density together with the method for computing STUDIES OF EVAPOTRANSPIRATION the volume of foliage are shown in the following example : Transect across greasewood tamk 2, southeast to northwest corners, Aug. 4, 1966 Vegetation Height Interval times (ft) Species ! Height _ Intercept intercept (ft) (ft) (sq ft) 0.040: 1.0... 00290 Dy,. Aan altel ana 1:0 to: G 1. 1 1. 0 1. 10 20140: BB. cic By: SINCE ULI Cea na eae 8.9 10, 00.0. G 1.1 2:4 2. 97 $210 B x Terra buns can to G 2. 0 2:7, 5. 40 12210 G 3. 2 4. 6 14. 72 10:8 to 217.2. e en's BX 21.7 A0 20.0.2 .n il. c o. G 3. 0 4. 3 12. 90 26:0 G 2. 5 5. 0 12. 50 31.0 to 95.9.1 B:!" eden a 35.0 t0 G 1. 3 2. 1 2. 73 38.0 to 38:5. G (Dead plant) $5.0 to B vei cee Erbe cn abe a malsle 39.019 G 1.7. 2.1 3. 57 41.1140 D!: irons eens 41.9 10 42.0.:...0....2.0. G 8 1. 0 80 cer: ue ane au nes ap une e's fale mal mre 25. 5 56. 69 'B, bare; G, greasewood. NoTE.-Length of transect 42.9 ft. Other computations are as follows: Computations for transect: Average weighted height= 25.5100 12.9 =59.4 percent. Computations for tank, based on four transects: Average weighted 2.05 ft. Average cover dengity.L....... .... - ccceweaksed 49.55 percent. (ATON OI T2. ILE. L {lor inne un on E- 4h we aul as an n 900 sq ft. Foliage volume: 2.05%900%0.4955=915 cu ft. 56.69 253—222 ft. Cover density= Variation occurred in the height of the plants as well as in the height of different parts of individual plants, for example, between the center and perimeter of the crown. These variations necessitated refinement of the method of height measurement so that an average plant height for each transect could be determined. For indi- vidual plants, the results of several measurements of the height of the crown along the transect were averaged. For plants of different heights, the transects were seg- mented into intercepts across a plant or group of plants of approximately equal height. An example of the variation in height is afforded by the willow tanks. During the winter of 1963-64, rab- bits-present despite the ostensibly rabbit-proof fence- gnawed the bark of the stems girdling and killing some plants and severely damaging others. Regrowth from the crown was rapid, and by August 1964 the new growth was about one half as tall as the undamaged growth. This combination of new and old growth made it quite difficult to obtain a measure of the average EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY height of the plants. In computing the foliage volume, a weighted average plant height was used. The weighted average plant height for each tank was obtained by dividing the sum of the products of crown intercept and height of the plant for the transects by the total length of the transects. The results of the transect measurements for each season for the various tanks and the foliage volume com- puted from them, together with the depth to the operat- ing water levels for that season, are given in table 1. Tasu® 1.-Growth and development of woody phreatophytes grown in evapotranspiration tanks at the Winnemucca test site Cover Plant - Foliage - Depth to Year density - height - volume water level Remarks (percent) (ft) (cu ft) - below surface of tank (ft) Greasewood tanks 1961 Sept. 14:.........:. 25.5 1.21 276 5.0 Tank 1. 29. 5 1.35 358 5.0 Tank 2. 1962 June ... 49.6 1. 37 610 5.0 Tank 1. 39.1 1.43 508 5.0 Tank 2. Aug. 55. 4 1. 58 784 5.0 Tank 1. 38. 9 1.43 501 5.0 Tank 2. 1963 July 61. 9 1.97 1, O91 5.0 Tank 1. 47.9 1. 91 823 5.0 Tank 2. Sept. 0...1........ 56. 4 2. 01 1,015 5.0 Tank 1. 46, 4 1.75 730 5.0 Tank 2. 1964 Aug. 62. 4 2.14 1,195 6.0 Tank 1. 47.3 1.97 839 6.0 Tank 2. 1965 Aug. 59. 2 2.19 1,161 7.5 Tank1l 50. 4 2.06 (O Tank 2. 1966 Aug.: 81.7 2.26 1, 046 7.6 Tank1. Ok 49. 6 2.05 915 6.2 Tank 2. 1967 July 26>...2..0...0. 52.1 2. 24 1,045 7.8 Tank1. 49.3 2.12 940 7.8 Tank 2 Rabbitbrush tanks 1963 July 17,18.......... 37.8 1.20 197 5.0 Averageof 3 tanks. Sept. - 46. 2 1.45 272 5.0 Do. 1964 Aug. §: 0100000... 0. 51.0 1. 64 335 5.0 Do. 1965 Aug. 64.1 1.90 411 5.0 Do. 1966 Aug. 4,6.2.:020001.} 56.0 2.05 460 5.3 Do. 1967 July 58. 8 2. 35 558 6.2 Do. Willow tanks 1961 Sept. HMH:2...2...;._ 81.6 3.26 266 5.0 Tank 1. Ad 2.82 219 5.0 Tank 2. 78.8 2. 04 282 5.0 Average of 3 tanks. 1962 June 14:2} 0.0.00... 84.8 3. 21 272 5.0 Tank 1. Wung. 96. 8 4. 24 411 5.0 Do. 86.9 4. 38 381 5.0 Tank 2. fare 92. 8 4.32 400 5.0 Average of 3 tanks. Sept: 88. 3 4. 40 389 5.0 Tank 1. 87.2 5.06 441 5.0 Tank 2. 89. 5 4. 68 419 5.0 Average of 3 tanks. 1964 Aug. 410.205.0000. 278.9 2 4. 33 2 342 5.0 Tank 1. 2 90. 5 2 4. 97 2 450 5.0 Tank 2. 287.7 4. 51 2 306 5.0 Average of 3 tanks See footnotes at end of table. 372-487 0O-10--3 D13 1.-Growth and development of woody phreatophytes grown in evapotransptration tanks at the Winnemucca test site-Con. Cover Plant - Foliage - Depth to Year density - height - volume - water level Remarks (percent) (ft) (cu ft) - below surface of tank (ft) Willow tanks-Continued 1965 83.7 4. 34 363 5.7 Tank 1. oer eer 94. 3 5. 60 528 5 3.5 Tank 2. ______________________________ (3 1966 Aug. 58.1 3. 74 217 5.8 Tank 1. 74.8 5.23 301 4.2 Tank 2. 1967 July 27.2 2.0.0.0... 56. 7 3. 74 212 5.4 Tank 1. 75.2 5.72 430 4.1 Tank 2. Wildrose tanks 63. 5 1.73 110 5.0 Tank1. 24.1 1.20 290 5.0 Tank 2. 50. 5 1. 54 78 5.0 Average of 3 tanks. 77.7 2.11 164 5.0 Tank 1. 60. 6 1. 67 101 5.0 Tank 2. 108 ; M Meaca one 72.2 2.10 154 5.0 Average of 3 tanks. 4 Aug 20. 79.8 2. 34 187 5.0 Tank 1. 78. 6 2. 57 189 5.0 Tank 2. jops 80. 6 2. 64 214 5.0 Average of 3 tanks. Aug 75. 5 2.13 161 5.0 Tank 1. 72.7 2. 65 193 ( 4.2 Tank 2. A Ep M Upasana eta seule 3 1966 Aug.: (b...... 64. 0 1.73 111 5.9 Tank 1. 60. 0 2. 42 145 4.2 Tank 2. 1967 July 76. 6 1.77 136 6.1 Tank 1. 73.0 2.81 205 4.4 Tank 2. ' Variable water level in tank 2, resulting from leak in plastic membrane. ? Plants damaged by rabbits gnawing bark; some stems died. 3 Tank 3 discontinued on account of leaks in membrane. The variations in growth and development of the plants are reflected in the changes in foliage volumes from one season to the next, as shown in figure 12. Only those tanks are shown for which the record of foliage volume and operating water level are continuous for the period of record. Curves for willow tank 3 and wild- rose tank 3 are not shown, as they were discontinued in 1965. For the most part, the foliage volumes showed a rather uniform increase during the early part of the study when the plants were becoming established and the operating water levels in all the tanks were main- tained at a depth of 5 feet. In order to observe the effects of differences in depth to water on the growth rates after the plants had be- come established, the operating water levels were changed from the 5-foot depth. The relation of foliage volume to those changes in water level is shown in the curves of the foliage volumes of greasewood, willow, and wildrose. For the rabbitbrush tanks, in which the 5-foot water level was maintained after the plants had become established, the foliage volume increased at nearly the same rate over the 3-year period 1963 through 1965; practically the same rate continued in 1966 and D14 1200 G@REAsEwOOD /\\\ |.1000|- [r ele _c: I} w. Tank 1 Tank 2." xa o taa tn s " - 8 800 l/I 4 % u 600 r 3 / 3 / Tanks a ,// 1 and 2 1 400 /}t Tank 2 G lk. op Li Roo o 1s s 4 -! [el LL 200 |- Tank] 1 Tanks 1 and 2 0 600 WILLOW - 500 - Tank 2// \ to p \ Li y A. Ina U avin ied \ 9 ,// \\ Pret c Bh 1 3 409 L "to- Tank 1 [ L2) /. Rl Tank 2 = / a e ess $* 3 / 4. anks Tank 1 3 7 1 and 2 9 / u 200 |- 0 < ha & 100 |- 1961 1962 1963 1964 1965 1966 1967 STUDIES OF EVAPOTRANSPIRATION RABBIL’BRUSH Average of 3 tanks // -| 6 $ DEPTH TO WATER LEVEL, IN FEET BELOW SURFACE OF TANK WILDROSE | to -—| [3 3 x N | p Tanks 1 and 2 | wn \\ | o DEPTH TO WATER LEVEL, IN FEET BELOW SURFACE OF TANK 2T Sr" 1 Tank 2 X | w \ # # 1962 1963 1964 1965 1966 1967 FIGURE 12.-Seasonal foliage volumes and operating water levels for the four species of phreatophytes grown in evapotranspiration tanks. 1967, when the water level was lowered from 5.0 feet to 5.3 feet in 1966 and from 5.3 to 6.2 feet in 1967. There was a moderate to small increase in foliage volumes for the greasewood tanks in 1964 when the water level was lowered from 5 feet to 6 feet below the surface of the tank; however, the amount of increase was less than with the 5-foot water level from 1962 to 1963. The record for tank 2 was interrupted during the 1965 season because a leak in early August resulted in a variable water level for the remainder of the season. When the leak was repaired in the spring of 1966, the tank again became operational. The foliage volume (as shown) was measured in 1965, but the water level is not shown because of its variation. Foliage volume for tank 1 decreased in 1965 and 1966, when the water level in tank 1 was lowered from 6.0 to 7.5 feet, and in tank 2 in 1966, when the level was lowered to 6.2 feet, follow- ing repair of the ruptured membrane. In 1967, however, with the water level in both tanks at 7.8 feet, there was no further decrease in the foliage volume of tank 1, and there was a small increase in tank 2. It is inferred from this fact that the plants had become established and adjusted to the deeper water levels. If such is the case, it points up an important principle of phreato- phyte growth ; namely, that a lowered water level affects the plant only temporarily pending readjustment of the root system to the new environment, provided the lowered water level is not beyond the reach of the roots. The effect of the damage to the willow plants by rabbits during the winter of 1963-64 resulted in a de- crease in the foliage volume for tank 1 and a reduc- tion in the rate of increase in foliage volume in tank 2 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY during the 1964 growing season. The foliage volume in tank 1 increased during the 1965 season despite a water level lower than that in 1964. The increase in foliage volume for willow 2 was much greater in 1965 than in 1964. Part of the increase was attributable to the higher water level, raised from a 5-foot depth to a 3.5- foot depth, and part to a continued recovery from the damage by rabbits. On the basis of the differences in the increase of foli- age volume in the two willow tanks having different water levels, it is inferred that the plants did not recover completely in 1965 from the damage in the winter of 1963-64. During the second growing season, the plants evidently were still recovering from the damage. The decrease in foliage volume in tank 2 in the 1966 grow- ing season is believed to reflect the effect of the greater depth to water level in 1966 when the water level was lowered 0.7 foot from 3.5 to 4.2 feet below the surface of the tank. The cause for the sharp decrease in foliage volume in tank 1 in 1966 is uncertain. Several plants died in 1966, and this is reflected in a decrease in cover density from 83.7 percent in 1965 to 58.1 percent and 37 percent in 1966. The cause of death was not apparent. There was only minor damage by rabbits and none by insects. In 1967 the water level was raised 0.4 foot in tank 1, from 5.8 feet to 5.4 feet below the surface of the tank, while in tank 2 the raise was only 0.1 foot from 4.2 to 4.1 feet. The foliage volume of tank 1 was almost unchanged from 1966, but in tank 2 there was a sub- stantial increase. The halt in the downward trend in tank 1 and the increase in tank 2 are believed due in part to the higher water level and in part to the adjustment of the roots to the changed water environment as de- scribed for the greasewood tanks. The plants in the wildrose tanks had in general a similar pattern in that large foliage volumes were pro- duced when the water levels were shallow, and small volumes when the water levels were deep, except in 1965. In that year, with no change in water level, the foliage volume in tank 1 decreased about 14 percent. The reduction resulted from the death of stems of several plants in July. Many variables affect plant growth and development such as climate, soils, salts in the soil, and water supply. However, none of these are responsible for the change in foliage volume of the tanks from year to year, as condi- tions were the same for each pair of tanks for each species. The differences in the paired tanks, shown in figure 12, can be due only to changes in the depth to water level during the growing season. In order to have some basis for comparing greasewood growth in the tanks to growth in the field outside the tanks, several 100-foot transects were measured in the field. The transects, measured during the 1964, 1965, and D15. 1966 growing seasons, sampled three localities having different growth conditions. The first locality was within the test-site enclosure near the north side; there the growth had not been disturbed by livestock grazing since the site was fenced in 1960. Although greasewood is not very palatable, cattle will browse on it at certain times of the year when hungry. At the same time, the plants may be damaged through trampling or by breaking of plant stems. The second locality was outside the test-site enclosure and about 30 feet away from the first locality. There browse conditions were the same as those within the enclosure at the time it was fenced. This locality is in a pasture that is grazed moderately during the winter months. The third locality, also outside the test site, was in a pasture that was heavily grazed and had frequently held small herds of cattle for periods of as much as 1 month. Damage to the plants from grazing and trampling was quite apparent. The results of the transect measure- ments, together with comparative data on average foliage volume of the two greasewood tanks, are shown in table 2. 2.-Comparative data on greasewood growth in the field under different growth conditions and the average in the two evapotranspiration tanks Plant Cover Foliage Year height density volume (ft) (percent) _ (cu-ft per acre) Inside Winnemucca test-site enclosure 1964... c eni enc lc ers 2. 37 51. 3 52, 960 19052 .e 22 2 = oe .o s aed an a aks 2. 39 57. 3 59, 650 19662 c e 2 LNC ALLIE 2. 30 57. 8 57, 910 Outside test-site enclosure (grazed moderately) 1904 . . L022 el nen ocean cleans 2. 85 41. 3 42, 280 Outside test site (grazed heavily) 19064: euler iets. aod 2. 08 28. 0 25, 370 1905. . c c. alden en a- 2. 04 33. 1 29, 410 1900-2 CD0. ne ert r 2. 04 36. 8 32, 700 Average of two evapotranspiration tanks 1964. cs cn celles sou ules 2. 06 54. 8 49, 170 1965... 2 el eee eau o aed 2. 12 54. 8 50, 620 1900... :.. er. iud 2. 16 50. 6 47, 590 Comparisons of plant growth on the basis of foliage volumes show that the undisturbed growth in the en- closure is about 15 percent greater than the average of the growth in the two tanks. This difference is not sur- prising in view of the fact that these plants were well established at the time the tanks were planted and have D16 been able to grow undisturbed since that time. The growth outside the enclosure, in the moderately grazed pasture, was about 10 percent less than the average of the two tanks. The difference in growth between the two localities, for which the plant growth was about the same in 1960, show that there was a substantial improve- ment in growth when the plants were protected from livestock grazing. The effect of heavy grazing is reflected in the foliage volume of the transects in the heavily grazed pasture. Plant growth there was less than half that in the en- closure, about 40 percent of that in the moderately grazed pasture, and about 65 percent of that in the tanks. These data indicate that man's activities in the man- agement of livestock operations adversely affect grease- wood growth. Other woody phreatophytes may be af- fected to a greater extent, as the palatability of some, such as willow, is higher than that of greasewood. PROBLEMS In the planning of the evapotranspiration studies, some problems were anticipated in the operation of the tanks, in transplanting, and in establishing growths representative of those outside the tanks. Some problems arose as the result of nature's handiwork, whereas others resulted from the disturbance of nature's balance. An important objective of the project construction and development was the maintenance of the natural en- vironment insofar as possible. In the tank construction, disturbance of the soils could not be avoided, and the soils were mixed to some extent and could not be re- placed in the same layered sequence in which they oc- curred naturally. The problems caused by nature included damage by insect infestation during the growing season and by rabbits during the winter months; they were not re- lated to construction or operation of the tanks. Insect infestation was not restricted to the plants grown in the tanks but was widespread over the countryside. That condition recurs at intervals of several years. Damage by rabbits gnawing the bark of willow plants may be local or widespread depending on the severity of the winter and the scarcity of food available to the rabbit population. The accumulation of deleterious salts in the root zone in several of the tanks was one of the problems resulting from the disturbance of nature's balance. DAMAGE BY WEBBING INSECTS In early July 1960, a webbing insect spun webs and laid eggs on the greasewood plants in the tanks and on about 90 percent of the area of greasewood growth STUDIES OF EVAPOTRANSPIRATION in the Humboldt River valley, including the Winne- mucca reach. The insect was identified as belonging to the family Pyrolidae, genus Zuwmysic sp. Infestations of the insect appear periodically throughout grease- wood areas in Nevada and Idaho, and doubtless other western States. However, their only appearance during the period of study was in July 1960. sp. was the night-flying insect that was discovered in 1950 by George Zappettini, State Forester for Nevada, and re- ported on in his dissertation for a master's degree at the University of Idaho. The larvae has a snout mouth, is from 10 to 16 millimeters long, is 3 millimeters in diameter, and has a voracious appetite. The adult has a wing span of about 23 millimeters and a body length of about 10 millimeters. | Webbing was confined almost entirely to greasewood and associated species such as rabbitbrush, shadscale, and occasionally a sagebrush plant. No webbing from this insect was observed in willow growth. The webs had a white or silver luster and were so numerous that in some areas when the sun was low on the horizon, as in early morning or late afternoon, they gave a soft silver sheen to vast areas of greasewood growth. Damage due to the larvae feeding on the leaves was extensive; many of the leaves on the plants were shriveled and brown and presented a dead appearance. Serious damage to the greasewood plants in the tanks was avoided by spraying with the insecticide DDT on July 16. The plants in the tanks were only about 3 months old, and it was evident that unless the ravages by the insects were checked, the plants would be severely damaged and perhaps die. Although damage was extensive, the plants recovered rapidly following the spraying; by early August the plants presented a healthy green appearance and extensive new growth was apparent. A different webbing insect, the so-called "tent caterpillar", spun webs, laid eggs, and produced larvae that damaged the willow plants. This insect, although also widespread, did not blanket the willow growth, but occurred randomly in groups or clusters. With the exception of the 1966 growing seasons, webs of the insect were found on one or more of the plants in the willow tanks beginning with the 1962 season. Serious damage again was avoided by spraying with DDT. It could be argued that preventive measures should not be taken to guard against damage by insects, that they are a part of nature's checks and balances, and hence that the studies of evapotranspiration that use such measures do not simulate natural conditions. However, the newly transplanted greasewood shrubs were in grave danger of high mortality, and unchecked damage would have delayed the studies a full year. EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY Furthermore, the occurrence of the infestations at inter- vals of severals years, rather than yearly, reduces their impact on the average annual evapotranspiration loss. As the tent caterpillar does not damage the willow growth everywhere, but only in spots, evapotranspira- tion losses from undamaged plants would be a closer representation than that from damaged plants. DAMAGE BY RABBITS The fence enclosing the test site was constructed to be rabbit proof, but rabbits did manage to enter. The rabbits probably came through the gate at times when it was open and workmen were in a remote part of the site. During the severe winter of 1963-64, rabbits dam- aged the willow plants in the tanks so badly that a number of the stems died. The rabbits gnawed the bark of the stems just above the surface of the tank and those that died were girdled or nearly so. Although vigorous regrowth occurred from the root crown during the 1964 growing season, a comparison of foliage volumes of the 1963 and 1964 seasons indicated that it was not enough to compensate for the loss of foliage from the dead stems. Damage to the plants by rabbits was noted also during the winters of 1964-65 and 1965-66, but it was not extensive and only a few stems died. During the 1963-64 winter, there was damage also to the willows growing in the site outside the tanks, but it did not appear to be as severe as that in the tanks. Possibly the bark from the younger trees in the tanks was more palatable than that from the older trees. In an effort to prevent or at least reduce further damage, a program to remove the rabbits from the test site was begun in the fall of 1964. Using a humane trap, in which the rabbits were captured alive, several rabbits were caught and then released outside the fenced site. The trapping program is believed responsible for the lesser damage in the following years. SALT CONTENT OF THE SOILS The soils of the terrace deposits, used in the grease- wood and rabbitbrush tanks, were suspected to have an appreciable content of alkaline salts. However, because there was a luxuriant growth of greasewood on these soils, no difficulty was expected in growing greasewood plants in tanks containing the same soil, particularly as greasewood is known to have a high tolerance for alkaline salts. The soils of the flood-plain deposits, used in the wil- low and wildrose tanks, were considered to have a low salt content. These deposits are subject to repeated leaching by overflow of the Humboldt River during spring runoff. The flood plain supported a dense and D17 luxuriant growth of willow, wildrose, and meadow grasses. The higher salt content of the terrace deposits was confirmed by the appearance of an incrustation of salts on the surface of the greasewood and rabbitbrush tanks as the result of maintaining a high water level while the plants were becoming established, whereas there was no incrustation on the willow and wildrose tanks. A water-soluble portion of the incrustation samples col- lected in August 1960 was made by treating 1.015 grams of the ovendried sample with a liter of water. The sample gave the following concentration in the extract : mg/l mg/l OH) e Ere 2.4 PH : cscs oce 10.1 ME. . .._ lel eens T OOS cloe nian o ey 93 NA NC. level 115 SO _o miners 60 Dissolved solids_______ 850 0 eae pano 28 In April 1962, samples of water from the saturated material below the water level were collected, for chemi- cal analysis, from all but rabbitbrush tanks 1 and 3 and the bare-soil tank. The samples were obtained by with- drawing water from the water-distribution system of the tank by pumping. In order to have a representative sample from the 10- by 10-foot tank, at least 50 gallons was pumped to waste before sampling ; greater volumes were pumped from the larger tanks. A sample of water was also collected from the well supplying water for the evapotranspiration tanks. Later, in August 1962, addi- tional samples of water were collected from greasewood tank 2. The results of the chemical analyses, given in table 3, show clearly that the mineral content of the soils of the terrace deposits is considerably greater than that of the flood-plain deposits. BORON TOXICITY A wholly unexpected and serious effect on the growth of, and water use by, the greasewood plants occurred during the 1962 growing season. The first signs of dis- tress, noticed in July 1962, were tip burn of the leaves and a change in color of the foliage. The color change from a normal dark green to a yellowish green became more pronounced as the season advanced. In addition to those symptoms of distress, there was progressive de- foliation during August and September, and by early October some plants had lost more than half their leaves. Chemical analysis of leaf samples and of the soil in the tanks indicated that the difficulty was probably due to toxic concentrations of boron in the root zone. The first indication that boron was responsible for the difficulty was the chemical analysis of a second sam- ple of the incrustation on the soil surface of greasewood tank 1 collected in the fall of 1962. The results of the second analysis were startling, for it showed a boron D18 STUDIES OF EVAPOTRANSPIRATION TaBu® 3.-Chemical analyses of water from the saturated soil in the evapotranspiration tanks of the Winnemucca test site [Constituents given in milligrams per liter] Gallons Specifi pumped Calcium Magnesi- Sodium Potassi- Lithium Bicar Carbon- Sulfate Chloride Phos Boron Dissolved 005?!ch- Date collected 1962 _ before (Ca) um (Mg) (Na) um (K) (Li) bonate ate (804) (C1) phate (B) solids ance pH collecting (HCO) (CO;) (POA) (micromhos sample at 25°C) Project supply well Abr 1,000+ 48 14 103 t cn 320 0 78 $> nis: $4 sot 785 0 7.81 Greasewood tank 1! Apr. 153 43 1, 260 31 0. 29 2, 750 0 518 384 5.5 13 3, 800 5,190 8. 14 79 28 1, 030 25 . 24 2, 020 50 388 312 6. 4 9.2 2, 970 4,190 8. 42 56 19 942 28 . 21 1, 840 39 326 274 7.0 8.8 2, 650 3, 800 8. 40 Greasewood tank 2 ! 500 99 30 768 1, 680 0 338 3, 460 8.19 50 51 15 119 360 0 88 877 7.65 960 56 15 208 700 0 156 1, 600 8. 13 1, 470 53 15 350 798 0 171 1, 780 7. 80 , 430 61 14 433 959 0 176 2, 050 7.70 , 650 51 14 449 1,020 0 179 2, 150 8. 13 Rabbitbrush tank 2 ! ADL. 100 197 5 768 20 0. 24 1, 260 0 672 492 0. 57 2.1 2, 910 4,170 7.95 Willow tank 1 ? 50 107 27 143 13 0. 09 668 0 73 50 0. 09 0. 47 799 1, 220 7. 82 Willow tank 2 ADI. -le ce nos 50 115 31 143 13 0. 09 632 0 78 97 0. 05 0. 40 836 1,320 7.42 Willow tank 3 ? abt 50 117 34 140 13" - 0:09 ces 0 73 so 'ol" ' o 4s 850 1,340 _ 7.55 Wildrose tank 1 * lX." so 79 46 162 B ~ an 749 0 53 =_ ogy, -. oft 83s 1,310 _ 7.60 Wildrose tank 2 Apris. 50 99 36 183 # 01 815 ( 49 67. ; 006 - 0.61 set 1,380 |_ 7.87 Wildrose tank 3 : ADI. 50 120 49 140 11 0. 10 792 0 57 80 0. 08 0. 38 896 1, 410 7. 44 1 Terrace soil. 2 Flood-plain soil. content of 416 milligrams per kilogram. To assess more | dried portions of the leaf and root samples are given exactly the amount of boron in the plant and in the soil, | below : samples of leaves and of soil were collected in October [Leaves were collected on Oct. 18, 1962] 1962. The soil was sampled at 1-foot intervals to a depth Greasewood tank 1 wE P + U of 8 feet in greasewood tank 1 and to 6 feet in tank 2.) shed teaves from aix plants:. ol cle c l_. 120 and to a depth of 4 feet in rabbitbrush tank 1. Samples | Green leaves, from the tips of branches________________ 233 “fiare colleictefl with a 3-1ncdll1 earlih auger. In augslrlng for iressewood Tank 3 the s in gre od tan consi uan- P E {TSTWO 2’ A [S derable I all; Shed leaves from two plants_________________________ 125 tity of root material was encountered 'between depthS | qreen leaves from tips Of 196 of 1 and 6 feet. The small roots and rootlets were care- | Root material between the depth of 1-6 feet___________- 54 fully dsefiarated from thelseyerai silfi-arlrjples and pre- plants arowing ontside of tanks serve £ arate analysis o ir boro ntent. g: AnyInS the n conte Shed leaves from several plants_______________________ 107 Samples of the leaves shed were collected from both tanks and from plants growing outside the tanks, and The boron content of the shed leaves was about 20 samples of green leaves were plucked from plants | percent higher for the plants in the tanks than for growing in the two tanks. those on the outside, and the green leaves had nearly The results of the analyses for boron content of air | twice the content of boron that the shed leaves had. A EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY sample of green leaves was collected from plants grow- ing outside the tanks, but unfortunately it was lost. The soil samples were analyzed for their water soluble boron content by the Nevada Soil Testing Laboratory at the University of Nevada. The greatest concentra- tion in both greasewood tanks occurred between the depths of 3 and 4 feet, as shown in figure 13. In the graphs of figure 13, the boron content is plotted as the midpoint of the sample interval, 0.5, 1.5 feet, and so forth. The four samples from the rabbitbrush tank all had the same boron content, 5 mg/kg. SOURCE AND ACCUMULATION OF BORON IN THE TOP SOIL The terrace deposits are rich in boron, as was estab- lished by the chemical analysis of soil samples. The mineral form of occurrence is not known, but the results of soil analysis indicate that some of the boron is readily soluble in water and that a larger amount is in a rela- tively insoluble or only slightly soluble form. Boron in the water soluble form is presumed to be readily available to the plants and was given the most attention. 0 Greasewood 1 ba Greasewood 2 . Greasewood 1/1 p I Water level EXPLANATION DEPTH BELOW SURFACE OF TANK, IN FEET o I Oct. 22, 1962, before leaching April 15, 1963, after leaching $ | | 0 10 20 30 BORON, IN MILLIGRAMS PER KILOGRAM FIGURE 13.-Soluble boron content of the soil in greasewood tanks 1 and 2 before leaching and in tank 1 after leaching. In addition to boron, the sample was found to have a lithium con- tent of 20 mg/kg. D19 In order to obtain information about the total con- centration of boron in the deposits, four soil samples were analyzed. The samples were collected from grease- wood tank 1 with a Veihmeyer soil-sampling tube. The sampling on October 23, 1963, was from the surface to a depth of 5.5 feet, at 14-foot intervals. Determina- tions were made of the soluble boron in all the samples and of total boron in four samples at depths from 0.5- 2.5 feet. Total and soluble boron content of the four samples are given in the following tabulation : Depth interval (ft) Total boron _ Water soluble (mglkg) boron (mg/kg) 0.0 to 1.0.5 700 1. 7 1:0 10 875 . 18. 1 1:5 10 1, 140 7. 34 2010 280 11:7 The soluble boron content probably would have been higher had not the soil in the tank been leached the previous October. The analyses indicated that a large supply of boron is stored in the soil in some form not readily soluble; however, this material may be slowly decomposed by chemical reactions with water and solutes in the tank, and accumulations of the decomposition products in the tanks could lead to toxic quantities of soluble boron. The analyses for insoluble boron content illustrate the magnitude of the total boron supply in the tanks. The depths at which the samples were collected have no relation to natural soil profiles, as the soil in the tank had been disturbed. In addition to the sampling in the tank, three sets of samples of the soil profile and two sets of surface- soil samples were collected for soluble boron determina- tion from the undisturbed soil outside the tanks. The sampling points were in the midst of a thicket of about 15 greasewood plants. The starting point for the lines of lateral samples was the central stem of one of the large plants. Figure 14 shows the location of the sam- pling points, the lateral lines of sampling, and the sur- rounding greasewood growth. The dates, depths, and sampling interval are given in the following tabulation : Vertical sampling Sampling point Date Depth Sampling sampled interval (ft) (t) Ress ehe rian Oct. 22, 1962... 0-9 1. 0 Deke Apr. 9, 1963.____ 0-1. 5 0. 25 ome s Oct. 23, 1903... 0-5. 5 0. 5 Lateral sampling Distance from Sampling Sampling line Date central stem _ interval of greasewood (ft) plant (ft) Apr. 9, 1963... 1=s 1 Oct. 25, 1963... _ 0-8 1 D20 STUDIES Surface soil sampling lines, 3-inch depth D-F. 9 samples at; I F / BPD D-E. 8 samples at J F 1-foot intervals F a 2 _, 1-foot intervals " f fi I E. a Enlarged view of a part of A B 1 in. = 10 ft Soil sampling points A, B, and C (of soil profile) C. Sampled 0-5.5 feet . at 6-inch intervalsQQQ E taha @E.@—— . Sampled 0-1.5 fee 0 at 3-inch intervals A. Sampled 0-9 feet 7 (o) o at 1-foot intervals (Q? © Greasewood tank 2 Greasewood tank 1 A 1 in. = 20 ft FicurE 14.-Location of points and lines used in sampling undisturbed soil, in relation to clumps of greasewood. The nine samples collected from sampling point A by means of a soil auger were analyzed by the Nevada Soil Testing Laboratory at the University of Nevada. All the other samples were analyzed by the Water Resources Division Laboratory of the Geological Survey at Menlo Park, Calif. The samples collected from point B were taken from the side of a 1.5-foot dug pit, and those collected from point C were taken with a Veih- meyer soil-sampling tube. All the samples were analyzed for water-soluble boron ; in addition, the pH and specific conductance were determined for the samples for sam- pling point A, and sodium for the samples from line D-F. The results of these analyses are given in tables 4 and 5 and are shown graphically in figure 15. These analyses demonstrate that there is an accumula- tion of soluble boron in the surface and near-surface soil. The concentration in the surface soil is greatest in the vicinity of the greasewood plants, and that in the near-surface soil is greatest between the depths of 1 and 2 feet. The data suggest that evapotranspiration by the greasewood is largely responsible for the accumulation and concentration of boron in the soil. The chemical analyses of the samples of the soil, roots, and leaves show that there is uptake of soluble boron by the roots, trans- location of the solute through the roots and stems to the leaves, temporary storage in the leaves, and release to the soil. The process of release may include guttation, leaching by washing of the green leaves by rainfall, decay and leaching of the leaves after being shed, or a combination of these processes. Experiments with barley OF EVAPOTRANSPIRATION Tasur 4.-Results of analysis of samples of undisturbed soil profile [The nine samples collected from sampling point A were analyzed by the Nevada Soil Testing Laboratory, Uni- versity of Nevada, Reno, Nev. All other samples were analyzed by the Water Resources Division Laboratory, U.S. Geological Survey, Menlo Park, Calif.] Specific Depth interval (ft) pH conductance Boron (micromhos (mg/kg) at 25°C) Sampling point A [Sample Oct. 22, 1962] 0 to 9. 6 29, 000 24 116 9. 6 24, 000 10 2 £09... 9. 6 7, 500 8 Sto 4.........~ 9. 6 3, 500 5 4 10 9. 5 1, 500 4 210 0°....._.l... 9. 2 1, 100 6 C10 7... 9. 0 700 5 T 9.1 800 5 510 9. 0 700 5 Sampling point B-Dug pit [Sampled Apr. 9, 1963] 0:0 400,251. 208.0 - nissan a awl ale ny 13 0:25 10 0.5. .._ L L2 cense Ccs aud 18 0:8 to c_ . 29 0:75 10 1.0. ... uas. 26 1.010 1.25... ln uo 32 12540 1.5. :..... ce as 25 Sampling point C [Sampled Oct. 23, 1963] 0.010 0.:5-0: 08: IIL OLL 10. 5 056 20. 7 1:0 to 1.5.2.9. 009000 32. 0 1.540 2.0... 002002 ee ann 20. 3 2:0 10 2.50202 III one an aes 13. 5 2:9A0 9.0... .s le L sak 4. 3 oeil ces 1.8 0-0 10 4.0. .._: c ine 1. 4 4:0 10 4.5. .:.. LILAC e aa ak ank 1. 0 4.0 ... Le dept une 1. 0 5.0 19 5.5.02 ceca a nei . 9 plants grown in a boron solution show that boron may be lost from the leaves by both guttation and rainfall washing the leaves (Oertli, 1964). Following release from the leaves, boron seemingly is moved downward in the soil by infiltration of rainfall or melted snow water. As about two thirds of the annual precipitation falls during the winter months when evapotranspiration is low, the downward movement probably occurs at that time. During the spring and summer months, evapo- transpiration dissipates the winter accumulation of water, leaving behind the boron and mineral salts. The boron and specific conductance curves for sam- pling point A, shown in figure 15, illustrate this condi- tion. The concentration of boron in the top 1.5 feet of the soil column is shown in greater detail by the curve for sampling point B, where the samples were collected at 3-inch depth intervals. The distribution of boron in EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY % 0.5 : Point B > # DEPTH BELOW LAND SURFACE, IN FEET 0 10 20 0 40 BORON, IN MILLIGRAMS PER KILOGRAM SPECIFIC CONDUCTANCE (EC x 10° at 25°C) 00 10,000 20,000 30,000 Point C.\Boron / \\ Point] A. Specific confluctance Point] A. Boron DEPTH BELOW LAND SURFACE, IN FEET 10 0 10 20 30 40 BORON, IN MILLIGRAMS PER KILOGRAM 120 1200 Line |D-F. Borory\ 800 C o Line D-F. Sodium, & hy 400 BORON, IN MILLIGRAMS PER KILOGRAM SODIUM, IN MILLIGRAMS PER KILOGRAM fl D-E. Boron 0 2 4 6 DISTANCE FROM CENTRAL STEM OF GREASEWOOD PLANT, IN FEET co FiGurE 15.-Results of boron, sodium, and electrical-conductance deter- minations of soil samples collected from sampling points shown in figure 14. D21 Tapur 5.-Results of analysis of surface soil sampled to 3-inch depth at 1-foot intervals from central stem of a greasewood plant [All samples were analyzed by the Water Resources Division Laboratory of the U.S. Geological Survey, Menlo Park, Calif.] Distance from central stem (ft) Soluble boron - Sodium (mg/kg) (mg/kg) Sampling line D-E [Sampled Apr. 9, 1963] 100.0 L.. c It 19 ! erent es bale 10 } nases tne rels aeons oe a 11} : nea awed 19; cin }... 0.2.2. - «aka naan sen aed 11) Tas irene a cate tow 3 A bss 0 o es o Sampling line D-F [Sampled Oct. 23, 1963] 0 back 4. 4 86. 2 I- sin .n oren eben aus 4. 7 150 e eran cr anagem s o 8. 2 467 O 14. 5 707 Aue nne dee cai s 45 575 vg. 523 6- 90 487 -= tesa 38 477 Bins ue ie euln ake aad aand 18. 6 623 the soil column according to depth, with the concentra- tion greatest in the top 2 feet, then decreasing with depth, suggests that little if any boron is moved down- ward to the water table by percolating rainwater. Thus, recharge to the ground water by direct precipitation seems to be negligible. If the boron had been more uni- formly distributed throughout the soil column, with perhaps a slight concentration at the surface, downward movement of boron by percolating rainwater and of re- charge to the water table by direct precipitation would be indicated. Boron in the near-surface soil is believed to be re- sponsible for the difficulty with growing crops on the terrace soil. Mr. H. T. Harrer, owner of the land of the Winnemucca test site, reported (oral commun., 1963) that 3 years of intensive irrigation were needed in order to establish a field of alfalfa on a nearby parcel of land. The heavy irrigation each year seems to have leached and reduced the boron content in the root zone to a con- centration that did not affect growth of the alfalfa. The large difference in boron between the two sam- pling lines D-E and D-F is believed to result from sampling at different times of the year. Line D-E, for which the boron content is the lower of the two, was sampled in April at the end of the winter season, whereas line D-F was sampled in October at the end of the grow- ing season. A comparison of the precipitation for the D22 6-month periods preceding the sampling dates shows 3.37 inches for line D-E and 2.82 inches for line a difference of about half an inch. The mode of precipi- tation was also different. The 3.37 inches preceding the April sampling of line D-E occurred largely in the form of snow that accumulated during each storm and then melted slowly at temperatures above 32°F. Such a con- dition provides for a maximum infiltration opportunity that allows wetting of the soil to considerable depth and downward movement of boron. The 2.82 inches of pre- cipitation preceding the October sampling of line D-F}, fell as rain, in widely separated showers. With the ex- ception of 1 day in July when 0.48 inches fell, the showers were light, none of more than 0.3 inch and most of them less than 0.25 inch. During the summer months, when the rainwater is dissipated rapidly by evapotran- spiration, the soil is wetted only superficially, and so there is less opportunity for boron to migrate downward into the soil column than during the winter months. LEACHING AS A CORRECTIVE MEASURE The results of the chemical analyses of the soil and leaf samples indicated that the soil in the tanks was rich in boron, and that there was uptake of boron by the grease- wood plants in quantities that damaged the plants. To assure continued growth and survival of the plants, a reduction of the boron content in the root zone was essen- tial. The measure considered to be the simplest and most effective was to leach by backwashing. In this procedure, water was introduced into the bottom of the tank through the distribution system until the entire soil mass was saturated and water overflowed the tank. As the water moved upward through the soil, boron and other mineral salts were taken into solution and carried away in the effluent. As the soil in both the greasewood and rabbitbrush tanks was similar in texture, and presumably similar in boron and mineral salts contents, both sets of tanks were leached, even though the rabbitbrush plants had not shown any signs of distress. Backwashing of grease- wood tank 1 and rabbitbrush tank 1 began in late Octo- ber 1962, and was nearly completed on November 5 when freezing temperatures effectively halted further opera- tion for the year. Backwashing of the three remaining tanks, greasewood tank 2 and rabbitbrush tanks 2 and 3, was completed in April 1963. During the summer of 1965, the rabbitbrush plants in the tanks began to show symptoms of distress. Leaftip burn was quite noticeable, and toward the end of the growing season, there was some defoliation. The plants in the greasewood tanks, however, did not show any STUDIES OF EVAPOTRANSPIRATION symptoms indicating boron toxicity. The backwashing evidently had not been as effective in the rabbitbrush tanks as in the greasewood tanks, or the greasewood plants had a greater tolerance to boron than rabbit- brush. The latter explanation seems more appropriate. These symptoms of boron toxicity indicated that the rabbitbrush tanks needed to be backwashed a second time. Backwashing of all three tanks was completed in early April 1966. Sampling of the soil in the tanks to determine the reduction of its boron content as leaching progressed was not practicable. Collection of soil samples at depth would have been exceedingly difficult and facilities for analytical determinations were not available at the site. In lieu thereof, the specific conductance (in micromhos at 25°C) of the effluent water was used as a measure of the effectiveness of the treatment. The volume of water used for leaching and the rate at which it was added to each tank was measured with water meters. As soon as the soil mass in each tank was saturated, a 5-foot extension was attached to the supply pipe of the tank to provide a head for the upward movement of water through the soil. The rate of flow through the soil was not the same for all tanks of equal surface area ; however, an approximately uniform flow through each tank was achieved by a rate of inflow adjusted to maintain a head of about 5 feet above the surface of the tank. The rate of inflow to greasewood tank 1 and rabbitbrush tank 1 in October 1962 was less uniform than the rates for greasewood tank 2 and rabbitbrush tanks 1, 2, and 3 in April 1963 and 1966. The difficulty was due largely to the low temperatures during the night causing ice to form on the surface of the tanks and in the surface soil. Samples of the effluent, for analysis, were collected at the overflow point of each tank. Field determinations of the conductivity of the samples were made as they were collected. The results of the conductivity determina- tion were used as a guide in selecting the samples of the effluent for chemical analysis. As shown in figures 16 and 17, the conductivities were high initially and decreased as the leaching progressed. Leaching was continued until the conductivity of the effluent had decreased to about 2,000 micromhos. The value of 2,000 micromhos was arbitrarily selected on the assumption that at that conductivity, the boron content of the soil in the root zone had been reduced to a level that was not harmful to the plants. The results of chemical analysis of the effluent from greasewood tank 1 and rabbitbrush tank 1 are given in table 6. These analyses show a general decrease in con- centrations of all the constituents as the leaching prog- ressed. The results for rabbitbrush tank 1 are somewhat SPECIFIC CONDUCTANCE, IN MICROMHOS AT 25°C 16,000 14,000 12,000 10,000 SPECIFIC CONDUCTANCE, IN THOUSANDS OF MICROMHOS AT 25°C 8000 6000 4000 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY D23 T ~ Greasewood tank 1. -- ~~ Greasewood tank 1. Boron T Greasewood tank 2. Boron 1 | ho t &n o 1 8 BORON, IN MILLIGRAMS PER LITER s - Mice 2000 |- Conductivity ~ ~a Greasewood tank 2." : -~ t* 315 Conductivity 0 1 1 1 1 I 1 1 0 0 4 8 12 16 20 24 28 32 EFFLUENT, IN THOUSANDS OF GALLONS FIGURE 16.-Decrease in boron and specific conductance of the effluent from greasewood tanks 1 and 2 during leaching. 12 T T T T T T T 60 RABBITBRUSH TANK 2 10 - 1/5 EXPLANATION 9 8|- Specific conductance - 40 Boron -| 30 - 20 a ud e ix 110 y a #¢0\. L 1966 a 0 1 1 1 i 1 1 1 0 3 ® el 3 12 T T T T T T T 60 3 RABBITBRUSH TANK 3 Z 10 -| so Z EXPLANATION "g &--&---&--s 8 al- Specific conductance - 40 30 20 10 0 1 1 T I T 1 | 0 2 4 6 8 10 12 14 16 $ EFFLUENT, IN THOUSANDS OF GALLONS FiGURE 17.-Comparison of specific conductance and boron of the effluent from rabbitbrush tanks 2 and 3 for the leachings of 1963 and 1966. DXM erratic, because of different rates of circulation. Owing to the below-freezing temperature at night during leach- ing, regulation of the low rate of flow of the leach water was difficult. Regulation of the higher rate of flow into greasewood tank 1 was not a problem until near the last 48 hours of the operation. During the latter period, the rate decreased from about 79 gallons an hour to about 46 and then to about 32 gallons an hour. The ef- fect was to increase the concentration of most of the constituents in the effluent, as shown in table 6. The in- crease in boron and conductivity is shown graphically in figure 11. Boron increased from 8.7 to 16 mg/l and conductivity from 1,720 to 2,450 micromhos with the decrease in rate of inflow from 79 to 46 gallons an hour. The increase was not unexpected, for with slower move- ment through the soil, the leach water was in contact with the soil grains for a longer period of time; thus, greater opportunity was afforded for the constituents to be taken into solution. Less difficulty was experienced in maintaining a uni- form flow through the tanks leached in April 1963. In greasewood tank 2, the rate averaged 224 gallons per hour, varying only about 2 gallons per hour during the period of leaching. In rabbitbrush tank 2, the rate averaged 70 gallons an hour, and for rabbitbrush tank 3, it was 99 gallons an hour. Comparison of the specific conductance curves of the effluent for the leachings of April 1963 and April 1966, shown in figure 17, indicate that the effluent in 1966 was less concentrated. Boron determinations were made only for rabbitbrush tank 3 in 1963 and for tank 2 in 1966; so a direct comparison of the difference in the boron STUDIES OF EVAPOTRANSPIRATION content of the effluent from the two leachings is not available for either tank. It seems reasonable, however, to assume that, like the conductivity, the boron content of the effluent was less in 1966 than 1963. The rates of flow through the tanks were smaller in 1966 : 54 gallons an hour for rabbitbrush tank 2 and 81 gallons an hour for tank 3, a decrease of 16 and 18 gal- lons an hour, respectively. As a result, the leach water had a greater opportunity to take the soluble salts into solution. The decrease in concentration of the effluent in 1966, even though the opportunity to take more salts into solution was greater, indicates that the leaching of 1963 was effective in removing most salts; however, ap- parently the 1963 leaching did not reduce the boron con- tent to a concentration that was not damaging to the plants. Because the plants in the greasewood tanks that were leached in 1962 and 1963 did not show any evidence of boron toxicity in 1965 or in 1966, the greasewood evidently has a higher tolerance to boron than does rabbitbrush. FACTORS INFLUENCING EVAPOTRANSPIRATION RATES Evapotranspiration was defined earlier as water withdrawn from soil by evaporation and plant transpi- ration. Evaporation from the soil surface generally is the smaller fraction of evapotranspiration. The results from the bare-soil tank studies at the Winnemucca test site indicated that the evaporation from soil was less than half the evapotranspiration from the vegetated tanks. Transpiration by plants and evaporation from the soil TaBus 6.-Quantity, rate of inflow to tank, and chemical analyses of samples of the effluent from greasewood tank 1 and rabbitbrush tank 1, October 1962 [Constituents in milligrams per liter] Effluent Elapsed - Rateof Conductivity Calcium Magnesium Sodium Potassium Bicarbonate Carbonate Sulfate Chloride Phosphate Boron pH (gal) time inflow to (micromhos _ (Ca) (Mg) (Na) (K) (HCOq) (CO3) (804) (CD (PO4) (B) (br) tank (gph) - at 25°C) Grease wood tank 1 0 0 - 11, 300 68 25 2, 830 326 4,060 59 1, 680 1,270 24 20 8. 32 2, 150 20. 5 105 9, 400 50 21 2, 500 243 4,720 143 854 653 44 33 8. 42 3, 780 37.0 99 8,170 30 16 2,180 191 4,210 245 606 475 52 37 8. 62 6, 140 60. 0 103 6, 180 15 8.5 1, 630 137 3, 090 280 358 285 48 38 8. 80 8, 440 85.0 92 4, 750 11 4.8 1, 190 101 2, 240 200 225 182 39 8. 91 10, 700 109. 0 94 3, 740 w -- - -- - - =- l - = = 12, 570 133.0 86 3, 020 11 3.9 740 60 1 420 172 135 96 27 19 8. 90 14, 570 157.0 83 2, 820 - - - - - -- - - «- - - 6, 460 181. 0 79 1,720 28 9.8 378 32 818 59 102 66 13 8.7 8. 63 17, 570 205. 0 46 2, 450 11 3.1 587 42 1, 300 54 116 68 28 16 8. 40 18, 330 228. 5 32 2, 490 7.2 1.9 597 41 1, 340 44 117 66 23 16 8. 40 19, 080 253.0 31 2,370 6.5 2.3 567 41 1, 160 103 104 67 23 15 8. 78 Rabbitbrush tank 1 0 + == 6, 550 13 18 1, 630 90 2,320 202 793 508 9.0 12 8. 83 550 28. 5 28 6, 430 18 11 1, 710 79 3, 330 108 507 356 12 18 8. 48 1, 820 51. 5 45 6, 620 24 10 1, 720 82 3, 070 103 591 488 15 25 8. 48 3, 060 116. 0 19 3, 460 47 16 766 37 1, 330 0 333 206 6.5 8.6 8. 10 3, 710 163. 5 14 4,150 43 15 972 42 1, 600 0 464 384 8.5 11 8. 04 4,320 211. 5 13 1, 890 54 16 366 22 750 0 198 152 2.9 3.3 7.99 4,530 285.5 9 5,210 43 11 1, 250 56 1, 870 34 645 528 12 13 8. 33 5,160 284.0 13 3, 840 44 17 864 40 1, 410 7 443 330 7.9 8.0 8. 28 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY occur in response to the same energy sources as evapora- tion from a water surface. The response, however, is modified by the physical characteristics of the soil and the physiological characteristics of the plants; the ef- fects of the modifications are not fully understood. For a given climatic condition, with water nonlimiting, as for phreatophytes, the rate of transpiration depends on the species, cover density and plant size, stage of matu- rity, and tolerance to mineral salts in the soil and water. For a given plant species, the annual rate is affected by climatic conditions such as temperature, wind move- ment, humidity, solar radiation, rainfall, and length of growing season. Of these, temperature is the most im- portant for it determines the warmth of the growing season and controls the length of it. The growing season has been defined by the Phreato- phyte Subcommittee of the Pacific Southwest Inter- Agency Committee (1966) as "the season that is warm enough for plants to grow." Generally it is considered as the period between the last killing frost in the spring and the first killing frost in autumn. The minimum temperature that constitutes a killing frost for one species may have little or no effect on more hardy species. The four species of phreatophytes studied were hardy plants, native to the Humboldt River basin, and accli- mated to the basin. Data on the minimum temperature conditions that would constitute a killing frost for these species are uncertain. A killing frost for them may be interpreted as that which damages the foliage severely enough to cause defoliation, thus restricting the rate of water use. It is not a killing frost in the sense that the plants are killed, for the shrubs are perennials that persist from year to year and survive below-zero tem- peratures during winter months. However, some gen- eralizations may be made. Observations in the area in- dicate that the four transplanted species withstood, without apparent damage, temperatures of 32°F or slightly lower that severely damaged or killed less hardy plants such as alfalfa, garden and flowering orna- mental plants. Threshold temperatures of 32°F, 28°F, 24°F and 16°F are reported regularly by the Weather Bureau as an aid for those people concerned with dam- age to plants of different degrees of hardiness. The threshold temperature that appears to stop growth in the four species studied is 28°F. This interpretation is based on observations of the plants and on a decrease in water use following minimums of 28°F. The yearly periods between minimums of 32°F are shorter than those between minimums of 28°F. For comparisons of the lengths of the growing seasons con- trolled by minimums of 32°F and of 28°F the earliest and latest dates of these temperatures and the lengths D25 of the periods between them are shown in the follow- ing tabulation : Threshold temperatures 32°F 28°F Earliest Days Latest Earliest Days - Latest June 5 Aug. 28 84 Apr. 30 Sept. 19 142 May 12 Sept. 24 135 Apr. 20 Oct. 24 187 * Aug. 30 101 May 7 Sept. 19 137 Sept. 9 117 May 6 Sept. 17 134 June 3 Sept. 2 91 June 3 Oct. 4 123 May 14 Sept. 14 123 May 14 Sept.14 123 22. e eous ool bok ae an ane Bode 100 os crane. 141 The warmth of a growing season, of a month, or of any period of time may be described and compared on the basis of degree days-a degree day being 1° of the average daily temperature above a base of 32°F for 1 day. Thus an average temperature of 42°F for 1 day is equivalent to 10 degree days. The degree days by months and for the period April through October at the test site are given in table 7. The data in the table show that the warmth in the early and late months varies widely. For example, April 1967 was cooler by 354 degree days than April 1962 (equivalent to an average daily tem- perature difference of 11.8°F ; October 1962 was warmer by 456 degree days than October 1966 (equivalent to an average daily difference of 14.7°F). In the midsummer months of June, July, and August, however, the range of difference was much less, being about 125 degree days for each month. The relation of monthly draft on ground water (water added) to temperature during the period April through October 1966 for three of the species grown in the evapotranspiration tanks is shown in figure 18; the values are expressed in percentage of the seasonal total. The appearance of some new leaves and buds on the rabbitbrush and greasewood shrubs in April indicated plant activity and water use. Growth activity by the willows, however, was barely discernible. During April, the draft by the rabbitbrush and greasewood plants was a little more than 2 percent of the total use for the season, while that for the willow plants was less than 1 percent. Evapotranspiration during TaBur 7.-Comparative warmth for the period April 1 to Octo- ber 31, 1962-67, by months, at the Winnemucca test site, in degree days above a base of 32°F Month 1962 1963 1964 1965 _ 1966 1967 - Average lac s. 591 363 360 507 495 287 426 May.... Tork 657 831 645 589 871 694 714 June.... sts 942 819 834 891 903 867 876 July.... eau A,147 04,008 | 1,215. 1 Isd (1,110 ~ 1,208 1,163 August.... ---- 1,029 1,078 1,060 1,051 1,162 1,268 1, 107 September...::l.:........ 885 945 728 660 882 960 842 October... 958 TOL 688 605 502 558 669 Total degree days..... 6,209 5,820 5,525 5,487 5,934 5,807 5.797 D26 s T T T T I T irst Rabbitbrush a AEvapotransplra ion ger x Pa T =-- Temperature "asp / 't ort | | | \| | a < |— 0 +- I I | T | | g Evapotranspiration ® Greasewood ake K / a 20 |- eat n uu. o e Temperature Sor // = 3 =p | | | | | | < u o a a. Rr | | | | Will :Evapotranspiration How daal 20 ae. s = m ea X 0 z | | Apr. May June July | Aug. Sept. | Oct. 1966 FIGURE 18.-Relation of monthly draft on ground water to tem- perature April through October 1966 for three species of phreatophytes. Monthly values of evapotranspiration and warmth in degree days are expressed in percentage of the seasonal totals. May, June, and July increased at a rather uniform rate for the three species. The peak rate of ground-water use for rabbitbrush and willow occurred in July, but that for greasewood was not reached until August. The decrease in rate of use following the peaks was rapid for rabbitbrush and greasewood, but much slower for willow. The ground-water use by the rabbitbrush and greasewood during the peak month was 28 percent of the seasonal total ; the ground-water use by the willows was about 27 percent. During the period June through August, the ground-water use by the rabbitbrush and willow was about 67 percent of the seasonal total and that by the greasewood about 72 percent, whereas the corresponding warmth was only 54 percent of the total for the season. The graphic representation of evapotranspiration in figure 18 demonstrates for the three species of phreato- phytes the rates of ground water use, the differences in rates at the beginning and end of the growing season, and the months of peak use. These differences emphasize the need for studies of evapotranspiration by species, for proper assessment of evapotranspiration discharge and, more importantly, the period and rate of draft on the ground-water reservoir. Wind movement affects evapotranspiration by re- moving the humidity-laden air adjacent to the trans- piring leaf and replacing it with air of lower humidity. STUDIES OF EVAPOTRANSPIRATION Wind also affects evaporation from soil and water sur- faces in much the same manner. Rainfall, as described earlier, is scant and generally occurs as showers of less than 0.5 inch. During rain pe- riods, there is an increase in the humidity of the air that results in a reduction of the evapotranspiration rate. The principal effect of rainfall in the Humboldt River valley is its influence on the draft from the ground water by phreatophytes. During seasons of high rainfall, the draft from the ground water is reduced by the quan- tity of rain that enters the soil and becomes available to the plants as soil moisture, as well as by the lesser evapotranspiration rate resulting from increased hu- midity. Conversely, during periods of low rainfall, the draft on the ground water is greater, as there is little opportunity for replenishment of the soil moisture. As evaporation from a water surface occurs in re- sponse to the same energy source as evapotranspiration, the seasonal evaporation from a standard evaporation pan may be considered as an index to the relative eva- potranspiration for that growing season. Thus, when pan evaporation is high, evapotranspiration may be ex- pected to be correspondingly high. The values for the elements of climate (wind, rain, and pan evaporation) as observed at the Winnemucca test site during the April through October periods from 1962 through 1967 are given in table 8. EXPRESSION OF EVAPOTRANSPIRATION LOSSES AND EXTRAPOLATION TO GROWTH AREAS The transposition of evapotranspiration losses by phreatophytes from the place of measurement such as evapotranspiration tanks to natural growth areas in- volves many variables. The two most important are dif- ferences in climate and in plant growth. Other variables include differences in soil texture and fertility, kind and amount of salts in the soil, depth to the water table, and quality of the ground water. Consideration must also be given to the relative stage of plant development at the two locations. Studies at the Winnemucca test site of the relationship of water use to plant develop- ment of the four species of phreatophytes indicate that evapotranspiration was generally greater after the plants had become established and had entered a pe- TaBur 8.-Climatological data for the April 1-October 31 period, 1962-67, at the Winnemucca test site Year Wind move- Rainfall Pan evapora- (April through October) ment (miles) (in.) tion (in.) 196220020; .c unua sass aioe 11, 383 1. 36 62. 04 1908. cul oen cer eee be aia 10, 803 6. 66 53. 94 1904. Le ceo eeu ics aus 9, 555 5. 21 56. 60 1965.0: rea coun Norte . 11, 487 3. 63 52. 72 1906.2 0 ace. : i eis 9, 120 1. 64 66. 79 1907. ._ lll ous 9, 487 3. 60 56. 10 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY riod of vigorous growth than after the plants had reached maturity. The extrapolation of evapotranspira- tion losses from immature phreatophytes to areas of ma- ture growth would result in estimates that would be too large. Realistic transposition of values for evapotranspira- tion losses from the place of measurement to nat- ural stands of phreatophytes is as important as deter- mination of the losses. The two methods commonly used are based on (1) area and (2) volume of foliage. AREAL METHOD Until about 1950, the usual method of expressing evapotranspiration was on an areal basis, that is, depth over a unit area, as acre-inches or acre-feet per acre. Evapotranspiration expressed on an areal basis de- scribes the water use of a vegetated area for the existing growth condition without any indication as to the growth condition. Thus extrapolation on an areal basis is valid only when the conditions of growth, climate, soil, water supply, and water quality are similar. Gen- erally, when the experimental studies are situated with- in or adjacent to the growth area under consideration, the only variables that need to be considered are the variations in plant growth and depth to water. Varia- tions in plant growth are not uncommon in stands of phreatophytes. The plants may range in size from seed- lings to mature plants, and growth density may range from a few percent to 100 percent. As a result, an appre- ciable error may be introduced into the computation of the water use by a natural stand when the conditions of growth for the measurement-area values are not known. In the past, adjustments for marked differ- ences in growth conditions were made arbitrarily by assuming a linear relationship between evapotranspira- tion and variation in cover density. The validity of this assumption is doubtful, as available fragmentary data suggest a greater use when cover density is in the 80 to 90 percent range than that indicated by a direct propor- tion. This may be due, in part, to the "oasis effect." More information is needed on the relation of evapo- transpiration rates to variations in cover density so that the extrapolation may be made more realistically. VOLUME-OF-FOLIAGE METHOD To avoid some of the difficulties and uncertainties in- herent in the areal method, the volume-of-foliage meth- od was developed. In this method the evapotranspira- tion loss or water use is expressed as a unit quantity per unit of foliage volume. The method presumes that transpiration, by a plant species, is proportional to the total transpiring leaf area and thus is proportional to the foliage volume. Transpiration rates vary for differ- D27 ent plant species, in proportion to the leaf area and to the rate per unit of leaf area. The transpiration rate per unit of leaf area has been found to differ markedly among species ('Tomanek and Ziegler, 1962). The volume-of-foliage method requires detailed measurements of cover density and thickness or canopy depth for the computation of foliage volume. The meas- urements for volume determination of the growth in the tanks were made by the techniques and to the stand- ards outlined in the manual by Horton, Robinson, and McDonald (1964) for surveying phreatophyte vegetation. Evapotranspiration on a volume-of-foliage basis may be expressed as acre-feet of water per acre-foot of foli- age or in cubic-foot units. This method obviates the corrections for differences in plant growth, except possi- bly for growth areas of low density. In these latter areas, the method may have limitations due to the "oasis" effect, and this facet needs further study. The evapotranspiration-tank studies were designed and the required data were collected so that the evapo- transpiration rates could be calculated by both methods. RESULTS OF EVAPOTRANSPIRATION STUDIES The evapotranspiration data for the four species of phreatophytes and the bare-soil tank are tabulated in table 9 for the two methods and for the different water levels. The two sets of graphs in figure 19 depict evapo- transpiration-computed by the areal and the volume- of-foliage methods-during the five seasons 1963 through 1967, together with the operating water levels. Graphs are shown only for those tanks for which the records of water use were unbroken over the 5 years; those for greasewood tank 2, willow tank 3, and wild- rose tank 3 are not shown. In the graph of evapotrans- piration based on the areal method, the amounts of the three sources of water that make up the total evapo- transpiration loss-ground water, soil moisture, and rainfall-are shown for each tank. The graph of the volume-of-foliage method shows only the total evapo- transpiration loss. The evapotranspiration rates for the different tanks and the factors that influence them will be discussed separately by species. GREASEWOOD The decrease in evapotranspiration from 1963 to 1964 is believed to have resulted from the 1-foot lower water level in 1964 and the shorter and cooler growing season. Both tend to reduce the water use. The period between the minimum temperature of 28°F was 50 days less in 1964 than in 1963, and the period April through October was 295 degree days cooler. The slight decrease from 1964 to 1965 seems to be due largely to the lower- STUDIES OF EVAPOTRANSPIRATION Tapur 9.-Evapotranspiration by four species of phreatophytes grown in tanks at the Winnemucca test site and evapora- tion from bare soil during the growing seasons 1963-67 Depth to Sources of water in acre-feet per acre Evapotranspiration € Year watt}; level Soil Water Acre-feet Acre-feet of Remarks in feet Rainfall moisture added per acre water per acre- foot of foliage Greasewood 1963 :-. _ n..... 5. 0 0. 44 0. 48 0. 89 1. 81 1. 70 - Average of 2 tanks. 1968 5. 0 . 44 . 44 90 1. 78 1.45 Tank 1. 1964: 6. 0 . 40 . 82 53 1. 25 1. 06 Average of 2 tanks. 6. 0 . 40 . 26 56 1. 22 . 91 Tank 1. 1965 .., 1.5 . 30 . 42 42 1. 14 88 Do. 1906: 7. 6 . 14 . 24 83 1. 21 1. 04 Do. 6. 2 14 . 25 98 1. 37 1.35 Tank 2. e Ale 7. 8 . 80 . 43 56 1. 29 1.10 Tank 1 7. 8 . 30 . 64 62 1. 56 1. 49 Tank 2 Rabbitbrush 1968 j 5. 0 0. 44 0. 23 1. 52 2. 19 3. 22 Average of 3 tanks. 5. 0 40 . 13 1. 09 1. 62 1. 94 Do. £965..:2! 5. 0 30 15 . 62 3 1. O7 3 1. 04 Do. 5. 8 14 43 1. 06 1. 63 1. 42 Do. A 6. 2 30 43 1. O1 1. 74 1. 25 Do. Willow 1963 ..c.c..._l_:..us 5. 0 0. 44 0. 27 3. 23 3. 94 0. 94 Average of 3 tanks. 5. 0 . 44 : 27 2. 89 3. 60 . 93 Tank 1. 5. 0 44 . 28 2. 97 3. 69 . 84 Tank 2. 5. 0 40 . 20 1. 81 4 2. 41 4 . 61 - Average of 3 tanks. 5. 0 40 . 10 1. 76 A 2. 26 *. 66 Tank 1. 5. 0 40 . 28 1. 80 A 2. 43 4.54 Tank 2. 5.7 30 . 29 1. 48 2. O7 . 57 Tank 1. 8. 5 30 . 18 2. 13 2. 58 . 49 Tank 2. 5. 8 14 . 28 2. 30 2. T2 1.26 Tank 1. 4, 2 14 . 22 3. 19 8. 55 . 91 - Tank 2. 5. 4 30 . 25 1. 55 2. 10 .99 Tank 1. 4. 1 30 24 1. 85 2. 39 . 56 Tank 2. Wildrose 1968 1:: 2.2.0 000. 5. 0 0. 44 0. 20 0. 77 1. 41 0. 92 Average of 3 tanks. 19083... 5. 0 . 44 . 14 1. 04 1. 64 1. 00 Tank 1. 19035... 02> ewa , 5. 0 44 . 21 . 69 1. 34 1. 33 Tank 2. 5. 0 40 . 15 . 84 1. 39 . 65 Average of 3 tanks. 5. 0 40 . 10 . 86 1. 36 73 Tank 1. 5. 0 40 14 . 89 1. 43 .76 Tank 2. 1965: .c Laude 5. 0 30 10 . 83 1. 23 . T6 Tank 1. 1965 4. 2 . 80 18 1. 26 1. 74 .90 Tank 2. 5. 9 . 14 19 1. 23 1. 56 1. 40 Tank 1. 4. 2 14 24 1. 75 2. 13 1. 47 Tapk 2. 6. 1 30 21 1. 08 1. 59 1.17 Tank 1. 1907-2 .L. 4. 4 . 30 16 1. 46 1. 92 . 94 Tank 2. Bare soil 1963 1. ng.. 2. 2 0. 44 . 03 0. 30 0; T7. 2. 0 eis ns 1904 1.9 . 40 -. 08 . 63 1:00 ... 1965.00.24. 400. 2. 3 . 30 -. O1 . 23 102, Lans 1066 °_.-.}...s.s.=.Ll-.s 4. 0 . 14 04 18 30 1 May 1 to October 20. i Plants damaged, and some stems dead, as a result of rabbits gnawing 2 Membrane in tank 2 perforated in August 1965. the bark of the plants. 3 The decrease in water use in 1965 may reflect slower growth caused by 5 Tank 3 discontinued in 1965. an accumulation of salts in the root zone. Their presence was indicated ® Unmeasured water entered tank 1967. by tip burn of the leaves, and by a change in color of the foliage during the growing season. EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY D29 4.0 3 - i t y P Gressswaod'r : |-. |). Rabatbnish Willow 1 Wildrose 1 wildrose 2 _| , {} S 3 tanks) i’ red & € 3.0 --: E-~ por- i ) d = = o. Em o 0 © & - m © & al 1963} 1963 | 1967 EVAPOTRANSPIRATION PER UNIT OF FOLIAGE FIGURE 19.-Evapotranspiration by four species of phreatophytes dufiing1 the growing seasons 1963 through 1967 for indicated depth of water level. ing of the water level from 6.0 to 7.5 feet below the surface of the tank, as the length and warmth of the two growing seasons were comparable. The differences in the period between minimums of 28°F was only 3 days and in growing season warmth was 38 degree days. Evapotranspiration increased a small amount in 1966 and again in 1967. During these 2 years, the water level was only slightly deeper than in 1965; the periods be- tween 28°F averaged 11 days less than in 1965, but the growing season was warmer by 447 degree days in 1966 and 320 degree days in 1967. During at least a part of 1965 and in 1966 and 1967, the root system of the plants in all probability had become adjusted to the water environment of the .5- to 7.8-foot water levels. Thus it seems that the length and warmth of the growing season exert a greater effect on evapotranspiration of grease- wood than does a lowering of water level in the 5.0- to 7.8-foot range. The evapotranspiration values for greasewood are believed to be representative for plants in the Hum- boldt River basin having approximately similar growth and water-level conditions. The average value for tank 1 for the four seasons 1964 through 1967 for water levels from 6.0 to 7.8 feet was 1.21 acre-feet'per acre and 0.98 acre-foot per acre-foot of foliage. Ground water sup- plied 50 percent of this water, rainfall 23 percent, and soil moisture 27 percent. For tank 2, which had about 100 cubic feet less foliage, the evapotranspiration aver- aged 1.46 acre-feet per acre and 1.42 acre-feet per acre- foot of foliage during the 1966 and 1967 seasons, for water levels in the depth range of 6.2 and 7.8 feet. Ground water supplied 55 percent of the water, rain- fall 15 percent, and soil moisture 30 percent. RABBITBRUSH The decrease in the evapotranspiration rate from 1963 through 1965 probably resulted largely from the ad- verse effect of boron that still remained in the root zone in toxic amounts after leaching. Climatic effects such as the shorter and cooler growing seasons were doubtless contributing factors, but the impact wa masked by the deleterious effect of the boron. The increase in water use in 1966 and 1967 is largely the result of leaching of the tanks and removal of the boron in the spring of 1966. The warmer growing seasons of 1966 and 1967 were an important factor also. - The evapotranspiration values for rabbitbrush, like greasewood, are believed to be representative for that plant in the Humboldt River basin under similar condi- tions. The average value, for the three tanks for the 1964, 1966, and 1967 seasons for water levels of 5.0-6.2 feet is 1.66 acre-feet per acre and 1.54 acre-feet per acre- foot of foliage. The 1963 and 1965 seasons were not included in the average because the plants were not mature in 1963, and the adverse effect of boron was ap- D30 parent in 1965. Of the 1.66 acre-feet per acre-foot, 64 percent was supplied from ground water, 17 percent from rainfall, and 20 percent from soil moisture. WILLOW The large decrease in evapotranspiration in the two willow tanks from 1963 to 1964 was caused largely by damage to the plants by rabbits during the winter of 1963-64. Evapotranspiration continued to decrease in tank 1 in 1965 while it increased in tank 2. The differ- ences are ascribed to the changes in water levels. In tank 1, the water level was lowered from 5.0 to 5.7 feet, while the water level in tank 2 was raised from 5.0 to 3.5 feet below the surface of the tank. As other conditions were the same, the difference in evapotranspiration can be accounted for only by the lower and higher water levels. Evapotranspiration in 1966 was appreciably higher in both tanks than in either 1964 or 1965. The water level in tank 1 was virtually unchanged from 1965 and in tank 2 was lower by 0.7 foot. The increased evapo- transpiration was due to the warmer growing season, especially during the months May through September, the period of active willow growth. This period was warmer by 562 degree days in 1966 than in 1965. Evapotranspiration in both tanks was markedly less in 1967 than in 1966. As the differences in water levels and in the warmth of the May through September period were relatively small, some other explanation must be sought to explain the decrease. As shown in table 1, the differences in cover density and foliage volumes between 1966 and 1967 were small. Maturation of the plants or concentration of deleterious salts in the root zone in toxic amounts provide the best explanations of the decrease. The decrease is attributed, however, to the adverse effect of the salts because the plants, which had been planted in 1960, were considered to have reached maturity by 1963 and certainly had by 1964. The water supplied to the tanks shown in table 4, al- though low in dissolved solids, is highest in sodium. It is conceivable that the alkali salts in the supply water may have accumulated in the root zone and reached a concentration in 1967 that affected the use of water by the plants. Willows have a low tolerance for alkali salts, and the threshold tolerance may have been exceeded in 1967. Unfortunately data are not available to indicate the threshold tolerance of willows to alkaline conditions. The striking feature of evapotranspiration by willow shown in figure 19 is the heavy draft on ground water, which was the highest for the four species studied. For the 2 years when evapotranspiration was highest, ground water supplied to the tanks averaged 83 percent of the total water use; in the 3 years of lower evapotranspira- STUDIES OF EVAPOTRANSPIRATION tion-1964, 1965, and 1967-ground water supplied to the tanks averaged 76 percent of the total water use. These data indicate that draft on the ground water by willow is relatively independent of rainfall. Thus, in 1963, the year of highest seasonal precipitation, rain ac- counted for 12 percent of the total water use, and in 1966, the year of lowest seasonal precipitation, rain accounted for only 5 percent of the water use. In the other 3 years, seasonal rainfall averaged 15 percent of the water use. Excluding the 2 years 1964 and 1967, when water use seems to have been adversely affected by damage by rabbits and an alkaline condition in the root zone, the average evapotranspiration for the two tanks was 3.03 acre-feet per acre and 0.83 acre-foot per acre-foot of foliage. Ground water supplied 82 percent, rainfall 10 percent, and soil moisture 8 percent of the total evapo- transpiration during these 3 years. WILDROSE The causes of the differences in evapotranspiration in the wildrose tanks are not as apparent as for the other three species. The decrease in tank 1 from the 1963 to the 1965 season may be the result of the cool seasons of 1964 and 1965. The depth to the water level remained unchanged at 5.0 feet during this period. In tank 2 there was a slight increase in evapotranspiration from 1963 to 1964, while the depth to the water level remained unchanged. The slight increase, in contrast to the decrease in tank 1, may have been due to the 90 cubic feet increase in foliage volume. In 1964 the water level in tank 2 was raised 0.8 foot, from 5.0 to 4.2 feet below the surface of the tank, and remained at that level through 1966. The increase in evapotranspiration in these 2 years is believed due to the higher water level in 1965 and to the warmer growing season in 1966. In tank 1 the water level in 1966 was lowered 0.9 foot from 5.0 to 5.9 feet. Normally this would have caused a de- crease in evapotranspiration ; however, the effect of the warm growing season seems to have more than compen- sated for the effect of the greater depth to water. In 1967, with slightly lower water levels in both tanks, evapotranspiration increased slightly in tank 1 and decreased slightly in tank 2. The average evapotranspiration for the 5 years of record was 1.48 acre-feet per acre and 1.01 acre-feet per acre-foot of foliage for tank 1, which had the deeper operating water levels, and 1.71 acre-feet per acre and 1.08 acre-feet per acre-foot of foliage for tank 2, which had the higher operating water levels. Over the 5-year period, ground water supplied 70 percent, rainfall 20 percent, and soil moisture 10 percent of the total evapo- transpiration. Wildrose was second to willow in its use EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY of ground water and utilized less soil moisture than any of the other three species. BARE SOIL The values of evaporation from bare soil given in table 9 show that evaporation from a bare-soil surface is less than half the evapotranspiration losses shown for the phreatophytic vegetation. During the 4 years of record, with the depth to the water level ranging from 1.9 foot to 4.0 feet below the surface of the tank, draft from the ground water, based on the water added to the tank, was 51 percent of the total loss, while rainfall supplied 48 percent. Evaporation for the 4 years ranged from 0.36 to 1.00 acre-foot per acre and averaged 0.66 acre-foot per acre. As expected, the loss was least at the deepest water level, 0.36 acre-foot per acre from a 4.0- foot water level, and 1.00 acre-foot per acre with a 1.9-foot water level. Loss from the ground water ranged from 0.18 to 0.63 acre-foot per acre and averaged 0.34 acre-foot per acre. During June of 1963, 1964, and 1965, some recharge to the ground water in the tank resulted from showers of 0.50-0.75 inch, as indicated by small rises of water level in the tank following showers of one- half inch or more. No large showers occurred in 1966, and water from the light summer rains seemingly did not percolate to the ground water in the tank. SUMMARY One of the largest unknowns in the water budget of the Winnemucca reach of the Humboldt River is the consumptively wasted water from areas of phreato- phytes of low-beneficial usefulness. Studies were begun in 1959 to evaluate the unit annual consumptive waste of four of the common woody phreatophytes-grease- wood, rabbitbrush, willow, and wildrose-growing in the reach. The water use by these shrubs was determined by growing the plants under controlled conditions in 11 evapotranspiration tanks that ranged in size from 10 feet square and 7 feet deep to 30 feet square and 10.5 feet deep. In addition, evaporation was determined from unplanted bare soil in a tank 10 feet square. Evapotranspiration was computed as the total quan- tity of water added to the tanks, the rainfall on the tanks during the growing season, and the reduction in soil moisture between the beginning and end of the growing season. Plant growth, development, and water use were adversely affected during some years by dam- age to the plants and by boron toxicity. Damage re- sulted from rabbits gnawing the bark of willows and from insects feeding on the leaves of greasewood and willows. Boron toxicity resulted from concentrations of soluble boron in the soils at the root zones of greasewood D31 and rabbitbrush. The causes of plant damage were cor- rected by catching and removing rabbits from the test site enclosure and by spraying the insect infestation with insecticide. Boron toxicity was corrected by re- ducing the concentrations of boron in the root zone through backwash leaching. Foliage volumes, a measure of plant growth and development, were computed from transects across the tanks. They provided a basis for comparison of growth by species from year to year and seasonally between in- dividual tanks of the same species. Foliage volumes also provided a basis for expressing water use in terms of foliage. Climatic conditions and the lengths of the growing seasons affected the annual evapotranspiration rates. The most important climatic element was temperature. The highest water use by the plants occurred in June, July, and August when more than two thirds of the seasonal use took place. The least water use occurred in April and October when the use each month was ap- proximately 2 percent of the seasonal use. Evapotranspiration rates were computed by two methods and expressed in two different units-on an areal basis (in depth over a unit area of land) and on a volume of foliage basis (in volume of water per unit volume of foliage). Evapotranspiration expressed areally gives no indication of the growth conditions for which the information was obtained. When expressed by volume of foliage, however, the growth conditions, which are represented by the product of the cover dens- ity and thickness of foliage for a unit area, are inherent in the expression, because the evapotranspiration is presumed to be proportional to the transpiring leaf area, and thus is proportional to the foliage volume. In the extrapolation of experimental data to field areas of dissimilar growth, the volume-of-foliage method is preferable as less uncertainties are involved than in the areal method. The results of the studies ranged rather widely be- tween species. For the same species the seasonal results varied, differences being due to the operating water level, the warmth of the growing season, and plant re- sponse to the effects of damage or alleviation of damage by rabbits, insects, and boron toxicity. Draft from the water table (equivalent to the water supplied to the tanks) varied with the seasonal rainfall, being greatest when the rainfall was scant and least when it was copious. The data obtained in the evapotranspiration tank studies at the Winnemucca test site indicate that during 1963-67 average water use by greasewood ranged from 1.21 to 1.45 acre-feet per acre in tanks 1 and 2, of which D32 50 to 55 percent was supplied by ground water. The average evapotranspiration by rabbitbrush for 3 years, 1964, 1966, and 1967, was 1.66 acre-feet per acre, of which 64 percent was supplied by ground water. Evapo- transpiration by willow was the highest of the four species, amounting to 3.03 acre-feet per acre for the two tanks during the 1963, 1965, 1966 seasons. It was also the highest user of ground water, obtaining 82 percent of its water from that source. Wildrose was the second highest user of ground water and the smallest user of soil moisture. Evapotranspiration by wildrose averaged 1.48 acre-feet per acre in tank 1 with operating levels ranging from 5.0 to 6.1 feet below the surface of the tank, and 1.71 acre-feet per acre in tank 2 with operating water levels ranging from 4.2 to 5.0 feet below the sur- face of the tank. On the average, ground water supplied TO percent of the total use, and soil moisture only 10 percent. SOIL-MOISTURE DETERMINATIONS By A. O. WAANANEN The woody phreatophytes under study in the evapo- transpiration tanks at the Winnemucca test site of the Humboldt River Research Project receive part of their seasonal water supply from soil moisture in the unsatu- rated zone. Precipitation and water added to the tanks during the growing season constitute the principal part of the water supply, but water from winter precipitation stored as soil moisture above the water table may repre- sent a significant part of the water budget. Evaluations of evapotranspiration water use thus would be incom- plete without information on the quantities of water provided from this soil moisture. Initial observations of the water content of the soils in the evapotranspiration tanks were made in Septem- ber 1961 using a neutron-scattering soil-moisture meter in access tubes installed in each tank for this purpose. A regular program of observations was started in April 1962. The purpose of the soil-moisture observations was to determine the water content of the soils at the beginning and end of the growing season and at selected interven- ing times to define seasonal variations. The change in water content during the growing season thus provides a measure of the volumes of water provided to the plants from this source. Data were obtained also at several sites on the Humboldt River flood plain near Winne- mucca, including one tube installed within the Winne- mucea test site, to explore the range of variations in water content in the zone of fluctuation of the water table adjacent to the river and in the unsaturated soils near the land surface. STUDIES OF EVAPOTRANSPIRATION SOIL-MOISTURE OBSERVATIONS AT THE WINNEMUCCA TEST SITE EQUIPMENT AND PROCEDURE The neutron meter provides a convenient means for determining changes in the moisture content of soils. After installation of suitable access tubes, rapid and repetitive observations can be taken in the tubes at any time as the soils are not subjected to further disturb- ance. Differences in the quantity and distribution of moisture in the soils as shown by subsequent observations represent a measure of the changes in the water content. The neutron-scattering soil-moisture meter used con- sists of a depth probe equipped with a 28-milligram actinium-beryllium neutron source, detector tube, and preamplifier connected by cable to a portable counting device (scaler). The probe is stored and transported in a shield that serves also as a standard for relating meter counts to water content and for checking meter operation. In normal use, the probe is lowered in an access tube to desired depths. Fast neutrons emitted by the source enter the surrounding soil materials and are moderated by hydrogen ions present principally in the moisture in the soil. Thermal (slowed) neutrons are detected and counted. The observed count varies directly with the number of hydrogen ions in the soil, contained principally in the water. The effective diame- ter of the sphere of influence of the neutron source varies inversely with the water content of the soil. Appropriate calibration relations permit conversion of the observed counts to moisture content in percent by volume or to weight of water per unit of volume. Figure 20 presents a view of the neutron-meter scaler, probe, and shield as well as soil augers and types of access tubes used in the soil-moisture studies. A typical application of the neutron meter in the tank studies is illustrated in the generalized section of an evapotrans- piration tank shown in figure 21. Initially only one access tube was installed in each of the tanks planted to woody phreatophytes and in the bare-soil tank. These were placed near the center of the tanks but not over any conduit of the water distri- bution systems. Additional tubes were installed later in the two 30-foot tanks planted to greasewood and in one of the 20-foot tanks planted to rabbitbrush to provide information also on the lateral distribution of moisture in the tanks. Four tubes were installed in greasewood tank 1, of which two were placed directly over units of the water system. Tubes of aluminum (2-inch outside diameter, 0.065-inch wall) or alloy steel (1.75-inch out- side diameter, 0.035-inch wall) were used, and these were sealed at the bottom so that observations could be taken at depths below the water level in the tanks. The tubes in the larger tanks were of sufficient length to EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY FIGURE 20.-Neutron-meter scaler, probe, and shield, with soil augers and typical access tubes. permit sampling nearly the full depth of the tanks, but those in the shallower 10-foot tanks limited sam- pling to 60-inch profiles. fTo portable scaler HI “|/Shield D33 Figure 22 presents a view of the neutron meter set up for use in greasewood tank 1 in July 1962. The access tubes installed in the rabbitbrush tanks are vis- ible in the photograph (fig. 23) taken July 1963. These photographs also give some indication of the relative plant growths at the respective times. Soil-moisture observations were taken in the access tubes at half-foot depth intervals below the land sur- face, and moisture contents were computed in successive 6-inch-thick zones. The observed values of water con- tent, representing the integrated result for spherical volumes of soil ranging from 12 to 30 inches in diam- eter depending on the moisture present, were used as the mean for each zone. Near the land surface, the sphere of influence of the source intersects the air-soil inter- face, and the inclusion of some air in the observed volume results in a reduction in the neutron count. Therefore, moisture in the top increment was computed for a 9-inch depth using the moisture value observed Water content THE |, Neutron probe I I I | 1] rM s y *A Una > / siti (1, 4 ~] «a R ye 2 YJ E te~ Sphere of influence \\ NaF al s s of neutron source \ > ¢ T ae! # fizz/l” U \\\F\~\\\ hang Aal Ktr® re Capillary fringe _ Water level dad Plastic membrane tank liner Sealed access tube SECTION OF EVAPOTRANSPIRATION TANK | | | SOIL-MOISTURE PROFILE FiGuUrRE 21.-Generalized section of evapotranspiration tank showing neutron-meter depth probe in access tube for soil- moisture observations, and typical soil-moisture profile. D34 at the 6-inch depth. Tests had indicated that this pro- cedure yielded reasonably reliable values as the reduced count was offset by the moisture differences at the 6-inch and the mid-increment depths. This procedure obviated the need for the precise determination of mois- ture at the surface. Calibration relations for the specific soils at the Win- nemucca test site were reviewed. The heavy flood-plain soils in the meadow area are well leached, and the normal calibrations for the tubes used were deemed applicable. The terrace sands and gravels, in place, contain alkali salts and some boron, as discussed in the main report, that vary in concentration with depth and that have the greatest concentrations in the upper 24 feet of the soil profile (fig. 13). Boron is a neutron moderator, and its presence in appreciable concentra- tions may cause sufficient absorption to reduce the meter count for given water content. Some other salts present in the soils also may have some moderating influences. The materials in the greasewood and rabbitbrush tanks (fig. 2), however, were mixed during the construction of the tanks, and the pattern of the concentrations of salts is uncertain. Gravimetric sampling indicated a F1GURE 22.-Neutron meter set up for use in greasewood tank 1. small effect on the calibration relations, but definition was inconclusive owing to the mixing of the soils and the variations with time resulting from concentration of boron in the root zone by evapotranspiration and subsequent leaching of the tanks to reduce the concen- trations of boron and other salts. Accordingly, the nor- mal calibration curves were used without adjustment, but with recognition that the resulting values might be small. Determination of changes in water content, as shown by differences in moisture rather than the STUDIES OF EVAPOTRANSPIRATION 23.-Rabbitbrush tanks and access tubes installed for soil- moisture observations. total content, was the principal objective. The effect of small errors on the seasonal water budget is minor. WATER CONTENTS OBSERVED Soil-moisture observations were made at the begin- ning, middle, and end of each of the growing seasons from 1962-67. Two additional sets of observations were made during the 1962 season and three in 1963. Post- season observations were made also in December 1963. Water-use evaluations for the first two seasons indi- cated that about one-half of the seasonal water use in the tanks had occurred by the end of July or early August. The summer observations therefore were sched- uled usually for that time to provide a convenient mid- season check on tank operations and an index of midsummer moisture conditions. The results of the spring and autumn observations during the 1962-67 seasons, in the individual tubes in the tanks and in the tube in the meadow area at the test site, are listed in table 10. The changes in water content occurring during the summer and winter periods are also indicated. The losses in the summer season indicate the part of the soil moisture that was discharged by evapotranspiration. These data and the tabulations in table 9 in the main report indicate that soil moisture supplied 26 percent of the total water used in the grease- wood tanks during the study period, 15 percent in the rabbitbrush, 12 percent in the wildrose, and 8 percent in the willow tanks. The contributions varied widely from year to year owing to differences in plant growth, water level, and the availability of soil moisture. The changes in water content in the bare-soil tank were very small, and the soil-moisture contributions to the seasonal evaporation were minor. The data listed in table 10 represent net changes in water content for the full profile sampled, although the principal depletions of soil moisture occurred in the unsaturated zone. Expression of the moisture contents EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY in terms of depth per unit of area permits ready com- parison among tubes and tanks or with precipitation. The results are expressed in inches, but the seasonal water-use values were reported in equivalents of acre- feet per acre in the section on soil moisture and in table 9 in the main report. Profiles of soil moisture in the tanks and the meadow at the beginning and end of each season are shown in figure 24. The water content is expressed in percent of volume, and the differences between the spring and autumn profiles delineate the zones of moisture deple- tion. The profiles for each of the four species of plants are composites of the data from several tanks and tubes for a particular species. The depths of water in the tanks shown also are average levels at the time of sampling; owing to variations in evapotranspirative draft and early season water-level adjustments, these depths may differ from the operating levels for the season. In 1964, the water content observed in the first foot of the capil- lary fringe above the water table in the tanks with plants was about 1 percent by volume greater in October than in April. This condition might have resulted from an increase in soil-moisture tension as available moisture higher in the capillary fringe was reduced by transpira- tion. The additional profiles shown for the meadow area represent the maximum and minimum conditions ob- served during the period September 1961-October 1967. The maximum occurred June 11, 1962, when flooding of the meadow caused essentially full saturation of the soils. Winter precipitation was the principal source of water for replenishment of moisture in the soil above the capillary fringe. The annual replenishment varies widely as a result of variations in the type and rate of precipitation and opportunity for infiltration. The winter changes in water content shown in table 10 are a measure of the part of the winter precipitation that was stored during the nongrowing season. The remain- der of the precipitation was lost by evaporation, some evapotranspiration, and sublimation of snow and ice. WATER-CONTENT VARIATIONS IN 1963 Variations in the water content of soils in the evapo- transpiration tanks during a typical season are shown by the seven sets of observations made in 1963. The water contents for the depths sampled in the tanks, expressed in inches depth per unit area, and the depths to water at the observation times are listed in table 11; the aver- age contents are shown in figure 25. The September 4 observations showed the least content for the season, except that in the bare-soil tank. The increased content in October and December may be both the result of in- crease in capillary water above the water table on the D35 reduction or cessation of evapotranspirative draft and the result of small additions from precipitation. Capil- lary rise above the shallow water table in the bare-soil tank kept the surface soils moist. Evaporation of the small quantity of free water in these heavy soils caused a 1-foot lowering of the water table and a small reduc- tion in total water content between October 20 and December 16. WATER-CONTENT CHANGES IN SHALLOW FLOOD-PLAIN DEPOSITS The flood-plain deposits in the Humboldt River val- ley provide storage space for large volumes of water in the ground-water reservoir and in the unsaturated soils near the land surface during floodflows. Water stored in these deposits during periods of rising river stage and released to the stream when the river stage falls is one of the principal sources of water that sustains low flows in the Humboldt River. The water may be stored as soil moisture in the unsaturated zone, includ- ing the capillary fringe, and as ground water in the saturated zone. The discharge of water from the flood-plain deposits in areas where the water table is at shallow depth may occur by evaporation from the land surface, by tran- spiration from riparian and flood-plain vegetation- commonly woody phreatophytes-and by underflow to stream channels. All the water going into storage in a given season may not be released in the subsequent low-water season, or for several seasons of copious pre- cipitation or heavy streamflow. The quantity of water stored or released seasonally, or carried over, may be substantial. Data on changes in the water content of the flood-plain deposits have been obtained during the period 1962- 67 at three sites in the Humboldt River valley ; one is at the test site 4 miles southwest of Winnemucca, a second at Winnemucca, and a third at the Kearns Ranch 6% miles northeast of Winnemucca. The access tubes at these sites permit sampling the soil profiles for depths of 81, 100, and 90 inches, respectively, which are ade- quate to cover the full range of the zone of soil-moisture change. Observations were made at the same intervals of time and depth as those in the evapotranspiration tanks. Water-content profiles at these sites, based on five sets of observations made in 1962 as reported by Waananen (1965), showed seasonal changes in a year of high flow in the Humboldt River. Data obtained June 11, 1962, after the meadow had been flooded, rep- resent the largest content observed during the study and indicate the water content of the soils under nearly full saturation. The initial measurement in September 1961 D36 STUDIES OF EVAPOTRANSPIRATION TABLE 10.-Soil-moisture changes in evapotranspiration tanks at the Depth of Water content and seasonal gain (+) or loss (-), in inches depth Tank and tube profile observed, _ Apr.8-10, Change Oct. 18-20, Change May 1-2, Change Oct.20-21, Change Apr. 13-14, Change in. 1962 1962 1963 1963 1964 Grease wood -AL Tens 81 22.16 -4.97 17.19 +4.71 | 21.90 -5.33 16. 57 +1. 78 18. 35 -3. 31 TCB Len pei cn e ne we 99 24.60 -B.91 20.78 +60.54 27.82 -5.47 21.85 41.91 23.76 -4.73 T- NTU i d a m tim ale 09. ee ue es in on 16. 62 4+4.68 21.30 -5.450 15.85 +. 83 16. 68 -2. 22 1 BD: evi an. -w ake 03 Memes be fina resis - 19.90 +4.11 24.01 -4.71 19.30 +.60 19.90 -2 02 Average... <4. 44: L2 +B5.01 . -B. AC A. 28 L cl nn. -9. 07 gyri t niid. T5 17.30 3.65 1355 - aso -644 pas 19356 Is so ~4.03 PSB: 2.002 uk abe nie 99 26.81 -538 21.48 F605 28.50 -6.21 2215 43.16 25.31 -4.79 ST acc Ll -4. 02 ._: 4T. AD s -6,82 :2..l.L.. -4. 71 Rabbitbrush I- Ars ITIN il ate 93 26.02 -1.08 24.99 -225 22.74 -2.857 20.37 F1.00 21.43. -1. 90 J-B- If LucoLn .I UL ST e osa delas 22. 70 -240 20.30 -3. 01 17.29 +159 18. 88 -2. 65 03 LLL Sel io rak aisle ad a ale araw beled wine aie aid o mrad ia ma ut ave ie io in ie haul ie a a o ie aie ale s a a Ga Wd a avs ale where a ae tn an ane Pela ei aN aI u olo whe 93 26.94 -1.08 25.86 -289 22.97 -2.66 20.31 -+} 21.72 -1.12 Spoon din tls 93 27. 06 -1.25 25. 81 -1. 84 23.97 -2.78 21. 19 +>. 356 22, 55 -1.40 92 L0 -A, 12; . -25890 RTN . ELLE 1 o whs we a n't ale ale niels -1. 61 Willow Tell ranna ean ie a aan ke 60 22.00 -4. 12. 17.97 - +2.02 20.590 -53. 25 17.3 - 2. 18 19. 52 -1. 21 P ade e na ale a e'n alain aos 60 22. 47 -6.11 16. 36 +3.20 19.56 -3.32 16.24 +2. 79 19.03 -2.76 e sa n e co a aan 60 22.150 -4.57 17.58 20.93 -3.05 7. 8t +2. 47 - 20.35 -8. 06 Average.... 60. aan oe ole 4 o a a % ated lial maite H cn a i us a fale fat on ae e ie t e - PRT ..a ined (ak a aran anns nest -2. 34 Wildrose 3 oar a coa atv ki dance lee emale b iud a 60 $24.26 4+0.00 24.85 23.46 -1.02. 21.84 +0.70 - 22.54 -1.26 a i aa da e oe wie ne me a a we halle ie 60 _} 23. 48 -. 31 23. 17 -. 98 - 22.19 -2.55 19. 4 +129 20.93 -1.70 e nn Ane nls aan aln as ten 60 324.65 -1.82 22. 83 -. 45 22.38 -3.19 19. 19 +170 20,89 -2.44 60:3 e e a aa n a ieee are ain dl are a l o aie io er ao ein nn n e n as ae -B A3 contre nre snene ank s - 1. 80 Bare soil B0 ° * 28.28 -i1.72) 21.51 (28:10 -O Sp., 22.71 +0.04 22, 75 . +0. 37 Meadow 81 *26.88 -5.59 21.24 +8.13 24.37 -. 21.05 4+3.84 24.89 -2. 52 Oo ho 1 Data not representative owing to puncture of tank membrane liner. 2 Change in depth of profile from 87 to 93 inches. EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY Winnemucca test site during summer and winter periods 1962-67 D37 Water content and seasonal gain (+) or loss (-), in inches depth-Continued 00:1. 9131.21, Change Aprl. 922-21, Change Octl. 92:555—26, Change Maggi—3, Change 00ti9l635-20' Change Aprl. 93-20, Change 00ti916%_20’ Grease wood-Continued 15. 04 +3. 05 18. 09 -5. 14 12. 95 +1.60 14.55 -8.32 11.23 4+4.91 16.14 -5.27 10. 87 19. 03 +3. 68 22. 71 -6. 15 16. 56 +1. 84 18.40 -8.39 15.01 4+4.51 19.52 -6.50 13. 02 14. 46 +2. 07 16. 53 -4. 13 12. 40 +. 835 12.75 -2.08 10.67 +83.22 13.89 -4.62 9. 27 17. 88 +2. 11 19. 99 -4. 63 15. 36 +. 99 16.35 -2.59 13.76 +83.09 16.85 -4.22 12. 63 _________ F240 cc_..ll.l.. p20. .f... -2 84 08 13. 87 +3. 61 17.48 ! -7.72 1 9. 76 Ca +4. 49 14.25 -8.26 10.99 +5.36 16.35 -7.60 8. 75 20. 52 +4. 14 24. 66 ! -9. 40 1 15.26 ! 20.85 -2.86 17.99 +4.97 22.96 -7.71 15. 25 _________ 4-o88 1 -B8,00 ....s..... A 45 04. <8, 06 -........ +5, 10 ::...... -7.60 Rabbitbrush-Continued 19. 44 +1. 73 21. 17 -2. 64 19. 03 +4. 54 23.57 -7.19 16.38 41.67 18.05 -4.06 13. 99 16. 23 +2. 47 22 eee ene al an nie al un a a s in hle bale wie ain o al be bn a o tu un ole aln ate a ae on belies ies a a aul bie ilar ae ___________________ 2 20. 48 -2. 98 17. 45 +5. 84 - 283.29 -8.46 14.83 43.58 18.41 -5.48 12. 93 20. 60 +2. 00 22. 60 -1. 69 20. 91 +8. 05 283.96 -3.33 20.63 +1.82 22.45 -4.40 18. 05 21. 15 +2. 12 23. 27 -1. 04 22. 23 +2. 39 24.62 -4.30 20.32 +2083 22.35 -6. 23 16. 12 _____________________________ cl c_. L IC =a ness saan ss bin ss sunless. T" rily nets eins Willow-Continued 18. 31 +1. 85 20. 16 -8. 45 16. 71 +1.68 18.39 -8.30 15.09 4+3.99 19.08% -3.05 16. 03 16. 27 +3. 37 19. 64 -1. 75 17. 89 +2 21 20.10 -2650 17.45 43.58 21.03 -2.90 18. 13 17. 29 +3. 44 20. T9 Le n ae ne elect n o a a's be alse a it 21.82 ...... 12.70 :4{3.48. . 10.18 Wildrose-Continued 21. 28 +1. 54 22. 82 -1. 17 21. 65 +0. 65 22,30 -2 27 20.08 +212 22.15 -2.56 19. 59 19. 23 +2. 46 21. 69 -2. 18 19. 51 +3. 39 22.90 -284 20.06 +255 22.61 -1.87 20. 74 18. 45 +2. 78 BL TE LCA ool o ule aln o ale al ib a ae a alain le m nares 14/11 > +4.82 18; 09 L_. -< usu yy Bare soil-Continued 23. 12 - 0. 66 22. 46 +0. 09 22. 55 -0. 80 - 21.75 -0.42 21.33 41.25 22.58 -0.99 21, 59 Meadow-Continued 22. 37 +4. 56 26. 93 -1. 86 25. 07 +8. 63 28.70 -13.51 15.19 4+4.65 19.84 +0. 34 20. 18 3 Apr. 4, 1962. STUDIES OF EVAPOTRANSPIRATION D38 3WN7TOA A8 LN3JOH3d NI M3LVM laaa7 aszem _| I Tz lady | = ~ >-: jaas7 6 judy sa 1 _ 1 | aoOom3sv349 L961 9961 S961 P961 €961 2961 O8. "ov Of Of Or 0 jos Of O6 - Os Or o.: Os or of oz, Of o. Os or os oz "Of o os or of -or or o 'os or 'of of (Ot o f- t- T -= ~~ t 7 [r Janay - s < gem - 7 an soz 1°0 - ~ _ 7 ke $s ~ tz judy # pT indy =1 € fidy = csi | (eT sj | #¥A pse {_" moOTIIM I I I I * €13°0 ~ -__ o NI MOT38 H1Ld3G 06 O€ D39 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY "sos au} 30 uofpein}jes [Inj squosordor pure poarosqo 9} St 'I96T 'IT oung 103 argoid Moptapt 'T96T 'FT 3093009 pus j097000 J9]8M UooMjoqg aouoJoJIP soIJoId Mopeour Jof storie '1u0J000 dojsm uf [euostos sSyut} ou} uf sorgoid uo popeyus ©0318 3807 BoontouutM oy} je upe;d poop oy} pus syurj uf; 19-z96T suosuas Suzmo13 ou} Jo puo pus Supuuf89q oy} 36 sotgoid orn3sfOu-[Og-'$J SuUNDLI JWNTOA A8 LNJOU3d NI M3LVM L961 9961 S961 v96T €961 2961 omo¢omomofioomo¢omowofioomovomomofioomowomomofioomovomomofioomoeomowoflo I I I I I I I I I I I I I I I I f. =<+ | T T Afififih 7 ee &\\h S 3 l 3 T2 _M>w._ 2 # N 3 aung ET _ dy of 390} . s 1 \\ 02 31°00 /% ---I | 3 | | | | | _ | | 00 z Rew \; p | te tet -I | s :- 3 oz 390 3 oz 390 - eee -> sz 700 1 may | NI MOT38 H1d3G jaaaq - O€ a2lem | | | | | 1 | o T110S 34v8 -- 09 eldon ten. ¢ C | 3SOHG07TIM D40 in the meadow area at the test site fortunately provided information on the water content at the end of a 3-year dry period, thus perhaps representing the minimum that might be expected. The water content observed at the end of each season from 1961 to 1967 at the three sites is shown in figure 26, together with the maximum content as observed in June 1962. These data demonstrate that the storage in the 2C t T T Water GREASEWOOD & - Content Depth of profile, 90 inches 24 |- Depth to water, 60 inches - 221- - 20 |- - 18 23 I I T RABBITBRUSH b Depth of profile, 92 inches 21} Depth to water, 60 inches - 19 4 24 | | I WILLOW i Depth of profile, 60 inches til 22 |- Depth to water, 60 inches - el S é z € = 20 -I Ra Z u - 4 18- -I a o a € 16 S 26 I I I WILDROSE Depth of profile, 60 inches 24 |- Depth to water, 60 inches _- € 22 - s 20 |- - 18 I I I BARE SOIL Depth of profile, 60 inches 24 Water Depth to water, 60 inches - KContenti 22r- £ 20 - April May June July _ July Sept. Oct. Dec.Dec. 1 1 it' 1 26 4 20 16 31 FIGURE 25.-Variations in water content of soils in evapotranspiration tanks at the Winnemucca test site during the 1963 season. STUDIES OF EVAPOTRANSPIRATION flood-plain deposits increased more than 1 acre-foot per acre between September 1961 and June 1962, but in Oc- tober 1962 only about one-half acre-foot per acre of this increase still remained in storage. In subsequent seasons the storage increased slightly, but in the dry season of 1966 the soil moisture was depleted to the 1961 levels; in 1967 the storage increased again, but less than one-half acre-foot per acre. The distribution of moisture in the soil profile in the meadow area at the test site is shown in figure 24. The profiles for September 1961 and June 1962 are repeated in each graph to indicate the relation of the water con- TaBur 11.-Water content of soils in evapotranspiration tanks at the Winnemucca test site, as observed during the 1963 season Depth of Water content, in inches depth (upper number) and profile _ depth to water, inches below land surface (Jower number) Tank observed (in.) May June July July Sept. Oct. Dec. 1-2 11 1 26-27 4-5 20-22 16-17 Greasewood :set 9114 23.63 22.10 21.26 19.23 18.00 18.30 17.53 59 60 60 60 60 60 91 ves 87 24.98 23.41 22.53 19.53 18.26 18.65 19. 52 87 60 60 60 60 60 72 Average water content.... 90 _ 24.30 22.76 21.90 19.38 18.13 18.52 18.52 Rabbitbrush 90 21.52 20.83 19.92 18.48 18.54 18. 18. 87 61 60 60 60 60 60 74 93 22.97 21.91 21.38 19.57 19.81 20.31 19. 87 59 60 60 60 60 60 TL 93 28.907 24.30 2244 20.65 2119 21.07 61 60 60 60 60 60 72 Average water content...... 92 2282 22.08 21256 190.74 19.07 20.11 19.94 Willow 60 _ 20.59 19.97 16.51 17.34 | 18.11 56 60 60 60 (1) 60 _ 19.56 19.07 15.52 16.24 16.94 57 60 65 62 (®) 60 20.93 20.19 17.43 17.88 18. 90 61 62 64 61 (3) Average water content...... 60 20.36 20.73 223.84 19.74 16.49 17.15 17.98 Wildrose 1 a een een tana ans 60 _ 23.46 22.95 224.00 23.05 2112 21.84 2212 59 60 59 61 61 76 BN AV. A chelvens 60 22.19 21.96 19.12 19.64 20. 35 56 59 62 63 T7 BLED re 60 22. 38 22.58 18.77 19.19 19. 69 55 B ° 56 62 62 79 Average water content.... 60 - 22.68 22.56 224.11 22.53 19.67 20.22 20.72 Bare soil 60 23.10 2282 2272 2227 22.16 22.71 22, 00 26 18 23 28 28 28 40 ! Inlet tube dry. : 2 Tanks flooded; water content from 1962 data under full saturation. 3 Dry. D41 EVAPOTRANSPIRATION BY WOODY PHREATOPHYTES IN THE HUMBOLDT RIVER VALLEY 5° T T T I I T T I I I I T I T I I T I I I I 40 |- a - w LJ i ps Maximum Winnemucca? o + °C contents f S Lee ae meee Sieg F Z --f -__- -f ~ t observed esis -- \\ Z June 11 € ~ xs - e w . % --a T 30|- o o m- 1~ ~- teal i- A ® a (Z) m e o Kearns 8 o ve i N Ranch _A in _ 6 -- -L 6 ~ a E o "/// \\\\o/// g z A / Winnemucca ¥ - test site [e] 20 |- < g gol H 8 a* 10 1 1 1 1 1 1 1 1 1 1 1 1 1 1 T 1 1 1 1 1 1 1961 1962 1963 1964 1965 1966 1967 FicurE 26.-Water content at end of each season and maximum content observed in June 1962, in inches, as observed at three sites in Humboldt River flood plain near Winnemucca, Nev. tent each spring and fall with the extremes observed. The shaded areas indicate the distribution in the profile of moisture in excess of the minimum storage. The Oc- tober 1966 data differ slightly from those for 1961 be- cause the initial observations were taken at different depths. It might be inferred from the slight increase in the October water content from 1963 to 1965 that water storage in the flood-plain deposits had only a minor effect on the annual water budgets for the Humboldt River in those years. The large storage increase in 1962, sharp decrease in 1966, and lesser gain in 1967, however, when related to the full extent of the flood plain affected in the basin, may represent significant differences in the relation between annual precipitation and stream- flow, as well as in the water available to plants on the flood plain. The depletion of soil moisture and ground water in 1966 may have resulted in a larger sustained summer flow in the Humboldt River than would have been produced by the annual precipitation alone, whereas the increased retention in 1962 and 1967 reduced the streamflow. The seasonal water-content determinations thus pro- vide an index of storage capacity and a means for esti- mating the volumes of water that could be accepted by and stored in flood-plain deposits during subsequent floodflows, or that would be effective in maintaining streamflow during dry seasons. REFERENCES American Society of Civil Engineers, 1962, Nomenclature for hydraulics: Am. Soc. Civil Engineers Manual No. 43, 501 p. Cohen, Philip, 1965, Water resources of the Humboldt River valley near Winnemucca, Nevada: U.S. Geol. Survey Water-Supply Paper 1795, 143 p. Horton, J. S., Robinson, T. W., and McDonald, H. R., 1964, Guide for surveying phreatophyte vegetation: U.S. Dept. Agriculture, Forest Service, Agriculture Handbook 266, 37 p. Lee, C. H., and others, 1926, Evaporation on the United States Reclamation Projects: Am. Soc. Civil Engineers Proc., August 1926, p. 1196. Oertli, J. J., 1964, Loss of boron from plants through guttation : Soil Science, July-December, v. 94, p. 214-219. Pacific Southwest Inter-Agency Committee, Phreatophyte Sub- committee, 1966, Glossary of terms relating to the phreato- phyte problem in Vegetation management on flood plains and riparian lands-symposium presented by the Phreato- phyte Subcommittee at 66-3 meeting of the Pacific South- west Inter-Agency Committee, Albuquerque, New Mexico, Aug. 30, 1966, p. 48-54. Sondregger, A. L., 1929, Water supply from rainfall on valley floor: Am. Soc. Civil Engineers Proc., v. 55, no. 5, p. 1139-1165. Tomanek, G. W., and Ziegler, R. I., 1962, Ecological studies of saltcedar : Div. Biological Sciences, Fort Hays Kansas State College, Fort Hays, Kansas, 128 p. Waananen, A. O., 1965, Water-content changes in shallow flood- plain deposits at three sites in Water resources of the Hum- boldt River valley near Winnemucca, Nevada: U.S. Geol. Survey Water-Supply Paper 1795, p. 104-108. U.S. GOVERNMENT PRINTING OFFICE: 1970 - O-372-487 I~. e J DAY vi #4191 - E Water Use by Saltcedar as Measured by the Water Budget Method mCEs ry 3 GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-E DOCUMENTS DEPARTMENT BEC L9 1976. - ] yc LIBRARY unversity 0f Gacrormy - "MN Water Use by Saltcedar as Measured by the Water Budget Method By T. E. A. van Hylckama STUDIES OF EV A TION GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-E UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1974 UNITED STATES DEPARTMENT OF THE INTERIOR ROGERS C. B. MORTON, Secretary GEOLOGICAL SURVEY V. E. McKelvey, Director Library of Congress Cataloging in Publication Data Van Hylckama, T. E. A. Water use by saltcedar as measured by the water budget method. (Studies of evapotranspiration) _ (Geological Survey Professional Paper 491-E) Supt. of Docs. No.: 119.16:491-E Bibliography: p. 1. Tamarix pentandra. 2. Evapotranspiration. I. Title. II. Series. III. Series: United States Geological Survey Professional Paper 491-E. QE75.P9 no.491-E [QK495.T35] - 557.3°08s [551.5572] - 74-17416 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price $1.30 (paper cover) Stock Number 24011-02563 CONTENTS Page Page MADsLTact adn dane varenes obs ain rea cid R16 n £9 g a E1 - Water use, large tanks.-Continued Background of the Buckeye Project ........................ 1 1964: Floods n/ ian i firth a Lisa E11 Reduction of water loss :.:... .psd ne 1 1965: Salinity -. ........ cisa apna ye on caulk aids s 12 Water use by saltcedar. sng. 1 1966: Density of stand :.... s. 13 Measurement of water loss : ..;... ...... c.. r 2 10907: Finale: ...:... .0n y cul acre iia a nan alk am ag 16 Evapotranspirometers ..................................... 2 Soil moisture and soil-moisture fluctuations ................ 16 Definition All na is ands 2 Method of measurement .f 16 Oasisieffect ..:. n rin as heet nas bian natin ign a 2 Ambient air temperature 17 The Buckeye test site., .} ....0, .I mais lain F Salinity. xl l iy anar foal £294: 18 Location.. 2..:;..n. c sls LMA ras . sev a aa canvas digo 3 Chloride content.. :. sira eer ng caer ramas 18 :. :,. %...... cease ee aloe race aig a 3. Vegetation growth and development ..,..................... 19 Climate >.. Me. sts sry ria ri ean dran aa aran d 3 Transpiration rate -..; can sas h nals a's 19 Floods :..... ..:. sis ieri rir y een aa to an rag ad 4 Water tise =/ (c {ivi tos Taak ya ke aie bane s bo a hod 19 Cicadas .: .sis arn an ai ras malas 4 Sunlight. salinity, stand density 19 Fank Construction.. (yn gran rl nig ks 5 Plant variations "sav. ios walk baa a 20 instrumentation ..... 2. me teairra eny ies 7 / ~ Water use, small- tanks ;-. . .tc D. IVs Ni . 20 Floatiess control system s. 7 Harmonic analysis.: FIS SA Nines careless 20 Adverse conditions ...... cy .us y ridin avi 8 Oasis tanks. ~ {sn renal ru ara l va k a caa ae as 23 Water devel eo ys iaa tat 9 Bare-soil tanks . ia aia ev. 24 Soil moisture and temperature 00... 9 Water table fluctuations »...... Cs AQ. .m nC aN i. 25 Vegetation. ..... n ise thn y - Discussion:! n s ".f. IW ITL ira rik s aa ne raid a 27 Tank planting .. ...,. 0. ..i rence alty oy a c ais eight 9 Varying factore of water r... 27 Transpiration volume 9 Methods of evaluating water use 29 large tanks :... ..: 7... .a sil sll rs bin alee raps vae aoe 10 ~References ........ ...:} ...... 9. nian ys, 29 1961°:60: Depth to water f s 10 _ Appendix: Analysis of variance with multiple classification .. 30 ILLUSTRATIONS Page Figure T> Photograph showing saltcedar freely exposed :..., . : .. . 2a reine ay a vas a an ad hain v aan ae nin bae Petrine ba ne wa balsa ay waa E3 2. «Photograph showing saltcedar as part of a thicket . .... ... isis . sa rie rav ar kia pack aind pied nald bis ale a sian g 3 3; Map showing Jocation of the Buckeye test site" .~ .; A. :.. sy one salt von agia wh Ree drag a ph anna ecole bp arie n a naik old ae wie o 4 4-8. Photographs showing: 4.~- Frost damage on saltcedar twigs. ss. ...; 1, !s Sy cul. s sulk aome T awe a naan na tinh ahaa a eed a aik ain dalek wl ail nach 4 5... Two scenes of-results of typical floods . :.. .;. ... .au Es ench vic naa aa daa val han g abn an aat tio a 5 6. Cicada :egress :... .. 2; ..; 2. sa. g L y id a hoa sa ea a ean alah a elin a ae aie nee ane ain d aaa g sos i ae Br a opie winn anid ne 5 7. Saltcedar twigs damaged by cicada ege lay ing: ...s... ..V. 0 . mS sak naaa means c 2a nev ana Wie aia dinah hicle i ank 5 $:. An evapotranspirometer newly planted to saltcedar ...... iis bane ena inane aed naal wah ner s 8 9; Diagram showing docation of tanks and InstTuments ... or yey rac saan s rara aind asl papa ais aige nay ay ad 8 10; Photograph of instrumentation for the water budget .; .. : . 2... .u. all null ana hy oe wine e winn oa aln ng pace a nannini e a neer 8 11.~ Schematic cross section of an evapolranspirometer .. lar vidi as ral ad nline nr sian balls rina ra ned ea enne a o 8 12. Graph showing soil moisture by volume in the six large tanks on a typical summer day in 1965 ..................... 13 13. Bar graph of total water use during 1965 in six evapotranspirometers showing the effect of depth to water and flushing of saling ground water cava sak salva a a Fanless a an nage alie (ia an calan ee righ niall alk ana aaa aind o 13 14. Photographs comparing plant growth for evapotranspirometers 7 and 8 13 15. Photographs of three of the six evapotranspirometers after treatment of vegetation in March 1966 ................... 15 16. Graph showing accumulated water use in six evapotranspirometers during 1966 showing the effects of density of stand and salinity of artifically-maintained ground water . ... . s 00.2 suk ai erice a nia ace dale 1 tie ira aan oda an waa wank laa + 15 17. Graph showing accumulated water use in tanks 1 to 6 between April 1 and September 1, 1967 ......... 16 18. Calibration curve of soil moisture measurements by the neutron method ... kell... 17 19-22. Graphs showing: 19; Soil moisture wariation with -depth in one tani .-..; :.. 23. 0 To ei Paks rare iga wien cle ape anns aln pals on vine 17 20. - Apparent fluctuations in soil moisture during a three-day period in one evapotranspirometer ............... 18 21. Effect of chloride concentrations on the counting ratios of a neutron logger ............................... 18 22. "Boil moisture-in tank 4 in May of-1905, 1965; 1966. -and 1907. ... vives be r rich cans nite an hibs a 18 II IV Ficure TABLE 28. 24. 25. 26-35. gin ® i iio. be f CONTENTS Page Photographs of two- to three-month-old branches of saltcedar from the Buckeye Project site along the Gila River in Arizona E21 Photograph of the small-tank area at the Buckeye Project, March 1963 ...... kkk kkk kkk ek ee eee ees 22 Recording chart of tank 7 showing frequency of filling during a week in May 1967 k...... 0s 22 Graphs showing: 26. Water use in evapotranspirometers 7, planted to saltcedar, and 11, bare kkk... k. 22 27. First and second harmonics of the analyses of average hourly water use in evapotranspirometers 7 and 11 ....... 22 28. Second harmonics of water-use fluctuations of bare-soil tank 11 ll kkk kkk kkk eke +e 23 29. Water use in tanks 7 and 8, planted to saltcedar, surrounded by bare soil (oasis tanks) ................... 24 30. _ Water levels in evapotranspirometer 9 and inverted barometric pressure kk kkk. P 25 31. Water levels in evapotranspirometer 1 and Barometric pressure . ..,... ak 26 32. Water levels in evapotranspirometer 2 and barometric preSSUIe . .ll. ea rk rk rk rk rre ke 26 33. Average water levels, average barometric pressures, and average observed and computed water use in tank 6, planted to saltcedar, for June 30-July 4, 1966 cer rikki nar serts eae tors 27 34. - Yearly water use (1961-1963) versus depth to ground water in six evapotranspirometers at the Buckeye Project 28 35. Total water use per year in six evapotranspirometers versus specific conductance of the saturation extract of the soil samples taken from the root zones in July or August of each year ...................... .. k.} }} 28 TABLES Page Meteorological data for project site near Buckeye, su uaa kake ak wasi ks bene oes rev viable nels E6 Dates of flooding of the Waterman Wash Fir rst ac ab nk s aia bien a's ak aik sible nih nie wie le a a aries 7 Results of surveys of saltcedar, taken in the fall of each year, in the evapotranspirometers .... ................. ...... 10 Monthly water use in evapotranspirometers, excluding rainfall kkk kkk kk kkk kk kkk rks 11 Yearly water use in evapotranspirometers, excluding rainfall (¥ ks kr ks ck ck 6+ 12 Quality of ground water In evapotrankpiromet@rs ..}... .................. {.r cana nere nin an ae anna ea's ain a bia bd n a+ 12 Water use, excluding rainfall, in evapotranspirometers during 1965 66666 rks ks ks 14 Analysis of project water and better-quality water from a new well south of the project site ......................... 15 Water use in six evapotranspirometers between April 1 and September 1, 1966 kkk kkk} lke 16 Water use in six evapotranspirometers between April 1 and September 1, 1967 kk kkk kkk }}}} 16 Change in soll moisture content.... .a 21.2160. 0}. pa na nh haka cae als ae cake wine pas at 8 b a 9g eine a bn ane mig n a 17 Increase in twig length of saltcedar inside and outside evapotranspirometers between April 5 and June 22, 1966 ...... 20 Mean and amplitudes of the harmonic analysis of water use in tanks 7 and 11 kkk kk kkk} }}} 23 Water use-in oasis tanks T -and 87... /: 0... .o. nd hd 1 ds a nl nair b anl ak ak 1s cale wie Ga n son' n a ah Piola a nid aie na ml al a a 24 Evaporation from bate soils in tanks 10 and 11 >:. fii iv il nah anes a aks sak bie me's anl sively 6 oa alain ule a ale nie n 'e 25 Monthly evaporation, excluding rainfall, from ground water in bare-soil tanks at Imperial Camp and near Buckeye, Ariz 26 Summary by years of water use in and treatment of six evapotranspirometers kkk k kkk ks keke 27 STUDIES OF EVAPOTRANSPIRATION WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD By T. E. A. van Hyrckama ABSTRACT Water use by saltcedar (Tamarix pentandra) was studied from 1961 through 1967 near Buckeye, Ariz. The test site, in the flood plain of the Gila River, was surrounded by a kilometer-wide dense strip of saltcedar thickets. Areas to the north and south of the strip were mainly desert and a few cotton fields. The test site growing season lasts 8 to 9 months, the humidity is low, winds are strong, summer temperatures at the site often reach 50°C, and evapotranspiration rates are among the highest in the United States. Evapotranspiration rates and quantities were observed in six plastic- lined evapotranspirometers (tanks) with 81-m' (square-meter) surfaces. Analyses were made on the effects of depth to ground water in the tanks, of salinity of soil moisture, and of vegetation density. Rates of water use from bare and vegetated soil were observed in five smaller tanks (36 m* each). When the depth to ground water, or water table, was 1.5 m (meters) the average water use was about 215 cm/yr (centimeters per year); when the water table was 2.1 m, the use diminished to about 150 cm/yr; and when the water table was 2.7 m, the yearly water use was less than 100 cm/yr. Water use varies greatly with salinity of the soil moisture. Salinity may be expressed in terms of specific conductance of the saturation ex- tract (ECs) in mmho/cm (millimhos per centimeter) at 25°C. In tanks which measured ECs=20, the water use was 70 percent; in tanks which measured ECs=30, the water use was only half that in tanks with an ECs=10. When the vegetation was cut twice a year from an original average height of 3 m. to a height of about 50 cm the water use decreased to about half that in tanks where the vegetation was not cut. However, when the vegetation was thinned to 50 percent of the original density the water use diminished by only 10 percent. The maximum yearly water use (311 cm) was measured in 1965 in a tank with a high water table, a dense vegetation, and an ECs less than 10. Although in half of the 36 cases (6 tanks X6 years) the yearly water use was 150 cm or less, there were 11 tank-years with a water use of 200 cm and more-when the table was high, the salinity was comparatively low, and the stand density was medium to high. The daily fluctuations in water use from bare soil showed that in summer the evaporation at midday diminishes because of the forma- tion of a vapor barrier; but, evaporation continues from the soil un- derneath a dense vegetation. Atmospheric pressure fluctuations which affect the water level in the plastic-lined tanks must be considered when such levels are used to determine water consumption quantitatively. BACKGROUND OF THE BUCKEYE PROJECT REDUCTION OF WATER LOSS The increasing population of the arid and semiarid regions of the southwestern United States and the ac- companying need for more water has continually focused the attention of hydrologists and water managers on ways and means to salvage water when and wherever possible. Current studies reflect efforts to conserve water by: reducing water loss from lakes and ponds (and even plants) by covering the surfaces with chemicals that reduce evaporation; determining the most economical methods of irrigation; eradicating plant growth along arroyos and rivers; condensing moisture in the air to in- duce precipitation; and desalting saline water for in- dustrial, domestic, and agricultural use. The southwestern United States and many other parts of the world will be able to support their present and future populations only if some or all of such studies eventually make more water available. This paper discusses reduction of water loss by conver- ting saltcedar jungles, vegetation which uses a lot of water, into less thirsty pasturelands or bare soil. WATER USE BY SALTCEDAR The Latin name for saltcedar (Tamarix pentandra)," and the often used name "tamarisk", were derived from the name of a river in the Pyrenees. Maybe this led to the belief that saltcedar was imported into this country by early colonists from the Mediterranean regions. Later studies (Horton, 1964) raise doubt on this supposition. (There is, for instance, evidence that in 1823 saltcedar was imported simply as a garden plant, at least in New York.) Whatever happened, saltcedar was introduced and started to spread. No attention was paid to this until construction of reservoirs and excessive use of ground 'Dr. B. R. Baum (1967), botanist at the Hebrew University, Jerusalem, discovered that T. pentandra Pallas might be the wrong name for T. chinensis or T. ramosissima. However, J. S. Horton (written commun., June 1971) is of the opinion that most and perhaps all of the five- stamen tamarisk in North America are in either the T. pentandra or the T. gallica group of genotypes. E1 E2 water began in many places to lower the water table | along rivers. As a result, the native vegetation died, and saltcedar, with its deep rooting system and salt exuda- tion, was left in sole command of the water-depleted areas. During the 1920's people began to realize that the plant might well be using copious amounts of ground water. Thus began the studies on saltcedar and many other phreatophytes. Although saltcedar has some value for the control of soil erosion and for wildlife habitat, these beneficial features are offset by its lavish use of water. The sometimes remarkably deep rooting system developed by this species enables it to use ground water from depths as great as 10 m (meters) or more (30 ft (feet) or more) below the land surface, a feat equaled only by a few other species, usually known by the generic name of phreatophytes or "well plants'" (Meinzer, 1927). Because of this deep rooting capacity, saltcedar may have free agcess to water and, therefore, may consume by evapotranspiration as much or more water than the amount that would (other things being equal) evaporate from a lake surface. For instance, the U.S. Bureau of Reclamation (1964) estimates that along the Colorado River 67,000 hectares (167,000 acres) of saltcedar and other water-loving plants' consume as much as 700 million m* (cubic meters) (568,000 acre-ft) of water per year. It is, therefore, not surprising that eradication of saltcedar has been undertaken over vast areas of the Southwest. MEASUREMENT OF WATER LOSS Water use by vegetation or evaporation losses from bare soil can be measured in several ways. There is a great demand for techniques that use portable or semiportable instrumentation with which one can measure the evapotranspiration indirectly. The advan- tages are obvious: the hardware can be moved from place to place, and, once proper correlation between direct and indirect methods is established, information can be ob- tained in a comparatively short time. Such methods have been successfully tested over open water (U.S. Geological Survey, 1954; Harbeck and others, 1958) and over low vegetation (Rider, 1956; Tanner, 1960) but rare- ly over such high stands as saltcedars. The Buckeye Project was equipped with instruments to observe the radiation balance as well as to collect data on mass transfer. These instruments and their functions will be explained and the data will be presented in another report in this series. The most direct method of measuring water use is the water-budget method, in which an account is kept of the amounts of water applied to, and lost from, a particular container, area, or type of surface. Such a method is ex- pensive and time consuming but generally gives the 'Some of these water-loving plants sharing saltcedar's notoriety have been extensively studied by McDonald and Hughes (1968). STUDIES OF EVAPOTRANSPIRATION most accurate data, provided the physical surroundings are properly maintained and controlled. This paper dis- cusses the water budget and presents data on water use as measured in evapotranspirometers. This project was established as a joint effort by the Geological Survey and the Bureau of Reclamation. The author gratefully acknowledges the generous assistance given by the Bureau and its personnel, especially by Cur- tis W. Bowser. Also the assistance and advice of other in- dividuals too numerous to mention is acknowledged with gratitude. EVAPOTRANSPIROMETERS DEFINITION According to the "Glossary of Meteorology" (Huschke, 1959), evapotranspirometers are instruments which measure the rate of evapotranspiration, the loss of water from the soil both by evaporation and by transpiration from plants growing on that soil. Evapotransirometers consist of a vegetated soil tank designed so that all water added to the tank and all water remaining after evapotranspiration can be measured. Some are quite simple, such as oil drums filled with soil and inserted into the ground. Others are large elaborate structures at- tached to recorders which indicate gains and losses of weight due to gains and losses of moisture. Some have perforated bottoms and the water seeping through can be tapped off, weighed and chemically analyzed; they are called lysimeters, a word derived from the Greek "Arger'" which means "to dissolve'"'. Whereas a lysimeter can nearly always be used as an evapotranspirometer, the reverse is not true. The size of evapotranspirometers is partly determined by the type of vegetation to be studied. Obviously, a small container might suffice for grasses, but in- struments like those built in the Netherlands, which have an area of 625 m* (square meters) (6,725 ft* (square feet)) and are 5 or more meters deep (15 ft and over), may be needed for studying trees. The larger the size, the more difficult it becomes to detect malfunctioning such as leakage (Penman and Schofield, 1941) and to ac- curately maintain ground water at intended levels. OASIS EFFECT It is difficult to imitate natural conditions inside and outside the lysimeters or evapotranspirometers. Often the failure to maintain a representative test environment has resulted in grossly overestimated amounts of water used by similar plants in a natural environment (Mather, 1954). Figures 1 and 2 for example illustrate variations of plant density. The first photograph shows a saltcedar plant standing alone with the fronds (as the terminal branches with their scalelike leaves are called) all green down to the ground. Such plants have a large active surface and therefore are capable of transpiring more water than plants shown in figure 2. Owing to in- WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD w Ficure 1.-Saltcedar freely exposed. Rod is a standard stadia type (feet and tenths of feet). Figure 2.-Saltcedar as part of a thicket. Rod is a standard stadia type (feet and tenths of feet). tolerance to shade and possibly to lack of moisture, these plants have shed many of their fronds. Each plant has much less active surface and consequently can transpire less, other things being equal, than a single plant stan- ding as if in an oasis. Clearly, if the use of water by a single plant such as shown in figure 1 is measured, it is not warranted to apply E3 results on a per-plant basis to an acre, much less to thousands of acres of dense growth. To dispel the oasis effect, an evapotranspirometer must be surrounded by a buffer zone planted with the same vegetation as that of the tank, and all other conditions should be as similar as possible to those of the instrument. The size of such a buffer zone depends on climatic conditions. Thornthwaite and Mather (1955) point out: "In a moist climate such as Ireland a square 50 meters on a side should be sufficient, but in the desert probably a square 400 meters on a side would not be too large." In semiarid climates, however, the riparian vegetation is subject to some oasis effects anyway, especially when winds blow normal rather than parallel to the stream. The actual size of the buffer zone becomes relatively insignificant compared with the requirements that the vegetation is of equal height and density and that the surrounding soil is kept as moist as the soil in the tanks. THE BUCKEYE TEST SITE LOCATION The test site is located in the southeast corner of sec- tion 11, R. 3 E. and T. 1 S. of the Gila and Salt river base line and meridian (33°21" N. and 112°31 W.), as in- dicated in figure 3, and its elevation is about 260 m (855 ft) above mean sea level. The area was inspected in the fall of 1958 by members of the U.S. Geological Survey and the U.S. Bureau of Reclamation. At that time, most of sections 11 and 12 as well as the land furthur up and downstream along the Gila River presented a nearly homogeneous stand of very dense saltcedar. For this reason, and because electric power lines were near, the site was considered ideal for the phreatophyte studies. The low-flow channel of the Gila River was remote enough to eliminate danger of flooding; but as the map (fig. 3) shows, the low-flow channel of the Waterman Wash curves very closely around the project site and minor flooding could be expected. TEST ENVIRONMENT CLIMATE In general, the climate at the project site is typical of that of the Sonoran Desert, but it differs in detail con- siderably from the average climate. As mentioned before, the area lies in the flood plain of the Gila River. Five km (kilometers) (3 mi (miles)) to the south are the Buckeye Hills which rise to 270 m (900 ft) above the valley floor; 16 km (10 mi) to the north are the White Tank Mountains, rising slowly at first and then steeply to 600 m (2,000 ft); and to the east the Sierra Estrella towers 1,100 m (3,600 ft) above the project site. Daytime temperatures can be very high, indeed, compared with those observed at standard weather installations outside the area. But, as a result of the surrounding mountains, there is considerable cold-air drainage on quiet nights, F4 R. 3 w. fl 112°31' J * LOWER RIVER ROAD | f I 11 P ARIZONA l 2000 4000 FEET 0 1 KILOMETER Salt R COLORADO RIVER Ficure 3.-Location of the Buckeye test site. and, even during the summer, nights are often cool. Freezing in the dawn hours of the early spring sometimes damaged young fronds, as shown in figure 4. However, it is not likely that the frost affected evapotranspiration because the plants quickly outgrew the damage. The dense vegetation and large irrigated areas north and south of the Gila River can create relatively high humidities. Table 1 presents monthly data on total precipitation, mean maximum and mean minimum temperatures and relative humidities, average wind speed, and average daily solar radiation. At first it was planned to install a U. S. Weather Bureau class A evaporation pan, but the dusty winds in the area together with dead leaves and other trash falling from surrounding saltcedars would make pan data practically worthless. FLOODS When the project site was chosen there was some concern about the frequency of flooding of the Waterman Wash. Based on information obtained here, it was decided that floodings were so rare that the risk of damage to the site would be small. However, during the STUDIES OF EVAPOTRANSPIRATION F ioure 4.-Frost damage on saltcedar twigs. second half of 1959 the project site was inundated nine times. When flood conditions such as shown in figure 5 again occurred repeatedly during 1960, it became necessary to construct levees capable of preventing small floods from upsetting the records. All floodings between July 1959 and September 1967 and the times that levees had to be strengthened or repaired are listed in table 2. On November 2, 1963, a severe storm with hail stones as large as 2.5 cm (centimeters) (about 1 in. (inch)) in diameter hit the project site. One pyrheliometer was smashed and anemometer cups were so severely dented that five sets had to be replaced. The Gila River flooded a few times, usually only mak- ing access roads north of the project site impassable. But, in December, 1965, the Salt River Project was forced to release water from behind Roosevelt and other dams, and the usually dry bed of the Gila became a "mile-wide" river; however, water reached just the northern row of evapotranspirometers, causing some gullying which was easily repaired. CICADAS During the latter half of May and the beginning of June each year, thousands of cicadas crawled out of the ground (fig. 6) and invaded the saltcedar stands. Damage was done by the females who laid eggs on the young branches after making an incision in the bark for each egg. This often resulted in girdling of the branches; the parts above the girdling died as shown in figure 7. However, regrowth from the plant underneath the WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD (iv [ + Ficure 5.-Two results of typical floods. Top: December 1959; tank 5 in foreground has not been planted yet. Bottom: September 1966, mud deposited on and in one of the recorders; high water mark about 95 cm (3.5 ft) above ground level. girdled areas was so vigorous that the cicada damage could not possibly have affected water use significantly. TANK CONSTRUCTION In May 1959 construction started on the first of six tanks, 9X9 m (30%X30 ft) in surface and about 4.25 m (14 ft) deep. To reduce cost and construction time, large sheets of plastic were used to line the evapotranspirometers, as suggested by Robinson and Bowser (1959). Robinson (1970) has described the con- struction and the plumbing in detail. It is necessary to mention here only that since the tanks were dug without shoring and the soil was fairly dry even at great depths it E5 F icure 6.-Cicada egress holes; pencil is 15 cm (6 in.) long. i\ $ i/ 4 \¥p4 Fioure 7.-Saltcedar twigs damaged by cicada egg laying. was not surprising that cave-ins occurred. As a result, the sides of the tanks were not straight walls, as suggested in figure 5 of Robinson's paper. The surface, too, deviated somewhat from the intended 9X9 m. After each tank was finished, it was planted to saltcedar. Vigorously growing bushes were selected from the surrounding stands and carefully dug up. Branches and roots were pruned to about 60 em (2 ft). Twenty-five crown cuttings were planted in each tank, and when the last one was finished in October 1959, the surroun- dings of the tanks were similarly planted. Figure 8 shows a newly planted tank. In 1962, five small tanks were constructed north of the existing ones as shown in figure 9. These new tanks were 6 meters square (20X20 ft) and only a little more than 2 m (6 ft) deep. They were lined with heavy butyl rubber instead of plastic. Tanks 7 and 8 were planted to saltcedar in February, 1963, but the others were kept STUDIES OF EVAPOTRANSPIRATION TABLE 1.-Meteorological data for project site near Buckeye, Ariz. Mean temperature Relative humidity Wind speed Solar Precipitation (°C) (°F) (percent) at 4 meters radiation Month (in.) (cm) (cm/sec) (mph) Langleys/day Max Min _ Max Min Max - Min 1961 Jan 0.43 1.1 21.2 4.7. 70.1. 40.5 98.5 38.7 121. 2.7 282 Feb ..::@:+, 0.00 _ 0.0 23.9 -0.9 75.1 30.3 82.7201 ay £34 421 Mar.; ....%. 0.24 - 0.6 21.5 - 22 81.5 35.9 68.5 21.0 130 _ 2.9 514 Apt ...>... 4 0.00 _ 0.0 31.9 (4.2 89.5 39.6 46.0 17.0 108 2.5 682 May 0.00 _ 0.0 38.6 9.3 101.5 48.7 59.5 19.5 80 . 1.8 717 0.00 _ 0.0 44.4 14.5 111.9 58.1 83.4 23.8 94 2.1 698 July ::...;.5 0.67 1.7 44.56 22.5 112.1 72.1 87.0 26.0 121 Hi 648 Aug .:.... 0.55 1.4 42:7 . 22.0 108.9 72.1 86.0 25.2 125 - 2.8 567 Sept 0.08 _ 0.2 39.2 15.2 102.5 59.4 84.5 24.4 M9 2.6 529 Oct .si... 0.20 - 0.5 83.1 - 6.9 91.5 44.5 74.0 22.0 130 - 3.1 429 Nov :«:y.¢.. 0.00 - 0.0 23.8 1.1 748 35.9 87.0 26.0 89 2.0 305 Dec: .:....." 1.65 4.2 18.8 -0.8 65.9 30.5 96.0 33.6 80. . 1.8 259 Total or mean 3.82 _ 9.7 32.5 8.4 90.4 47.1 79.4 25.0 109 _ 2.4 504 1962 Jan i. .a ss 1.61 -~ 4.1 19.7 --1.2 607.4 . 29.8 93.6 28.0 85 .: 1.9 307 Feb ;i.:.:.:. 0.15 1.9 29.2 $.1 78.1 Shobo 92.0 30.0 119 : 2.0 369 Mar :..... :; 0.63 _ 1.6 24.8 05 176.6 32.9 81.0 23.3 119 2.6 505 Api 0.00 0.0 34.17 6.4 94.4 43.6 78.0 22.6 1359 3.1 630 May: +s. 0.00 _ 0.0 36.0 8.3 96.8 47.0 60.4 21.6 94 2.1 696 Jung..;.:... 0.20 0.5 42.3 12.4 108.1 54.4 67.5 21.0 67. 1.5 654 July ::.: 0.24 - 0.6 45.1 19.1 115.1 66.3 51.5 18.0 134 - 3.0 679 Aug s... s +.. 0.08 _ 0.2 46.1 19.4 114.9 66.9 62.1 18.4 148 - 3.3 623 Sept .-..... 1.50 - 3.8 41.3 17.4 106.4 63.3 82.6 30.3 94 2.1 488 Oct 0.00 _ 0.0 85.9 175 95.0- 45.5 84.17 21.6 67: 1.5 450 Nov .;...s .. 0.16 0.4 28.4 3.3 83.1 38.0 83.17 25.6 16 - 1.17 321 Dec :.:;.... 0.35 0.9 226 0.3 72.7 826 87.0 28.7 aas e 268 Total or mean 5.52 14.0 83.3 8.0 91.9 46.5 77.0 24.1 104 - 2.3 499 1963 Jan 44.4 is: 0.10. 0.4 19.34 -8.7 66.17 25.3 85.2. 23.9 ts Nes! 314 Feb .:.:s :> 0.32 0.8 26.5 (3.4 82.0 98.2 82.1 22.6 mH fea 403 Mar.. ;. :/. 0.28. 0.7 27.4 1.6 81.3 34.8 77.17. 20.4 Pas "s 530 Apr r; 0.00 _ 0.0 30.8 3.8 87.5 38.8 83.4 16.4 rates 628 May %%... 0.00 _ 0.0 39.8 11.9 103.6 53.4 50.3 16.2 iss 680 June.. .;... 0.00 0.0 40.8 12.3 105.4 54.2 53.4 19.8 s i% 732 July :;;... 0.00 0.0 45.9 22.1 114.6 71.7 58.5 24.9 My rya 656 2.60 _ 6.6 41.9 22.0 107.4 71.6 81.3 25.4 ax R 558 Sept ...... 0.04 _ 0.1 41.7 18.4 107.0 65.2 76.9 22.3 124 2.8 582 Oct" siya y?. 0.91 28 36.6 11.6 97.9 52.8 83.2 25.9 105. 23 435 :s 0.81 -. 22 26.8 5.2 80.3 41.4 90.0 28.0 153 - 3.4 437 Dét Pixs X+ 22.8 -8.1 78.1 25.4 87.1 21.8 ake @ 319 Total or mean 5.18 13.1 33.5 8.1 92.3 47.1 74.1. 22.1 127 2.8 519 1964 0.28 0.7 19.4 -4.4 67.3 24.1 76.17 23.1 he?" Fisk 290 Feb ..:;.¥.. 0.08 _ 0.2 22.4 -4.3 72.3 24.3 65.4 19.1 Tix Faz 407 Mar., 0.81 . 2.2 24.4 1.5 76.0 34.7 79.4 19.6 x4 521 0.00 _ 0.0 82.0 5.9 89.6 42.7 66.5 21.0 'ey ¥2 598 May %.... 0.00 _ 0.0 36.8 9.6 98.2 49.2 58.4 17.3 152 3.4 683 June. ..,... 0.00 _ 0.0 41.3 14.6 106.4 58.3 58.8 21.4 120 : 2:7 685 July ..:..+., 087 . 22 44.5 226 112.1 12.5 63.0 22.6 1292.9 632 Aug :;... 2.60 - 6.6 41.2 21.6 106.1 70.8 84.0 26.7 129 239 517 Sept ; ..... 1.54 3.9 36.1 16.7 97.0 62.1 85.9 24.2 118 => 2.6 478 Oct :...... 0.35 0.9 83.6 10.6 92.5 51.1 83.0 22.2 80 1.8 409 Nov .. /: :y % 0.483 1.1 22.1 0.5 -7L8 - 32.9 90.3 26.1 vale ay 320 0.71 1.8 20.2 -0.7 68.4 30.7 85.9 30.3 sab Yass 245 Total or mean 7.73 19.6 81.2 . 78. 88.1 46.1 13.9. 22.8 121 2.7 482 WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD E7 Table 1.-Meteorological data for project site near Buckeye, Ariz-Continued Mean temperature Relative humidity Wind speed Solar Precipitation (°C) (°F) (percent) at 4 meters radiation Month (in.) (cm) (cm/sec) (mph) Langleys/day Max Min Max Min Max Min 1965 1.05 - 4.2 19.0 1.3 66.2 34.4 91.5 28.4 156 3.5 267 Feb ......;. 1.54 3.9 21.2 0.6 170.2: 81.0 88.6 23.4 163 3.6 367 Mar.:.;.:.: 0.79 - 2.0 23.1" 20 178.6 85.6 86.4 22.7 162 3.6 433 Apr .:....;. 1.54 3.9 28.6 6.1 83.5 42.9 85.1 19.6 190 - 4.2 518 May ...:;:: 0.43 1.1 34.1 7.6 98.4 45.6 11.0 18.1 220 4.9 592 Jung :...: ;. 0.08 _ 0.2 31.4 | 9.1 99.4 49.4 59.5 18.4 162 3.6 639 July 0.35 0.9 42.3 20.3 108.2 68.5 66.1 21.1 171 3.8 577 Aug :-\. .... 0.08 _ 0.2 41.7 18.8 107.1 65.8 70.4 20.7 181 4.0 582 Sept ...:... 0.35 0.9 40.9 12.6 105.7 54.6 78.8 21.5 149 3.3 465 Oct:.....¢s. 0.00 _ 0.0 35.2 - 1.1 95.8 44.7 71.8 22.0 150 _ 3.4 432 Nov..::.:.:. 0.67 1.7 26.6 3.8 79.8 38.8 85.2 29.2 113 2.5 260 Dec 3.78 9.6 19.9 2.90 67.8 87.2 96.5 46.7 117 2.6 194 Total or mean 10.83 28.6 30.8 7.6 87.5 45.7 79.1 . 24.3 161 3.6 444 1966 Jan;... .... 0.55 1.4 178 2.9, 641 27.8 97.5 31.6 e ow 253 Feb :.:... 1.54 8.9 18.3 -1.1 64.9 30.1 97.6 28.3 151 3.4 360 Mar..:7:%..; 0.16 0.4 27.1 3.9 80.8 39.0 92.7. 21.0 170 . 9.8 440 Apr ...:...., 0.00 _ 0.0 382.2 5.4 89.9 41.8 81.5 29.9 258 - 5.8 575 May ... /:. trace trace 38.1 10.8 100.6 51.4 76.8 24.4 172 - 8.8 649 June ...... 0.00 _ 0.0 41.6 14.3 106.8 57.7 71.7 - 28.9 1390 3.1 677 July. ;.; /..; 0.83 2.1 43.3 20.8 109.9 69.5 80.1 28.0 166 . 3.7 576 Mig 0.51 1.3 39.6 22.2 103.2 72.0 83.8 24.8 186 - 4.2 527 Sept 4.49 11.4 37.8 16.6 100.0 61.8 89.2 26.1 131 2.9 457 Oct...... 0.28 - 0.7 31.2 8.1 88.1 46.6 93.8 23.7 rigs Tuse 378 Nov ...:... 0.43 1.1 26.7 4.0 80.0 39.2 95.4 25.5 270 Dec::.::;.: 0.00 _ 0.0 21.2 < . 70.1 27.8 86.9 27.6 255 Total or mean 8.79 22.3 31.2 8.4 88.2 47.1 87.1 25.4 172¢ 8.8 451 1967 Jan wires.. 0.24 - 0.6 20.1 ~ 9.7 68.1. 25.8 79.6 23.2 281 6.3 285 Feb .::.... 0.00 _ 0.0 24.8 -2.0 76.6 28.4 63.6 20.7 168 - 3.8 487 Mat....::; 0.24 - 0.6 204" 2.8 83.1 81.1 74.4 24.5 164 3.7 512 0.20 - 0.5 21.4 :- 3.4 81.4 38.1 85.1 21.5 156 - 3.5 636 May. 0.00 _ 0.0 34.8 7.9 94.6 46.2 724%. 21.7 187 - 4.2 696 June. 0.16 0.4 39.6 14.2 108.3 57.5 74.1 24.4 o Tos July.. :. 2:0 0.24 _ 0.6 44.9 22.9 1129 73.3 89.5 27.5 Aug .... i.; 0.20 - 0.5 48.1 23.3 109.5 74.0 72.6 25.1 Total or mean 1.28 _ 3.2 329 8.0 912 47.5 76.4 23.6 191. 43 513 Taste 2.-Dates of flooding of the Waterman Wash bare. Tank 11, however, was surrounded by a triple hedge of saltcedar about 3 m (10 ft) wide. The purpose es Past T! | was to determine whether an "oasis-in-reverse" situation 1989 . . . July 18, 30; Aug. 2, 6, 11; Oct. 20; Dec. 18, 95, ...l... » | would have any effect on evaporation from bare soil. 1960 ... Mar. 1; July 30; Aug. 10, 22. Levees built in November ................ 4 19819 :~» duly 9,20; Aug.18, 20, 25; Septs 18 ... . . .. .. lili poto arine n rvs 6 1902 ... J:n).' 24; Feb.u2gl; Aug. 19; Set)? 22. Levees strengthened in:October ..... 4 INSTRUMENTATION 1963)... : Aug. (6, 22, 26, 80; Bopt. 18; Oct. 25;-Nov. 7. .. ).. ..o l insu i nuliey 7 1964 ... Jails. 23; Mar. 2; JulyeZlZ, 31; Aug. 4, 12, I??? 26; Sept. 13; 14; Oct. 17; Dec. 19. FLOATL'ESS CONTROL SYSTEM Levees repaired in November ............................ tf 12 The water level in the tanks was regulated by a 1965 . . . Feb. 7; Mar. 11; July 18, 29; Sept. 4, 18; Nov. 23; Dec. 23 .............. 8 A a k 1966 :. . May 8: July 24 Aug. 9, 11; Sept. I® s | floatless control system in which a valve automatically r ttt fon eate ccr an * __ | opened as soon as the water level fell below the lowest of Tots seis elses mae: tos aunt nae ta oa nen saben reas rae ag. s | two electrical contact points. Water then entered the R P f ; tank until the level reached the upper contact point, Breezfisczzdtvhmh the evapotranspirometer site was partly or totally inundated Wthh was about 5 mm (millimeters) (02 1n) above the "After July 1 only. . rif? o meee meretor foods chutes dan 2000 sss. 1<3wer point. A pen on an event recorder indicated the 'Before Sept. 5 only. time and duration that the valve was open, and the Ficure 8.-Man standing on surface of an evapotranspirometer newly planted to saltcedar; original vegetation in the background. o Bare soil t me Be er ee Ue meron ae seo e hind aet mild hie ane ae mie nee hae ie hen fe has ins nee Hee as ine ins mas are has sn nace aa al J o o o o 2 1 *p 0 50 100 FEET s 0 10 20 30 METERS 4 e 3 R 6 o 5 ®w Figure 9.-Location of tanks and instruments. Tanks 1-6 are inside a dense stand of saltcedar; tanks 7-11 in an open area, 7 and 8 planted to saltcedar; tanks 9,10 and 11 are bare but 11 is surrounded by a hedge of saltcedar. P, R, and W, solid dots, are instrument masts. Circles indicate locations of access tubes for determining soil moisture outside the tanks. Tanks 1, 2, 3, 4, and 6 were constructed in May and June of 1959, tank 5 in October, 1959, and tanks 7-11 in November and December of 1962. quantity of water (to the nearest tenth of a gallon) that entered the tank could be read from a standard water meter. Routinely, the meters were read every morning from Monday through Friday, and frequently, on Satur- days and Sundays. Occasionally (quite often in 1966 and 1967) readings were taken at 2-hour intervals for periods of 72 hours or more. Figure 10 shows the instrumentation at one of the tanks. One pen on the recorder, a, indicates the temperature of the water; the other pen marks the time that the magnetic valve b is open and water via STUDIES OF EVAPOTRANSPIRATION F iGurE10.-Instrumentation for the water budget: a, on/off and water temperature recorder; b, magnetic valve; c, water meter; d, valve-control mechanism. meter C enters the tank. The switching mechanism, d, opens and closes valve b. Figure 11 is a schematic cross section of a tank, show- ing contact points of the electrode rods at the water table and other features. ADVERSE CONDITIONS It is not surprising that the rigors of the climate often adversely affected the contol system. Temperatures as high as 50°C (122°F) combined with high humidity were not infrequent. Dust storms interfered with the electrical Recorder T an Stand Electrodes for ro Well = pipe water-level point control Water level Rods w L Water table E Float tu TContact S s ! | points 14 . H x a _ : #. | H Id | % $= q H td | H | ti 4 Ld U Perforated pipe _____________________________ q 9 METERS Ficure 11.-Schematic cross section of an evapotranspirometer. Access tubes for soil moisture determination, tensionmeters and thermocouples not shown. (Instruments and plumbing not to scale.) WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD contacts. During the colder seasons, frost damage oc- curred in the plumbing, even though heat bands were in- stalled to protect the pipes from bursting. The well water often contained a large amount of sand, and, owing to the comparatively high content of dissolved solids, the metal pipes and aluminum tubes and rods corroded easily. As a result, valves stuck, water meters did not turn, and contacts failed. Various quantities of water which entered the tanks were sometimes unaccounted for, but the amount of water was always in excess of that used during such periods through evapotranspiration alone. The method which was devised to correct for these un- desirable effects is discussed in the following section. WATER LEVEL In order to have a certain amount of control on malfunctioning which might escape attention, 1%-in. (about 4-cm) well points were installed about 3 m (9 ft) from the stand pipe (fig. 11). Depth to water in the well points was measured whenever the water meters were read. Corrections for apparent excessive use could reasonably be made by plotting the actual water-level fluctuations and assuming a 40-percent porosity of the subsoil in the tanks. Effects of unintentional flooding were also corrected sometimes by pumping water out of the tanks until the intended depth to water table was reached again. Similar corrective measures were taken each time one or more of the evapotranspirometers was inundated by flooding of the Waterman Wash. In 1964 flooding oc- curred so frequently (see table 2) that only water-use data for short periods of controlled conditions could be considered reliable. SOIL MOISTURE AND TEMPERATURE The tanks were equipped with aluminum access tubes, each with 2-in. (about 5-em) outside diameters and 1% - in. (about 4.5-cm) inside diameters, for measuring soil moisture by the neutron scattering method. (These tubes are not shown in fig. 11.) Most of the tanks had 8-inch (about 20-ecm) plastic pipes which extended down to the water table. At times, water-level recorders were placed on these pipes. Batteries and tensiometers were installed in a few tanks in 1963 and in all tanks in 1965. The batteries made it possible to estimate moisture stress in the soil at different levels and also to estimate directions of soil- moisture flow towards or away from the land surface. Tanks 1 and 6 were equipped with heat-flow plates to determine the incoming and outgoing sensible heat; also, several sets of thermocouples were installed inside and outside these and other tanks to measure soil temperature at two or more levels. Finally, most of the tanks were equipped with a small rain gage to measure throughfall. EQ VEGETATION TANK PLANTING Usually, phreatophytic vegetation along streambeds or around reservoirs and lakes is mapped by a set of stan- dard survey procedures (Horton and others, 1964). These methods are very useful and effective in providing data which can be subjected to statistical and other analyses; they are unsatisfactory, however, for a small area such as an evapotranspirometer. The standard methods are not sensitive enough to elicit the possibly small differences that may exist between two tanks. In this experiment, the size of the tanks allowed detailed observation of the vegetation to be made conveniently by the procedures described below. The surface of each tank was divided into twenty-five squares. Each square supposedly contained: one originally planted crown cutting, growth resulting from the sprouting of pieces of stem and root (buried during the construction of the tanks), and growth from seeds that were blown or washed in after construction. Growths of the latter two types sometimes reached the top of the vegetation, especially in the few areas where the crown cuttings did not succeed in dominating the surroundings. The leaf area completely shading the ground was estimated by percentage of each square, and the area of the canopy for each bush or clump was similarly estimated and then converted into square meters. TRANSPIRATION VOLUME It has often been assumed that consumptive water use by phreatophytes is more or less related to the volume of transpiring foliage. A method was therefore sought that would give a measure of volume of transpiring foliage. The total volume of the vegetation taken as the product of the area of the canopy times the average height, sometimes corrected for crown depth, has been used as a parameter (Gatewood and others, 1950). This method has, at least for saltcedar, a decided dis- advantage. The plants are considered to be a set of cylinders, and as a result, the actual transpiring volume is greatly over-estimated. Nonetheless, such data can be used to compare one area with another. A more natural geometric configuration is obtained, however, if one assumes the transpiring volume of the bush to be the up- per half of an oblate spheroidal shell. In the tanks at the Buckeye Project, such a shell is about 50 ecm (about 20 in.) thick. The horizontal radius is that of the mean radius of the more-or-less circular area shading the ground, and the vertical radius is about one-third of the total height of clump or bush. The volume of half an oblate spheroid is (27/3) a*b in which a is the longer, in this case horizontal, radius, and b is the vertical or shorter radius. The volume of the shell is then (27/3) [(a-50)%(b-50)], and represents a transpiring volume E10 based on a more realistic shape than that of cylinders. Volumes for 1965 computed by the method of Gatewood and others (1950) were between 150 and 250 percent larger than those obtained by the spheroidal shell method. The average height, the mean coverage in percent of the total area, and the total transpiring volume of the canopy are shown in table 3 for each tank surveyed in the fall of the years 1962-66. As can be seen, equilibrium seemed to have been reached by 1964. It appears that although there were differences in rate of growth (to be discussed later), these differences were offset in each tank by the dying of parts that were no longer exposed to sufficient sunlight. In addition to being a phreatophyte, saltcedar is obviously a heliophyte. TaBis 3.-Results of surveys of saltcedar, taken in the fall of each year, in the evapotranspirometers Mean Mean Total Date of height cover spheroidal survey (m)! density volume (percent) (m) Tank 1 80 74.0 90 90.8 100 98.7 100 100.8 99 99.5 80 81.2 80 88.3 90 96.4 100 98.7 95 96.9 80 78.4 80 92.4 100 110.3 95 108.2 100 113.1 75 75.4 90 94.3 105 104.8 100 110.5 105 112.8 70 82.3 80 90.2 100 107.6 100 108.0 100 111.0 T5 79.7 T5 93.8 105 101.2 95 105.2 100 100.5 'To nearest 5 cm. STUDIES OF EVAPOTRANSPIRATION Finally, table 3 shows that there are considerable differences between the observed variables in the tanks (actually more than 10 percent), but they are not statistically significant. No attempt, therefore, has been made to relate water use to these variables. Moreover, the relationship between transpiring volume and water use, for example, is not so simple as is often assumed. This will be discussed under "Vegetation Growth and Development". WATER USE, LARGE TANKS 1961-63: DEPTH TO WATER During 1961, 1962, and 1963, the water levels in tanks 3 and 5 were maintained at 1.5 m (5 ft) below the land surface; in tanks 4 and 1, at 2.1 m (7 ft); and in tanks 2 and 6, at 2.7 m (9 ft). The water use by months is given in table 4. These 3 years are tabulated separately because the treatment did not change during this time; after 1964, some of the tanks were flushed to reduce the salinity of the ground water, and in 1966 other changes were made as explained below. Table 5 summarizes the total use per tank per year for 1961, 1962, and 1963 and includes the results of an analysis of variance (Fisher, 1944; Fisher and Yates, 1943). (See "Appendix: Analysis of Variance" for ex- planation of least significant difference.) The analysis of variance, as can be seen from the table, shows a signifi- cant effect of interaction between years and depth to water. The table gives an F-value which is significant at the 1-percent level for effects of depth to water. Although some tanks show an increase in water use between 1961 and 1963, such increases are not statistically significant and are completely overshadowed by the depth-to-water effect. Data from tanks 2 and 6 show a gradual decrease in water use probably due to an increase in salinity of the ground water. Unfortunately only a few measurements of specific conductance were made prior to 1963, but it may be assumed that the conductance in all the tanks was the same. Table 6 shows that the conductance of the ground water in tanks 2 and 6 had increased much more than in the other tanks. The low water use in tanks 2 and 6 may have been caused by high salinity of the ground water rather than depth to ground water. The difference, however, in water use between tanks 3 and 5, with a water table depth at 1.5 m (5 ft), and 1 and 4, with a water table depth at 2.1 m (7 ft), must have been caused by the deeper water table in tanks 1 and 4 because the specific conductance of the ground water in all four was practically the same. Further evidence of the effect of salinity of soil moisture will be discussed in ©1965: Salinity." WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD TaBLe 4.-Monthly water use in evapotran. E11 spirometers by saltcedar, excluding rainfall Tank No. . .. 3 5 4 1 2 6 Depth to water (cm). . . 150 150 210 210 270 270 Date Cm In. Cm In. Cm In. Cm In. Cm In. Cm In. 1961 f Jan :...... 2.0 . 0.79 2.1 0.83 1.5 0.59 1.2 0.47 0.4 0.16 0.4 0.16 Feb.. ...... 2.6-,.. 1.142 4.1. . 36 142 18. 0.71 8. 07 £8: 071 Mar.:.:...:.. $6 . 2.20 6.0 - 2.36 as : Ao £0... 1.97 $8 - , 1.50 14 0.55 Apt 16.8 6.61 16.4 6.46 74 3.03 8.2 3.22 6.3 2.48 7.6 2.99 May ../... 20.0 10:23 gas : 8.08 14.9 . 5.86 al6 8.50 162 6.38 15.8 - 622 June ::..... 384 13.15 27.3 10.75 21.6 8.50 27.1 10.67 18.1 7.12 21.2 8.35 July ...... 37.3 14.68 39.5 15.55 27.8 10.94 30.3 11.93 27.2 10.71 21.2 - 10.71 Aug .. :.. ¢.. 30.6 12.05 33.9 13.35 21.1 8.31 17.8 . 1.01 19.3 7.60 21.3 8.38 Sept ...;.... 22.5 8.86 21.5 8.46 18.5 7.28 16.8 _ 6.61 8.7 3.42 6.8 2.68 Oct ..:;..:. 124 _ 4.88 10.2. : 4:01 123. | 4.94 11.0 : 4:83 29. 114 49 _ .T. Nov ..::... 5.8 2.28 9.4 3.10 7.9 3.11 3.7 1.46 0.0 0.00 0.0 0.00 2.9 1.14 6.3 2.48 1.4 0.55 0.9 0.35 0.0 _ 0.00 0.0 0.00 Total 198g. 7s.990 _ 100.5) Teddi c 4411. 5553 - 1454 57.98 104.7 41.92 - 108.0 -d2.62 1962 Jan: 1.7 0.67 2.9 1.14 0.8 0.31 0.0 _ 0.00 0.0 _ 0.00 0.0 0.00 Feb:. ~.... 26 tos 261.02 1+ 086 11 _ 0.43 11 043 11 0.86 Mar ..;...:; 5.4 2.12 4.1 1.61 2.5 0.98 2.5 0.98 3.4 1.34 3.4 1.34 Apr ... «:... 23.3 9.17 22.0 8.66 11.2 4.41 12.0 4.72 9.0 : 3.54 9.0 3.54 May :.!.... 40.0 15.75 40.4 15.90 25.9 10.20 29.5 11.57 19.1 7.52 15.2 5.98 June ...... 42.3 16.65 45.0 - 17.71 38.1 13.03 34.2 13.46 20.5 8.07 21.7 8.54 July . .. . .. 24.3 9.57 25.6 10.08 16.7 6.57 17.2 6.77 10.06 4.17 10.1 _ 3.98 Aug ::.... 22.0 8.66 22.4 8.82 12.7 5.00 15.3 6.02 7.8 3.07 8.2 3.23 Sept ::...... 20.6 -~ 811 20.7. - 8.16 11.7 _. 4.61 12.7 - £.00 79. _ 311 88. _ g.46 Oct ::.... .: 16.0 . 6:30 179. - 7:05 10.52 A18 11.4 _ 4.49 7A. . 2.91 g1 ~- $19 Nov ...:... 14.3 5.63 14.2 5.59 7.3 2.87 9.9 3.90 4.7 1.85 5.8 2.28 Dec :?... 5.8 2.28 3.9 1.53 3.5 1.38 4.5 1.71 2.3 0.90 2.8 1.10 Total 218.3 85.93 221.7 87.26 137.0 53.92 150.3 59.11 93.8 36.91 94.2 37.07 1963 Jan ......:. 0.8. :0.91 3.7 1.46 1.0 0.39 1.0 : 0.39 0.0 _ 0.00 0.8 0.31 Feb .... ga. of'ass 4.2. 1.65 1.4 _ 0.55 12 - G47 - 0.35 18. 071 5.0 1.97 5.7 2.24 2.1 0.82 2.1 1.06 2.6 1.02 2.7 1.06 ApI =;. .: ../. 16.4 6.46 16.8 6.61 7.9 3.11 9.8 3.86 5.3 2.09 6.1 2.40 May :...... 37.3 14.68 36.8 14.49 24.9 9.80 274 10.79 13.5 5.31 14.8 5.83 June :...... 42.5 16.73 17.01 86.2 - 14.25 85.8 14.09 16.9. 6.05 16.9 . 6.65 July :.;::;..: 41.8 16.46 36.5 14.37 28.2 . 11.10 92.8. 12.71 13.6 _ 5.35 12.7 5.00 Aug ..:..... 24.4 9.61 23.9 9.41 16.4 6.46 16.3 6.42 8.6 3.38 7.6 2.99 Sept: -....... 27.3 10.75 25.3 9.96 17.9 7.05 17.9 7.05 9.5 3.74 10.3 4.05 Oct :...... 191. -- 7.52 20.8 _ 8.19 145 - 5.71 115. _ 4.58 g 1 - g.88 10.8 _ 4.25 Nov 7.5. 9:05 10.6 - 417 61 - 2.40 £0 .. 1.97 41 . 'Ab1 py- 916 $o - 0.35 1.2. 047 29 TPH 2.4 ~ 0:94 24-094 24 .~ 0.94 Total 226.5 89.17 228.17 90.03 159.5 - 62.78 163.3 - 64.28 86.5 34.02 92.4 36.35 1964: FLOODS This treatment had two secondary effects which were Obviously something had to be done to improve the quality of the ground water, and in January and February of 1964 all tanks were flushed. Therefore, the water-level controls were disconnected and water was forced through the stand pipe and the laterals at the bot- tom of the tank, thus driving water from the bottom to the top of the tank. Specific conductance (mmho cm ~ at 25°C) of the effluent was measured daily, and backwashing was continued for 5 to 10 days until the conductance was about equal to that of the well water. foreseen but about which little could be done. First, the water content of the soil above the ground-water level was greatly increased, and it was expected that it would take considerable time before the water content in the soil would return to the 1963 levels, even though excess water was pumped out. The second effect, a result of the frequent flooding prior to the building of the levees, was that additional soil had been dumped on the evapotranspirometer site, and in many places the soil was higher than the plastic lining. As long as the top E12 TaBLE 5.-Yearly water use in evapotranspirometers, exluding rainfall Depth to Mean ground water. .m. . 1.5 2.1 2.1 of ATT 5 7 9 all Tank No ..... 3 5 4 1 2 6 tacks 1961; Cms... 198.9 199.5 141.1 145.4 104.7 108.0 149.6 In isis c..? 78.29 78.54 55.53 57.23 41.22 42.52 58.89 1962; Cm s .me. 218.3 221.7 137.0 150.3 93.8 94.2 152.6 85.93 87.26 53.92 59.11 36.91 37.07 60.03 1963: Om ...s; 226.5 228.1 159.5 163.3 86.5 92.4 159.5 Ina var. 89.17 90.03 62.78 64.28 34.02 36.35 62.77 Mean of 3 years Om 213.6 216.6 145.9 153.0 95.0 98.2 153.9 Ansa iand 83.36 85.28 57.41 60.21 37.40 38.65 60.38 Analysis of variance Source of Sum of Degrees of Mean F3 variance squares freedom square Years .......1. 308.91 2 154.5 9.6 Depth to water . 42,660.78 2 21,330.4 **1,324.9 Interaction ..... 1,282.81 4 320.7 *19.9 Error 144.79 H 16.1 Total >;. 44,397.29 TiN sid un an rara as 'F values: significant differences indicated by * at 5-percent level, ** at l-percent level. Least significant difference at 1-percent level is 25.0 cm. TaBus 6.-Quality of ground water in evapotranspirometers [In 1960 only three tanks were sampled; in 1963 all tanks were sampled but in three tanks samples were taken from two locations: a) where water enters the tanks, and b) where the ground water is more or less stagnant and less diluted with incoming water. Agency making analysis: U.S. Geological Survey Quality of Water Laboratory, Albuquerque, N. Mex.] Specific conductance Sodium plus potassium Chlorine (mmho em*' at 25°C) (mg/l) (mg/l) Tank N° _ May 11, - Mey 61963 _ May 11, May 6 1963_ - May 11, May 6, 1963 1960 a b 1960 a b 1960 a b deve ove m vac ts T - Opera en oan 11900. ia. rss Ch taas ce 2,140 Rays. +. 5.9 74 12.6 730 1,120 1,820 1,580 1,950 3,980 aisi reunions (o) e nb 1180-5 NIA ua 1,990 C ARO TLT. ols deren bae ale (id. Sol reat oue nate aa 11200. 07.100, 2,110 6.5 7.4 7.0 750 1,130 1,190 1,710 1,960 1,960 of aver 6.2 TA 31.0 740 1,150 5,180 1,640 2,040 10,800 layers of the soil were dry, this did not matter. However, when tops of the tanks and also the surroundings were saturated, root growth from trees inside the tanks could be expected to reach over the tank boundaries. Also, roots from surrounding trees might penetrate the tanks. In March 1964, trenches were dug to the plastic lining wherever necessary to prevent the roots from overreaching. As expected, water use increased enormously in 1964, not only due to the lowering of the salinity but also because of the high evaporation rate from the soil sur- faces of the saturated tanks. In addition, the area suf- fered three floods (see table 2) which not only partially refilled the trenches, but also kept the topsoils of the STUDIES OF EVAPOTRANSPIRATION tanks saturated with large quantities of freshwater. The water-use data for short periods could be used for com- parison between tanks and for correlation studies with climatological phenomena. However, the monthly and yearly totals were so heavily influenced by these catastrophes that any attempt to compensate for them was thought to be futile and would result in completely unreliable data. For these reasons, data from 1964 are not tabulated in this report. 1965: SALINITY After a sturdy dike was built in November 1964 to keep the Waterman Wash floods out of the study area, attention was once more focused on the water use by saltcedar. As was pointed out, by the end of 1963 it became apparent that the decrease in water use in tanks 2 and 6 might have been due to deterioration of the ground-water quality in those tanks. The vigorous growth and the tremendous increase in water use follow- ing the flushing in 1964 provided more evidence. To make comparisons, one of each of the pairs of tanks with equal depth to water was flushed and the other was not. In 1964, when the flushing was finished, the excess water was simply pumped out until the water level had reached the original depth; whereas in 1965 the water table was lowered as far as the pump could draw it down. The soil moisture was then allowed to drain and the tank was again pumped out. This was repeated until no more water collected in the stand pipe. This procedure reduced the soil moisture above the capillary fringe, and the water content became more comparable with that in the untreated tanks. After this drying process, ground water was allowed to rise to its intended level. Effectiveness of this treatment is shown in figure 12, where the percentages of soil moisture are plotted against depth for each of the six tanks. While the higher moisture content resulting from a higher water table is quite evident, there is no significant difference of moisture content between flushed and unflushed tanks. It seems, therefore, reasonable to assume that the differences in water use so clearly shown in table 7 are due to differences in salinity and depth to ground water and do not result from differences in soil moisture above the capillary fringe. Figure 13 summarizes the data from table 7 (third section). It is obvious that the differences between flushed and unflushed tanks significantly overshadow the influence of depth to ground water on the rate of water use. Another study provided corroborative material. As mentioned previously, two of the small tanks were planted to saltcedar in 1963. About the same time, a cot- ton farmer to the south of the project area drilled a new well in the vicinity of the Waterman Wash. Results of the chemical analysis of samples from this well water and of the project water are presented in table 8. Clearly, WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD i T T T T T T T T T Water levels below land surfaces 3\\\\ 3 and 5: 1.5 m (5 ft) 0.5 - 1 and 4: 2.1 m (7 ft) 3 2 and 6: 2.7 m (9 ft) 0 T N 0 I Numbers are tank numbers \\ DEPTH BELOW SURFACE, IN METERS o T 2, 3, 4 flushed to reduce the s 2.5 salinity of the ground p # 4 water 1, 5, 6 not treated 3.0 F 1 i 1 i | i | i 10 20 30 40 50 MOISTURE, IN PERCENT BY VOLUME Ficure 12.-Soil moisture by volume in the six large tanks on a typical summer day in 1965. 2 ado TANK _ NUMBER Fad w 2, 6 = EXPLANATION i- Least E E significant 0 w difference z » 200 Flushed . 5 Jan.,1965 t + Not 3 100 flushed CC Means r [_| '_ s 9 1.5 21 Means in €/m -~ mn DEPTH TO GROUND WATER, IN METERS Ficure 13.-Total water use during 1965 in six evapotranspirometers (tanks) showing the effect of depth to water and flushing of saline ground water. The bar denotes the least significant difference at the 5 percent level. Tanks 3, 4, and 2 were flushed in January 1965. The three columns to the right show the mean water use as a function of depth to ground water. water from the farmer's well is of far better quality than the water obtained at the project site. A (3,750-1 (liter)) (1,000-gal (gallon)) container was installed near tank 7 so that water from the new well could be stored and delivered either to tank 7 or to tank 8 or both. From June 1963 through March 1964, tank 7 was provided with water from the new well while tank 8 was fed with project water. Figure 14 shows that during this time, the saltcedar grew much more vigorously in tank 7 than in tank 8. It is not surprising that water use was also higher in tank 7. During the period mentioned, tank 7 used 264 cm (104 in.) in contrast to 170 ecm (67 in.) for tank 8. It should be E13 Ficure 14. -Evapotranspirometer 7, top, received better quality water between June 1963 and February 1964 (when photograph was taken). The plants in tank 8, bottom, grew on project water. noted that the vegetation in these two tanks stood isolated, and actual water use obviously cannot be com- pared with that of the other tanks. Nonetheless, because such isolated clumps do occur in nature, some further results are discussed under "Water Use, Small Tanks." 1966: DENSITY OF STAND Gatewood and others (1950) developed a method to compute estimated water use based on the assumption that water loss is directly proportional to the volume of green vegetative material growing on the area. The method implies that the denser the vegetation, the more voluminous is the water loss. Although objections E14 STUDIES OF EVAPOTRANSPIRATION TABLE 7.-Water use, excluding rainfall, in evapotranspirometers during 1965 Tank 3 5 4 1 2 6 Depth to water... 1.5 1.5 2.1 2.1 2.1 2.7 fies 5 5 7 7 9 9 Flushed; 45.7.3 Yes No Yes No Yes No Month Cm In. Cm In. Cm In. Cm In. Cm In. Cm In. Jan'....,...... 0 0 0 0 0 0 0 0 0 0 0 0 Feb 1.0 .39 A .16 A 16 0 0 0 0 0 0 Mar.. :sa1..... 9.6 3.178 4.4 1.73 4.1 1.61 2.7 1.06 0 0 2.7 1.06 27.1 10.67 16.8 6.61 21.3 8.39 6.9 2.172 18.9 7 44 3.7 1.46 May ......:... 52.1 20.51 32.0 12.60 48.0 18.90 27.3 10.75 43.0 16.93 16.4 6.46 63.6 25.04 42.9 16.89 60.7 23.90 36.8 14.49 52.6 20.71 21.2 8.35 duly .. 45.8 18.03 33.0 12.99 37.1 14.61 28.6 11.26 34.1 13.43 16.2 6.38 Mig 37.5 14.176 30.6 12.05 29.9 11.77 20.2 7.95 29.2 11.50 14.4 5.67 Sept ......:..: 30.5 12.01 26.8 10.55 24.9 9.80 18.6 7.32 24.2 9.53 12.4 4.88 Oct :... .s" 27. 4 10.79 25.1 9.88 21.7 8.54 17.7 6.97 24.4 9.61 10.6 4.17 Nov.. .~ 15.0 5.91 12.6 4.96 10.4 4.09 7.8 3.07 9.7 3.82 5.8 2.28 Dec 1.3 51 1.1 43 1.4 55 .9 .35 1.7 .67 2.1 .83 Total.... 310.8 122.36 225.8 88.90 259.8 102.28 167.5 65.94 238.0 93.170 105.6 41.57 Mean of tanks Nos .......... 3 and 5 4 and 1 2 and 6 3, 4, and 2 5, 1, and 6 Depth to water (m) .... ...... 1.5 2.1 2.1 Flushed Not flushed Month Cm In. Cm In. Cm In. Cm In. Cm In. Jan 0 0 0 0 0 0 0 0 0 0 27. .28 D .08 ( 0 .5 .20 2 .08 7.0 2.16 3.4 1.34 1.4 55 4.6 1.81 3.3 1.30 Apr :. : 2s 22.0 8.66 14.1 5.55 11.3 4.45 22.4 8.82 9.1 3.58 May .>. ..+...: 42.0 16.54 37.6 14.80 29.7 11.69 47.7 18.78 25.2 9.92 June :;:.;...... 58.3 20.98 48.7 19.17 36.9 14.53 59.0 23.23 33.6 13.23 Tully .:? -s :. 39.4 15.51 32.9 12.96 25.1 9.88 39.0 15.35 25.9 10.20 Aug :% :i. rae. 34.0 13.39 25.1 9.88 21.8 8.58 32.2 12.68 21.8 8.58 Sept .;..::.... 28.7 11.30 21.7 8.54 18.3 7.20 26.5 10.43 19.3 7.60 Oct 26.2 10.31 19.7 7.16 17.5 6.89 24.5 9.65 17.8 7.01 Nov : 13.8 5.43 9.1 3.58 7.7 3.03 11.7 4.61 8.7 3.43 Dec .:.:.....;: 1.2 AT 1.1 43 1.9 15 1.4 55 1.4 55 Total ..., 268.3 105.63 213.7 84.13 171.8 67.64 269.6 106.14 166.3 65.47 Depth to water (m) 1.5 2.1 2.7 Mean Analysis of variance! 5 Source of Sum of Degrees of - Mean °F Flushing Cm In. Cm In. Cm In. Cm In. variance squares freedom _ square Flushed ...... 310.8 122.36 259.8 102.28 238.0 93.70 269.6 106.11 Depth to water . 9,366.86 2 4,683.43 14.38 Not flushed ... 225.8 88.90 167.5 65.94 105.6 41.57 166.3 65.47 Flushed vs. not flushed ... 15,985.68 1 15,985.68 _ *49.08 Méan .. S6ss i06ea 15" adi iis or ......" .. .:.. Error a o (and lll i [2 . Total .... 26,003.89 boc suis s ra ice 1a e 'The analysis of variance indicates that no interaction between depth to water and flushing could be taken into account because of lack of replication °F value: *indicates significant difference at the 5-percent level. against this method have been cited (Coleman, 1953), Horton, Robinson, and McDonald (1964) stated, "No better method is yet known." The Buckeye Project provided opportunity to shed some qualitative light on the relationship between densi- ty of stand and water use. It is, after all, reasonable to assume that in fairly open stands wind can penetrate deeper into the vegetation, and, at least on favorable oc- casions, the wind may take away moisture, thus allowing "space'" for additional evaporation and transpiration. in 1966. For this reason the following experiment was conducted After tanks 2, 3, and 4 had again been flushed, the water level in all tanks was brought to 2.1 m (7 ft) and in WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD TaBLE 8.-Analysis of project water and better-quality water from a new well south of the project site. [Agency making analysis: U.S. Geological Survey Quality of Water Laboratory, Albuquerque, N. Mex., May 1963. Data in mg/l except when otherwise designated] Better quality water Analysis factors Project water Residueat180°C .............. 4,760 2,280 Hardness as CaCO; ............ 1,490 650 Percent Na ..;.:..........;... 63 63 Sodium adsorption ratio ...... .. 13 8.8 PHE: i iaa nt nine highs res 6.8 7.3 Specific conductance (mmho em' at 25° C) ........ 7.57 3.64 S10; FLCL. in sais + anise a's hire a 31 32 C 396 182 122 48 1,150 515 405 234 836 256 1,980 915 1.3 2.5 1.0 17 March of 1966 the vegetation in tanks 1 and 4 was thinned out to approximately 50 percent of the original density. This was accomplished by simply cutting half of the branches in each clump, taking the diameter of the branches into consideration. All vegetation in tanks 3 and 5 was shorn off at knee height. Figure 15 illustrates the general appearance of three of the six tanks, and figure 16 shows the combined results of the two treatments. As expected, the tanks that had been flushed (2, 3, and 4) used more water than those that were not treated. The water use in the shorn tanks was much less than in the vegetated tanks, but within two months, use in- creased sharply with regrowth of the vegetation. The shorn stumps, still having their root systems, sprouted very fast, especially in tank 3. In mid-July tanks 3 and 5 were again shorn and water use dropped for the second time. ---I T == --- -- T _2] 250 |- ///’ Z we € ya =A w 418) .- i- 25.0% w 200}- . A19) = 3 f 2 Z. 5 Vegetation in tanks at a o 3 and 5 cut during /// z 150|- 7. R £ March and July f. | i F ;?'/ § s // ~2(30) o 7 7 > eras. 2 so- e pes = * 2 3 # a L // /// 1(30) aA Fas e 2 //// P4 2 agl a> 8128) __" m o /// < [ 7 #4 -z - >B(38) < ty y 3 C /’/ /// [ g // /// 3 i es J’,//// 8 --f" | | | < APRIL may JUNE JULY August 1967 Ficure 17.-Accumulated water use in tanks 1 to 6 between April 1 and September 1, 1967. Dashed curves represent tanks flushed in Jan. 1966; solid curves represent tanks not flushed. See figure 16 for tank treatments. Tank numbers indicated on curves; numbers in parentheses indicate conductance of soil moisture in August 1967 (mmho em~' at 25°C). observed in August 1967 are shown in parentheses. Com- parison with the data of figure 16 shows that salinity in- creased very fast, especially in the tanks that used the greatest amount of water in 1966 and 1967. The rapid in- crease of water use in tanks 3 and 5 (shorn twice in 1966) is also worthy of notice. The last flood of the project history occurred on September 4, which made it virtually impossible to work in the project area during the rest of the month. In Oc- tober dismantling of the equipment began, and in November some of the tanks were partially excavated to determine the condition of the linings. It appeared that the plastic and the rubber had endured well, except that occasionally small holes made by gophers or mice were encountered in the top 30 cm (12 in.). However, since these holes were well above the water table and in the dry region of the soil blocks, effects on water use were suspected to be negligible. Below the level of the water table a fine network of roots clung to the outside of the plastic lining. The in- sides of the tanks below and slightly above the water table were considerably cooler than the dry soil at the same level outside the plastic. Even though the soil out- side the tank was very dry at these depths there was enough vapor to condense on the lining. Thus, roots com- ing in contact with this water developed into the fine network. SOIL MOISTURE AND SOIL-MOISTURE FLUCTUATIONS METHOD OF MEASUREMENT In reports on water use by vegetation in lysimeters or evapotranspirometers, the increase or decrease of water WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD during a period is often taken into account (for instance, Gatewood and others, 1950). If from one month to the next, there is a rise in the water table, the increase in content is subtracted from the amount of water measured in the tank, and the actual water use by the vegetation is obtained. The water levels in the project tanks varied little, usually less than 1.5 ecm (0.05 ft) from one month to the next, and there was not a discernible seasonal fluctuation. This is not surprising in view of the mechanism by which the water levels were controlled. (The daily fluctuations, to be discussed later, have no measurable effect on the monthly water use.) It was surprising that at times soil moisture contents above the water level varied considerably. Fairly large differences in water content from month to month oc- curred or, as explained below, seemed to have occurred in some of the tanks. It appears, however, that these changes may have been due to causes other than water use or lack of use by the vegetation, since the increases and decreases were not consistent. Table 11 shows ex- amples of changes in soil moisture content expressed as centimeter depth of water and as the percentages of measured water use during the indicated period. For in- stance, from May 16 to June 24, 1963, there is an increase in water content of 4.5 percent in tank 4, but in tanks 1, 2, and 6, a decrease is measured. From May 3 to June 13, 1966, there is a decrease in tanks 2 and 3 whereas there is an increase in tanks 1 and 4. TaBi® 11.-Change in soil moisture content [A=change expressed as centimeter depth of water, and percent= change expressed as percent- age of measured use of water during indicated periods; minus symbol indicates a decrease} Tank No. ... 1 2 3 4 5 6 A Percent A Percent A Percent A Percent A Percent A Percent 1963 May 16-June 24. ...... -1.0 22 -28 185 No data. 1.9 4.5 No data .-0.8 3.8 1965 July 21-Aug20 ...... L4 "43.1.4 04 00. 20 04 L1 _ 26° 70 1966 May 3-June 13... ... 25 9.1 -49 54 -22 7.3 0.7 L1 $.0550.0 14.0 48.0 'This increase resulted when a sprinkler hose burst between tanks 5 and 6 in early June. These soil-moisture measurements were made by the neutron scattering method with instruments calibrated at the project (Task Force, 1964). Figure 18 shows the in- strument calibration data. The correlation coefficient is, as the figure shows, highly significant, yet a considerable scattering is possible. A good example of what can happen is given in figure 19 where a few soil-moisture measurements by the neutron scattering method are plotted against depth. Soil-moisture readings obtained in August and Oc- tober of 1963 were 3-5 percent higher than readings taken in May and June, but the greatest differences were in readings about 0.5 m (1.6 ft) above and below the water E17 80 60 40 ACTUAL/STANDARD COUNT 20 RATIO o | 1 1 ! 0 10 20 30 40 50 MOISTURE CONTENT MEASURED, IN PERCENT BY VOLUME Ficure 18.-Calibration curve of soil moisture measurements by the neutron method; in- dividual readings indicated by dots. The correlation coefficient (r) is 0.911 and is highly significant. o T T T T ~ in O: S- -I E August and October f C 1963 w 3 Z 1.0 - = s o < u. & 1.51 I ad 2 May and June to 1963 & < x O = W w - i 2.0 T Ground-water level F— a. w o 2.91 A 3.0 1 1 1 1 0 10 20 30 40 50 soIL MOISTURE, IN PERCENT BY VOLUME Ficure 19.-Soil moisture in one tank showing the typically small variations. table, where the soil is saturated and such large variations are not likely to occur. AMBIENT AIR TEMPERATURE During two of the special observation series under- taken in 1966, soil-moisture readings were taken at 3- to 4-hour intervals over a period of 3 days. Data for one tank are plotted in figure 20. While the water-level fluc- tuations are negligible, there seems at first sight to be a considerable fluctuation in soil moisture, but these fluc- E18 4 T T T T T T T T T T T 2b 15 o 20 18, 16[- 14 28 281-75 T 24 = 22 36 34 )- =I 30 |- - 28) 7 26 IN PERCENT BY VOLUME 42 38 |- 36 |- e 34 )- { 32 SOIL MOISTURE, 44 42 |- - 40 |- 38 |- { 36)- A 34 I 1 I I | 1 | I 1 1 1 beoo N 1800 M O600 N 1800 Mm O600 N 1800 M Aug 1966 _ 3 4 5 LOCAL SOLAR TIME Ficure 20.-Apparent fluctuations in soil moisture during a 3-day period in one evapotranspirometer; true solar time, N=1200 and M=2400. Variations may be due to temperature sensitivity of the counting apparatus, the probe and the standard. Numbers at the left end of curves indicate depth of sampling, in centimeters, below soil surface. tuations are inconsistent. For instance, on August 3 soil moisture at all levels decreased between 1030 and 1600 hours, but between the same times on the next day the moisture increased. Temperature of the ambient air near neutron loggers can have a profound influence on the readings (Task Force, 1964), and this probably accounts for the apparent anomalies. Usually though, soil- moisture readings were taken in the early mornings and the effect of temperature differences could safely be ig- nored. SALINITY Another source of possible misreadings was the salini- ty of the soil moisture. It is known (Benz and others, 1965) that a high chloride content of the soil can affect the readings because chloride atoms are capable of ab- sorbing neutrons. The result is that, with increasing chloride content, the count ratio goes down. In order to measure this, several batches of rock salt and water were mixed and neutron readings were taken at different con- centrations. The batches were then analyzed for chloride content. In figure 21, counting ratios of the neutron logger are plotted against the chloride contents of the 40}|- § STUDIES OF EVAPOTRANSPIRATION [— Z fe $2.0 r --- e F rr o CC C ’\ T 0 °\ al O 0.8 |- © f - Z ' B <4 0.6|- C < « a 3 3 30.4— 2 g C ...... B28 d291 216 850 () (@) (@) () 33.7 i827 so00 iis | monthly water use. At times the water supply to this Nov: «si- 20.0 7.87 8.1 3.19 16.5 6.49 125 4.92 15.2 5.98 9.8 3.85 5 and bee ay. 126 ase 48 i8e 67 anm 29 iu _c) | tank was shut off and the rate of change in the declining water table was used in connection with energy budget Year ...... 1966 1967 'Not connected . analysis. For the same purpose, the tank was shaded oc- 'Ground water in tanks , i £ sok 3 } Z g __|flushed out with project | Ccasionally to study the different effects of drying of the ""ifank was fed better | tOP Soil on the rate of water loss. Some of the results are Cm In. Cm In. Cm In. Cm mm. [quality SC (see analysis | discussed under "Water Table Fluctuations"; other in table " Jaa yu o 0. anes y arly aca results will be discussed in a paper dealing with the ........ % if 5 0. is oBoly.. nl F...... Soc. t, o, 19 t of» 8.000" °C~"° | energy budget and mass-transfer methods for deter- pist.::. ...: 3. 0. I +5. ¥. fa. 196|~ , te. estan apres at 16.1 6.34 27.1 10.67 15.8 6.22 16.2 6.38 221505133 mining evapotranspiration. ....... 35.1 13.82 52.4 20.63 36.4 14.33 35.3 13.90| ;,; ; + + Mg i 111 1618 Ba9 31.84 16.93 ass 17.91]. "Ole myer floods Data from tanks 10 and 11 are summarized in table 15. duly -?. c. 42.4 16.69 47.3 18.62 39.1 15.39 40.5 15.94 i Aug 18.3 7.20 38.2 15.04 28.9 11.38 31.3 12.32 Here, as. in the vegetated tanks 1 through 6’_ the effect of Sept :.... (9 " ASL i- (Y 2 lic aer is. change in water level on water losses is quite apparent. Oct sai deans 16.92.0088 15.8- 602°. a . yoo on in in aes .. : Compare, for instance, the data for tank 10 with those of Pes rs." s 1 fo Aa t r car a>. tank 11 between June 1964 and December 1965. Before with better quality water, used more water than did tank 8 on project water. After the water qualities of the two tanks were switched in 1964 tank 8 began to use more water than did tank 7, but the differences were comparatively small. In 1965, tank 8 remained on better quality water, and was not flushed whereas tank 7 remained on project water but was flushed. Tank 7 measured higher water use than tank 8; although tank 8 was using better water, the salt accumulation resulted in a lower use. In 1966 June 1964 the water level in both tanks was the same and tank 10 seemed to use more water, which might be due to the fact that tank 10 did not have the protective belt which surrounded tank 11. However, in 1966, tank 11 used more water than tank 10. It is possible that the previous high water level in tank 10 resulted in a con- siderable salt accumulation in the top layers of the tank which, in turn, may have affected the monthly water use. It is interesting to compare the water use recorded for these tanks with data published earlier on water use in similar tanks near Yuma, Ariz. (McDonald and WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD TABLE 15.-Evaporation from bare soil in tanks 10 and 11 Tank 10 Tank 11 Date Nominal Nominal depth Water use depth Water use to water to water Cm Ft Cm In. Cm Ft Cm In. 1963 July 120. . 4.1 () (): 125: . 4.1 5.2 2.05 Aug .:.;....... 195 - A1. "2.1 80.82. 125 4.1286 "1.42 Sept .......... 125 »A1 ~90.5 8.14 - 125 4:1 6.0 2.36 Oct :...... ...} 125. 4.1 A56 f228 125 A1 '46 '1.77 Nov :;...:..... 125 _ :4.1 8b L188 125 (4.1 2:1 0.82 1964 Jan. rs .s 125): A.1 2:2 O81} 125, . 4.1 1.4 0.55 125 - 4.1 2.9 1.14 126 4.1 1.9 0.75 Mar:....:...... 125 - 4.1 8.7 145" . 125 4.1 2.3 0.90 ApH .... sir.... . 125 - 4.1 42 1.05) 125 4:1 3.17 1.46 May. .......... 125 - 4.1 (8) (): - 120 --4.1 - 4.8 - 1.00 June.:......... 100 3.3 10.0 5.94 120 . 4.1 7.5 2.96 100. ~$.8 "0.4 -'8.10 125. 4.1 © "7.5. "2.87 Aug :. 100. $3 "0.0 "5.54. 125 4:1 "'5.3 "2.08 Sept 100 _ 4.8 45.7 125 4.1 "4.2 '1.05 Oct 2. s..... 1 100 8.8 57 22 125 41 4.4 ; 1.78 Nov 100 - 3:5 $1 212 125" (4.1 () () Dec i.......... 100 .3.3. 42 1.05) 125: 4.1 () () 1965 100 > $3.8: 'A 'f'r.oll- 125) 4.1. 15.6 "2.20 May ....:..... 100 : 3.5 ° 11.6. 4:56} 125: 4.1 9.6 3.178 June .;... ..; 100; > 8.5: 121 446 126. 4.1 10.5 4:13 duly: 100; 3.9 11.4 4,49; 125 . 4.1 8.0 3.15 Aug ;. 100: "8.3 12.5 4902 125 4.1 10.5 _ 4.00 Sept: ......:...: 100 3.43 9.4 38.70. 125) "4.1. 6.0 2.50 Oct fs 100: 8.8~6.0° 2.96; 125. 4.1 6.1 2.01 Nov :s. 100. ~g.5= 4.7 «1851 125. 4.1 3.4 1.34 Dec :....}..... 100 _ 3.3 () (° 120. A1 ; 1.4. 0.55 1966 Apr ;. 150 - 4.9 (') ('): 150. - 49. . 5.2 2.05 May .:..:.;.:... 150 - 4.9 2.6 1.02) 150 - 4.9. -n.d. . nd. June..., .:.... 150 ~ 4.9 4.0 L657 150° M.9 6.5 2.48 July 150 _ 4.09. "3.3 ©1500, 150 4.0 . *0.0 "1.98 Aug :s. 100- 4.9 "4.7 "LAs 150 49 '2.44 1967 API 100 8.5 2.2. O81. 100 5.5 (4) (4) May .....:.... 100 _ 3.3 54 212: 100 - 3.3 (4) (4) Junei..:.;.... 100 4.8 7.8. 287: 100 - 5.3 4} (4) July ........... 100 - 10.0 8.94: 100 _ 3.3 4 (4) Aug :. 100. 8.3 12.2 "480. 100 _ 3.3 (*) C) 'Insufficient data to compute monthly use. 'Water use affected by rainfall. 'Chane of water level. 'Data suspect, excessive use probably due to leaks. Hughes, 1968) and in the Humboldt River Valley in Nevada (Robinson, 1970). The Humboldt River Valley report mentions water use by year but does not give data on temperatures and radiation. The growing season is undoubtedly shorter, radiation is less, and monthly temperatures are lower than those in either Yuma or Buckeye. It is therefore not surprising that the use is much less than that reported for the Yuma area. A comparison of the weather data in Yuma and Buckeye shows clearly that the situations are quite similar, yet the water use near Buckeye is very much higher. Table 16 allows comparison of the water use in three tanks near Yuma with the use in two tanks near Buckeye during the later part of 1963 and for 1964. Some E25 of the differences in water use might be explained by the differences in the soil. McDonald and Hughes (1968) pre- sent sieve analysis of the soils in the bare tanks at Imperial Camp. The percentage of particles less than 0.02 mm for the three Yuma tanks are 14, 8, and 6. The average percentage of particles for the soil in the bare tanks at the Buckeye Project is 28 for the top em and 25 for the top 50 cm, with respectively 6.5 and 5.8 percent clay (particles less than 0.002 mm). Obviously, the soil at the Buckeye site is much finer and, although this may slow down the upward movement of water, it also ac- counts for finer capillaries and assures a more con- tinuous supply of water for evaporation. WATER TABLE FLUCTUATIONS White (1932) and also Gatewood and others (1950) describe how one can make an estimate of evapotranspiration by analyzing the diurnal fluctuations of the water table. The fluctuations are partly caused by evaporation from the soil, but they are mainly caused by a pumping action as plants draw water from the water table or from the soil moisture above it. For this to be true, there should be a fall in water level during the day and a rise during the night. Although the phenomenon has been observed under natural conditions, the fluc- tuations of the water table in the evapotranspirometers showed the opposite: a rise during the day and a decline during the night. A typical example of fluctuation in water table for a bare tank with the water controls shut off is given in figure 30. There is a gradual decline in the water table, but contrary to what is expected, there is a rise in the water table in the afternoon. Afternoon rises, however, coincide with lowering of the barometric pressure. To make the comparison easier to follow, the barometer readings have been reversed and the scale adjusted. The distance between "low" and "high" represents 5-cm water pressure and is 40 percent of the distance between 1.50 and 1.55 on the left scale. This percentage is called the barometric efficiency (Ferris, 1959). Figure 31 shows a similar situation for a tank planted to saltcedar. The plants were dormant and the average iC iC u 149 [ [ ; Eff tw 1.50 _\\/’X1 \Water level Ei 22 | \\’// \\- Lug Zu OI—g ad. - z e CC T - Jn Fop in s 10 11 12 =C JUNE 1964 a. Ficure 30.-Water levels in evapotranspirometer 9 and inverted barometric pressure. No vegetation; water controls off; June 10-14, 1964. E26 STUDIES OF EVAPOTRANSPIRATION 16. -Monthly evaporation, excluding rainfall, from ground water in bare-soil tanks at Imperial Camp and near Buckeye, Ariz. Imperial Camp! Buckeye Tank No. ... BS1 BS2 BS3 10 11 Nominal depth to water in 1963...cm, ft... _ 60 1.9 90 3.0 105 3.5 125 4.1 125 4.1 Month Cm In, Cm In. Cm In. Cm In Cm In. July ..:.. 4.1 1.61 2-7. 1.05 1.6 0.65 5.2 2.05 Aug ..... ~ :20.04 "19. 10.46 20.4 _ 20.18 "0.82 ~H.4 Sept .... 9.5 3.74 6.0 2.36 Oct :;. a+ *L.1. -: 20.42 te: 5.8 _ 12.28 46 - Nov s. .&. 2.2 0.87 1.1 0.42 3.5 1.38 2.1 0.82 Nominal depth to water in 1964...cm, ft. 65 2.1 90 3.0 120 4.0 125} 4.1 125 4.1 Month Cm In. Cm In Cm In Cm In, Cm In Jan;... 2.6 1.04 1:7 0.67 0.4 0.17 2.2 0.87 1.4 0.55 Feb ..... 2.4 0.94 1.0 0.48 0.4 0.17 2:9 1.14 1.9 0.175 Mar .:.... 8.1 1.23 1.1 0.44 0.1 0.04 3.7 1.45 2.3 0.90 Apr :::%. 4.0 1.57 2.2 0.88 0.1 0.04 4.2 1.65 8.7 1.46 May 5.9 2.31 3.0 1.17 1.1 0.44 (4) (*) 4.3 1.69 June 9.2 3.62 3.0 1.19 0.7 0.28 10.0 3.94 7.5 2.95 July..., 7.0 2.74 2.4 0.96 1.5 0.61 *9:4 ~ "3.170 17.9 - Aug ..... 6.0 2.97 10.2 - 20.07 10.1 10.27 29.0 _ 23.54 - £2.08 Sept 5.2 2.05 1.4 0.56 2.6 1.01 "5.7 _. 22.24 A42 "1.66 Oct .;... 5.1 2.01 0.7 0.27 0.8 0.32 5.17 2.24 4.4 1.73 'Adapted from McDonald and Hughes (1968, table 6). 'Water use affected by rainfall. 'Depth to ground water was 100 cm (3.3 ft) after May 1964. 'Change in water level. a u a CC 2C g, 2.05 T U T I I T ® t [:P 2.05 [ l 1 T [ [ fig - a< | w - Water T fu Water **. ! Ew level e 3 2 level o Lol- 3 g & ff“ 93 2l bol | ~A Eo a 64) Pressure _§E 42 |- “£51 a thi § _ yo On | Ses 2i0)-.. 1 "I_ 7" A| pressure Low OK, + $5) 4s |y / _) /- [~ "A E| <2 ca 212 SF | 7 io sz X6 3 g {3 |®5 uo ~ oz | Cm ¥". NL ~. / o< i- f J (o s+ # High) wt} $ 1 1 1 1 1 L___l« s | $ 1 1 1 1 | C-~IfS 31 1 2 3 4 5 wus | S 26 27 28 29 30 31 u- JANUARY FEBRUARY 5 MAY 1966 5 1967 x = Ficure 31.-Water levels in evapotranspirometer 1 and barometric pressure. Tank planted to saltcedar but plants leafless and dormant; January 31-February 5, 1967. water level remained stationary, but the fluctuations followed the barometric ups and downs as before. The ef- ficiency is again about 40 percent. On vegetated tanks where saltcedar is transpiring, the barometric fluctuations and those of the water level are out of phase as shown in figure 32. The little wiggles were produced by the water-stage recorder in response to fill- ing and occur at exactly the same time that the recorders Figure 32.-Water levels in evapotranspirometer 2 and barometric pressure. Tank planted to saltcedar; May 26-31, 1966. indicate filling by the system illustrated in figure 25. Ob- viously the water level in the control pipes which contain the contact points (see "Instrumentation") did not res- pond to the atmospheric fluctuations. The water-level control pipes probably had a finer screen than the recorder pipes; also, the control pipes were installed shortly after the tanks had been constructed and a con- siderable amount of clogging could be expected. The WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD reaction to falling and rising water levels in the control pipes was therefore sluggish compared to that in the water-level recorder tubes which were installed in 1964. It is now possible to construct a hypothetical water level by adjusting the actual one for the atmospheric pressure changes, assuming a 40 percent barometric ef- ficiency throughout. Under high atmospheric-pressure conditions the water level would have been shallower had the pressure been lowered to the mean pressure; un- der low atmospheric-pressure conditions the water level would have been deeper if the atmospheric pressure were increased to the mean. Figure 33 shows the results of such a manipulation of data for a 5-day average of the changes in evapotranspirometer 6. The line representing the adjusted water level and the points showing observed hourly rates of water use are clearly in phase. CC oFu 1-5 =r -t -+«--+ --+ -+--r-1T-+T--1 Water table deep 1 a [=] Water levels E U NTI o T E R #B 0 in s | B o WATER USE, IN LITERS PER HOUR ‘\ \_ Computed water use ye P Observed water use t 0 U A R a o I N D o o -g 5 9 13 17 21 1 LOCAL SOLAR TIME FiGurE 33.-Average water levels, average barometric pressures, and average observed and computed water use in tank 6, planted to saltcedar, for June 30-July 4, 1966. The picture gets more complicated when there appear to be effects of salinity of the soil moisture. When the conductivity of soil moisture is low the plants seem to use that water first, and the ground-water levels are not affected until 2 to 4 hours after the transpiration has started. The result is a phase shift between the water- level curve adjusted for barometric effects and the curve drawn through measured points (van Hylckama, 1968). It is clear that diurnal atmospheric-pressure effects can be masked, and yet they may have influenced the water level in transpiration wells. Also it is possible that there was selective water uptake by the vegetation; without that the analysis of transpiration-well data could lead to wrong estimates, if not of the daily amount of water transpired, at least of the time of consumption. DISCUSSION In the foregoing pages two things seem to stand out rather clearly: (1) water use by saltcedar varies E27 with many factors, and (2) plastic-lined evapotranspirometers may be capricious instruments yielding data that should be considered with caution. Both statements, while true, are not very useful and should be reinforced with some further analytical studies. VARYING FACTORS OF WATER USE First of all, some of the data show that the water use by saltcedar can be enormous. Table 17 is a summary of all water-use data in rounded numbers, by years. The table shows that in 1965, tank 3 used a little more than 310 cm (122 in.) of water. This is nearly equal to the highest pan evaporation observed in the Lake Mead studies (Harbeck and others, 1958). A similar value was obtained by extrapolating data for tank 2 in 1966. Both tanks were flushed out in the spring of each year, and part of the high water use may be attributed to a high soil-moisture content, at least in the early part of the year. The next highest user for two consecutive years was tank 4 with 260 cm (102 in.) in 1965 and, again by ex- trapolation, 268 ecm (106 in.) in 1966. This tank also was flushed out both years. The other high uses of 200 or more centimeters occured from tanks 3 and 5 in the years 1961, 1962, 1963, and 1965, and are comparable to the use rates observed by Gatewood and others (1950). The 1965 total use from tank 2 was also high (238 cm or 94 in.). However, in 50 percent of the tanks the water use was less than 150 cm (59 in.). TABLE 17.-Summary by years of water use in and treatment of six evapotranspirometers Tank No: :- acs 3 5 4 1 2 6 Depth to water...m ... .... 1.5 1.5 2.1 2.1 2.7 2.7 Vegetation treatment .... None None None None None None Figshed }.. ...on 0.2 incr No No No No No No 1961... 199 200 141 145 105 108 78 79 56 57 41 43 1962... 218 222 137 150 94 94 86 87 54 59 37 37 1963... 226 229 160 163 86 92 i 89 90 63 64 34 36 Depth to water... m ...... 1.5 1.5 2.1 2.1 27 27 Vegetation treatment .... None None None None None None Flushed .... .... Yes No Yes No Yes No 1965..cm ....... 311 226 260 168 238 106 Wd essays. 122 89 102 66 94 42 Depth to water...m ...... 2.1 2.1 2.4 2.1 2.1 2.1 Vegetation treatment .... (*) () C) () None None 32s raz nes anl Yes No Yes No Yes No 1966s .em ...:. 154 51 268 122 294 145 din ¥ 61 20 106 48 116 57 Depth to water...m ...... 2.1 2.1 2.1 2.1 2.1 2.1 Vegetation treatment .... None None None None None None Flushed -:, .v... : No No No No No No 1907 SCM + 2T sea seas 137 86 161 109 156 134 lit redeco aes ares 54 34 63 43 61 53 'Cut to 50 cm . *Thinned 50 percent. 'Data extrapolated: 100/70 measured use, Mar. through Aug. Figure 34, a graph of the data from table 4, clearly demonstrates how the water use diminished with in- E28 400 [ T in 300 200 100 WATER USE, IN CENTIMETERS PER YEAR 0 1 L 1 0 100 200 300 DEPTH TO GROUND WATER, IN CENTIMETERS 400 Ficure 34.-Yearly water use (1961-63) versus depth to ground water in six evapotranspirometers at the Buckeye Project. creasing depth to ground water. The straight line drawn by eye through the points leads to two interesting in- tersections with the y- and x-axes. Extrapolation to the zero water level would indicate a water use of about 360 cm (142 in.), which just about equals the 1952-53 pan evaporation for Boulder Island, mentioned in table 19 of the Lake Mead studies (Harbeck and others, 1958). Ex- trapolation in the other direction leads to the absurdity that the saltcedar would stop using water with a water table at 360 cm (12 ft) below the land surface. It seems altogether reasonable, though, that the water use would continued to diminish with a declining water level, but more in character with the curved line in figure 34. The effects of salinity of the ground water and soil moisture can be illustrated in a similar manner (see fig. 35). Extrapolating the line to the left leads to an es- timated water use of about 360 cm (142 in.) for pure water, which again checks with the Lake Mead data. Ex- trapolating to the right end would lead one to expect a zero water use at a conductance of about 50 mmho em'; this is in the range of conductance of saturation extracts taken from the top soil. At such concentrations no seed, not even saltcedar seed, will germinate. Existing vegeta- tion would stop growing too, as implied in "Vegetation Growth and Development." The effects of stand density on water use in this study cannot be separated completely from the effects of salinity, but the interaction could be mathematically analyzed, as was done in table 9 with results shown in figure 16. STUDIES OF EVAPOTRANSPIRATION 400 T F T T t p + 1965 t t soo} a 3 P— z u < 0 w > > 200|- 4 __ CC W Io & B CC w 100} " fud < Es o I I | | (] 10 20 30 40 50 SPECIFIC CONDUCTANCE, IN MILLIMHOs PER CENTIMETER aT 25°C Figure 35.-Total water use per year in six evapotranspirometers versus specific conductance of the saturation extract of soil samples taken from the root zones in July or August of each year. The number 6 refers to an anomalous datum for tank 6 in 1965. The section "©1966: Density of Stand" refers to the practice of using a volume-density correction factor in estimating maximum water use by phreatophytes. For instance, if the volume density at time of measurement was 50 percent, it was assumed that the water use under 100 percent conditions would be twice as much. Under the circumstances of the experiments described in the previous sections, this is not necessarily so. Tanks deprived of 50 percent of their transpiring surfaces used only 10 to 15 percent less water than a control tank. Ob- viously, one has to consider these possibilities when es- timating water use under natural conditions. On one hand, the investigator knows the amount of water lost from a 100 percent volume density stand, and he divides this amount in half to estimate water use in an area with only 50 percent volume density. His estimates of actual water losses will then be too low. On the other hand, if a certain water use by a stand of 50 percent volume densi- ty is measured, a prediction as to what might happen when this stand develops to 100 percent volume density will lead to conclusions which may be grossly overestimated. Although a volume density correction factor is no longer used in current phreatophyte research (for exam- ple, McDonald and Hughes, 1968; Robinson, 1970), a search is still needed for a satisfactory way of expressing water use in relation to stand density on a quantitative basis. It would be pleasant to be able to estimate water use by simply measuring the vegetation (and maybe a few other factors, such as climatological and meteorological, which will be discussed in another report in this series). The evapotranspirometers deserve some further dis- cussion. Gatewood and others (1950) quote a list of sources of error in performing experiments with plants WATER USE BY SALTCEDAR AS MEASURED BY THE WATER BUDGET METHOD growing in tanks. Most of these deal with improper con- ditions of plant and soil compared with the natural en- vironment. The number of tanks, their size, and the duration of the project eliminated most of these sources of error, but others occurred that were not mentioned. Of these sources of error, the salinity build-up was probably the most serious. High salinity, however, does occur under natural conditions, and, although the situa- tion in the tanks may have been exaggerated compared to what happens under natural conditions, this source of error actually led to a better understanding of soil-, plant-, water-relationships. Also the water-level fluc- tuations due to variations in atmospheric pressure could and would have affected the water use in the tanks if the fluctuations had occurred in the water level control pipes. Another possible source of error was the choice of the location of the experimental site in the flood plain of the Waterman Wash. Flooding was part of the natural environment of the vegetation, and when levees were built the saltcedar that was not sprinkle-irrigated began to die during periods of prolonged drought. Flooding as a source of error was eliminated by interpolating observed water use from periods in which tank sites were not adversely affected by floods. METHODS OF EVALUATING WATER USE Properly built evapotranspirometers should have provisions for draining the soil column, but this makes them expensive-even more so when they have to be large enough for the study of tall vegetation. Whether the apparatus used is of the plastic lining type or of more sophisticated construction, studies of water use by the evapotranspirometer method is time consuming, es- pecially for such vegetation as saltcedar, mesquite, and other trees. If only a water budget method is used, the results can actually be applied only to areas quite similar in ecology. Studies to date on the effects of environment (radiation, temperature, winds, and so forth) on rate and quantity of water use have been encouraging and may eventually lead to a better method of evaluating water use by phreatophytes. Portable equipment could be set up in places where a knowledge of water use is required. Observations made during comparatively short periods of a few weeks or months would likely yield results which could be reliably extrapolated to yearly quantities. As a control, the water losses from the soil should be measured and the fluctuations of the ground-water levels should be observed. It is along these lines that further research in water use by phreatophytes (and other plant covers) would be desirable and possible. REFERENCES Arkley, R. Y., 1963, Relationships between plant growth and transpiration: Hilgardia, v. 34, p. 559-584. E29 Baum, B. R., 1967, Introduced and naturalized tamarisk in the United States and Canada (Tamaricaceae): Baileya, v. 15, p. 19-25. Benz, L. D., Willis, W. O., Nielson, D. B., and Sandaval, F. M., 1965, Neutron moisture meter calibration for use in saline soils: Agr. Eng., v. 46, p. 326-327. Brooks, C. E. P., and Carruthers, N., 1953, Handbook of statistical methods in meteorology: London, Her Majesty's Stationery Of- fice, 412 p. Chang, Jen-Hu, 1968, Climate and agriculture: Chicago, Aldine Publishing Co., p. 129-244. Colman, E. A., 1953, Vegetation and watershed management: New York, Ronald Press, 412 p. Douglas, E., 1967, Confusion among the saltcedars: Arizona Farmer- Ranchman, v. 46, no. 23. Ferris, J. G., 1959, Groundwater, chapter 6 of Wisler, C. O., and Brater, E. F., Hydrology: New York, Wiley and Sons, Inc., 408 p. Fisher, R. A., 1944, Statistical methods for research workers: London, Oliver and Boyd, Ltd., chap. VII, VIII. Fisher, R. A., and Yates, F., 1943, Statistical tables: London, Oliver and Boyd, Ltd., table V. Gatewood, J. S., Robinson, T. W., Colby, B. R., Hem, J. D., Halpenny, L. C., 1950, Use of water by bottom land vegetation in lower Saf- ford Valley, Arizona: U.S. Geol. Survey Water-Supply Paper 1103, 210 p. Harbeck, Jr., G. E., Kohler, M. A., Koberg, G. E., and others, 1958, Water-loss investigations: Lake Mead studies: U.S. Geol. Survey Prof. Paper 298, 100 p., 1 pl. Horton, J. S., 1964, Notes on the introduction of deciduous tamarisk: U.S. Forest Service Research Note RM-16, p. 1-7. Horton, J. S., Robinson, T. W., and McDonald, H. R., 1964, Guide for surveying phreatophyte vegetation: U.S.D.A. Forest Service, Agr. Handbook No. 266, Washington, D. C., 37 p. Huschke, R. E., ed., 1959, Glossary of meteorology: Am. Meteorol. Soc., 638 pp. Mather, J. R., ed., 1954, The measurement of potential evapotranspiration: Johns Hopkins Univ. Pubs. Climatology, VII, p. 7-225. 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. Meinzer, O. E., ed., 1927, Plants as indicators of ground water: U.S. Geol. Survey Water-Supply Paper 577, 95 p. Penman, H. L., 1948, Natural evaporation from open water, bare soil, and grass: Royal Soc. [London] Proc. A, v. 193, p. 120-246. 1955, Evaporation, an introductory survey: Netherland Jour. Agr. Sci., v. 4, p. 9-29. Penman, H. L., and Schofield, R. K., 1941, Drainage and evaporation from fallow soil at Rothamsted: Jour. Agr. Research, v. 31, p. 74- 109. Rider, N. E., 1956, Water losses from various land surfaces: Quart. Jour. Royal Meteorol. Soc., v. 83, p. 181-193. Robinson, T. W., 1970, Evapotranspiration by woody phreatophytes in the Humboldt River Valley near Winnemucca, Nevada: U.S. Geol. Survey Prof. Paper 491-D, 41 p. Robinson, T. W., and Bowser, C. W., 1959, Buckeye Project-water use by saltcedar: U.S. Geol. Survey and Bur. of Reclamation open-file rept. Tanner, C. B., 1960, Energy balance approach to evapotranspiration from crops: Soil Sci. Soc. America Proc., v. 24, p. 1-9. Task Force on Use of Neutron Meters, Committee on Hydrometeorology, 1964, Use of neutron meters in soil moisture measurement: Am. Soc. Civil Engineers Proc., Jour. Hydraulics Div., v. 90, p. 21-43. Thornthwaite, C. W., 1948, An approach toward a rational classifica- tion of climates: Geog. Rev., v. 38, p. 54-94. E30 Thornthwaite, C. W., and Mather, J. R., 1955, The water balance: Johns Hopkins Univ. Pubs. Climatology, v. VIII, p. 1-104. U.S. Bureau of Reclamation, 1964, Pacific southwest water plan, supplemental information report on water salvage projects: Lower Colorado River, Arizona-California: Denver, Colo., U.S. Bur. Reclamation, 115 p. U.S. Geological Survey, 1954, Water-loss investigations-Lake Hefner studies, technical report: U.S. Geol. Survey Prof. Paper 269, 158 p. van Hylckama, T. E. A., 1963, Growth, development, and water use by saltcedar: Internat. Assoc. Sci. Hydrology, Pub. 62, p. 75-86. 1966, Evaporation from vegetated and fallow soils: Water Resources Research, v. 2, p. 99-103. 1968, Water level fluctuations in evapotranspirometers: Water Resources Research, v. 4, p. 761-768. White, W. N., 1932, A method of estimating ground-water supplies based on discharge by plants and evaporation from the soil: U.S. Geol. Survey Water-Supply Paper 659-A, 105 p. APPENDIX: ANALYSIS OF VARIANCE WITH: MULTIPLE CLASSIFICATION If an investigator subjects units of his experiments to a variety of treatments it is possible by statistical analysis (provided the experiment is properly designed) to separate the effects of each treatment from random, or so-called error, effects. If the error effects due to unexpected influences are large, it becomes impossible to draw reliable conclusions about the significance of intentional treatments. For example, in this paper units are tanks or saltcedar twigs, treatments are depth to water, salinity of soil moisture, and so on. The random or error effects result from possible differences in exposure to wind or sunlight, from differences in the functioning of the plumbing ap- paratus, from mistakes in chemical analysis or measurements, and from other differences between plants or tanks.! Since there were only six large tanks the number of prescribed treatments was limited to two; otherwise we would have lost too many degrees of freedom and an analysis of variance would have been impossible. Our mathematical model thus becomes a relatively simple one. We might say that the magnitude of a particular observation is comprised of the following components: (1) the value of a mean that all observations in a par- ticular experiment have in common, (2) one or more components arising from particular treatments, and (3) a component resulting from errors and random effects. 'A lucid discussion is given by Fisher (1944). We must assume here that the reader is familiar with such terms as: mean, standard deviation, variance, and degree of freedom. STUDIES OF EVAPOTRANSPIRATION For instance if we have two treatments, differences in depth to ground water and flushing versus non flushing, we may write: Tij=m+dit+f j+ei, (1) where m is the overall effect or mean, d; is the effect of depth to water on water use, f; is the effect of flushing or nonflushing, ei; is the error, and T;; is the water use value in a tank treated with a depth to water value, i, and a flushing treatment, ;. Applying the method of least-squares we can arrive at an equation of the type: (2) in which the C's are constants. Thus, one obtains a set of actual and theoretical results to which a test of linearity can be applied. Then, a t-test to the partial regression coefficients would show that, for instance in table 5, the depth to water was the overriding influence on the quan- tities of water used and that the effect of years was only very small. This method, however, is slow and the error sum of squares is much more quickly computed by a technique known as the analysis of variance. In the method of least-squares each observation is represented as the sum of two or more components according to the mathematical model; in the analysis of variance the sum of such squares of the observation is partitioned accord- ing to components. This is illustrated in the analyses of variance given in tables 5, 7, and 12. The F-test is then applied, and the value obtained is compared with those in an F table. It is customary to mark F's that indicate significant differences at the 1-percent level with a double asterisk, those at the 5 percent level with a single variation and interaction that has been computed. Thus one has an estimate of those treatments which had significant effects on water use and those which did not. The final step is to compute the least significant difference, which is done by making use of the so-called t-table. The distribution of t is used to test the significance of a deviation when its standard error is es- timated from the data. Thus, t is the deviation divided by its estimated standard error. Then the least signifi- cant difference is computed by multiplying the ap- propriate t-value by the root of the mean square for error. This is the yardstick by which we can judge the significance of the differences due to treatment. See tables 5, 7, 9, 10 and 12. Water use per year=C:+Cd+C3f r US GOVERNMENT PRINTING OFFICE: 1974-677-304/7 Weather and Evapotranspiration Studies in a Saltcedar Thicket, Arizona By I. E. A. VAN HYLCKAMA SIP U D IE S -O F- EY A-P Oo F R A NSP LR A T L-O N EEOLOGICAL PAPER 49 -E UNITED STATES GOVERNMENT PRINTING OFFICE. WASHINGFON.: 19080 UNITED STATES DEPARTMENT OF THE INTERIOR CECIL D. ANDRUS, Secretary GEOLOGICAL SURVEY H. William Menard, Director Library of Congress Cataloging in Publication Data Van Hylckama, T. E. A. Weather and evapotranspiration studies in a saltcedar thicket, Arizona. (Studies of evapotranspiration) (Geological Survey professional paper ; 491-F) Bibliography: p. F48-F51. 1. Tamarix Chinensis. 2. Evapotranspiration-Arizona-Buckeye region. 3. Vegetation and climate-Arizona-Buckeye region. 4. Microclimatology-Arizona-Buckeye region. 5. Phreatophytes- Arizona-Buckeye region. I. Title. II. Series. III. Series: United States. Geological Survey. Professional paper ; 491-F. QE75.P9 no. 491-F [QK495.T35] 557.3s 80-11316 [583'.158] For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 CONTENTS Page Instrumentation-Continued Page I NUSLTACL n I- eil. le F1 Radiation :>. 20: erin enol neenee eo neb eras oo are oe F19 ._ -_ cor 1 Soil temperature, heat flow, and soil moisture __________-- 20 Saltcedar and other phreatophytes _____________________- 1 Soil 20 2 Heat flow plates £.... _-. ..:: ell -ne SPL uce 20 3 Soil moisture l ree bra dunit il 20 Methods of studying water tise -_.--_--..-__._-._.______.--_... 3 Water-tise 21 AWit .... ...l .es odc 3 The microclimate of a saltcedar thicket __________________---- 22 methods nn. 4 why the microclimate? 22 Micrometeorologic 4 'The special observation periods __._._._._.________________ 22 Empirical 5 Daily cycles of temperature and humidity __________------ 23 The Buckeye Lest Site 5 Winds {:... Bcit cll 27 Location; soil, and water. 5 Radiation --- 0 eee rit alain 30 caer sus ne o 6 Soil heat flux clo cee ccc er eae 30 Veselation caa 8 Soil temperatures LC- css 34 Evapotranspiration and its calculation 10 Heat flux and WALETAMSG adlut aas 34 10 Turbulent 35 T _-. -e mse o 11 Carbon dioxide 2s: snes aI ca sess 36 Fiow resistance 12 Predicting WALET USG 40 methods 12 Use of long-term means 40 The atmospheric water balance ____________________-- 12 Use of daily 42 The eddy-correlation method -___.__._._._______._______. 13 Short-term predictions z.... 43 Bulk-aerodynamic 13 Discussion ~ _s .t _o rit ience lien el ena e ut aoi emale 47 The Dalton 14 Referencesscitod 2}. LZ elin coli ec der n canis 48 The aerodynamic or profile approach _______________- 14 Appendices .= 53 Enersgy-budget method 15 A. Tables Combination methods .-___..-.:...l_._______l_l.l.l..l... 16 A-1 Hourly averages of windspeeds and wind direction instrumentation ..-... . cone 17 A-2 Hourly averages of air temperature and vapor pressure The atmosphere: Wind and temperature _______________-_-- 17 A-3 Radiation data ViAd _ : e 17 A-4 Windspeed, temperature, and water vapor data AXTFALEMpETALUTE - .L =.. ._ .cn cane eee eee ee ee s 18 B. A nomogram for data reduction Precipitation and water vapor _______._____________._____ 18 C. Symbols and dimensions Rainfall: :n 22 cn neon nce ancenenes 18 Water ...- o al oben 18 ILLUSTRATIONS Page Ficur®E 1. A reproduction of plate 1 in Hales (1727) "Vegetable Statitk®"" .- ._... «ou F4 2. . Map showing location -of the project Site .._. - - ns canes 5 3. Sketch of the project site, southeast of Buckeye, Ariz., showing position of main features _____________________---- 6 d..: Aerial photograph of the project Site -_ rok as sn ooc 6 6; Photograph of the project site from a 4-m-high platform 6 6: Graph «showing mean monthly temperature -. :c--.--L-. 1. outs ener 7 7; Circular graph showing wind roses -__._...._c-..._..___...:..__. e iain onan ban one ae akan eo ee 8 6: Circular graph showing vind directions ..-... >.. cns cur au nace ues s 8 9-12. Graphs showing: o.. Mean monthly wind spepds ---.. <.. ce dnes cies 8 10: Monthly rainfall col anes ss o 9 11: Solar radiation : Pollie nle ruen re denn sean s oan ane men 9 12. Dew-point LfemperalUifas -_ onc LL Gif dinar dn tabs ces anes sales 9 13; Schematic «drawings of leat stomata -_ den nle EIC eer va eee ana ani naudasbaes aas cas 11 14. Example of a 16-point strip chart ol. cele 17 15+ Example of a windspeed FECOMA 22 reece nce innuendo ench ner concn - beaches ae nan ces crane 18 III IV FIGURE TABLE 16. 17. 18. 19-22. 23. 24. 25. 26. 27-45. om. ni mm on im so po I- CONTENTS Page Skeich of an air-temperature sensing syStGM F19 Diagram of the ventilation system =s 19 Nomogram to compute vapor 20 Photographs showing: 19. Radiation mast showing instrument installed . _._ _. 21 20. Adjustable nomogram for computing heat flow .Q LLL 21 21.) Tensiometers Installed in one tank :-. __.. . L Ie ELI Lee ao ace B 21 22. Reading of tensiometers at Hight :-. cern 22 Sample of an event recording chart ...-... 22 Graph showing microclimate characteristics of the Buckeye project Site 23 Graph showing the distribution of windspeeds within and above a stand of saltcedar ________________________---- 27 Example of a border-punched card used for analyzing wind profiles 28 Graphs showing: & 27. How to determine roughness length, friction velocity, and zero displacement __________________________--- 29 28. Relation between modified roughness length and windspeed ratios 30 20. Modified roughness length versus stability ratio ...... op 30 30. March of radiation and heat flow for seven 3-day periods 31 $1. Methods 'of computing heat flow in the soil 32 32. Heat flow into and out of the ground in an evapotranspirometer near Buckeye, Ariz. __________________-- 33 33. Heat flow into and out of the ground in a lysimeter near Tempe, Ariz. ___________________ccccccccccc___- 33 84. Sinusoidal 24-hour heat flow into and out of soils 33 35. Relation between amplitudes of daily temperature waves, damping depths, and depths of 0.05 x amplitudes 35 86. Relation between soil heat Hux and water Uge -.-. c 85 37. Relation between windspeeds, temperatures, and vapor pressures 37 38. Fluctuations of CO; content of the air and windspeeds over a saltcedar thicket ________________________-- 40 39. CO, fluctuations over a saltcedar thicket and its harmonic analysis _____________________ccccccccccc__-- 40 40; Relation between. CO; and water vapor fluxes - -_-. ._.. .._... .... . Lull n coos a be a aB an me malas ale me 40 41. Average water use as computed by various formulas and compared with measured use ______________-_--- 41 42. Water use as predicted by the Konstantinov model versus measured use __________________________--_---- 45 43. Water use as predicted by the Rider model versus measured Use = 45 44. Hourly rates of evapotranspiration for two 3-day periods near Buckeye, Ariz. ____________________-_---- 46 45. Hourly rates of evapotranspiration as measured and as computed using equation 26 _________________--- 47 TABLES Page Analyses of soils at the Buckeye project _.. -n be F7 Analyses of irrigation water at the Buckeye project "-_. colle 7 Plants (other than saltcedar) observed growing at the Buckeye project site, 1961-66 __________________________-- 10 Values of y(AT/A¢) for a 24-hour period..May 5 and 6, 1966. } 16 Heat-flow plate calibration factors --.. L.. -...... .r: C- ans 21 Some results of harmonic analyses of soil-temperature fluctuations inside and outside tank 6 for two 3-day periods 35 Variables used in some models for predicting potential evapotranspiration 41 Comparison of different methods of estimating evapotranspiration for 1963 using the harmonic analysis data ___- 42 Water use in tanks 2 and 4 during special observation periods with some results of harmonic analyses of the hourly data -.2.2:2. cic ns o uk cod sas tsun an den nan deus ace o ae 48 Some climatic data pertaining to seven special observation periods of 3 days each _________________________----- 43 Water use during special observation periods of 72 hours each as measured in tanks 2 and 4 and estimated according to various equations ... L. sees 44 Water use by saltcedar Tor 3 days in 1966 -ne. 47 Multiply SI Unit Area square centimeters (cm") square meters (m*) square kilometers (km*) Electricity and Magnetism millisiemens (mS) Energy joules (J) Energy/Area Time watt/meter' (W/m?) Heat watt/meter*kelvin (Wm K-') kelvin (J kg~' K-) Length nanometers (nm) micrometers (m) millimeters (mm) centimeters (cm) meters (m) kilometers (km) Mass kilograms (kg) Mass (Density) kilogram/meter® (kg/m®) Power watt (W) Pressure kilopascal (kPa) Temperature degrees Celsius (°C) kelvin (K) Velocity meter/second (m/s) Volume cubic meters (m*) CONTENTS CONVERSION TABLE By 0.155 10.76 247.1044 1.0 9.4787 x 10 0.2388 5.2895 x 10~ 1.434 x 10% 0.1761 2.388 x 10~ 0.3937 x 10~ 0.03937 0.03937 0.3937 3.281 0.621 2.205 6.24 x 107 0.001 3.412 0.2388 0.2953 10 1.8°C +32 (K-273.15) 1.8 +32 2.237 0.81 x 10: To obtain inch-pound or egs units square inches (in.") square feet (ft.") acres millimhos (mmho) British thermal units (mean) (Btu) calories (cal) Btu/foot*minute cal/centimeter*minute (cal cm~* min~') Btu/ft hour°F Btu/poundmass °F mils n inches (in.) n foot (ft) miles pound mass (Ibm) poundmass/foot® gram/cm® (g cm ~>) Btu/hour calories/second (cal/s) inches of mercury (in. Hg) millibar (mb) degrees Fahrenheit (°F) miles/hour (mph) acre-feet (a/f) STUDIES OF EVAPOTRANSPIRATION WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA By T. E. A. van HyLCKAMA ABSTRACT Water use by saltcedar, Tamarix chinensis, was studied from 1961 through 1967 near Buckeye, Ariz. The test site was located on the rim of the Gila River flood plain and was bordered on the north, east, and west sides by fetches of dense saltcedar thickets 1 or more kilometers wide. On the south side, however, the fetch was less than 100 meters. The climate of the area, typical of the Sonoran Desert, is charac- terized by low humidities, strong winds, and temperature extremes of <-10°C in winter and >50°C in summer. Potential evaporation values are among the highest observed in the United States. Rates and quantities of evapotranspiration were observed in six plastic-lined evapotranspirometers (tanks) whose 81 square-meter surfaces were planted to saltcedar having heights and density equal to those of the surroundings. The test site was further equipped with instrumentation to meas- ure the following data: solar short-wave radiation; long- and short- wave net radiation; albedo; humidity, temperature, and wind profiles in and over the vegetation; soil-temperature gradients and soil-heat flux; moisture content of the soil; and carbon dioxide content of the air. Detailed studies were made of the microclimate in and over a typical saltcedar thicket. Analyses showed that, above the vegeta- tion, the wind profiles in more than 80 percent of the observations were logarithmic. Within the thickets considerable turbulence and irregular wind inversions (tunneling) occurred during daylight hours. It was concluded that transport constants for momentum, heat, and vapor are the same more than 80 percent of the time because plots of windspeed versus temperature at different heights and of windspeeds versus vapor pressure at those heights fall on straight lines. Fluxes of carbon dioxide and vapor are closely related. Vapor fluxes diminish, as do rates of photosynthesis, during hot afternoons when temperatures exceed 40°C, suggesting a variable stomatal re- sistance factor. Estimates of potential evapotranspiration rates using various models were plotted against measured values. For rough estimates of total yearly quantities of evapotranspiration, three independent methods gave values closest to those measured by evapotrans- pirometers with the shallowest (1.5 m) depth to water and with low salinity of soil moisture. (The effects of depth-to-water and of salinity have been discussed extensively in "Water use by saltcedar as meas- ured by the water-budget method" (van Hylckama, 1974, U.S. Geological Survey Professional Paper 491-E). For short-term estimates (of the order of 1 hour) the 1966 combina- tion method of C. H. M. van Bavel gave results that were too high during daytime hours. When appropriate corrections were made by taking stomatal and aerodynamic resistances into account, the calcu- lated values fitted the measured ones very well. This shows that saltcedar reacts to extremely high windspeeds and temperatures by stomatal closure, thus diminishing evapotranspiration even though water is freely available (as it is in evapotranspirometers with favor- able water and soil conditions). That riparian vegetation always uses water at a potential rate cannot be taken for granted, and quantita- tive estimates of salvageable water based upon that assumption may at times be far too large. INTRODUCTION SALTCEDAR AND OTHER PHREATOPHYTES Saltcedar, Tamarix chinensis," is one of more than 70 species listed by Robinson (1958) as phreatophytes, a word derived by Meinzer (1923; 1927) from the two Greek words ppexrog (phreatos = well) and pvrov (phuton = plant). These plants are called phreatophytes because they have root systems with ac- cess to ground water or to water in the capillary fringe above it. It has sometimes been assumed that they can use water at a rate equal to, or greater than, the rate of potential evapotranspiration, even during long periods of drought. Whether this is always true is debatable, as will be discussed later in this report. A fact is that many phreatophytes, especially saltcedar, grow dense- ly along rivers and reservoirs, and the expression ripa- rian vegetation (from the Latin ripa=border) seems to be more appropriate and descriptive. Phreatophytes were estimated by Robinson (1958) to cover 16 million acres (65,000 km in the 17 Western States and to consume nearly 25 million acre-feet (31 x 10° m*) of water per year; they were expected to spread and to consume even more water in the years after 1958. As for saltcedar alone, Robinson (1965) esti- mated that in 1961 there were 900,000 acres (3600 km") covered by saltcedar using 3.5 million acre-feet (4.3 x 10° m*) of water per year. However there are reasons to believe that these numbers may be greatly exaggerated. For instance, 'In earlier studies the names T. gallica L. and T. pentandra Pall. are used. The name T. chinensis Lour. is preferred now by taxonomists (Horton and Campbell, 1974). Fi F2 STUDIES OF EVAPOTRANSPIRATION Robinson reported that of the 900,000 acres (3600 km?) just mentioned, about half were in Texas and that nearly half of those, or about 218,000 acres (880 km*), were located along the Pecos River and its tributaries. But that is far larger an area than the entire flood plain of that river and its feeders, including all agricul- tural land (J. S. Horton, written commun., 1977). Nonetheless, Robinson's data served to create active, even alarmist' interest in the phreatophyte problem and have encouraged the development of many action programs. The data also created the desire to know how much water these plants actually use and whether there are means of making this amount available for more beneficial use. In this report procedures of studying water use are discussed, methods, of measuring evapotranspiration are presented, the microclimate of vegetation on a typ- ical flood plain is described, and methods of indirectly estimating water use by saltcedar are evaluated. LITERATURE O. E. Meinzer (former Geologist-in-charge, Division of Ground Water, U.S. Geological Survey) who did much to call attention to the economic significance of phreatophytes in arid and semiarid regions, might have been surprised if he could have foreseen the amount of study that would be devoted to evapotrans- piration and related topics. A single bibliography (Rob- inson and Johnson, 1961) emphasizing papers on evap- oration and transpiration from the United States from the early 1800's until 1958 contains more than 600 titles. Humphreys (1962) assembled 804 references (mostly American) dated between 1802 and 1960. Hor- ton (1973) collected over 700 references to evapotrans- piration as related to phreatophyte management. There is very little overlap in these compilations, and, although there is a lot of speculation and empiricism, there is also very little in all this that specifically deals with the physical relationships between water use by phreatophytes and climatologic and meteorologic events. Research into the ecology of phreatophytes in gen- eral, and of saltcedar in particular, has been going on now for more than 25 years. Gatewood and others (1950) were probably the first to draw attention to the potentially large amount of water that saltcedar could transpire under favorable conditions. Since that time literature and study projects have proliferated enor- mously. The Phreatophyte Subcommittee of the Pacific *In the literature, saltcedars have been described as "water hogs" (Douglas, 1954), "ag- gressive" (Robinson, 1965), "greedy"(Douglas, 1965), "insidious" (Sebenik and Thames, 1968), "thieves" (Robinson, 1952), and "water-stealing culprits" (U.S. Information Service, 1965). These expressions indicate the emotional and propagandistic attitude that is some- times taken towards the phreatophyte problem. Southwest Interagency Committee (1965) compiled more than 500 references on phreatophytes and related subjects, nearly all more recent than 1950, while Robinson (1964) listed 48 projects pertaining to phreatophyte research. By 1965 a large number of these projects had been in progress for 2 or more years. In spite of the dedicated efforts of many agencies, in- cluding the U.S. Department of Agriculture's Agricul- tural Research Service, the U.S. Forest Service, the U.S. Geological Survey, and several University Exper- iment Stations, the complexity of the problem is such that we are still unable to just go into the field, set up some, preferably simple, instrumentation and deter- mine in the course of 2-6 months how much water is used by phreatophytes in that area and to estimate with reasonable certainty if and how much water can be diverted to economical use by eradication or conver- sion of the vegetation. The present study clarifies some of the problems and suggests possible solutions, but also describes in detail the difficulties one can expect while trying to solve such problems. Many of them have been discussed in the following papers, reporting on one or more aspects of the Buckeye Project (in chronological order): van Hylckama, T. E. A., 1960, Measuring water use by saltcedar: Proc. Fourth Ann. Ariz. Watershed Symposium, Watershed Management Div., State Land Dept., State of Ariz., Phoenix, p. 22-26. 1961, Natural recharge of ground water, in Fletcher, J. E., and Bender, G. L., eds, Ecology of ground water in the southwestern United States: 37th Ann. Meeting of the Southwest and Rocky Mountain Div. of Am. Assoc. for the Adv. of Sci., Arizona State Univ., Tempe, Ariz., p. 21-41. 1963, Growth, development, and water use by saltcedar (Tamarix pentandra) under different conditions of weather and access to water: Publ. 62, Internat. Assoc. Sci. Hydrology, Comm. for Evaporation, UNESCO, Paris, p. 75-86. Shakur, Abdul, 1964, Evapotranspiration from a stand of saltcedar: Masters Thesis, Dept. of Hydrology, Tucson, Univ. of Ariz., 73 p. van Hylckama, T. E. A., 1966a, Evaporation from vegetated and fallow soils: Water Resources Research, v. 2, p. 99-103. 1966b, Effects of soil salinity on the loss of water from vege- tated and fallow soil: Publ. 83, Internat. Assoc. Sci. Hydrology, UNESCO, Paris, p. 636-644. 1968a, Water level fluctuations in evapotranspirometers: Water Resources Research, v. 4, p. 7 1-768. 1969a, Photosynthesis and water use by saltcedar: Bull. Internat. Assoc. Sci. Hydrology, v. 14, p. 71-83. 1970a, Water use by saltcedar: Water Resources Research, v. 6, p. 728-735. 1970b, Winds over saltcedar: Agr. Meteorology, v. 7, p. 217- 288. Ripple, C. D., Rubin, Jacob, and van Hylckama, T. E. A., 1972, Es- timating steady-state evaporation rates from bare soil under conditions of high water tables: U.S. Geol. Survey Water-Supply Paper 2019-A, 39 p. van Hylckama, T. E. A., 1974, Water use by saltcedar as measured by the water budget method: U.S. Geol. Survey Prof. Paper 491-E, 30 p. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F3 1975, Estimating evapotranspiration by homoclimates: Geog. Rev., v. 65, p. 37-48. The following papers were presented but only printed in abstracts: van Hylckama, T. E. A., 1962, Use of water by saltcedar as measured by the water budget method: Jour. Geophys. Research, v. 67, p. 3535. 1964, Evaporation from bare soil: Am. Meteorol. Soc. Bull., v. 45, p. 540. 1968b, Wind profiles over saltcedar: Am. Meteorol. Soc. Bull., v. 49, p. 769. 1968c, Carbon dioxide and saltcedar growth: Southwestern and Rocky Mountain Div. of the Am. Assoc. Adv. Sci., Proc. 44th Ann. Meeting, El Paso, Texas, p. 5. 1969b, Plant growth and water use as affected by salinity of soil moisture and density of stand: Arid Lands in a Changing World, Abstracts of contributed papers, Am. Assoc. Adv. Sci., Comm. on Arid Lands, Tucson, Ariz., 2 p. 1972, Leaf resistance in saltcedar, in Adams, W. P. and Hel- leiner, F. M., eds., International Geography 1972, Univ. Toronto Press, p. 204. ACKNOWLEDGMENTS Under a memorandum of understanding dated May 4, 1959, "concerning studies of consumption of water by phreatophytes and related problems," the U.S. Geolog- ical Survey "will take the leadership in the preparation of reports and the publication of results of the studies conducted under the agreement" while the U.S. Bureau of Reclamation would be mainly responsible for the construction of the tanks and the flood protec- tion of the area. In particular, the help and assistance of Curtis W. Bowser, Robert Mason, and Leslie C. Pampel, all members of the Bureau, is gratefully ac- knowledged. The late Thomas W. Robinson together with Curtis W. Bowser conceived the idea of using 20-mil (about 0.5 mm) black vinyl plastic as a cheap and easy-to- install lining for the evapotranspirometers and super- vised the installation of the first two. G. Earl Harbeck, Jr., gave general supervision to the work of the Survey. Oliver E. Leppanen designed and constructed the wind-recording equipment. Jack L. Sidders supervised the construction of the tanks and aided in the installa- tion of the complex plumbing and the electrical cir- cuitry required for the study. He also performed nearly all the maintenance during the 6%-year duration of the project. Many others, among them a number of graduate and undergraduate students, helped in taking around-the- clock readings for periods of 3 or 5 days and assisted in the very time-consuming transcription of the strip- recording data into numerical and calibrated values. I am grateful for very helpful criticisms of the first draft of this report made by J. S. Horton, U.S. Forest service, (retired), and D. C. Davenport of the Univer- sity of California at Davis (Dept. of Land, Air, and Water Resources). METHODS OF STUDYING WATER USE A BIT OF HISTORY All plants need water for growth and development and most of them must transpire considerable quan- tities in order to grow and develop. That plants need water has been known undoubtedly since the time our earliest ancestors began to practice agriculture. One of the first to measure this water use was van Helmont (1655), who found that a 5-pound willow sapling in 5 years grew up to a 169-pound tree and concluded that the 164 pound increase came exclusively from the wa- ter.} That plants transpire however, was first described by Stephen Hales (1727). He presented data on the quan- tities of water transpired by different plants under dif- ferent conditions. Figure 1 shows just two of the dozens of experiments described by Hales and gives an idea of his painstaking work. Hales' studies were not im- proved upon until nearly a century later when many physiologists and physicists began to study the plant- soil-water relationships more intensively. Initially the transpiration phenomenon was only of scientific or academic interest; later, when economic considerations became urgent, the amount of water transpired with- out apparent benefit to man became the center of at- tention and the concept that transpired water is wasted became prevalent. To find out how "wasteful" saltcedar is, one must know how much water the plant uses under various circumstances. The basic method is essentially the same as that used by van Helmont and Hales: measure the water applied and after a definite time reweigh container plus plant and compute from these data the water lost by transpiration and evaporation from the soil. The crux is that in this manner only one or at most a few plants are studied, and one cannot safely extrapo- late that use to hectares or acres much less to thousands of them planted to such crops or covered by such vegetation. Therefore, since vegetation conversions, of which eradication is the extreme, cost money, one must make an estimate of the quantities of water that are used by the total vegetation to be converted or eradicated for the purpose of computing cost-benefit ratios. *"Omnia***vegetabilia***ex solo aquae elemento prodire hac mechanica didici"-freely translated: "That all vegetable matter comes solely from elements of water, I proved by this experiment." Then follows the description of how van Helmont planted, weighed, watered, and reweighed his willow tree. The actual weight of van Helmont's pounds is not certain. F4 STUDIES OF EVAPOTRANSPIRATION 5. Gm magmas.” FicurE 1.-A reproduction of plate 1 in Hales (1727) "Vegetable Staticks." The drawings were used to illustrate experiments I and VI. Plant on the left is a spearmint, Mentha spicata; the one on the right is a sunflower, Helianthus annus. HYDROLOGIC METHODS The most straightforward way of determining water use by saltcedar by means of containers (the lysimeter method, mentioned in any textbuok on hydrology) was discussed in another chapter of the Professional Paper 491 series (van Hylckama, 1974). However, finding a suitable location for evapotranspirometers or lysime- ters is often difficult; the construction is expensive; and the maintenance, until reliable results are obtained, very time consuming, especially if one is, as in this case, dealing with tall vegetation. (Saltcedar can reach a height of 5-7 m by the time it is 10 or more years old.) There are two other ways to measure water use by hydrologic (sometimes called water-balance) methods. The first is the natural-catchment or watershed- hydrology method of which Culler (1970) described a good example. Troubles with this method are discussed in the reference and will not be dealt with here. The second, called the "change-in-soil-moisture" method, consists of monitoring the changes in soil moisture, usually over the depth of the root zone. However, in- strumentation difficulties do not make this method very reliable over periods of less than 7-10 days, and large errors may occur if drainage is not measured or goes unnoticed (van Bavel and others, 1968) or if fluxes from the water table are neglected (E. P. Weeks, writ- ten commun., 1978). MICROMETEOROLOGIC METHODS Answers could be obtained much faster, and cer- tainly with less destruction, if one could evaluate the climatic and meteorological inputs that determine the reactions of the plants and the amount of water used or transpired by them. Such methods of measuring water use or evapotranspiration are available, do not require lysimeters, and do not depend on empiricism or edu- cated guesses. They can be classified as vapor-flow, bulk-aerodynamic, energy-balance, and combination methods. There are two types of vapor-flow methods. The atmospheric-water-balance method is analogous to the hydrologic balance in that it considers the quantities of moisture entering and leaving a column of air (rather than that entering or leaving a mass of soil or a watershed). The other is known as the eddy- correlation, eddy-flux, or eddy-transfer method. In this method the vertical turbulent fluxes of vertical wind and vapor in the atmosphere are measured over very short intervals, a few seconds at most. The method re- quires extremely sensitive wind and water-vapor sen- sors and needs elaborate recording systems to assimi- late the enormous amount of data accumulating during a day or even an hour. Since neither the sensors nor the integrating circuits or computers were available at the Buckeye Project, this method was not used but it will be mentioned together with the theory of the other methods and their instrumentation. The bulk-aerodynamic and the energy-balance methods go back, in principle at least, more than 2000 years. When Aristotle noticed that water, put on a plate, slowly disappeared, he said, "the wind does it." Others, during his lifetime, accused the sun. While the wind causes the aerodynamic or turbulence process, the sun causes the energy- or heat-balance process. These two systems still operate today; they are no longer theory as they were in Aristotle's days because now we have the instruments to measure them and their effects. The fourth method is actually a fusing of the effects of wind and sun and a number of equations have been proposed to estimate water losses by this combination method. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA EMPIRICAL METHODS Faster methods (apart from more or less educated guesses) involve one of the many empirical equations for computing potential evapotranspiration from read- ily available climatological parameters. Because such equations are empirically derived, claims as to their accuracy or inaccuracy are often without experimental foundation. In order to test the claims, many field ex- periments have been set up, mostly in the southwest- ern parts of the United States, where vegetation con- version is considered to be most effective. However, too often only scant attention is paid to the effects of the ambient weather upon water losses, and efficient ver- ification of accuracy and applicability of an empirical relationship remain debatable if not impossible. For instance, only a few of the projects listed by Robinson (1964) and only a few of the articles compiled by the Phreatophyte Subcommittee (1965) are concerned with the input forces and ambient conditions that determine the rate of water use by phreatophytes, and even fewer articles if one considers saltcedar alone. Of these few the papers by Gatewood and others (1950), Blaney (1957), Dylla and Muckel (1964), McDonald and Hughes (1968), Robinson (1970), Culler (1970), Culler and others (1970), Hughes (1972), and McQueen and Miller (1972) are probably the most pertinent. THE BUCKEYE TEST SITE LOCATION, SOIL, AND WATER The study site and its characteristics have been de- scribed in detail elsewhere (van Hylckama, 1974), and only a brief statement is required here. Figures 2 and 3 show the location and the layout of the project. Figure 4, an aerial photograph of the area, and figure 5, a photograph of the test site as seen from a 4-meter-high platform, show the density of vegetation on the flood plain at the experimental site. In choosing a location for evapotranspirometers an important consideration is the extent of the vegetation surrounding the proposed area. As a rule of thumb, it is assumed that homogeneous vegetation should extend over a radius 100 times the height of the vegetation under study (see Monteith, 1973, for a detailed discus- sion). As figure 4 shows, the vegetation does not extend that far to the south, and the map of figure 3 indicates such a desirable fetch only in a southwesterly direc- tion. However, under natural conditions saltcedar pre- dominates in this region as flood-plain vegetation along rivers and washes or as strips of riparian vegeta- tion along lakes and reservoirs. Especially when winds blow normal to rather than parallel to the river, the fetch will quite likely be too short and oasis effects are F5 R 3 W 112°31' LOWER RIVER ROAD 33°21 ® 4 4/17 W H \\\Q: Kg > I \ #" I | o 1 KILOMETER 0 3000 FEET ARIZONA COLORADO RIVER Salt R FiGuUuRE 2.-Location of the Buckeye test site. to be expected. The large rates of evaporation one finds in the literature are most likely due to these facts of nature. Whereas then the location of the evapotranspirome- ters at the Buckeye Project was as good as one can get in an area of flood-plain vegetation, and whereas the instrumentation was quite adequate to provide reliable results, the construction had one big drawback: no pro- visions were made for draining the tanks. The conse- quences of the resulting increase in salinity of the ground water in the tanks have been discussed elsewhere (van Hylckama, 1974). However, when sa- linity was abated by proper flushing, reliable water- budget data could be derived for short periods of the order of days and hours as will be shown later on. Even though flushing made possible short-period de- termination of evapotranspiration, one can never be sure how representative such evapotranspirometers are. The Buckeye tanks probably gave results that would be valid for the flood-plain vegetation along the Salt River between Tempe, Ariz., and the confluence with the Gila River, from that confluence a few tens of F6 Dense saltcedar /§\\v/h\\_//——~\\//\//\/ lel \ Bare ground / ) o Low fence L o lel el N $ o 120m 4 © o R 7 Ts nly 85 28 C/S g) 2 o 3) o $5 © Bp o [ml a. Platform (e] w Oo & Office, pump, and storage f Low fence Cotton field o 50 METERS I 0 o 150 FEET I FiGur® 3.-Sketch of the project site, southeast of Buck- eye, Ariz., showing position of main features. P, R, W are instrument masts for wind and temperature profiles, radiation, and wind direction respectively; M and O, ad- ditional masts. Numbers 1 through 6 indicate evapo- transpirometers forming part of the saltcedar thickets, constructed in 1959. Tanks 7 through 11 (not indicated in the figure) were constructed in 1962 in the area marked "bare ground." Open circles indicate location of access tubes to determine soil moisture outside the tanks. kilometers upstream along the Gila and downstream to Gila Bend, Ariz., provided depth to ground water and soil-moisture salinity are taken into account. The soil at the project site is a sandy to fine sandy loam belonging to the Entisol soil order. It consists mainly of alluvial sediments deposited by floods of the Waterman Wash. Table 1 presents some physical and chemical characteristics. The physical ones are av- erages of 12 samples taken from evapotranspirometers number 1 through 6 (see fig. 3). The chemical ones are presented as the range of samples taken from the 6 tanks in 1963 and as the range of samples taken from all 11 tanks in 1966. The large range and the changes between the 1963 and 1966 data are mostly due to the treatments of the tanks, such as flushing, change in quality of water, and so on. Water for irrigating tanks and surroundings came STUDIES OF EVAPOTRANSPIRATION FIGURE 4.-View from an airplane of the project site shortly after planting the first five tanks (1959). Camera looks to the north. Plies Huge - FiGURE 5.-View of the project site in 1963 as seen from a 4-m-high platform. The main instrument mast is visible right of center. White Tank Mountains north of Buckeye in the left background. Camera looks to the northeast. Arrow points to the mast carrying radiation instruments. See fig. 19. from a well drilled to a depth of 25 m in 1959 and deepened to 43 m in 1960. The change in quality due to this procedure is clearly shown in the data of table 2. CLIMATE The climate at the project site is arid. Mean annual temperature at Buckeye is 20.6°C and mean annual rainfall is 190 mm. The monthly mean temperatures at the project site do not vary much from year to year. The means of the 6% years of operation are presented in figure 6, which also shows the long-time means for the town of Buck- WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F7 TABLE 1.-Analyses of soils at the Buckeye project [USWCL: United States Water Conservation Laboratory, (USDA), Tempe, Arizona, U of A: The University of Arizona, Agricultural Experiment Station, Tucson, Arizona; USBR: United States Bureau of Reclamation, Phoenix Development Office, Soils Laboratory, Phoenix, Arizona. Physical characteristics are expressed as mean percentage of 12 samples and standard deviations. Moisture characteristics are expres TaBLE 2-Analyses of irrigation water at the Buckeye project [All analyses made by the U.S. Geological Survey. All values in milligrams per liter except where otherwise indicated] Moisture characteristics Hz20 b volume. One millibar = 100 N/m = 100 Pa; a = standar deviatiiilpercentage s June 9, _ May 11, _ Feb. 20, _ Apr. 26, May 6, 1959 1960 1961 1962 1963 Mechanical analysis of sandy loam, USWCL SiOg ce- 27 31 46 35 31 Calc. ics donate n eae 290 475 505 475 396 Classification Particle diameter Percent MB econ e-- 39 91 126 125 122 Na 512 960 1200 1250 1150 clay less than 2 um 7.0 £ 0.8 | cnl 207 366 416 403 405 silt 2-20 um 9.2 + 3.0 SOjec .or irene efor doe coded 265 662 836 878 836 silt 21-50 um $15% 30 | C! 1140 1880 2250 2250 1980 sand over 50 um 59.2 + 7.4 Total dissolved solids .__.____-- -- 4280 4860 5210 4720 Si Hardness as CaCO; -_-__--- 880 1560 1780 1700 1490 s f PHX reese ree 74 7A 1.5 7.3 6.8 Chemical analysis Specific Chemical sample Depth U of A USBR conductivity (cm) 1963 1966 umho/em Ca T Me Megioo a ___ ap E- 03 1.9 At a c.: ..4.30 6.97 8.21 8.32 7.57 75 cow. 0.3-1.3 125 0.2-2.0 Na + K (Meq/lOO g) 25 1.5.7 0.6-4.7 40 [ T: | | | | | | | | 75 1. .8 0.8-4.4 125 1.6-3.5 1.0-5.3 <3 Ci(msAj 25 _ 1060-1850 e20-5950 | 2 75 530-2530 1110-7650 8 125 200- 830 1280-9650 u u Total dissolved E solids1g/L) 25 werk 1.7-14.0 iD 75 26-136. | 8 125 yo#o7 |'z .. l... 25 7.9-8.1 7.5-8:1 CC 75 7.9-8.1 7.5-1.9 2 125 8.0-8.1 7:0-7 8 ® CC W a. 3 u m Millibars Percent & NEA Aa -il ro te eon 40.0 2.6 1024... 33.5 2.2 SO yar Us dics s canton 29.2 1.9 120 nre 26.9 2.4 1602-2. ge 24.7 3.8 O0 HE alena aaron 22.2 4.7 5002.0 eno 17.3 4.1 S002 reas en ien in 13.2 3.2 1000 >-. 10.6 2.2 eye according to Weather Bureau data (U.S. Dept. of Commerce, 1955). Maximum temperature at Buckeye may reach 49°C, usually in July and August, but at the test site, tem- peratures as high as 52°C have several times been re- corded and, whereas the lowest temperature at the Buckeye weather station is given as -10°C, tempera- tures as low as -12°C at the test site were not uncom- mon. In winter, the above-ground plumbing had to be protected with electrical heating coils. Wind directions are highly variable. Figure 7 shows wind roses for the calendar year and for the growing seasons at the project, compared with long-term av- erages observed at Sky Harbor Airport east of Phoenix, | J.. _ F fof __ sl _ L _-L _ A -_-L _- 1 v_ 'A M' } J_ A 's. 0 . N -D FicurE 6.-Mean monthly temperatures (1961-67) at the project site and 30-year means at Buckeye, Ariz. less than 80 km to the east of the test site. It is clear that the main directions at the airport are E and SE, while those at the project are NE and SW or. W. This difference is undoubtedly due to the effect of the mountains to the north and east of the project site. The wind directions also vary considerably by time of day as shown in figure 8. The significance of these wind directions will be shown in the discussion of the advec- tive term in the energy- and mass-transfer equations. Windspeeds also are quite variable. The monthly means for the 2 years for which complete data were available are given in figure 9; the data are compared with long-time averages observed at the Sky Harbor Airport. That the speeds at the airport are generally higher is not surprising because they were measured at the control tower level and those at the project at only 4 m. F8 s STUDIES OF EVAPOTRANSPIRATION Whole year April-September Sky Harbor 28-year mean FiGurE 7.-Wind roses for the Buckeye project site (1962, top; 1965, middle) and for Sky Harbor Airport, Phoenix, Ariz. (1930-58 mean values, bottom). Means for the calendar year (left); for the 6 warmest months of the year (right). Numbers in the centers are mean windspeeds, in meters per second. Rainfall, as shown in figure 10, is quite erratic and scanty although 1965 and 1966 showed annual totals of more than 200 mm each and are higher than the long- term averages given by the broken line in that figure. Solar radiation is intense, as can be expected. Figure 11 shows the observed radiation as a percent of the maximum possible radiation at the project-site latitude; the atmospherical transmission coefficient is assumed to be 0.9. Note the dip in radiation intensity due to thunderstorm activities in the late summer. Finally, the fluctuations in the dew-point tempera- ture, given in figure 12, were computed from observed air temperatures and relative humidities. VEGETATION As can be seen from figure 4, a considerable amount of clearing of the saltcedar vegetation was necessary FIGURE 8. -Mean wind directions for 6-hour and daily periods for the months of May, June, July, and August, 1965. Numbers in the center are windspeeds for the periods, in meters per second. 3.0 |- \ \ __-28-year mean values (deter ced a rgtt--r-r-1-1 2.0 n Ty F T3" 4 Lf -_ mo a _s estee eat ots | J- A: M. ~ J ~J A' $s _ N -D MEAN MONTHLY WINDSPEED, IN METERS PER SECOND FicurE 9.-Mean monthly windspeeds at the Buckeye project site in 1962 and 1965, and at Sky Harbor Airport, Phoenix, Ariz. (1930-58 mean values). before the evapotranspirometers could be installed. But as figure 5 shows, by 1963 vegetation on the cleared area was nearly equal in height and density to the original one. In open places, inside as well as out- side the test area, a number of grasses and forbs grew WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F9 eod 4 1rd - cr cE T T9 d Rares 1961 1962 1963 100 |- = 50 |- PROJECT SITE BUCKEYE, X yp ARIZONA \_ {-- f ~ ams. $l ~ §... 1963 1964 1965 100 |- --I RAINFALL, IN MILLIMETERS PER MONTH ang! 1988 1966 1967 _-PROJECT SITE 50 |- sA BUCKEYE, ARIZONA | _ < mBau+#7 $ 'N J MOM 3. s -N -j M M yj FicurE 10.-Monthly rainfall at the Buckeye project site and monthly mean (1930-58) at Buckeye, Ariz. up from seeds washed in with floods of the Waterman Wash. Some of these are roadside weeds, others are members of the mesquite associations occurring out- side the riparian areas along the lower Gila River and its tributaries. Table 3 gives a list of these plants to- gether with some characteristics. According to Kear- ney and Peebles (1960), quite a few of the species nor- mally grow at higher altitudes. They are washed down in the frequent floodings of the Waterman Wash. It is also noteworthy that none of these plants is known to be an indicator or edificator' (Chikishev, 1965) of soil- or soil-moisture characteristics, except possibly pig- weed, Chenopodium rubrum, and the quail plant, Heliotropium curassavicum, which seem to prefer al- kali or saline soil. Soon after the replanting of the cleared areas (fig. 4 is an example) most of this herbaceous vegetation dis- appeared, leaving saltcedar as a nearly 100-percent dominant species. There is undoubtedly a shading ef- fect, but an important cause of this phenomenon is the capacity of saltcedar to exude salt (Campbell and Strong, 1964; Hem, 1967) through openings in the * An edificator is a plant species that is an indicator of more than one soil or soil-moisture characteristic. «© 0 SOLAR RADIATION, IN PERCENTAGE OF MAX ImUM co 0 I | 1 1 1 1 1 1 1 1 1 1 M- A M =J: J .A 'S -O- N- D ~ le] J~_F FicurE 11.-Solar radiation in percentage of maximum possible; transmission coefficient assumed to be 0.9; plot based on mean values for 1961-67. 20 _ o 0 DEW POINT, IN DEGREES CELSIUS {-l -__- E=~M A 'M j 'J 1 wms o LLL asl A. $8 0 N DB. J LL FiGur®E 12.-Dew-point temperatures at the Buckeye project; plot based on mean values for 1961-67. scalelike leaves that are constructed very much like the water-guttating hydathodes one finds in hydro- phytes and other plants growing under conditions of frequent high relative humidities. Since saltcedar is a halophyte, such glands could well be named halohodes (van Hylckama, 1966). The exuded crystals are hygro- scopic and on cool mornings tiny drops are formed on the leaflets. This salty "dew," from which the saltcedar gets the first part of its common name, falls on the ground and kills any emerging seedling. One rarely finds an intruder in a saltcedar thicket. The conductiv- ity of the soil-moisture extracts taken from the sur- faces of the soils at the project site was often as high as 45 mmho/em at 25°C (millimho per centimeter), or 45 mS/cm (millisiemen per centimeter) at 298°K. Another point of interest is the genotypical variabil- ity of saltcedar. The plant can have the appearance of a small tree, as along the lower Colorado River near Yuma, Ariz., or it can take on a more shrublike ap- pearance as at the Buckeye Project site, or the plants can be tall but spindly, as along the Rio Grande near Bernardo, New Mex. The color of the flower can also vary from red through shades of pink to near white; the time and character of bloom can also vary greatly. Var- iations are mostly due to genetic characteristics and not to environmental factors (J. S. Horton, written commun., 1977). F10 STUDIES OF EVAPOTRANSPIRATION TABLE 3.-Plants (other than saltcedar) observed growing at the Buckeye project site, 1961-66 [Abbreviations used in columns are the following: For value of plant: W symbolizes weed and F forage plant. For abundance: a, abundant and c, common. For usual habitat: as, alkali or saline soils, dr, dry areas; gs, gravelly soil; ma, moist areas; rs, road sides; ss, sandy soils; wl, waste lands; ws, washes. For origin of plants: E, European; I, introduced from places other than European; and (-), no specific information available. Nomenclature follows Kearny and Peebles (1960)] Plant Value Quantity Habitat Origin Annuals and Biennials Grasses: Eragrostis cilianensis (love W a wl E Schismus barbatus (six week grass) ..-__.___._......_;_L_LLLLLc..- F a ss E Polypogon monspeliensis (rabbit foot grass) _______________________-- F c ma I Echinochloa colonum teock spur) F c ma E Forbs: Eriogonum deflexum (skeleton weed) - a ss - Chenopodiuni incisum (goose fool). " - - dr - . Chenopodium rubrum (pig w - as Chenopodium leptophyllum (goose foot) ________________________-- - - dr - iSalsola kalt (Russian Ww - wl Acanthochiton wrightit (green stripe) F - dr - Abronia villosa (sand verbena) cc lee uss - - ss - Argemone intermedia (prickle poppy) - - ss - Tribulus terrestris iL W a rs E Mentsclia albicauits (blazing star). - - dr - Nama 2. .L. ICA UIL Yage aaerdan bn naan as cea ble nene s - - ss - Amsinckin tessellata (fiddle neck) ._.-_._._.._.______LLL__LLL_____ - - so - Plantago purshit (plantain) - a dr - Verbestna encelioides (crown beard) .._.__-.___._._.._.___________Q_ - a wl - Paloforia nears _ (en- ai and - - ss - Pectis papposa (fetid marigold) ci - - gs - iSonchus pleraceus {sow thistle) 0000. W - wl E Perennials Grasses: Cynodon dactylon (Bermuda grass) ___._.__.______.__LL____________ W a - - Sorghum halepense (Johnson grass) Ww a wl Forbs: © ' Sphaeralcea ambigua (globe mallow) ____________________________ - - dr - Petalonyx thurberi (sandpaper plant) ._-._._._.___;_..____._._____ - - ss - Mentgelyia pubertl@ {SLICK 1G@f) =d aes s - - gs - Heliotropium curassavicum (quail plant) - - as - Solanum elaeganifolium (horse nettle) ___________________ccocc___- - - ss - Nicotiana irigonophylla (LobBacto) . - - ws - Shrubs: Prosopis Jultflona (mesquite) = c ws - (palo verde) - - ws - Lycium fremontit (desert thorn) - - ss - Encelia faringsa (brittle bush) - - gs # EVAPOTRANSPIRATION AND ITS CALCULATION Although evaporation, transpiration, and evapo- transpiration all involve the same process, namely the change of water into vapor, a distinction should be made between evaporation and transpiration. Such a distinction has not always been made. For instance, Thornthwaite (1948), Penman (1956), and many other workers, at times considered vegetation to be a mere inert transporter of water or vapor, but later inves- tigators found that at least some plants exercise a pro- nounced control over the rate of their transpiration. See also Monteith (1963), Tanner (1963), and (for lively discussions) Lee (1967; 1968a; b), Idso (1968), and van Bavel (1968). EVAPORATION Originally two methods to compute evaporation in- directly were tested over open water surfaces. Both methods consider the transformation of water into vapor as a surface phenomenon, similar to what hap- pened to the water on Aristotle's plate when exposed to sun and wind. If the plate is covered and remains at a constant temperature, the number of molecules leav- ing the water surface eventually become equal to the number entering the surface, at which time the air above the water is said to be saturated with water vapor. Any increase in water molecules in this space would cause condensation and the moisture would fall back to the water surface. If the cover is removed and wind blows over the plate, vapor is taken away. The original balance is upset, an excess of molecules leaves the water surface and the supply is depleted at a rate which (among other things) depends upon the speed of the air moving over the surface. This in brief is the principle of the aerodynamic method. On the other hand, assuming that there is no wind but that there is a change in temperature, (which WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA under natural conditions would be caused by the sun's energy), the water is heated, the molecules move faster and more leave the surface. The air in turn also be- comes heated and capable of absorbing more water molecules since the solar heat represents an incoming supply of energy which is transposed into the energy represented by warmer water and warmer air. Compu- tations based on this principle are called energy- budget methods. Thus, if the wind during the mass-transfer process does not heat or cool and if the air flowing over the surface is already saturated, no evaporation will take place due to the movement of air. Under such circum- stances one could use an energy-budget method pure and simple. Under natural conditions, however, one must combine the principles of energy budget and aerodynamics if one desires to arrive at reliable re- sults. This is especially true in areas where the air is dry and the wind very often strong-in other words, under conditions typical for the Buckeye Project and for the vast majority of areas in the arid and semiarid southwest. TRANSPIRATION If evaporation takes place through the surfaces of plants, it is called transpiration. In most plants grow- ing in or under water or growing under very moist conditions, much of this transpiration takes place through the epidermis and its waxy covering, the cuti- cle (from the Latin cutis = skin). This water loss through the cuticle of stems and leaves is called cuticu- lar transpiration. But in most land plants the outer surface is covered with dense hairs, with heavy layers of wax or, on stems, with corky bark. Under such cir- cumstances cuticular transpiration accounts for only a small fraction of the total. Most transpiration then oc- curs through little openings in the leaves called stomata-plural of the Greek word roua« (stoma = mouth)." It is through these openings that during day- light carbon dioxide is taken up for the process of photosynthesis and water is transpired. It should be noted that a small exchange of carbon dioxide and vapor can also take place through much larger open- ings (lenticels) in the outer surface of branches and stems. The size of the stomata varies with the plant species but those of saltcedar measure, when open, about 7 x 20 um with an open-pore size of about 60um* (Campbell and Strong, 1964). Figure 13 shows schema- tic drawings of a stoma and a typical distribution of stomata over a leaf surface. Unlike most deciduous * Saltcedar can also exude some liquid water through hydathodes at the end of vascular bundles in the leaves. Hydathodes in saltcedar usually are found about 1 mm below the leaftips (Kisser, 1956). For the salt glands, see section on "Vegetation." F11 100um FigurE 13.-One stoma in surface view (top left) and in cross section (top right); view of some stomata distrib- uted over the underside of a leaf (bottom). The dotted lines in the cross section (arrow) indicate position and shape of the guard cells (a) in closed position. Schemat- ic after Fitting, Schumacher, Harder, and Firbas (1951): a, guard cell (one of two); b, epidermis cells with cuticle (c); d, nucleus in cell; e, substomatal air space; f, parenchyma (thin-walled tissue cells); g, open stomata. Dots in cells indicate chloroplasts where water and carbon dioxide are combined to form sugars with the help of solar energy. trees, which usually have their stomata only on the underside of the leaves, the scalelike leaves of saltcedar have stomata all over the surface. These leaves are about 1.6 mm long and somewhat egg shaped; the largest diameter is about 0.8 mm. Since there are about 250 stomata per mm" (Tomanek and Ziegler, 1960) it follows that at most about 1.5 percent of the leaf surface is open for the passage of water vapor. Stomata react differently to ambient influences, de- pending on the species of the plant. Some stomata close partially or completely at night, others (for instance pineapple, Ananas sativas)® are wide open during that time. In some species they open during periods of high ©Cacti and other succulents are capable of taking up CO, during the night with little loss of water. Then, during times of sunshine, the CO; is photosynthetically combined with water and other chemicals in the plant cells to build plant material (Walter, 1973). F12 wind, in others they close, but all close when the plant is under water stress, at which time the turgor in the guard cells (see fig. 13) diminishes, resulting in closure of the openings. FLOW RESISTANCE One can now visualize a series of resistances against the movement of water and vapor. Disregarding the resistances against water movement in the soil, from the soil to the roots, and from the roots through the plant, which do not concern us here, it is intuitively clear that a resistance develops when water moves through the thin-walled parenchyma cells surrounding the substomatal cavity and vaporizes. The vapor then has to pass through a comparatively small opening to reach the ambient air. This is called the stomatal resis- tance (Monteith, 1963; Smirnov, 1963). Just outside the stomata and along the surface of the leaves is a layer, a few tens of molecules thick, where the flow is only laminar and the diffusion molecular, causing an external resistance. Further resistances occur within and above the canopy of a vegetation. It seems clear from the above that transpiration must be considered a phenomenon decidedly different from that of evaporation. But because evaporation from soil surfaces and from wet leaves (after dew or rain) can occur simultaneously with transpiration, the term evapotranspiration is commonly used when the loss of water from vegetated surfaces is considered. In the following pages it will be shown that the in- clusion of a resistance term in evapotranspiration equations used at the Buckeye project leads to water- loss data that fit the measured ones better than when no resistance term is used. In the discussion of the methods used to estimate evapotranspiration only the principles are considered. There are many variations proposed and in use. Only those tested at the Buckeye Project will be mentioned during the data analysis. VAPOR-FLOW METHODS Two methods are considered to belong in this group, and they differ greatly in approach. One, the atmos- pheric water-balance method, measures the amount of vapor that enters and leaves a tall column of air over a large area. The other, the eddy-correlation method, measures vertical vapor flow as close as possible to the evaporating surface. THE ATMOSPHERIC WATER BALANCE Basically, the estimation of evapotranspiration by the atmospheric water balance method is analogous to STUDIES OF EVAPOTRANSPIRATION the soil-water balance method. For the latter we have the equation: P:=RO +E + AS (DP in which P is precipitation, RO is runoff, E is evapora- tion, and AS is the change in soil-moisture storage. Deep percolation and underground drainage are ne- glected. Instead of considering a block of soil with the surface at its upper boundary, one can consider the volume of air with the soil surface at its lower boundary and the top, say, at the 500 millibar level (50 kPa [kilopascal], about 5.5 km above sea level). Then AS becomes a change in the quantity of moisture in the air which can be due to evaporation from soil or plants (vertical com- ponents) or to moisture entering or leaving the air col- umn more or less horizontally. This "horizontal" com- ponent is an advective term and has to be considered not only in the atmospheric water balance but also in the combination methods to be discussed later. With this in mind, one can write: E =P =- (V -M + AW) (2) in which V -M is the difference between the moisture in the atmosphere entering and leaving the column. The quantity V M is usually called "divergence" if it is positive and "convergence" if negative. It can be meas- ured by radiosonde observation. The change in pre- cipitable water (analogous to AS in equation 1) is de- noted by +AW. Since precipitation in this procedure is considered to be a measurable quantity, £ can be com- puted and compared with data derived from more di- rect measurements such as lysimetry. Taumer (1955) was probably the first one to make complete investigations over a 10 000 km area for in- dividual months and to compare meteorologic results with lysimeter data published by Kalweit (1953). More recently Rasmusson (1967; 1968) and Malhotra (1969) used the dense radiosonde network over the North American continent (an area at least 1000 times as large as Taumer's) to compute monthly values. It is obviously difficult to check the results of this type of studies against those obtained from conventional methods and one must resort to more esoteric methods such as seasonal changes in latitude or in the wobble of the earth caused by the movement of large air masses (Munk and McDonald, 1960; van Hylckama, 1970b). It is therefore not surprising that an attempt to apply the principle of equation 2 at the test site failed. The distance (92 m between the masts O and M in fig. 3) was far too short and the height of the sensing ele- ments (2 m above the vegetation) was far too small to enable the successful integration of windspeed, tem- WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA perature, and vapor-pressure data for the computation of an increase in water vapor due to transpiration. THE EDDY-CORRELATION METHOD Since evaporation and transpiration are basically surface phenomena, the first from the surfaces of free water and the second from the surfaces of plants, it is logical to measure these phenomena as close to the surface as possible. But technically this is difficult to do. However, during the last 20 years sensing and re- cording equipment has been greatly improved (L. J. Fritschen, oral commun., 1977). Although the Buckeye Project was not equipped for this model, it shall be mentioned here because of its promising potential. The eddy-correlation method can be considered the most fundamental approach to the measurement of evapotranspiration because the turbulent air movements that give rise to eddy diffusion are meas- ured together with the fluctuations in humidity. An eddy can probably best be visualized as being a parcel or "glob" of air with a long or short history of existence of its own. Brunt (1952, p. 219) did not con- sider it likely that any definition could be given which would be universally acceptable. He wrote, The eddies which form the edge of a stream flowing into a mill pond are of the nature of vortices with vertical axis, but the name 'eddy' as used in discussing motion in the atmosphere is not restricted to circu- lar motions. We can only define eddy as a physical entity which disturbs the uniform flow of air and this definition will include rotat- ing eddies, convection currents, and any other type of disturbance. This method, also known as the eddy-transfer method, rests on the idea that a moving parcel of air carries with it, on its journey up, down, or sideways, the heat, water vapor, dust, and momentum it con- tained at the start. That this is not necessarily so is one of the difficulties in applying this method and some other methods of estimating water losses or evapo- transpiration, as will be discussed later on. If measured close to the surface and with sufficient frequency (of the order of once per second or more of- ten) the mean upward flux of humidity (or any other property of the atmosphere) per unit mass can be given as: F = paw (3) where p,, is the density of the air, w its vertical velocity, q its humidity, and where the bar indicates the average condition during a selected period. Now any change can be expressed as an average and the deviation from that average during such a period. Hence: Prs bi rp, w =w + w'. and g -g +9 -_ (4) F13 where a prime gives the instantaneous deviations from the mean. Thus we can write: F = (p, + pi) (w + w') (q + q') (5) which can be expanded to: F = fiflwé+fiawa+bnqu+fiawhl rR a Ph 1 1 15 1 1 1 (6) +pu0q + puog' + piw'q + piw'q Because the sum of the deviations from the mean is zero by definition, all terms in this equation that have a single prime go to zero but the product of two such primed quantities does not. With this in mind, and also neglecting density fluctuations which are small close to the ground, we have: F = + paw'g' (7) The effective rate of upward diffusion of water vapor past the measuring point is now given by w'q', which is called the eddy flux, and this can be obtained by sepa- rately measuring w and q. These values can be multi- plied and integrated over a measuring time period, and the product of the individual mean values over that period can be subtracted. The instrumentation for direct measurements of the eddy and its properties has been under development since the work of Swinbank (1951). (Also see Perepel- kina, 1959.) These instruments use hot-wire anemometers and fine-wire dry and wet thermocouples and even more sensitive instruments are under study (Dyer, 1961; Goltz and others, 1970; Hicks and Good- man, 1971). The fact that enormous quantities of data have to be converted into required output over periods of 5, 10, or more minutes and the resulting require- ment for integrating circuits, analog computers, or di- gital microprocessor networks, makes this method very expensive. This discussion of the eddy-correlation method is in- cluded here to emphasize its importance. It is the only method that is independent of the characteristics of the surfaces over which measurements are made. The in- strumentation can be easily moved about and installed for any desired time over any surface to determine evapotranspiration rates and, if suitable, to provide ground-truth data for remote-sensing (including satel- lite) studies. BULK-AERODYNAMIC METHODS There are two methods that belong in this category: the Dalton approach and the profile approach. F14 THE DALTON APPROACH This method is so called because it was Dalton (1802) who recognized the principles of the atmospheric vapor flow close to a surface at which evaporation occurs. He reasoned that vapor transfer had to occur along a gra- dient of moisture concentration and that the efficiency of this concentration depends on the turbulence of the air (hence on the windspeed) above the surface. In other words, E, = fM) (€o - 2) (8) in which f(@) is a function of the mean windspeed, often given as: f(x) = a (1 +bu) (9) or = cu (10) E, is the evaporation rate, e, and e, are respectively the pressures at the evaporating surface at height z = 0 and in the air above it (for practical purposes 1 or 2 m), & is the average windspeed, and a, b, and c are con- stants." Several investigators have given numbers to these constants, for instance Rohwer (1931) has: o = 0.4 (1 + 0.27 hi,) (es - e,) mm/day_ (113) where @, the mean wind at the surface, is in miles per hour and the vapor pressures are in mm Hg. Penman (1948) modified this to: E0 -= 04 (1 + 017 122) (en Cs ell) mm/day (llb) with & at 2 m above the surface. Other variations are given by Kuznetsov and Fedorova (1968). Harbeck (1962) developed this variation for estimating evapora- tion from reservoirs: E, = NU (65 - ta) (11¢c) where N is a coefficient related to the reservoir's sur- face area. Since the variables are comparatively easy to measure close to the evaporating surfaces, average values over periods of half an hour or longer can be obtained. The method is an integral part of the combi- nation appraoch which will be discussed later. "The use of two constants in the wind function a(1 + b) can be considered an admission of the fact that many cup-type anemometers have considerable starting and stalling speeds, that is, it takes a comparatively strong wind to get them going and they keep turning for a considerable time after the wind has died down (G. E. Harbeck, Jr., written commun., 1955). STUDIES OF EVAPOTRANSPIRATION THE AERODYNAMIC OR PROFILE APPROACH The method is called profile approach because it con- siders temperature, humidity, and wind speeds at two or more levels above the surface, rather than at the surface alone. This sounds like a simplification, but the shape of the profiles depend so much on the stability or instability of the air that, unless the wind profile is logarithmic, large deviations from observed evapo- transpiration values can be expected. In the aerodynamic methods, we distinguish three vertical fluxes which result from the turbulent diffu- sion in the lower layers of the atmosphere. They are 7, the flux of momentum, A the flux of sensible heat, and V the flux of water vapor. They are expressed as fol- lows: T = PaKn Ju_ (12) Oz A = pit; Ka OT __ and (13) Oz ¥ # pk. * (14) Oz The K's are exchange coefficients respectively for mass (m), heat (A), and vapor (v); c, is the specific heat for dry air; T is the temperature; and the other symbols are as defined before. Any other type of atmospheric flux can be expressed in a similar manner. For instance, for carbon dioxide flux, F., we would have ac F,. B az (15) where C is the CO, content as mass per volume. Intui- tively it seems reasonable to assume that the exchange coefficients (K) would be the same for all four equa- tions. That this is not always so causes difficulties in applying the aerodynamic approach. For instance, as- suming that K,, = K, and that the wind profile is logarithmic, Thornthwaite and Holzman (1942) pro- posed the following equation for evapotranspiration (E): (g> - qi) (ug - ui) 16 [1n(Zz/21)]2 GO" E = pak" where k is von Karman's coefficient, often taken as 0.40 or 0.41, (see section on "Winds"). Later, other workers (for instance Pasquill, 1950 and Pierson and Jackman, 1975) attempted to correct the aerodynamic WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA equation for conditions of instability. The methods then became very complex, requiring very accurate and frequent measurements of windspeed, specific humidity, and temperature at three of more elevations above the surface. Webb (1965) presents a good review of this method. ENERGY-BUDGET METHOD Penman (1956, p. 16) starts his discussion of the en- ergy balance as follows: "The fundamental basis of the energy balance approach is unchallenged: the chal- lenge is to our ability to measure or estimate all the quantities needed to exploit the principle of the conser- vation of energy." The energy-budget, or heat-budget method, is one of the theoretical methods based on the principle of the conservation of energy. Evaporation and transpiration are essentially the same thing. In both cases about 2450 joules (585 calories) are needed to evaporate 1 g of water at summer temperature. The energy to do this comes directly or indirectly from the sun. At the top of the atmosphere the earth receives on each square meter perpendicular to the sun's rays about 1400 W or 2 cal cm-" min~', but a large part of the radiation is reflected back into space by dust, water vapor, and clouds; another part is scattered into the air; and a third part heats the soil and therewith the air. How much reaches the plants and the ground de- pends on latitude and season, on weather conditions, on elevation of the place, and on its exposure. The quantity may vary from practically zero under heavy overcast to more than 1000 W m-" on high mountains above the clouds. If we now measure all those different types of radia- tion, incoming and outgoing, and also the changes in temperature in the soil and in the air, and take into account that at times considerable quantities of hot air are brought into the test site by advection, we can set up a balance sheet. On one side we have all the incom- ing radiation, on the other the sum of all that goes out plus what is stored in the soil, in the air, and in the vegetation. If we do this, we find that the two sides do not balance. The difference is that part of the radiation which is used to evaporate water either from the sur- faces of lakes or from moisture on and in the soil, or from water on and in the plants. Thus we have a method of indirectly measuring the amount of water transpired, which is the only item in the balance sheet that we cannot measure directly. If the energies consumed in photosynthesis or liber- ated by metabolic processes are neglected we can write: R,, = LE + A +G (17) F15 in which R,, is the net radiative flux received by the surface, G is the heat flux into or out of the ground, A and LE are the sensible heat flux and latent-heat flux into the air respectively, and L is the latent heat of evaporation. In this equation R, and G can readily be measured, but the partitioning of the energy between evaporation and heating of the air has to be approached indirectly. The ratio A/LE, usually called the Bowen ratio (B), was found by Bowen (1926) to be: Tn r Tn > C p f r as [ 2-4 ] 1000 (18) where c, according to Bowen, varies between 0.58 and 0.66 depending on the state of the atmosphere. Note that this ratio was used to compute evapotranspiration over open water. T, and T,, are the temperatures of the water surface and of the air above it; e, is the satura- tion vapor pressure at T,, and e,, the actual vapor pres- sure at the elevation of T,. Over Lake Hefner (U.S. Geological Survey, 1954), where T,, and e, were meas- ured at 2 m above the surface, single-day average val- ues were found to vary from -20.02 to +31.50 al- though most of the time they lay between -O0.5 and +0.5. From the flux-gradient equations discussed (13 and 14) it follows that, if K, = K,, and atmospheric pressure is assumed to be constant, B can also be expressed as tess £361 AT B = or = y -- 19 L K, aq Yy ke (19) Oz where y is the psychrometer constant, which has a value of about 0.65 °C-'. If vapor pressure (e) differ- ences are used instead of specific humidity (q) values the dimensions are mb/°C. If B can be considered con- stant with height, the energy balance can also be writ- ten as: Ean—G LE _R”—G 1 FP A £LE |: _ 4. 1 + B or LE = R,,/ (1+$8) if G can be neglected. Table 4 shows the values of yAT/Ae for a 24-hour period at five different paired levels. The values are quite erratic, although most of the time when they are negative or positive together, they indicate some con- sistency with height. However, there are large de- viations: for instance between 0300 and 0400 the val- ues range from -0.78 to -13.65 and between 1400 and 1500 from -0.07 to +0.33. (20) F16 STUDIES OF EVAPOTRANSPIRATION TABLE 4.-Values of y (AT/Ae) for a 24-hour period, May 5 and 6, 1966 [AT and Ae are differences in temperature (°C) and vapor pressure (mb) between a lower and higher level; y is the psychrometer constant taken to be 0.65 mb/°C. The vegetation is 3 m tall and forms an even stand] Instrument heights above ground in meters Start of measuring periods (time in hours) 11 and 7 11 and 5 7 and 5 7 and 4 5 and 4 5 -2.92 2.00 -1.95 -3.25 -4.55 6 =-2.B1 =-2.31 -1.30 -2.16 -3.90 ¥ 1.30 0.65 0 -0.65 -1.30 8 0.21 0.26 0.33 0.21 0 9 0.07 0.08 0.09 0.21 0.39 10 0.37 0.49 0.65 0.81 1.09 11 0.05 0.25 0.51 0.56 0.65 12 0.33 0.23 0.12 0.47 1.02 13 0.33 0.33 0.65 0.31 -0.18 14 -0.07 0.08 0.26 0.29 0.33 15 0.52 0.14 -0.33 -0.05 0.39 16 0 0.17 0.39 0.10 0.21 17 0.37 0.12 -0.33 -0.49 -0.65 18 -0.54 -0.59 -0.65 =0.92 -1.63 19 -0.46 -0.49 -0.52 -0.86 -1.30 20 -0.65 -1.03 +1.56 -1.54 =1.51 21 -1.69 -1.37 -0.98 -1.30 =1.178 22 -0.65 -0.91 -1.30 -0.43 -1.30 23 -0.65 -1.02 =1.51 -1.78 -2.60 24 -8.58 -4.11 -5.20 -3.25 -1.80 1 -2.44 -1.95 -1.46 -1.51 -1.63 2 -5.20 -8.90 -2.60 -2.81 -8.25 3 -0.78 -0.92 -1.30 -5.41 -183.65 4 -6.83 -6.28 -5.20 -3.90 -2.60 The effect of the value of B on the computed LE val- ues can be shown as follows: Assume R,, = 1 cal cm- min~' or 60 cal em- hr-' (a nice summer mid-day value). Then if 0. :=-1.0 =0. t *~ 2 2 56 ~0 10 : 0.1 - 0.5 +20 1.1 10 "20 ; 1.0 .091 0.67 0.50 0.33 where ? indicates that the LE value is undetermined. Of course, during daylight, positive values are to be expected since temperature and vapor-pressure differ- ences are nearly always negative. Nonetheless, as B increases from 0 to 2.0, LE diminishes from 1.0 to 0.33. Such differences can be avoided by integrating tem- perature and vapor pressure data over a large area and avoiding instrument bias by frequently changing the position of the sensing devices. Also, observations have to be made very close to the transpiring surface. Suomi and Tanner (1958) (p. 301) present a fascinating de- scription of such instrumentation and mention that "on occasion plants even brushed against the bottom sen- sors". At the Buckeye Project the system was rigid, and the lower sensor was placed 1 m above the top of the vegetation. These are the main reasons that the energy-budget, using the Bowen ratio, could not be used at the Buckeye Project. COMBINATION METHODS The instrumentation seemed better adapted to the use of a combination of the energy-budget and a mass- transfer term. Penman (1948) combined the energy balance and the aerodynamic balance into the follow- ing equation: .. (Aly Hy + E4 £s Aly +1 dae. in which the saturation vapor pressure at the surface replaces the actual vapor pressure. The actual vapor pressure is very difficult to measure. In this equation, E, is the potential evaporation from open water; H, is the net gain of radiation at a water surface; E*,, = f (u) (e, - e,) (see equation 8); and A is the slope of the curve when saturation vapor pressure is plotted against tem- perature. Modifying this method, van Bavel (1966) de- rived the following expression for the instantaneous evaporation rate: 2 Aly H + L Bd,, min-. . (22 LBs Sy Ti cal cm~*min (22) in which B,, is defined as B,; = _pae R etic oce gcm min- 'mhb ' (23) p [In(e/zo)]" where e is the ratio of the molecular weights of water and air, (0.622), z, the roughness parameter, H the sum of energy inputs at the surface (except A and LE), and d,, the vapor pressure deficit (e, - e,), with e, the sat- uration vapor pressure. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA This equation has been applied in the analysis of the Buckeye data and often has been given values at vari- ance with the measured ones. Better agreement can be obtained if resistance terms are incorporated in the equation. Monteith (1963 and 1965) showed that the external resistance can be expressed as r,, = fin (G. /u,k*" sec cm -' (24) and the stomatal resistance as r, = (Aly + 1) (E,/E, - 1), see cm-' (25) where E,, is the actual evapotranspiration as measured with a lysimeter, for instance. Because the above equation contains both E, and E,,, the stomatal resistance can be computed from mea- sured values of these terms. The values so obtained can then be used to compute the evapotranspiration rates according to the following equation (the equation con- tains both resistance values): (A/y) H + [(pe)ip] (da/r.) Aly + 1 + rgr; (20) E0 an l/L Examples of the results will be given later. INSTRUMENTATION Basically the instrumentation chosen to measure the various components in the aerodynamic and energy budgets was the same as that used in earlier investiga- tions by the U.S. Geological Survey at Lake Hefner and Lake Mead (U.S. Geol. Survey, 1954; and Harbeck and others, 1958) and described in detail by Anderson, An- derson, and Marciano (1950). These studies, however, dealt with evaporation from open water, and for studies over tall vegetation certain adjustments had to be made. The characteristic high temperatures and strong winds of this area frequently produce blowing dust and dust devils. This dust made it highly unlikely that re- liable results could be obtained by the use of evapora- tion pans or a Cumming's radiation integrator. These instruments, therefore, were not installed. Unlike the lake investigation just mentioned, where recording milliammeters were used, strip-chart re- corders were used which had built-in amplification and automatic zero references, so no ice buckets or Dewar flasks were necessary. The two kinds of recorders used printed a record every 60 seconds, sampling in sequence the electrical output from a series of 12 or 16 different sensors. An example is given in figure 14. F17Z I I FiaurE 14.-Example of a 16-point strip chart about % of actual size, with date and times. The numbers, millivolts, must be converted into degrees Celsius by means of a template or by graphical means, because the relation between millivolts and degrees Celsius is not linear. The instrument prints four symbols in four colors (not shown here). Examples: Arrows 1 point to brown crosses, the dry- bulb temperature 11 meters above the ground; arrows 2 point to red crosses, the wet-bulb temperature, both recorded as millivolts. During the Lake Hefner investigations, anemome- ters were usually placed on top of the radiation shields. In the Buckeye Project, as in the Lake Mead studies, cup rotors were placed at the level of the ther- mocouples. Other variations will be mentioned below. THE ATMOSPHERE: WIND AND TEMPERATURE WIND Because it was unknown what wind and tempera- ture profiles would look like over the rough surfaces of saltcedar thickets, wind speeds and temperatures were measured 11, 7, 5, and 4 m above the ground, the aver- age height of the vegetation being 3 m. A fifth anemometer was installed underneath the canopy about 2 m above ground. To the southwest (from which the prevailing wind presumably blew®) two masts were erected (0 and M in fig. 3), each with an anemometer 5 m above ground. The transmitters originally installed were of the three-cup variety; they were made of aluminum and "To determine the location of the auxiliary masts (0 and M in fig. 3), information on winds was obtained from the Litchfield Park Naval Air Facility, 21 km east of the project site, but the data proved not to be representative of those at the project. See the section on "The Atmospheric Water Balance." F18 weighed about 140 grams each. The system was con- nected to counting devices regulated by mechanical and magnetic switches as described by Anderson, An- derson, and Marciano (1950). However, the starting and stalling speed of these transmitters (explained in the section on "The Dalton Approach") was far too high to obtain reliable wind profiles over tall vegetation. Moreover, the counting devices broke down frequently, because they could not withstand the rigors of the cli- mate. All anemometers, therefore, were replaced with a much more sensitive system. A new type of transmitter, consisting of three light plastic cups of 5 cm diameter attached to 5 cm long arms, weighed only 32.5 g. The counting devices were driven by photoelec- tric cells. These sensitive anemometers performed very well. Those at the auxiliary masts (O and M in fig. 3) were at times installed inside the vegetation and made it possible to construct wind patterns inside the thic- kets. See the section on "Winds." The counting mechanisms indicated every fifth and every fiftieth turn of the rotor on a strip chart. An example is given in figure 15. The number of "pips" per hour was manually converted to meters per second using a calibration table. AIR TEMPERATURE Air temperatures were measured with copper- constantan thermocouples connected to the strip re- corders. The thermocouples were protected from direct sunlight with aluminum shields of the type described by Anderson, Anderson, and Marciano (1950); they are outlined in figure 16. At this site these "coolie hats" had a tendency to corrode, to lose their reflectivity, and to become less effective shields. Covering them with adhesive aluminum foil proved to be a very practical and lasting remedy. Mean hourly temperatures were found by averaging the five or six values on the strip charts (see fig. 15) over 1-hour periods. Most of the averaging could be done graphically. Especially valuable was a commer- cially available data-reading device connected to a counter which greatly facilitated transposing data manually from the strip charts to graphs and tables. PRECIPITATION AND WATER VAPOR RAINFALL Precipitation was measured with three standard- type recording rain gages, one placed outside the saltcedar thickets, one on the 4-m platform, and one inside the thickets. Plastic wedge-shaped rain gages were installed on all tanks at about 1.50 m above ground. STUDIES OF EVAPOTRANSPIRATION BPM 7PM GPM FIGURE 15.-Windspeed record for June 21, 1965 between 1200 and 1600 (marked 6 p.m. and 8 p.m. on this chart). The two columns on the left are the windspeed "pips" for the anemometer 11 m above ground; the 11th to 14th columns are the records for the anemome- ters inside the thickets counting five times as fast as the others; the four columns on the right give the wind directions, N, E, S, and W, respectively. About % actual size. Precipitation data show that the rainfall was typical of the Sonoran Desert. During the summer, heavy thunder storms did occur, and 24-hour rainfalls of 30 mm and more were not uncommon. The heaviest rain- fall during the project period, which resulted from hur- ricane Helga, was 115 mm in 3 hours, on September 16, 1966. WATER VAPOR Water vapor, expressed in millibars (1 mb = 100 Pa) was computed from the dry and wet bulb ther- mocouples records. The wet thermocouples were kept permanently wet with wicks fed from reservoirs (fig. 16). All thermocouples above the vegetation had suffi- cient ventilation due to the system illustrated in figure 17. Those close to and inside the vegetation were venti- lated with fans driven by small electric motors. A nomogram was constructed (fig. 18) to convert dry and wet bulb temperatures into vapor pressures, under WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA b b b => c d d e | C E % & e bd’o' mill} f FiGURE 16.-Sketch of the air-temperature sensing system. a, radia- tion shield; b, plates holding the ventilator systems (see fig. 17); c, wet and dry thermocouple assemblies (enlarged below); d, dry and wet thermocouples, the wet one covered by wick e, the ends of which are floating in the reservoir; f, wire connectors to the rec- order; g, vent screw to allow refilling of the reservoir. the assumption of a constant atmospheric pressure of 985 mb (98.5 kPa) (the test site is 260 m above sea level). RADIATION The total hemispherical incoming solar radiation was measured with a pyrheliometer designed for measuring the intensity of solar radiation from 0.3 wm to 5.0 um. Nearly all the solar energy received at the earth's surface arrives in these wavelengths. The in- strument consists of two concentric rings, the inner one is painted with lampblack, the outer one has a cover of magnesium oxide which strongly reflects solar radia- tion. The difference in the temperature of the two rings gives a measure of the solar radiation. This sensing element is hermetically sealed in a bulb of soda-lime glass. The total reflected solar radiation was measured with a similar instrument mounted upside down. F19 FiGurE 17.-Schematic diagram of the ventilator system. Both dry and wet thermocouples are ventilated; the dry one between top and middle plates (see fig. 16), the wet one between middle and bottom plates. (From Anderson and others, 1950.) The total short- and long-wave incoming and re- flected radiation was measured with a radiometer simi- lar to the one used for the lake studies. Unlike that instrument which had a reflecting surface on the underside, this one was painted black on both surfaces and therefore acted as a net radiometer, recording the differences between incoming and outgoing radiation. The sensing element measured about 100 x 100 mm, four times larger than that of the instruments used in the lake studies. This plate was ventilated by a fishtail nozzle blower to eliminate the effect of gusts of wind on the calibration. The accuracy of such a radiometer de- pends greatly on the uniformity of the blackness on both surfaces. Amusingly, the cooling wind was appar- ently very attractive to doves nesting in the neighbor- hood and frequent cleaning and repainting of the top surface was necessary. Painting the retaining equip- ment with a sticky substance discouraged undesirable visitors with some measure of success. Figure 19 shows the radiation mast shortly after installation. Net radiometers of the type used here, and also the soil heat-flow plates to be discussed later, measure heat flow by means of thermopile voltages. The voltage needs a correction for temperature of the plate and each instrument has its own calibration factor. This makes transcribing of heat-flow data tedious and very time consuming. An adjustable nomogram was con- structed to facilitate this procedure (fig. 20). Details and theory are given in Appendix B. F20 SOIL TEMPERATURE, HEAT FLOW, AND SOIL MOISTURE SOIL TEMPERATURE The term G in the energy budget (equation 17) is often neglected when data are computed for periods of a day or longer. Over bare ground they certainly cannot be neglected over shorter periods. However, when the soil, as at the Buckeye project, is thoroughly protected by a dense stand of tall vegetation, the hourly and even daily changes become very small. Nonetheless, to have a complete picture of the situation, thermocouples were installed at 0,10, 45, and 90 cm below the surface in tanks number 2 and 6, and thermocouples were also installed outside tank 6 at the same depths. If the moisture content is known (as it was) and a thermal conductivity can reasonably assumed to be known (as it could), then measurement of incoming or outgoing heat is possible by a graphical analysis to be discussed later. HEAT-FLOW PLATES In addition to the thermocouples, heat-flow plates were installed in tanks 2 and 6 at 1 cm below the sur- face. They work on the same principle as the net radiometer but are much smaller (about 29 x 86 mm) so as to interfere as little as possible with the soil- moisture movement. Because such a small surface may not produce representative data, three plates were in- stalled in each tank and connected in such a way that the strip-chart recorder printed the mean value of each of the triplets. Unfortunately, these plates were very difficult to seal against the corrosive action of the soil moisture. Hence only a few sets of reliable soil heat- flow data could be obtained from these plates. Blackwell and Tyldesley (1965, p. 142) report that their tests reveal a variation in the calibration of their soil heat-flux plates of about +30 percent, and they continue: "Allowing for the use of unrepresentative calibrations, storage errors, and sampling difficulties, it would be unrealistic to claim an overall accuracy of much better than + 20 percent." The experience at the Buckeye Project was similar. Table 5 illustrates some of the difficulties. It gives the calibration factors for four plates as provided by the manufacturers at time of purchase and at two later dates. Neither the drastic change in factors nor the drop in significant figures underpins trust in the heat- flow records. Luckily, as will be shown later, the flows are small anyway and the actual values were of little importance in the energy budgets. SOIL MOISTURE Soil moisture and soil-moisture changes were meas- ured by the neutron-scattering method (van STUDIES OF EVAPOTRANSPIRATION & @ A sel- «© C & F 7 sol- € E Ty+°e 48.- --35 C e, mbar |-16 46t- Ase I a«[- -as - + azl- we Lse 40|- -40 L Ls n -se sel- Iso ® 361- Les T“ R Lze -* 10.- [o nd lE -:l Lo t __ HYGROMETRICCHART _ t, £ [ FOR THE { [s - BUCKEYE PROJECT T. |- 4}- Lo Ta F {. F 2l - -a L s € o- r-O FIGURE 18.-Nomogram to convert a given set of dry and wet bulb temperatures into vapor pressure expressed in millibars assuming a constant atmospheric pressure of 985 mb (98.5 kPa). Connect air temperature (T,) with a straight line to the wet-bulb temperature (T..) and continue this line to the e,, line to read vapor pressure. Example: T, = 28.5, T,. = 18.0, e, = 14.0. Hylckama, 1974, p. E16-E18). Readings were taken at least once a month, but during periods of intensive ob- servations, the frequency was increased to two or four times per day. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA a . as "he's w Figure 19.-Radiation mast with pyrheliometers up (a) and down (b) and with the net radiometer (c). Mast is 7 m tall or 4 m above the mean height of the vegetation. At the time the photograph was taken the soil was still bare. Compare with figure 5. A FigGurE 20.-Adjustable nomogram for the computation of heat flow. E, and E;: milivolt scale (data read from strip charts, see fig. 14); F, movable scale to adjust for plate temperature; M, and M;: heat- flow scales to read calories per square centimeter per minute, ad- justable to the calibration constant scales C, and C,. See Appendix B for details. F21 5.-Heat-flow plate calibration factors Recalibrated on Plate Nos. At purchase time December 1961 February 1963 314 0.09666 0.0747 0.072 315 0.13872 0.0828 0.075 316 0.12021 0.08365 0.081 317 0.14259 0.0849 0.080 Batteries of tensiometers were installed in all the tanks at depths commensurate with the water levels in the various tanks (fig. 21). Readings were taken 5 days a week and at 2-hour intervals during periods of inten- sive observations. Figure 22 shows such an event at night. In this way, periodic changes in soil moisture could be detected quickly. Directions of soil-moisture movement can also be de- tected with these instruments by converting the ten- siometer readings to manometer readings. Some of the results were used in the studies by Ripple, Rubin, and van Hylckama (1972), so they will not be discussed here. WATER-USE MEASUREMENTS The instrumentation to measure the water use is de- scribed extensively by van Hylckama (1974, p. E7- ~ Ficure 21.-Tensiometers installed in one of the tanks. Instruments were made at the project from commercially available parts. F22 FicurE 22.-Taking tensiometer readings at night during one of the intensive observation periods; observer (back to camera) is using a headlamp. See the section on "The Special Observation Periods." ES.). As mentioned on p. E20-E22 of that report, it was possible to analyze the frequency of fillings and to con- struct curves of water use over 24-hour periods. An even more refined analysis could be made if the on-off recorders were set for daily instead of weekly charts. An example of a daily chart is given in figure 23. Not only could the times of filling be more accurately de- termined, the duration of a filling could also be known and a much more detailed picture could be obtained to construct curves of hourly water use. Daily charts were always installed during periods of intensive observa- tions when the water meters were read at 2-hour inter- vals. THE MICROCLIMATE OF A SALTCEDAR THICKET WHY THE MICROCLIMATE? To fully understand and appreciate the methods of estimating evapotranspiration it is helpful to discuss some aspects of the weather inside and above saltcedar thickets such as occur near Buckeye, Ariz. After all, a saltcedar thicket, or any other flood-plain vegetation, comprises a much more complex environment than that on which most evapotranspiration equations are based. For example, Penman (1956, p. 20) defined potential evapotranspiration as "the amount of water transpired in unit time by a short green crop completely shading the ground, of uniform height and never short of wa- ter". Only the "never short of water" applies to the saltcedar in and around the evapotranspirometers, and the evaporating surface is not so well modeled by a two-dimensional surface as is a close-cropped lawn (or STUDIES OF EVAPOTRANSPIRATION XC s X f s se Uf NXT fig 5M§5 §®®\\ a -~ ' ile sz ssa %Z¢I%¢Z%%%%§‘3E & -t. m J "laffifi 0%”0[’I{'¢'®%‘/ Ree: (/ ”alwllgllz'la'flz'f'ldfiz’fi S t fills ”Ii/”II ill 1 "ll/ffmfi' t (t" m / [fiwfiflnfi'fififim O (. fl nestum fififi‘fifiw FIGURE 23.-Event recording chart indicating time and duration of tank filling. The jagged line near the center indicates water tem- perature. About % actual size. turf as Penman calls it). Thornthwaite and Mather (1955, p. 20) specify a number of conditions necessary to fit the definition of potential evapotranspiration: "First, the albedo of the evaporating surface must be a standard. Second, the rate of evapotranspiration must not be influenced by the advection of moist or dry air****." Certainly these conditions are not met in any riparian saltcedar thicket or flood-plain vegetation. Furthermore, temperatures, vapor pressures, and wind profiles inside the thickets may affect the evapo- transpiration rates just as much as do those above the vegetation where they dre usually measured. Also, the roughness length is often assumed to be constant but can, as will be shown, vary greatly with changing wind speeds and even seasonally. THE SPECIAL OBSERVATION PERIODS The instruments described earlier, by and large, worked faithfully, but did not always record truthfully. Bees and sometimes ants, discovering a source of moisture in the wicks of the hygrometers, caused strange fluctuations in the wet-bulb temperatures. So did dust devils which in the course of a few minutes could turn the wicks from white to brown. The actions of doves changing the reflectivity of the net-radiometer plate was mentioned already. Finally, the very sensi- tive anemometer system was open to undesirable infes- tation by spiders and other "bugs." WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA At this point an explanation is in order on the kind of recording used at the Buckeye Project (see fig. 14). As already mentioned in the section on "Instrumenta- tion," one item is recorded every minute, which means that on a 16-point recorder one phenomenon (for in- stance, the dry-bulb temperature at 11 m) is recorded less than 4 times per hour. On a 12-point recorder it is recorded 5 times per hour. There is, therefore, a time lapse of 16 or 12 minutes between readings of each transmitter. During the morning and evening hours quite a few "things" change rather rapidly. At such times it is easily possible that anomalous data show up caused by infrequency of recording rather than by in- strument failures or unusual happenings in the atmo- sphere. For long-term averages all this is not serious because such anomalies are easily spotted and can be elimi- nated before monthly, weekly, or even daily means are computed. However, to compute potential evapotran- spiration from energy-budget and mass-transfer equa- tions, short-term, even instantaneous, values are needed. For this reason special observation periods were or- ganized. Each period lasted 3 days, during which time F23 all instruments were checked every 2 hours around the clock (see fig. 22) and records were kept of those in- struments that did not record automatically, such as tensiometers and, most important, watermeters on the tanks. Before such periods, all instruments were checked for proper calibration, cleaned and adjusted if necessary, and coworkers were instructed into the proper methods of checking and recording. Also clocks were set to true local solar time so that it was exactly 1200 (noon) when the sun was exactly at 180°. The hourly averages of the data so collected are presented in Appendix A for the convenience of those who desire to analyze the information further. In addition, 15 5-day periods, selected from the records for their likelihood of accuracy and completeness, are preserved for future reference. DAILY CYCLES OF TEMPERATURE AND HUMIDITY Figure 24 shows the cycles of temperature and humidity. The vegetation was, on the average, 3 m tall, so the lower readings were taken inside the vegetation. It is clear that these readings were generally higher during the day and very much lower during the night; in other words, the vegetation was exposed to greater AIR TEMPERATURE IN DEGREES CELSIUS WINDS, IN METERS PER SECOND VAPOR PRESSURE, IN MILLIBARS % 9 a & 7 |- 2 w < 5r o g:- 8 g 5 Z <6_E> <6 mbar> 9 0 z g 0€ IT 2- D 05,(5) _ 09, (6) 13.(7) 17,(5) me) 01 (8) . 09, 13 17, 21 6 12 18 24 TRUE SOLAR TIME FIGURE 244. -Some microclimatic characteristics of the Buckeye project site, May 1966. A, March of temperature (degrees Celsius), 11 m (open circles), 5 m (solid line), and 2 m (dots) above ground; B, Windspeeds (meters per second) 7 m above ground; C, March of vapor pressure (millibars) for the same days 11 m (open circles), 5 m (solid line), and 2 m (dots) above ground; D, (left) six temperature profiles at 0500, 9000, 1300, 1700, 2100, and 0100 hours (true local time) on dates 1nd1cated by the number in parentheses; D, (right) six x vapor-pressure profiles for the same tlmes F24 STUDIES OF EVAPOTRANSPIRATION TRUE SOLAR TIME 6 12 18 24 6 12 18 24 6 12 18 24 T p o G o T 20 |- AIR TEMPERATURE, IN DEGREES CELSIUS i9F 4 -, JBAF June 1966 3 3 L 1 1 14 1 1 1 15 1 1 1 (6 WINDS, IN METERS PER SECOND VAPOR PRESSURE, IN MILLIBARS GROUND, IN METERS HEIGHT ABOVE T I T T we T -# s(G) g(4) 13(5) 17(3) 21(4) o1(6) 5, 9, N t o T ho o T - o T AIR TEMPERATURE, IN DEGREES CELSIUS £ * July 1966 Apanel 1 1 B 1 1 1 9 | 1 10 WINDS, IN METERS PER SECOND 20 |- VAPOR PRESSURE, IN MILLIBARS HEIGHT ABOVE GROUND, IN METERS -r--" \ | -t- 1 1 12 18 24 T T T I I T a <6 mbar > -> (9) (7) 9f(8) 13, 17 21(8) / 01 (10) 5, 9 13; 17 21 01 6 12 24 6 12 18 24 TRUE SOLAR TIME FicurE 24B.-Some microclimatic characteristics of the Buckeye project site, June and July 1966. A and E, March of temperature (degrees Celsius) 11 m (open circles), 5 m (solid lines), and 2 m (dots) above ground; B and F, Windspeeds (meters per second) 7 m above ground; C and G, March of vapor pressure (millibars) from the same days, 11 m (open circles), 5 m (solid lines), and 2 m (dots above ground; D and H, (left) six temperature profiles at 0500, 9000, 1300, 1700, 2100, and 0100 hours (true local time) on dates indicated by the number in parentheses; D and H, (right) six vapor-pressure profiles for the same times. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F25 TRUE SOLAR TIME 12 18 24 6 12 18 24 30 [- AIR TEMPERATURE IN DEGREES CELSIUS August 1966 *. C 1 (5 1 | 4 (6 WINDS, IN METERS PER SECOND T I I T T I VAPOR PRESSURE, IN MILLIBARS ho B T T - &n T as sa T IN METERS an w T T HEIGHT ABOVE GROUND 51(3) g(4) 135) 214) <6 mbar> -> 01,(6) 5, 9, 13, 17, 21, __ 01 T T T T T 40 20 AIR TEMPERATURE IN DEGREES CELSIUS Pa o p WINDS, IN METERS m MILLIBARS PER SECOND m 0 VAPOR PRESSURE, IN - o HEIGHT ABOVE GROUND, IN METERS (8) (9) 9 (9) 13(10) 17, 21, 1 <6 mbar> -> 04011) 5, 9) 13, 17, 2%: 01 12 18 24 6 12 24 6 12 18 24 TRUE SOLAR TIME FIGURE 24C. -Some microclimatic characteristics of the Buckeye project site, August and September 1966. Legend see figure 24B. temperature extremes than the air above it. The humidity was also much higher under than above the canopy, as is to be expected from a vegetation transpir- ing at the potential rate. September shows a decrease in vapor pressure inside the vegetation, and so does March, as can be expected. However, May 1967 shows a much lower vapor pressure inside the vegetation than May 1966. This lower vapor pressure was due to an increase in salinity of the ground water in the evapo- transpirometers; it thus indicates a lower evapotrans- piration rate. Figure 24 also shows selected temperature and vapor profiles illustrating better than the graphs the differ- ence in steepness of gradients between the values in- F26 STUDIES OF EVAPOTRANSPIRATION TRUE SOLAR TIME 6 12 18 24 6 12 18 24 6 12 18 24 T T T T T T T T T T T T £530” 3 292 io w T 0 £ 20. qu fl w & w = x 2G a ar So < Z , H F a 4 . March 1967 ** 23 1 Ae L- 124 1 1 1 (25 £. 1 | es o ng T T T T T T T T T T T T #B o 4, a Q j- UJ 2 w I g x a 55g "| ¢ a 1 1 1 1 1 1 1 1 1 1 1 1 f w ¢£ g10F o 2 a 4 95°13 5} | Sp 2 °C a. 3 T T T T T T T T T T ( f w 2 w 205 (DP—7 EOE <6 mbar > 65 z ° -> a ° 2 I 2 1 1 1 1 1 1 1 1 T T Te T T T T %:: T T T T T T T £330» 2 3 i w C is 20T- - w w & w = c & 0 w 10 }- # a o < Z A ** & = E May 1967 +. 12 | 1 1 13 1 1 1 (14 1 1 1 15 o T T T "ae" T T T T T t r r ; . £9 a} # w u 0 o |- WJ 2 w 9 2) * BH F . L 1 1 1 1 1 1 1 1 1 1 1 f w w ra © C < 10 |- s $q. KW a 9C | a. = T T T T T 3: T T T T u at . 4 2 6C | -o ® I g < <6 mbar > 9 0 z2 sage I 5 .~ < T | 1 1 1 1 1 3 1 1 l 6 12 TRUE SOLAR TIME FicurE 24D.-Some microclimatic characteristics of the Buckeye project site, March and May 1967. Legend see figure 24B. side and those above the vegetation. Some anomalies should be pointed out: for instance at 0100 on Sep- tember 11 there is a large increase in vapor between 4 and 5 m. This increase seems impossible because it would indicate a downward movement of moisture and possibly dew formation when the air temperature in- side the vegetation is 20°C! Similar but smaller anomalies can be observed in the March and May 1967 WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA profiles and can probably be explained only by some malfunctioning of the instruments. WINDS Figure 24 further illustrates the fluctuations of wind speeds 7 m above the ground (about 4 m above the vegetation). The curves show that during the observa- tion periods wind speeds were in the light to gentle breeze range on the Beaufort scale (less than 4 m/s). Notable exceptions occurred in the afternoons of July 8 and August 3 when wind increased to a moderate breeze of more than 6 m/s, and another wind, perhaps more unexpectedly, increased to 6 m/s, in the evening of August 5. The profiles of the winds above the thickets will be discussed presently. What concerns us first is what happens inside the thickets. In September 1966 two anemometers were installed 0.5 and 1 m above ground inside the saltcedar on tank 6. By marking the values for the windspeeds at appro- priate heights for each hour, one can draw lines of equal windspeed (isotachs). Together they give a clear picture of what happens during a day. Isotach curves peaking upward indicate that winds are slowing down; those pointing down indicate that winds increase. Most interestingly during some times of the day winds of higher velocity seem to "tunnel" through the vegeta- tion. Figure 25 shows five examples. Another example was given by van Hylckama (1970a). Unfortunately, related data such as vertical windspeeds and vapor pressure fluctuations were not observed. The tunneling phenomenon is true through, and has been reported elsewhere (in particular by Allen, 1968). Such ventila- tion makes it possible for large quantities of sensible heat to be removed, and thus may well affect the rate of evapotranspiration. Nearly all energy-budget and mass-transfer methods, or a combination of these, assume that the plots of instantaneous windspeeds (or of those averaged over periods of 4 hours or less) versus the logarithm of height lie on straight lines. It was therefore desirable to analyze the wind data to verify the occurrence of such conditions at the project site. About 2000 profiles were studied using border-punch cards of which figure 26 is an example (Casey and Perry, 1951). This system allows rapid selection of information by grouping pro- files according to windspeed, date, temperature, roughness length (¢,), friction velocity (w.), stability ratios (SR), and zero displacement (d). For each profile the stability ratio was determined, expressed as (Deacon, 1953) $n » Airlit- or (97) 0.5 <0.5 -I May 11/12, 1967 I 1 | T I 3.0 T I 0.5 A J Hy is | | S2 _ L0 3 PXE HEIGHT ABOVE GROUND, IN METERS Sept. 10/11, 1967 1 1 18 20: 22 24 2 4 a 6: s 16 1z 11. 16 TRUE SOLAR TIME FicurE 25.-Distribution of windspeeds (isotachs) within and above a stand of saltcedar near Buckeye, Ariz. The vegetation is about 3 m tall. The numbers near the isotachs indicate windspeeds in me- ters per second. Isotachs above 11 m are extrapolated. F28 © z 0 0 o 0 0 o o 0 o e cA: rg Ind R$ Ing ri In? r4 rE al Year _ Day.. Hour "A4 BUCKEYE PROJECT 3 A a~ wIND ANALYSIS $ n F Zo == M cm o ~ 77.3 eo~3* Uy = cm/sec @~ winp pirection __ W ___ _C o G ©% C o o ~ d =o . @ ol «[* o" "JF @ -52 s|:F-@ $i FTT ® o o e 0 0 0 © o o 0 0 0 FicUurE 26.-Example of a border-punched card used for the wind- profile analyses at the Buckeye Project. in which the subscripts give the heights above ground (m), T is temperature (°C), and u is windspeed (m/s). SR is indicative of the stability of the air. Negative numbers occur when the temperature at high elevation is lower than below, creating unstable conditions if the STUDIES OF EVAPOTRANSPIRATION difference is larger (more negative) than the lapse rate. The latter, however, is of the order of 0.06°C per 6 m, and usually, can be neglected. Strong inversions usually connected with low windspeeds give rise to high stability ratios. More than 50 percent of all cases showed an SR between -0.10 and +.0.10 with about half of them negative and half positive. The variables needed to describe the logarithmic profile are the roughness length and the friction veloc- ity. Both can be determined graphically; Lemon (1963) and Sellers (1965) explain the procedure. If neutral stability conditions exist, that is, when SR is close to zero, windspeed increases linearly with the logarithm of height and we have (Sellers, 1965), Au Au mrsh? > = constant, - (28) In (@3/z4) in which Au is the difference between wind speeds at heights z;, and z,. However, there is an elevation at which the wind speed becomes zero due to friction be- tween the air and the vegetation or the soil. This height is called the roughness length (z,), and we have u,. = (constant) In (2/z,), (29) where u, is the wind speed at height z. If the plots of windspeed fall on a straight line, the roughness length is easily determined graphically by extending this line to zero. Figure 27 shows a roughness length of 0.2 m for the windspeeds plotted on line C. The "constant" referred to is only constant for one particular profile, but increases with an increase of Au. The flatter the profile, that is, the smaller the angle between the profile line and the abscissa, the larger the constant becomes. Laboratory experiments of turbu- lent flow in pipes show that this constant depends on the rate at which momentum is transferred (which is fast for a "flat" profile) and on the air density. Thus (30) constant = 71— I: f; ill/2 . L3: where 7 is the vertical transfer of momentum, p, is the air density, u. is called "friction velocity," and k a con- stant of proportionality. The constant was named after von Karman who first studied this phenomenon. The von Karman constant actually varies with the temperature gradient. Sheppard (1946) shows that k varies from 0.32 at a positive gradient of 0.56° to 0.61°C/m at a negative gradient of 1.66°C/m. The gra- dient at the project varied from -0.34° to +0.56°C/m, but two-thirds of the observations showed a negative gradient and the bulk of these was on the order of WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA 2 4 m s! 6 sal w CC w '— w 3 e: B o Z 2 O CC O w > O c < 3 - C C w T a (49°) 0.2 {2 - 0.1 | 1 | 1 2 4 6 WINDSPEEDS, IN METERS PER SECOND Ficur® 27.-Graphical determination of roughness length (z,), fric- tion velocity (x,), and zero displacement (d). Curve A is a hypothetical set of windspeeds at 11, 7, 5, and 4 m extended to 0 m/s. Line B is derived by plotting the indicated speeds of curve A 1.5 m (d) below those "observed" on that curve, giving z, = 0.5 m. Line C connects another set of hypothetical windspeeds at indi- cated elevations. This line is straight and the extension gives 2, = 0.2 m. The angle (@) is about 49°, hence u,, = 100k/tan 49 = 35.6, k being von Karman's constant, taken as = 0.41. -0.10° to -0.20°C/m. Thus, use of an average value of 0.41 seems reasonable. Monteith and Szeicz (1960) and van Bavel (1966) used the same value. With this in mind we can now write Us u:_¥_ln Z k Z6 (84) The friction velocity can be obtained from the slope of the wind profile. As an example, line C in figure 27 makes an angle of 49° with the abscissa, hence we have u_ = 100 k/tan 49 = 35.6, remembering that In z, is a constant. Actually, z, is the integration constant for duldz = u.lk z (32) the differential equation of the profile of mean wind F29 velocity above ground. This equations strictly holds over flat surfaces such as ice and snow or very short smooth vegetation. Under such conditions z, can be considered to be an invariable property of the surface. Tall vegetation waving in the breeze does not offer such a convenient roughness length. Equation 31 therefore has been empirically modified to U u-= *- In z - d R Zo (32) in which d is the zero-plane displacement. It can be regarded as a measure of the wind penetrating the vegetation or as a level above which the normal turbu- lent exchange takes place. Sutton (1953) points out that, if d and z, are regarded as independent arbitrary constants, equation 33 cannot be derived from the orig- inal differential equation 32 because that is an equa- tion of the first order whose solution must involve not more than one arbitrary constant. A zero displacement also can be determined graphi- cally as illustrated in figure 27. Assume a curve A drawn through 4 points at which windspeeds are measured; continuing this curve to the zero line would indicate that at 2 m above the ground the velocity is zero. By lowering the points, we find by trial and error that if each of the speeds are plotted at 1.5 m below the original elevation, the points do lie on a straight line and form a logarithmic profile, with a z, at 0.5 m and 1.5 is the zero-plane displacement (d). The d values for the Buckeye Project data, however, are so variable and random and even negative (when a curve of type A in figure 27 is convex instead of concave) that, as Sellers (1965, p. 150) states: "it is almost impossible to attach any real physical significance to them" and "in practi- cally all cases and within the accuracy of the meas- urements the wind profile above the vegetation can be described just as well by equation 31 as by equation 33." This is not surprising, remembering that d is an empirical constant. The profile above tall vegetation, however, is af- fected by that vegetation, and z, actually becomes an adjusted roughness length. Such an adjusted rough- ness length changes with the steepness of the profile and also with the stability ratio. For instance, by grouping wind ratios u;/u, (the ratio of the wind speed at 7 m over that at 4 m) and evaluating the modified roughness length in each of these groups (which can be accomplished quickly by sorting the border-punch cards), one can plot the means so obtained as shown in figure 28. Similarly, the modified roughness length can be plotted against stability ratios, as shown in figure 29. These are the roughness lengths that have been used in the mass-transfer and combination equations to be discussed later. F30 2.00 |- a 1.80 |- 1.60 |- 1.40 |- WIND RATIOS u;/uy 1,20 |- e eles frst --- -~ 40 60 80 100 120 140 160 180 200 MODIFIED ROUGHNESS LENGTH (2g), IN CENTIMETERS FicurE 28.-Relation between modified roughness length and the ratio of windspeeds at 7 m to those at 4 m. Only the means of z, and u;/u, for each stability group are plotted. 14 F- L & 1.2 % 0.8 |- X 0.4 0.2 STABILITY RATIOS (7) 4 1 o & I L2 1 __3 " f 6 N 5 p % 1. t I 8 L.. $ 50 100 150 200 ROUGHNESS LENGTH zp, IN CENTIMETERS FicurE 29.-Modified roughness length versus stability ratio. A dot on a horizontal short line indicates the mean z, for each stability group. A dot on a vertical short line indicates the mean SR for each z, group. The length of the lines indicates the stand- ard deviation. Numbers on the inside of the axes indicate the groupings of the ranges of z,'s and SR's for which means were computed using the border-punch cards (see fig. 26). STUDIES OF EVAPOTRANSPIRATION RADIATION Figure 30 shows the march of radiation and heat flow for the seven 3-day periods mentioned earlier. The curves of the total short-wave (direct and diffuse) solar radiation show that most of the time there was a clear sky. Variable cloudiness occurred on July 7; August 3 was overcast as was March 24. Reflected radiation is also quite regular, but the change in reflectivity with the change in cloudiness is clearly shown in the record. Net radiation, consisting of the total short- and long- wave radiation coming in, minus that going out, fol- lows the march of the total short-wave radiation quite faithfully, except on May 5, 1966, where between 1200 and 1500 an unusual pattern can be observed which is not easily explainable. A warm rain could cause this kind of record, but there was no rain during any of the observation periods; nor do temperature and wind pro- files show anything unusual. SOIL HEAT FLUX The soil heat-flux term appears in the energy budget because part of the incoming radiation is used to heat the soil during the daylight hours. Conversely, at night the soil loses heat by outgoing long-wave radiation and by conduction. The soil heat flux is considered positive when heat moves through the soil toward the surface and nega- tive when it moves from the surface into the soil. Fig- ure 30 shows that the night-time flows are quite regu- lar but during the day they are not. Some of this ir- regularity is undoubtedly due to a variation in shad- ing, for instance the increase in flow at 1400 on all three days in June 1966 and also at 0700 on September 8, 9, and 10. Other sudden changes may well be due to temporary accumulations of moisture on the heat-flow plates. Measuring this flux is troublesome because the sens- ing elements must be small in order to allow water and vapor movements which otherwise may affect the tem- perature of the elements and give wrong flux values. Heat flux through a small area, therefore, must repre- sent the soil heat flux through a large area. At the Buckeye Project a measuring area of 0.00735 m* sup- posedly represented an area of 81 m*, which is the size FiGcurE 30.-March of radiation and heat flow at the Buckeye project for 3-day periods as indicated. Total short-wave radiation (direct plus diffuse): open circles (0000); short-wave radiation reflected by soil and vegetation (albedo): dots (.....); net short plus long-wave radiation (incoming minus reflected and outgoing): xxx; heat flux into and out of the soil: AAA. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA LANGLEYS PER MINUTE 1:2 1.2 1.2 1.2 1.2 1.2 6 12 18 24 6 12 18 24 6 12 18 24 I T I T I T T T T T T T J 10 C5 g°° 6 go* 7 s %o 8 2) o x o *g* 5 .** o 4 o $ *X :* May 1966 _ _ |- x - x x6 x 00 0 -I o x Xo ox he $& Xxo a. ox o * a 2 * aaaAs ab x *e AadAAaad44 q* *g Aa A4 ,a & AA 0 Baaaad Q'QA,“ s __AABQ?x§xxxxxx9X¢._H_ & AbiQQXxxxxxxxW‘éA‘ 7 “AAZRQxxeQXX‘x—O Afr A A tA - z AA A AA A a4 £8 aA? - 3 °° 4 $*"s 5 0°" 6 10 - o M 0 o o o x**x o .x ** x 5 , June 1966 % s P M x % « 9x Ox * Ox * x0 xo x8 3 3 px x 46 ox To o 8 y 8 a X apAAAAAAq X aA & a A A A AA a X Aad AAA 0 A A A 0 A 0 A Sa o* AX xxx xxx A22 ccc cx xR* 2a AigfixxxxxxXx—o ‘AAAAQ'AA X X ‘AAA‘Q‘AA X x X ‘AAAAO'AA # a & &A 8 A 110 6 $°°5 7 40° ° 8 0° 9 I 6 # o o x*x29 **§x* o xxx, ° July 1966 is x xo x x° 0x x * ox 0 o *o 7 - "ox x ° xy 0 o & Br o yo - ag x @ga aa8448%2 x x28 ag88ana $2 aaaao 0 & fi'?.AA A . aa Ak R3 gxooxxx x ® '4A4AAA'AAA #8xtxx X8 -?A90A9‘A9¢A‘9xxxxxxx§ - a ~AfAa A A A 7110 3 4 s 5 58 6 [~ 09" o o o C 3x Xie OOXXXO Aug. 1966 : Sox® & xo x es x0 X ie h o xo 2x Oxxx M 0 x x x 0 8 o 7 - 0 x ip 0 x x30 Xx X0 l aag X xo AAAAAAAAAA€%X *o AAAAAAAAagéo xigeAAAAAAAAA **: 8. AAlixxxxxxxxxxgx 44.15 "AAAQxXXXXxXXXXXX 85,4, g AA "KX x yx x**x xX 0 | A AAA A Af A a 110 23 24 25 26 [s 0° s 5° ° # 00 Mar. 1967 - y*® °,0,00 *y ® ® x 0 0 x x 0 0 x Xo ~I R x x XX ox o xo o* X0 0 X xO - A 0 d A A A aA A a 09§£ xggégAAAAAAAA 8§A gAAAAAAAAAAooeéAA Egéo Aaa A Io 5844 .+ 338 12009,3,0000000 4 e yue rua * 0 * + 223 , , . 64 * 3x Xxxx x x XX - A 4A" aA aad - A? a 00 00 . 00 110 _11 os, 12 $°°% 13 goals 14 X x 0 0 Ao*gt x" * x x* x May 1967 - °% o SX xo *x *x x 0 X x o o* *s ox wo 4 - o* x x o X *s u a a 0 a AA AA A a °,, A daa A4 x A044 Y"? a aa ck268 xxxx Rx 84 a o* AxxA'ASSQA AAégeAAAA x7 0 +»@Pat git 1/%xx*** +00$,,,.0°" *xX*3XX*IX**~ * ansag cen-" xxx x xxA* * 1 1 1 1 1 1 1 1 1 1 1 1 6 12 18 24 6 12 18 24 6 12 18 24 TRUE LOCAL SOLAR TIME HUNDRED WATTS PER SQUARE METER F31 F32 of one evapotranspirometer. Measurements made on bare soil might actually be representative, but on vege- tated soil, where sunlight and shade may alternate in unpredictable patterns, one can expect large variations which even show up in hourly averages. On the other hand, when the soil is vegetated and mostly shaded, the amounts of incoming and outgoing fluxes are very small compared to those of solar radiation; so they are often neglected (Penman, 1956). If the density of the soil, its heat capacity, and moisture content are known, and if temperature rec- ords at different depths in the soil are available, it is possible to compute the heat flux. We start with the equation for heat flux derived from Fourier's conduc- tion equation in one dimension: 2 "T 0z* (34) where T = temperature, in degrees Celsius, t = time, in seconds, z = depth, in meters, m= \/p c = meters squared per second (thermal diffusivity), A = thermal conductivity (watts per meter per degree Celsius), c = heat capacity-(joules per kilogram per de- gree Celsius), p = density (kilogram per cubic meter). At any given depth (¢) the heat flux (G, in W m -*) is: S (35) Oz or pe OT _ 0G, Ot Oz (36) Following Portman (1954) we can now multiply both sides of equation 36 by 8z and integrate from the sur- face where z = 0 to a depth z: 3 oT S pC $82 cae G0 = G2 0 (37) If, from temperature gradients, we can determine the depth where 8T/8z approaches 0, we can determine G, by evaluating the integral term from the surface to a depth where G, = 0. The term pc is the product of density and heat ca- pacity, and both variables change with the moisture content of the soil. The moisture content was regularly measured and expressed as moisture percent by vol- ume (My); Moisture content by weight (M,,) is then: ~ - "100 M, Ms = 140 + M, 89) STUDIES OF EVAPOTRANSPIRATION where 140 is 100 x the density of the dry soil as deter- mined by numerous weighings of Buckeye soils. The specific heat, according to Kersten (1949), varies from 0.176 for Northway silt loam to 0.197 for North- way fine sand. Because the texture of the soils at the Buckeye Project is slightly sandier than that of North- way silt loam, a value of 0.18 calories gram~' °C-'(753.5 J kg-" °C-") was adopted for dry soil. Hence, considering the specific heat for water being unity for all practical purposes, the volumetric heat capacity is: pe = 140 (018 4 -S-) calem ""C - (go) 1 - M., One can now plot two temperature profiles, say, 1 hour apart, on each of the layers of soil for which the moisture-dependent volumetric heat capacity can be considered uniform. The product of depth times a change in temperature can then be measured, for example, by planimeter, and this product multiplied by the pc for the particular depth gives the heat flux in calories per cm* (or joules/m*) per hour. For the exam- ple shown in figure 31, the profiles were plotted on DI Form 213-F paper where 1°C was represented by 0.5 in. and 1 cm depth by 0.05 in. Because the planimeter measured areas in cm", an adjustment factor of 6.2 was necessary to convert the data to calories cm ~*. Figure 32 shows the march of heat flow during one day computed from temperature measurements com- pared to that from heat-flow plates and also shows the 24-hour sinusoidal wave of these fluxes. It is im- mediately striking that these waves have roughly the $p 18 20 22 24 _ 20 one "pc My ( _ |__. '-- - ore a CC -O. u 0.276) 2.4 m 0.2 1 § ~~: SSC t 1 c" < 0.7 i TOTAL FLUX Q 04} -0.4x0.276x f 0.426)17.4 i: e sex oprexr _ . [f- =- 3 +0.7 x 0.426 x f 0 g 0.6 - =334cal em- EK * A w $ / >._o 5757—1" 20. & \ z w I '* \\ TEMPERATURE x" it [m PSL a: % h T a. *y -e-" % J w 0 |- -] 0 0 a & 3 fer g a. - -20 {2 4 2! C T S - -40 -4 |- - -60 -6 f: 5 0 4 8 12 16 20 24 TRUE SOLAR TIME FicurE 32.-Heat flow into (negative) and out of (positive) the soil in an evapotranspirometer near Buckeye, Ariz., recorded from heat- flow plates and computed from temperature records. The dashed lines are the curves of the first harmonics of the heat-flow and the temperature data, respectively. same amplitude, but their phase angles are about 4.5 hours apart. The sinusoidal wave for the fluxes com- puted from temperature profiles peaks at a little before 0800 while the one computed from the heat-flow plates peaks, as could be expected, at about noon. This dis- crepancy was quite consistent for all the periods analyzed. To find if more accurate or closely spaced ther- mometers could give closer agreement, a set of similar data, published by the Water Conservation Laboratory of the Agricultural Research Service in Tempe, Ariz. (van Bavel, 1967), was analyzed. The results, given in figure 33, show the close agreements between heat-flow data from heat-flow plates and those computed from temperature gradients. However, the temperatures were measured at depths 0.5, 1, 2, 4, 8, 16, 32, 64, and 128 ecm whereas temperatures at the Buckeye site were only measured at 0, 10, 45, and 90 cm. Also the Tempe thermographs were reported to be accurate to at least 0.1°C, while the Buckeye ones were only accurate to 0.5°C. Notice further that the Tempe fluxes are an order of magnitude larger. The data refer not to flow into or out of vegetated soil but into or out of bare ground. Especially the uncertainty as to what happens in the top 10 ecm of the soil during the course of a day must lead to grave errors. Hence soil-temperature data, as collected at the Buckeye Project, could not be used to compute heat fluxes on an hourly or more frequent F33 basis. They could, however, be used on a daily basis because the sinusoidal waves were very similar in shape. Figure 34 shows four examples of waves com- puted from temperature profiles that have been shifted over to coincide in phase with the ones of the heat-flow plates. -20 t T T T T T -| -2 CC LJ 10 [m < c ace CC w > CC o T I e a O 0 0 iG l a. CC a u w 1 & -1- 10 - 0 -- Temperature I- a records { Q 20 |- plates fra o - > 43 . I 30 1 l I l f | 4 8 12 16 20 24 LOCAL SOLAR TIME FicurE 33.-Heat flow into (negative) and out of (positive) the soil in a lysimeter near Tempe, Ariz., recorded from heat-flow plates and computed from temperature records. The dashed lines are the curves of the first harmonics of the heat-flow &nd the temperature data, respectively. (Plate and temperature data from van Bavel, 1967.) sd I I I T T T T I I T I I 3 |- May 5-8, 1966 |- June 3-6, 1966 LANGLEYS PER HOUR WATTS PER SQUARE METER 6 12 AB 24 6 12: 18 24 TRUE LOCAL SOLAR TIME 34.-Sinusoidal 24-hour waves of heat flow into (nega- tive) and out of (positive) the soil computed by harmonic analysis of heat-flow plates data (dashed lines) and from data computed from soil-temperature gradients (solid lines). F34 More than 20 years ago McCulloch and Penman (1956, p. 279) ran into similar difficulties in their at- tempt to compute heat flow in soils. They emphasized the need for precise thermometers and concluded: "If only one set of observations is to be used, then phases are preferred to amplitudes, that is inaccurate ther- mometers with an accurate clock are better than good thermometers with a bad clock." SOIL TEMPERATURES We have seen that the soil-temperature data of the Buckeye Project usually were inadequate for the com- putation of heat fluxes into or out of the soil. They are, however, adequate for computing some soil charac- teristics, using data derived from the harmonic analyses of the temperature waves. The method assumes that the data comprise only perfect sine curves of known periods P (in this case, days) or submultiples of P. Each curve explains a part of the variance of the data and it is possible to deter- mine how large a part of the variance is expressed by a first, second, etc., sinusoidal curve after the amplitudes (a) and phase angles («) of each curve have been com- puted (see Panofsky and Brier, 1965). Analyses of the temperature data of the Buckeye Project showed that nearly always more than 85 per- cent of the variance could be explained by the first harmonic, that is the sine waves with a 24-hour, or 86 400-sécond period. Using these phase angles and amplitude, one can compute thermal diffusivities (1), timelags (¢), and other characteristics." If a,. is the amplitude at a depth z and a, the amplitude of the surface temperature wave, we have (van Wijk and de- Vries, 1966): (40) a, = a, exp { 3% Z[LP:| 1/2} n where P is the length of the period (in this case 24 hours or 86 400 seconds.) Once 7 is computed, the time lag between the occur- rence of the maximum temperature at the surface and the maximum temperature at a distance below the sur- face can be computed using t=z/2(—P—— e mn (41) This time lag can then be compared with a time lag "Thermal diffusivity (7) is defined as K/pc, where K is the thermal conductivity (cal cm~' sec =C"), p(g cm-*) the density of the soil, and c (cal g~'* C~') its heat capacity. It denotes the temperature change in any layer of the soil as heat flows into it from an adjacent layer (Baver and others, 1972). STUDIES OF EVAPOTRANSPIRATION observed between peaks of the sinusoidal graphs or ac- tual curves at the two depths. A similar procedure can be followed for other depths, such as between z, = 10 em and z, = 45 ecm. The term z in (40 and 41) then becomes Az =z, - 2;. Van Wijk and deVries (1966) refer to the quantity D11 :[_———nP :| 1/2 TT (42) as the damping depth. It is the depth at which the amplitude is 1/e or 0.37 x the amplitude at the surface. The quantity arises as a constant when integrating the heat-flux equation (34) under suitable boundary condi- tions. At a depth of 4.61 D,,, the amplitude equals only 0.01 a, and the temperature (on a daily basis) can be considered constant. Table 6 shows the results of harmonic analyses for two locations and two 3-day periods using temperature data at three depths. With a few exceptions, observa- tional and computational data of the time lags agree quite well; this agreement gives some confidence in the computations of diffusivities and damping depths. The diffusivities increase as the moisture content increases with depth, and those computed for the three layers are in agreement with those given by van Wijk and de Vries (1966) and also by Baver, Gardner, and Gardner (1972). Figure 35 illustrates how damping depth and depths of near-zero amplitude can be obtained graphically by plotting the amplitudes computed for 0, 10, and 45 cm depths on semilogarithmic paper. The short horizontal lines indicate the depth of D, and the arrows are at the depth where the amplitude is about equal to 0.05 x the amplitude at the surface. HEAT FLUX AND WATER USE It follows from the first law of thermodynamics that if an evapotranspirometer for one reason or another uses less water, the energy of vaporization that be- comes available has to be used somewhere else. Prob- ably most of it is used to create turbulence in the air, and a certain amount will raise the temperature of the foliage, the branches, and the stems of the shrubs. Though the last three quantities are measureable, they were not determined at the Buckeye Project. The first quantity would be very difficult to distinguish from other turbulences. However, data were collected that showed an increase in the amount of heat stored in soils of evapotranspirometers using less water. Figure 36 shows the heat flux in two tanks having identical vegetation covers. One tank used less water than the other because of a difference in depth to WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F35 TABLE 6.-Some results of harmonic analyses of soil-temperature fluctuations inside and outside tank 6 for means of two 3-day periods 1 / [Diffusivity between 0 and 10, 0 and 45, and 10 and 45 cm, respectively. Damping depth (D,) = (i) ' at which a = 0.37 a,, between 0 and 10, 0 and 45, and 10 and 45 cm} T First harmonic (°C) Time lag Depth Mean Phase Time of highest Depth temperature - Amplitude - Fluctuations angles temperature Diffusivity Computed Observed Damping a~ 0.05 a, Tank (cm) (°C) (a) accowgted for (°C) (br; min) (1) mm's~' (br; min) (br; min) (cm) (em) %o) May 3-6, 1966 Inside 0 30.3 9.92 80 318 13:49 0.21 5:02 4:19 7.6 23 10 27.4 2.66 89 253 18:08 0.45 15:27 12:05 11.1 33 45 24.0 0.17 72 137 01:54 0.59 10:30 7:46 12.7 38 Outside 0 24.6 6.11 79 308 14:27 0.47 3:22 2:49 11.4 34 10 23.6 2.55 89 266 17:16 0.75 11:58 11:14 14.4 43 45 21.6 0.27 89 140 01:46 .88 8:59 8:25 15.6 47 July 6-9, 1966 Inside ) 32.4 6.20 87 314 14:06 0.32 4:04 3:47 9.4 28 10 30.2 2.15 87 257 17:53 0.68 12:34 12:36 13.7 41 45 27.4 0.23 90 124 02:42 0.89 8:33 8:49 15.6 47 Outside 0 33.1 8.53 85 323 13:28 0.20 5:09 4:06 7.4 22 10 29.9 2.24 90 262 17:34 0.58 13:37 12:46 12.6 38 45 27.4 0.25 91 132 01:14 0.93 8:22 8:40 16.0 48 x 0.6 8 o1 0.2 0.4 1 3 "4 6 810 - -- < 3 | lt. 52. € g- - € E* 44 a # oal #7. fret |- -t 30 {§ w V+ 3 LQ a. z f >>, Near Tank 6 we 5 > e /* oa 6k 4 & (+/ May 3, 1966 w _ T 5 oc 4 £ !. » 2 a m 0 6 <120 KC a 0 + 3 4|- I al ® o 3 y it @ w co F- \ s -3 € 0.7 s a -<10 2 20 2}- \\ 0 € a AX 9 0 w X at co 2 3 [- o 5 09.4 x - z - s C gp --L--L- fj 1-1 NLL _L _I E I € Near T@nk 6. - = _| f+ 0.5 1.0 1.5 & e" July 5, 1966 T 8 o 0.6 + - 0.6 S SOIL HEAT FLUX: 10° CALORIES PER DAY 1 1 cd i Ltd PER TANK 0.1 0.2 0.40 6 1 2 4 68 0.2 0.4 1 2 A 68 f 10 0.6 10 FicurE 36.-Soil heat flux versus water use expressed as heat AMPLITUDES OF DAILY SINUSOIDAL WAVES, IN DEGREES CELSIUS FiGurE 35.-Relations between amplitudes (a) of daily temperature waves computed from data at 0, 10, and 45 cm depth, damping depths (-), and depths of 0.05 a («-) for two locations and two dates near Buckeye, Arizona. ground water. The heat flux in the tank using less water is larger than that in the other. Note that for better comparison, the water use is expressed in calories per day per tank or joules per day per tank. The slopes of the lines connecting a tank with higher water use to one with lower water use indicate that each time the water use decreases, the soil heat flux increases. The dashed line, which is the mean slope of all curves, shows that an increase of 10° units. Two tanks using different amounts of water in each period are compared. 1 : May 1-6, 1963; 2 : June 5-10, 1963; 3 : September 8-13, 1963; 4 : October 8-13, 1963; 5 : May 5-8, 1966; 6 : August 3-6, 1966; 7 : September 8-11, 1966. Dashed line: mean slope of all curves. calories per day per tank is the result of a decrease of water use of 6 x 10° calories per day per tank. Obvi- ously considerable energy is used elsewhere, but it was impossible to show this transfer inside the thickets at the Buckeye Project. TURBULENT TRANSPORT As noted earlier (see section on "Empirical Methods"), most of the theoretical and even some of the empirical equations to determine evapotranspiration F36 are based on the profile theory. We saw that at the Buckeye Project the vast majority of the wind profiles could be described by equation 31 with z, being the "adjusted" roughness length and d (the zero displace- ment) suppressed. Also mentioned were the transport constants (see equations 12, 13, 14, and 15) for momentum, heat, vapor, and carbon dioxide which, de- pending on the type of turbulence, may or may not be the same. If they are the same, temperature and vapor-pressure values plotted against windspeeds at a number of elevations should fall on straight lines re- gardless of the shape of the profile (Penman and Long, 1960). Figure 37 shows examples for 6 of the periods used in figures 24 and 30. Note that 4-hour averages have been computed to avoid embarrassing anomalies. Nonetheless there are some glaring ones, for instance period 6 of September 10-11, 1966, which apparently can only be explained by instrumental or observational errors. Also, the data for March 23-26, 1967, are quite erratic. During that period the actual evapotranspira- tion is low and the transport constants for momentum, temperature, and vapor pressure might very well not be the same. Swinbank and Dyer (1967) showed that the transport constants for temperature and vapor pressure are often equal, but they are not equal to the momentum transport constant. If this is so, then the deviations from a straight line should be the same for temperature as well as vapor pressure. This actually seems to be the case as shown by the data for March 1967; wind profiles for that period deviate most from the logarithmic ones. CARBON DIOXIDE During the growing seasons of 1960-61 it was ob- served that the rate of growth and development of saltcedar grown in evapotranspirometers is correlated with the use of water by that plant. The less vigorous the growth, the less water was used. This raised the question whether the growth rate affected the rate of transpiration or whether other environmental factors affected both. In one attempt to explain the diminish- ing rates of growth and development, van Hylckama (1963) mentioned that maybe the carbon dioxide had become the limiting factor, especially at times of low windspeeds. If, during times of vigorous assimilation, the CO, content were to drop from a mean of about 300 ppm to 100 ppm, rates of growth and development would be diminished. Hence, the use of water might be diminished also. Such observations have been reported by Bonner and Galston (1952). Also the fact that rates of photosynthesis and of transpiration are related is well known. Bierhuizen and Slatyer (1965) found that, under laboratory condi- tions, an increase in carbon dioxide content of the air to 0.10 percent or more results in less transpiration per STUDIES OF EVAPOTRANSPIRATION gram of carbon dioxide assimilated by cotton. One could say, therefore, that the plants become more effi- cient producers. But such conditions do not exist in the field, where the CO, content of the air rarely exceeds 0.02 to 0.04 percent (200 to 400 ppm). Pallas (1965), also under laboratory conditions, observed less transpiration when the CO; content of the ambient air is high. He noticed that the stomata of his plants closed when the CO, content went up to 4,000 ppm. One must assume that this impaired the passages of both CO, and vapor. Therefore, the CO;, content of the air within and above the saltcedar thickets was measured with an in- frared gas analyzer (van Hylckama, 1969). Figure 38 is an example of the results of such meas- urements. During the day, air turbulence caused mix- ing and the CO;, content remained about the same in, as well as over, the thickets. During the night, sharp gradients occurred as the CO,; content inside the thick- ets rose at times to more than 380 ppm. The figure also shows rather large fluctuations in CO; content at all levels and times. Much of this was undoubtedly due to turbulence and possibly also to shifts in calibration of the analyzer. When photosynthesis takes place, one can expect to find smaller CO,; contents inside, or near, the vegeta- tion than above it. Diffusion and turbulence tend to equalize the CO, content of the air at all levels just as they equalize temperature, water-vapor content, and other properties of the atmosphere. These fluxes can be expressed as g m- s' or cal m-" s-". There are several empirical methods for computing them, and the ones used here are: F, « _E. Uta - m (C. = Cy) (43) [In{ (a =- 2,) / (b =z) }J° for CO, and V k" (, = Hy) (ea (44) t [In {(a S 20) / (b = 20) } ]2 for water vapor. F,. is the CO; flux, k the von Karman constant (here taken as 0.41), u, and u, (m/s) are the windspeeds at levels a and b (m), respectively, C, and C,, the CO; contents (ppm), 2, is the roughness length (m); and V the vapor flux with e, and e, the vapor pressures (mb) (Lemon, 1960). To convert to g/m", it was assumed that 1 ppm CO; at 985 mb (98.5 kPa) and 25°C equaled 1.77 mg/m and 1 mb vapor equaled 724 mg/m*. Figure 39 is an example of CO, fluxes during the day. Although short-time fluctuations resulting from those shown in figure 38 were somewhat smoothed out by averaging 3 days of data, there still are considerable irregularities. WINDSPEED (w), IN METERS PER SECOND N WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F37 10 16 10 16 10 14 11 11 10 TET Fars Fat: 399 FTS 3013 fas- -q 9 May 5-6, - 4 |- May 5-7, - t 3 4 \5 1966 s 1966 ~ 3 4 6 e ke s \ \ 6 6 - 9 1 2 a \ < 3. 5 - % * x \\ _| _\\ \\\:\\\\.\\ & \\ 4 EI {AEE ttt t ATEX 3 iew ESE rda fet 11 14 10 16 10 14 11 14 10 17 11 16 32 35 29 34 16 25 30 33 29 35 16 -_ 23 Lata d ms Paa EYG 4 r- 9 -r f ferro T ACR T T3 E9 2 5 r _ zat 7 - A - 1 3 s - aI E / 3 \ 4 / 6 2 al 1 2 3 4 6 S a/s 17. ~ x" PA* /Z i F: \ /,/ ,/’ ces = pes Ny iil tea e*". e a PLLA 4 1p dei" ALG LA-~~ Etof ict t LAZ, Later LIA. ( S-; 41" 19 22 55 38 24 29 18 23 37 18 27 10 __ 14 18 10, 13162 9 14 10 17 10 o. 12 fera T HRert TF to fat ra 43 G T4 13 Ss I 4 |- May 7-8, - 4 |- Means of 3 days - | \ \ 1966 | 3 $ * $ J e Mo > \ *T \ \\ \ \ \ Nod ||x\|\‘1 wig)?” \' ill fA ACFT _ '111\|\’1 x 1 14 9 15 18.7 9 12.2 11 14 17 10 14 4-3 TF Ger rG Ear Ed- 10937 E A 5 T 4 |- T* 4 E C 4 5 - | 45 a JA pt i 3 c gA) 7 -; § +- > /A ¥ / . - its ' s i yo g in: / / | |'/|/| Lita > I Lt Lot f { 14 dl {LAT 4 4 r ES - ral 4 8 7 j 5 9 EF T T3 On CIR: TT OTT ST-T OTT F343 7 - 3 4 June 3-4, - 4 |- June 4-5, - t ‘\ \ 1966 \ \ 1966 1 2 e 5 # . \ m . : ‘ 5\ 6\ lae \‘ \ \ X 3 w N & \\ \ ® 7 \ "I rar ' Menta s 14 Brg c L141 LAME I 1.4 '||T‘-| s. 9 5 -_ B 6 -$ 4 10 6 9 30 33 1093, 17 23 30 ito.". 17 TT OT TOE 4 rt EAT 3 FTT Fort A4 EAT TOPI T (> 3 4 T 73 ~ 41 3 4 T+}: al S (fp rg. / 5 R - j / 5 6: 3 / 1 \ § e { 5 5 / / 3 . (I, / / 4 4 l £15 L 1 I | + nI LADA L 1 _J_'_;=L 1 _| L | b IInI l Et Ete" 29 1 jg: 35 147 50 _ 25 141 4 35 15.5 25 13 5 11 5 11 4 9 5 l TST T4 Rid Sr T4 TCH TT-1 119-93 |- \\ “3562-6, 1. .4 +- 3\\ 4 Means if 3 days - 7 \ \\ $* sL £ '\ §° 4. } \\ \\ | 3 \\- \\\ \\\\ \\ \. ] LL 4 l 1 {oto tots d t ted "1 Ole \\ 1 1 1'1 FALL 1 4 1 4 & 1 | ‘l N+ I \\. 27 11 4 11 7 11 7:30 10 6 10 31 124 18 29 32 29.1 33 11.4 \? Lii 19 E Pad VAT P Ty" Fd $13 T CET TT TTT |- 3 June 5-6, 4 _ 4 |- 8 7 4 f A 1966 § s Means of 3 days _ 1 2 1 2 § / § 3 5 l} 3 ~ 24 / ¥: / { I L- fea Fa < a > 3 \ y | EA LLA § l‘l E. e Elfin eZ 41—11 14 § Har 3 (AAL + / 16.83 22 37 16.2 73 27 15.3 19 36 155 41 26 19 FicurE 37.-Plots of windspeeds (w) in meters per second versus vapor pressures (e) in millibars and versus tempera- tures (T) in degrees Celsius at the same levels. Data presented are 4-hour means: 1 = hours 5-8, 2 = hours 9-12, etc. Note: e and T scales on the abscissa are sometimes broken (SS) to save space. The windspeed scale (ordinate) does not change. The e (mb) and T (°C) values at the top of each box refer to time periods 2, 4 (1700-2000), and 6 (0100-0400), those at the bottom to 1, 3 (1300-1600), and 5 (2100-2400). F38 WINDSPEED(w) , IN METERS PER SECOND STUDIES OF EVAPOTRANSPIRATION 13 13 18 9 TT TTS (Rac TTT July 6-7, ] § 2 3 1966 . *f \\ a t £. < I. ® i 20 28 3|5 33 37 ort T -I - T rr ' \3 - 4 |- : -I a / ye T - + ose y SLA ALAEA TLL ~ L'rllli L4 O . & 34 23 286 ©3909 25 32 19 12 16 9 T £. & < co & 1 p T 14 ‘\ lobe _-. a 2\\ 3\\\ 4\\ er ® 1 N aA 13 16 T STS Means of 3 days 29 38 r «s 'at \3 4 6 - G et ? / f l / fica at g" ba cl 5 / //6— a Fra 13 T £11 1 4 4 4 fill/l Ld t t d |'//| [-p 4 ; L134 £ |-{’/l Fd f b -’/ as 32 39 ' 27 34 7 5 7 Trat TTT TTT Ain Bt Mar. 23-24, - - L3 4 Mar. 24-25, - 1967 |. 5 1967 | § e A 2 h \ e 6 -_ (913) L : i \ L a 6 ? : } 3 yoy a iP. § ‘ i Alt LOE 4 11 | 3-1 8 1 19 23 26 24 13 16 FIT-3 3 I 1999-3 21 6 5 TT OT Means offiL‘ 6 days a= 4 1.1 10 5: ":e ////— yas l‘l’llllll 1 3 - 141 4 - - 1 \ f 6 2 x ond * ~ hes" ' EAA t t 4 {L141 13 19 26 30 7 6 TTT ETT 3 F3 . _ Mar. 25-26, . (- 3 \ 4 *~ ~6, 1967 1 \} n a/ \ 1 1 1 1 1 LOC t 7 6 12 21 24 22 26 GF [Illa CTE T3 C 3 4 . 4 . 93 rrr 5 16 8 15 26 30 +15 FigurE 37.-Continued. 21 WINDSPEED (w), IN METERS PER SECOND WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA 20 25 22 20. 23 20 24 15 21 20 23 tr 3 TTT f30 GF FST r a FF ETTTT EAST 6 |- 4 Aug. 3-4, 1 g} Aug. 4-5, - & 1966 | 1966 _| G e 4 |- § 3. "4° 4 F4 2 \ \ k 6 a 3\ j \ | if Ca fall & & \ & yar ce a ~" a By \ - % C + \ a E2 ; & \‘. No a Wa 4 14 LAM on R -f '11|| 20 24 20 24 20 23 22 17 21 17: 20 332 37 37 28 35 39 38 _ 42 28 31 tor Te -t T mort FI TTF 6 |- 4 T 1 ef T -I C 5 4 - 4 * i , / 4 3 4 5 = \ I 2 2 +- \ f F 3 2 - ; - - J \'§ \ i f // - E 2 \\ * 4 7% ,/ * 11,11 R LOLOL _ C4} f. '/1| ‘Illll'l'l |/||11 30 323 37 40 33 28 40 45 20 24 19 23 29 20 24 18 22 21 24 6 TCT TAT TOT OT 5 L 6 CT TTI NCS fF . Aug. 5-6 4 e 4 R 4 Means of 3 _| 1966 3 5 _ days 4 3 3. 4C \\ > 7 C \\ I _ L- - 9 \\ 5 I 6 4 2 \ 6 fast f s at v yat E x 3 \ I £ 4 -or s \, § & 1 I'| LAOH 1 1 | ~ 1 1 ‘I N A d 44 L4 4-4 > 21. 24 33 22 22 0. .: 23 36 40 39 29 36 _ 40 39 40 2! 6 NF- 4 V4 T 5 rta 6L or TTT 3 TTT & e 4 A 4 |- 3 ~ 4 |- \ ig > 2 \ [I] I: 6— [- 1 2 I, b 6 \{] 3. __ ~. 34 fod f! 1. ~ yl} aya } ] LA {LL 4 L 1 JL (G4 E4 41 28 44 33 29 40 _ 44 32. 356 16 15 18 16 15 16 13 TAT T Taa TT 3 P TST d ah \ 3 Sept. 8-9, 1966 | _ | 1 _ 2 | a |. 5 2 |- k } \ 4 51" e Is | k \ ; \ g 6\1966 f i . sx \ |:. [ \ ; # 2 143, free x H 1 E44 \ ita \ 15 14 14 17 16 14 16 33 36 31 36 19 25 34 37 31 36 19 TAT T ri TTOT T meus T |- \ 4 \2 3 s E 1 f f PH y \% /4 / / ay 5 \\ / / / 421,111! |11.|" II'JIIlIIJ'f '/|I LOL _ ||11||y./ 20 25 38 40 22 30 21 24 38 41 23 29 14 13 16 14: 17 15 13 16 16 an an erp : an ind me AF TT -10 PSF tA A 3 SED:9:5% 11, L g a s Means of 3 days - 4 4 - 6 L e 316 # _I_L\1 E sp ix 1 | L I 1 4 1 1 1 St |\|’1 \ 13 14 1 14 15 ° 14 16 33 36 21 28 35 31 36 19 26 lns HU ICT 2| T |3 .T m 5 € : X 1 2 f a / '% 6 1,\ \ 4 "I / x \ P22 / | | / 7 / // 1177—1 4 4 \'| y1 I/,~’ r"; af j/I Etot I F,” F3 acd la" 19 26 39 2: 33 20 25 38 41 22 30 Figure 37.-Continued. F39 F40 t 0 =] t ® 0 CO, CONTENT OF AIR IN PARTS PER MILLION w t N p o [=] & o o J T 1 IN METERS PER SECOND to T 1 WINDSPEED, n T 1 |- TRUE SOLAR TIME FiGurE 38.-Top. Fluctuation of carbon dioxide content (parts per million) in and over a stand of saltcedar near Buckeye, Ariz., June 3-5, 1966. Carbon dioxide content at 6 m, at the top of the vegeta- tion, 3 m, and at ground level. Bottom: windspeed (meters/second) at 7 m above the ground. 10leIIIIIIIIIIIIIIIIIII ///&First harmonic / First and third harmonic CARBON DIOXIDE FLUX, IN GRAMS PER SQUARE METER PER SECOND x 103 1 N T -4 - - -6 - - -g - % 10 - 3 ~12 1 I C3 _L i l __ 3.4 J_l O4 3 l £24014 6 12 18 24 6 TRUE SOLAR TIME FicurE 39.-Carbon dioxide fluxes in a saltcedar thicket near Buck- eye, Ariz. Observed data, mean of June 3-5, 1966, compared with the computed first harmonic and with the combined first and third harmonic. By applying harmonic analysis, the data can be smoothed out further. Figure 39 shows that the first harmonic alone represents the data rather well, al- though by adding the third (8 hour) harmonic a sur- prisingly faithful picture is obtained. The reason for this has yet to be explained. Using first harmonics of both water vapor and CO,; fluxes an interesting curve is obtained showing the STUDIES OF EVAPOTRANSPIRATION correlation between the fluxes as exemplified in figure 40. The vapor flux, except during times of dew forma- tion, would always be negative, but it is more negative during daylight hours, when the CO, flux is positive and much less so during the night, when CO,; flux is also negative and large amounts of carbon dioxide are formed inside the thickets as was shown in figure 38. Figure 40 also makes it possible to estimate the water-use efficiency (the amount of carbohydrate pro- duced per unit of water) by considering the amount of water transpired per unit of CO; absorbed. Ratios of over 600 occur in the early morning when light inten- sity is low, while values around 160 occur during the middle of the day and increase rapidly towards the evening. These values agree with data presented by Arkley (1963) and Bierhuizen and Slatyer (1965). PREDICTING WATER USE USE OF LONG-TERM MEANS For the purpose of estimating how much water may be saved by managing riparian vegetation, it is usually and reasonably assumed that riparian vegetation transpires at the potential rate. For flood-plain vegeta- tion less close to stream or lake, the situation grows CARBON DIOXIDE FLUX, IN GRAMS PER SQUARE METER PER SECOND x 10° 24 ( -10 -60 -50 -20 VAPOR FLUX, IN GRAMS PER SQUARE METER PER SECOND x 10° FIGURE 40.-Relation between carbon dioxide and water vapor fluxes during a day over a stand of saltcedar near Buckeye, Ariz., means of 3 days June 3-5, 1966. Ordinate: CO, flux to- ward the vegetation positive; away from the plant negative. Abscissa: water vapor fluxes away from the vegation negative. Numbers indicate hours of the days. Flux data are quantities resulting from harmonic analysis of the fluxes and represent first harmonics only. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA more complex. Salinity of soil moisture, increasing depth to ground water, topography, and exposure af- fect, and usually diminish evapotranspiration. The presumption of potential evaporation is then no longer valid. However, in most cases, the approach to the es- timation of actual evapotranspiration is through the potential. The main interest therefore remains to pre- dict or be able to predict potential evapotranspiration. The foregoing description of the microclimate of a saltcedar thicket shows it to be a very intricate phenomenon. If one wants to know how this "compli- cated mess we call a canopy" (Tanner, 1963, p. 148) affects evapotranspiration rates, all variables perti- nent to the problem will have to be taken into account. However, models constructed for the prediction of water use often do not include a sufficient number, let alone all, of the variables that determine actual or po- tential evapotranspiration. Many models have been used, and even more proposed, that attempt to simplify the computation; often only one or two variables are used because of a lack of more information on the one hand and perhaps an overanxious desire to predict on the other. A few examples will be presented, and the results compared with one another and with data from the analyses of water use by two Buckeye evapotrans- pirometers. No detailed description of the methods are given because they are readily available in the litera- ture (see, for instance, the discussions by Cruff and Thompson (1967) and by Jensen (1973) ). Table 7 lists the references and also the variables used by the var- ious authors. Note that temperature (T') and radiation (R,) are the most frequently chosen variables. R,, when not measured, is supposed to be taken from climatic tables. Actual hours of sunshine relative to possible sunshine duration for that latitude and time are occa- sionally used instead of radiation. Vapor pressure (e) TABLE 7.-Variables used in some models for predicting potential evapotranspiration [T: mean temperature or maximum and minimum; e: vapor pressure, relative humidity, saturation vapor pressures (from T), wet-bulb depression or saturation vapor density; R,: sqlag radiation; Lat: latitude term to indicate duration of sunlight; and w: mean windsp Other N a R, Lat u Lowry-Johnson (1942) PRagleman (1967) ............._.... Jensen-Haise (1963) ________------ Lane (1964)..........:.. Thornthwaite (1948) - Ture (1961) Blaney-Criddle (1962). . .. babe be babe bebe Crimp Factor Weather Bureau (Kohler and others (1955) ...... Christiansen-Hargreaves (1969) 22 24.2 ALL. AAG gens an anis X Morton (1975) .......c..c..c« Papadakis (1966) --- ove Makkink (1957) --- Hamon (1961) .._...---. Olivier (1961) .___..._.. X 5% Albedo MMMMXMDE > MML | F41 can be computed from temperature and relative humidity records. Sometimes saturation vapor pres- sures are used; these are obtainable from temperature data and vapor-pressure tables. Such models are incapable of depicting the (often sudden) changes in any of these and other variables in and over the vegetation. They are typically applied not to short-term periods (although some of their authors so do), but to average monthly data, preferably taken as means over a series of years. Figure 41 gives the mean potential evapotranspiration per year as com- puted by the models mentioned in table 7; the results are compared with measured data taken as the average from tanks 3 and 5. These tanks had, between 1961 and 1965, the smallest depth to water, and, without a doubt, evapotranspired at the potential rate, the more so, because the soil-moisture salinity was also low dur- ing that period. Effect of depth to ground water and of soil-moisture salinity will be discussed presently. Another way of comparing models with measured data is by smoothing out the fluctuations throughout the year by using harmonic analyses. This is done for the year 1963 as shown in table 8. It is obvious from the percent column that the first harmonic describes the 3000 |- C [< -I m E- e grg —1a__ G _-§ '@ ___ Measured water use-means of 947 984 A y - a & 4 years and two tanks T -| £ #5 sa 1° E 2000 [- ——————————{j w s |- 1787. - 4 e 6s1 69 E f s > 60 a 55_I_55_I_ I- _ 50 The numbers express the estimated use - 1000 ‘I‘48 in percentages of the measured ones E § &oS gos £ $% sg ® hea: ~ € > yo 4° £ 2 5 a 9G © I p C o f (s) 5 € " g 5° r i a' 3 3 a 2 g a. = S FIGURE 41.-Water use (millimeters per year) as measured and as computed by various models. The horizontal lines indicate the av- erage use computed from 4 years of data and the vertical lines indicate the standard deviation. The measured use is computed from the means of tanks 3 and 5, which had, from 1961 to 1965, the smallest depth to water (1.5 m). The references are: CWT: Thornthwaite, 1948; Hamon, 1961; Makk.: Makkink, 1957; L./J.; Lowry-Johnson, 1942; Ch./H.: Christiansen-Hargreaves, 1969; Oliv.: Olivier, 1961; Turc, 1961; W. B. lake: Kohler and others, 1955, lake evaporation; B./C.;: Blaney-Criddle, 1962; Lane, 1964; J./H.: Jensen-Haise, 1963; Eagl.: Eagleman, 1967; Mort.: Morton, 1975; W. B. pan: Kohler and others, 1955, pan evaporation; Pap.: Papadakis, 1966. F42 STUDIES OF EVAPOTRANSPIRATION TABLE 8.-Comparison of different methods of estimating evapotranspiration for 1963 using the harmonic analysis data [/ = mean and a, = amplitude of first harmonic, both in millimeters. 6, = phase angle of first harmonic (in degrees); % = percent of all fluctuations accounted for by first harmonic and peak 1 = approximate day of maximum use, computed from <,; total/year = total water use rounded to tens of mil and 5 which had the smallest depth to water (1.5 m)] imeters. Measured use was taken as the mean of the use measured in tanks 3 Total a per Method h a, % d Peak year Measured .-. 2:;..-: ._ 190 104 82 277 July 8 2280 Thornthwalte <1 -.... -.:. 97 104 87 277 8 1170 ence 102 92 83 271 14 1230 113 57 89 285 June 30 1350 Lowry-Jolinson en. No monthly altes - - =-- s 1300 Christiansen-Hargreaves ___________---- 130 77 90 287 June 28 1560 {L 133 110 85 288 27 1590 TUIC arr inanna ann 141 84 86 288 27 1660 Lake evaporation Kohler and'others 139 90 91 291 24 1660 Blaney:Criddle 152 68 89 272 July 13 1820 Lane:. ln ell 172 114 90 278 7 2070 177 128 90 276 9 2120 EagiemAN . sauces 195 141 76 264 21 2340 Morton... cel - aad 195 105 89 282 3 2350 Pan evaporation Kohler and others ....:..:z..--....- 204 126 91 286 June 29 2450 Papadakis 254 140 84 269 July 16 3040 fluctuations accurately enough for all practical pur- poses. The second and third harmonics (not shown in the table) contribute from less than 1 to almost 5 per- cent to the fluctuations. Also the phase angles, and hence the time of maximum use, are quite similar for both measured and computed use, but there the simi- larity ends. The means (A) of the Eagleman and Morton models are closest to the measured means, but the amplitude (@,) of the Eagleman model is much larger than the measured a,, while the Morton model shows an amplitude about equal to the measured one. Russian authors are notably absent in table 8 and figure 41. There is good reason for this. Such methods as proposed by Budagovskii and Savina (1956); Ognev and Kozlov (1962); Konstantinov (1963a); and Bavina (1967) are all based on the one proposed by Budyko (1948; 1956). Budyko's method is basically a combina- tion method similar to Penman's (See sections on "Bulk-Aerodynamic Methods" and "Energy-Budget Methods" in this report.) Since such methods are essen- tially models of instantaneous phenomena, they cannot successfully be used for daily or monthly averages, but one will be discussed in the section on "Short-Term Predictions." USE OF DAILY AVERAGES It was earlier explained (see section on "Water Use Measurements") how hourly water use could be ob- tained from the on-off recorders and the watermeters connected to the evapotranspirometers. These data, too, can then be smoothed out by harmonic analysis. An example is given in table 9. With one exception, all first harmonics account for more than 90 percent of the fluctuations. Tank 4, which before 1966 had the same density of stand as tank 2, was thinned in March of that year to 50 percent of its original density, and the data show the slight effect of this treatment on water use, except in March 1967 when instrument failure was, if not the likeliest, the easiest explanation for the difference. Table 10 shows some of the climatic data for the spe- cial observation periods mentioned in table 9. The daily mean temperatures (T,,,) did not vary much dur- ing each of the 3-day periods, but there are consider- able differences in radiation (R, and R,) and in windspeed. It is also of interest to note that, whereas the air temperature during the August period was by far the highest, the radiation was lower than in any of the earlier months. This explains in part the drastic decline in water use during August as compared to May, June, and July. Since most of the empirical models have mean temperature as their main input, one could expect a better fit between measured and computed data in that month than in any of the others. Table 11 shows this to be the case. Six of the 15 models in table 8 have been used here to estimate evapotranspiration by days because only those models were used whose authors have applied them to esti- mate daily evapotranspiration values. As in table 8 and in figure 41, the fit is very poor, the least poor one being in August. The reason for discussing some of the methods for estimating evapotranspiration was to emphasize the fact that a saltcedar thicket is a complicated entity. It is subject to (and reacting to) continuously and at times rapidly changing inputs. The three empirical models WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F43 TaBLE 9.-Water use in tanks 2 and 4 during special observation periods with some results of harmonic analyses of the hourly data [Means of three day, mm per hour. J = mean; a; = amplitude of the first harmonic (mm per hour); % = percent of total fluctuation due to first harmonic, 6, = phase angle of first harmonic, rounded to whole degrees. Peak = time of maximum use computed from 431] Total Mi pet Period Tank h r % di Peak day (mim) 1966 May 5-8... o :l ecco clon eco ca n 2 0.70 0.47 91 324 13:25 16.8 4 N 0 D A T A USF an cung aes eel a ale 2 0.80 0.54 ~f 8 325 13:21 19.3 4 0.73 0.43 94 324 13:25 17.5 JalyiM=0 ....c:. L L olla 2 0.87 0.65 93 341 12:17 21.0 4 0.86 0.60 92 335 12:41 20.5 AUp 0-0 ._. Lc. CL. cell Jigen educeateds 2 0.51 0.34 90 334 12:45 122 4 0.46 0.25 92 334 12:45 11.1 " lc e eB ana nes ances 2 0.50 0.44 88 342 12:13 11.9 4 0.51 0.38 92 339 12:25 12.2 1967 Mar 23-20) - .._... Ero 2 2.3 No hourly values available 4 0.9 MayAl-14 2 0.36 0.22 95 327 13:18 8.7 4 0.35 0.24 94 343 12:09 8.4 TABLE 10.-Some climatic data pertaining to seven special observation periods of 3 days each [R,: total incoming short-wave radiation in langleys per day (ly/d), watts per m* (Wm ~*) and millimeters water equivalent per day (mm/d); R,,: same but for net radiation; T,, and T,,; dry and wet bulb temperature at 4 m above the soil (°C, mean of 24 hours); e,: vapor pressure at 4 m, determined with fig. 18 (1 kilopascal = 10 millibar); u, windspeed at 4 m (kilometer per day); rh: relative humidity (percent)] R, R. Ta Ta: €, u rh Period ly/id Wm ~* mm/d ly/id Wm ~* mm/d "C °C kPa km/d % 1966 Mayib-5 1 593 287 10.2 387 187 6.7 28.6 17.0 1.21 144.6 _ 31 2 641 310 11.0 410 199 7.0 27:1 17.2 1.28 93.2 34 3 596 289 10.3 374 181 6.4 28.9 16.7 1.13 123.0 28 1 716 347 12.3 414 200 7A 25.3 12.5 0.66 121.4 20 2 713 345 12.2 415 201 7.1 25.5 13.3 0.76 122.7 23 3 699 338 12.0 416 201 7-1 26.7 14.2 0.85 115.9 24 1 668 323 11.5 356 172 6.1 32.6 16.7 0.91 130.4 18 2 579 280 10.0 342 166 5.9 31.8 19.1 1.42 169.7 30 3 655 317 11.3 415 201 7A 32.4 21.0 1.71 140.9 36 Augrd-B 1 417 202 y 260 126 4.5 35.1 23.6 2.19 193.6 39 2 580 281 10.0 382 185 6.6 35.6 22.8 1.97 176.7 34 3 525 254 9.1 344 167 5.9 35.1 23.6 2.18 170.2 39 Sept 1 539 261 9.3 275 133 4.7 30.2 19.4 1.57 115.2 37 2 482 233 8.3 240 116 4.1 30.3 19.1 1.50 126.3 35 3 480 232 8.3 224 108 3.9 31.4 18.7 1.37 101.1 30 1967 Mar 22-260. 1 421 204 7.2 179 87 3.1 20.8 10.2 0.58 113.0 24 2 460 223 7.8 178 86 3.0 20.9 10.5 0.62 198.1 25 3 508 246 8.7 251 122 4.3 16.2 8.7 0.65 106.2 35 1 685 332 11.7 385 186 6.6 20.2 10.8 0.71 133.7 30 2 662 321 11.3 375 182 6.4 20.5 10.7 0.67 162.3 28 3 685 332 11.7 358 173 6.1 19.9 9.9 0.60 187.5 26 using monthly mean values that give estimates close to the measured values were also the ones that used four instead of two or three variables as inputs (fig. 41), but For short-term predictions the above mentioned em- these three models are not supposed to be valid for | pirical models have no fit. On the other hand, models short-term periods; hence, they were not tested against | capable of predicting water use for hourly or even daily water-use data. shorter periods should not be applied to monthly or SHORT-TERM PREDICTIONS F44 STUDIES OF EVAPOTRANSPIRATION TABLE 11.-Water use during special observation periods of 72 hours each as measured in tanks 2 and 4 and estimated according to various equations [Water level in tanks 2.1 m below the surface. Means of 3 days, millimeters per day) 1966 1967 Method May June July Aug. Sept. Mar. May 5-8 3-6 6-9 3-6 8-11 22-26 11-14 Measured in tank 2 _.... -_. -_- n ns 16.8 19.3 21.0 12.2 11.9 2.3 8.7 Measured in tank d -__-... z.. --. _ lentil Gunes 17.5 20.5 11.1 12.2 0.9 8.4 OlIvIer: nc ee oen ein 9.0 10.4 A 9.2 7.8 6.3 6.3 Jensen-Halge "ic..._.c2en.ciccc ccc nie cuse ate ne aa ngs i 8.3 8.9 9.7 8.5 7.3 4.4 6.8 Christiansen-Hargreaves 5.5 6.1 6.5 5.9 4.7 3.5 5.4 >. . £- neige ceca iene nea nene wake 5.7 8.1 6.9 7.0 5.8 2.3 2.8 Hamon ccc. Clin lorn iaa ee eo aol aul 5.1 4.8 6.9 71.2 4.8 2.4 3.2 Makkink : - cl ene ct ec re a mas pink 4.8 5.3 4.7 4.3 3.6 2.3 4.9 weekly periods by inserting into the equations mean monthly or weekly inputs such as mean temperature, mean windspeed, etc. They can, of course, give good results by adding 24 hourly values to give daily quan- tities of water use, and a compilation of 30 daily values will give a good monthly value. The better a prediction equation represents the ac- tual proceedings in and over a saltcedar thicket, the better the accuracy of prediction. The principles and theories of the eddy-diffusion, the aerodynamic, and the combination methods were discussed earlier (see the section on "Evapotranspiration and its Calcula- tion"). Since no instrumentation for it was available, the eddy-diffusion method could not be tested at the Buckeye Project. Also energy-budget methods that do not include advective terms are omitted in the follow- ing discussion because, as was shown in the sections on "Climate" and "The Microclimate of a Saltcedar Thick- et", advection is a very important and persistent part of the proceedings in and over any riparian or flood-plain vegetation in a dry and therefore usually windy cli- mate. Finally, the Bowen ratio could not be tested be- cause of its erratic nature, as was discussed in the sec- tion on the "Energy-Budget Method." Three methods for predicting short-term water use will be discussed: one by Konstantinov (1963b), one by Rider (1957), and one by van Bavel (1966) which is modified by the inclusion of resistance terms as de- scribed by Monteith (1963; 1965). Konstantinov's'' method is based on aerodynamic approach including gradients of windspeed and vapor pressure and a Richardson number defined as Tog ea Tin (Msn = Ups)" Ri = -0.078 (45) This is similar to the stability ratio of equation 27. The subscripts indicate the height (in meters) of the in- ' The reason for including this model in the present discussion is that Konstantinov considers his methods to be an improvement to be used with the official GGO manual of 1957 (Glavnaya Geophysicheskaya Observatoriya = Main Geophysical Laboratory) for making observations by the gradient method. struments above the soil. These heights are much too low for saltcedar studies, but Konstantinov gives in- structions on how to reduce higher observations to lower ones. To compute windspeeds, for instance, we have nip. S Ae fl _ __ log (CGxz1) £0 Uza.q ~- UWo.2 where z, and z,, are elevations other than 2.0 and 0.2 m. Of course, such a transform is only admissable if there exists a logarithmic wind profile throughout. Such a profile may not always exist over a saltcedar thicket. The computations give results that, hour by hour, deviate considerably from the measured values. For 4-hour periods the outcome is a little better, but still leaves a lot to be desired. Figure 42 is a small example. The actual temperatures, windspeeds, and vapor pres- sures were taken from the 7 and 5 m levels and re- duced, according to Konstantinov's instructions, to 2 and 0.2 m. This procedure may not be valid where tall vegetation is involved and may well be the reason for the poor fit. The daily totals for the day, illustrated in figure 42, are 24 mm for the Konstantinov model and 20 mm as measured. The correlation coefficient, 0.72, is only significant at the 5 percent level. In 1957, Rider published a study of water losses from various land surfaces in which he used the aerodynamic method for predicting evapotranspira- tion. His model was based on the Thornthwaite- Holzman (1942) equation as modified by Pasquill (1949): 346 x 10 * (G,; (4; =') (47) T [In g- d ] d 24 ~- E0 = A comparison with equation 16 shows that the modification consists in the introduction of a zero-plane displacement, d, (discussed in the section on "Winds"), air temperature (T, in degrees Kelvin) instead of air density, and vapor pressure (e) instead of specific hu- midity (q). WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA PREDICTED WATER USE, IN MILLIMETERS PER 4 HOURS to I 0 1 | 0 4 2 3 4 5 MEASURED WATER USE, JIN MILLIMETERS PER 4 HOURS FiGurE 42.-Water use as predicted by the Konstantinov model (or- dinate) versus the one measured in tank 2 on June 3-4, 1966. Correlation coefficient = 0.72 *. Regression y = 0.76 + 0.88x. The numbers at the dots give the 4-hour periods 1 = 5-8, 2 = 9-12, etc. Rider determined hourly values for 18 days for which he had measured hourly water-use values available. The fits are only fair, the main reason being, by Rider's own admission, the quality of his records. Also equa- tion 47 as well as equation 16 are strictly valid only under natural conditions of atmospheric stability. The often successful method of smoothing out erraticism is to use sums of a certain number of values and (or) to use means of a certain number of days. This was done here for 4-hour sums and 3-day means. Two examples are presented at the top of figure 43. The correlation coefficients are at the 1 percent level of confidence and the regression lines show a close to 1 : 1 relationship. The two curves at the bottom of figure 43 show the result of computations for only one day. On the right the air temperatures and humidity data are taken from instruments 7 and 4 m above the ground (4 and 1 m above the vegetation). On the left, the data at 11 and 5 m were used (8 and 2 m above the vegetation). The method is rather sensitive to the height at which the observations were taken, but both curves show a fairly good fit. Nonetheless, the instrumentation used at the Buckeye Project was not accurate or sensitive enough for a successful application of the aerodynamic model. The combination method promised to give a better prediction, therefore, it was subjected to more elabo- rate testing. The model itself has been presented in the section on "Combination Methods" in this report. (1 Je T °"B 1.939 "I_ 1" 13 7 - sL _ w r = 0.89 * & 5- y=9.34 +0907 aay. = 0 y = -0.75 + 1.16x E 5|- -|- - < § +- -- - a_ 2 s} ag & [m E 2 |- Means of 3 days -- - é 4 May 5-8, 1966 Means of 3 days a -. June 3-6, 1966 3 L2 L _ 1. -H £4 {t "4 = /| 1s" | S34 -H | s #21 71-9 2 6 r = 0.90 # r = 0.87 #5 'I-I_J y =-.52 + 1.11% y = -0.85 + 1.26x « 5 7r- R ag 3 e § +t a- - [— 2 3|- -- s o E 3 |_ 5-11 gradients _| _ f - a June 7-8, 1966 <% 4-7 gradients 1L- « Fal om June 7-8, 1966 _| -t ___A f_ _L _s 1. 2:3. :4 -5. 6 1 2: 3. 4 -b 46 MEASURED WATER USE,IN MILLIMETERS PER 4 HOURS FigurE 43.-Water use predicted by the Rider model versus that measured in tank 2. Gradients: the numbers indicate the eleva- tion above ground from which the input data were taken. Equations 22 to 26 are especially pertinent. A large part of the computations were done by longhand, using graphs and nomograms. Later a small programmable desk calculator was available. In both cases computa- tions were simplified by assuming a constant air pres- sure of 98.3 kPa. In that case, the A/y term in equation 22 is dependent on air temperature alone and it was found that A/y = exp 0.045 T., - 0.32 (48) was a very good approximation. Similarly, the first fac- tor in equation 23 could be made dependent on air tem- perature alone and %z 36.4 x 10° + (T, + 273) (49) proved to be of excellent accuracy. As van Bavel (1966) points out, the expression for B, in equation (23) is based upon standard wind profile theory and applies strictly to adiabatic conditions only. Under such conditions logarithmic wind profiles ob- tain. Analyses of the wind data from the Buckeye Proj- ect revealed that within the limits of accuracy of the anemometers the profiles were nearly always FA46 logarithmic but a variable modified roughness length had to be taken into account as was discussed in the section on "Winds." Inserting this modified roughness length, and using temperature and wind data from the 4-m level, gives computed values that are reasonably close to those measured. The results of two computations, using the combina- tion equation and covering 72-hour periods in May and June 1966, are displayed in figure 44 together with the water-use data. As can be seen, the computed potential evapotranspiration rate not only fluctuates much more than the one measured for tank 2 but, especially dur- ing the daylight hours, often seems excessively high. That the measured data do not fluctuate as much as the theoretical ones is partly due to the slow response of the vegetation to comparatively sudden changes in the ambient parameters and also due to the assumption under which the hourly water use in the tanks was computed; namely, a constant rate of water use be- tween one filling and the next. A perusal of these and other data shows clearly that large deviations occur when windspeeds are more than about 2 m/s and during periods of high temperature. This suggests the existence of a significant aerodynamic and stomatal or leaf resistance (see the sections on "Flow Resistance" and "Combination Methods" in this report, especially equations 24, 25, and 26). The aerodynamic resistance (r,) can readily be com- puted if windspeeds and roughness lengths are known. The stomatal resistance, however, depends not only on r, but also on A/y and on the ratio of potential over actual evapotranspiration. Stomatal resistance can be measured directly on broad-leafed plants with a porometer (Stiles, 1970; Byrne and others, 1970) but - MILLIMETERS PER HOUR 18 24 6 12 TRUE SOLAR TIME 6:12 1824 G - 12 18 24 FicurE 44.-Hourly rates of evapotranspiration for two 3-day periods near Buckeye, Ariz. Measured water use in tank No. 2, compared to that computed with equations 22 and 23. STUDIES OF EVAPOTRANSPIRATION these instruments cannot be used on the scalelike leaves of a saltcedar twig.'' However, E,, can be com- puted with equation 22 and E,, can be measured. Thus values for the stomatal resistance (r,) can be obtained by the use of equation 25. This was done for one set of data, and it was found that r, was highly correlated with windspeeds and also, but much less so, with tem- peratures and humidity. Equation 25 shows that negative values appear when E, < E,. This can happen if (1) the graphical analysis of hourly water use results in a high E,,, (2) an erratic set of (say) temperature records gives a low E,, or (3) if E,, is actually larger than the potential value. The data analysis showed that vapor pressure deficit affected r, very little, the effect probably being over- shadowed by that of temperature and wind. It was found that r, = -1.26 + 0.018 T; + 0.5 u; (50) It may seem strange that there is a negative constant in the equation but this is the result of the fact that T and u values were taken at 7 m above ground and not at the surface of the leaves where temperatures are higher (Gates, 1963) and windspeeds lower (Monteith, 1965). Thus, equation 50 is purely empirical, being de- rived from the 3-day data of August and then applied to other 3-day periods. Within the range of temperatures and windspeeds used in the regression analysis, r, ranges from -0.7 s/cm for T; = 8°C and u; = 0.9 m/s to 2.9 s/cm for T; = 42°C and u; = 7.3 m/s. Hughes (1972), who used average monthly data, re- ports slightly higher r, values and finds that r, = 0.128 us "/T,""", but restricts his analyses to temperatures between 12° and 24°C and windspeeds up to 3 m/s only. His relationship thus shows that r, increases with in- creasing windspeed, as in the present study, but de- creases with increasing temperatures (at least in the quoted temperature range) in contrast with the present findings. Whether the use of porometer data or of data ob- tained by other means of actually measuring the resis- tance term would give better results is a matter of speculation, but the data computed by equations 26 and 50 and illustrated in figure 45 seem quite accept- able with the glaring exception of those from the period May 11-14, 1967, shown at the bottom of the figure. Here salinity had a decisive effect. In 1974 van Hylckama discussed the effect of salinity on rate of water use and explained how some of the "D. C. Davenport (oral and written communs., 1977) reported successful use of the porometer on saltcedar twigs by careful standardization of twig length and volume, but admits that the method is rather subjective and chances of discrepancies increase if differ- ent people use the method. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA 2-0TTII—[ll1lllTlIIIIIIIIIIIIIIIITIIIIII May 5-8, 1 1.5 - Way's 966 X’x)< Computed-_,, x a * a** 'x " °./Measured fin}, ® 1:0 ;- Rd a x3. ya Xxgg y x ® e 0.5 5.5"? Condo faust" I Pie llJJLLLJlIIIXlLIILilTXTIfllllllllllll TT OT I THY +3 pin an ion tat I SRA I Marae f I -d I TT 2.0 |- June 3-6, 1966 < 1.5 f- Xfi‘k Computed\:: s Measured\,§§“' = gle, AT. aa *> % = ..o'0¢ xx)“ P Kg, “I 51; U L 0.5 X_xxx ‘kdé‘: x* xl’wfif $82,152!“ g (eterno el ee plore c ere Lisa itl} O [403 4 I fop I P94 9-177 I frass I TOT I 2.0 |- July 6—9 1966 f CC Wo 1.5 |- 9; & /Computed .-°"./Measured <";.. w o >< "*. xs ® Xx- —I ¢ 10- XX ** * JP %x *x z me u & x & *x *' x *A . x E 0.5 [*x z“. x lo..><’°‘ ”52:41“. x 0.0.f3—I e xX * Ager" x X x 5 3 L3 t cip T a cop aul t tt ort rat tt iri {r j m 113 I 1 4-T A I FTI Y I a T 451 I T 349 I Drama I T4 CI heal 5 2:0 |- Sept. 8-11, 1966 ar , *x-Computed x 23 U 1.0 [- __ mets e* > ~*", A e % gk Xx Tart" + 0.5} .>2< # ® ixx 3 s 52 J x, - "s xx * x Nt B ea k. "5090 3k **2 xyzfi‘ Pes 2.0 I- May 11-14, 1967 7 1.95, y x, Computed, Adpusted\Q 20, ZA 1.0 |- x0 6$° Measured “XXXpoo 00x y x°oo -I O 0.5 Iii-2X .an 00. omooxxx... omoobd ‘. nugggggc’poo— % i | |||1|'?T"’)‘||||1|||+'Y'P|||||l|1|+|| I2 24 12 24 12 24 Figure 45.-Computed evapotranspiration, using equation 26, versus that measured in tank 2 at the Buckeye project in mil- limeters per hour for five periods of 3 days each. The ©0000 (ad- justed) line on the bottom graph compensates for salinity buildup (see text). tanks were flushed in 1965 and 1966 to reduce the sa- linity of the soil moisture and the "ground water." In 1967 the tanks were given no such treatment. The rec- ords showed that the specific conductance in tank 2 increased between September 1966 and May 1967 from 18 to 32 millimhos per em at 25°C. Such an increase would result in a decrease of water use by about 50 percent. Therefore the measured water use, indicated by the solid line in figure 45, was doubled as depicted by open circles. Although the corrected values do not fit ideally, they obviously compensate for drastic effects of soil-moisture salinity on water use. F47 It was mentioned (see section on "Short-Term Pre- dictions") that models capable of predicting water use for hourly periods should not be applied to longer periods by inserting long-time means into the equa- tions. An example of the result of such faulty procedure is given in table 12. Daily totals of evapotranspiration computed as sums of hourly values fit the measured ones quite well, whereas those computed using daily means of radiation, winds, etc., are much too large. DISCUSSION A few important conclusions can be drawn from the foregoing. Flood-plain vegetation in the arid and semi-arid southwest is characterized by a very complex set of interactions between inputs of ambient weather, soil, soil-moisture characteristics, and plant reactions to such inputs. The instrumentation at the Buckeye Project had been used with adequate, if not excellent, results when applied to measuring evaporation from open water, but it was less adequate to cope successfully with the com- plexity of a riparian or flood-plain vegetation and its evapotranspiration. For a rough estimate of yearly water use, the Weath- er Bureau's pan evaporation method (Kohler and others, 1955), and the Eagleman (1967) and the Morton (1975) models seem adequate. That Kohler's pan- evaporation, rather than lake-evaporation method, produced values closer to the measured quantities em- phasizes the oasitic characteristics of riparian vegeta- tion where the advective term is very important, espe- cially when, as is often the case, wind speeds are high and humidity is low. For detailed estimates or predictions, the combina- tion model (equations 22-26) gives the best results al- though such estimates often will be too high if aero- dynamic and stomatal resistances are neglected. Ri- parian and flood-plain vegetation simply are not sets of wicks that release water on demand. The disadvantages of the use of closed-bottom, plastic-lined tanks for measuring evapotranspiration rates have been discussed elsewhere (van Hylckama, 1974, p. E29). Better lysimeters would not only be flood-prone, if they were properly located, that is, within the riparian vegetation, but they would be very TABLE 12.-Water use (millimeters/day) by saltcedar for 3 days in 1966 Computed as the Deviation Computed with Deviation Date Measured sum of 24 hourly from eq. 26 from mean from values (eq. 26) measured daily values measured April® ccc _. 0 10.4 11.5 +1:4 14.4 +4.0 AO EC Cc ol lc elin ee ia in eae we ne nio 15.4 15.6 +0.2 18.2 +2.8 May Si- 4.8.2. .. ol oon oe dor ie 13.5 12.6 ~-0.9 15.8 +2.3 F48 expensive. Moreover, like the plastic tanks, they would provide data that, strictly speaking, would be applica- ble only to areas quite similar in ecology. Such similar- ity is difficult to ascertain unequivocally. The present studies have shown that even under the very complicated circumstances of riparian and flood- plain vegetation, comparatively simple instrumenta- tion can provide fairly adequate data for the purpose of estimating evapotranspiration using the combination- type model. Improved sensors and recorders, especially those of humidity and radiation, would undoubtedly provide more reliable results. However, although not tested at the Buckeye Proj- ect, the eddy-correlation model and its instrumenta- tion seems to be the most promising of all models and methods in use so far because of the fact that the method is independent of the height or exposure of the vegetation and independent of lapse rates or tempera- ture and other profiles. By contrast, most aerodynamic or energy-budget models are valid only under adiabatic conditions. In future evapotranspiration research, the eddy- correlation methods and its further instrumentation development should have the predominant attention of hydrologists, botanists, and other researchers in the Water Resources Division and, actually, in all other Federal, State, and private organizations studying plant-soil-water relationships. REFERENCES CITED Allen, L. H., Jr., 1968, Turbulence and wind speed spectra within a Japanese larch plantation: Jour. Appl. Meteorology, v. 7, no. 1, p. 73-78. Anderson, E. R., Anderson, L. J., and Marciano, J. J., 1950, Lake Mead water loss investigation-A review of evaporation theory and development of instrumentation: U.S. Navy Electronics Lab, San Diego, Calif., Rept. 159, 70 p. Arkley, R. J., 1963, Relationship between plant growth and transpiration: Hilgardia, v. 34, p. 559-584. Baver, L. D., Gardner, W. H., and Gardner, W. R., 1972, Soil physics: New York, JohnWiley & Sons, 498 p. Bavina, L. G., 1967, Refinement of parameters for calculating evap- oration from bogs on the basis of observations at bog stations: [translated from Russian] in Soviet Hydrology, no. 4, 1967, p. 348-370. Bierhuizen, J. F., and Slatyer, R. O., 1965, Effect of atmospheric concentration of water vapor and CO, in determining transpiration-photosynthesis relationships of cotton leaves: Agr. Meteorology, v. 2, p. 259-270. Blackwell, J. J., and Tyldesley, J. B., 1965, Measurement of natural evaporation: comparison of gravimetric and aerodynamic methods, in Eckardt, F. E., ed., Methodology of plant ecophysiol- ogy: United Nations Educational and Scientific Cultural Or- ganization, Arid Zone Research, Paris, France, v. 25, p. 141-148. Blaney, H. F., 1957, Relationship of pan evaporation to evapotrans- piration by phreatophytes and hydrophytes: Phreatophyte Sym- posium, Am. Geophys. Union, Sacramento, Calif., February 1957, 7 p. STUDIES OF EVAPOTRANSPIRATION Blaney, H. F., and Criddle, W. D., 1962, Determining consumptive use and irrigation water requirements: U.S. Dept. Agriculture Tech. Bull. 1275, 59 p. Bonner, James, and Galston, A. W., 1952, Principles of plant physi- ology: San Francisco, W. H. Freeman & Co., 499 p. Bowen, I. S., 1926, The ratio of heat-losses by conduction and by evaporation from any water surface: Phys. Rev., v. 27, p. 779- 787. Brunt, David, 1952, Physical and dynamical meteorology: Cam- bridge, England, University Press, 428 p. Budagovskii, A. I., and Savina, S. S., 1956, Evaporation from the surface of vegetative cover: Meteorologiya i gidrologiya 1956, no. 8, p. 20-23. Translated by E. Feigle for the Office of Technical Services, Dept. of Commerce, Washington, D. C. Budyko, M. IL., 1948, Evaporation under natural conditions: Lenin- grad, Hydrometeorological Publications [translated from Rus- sian, 1963, Israel Program for Sci. Translations, Natl. Sci. Found. and the U.S. Dept of the Interior], 130 p. 1956, The heat balance of the earth's surface: Leningrad, Hy- drometeorological Publications [translated by N. A. Stepanova, 1958, U.S. Dept. of Commerce], 259 p. Byrne, G. F., Rose, C. W., and Slatyer, R. O., 1970, An aspirated diffusion porometer: Agr. Meteorology, v. 7, p. 39-44. Campbell, C. J., and Strong, J. E., 1964, Salt gland anatomy in Tamarix pentandra (Tamaricaceae): The Southwestern Natu- ralist, v. 9, no. 4, p. 232-238. Casey, R. S., and Perry, J. W., 1951, Punched cards-their applica- tions to science and industry: New York, Reinhold Publishing Corp., 506 p. Chikishev, A. G., ed., 1965, Plant indicators of soils, rocks and sub- surface waters: Authorized translation from the Russian, Con- sultants Bureau Enterprises, New York, 210 p. Christiansen, J. E., and Hargreaves, G. H., 1969, Irrigation require- ments from evaporation: Trans. Internat. Comm. Irrig. Drain- age, v. III, p. 23.569-23.596. Cruff, R. W., and Thompson, T. H., 1967, A comparison of methods of estimating potential evapotranspiration for climatological data in arid and subhumid environments: U.S. Geol. Survey Water- Supply Paper 1839-M, 28 p. Culler, R. C., 1970, Water conservation by removal of phreatophytes: EOS, Am. Geophys. Union, Trans., v. 51, no. 10, p. 684-689. Culler, R. C., and others, 1970, Objectives, methods, and environment-Gila River Phreatophyte Project, Graham County, Arizona: U.S. Geol. Survey Prof. Paper 655-A, 25 p. Dalton, John, 1802, Experimental essays on the constitution of mixed gases; on the force of steam or vapor from water and other liquids in different temperatures, both in a Torricellian vacuum and in air; on evaporation; and on the expansion of gases by heat: Manchester Lit. Philos. Soc. Mem. Proc., v. 5, p. 536-602. Deacon, E. L., 1953, Vertical profiles of mean wind in the surface layers of the atmosphere: Geophys. Mem. no. 91, London, Her Majesty's Stationery Office, 92 p. Douglas, Ernie, 1954, Phreatophytes-water hogs of the West: Land Improvement, v. 1, no. 3, p. 8-12. 1965, Surveyors gather material for war against saltcedar: Arizona Farmer-Ranchman, May 29, 1965. Dyer, A. J., 1961, Measurements of evaporation and heat transfer in the lower atmosphere by an automatic eddy-correlation tech- nique: Royal Meteorol. Soc. Quart. Jour., v. 87, p. 401-412. Dylla, A. S., and Muckel, D. C., 1964, Evapotranspiration studies on native meadow grasses: Soil and Water Conserv. Research Div. Agr. Research Service, U.S. Dept. of Agriculture and Nevada State Dept. of Conserv. and Nat. Resources, no. R 9, 29 p. Eagleman, J. R., 1967, Pan evaporation, potential and actual evapo- transpiration: Jour. Appl. Meteorology, v. 6, p. 482-488. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA Fitting, Hans, Schumacher, Walter, Harder, Richard, and Firbas, Franz, 1951, Lehrbuch der Botanik | Textbook of Botany |: Ger- many, Piscator-Verlag Stuttgart, 626 p. Gates, D. M., 1963, The energy environment in which we live: Am. Scientist, v. 51, p. 327-348. Gatewood, J. S., Robinson, T. W., Colby, B. R., Hem, J. D., and Hal- penny, 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. Goltz, S. M., Tanner, C. B., Thurtell, G. W., and Jones, F. E., 1970, Evaporation measurements by an eddy correlation method: Water Resources Research, v. 6, no. 2, p. 440-446. Hales, Stephen, 1727, Vegetable staticks: London,[Reprinted by] The Scientific Book Guild, 214 p. Hamon, W. R., 1961, Estimating potential evapotranspiration: Am. Soc. Civil Engineers, Hydraulics Div., Jour., v. 87, no. HY3, p. 107-120. Harbeck, G. E., Jr., 1962, A practical field technique for measuring reservoir evaporation utilizing mass-transfer theory: U.S. Geol. Survey Prof. Paper 272-E, p. 101-105. Harbeck, G. E., Jr., Kohler, M. A., Koberg, G. E., and others, 1958, Water-loss investigations: Lake Mead studies: U.S. Geol. Survey Prof. Paper 298, 100 p. Hem, J. D., 1967, Composition of saline residues on leaves and stems of saltcedar (Tamarix pentandra Pallas): U.S. Geol. Survey Prof. Paper 491-C, 9. p. Hicks, B. B., and Goodman, H. S., 1971, The eddy-correlation tech- nique of evaporation measurement using a sensitized quartz- crystal hygrometer: Jour. Appl. Meteorology, v. 10, p. 221-223. Horton, J. S., 1973, Evapotranspiration and water research as re- lated to riparian and phreatophyte management-An abstract bibliography: U.S. Dept. of Agriculture, Forest Service Misc. Pub. 1234, 192 p. Horton, J. S., and Campbell, C. J., 1974, Management of phreatophytes and riparian vegetation for maximum multiple use values: U. S. Dept. of Agriculture, Forest Service Research Paper RM-117, 23 p. Hughes, W. C., 1972, Simulation of saltcedar evapotranspiration: Am. Soc. Civil Engineers, Jour. Irrig. and Drainage Div., v. 98, no. 4, p. 533-542. Humphrys, C. R., 1962, Evapotranspiration bibliography: Michigan State Univ. Agr. Expt. Sta. Water Bull. No. 7, 33 p. Idso, S. B., 1968, Comments on a paper by Richard Lee, "The hy- drologic importance of transpiration control by stomata": Water Resources Research, v. 4, p. 665-666. Jensen, M. E., ed., 1973, Consumptive use of water and irrigation water requirements: New York, Am. Soc. Civil Engineers, 215 p. Jensen, M. E., and Haise, H. R., 1963, Estimating evapotranspira- tion from solar radiation: Am. Soc. Civil Engineers, Jour. Irrig. Drainage Div., v. 89, p. 15-41. Kalweit, H., 1953, Der Wasserhaushalt [Water economy]: Berlin, VEB Verlag, 240 p. Kearney, T. H., and Peebles, R. H., 1960, Arizona flora: Berkeley, California Univ. Press, 1085 p. Kersten, M. S., 1949, Laboratory research for the determination of the thermal properties of soils: St. Paul, Minnesota Univ. En- gineering Expt. Sta. Tech. Rept. no. 23, 227 p. Kisser, J. G., 1956, Wasserumsatz und Salzstoffwechsel [Water turnover and salt metabolism], in Rubhland, W., ed., Handbuch der Pflanzenphysiologie [Handbook of plant physiology ]: v. III, Berlin, Springer-Verlag, p. 669-684. Kohler, M. A., Nordenson,T. J., and Fox, W. E., 1955, Evaporation from pans and lakes: U.S. Weather Bureau Research Paper 38, 21 p. F49 Konstantinov, A. R., 1963a, Regime of evaporation from the surface of ponds and reservoirs: [translated from Russian by L. G. Rob- bins] in Soviet Hydrology, no. 1, 1963, p. 76-85. 1963b, Methods for the calculation of evaporation from the data of gradient measurements: [translated from Russian by E. F. Tourkofi'] in Soviet Hydrology, no. 6, 1963, p. 568-591. Kuznetsov, V. I., and Fedorova, T. G., 1968, Comparative estimate of methods of computing evaporation from bodies of water: (translated from Russian] in Soviet Hydrology, no. 3, 1968, p. 238-253. Lane, R. K., 1964, Estimating evaporation from insolation: Am. Soc. Civil Engineers, Hydraulics Div. Jour., v. 90, no. HY5, p. 33-41. Lee, Richard, 1967, The hydrologic importance of transpiration con- trol by stomata: Water Resources Research, v. 3, no. 3, p. 737- 752. 1968a, Reply to S. B. Idso comments: Water Resources Re- search, v. 4, p. 667-669. 1968b, Reply to C. H. M. van Bavel comments: Water Re- sources Research, v. 4, p. 1389-1390. Lemon, E. R., 1960, Photosynthesis under field conditions. II. An aerodynamic method for determining the turbulent carbon dioxide exchange between the atmosphere and a corn field: Agronomy Jour., v. 52, p. 697-703. 1963, Theoretical consideration of aerodynamic exchange in a turbulent boundary layer, in Lemon, E. R., ed., 1963, The energy budget at the earth's surface: U.S. Dept. of Agriculture, Agric. Research Service Products Rept. No. 71, pt. I, p. 3-27. Lowery, R. L., and Johnson, A. F., 1942, Consumptive use of water for agriculture: Am. Soc. Civil Engineers Trans., v. 107, p. 1243-1266. McCulloch, J. S. G., and Penman, H. L., 1956, Heat flow in the soil: Sixieme Congres de la Science du Sol, Paris, v. I, p. 275-280. 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. McQueen, I. S., and Miller, R. F., 1972, Soil-moisture and energy relationships associated with riparian vegetation near San Car- los, Arizona: U.S. Geol. Survey Prof. Paper 655-E, 51 p. Makkink, G. F., 1957, Testing the Penman formula by means of lysimeters: Jour. Inst. Water Engineering, v. 11, no. 3, p. 277- 288. Malhotra, G. P., 1969, Hydrologic cycle of North America formulated by the analysis of water vapor transport data: Ph.D. Dissert., Connecticut Univ., 159 p. Meinzer, O. E.,1923, Outline of ground-water hydrology, with defini- tions: U.S. Geol. Survey Water-Supply Paper 494, 71 p. 1927, Plants as in indicators of ground water: U.S. Geol. Sur- vey Water-Supply Paper 577, 95 p. Monteith, J. L., 1963, Gas exchange in plant communities, in Evans, L. T., ed., Environmental control of plant growth: New York, Academic Press, chap. 7, p. 95-112. 1965, Evaporation and environment, in The State and move- ment of water in living organisms: Cambridge, Symposia of the Society for Experimental Biology, chap. 19, p. 205-234. 1973, Principles of environmental physics: New York, Ameri- can Elsevier Publishing Company, 241 p. Monteith, J. L., and Szeicz, G., 1960, The carbon dioxide flux over a field of sugar beets: Royal Meteorol. Soc. Quart. Jour., v. 86, p. 205-214. Morton, F. I., 1975, Estimating evaporation and transpiration from climatological observations: Jour. Appl. Meteorology, v. 14, p. 488-497. Munk, W. H., and MacDonald, G. J. F., 1960, The rotation of the earth: Cambridge, Univ. Press, 323 p. F50 Ognev, B. I., and Kozlov, Y. N., 1962, Computation of evaporation from a water surface: [translated by D. B. Krimgold] in Soviet Hydrology, no. 1, 1962, p. 2-4. Olivier, Henry, 1961, Irrigation and climate: London, Edward Ar- nold Ltd., 250 p. Pallas, J. E., Jr., 1965, Transpiration and stomatal opening with changes in carbon dioxide content of the air: Science, v. 147, p. 171-173. Panofsky, H. A., and Brier, G. W., 1965, Some applications of statis- tics to meteorology: University Park, Pennsylvania State Univ., 224 p. Papadakis, J., 1966, Climates of the world and their agricultural potential, in Jensen, M. E., ed., 1973, Consumptive use of water and irrigation water requirements: New York, Am. Soc. Civil Engineers, p. 84, 85, and 97. Pasquill, F., 1949, Eddy diffusion of water vapour and heat near the ground: Royal Soc. London Proc., Series A, v. 198, p. 116-140. 1950, Some further considerations of the measurement and indirect evaluation of natural evaporation: Royal Meteorol. Soc. Quart. Jour., v. 76, p. 287-301. Penman, H. L., 1948, Natural evaporation from open water, bare soil and grass: Royal Soc. London, A, Proc., v. 193, p. 120-145. 1956, Evaporation: An introductory survey: Netherland Jour. Agr. Sci., v. 4, p. 9-29. Penman, H. L., and Long, I. F., 1960, Weather in wheat: An essay in micro-meteorology: Royal Meteorol. Soc. Quart. Jour., v. 86, p. 16-50. Perepelkina, A. V., 1959, On the determination of turbulent flow of the heat: Bull. USSR Acad. Sci. Inst. Geophys. Series, p. 727- 733. [translated by J. S. Sweet for the Office of Technical Ser- vices, Dept. of Commerce, Washington, D.C.] Phreatophyte Subcommittee, 1965, List of references on phreatophytes and related subjects: Pacific Southwest Inter- agency Comm., mimeo rept. p. I-1-VIII-2, available from the Rocky Mtn. Forest and Range Expt. Sta., Tempe, Arizona State Univ. Pierson, F. W., and Jackman, A. P., 1975, An investigation of the predictive ability of several evaporation equations: Jour. Appl. Meteorology, v. 14, p. 477-487. Portman, D. J., 1954, Determination of soil heat transfer at O'Neill, Nebraska: [Johns Hopkins Univ., Laboratory of Climatology ] Pub. in Climatology, v. 7, p. 320-325. Rasmusson, E. M., 1967, Atmospheric water vapor transport and the water balance of North America: Part I. Characteristics of the water vapor flux field: Monthly Weather Rev., v. 95, no. 7, p. 403-426. 1968, Atmospheric water vapor transport and the water bal- ance of North America: Part II. Large-scale water balance inves- tigations: Monthly Weather Rev., v. 96, no. 10, p. 720-734. Rider, N. E., 1957, Water losses from various land surfaces: Royal Meteorol. Soc. Quart. Jour., v. 83, p. 181-193. Ripple, C. D., Rubin, Jacob, and van Hylckama, T. E. A., 1972, Es- timating steady-state evaporation rates from bare soils under conditions of high water table: U.S. Geol. Survey Water-Supply Paper 2019-A, 39 p. Robinson, T. W., 1952, Water thieves: Chemurgic Digest, October 1952, p. 12-15. 1958, Phreatophytes: U.S. Geol. Survey Water-Supply Paper 1423, 84 p. 1964, Phreatophyte research in the Western States March 1969 to July 1964: U.S. Geol. Survey Circ. 495, 31 p. 1965, Introduction, spread, and areal extent of saltcedar (Tamarix) in the Western States: U.S. Geol. Survey Prof. Paper 491- A, 12 p. STUDIES OF EVAPOTRANSPIRATION 1970, Evapotranspiration by woody phreatophytes in the Humboldt River valley near Winnemucca, Nevada: U.S. Geol. Survey Prof. Paper 491-D, 41 p. Robinson, T. W., and Johnson, A. I., 1961, Selected bibliography on evaporation and transpiration: U.S. Geol. Survey Water-Supply Paper 1539-R, 25 p. Rohwer, Carl, 1931, Evaporation from free water surfaces: U.S. Dept. Agriculture Tech. Bull. 271, 96 p. Sebenik, P. G., and Thames, J. L., 1968, Water consumption by phreatophytes: Progressive Agriculture, March-April 1968, p. 10-11. Sellers, W. D., 1965, Physical climatology: Univ. of Chicago Press, 272 p. Sheppard, P. A., 1946, The aerodynamic drag of the earth's surface and the value of von Karman's constant in the lower atmo- sphere: Royal Soc. London, Proc. Series A, v. 188, p. 208-22. Smirnov. K. L., 1963, Evaporation from the soil in the Kustanay- skaya Province: [translated by D. B. Krimgold] in Soviet Hy- drology, no. 2, 1963, p. 156-174. Stiles, W., 1970, A diffusive resistance porometer for field use: Jour. Appl. Ecology, v. 7, p. 617-622. Suomi, V. E., and Tanner, C. B., 1958, Evapotranspiration estimates from heat-budget measurements over a field crop: Am. Geophys. Union, Trans., v. 39, p. 298-304. Sutton, O. G., 1953, Micrometeorology: New York, McGraw Hill Book Company, 333 p. Swinbank, W. C., 1951, The measurement of vertical transfer of heat and water vapor by eddies in the lower atmosphere: Jour. Meteorology, v. 8, no. 3, p. 135-145. Swinbank, W. C., and Dyer, A. J., 1967, An experimental study in micro-meteorology: Royal Meteorol. Soc. Quart. Jour., v. 93, p. 494-500. Tanner, C. B., 1963, Energy relations in plant communities, in Evans, L. T., ed., 1963, Environmental control of plant growth: New York, Academy Press, p. 141-148. Taumer, F., 1955, Eine Methode zur Bestimmung der Gebietsver- dunstung auf meteorologischer Grundlage [A method to deter- mine the areal evaporation based on meteorologic principles]: Leipzig, Inst. f. Agrarmeteorologie, v. I, no. 1, p. 1-71. Thornthwaite, C. W., 1948, An approach toward a rational classifica- tion of climate: Geog. Rev., v. 38, p. 55-94. Thornthwaite, C. W., and Holzman, Benjamin, 1942, Measurement of evaporation from land and water surfaces: U.S. Dept. of Ag- riculture Tech. Bull. no. 817, 75 p. Thornthwaite, C. W., and Mather, J. R., 1955: The water balance: Publications in Climatology, Centerton, New Jersey, v. 8, p. 1-104. Tomanek, G. W., and Ziegler, R. L., 1960, Ecological studies of salt cedar: Kansas State Coll., Div. Biological Sci., Hays, Kansas, 128 p. Turc, L., 1961, Evaluation des besoins en eau d'irrigation, evapo- transpiration potentielle [Estimation of irrigation water needs, potential evapotranspiration ]: Annales Agronomiques, v. 12 (1), p. 13-49. U.S. Dept. of Commerce, Weather Bureau, 1955, Climatic Summary of the United States-Supplement for 1931 through 1952: U.S. Govt. Printing Office, Washington, D. C. U.S. Geological Survey, 1954, Water-loss investigations: Lake Hefner studies, technical report: U.S. Survey Prof. Paper 269, 158 p. U.S. Information Service, 1965, U.S. combats water-stealing plants: U.S. Govt. Press release, Washington, D. C. van Bavel, C. H. M., 1966, Potential evaporation-the combination concept and its experimental verification: Water Resources Re- search v. 2, no. 3, p. 455-467. WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA 1967, Surface energy balance of bare soil as influenced by wetting and drying: U.S. Dept. of Agriculture, Water Conserv. Lab., Research rept. No. 397, 72 p. 1968, Further to the hydrologic importance of transpiration control by stomata: Water Resources Research, v. 4, p. 1387- 1388. van Bavel, C. H. M, Brust, K. J., and Stirk, G. B., 1968, Hydraulic properties of a clay loam soil and the field measurement of water uptake by roots: II. The water balance of the root zone: Soil Sci. Soc. Am. Proc., v. 32, p. 317-321. van Helmont, Joan Baptista, 1655, Ortus medicinae [The rise of medicine]: Posthumously edited by his son, Franciscus Mer- curius van Helmont, published by J. B. Nevenet, Leiden, the Netherlands, p. 68, paragraph 30. van Hylckama, T. E. A., 1963, Growth, development and water use by saltcedar (Tamarix pentandra) under different conditions of weather and access to water: Internat. Assoc. Sci. Hydrology Pub. No. 62, p. 75-86. 1966, Effect of soil salinity on the loss of water from vegetated and fallow soils: Wageningen Symposium Proc., v. II, Internat. F51 Assoc. of Sci. Hydrology Pub. No. 83, p. 635-644. 1969, Photosynthesis and water use by saltcedar: Internat. Assoc. Sci. Hydrology Bull., v. 14, p. 71-83. 1970a, Winds over saltcedar: Agr. Meteorology, v. 7, p. 217- 233. 1970b, Water balance and earth unbalance: Symposium on world water balance, Internat. Assoc. Sci. Hydrology, pub. No. 93, p. 434-444. 1974, Water use by saltcedar as measured by the water budget method: U.S. Geol. Survey Prof. Paper 491-E, 30 p. van Wijk, W. R., and de Vries, D. A., 1966, Periodic temperature variations in a homogeneous soil, in van Wijk, W. R., ed., 1966, Physics of plant environment: Amsterdam, North-Holland Pub- lishing Company, 382 p. Walter, Heinrich, 1973, Vegetation of the earth: London, English University Press, 237 p. Webb, E. K., 1965, Aerial microclimate, in Waggoner, P. E., and others, eds., 1965, Agricultural meteorology : Chap. 2, Meteorol. Mon., v. 6, no. 28, July 1965, pub. by Am. Meteorol. Soc., p. 27-58. APPENDICES Wind direction 5m 4m 2m 7m 5-6 May 1966 11m STUDIES OF EVAPOTRANSPIRATION 1-hr. period beginning A1-1.-Hourly averages of windspeeds and wind direction at the Buckeye Project during selected periods of 3 days each [Windspeeds in meters per second (m s~') at indicated levels (meters) above ground. Clocks set to true local solar time: 1200 (noon) when sun exactly at 180°] F54 Em- --- ® |calcalcol ® [calcalcal W RAAZZZZZZZZZZZZZZZEEREZZZ pi t- o aio w oin o 90 t on I~ NL xf D NN © D- b= 10 10 r- Cd 2 Coan inc wn on torn ini cosccooccoscccocccccccccccoce oescsosccsessceceso -b t o tD in N - 4D O H H mH & mH QI r 05 O 99 00 iD N - n§ESnqESNARANGESRIEEILSLIEER )q | $ olin | 3 | 3 - NNoo oo ad >. 3 3 § a «w ~ Tf 00 & N O NID NNY OL t O ~I HiQ I0 0) N o O O) iO I~ o9 tD & © 9p O 4 ap 9D (Y I ao 1D 1D 1D an ad - 90 4D i- 90 iq O) xh IQ iQ & Cy op tp tp 11 op & e4 op O) on o Noo o on AH HH /o HH HOH 69 73 19 54 50 85 59 59 53 51 95 erect annees cas ae an bg ae apa f ie ae b- rece al AIIRA inc one an emic ane ses nae sees pU Ieee ce cc c. uc ae ows I lector ily ducui ube cm arctic rect conal 6-7 July 1966 wn on n > z ~- 09 O ( O < sf O sf OI tD op b- N D Had 0d © H 10 tD A ~ 0 t In ® o N in 0 in on «o w o & t Sar 6 t or mao o OJ H O HoH - 9p I~ 90 90 D I~ 0D xf 00 Ch in or I~ ID in aa < G) 03 NY 4 - ode o « 00 0D 00 © MI I~ - 0D SH I~ b- O 0) Han G op t O ® IL 0 & 6 S in in © op - op «i IL L- 5 a; o; a o o a; in n 0 0 H O Hmm A Noo Tor 09 09 6 N N N HHH O HH Dna f evet i co uce sal. od ad OE tre iii ce. also I arene erie an ce sale ci ae news ae ad 9D O Arr iran 1. 17 23 23 80 19 13 46 95 58 T ef Lei oun u « n na ascend To Sere ent oue aos. - oal 1G cree - Serr 142 NPN AA cess cans and 19 EAL eee: coule coi 10 Eerie cind IQ cy: se c- 83 58 60 IP s e ec ail c 19 s- SLCC: L. cee s aed A rer 1.00 7-8 July ,1966 mG@@GNOM®®NQ@GiGzopr | 1 1 l lor tor rod | 1. ) 1 11 1 i } i g a a a i g g roa 9D © 9p 90 I~ in © - I~ 1D © N EE... _E. Eke _ - am. BAR elelefele z nzZzzinin>in;> in;» ininininzinz22iininin 72 2 iz iz iz tn tn tn tn tn tn tn tn tn un un tn an on m Z. 2. Z. NNNflNVSSVNWWWWWWW Pot popa a Oo or xf op apanan && 1 1 i 1 | C) 4 vA hosel o alone so, 109.010. 0 mea cue Corde me Shans w 2 a U. n ani reece AM EF a 464 © ound oid ma ene ode eaten Hoty Pria i 1s C a 0 ioe reat mae ot i + es pec Nis cie nel fon aes Tem tete 2 a nS W Pio d o ol soha d oon ees SD ale ets ee So A0 eela ro a CE A4 $2.0 0 Mnet rae s ie ng W W W i W . W W 00 i he RCC don ono ccd o nd a t 20 i al o Pico esi ab d+ rani CNC . a 04. 230040 Aesop Se o drei onthe ana ona nd sra ien per nds e mo seca cols .+ a Ease cal H0. ol tho Alera enol coal O e cia orp aia 1.2 e Aarni n alee in n n n reamer arora ad e 10009 (9400 S29 ceo ) Mct ae 1 f Denford Monel rede d Nue ad iy erie ted e Reco cog a M Mng nny Cou 20000 nicco e nowe Sas a (L 4.100 Vlora cater mancini) (4 2g ri han 4 1 Sean Eet ot (ened tee o She we elt ma opal ner s 9 ar 1. Cor reac ma 04.29 t as 2A ieee ien l cn o a an id mon Len . Ave his bn s al a i net Noe ilt mee ele i W bad lid ho Wiad Sf hedond W (ef Moo (Hasi on s U 2, orice te enon cy 24 700 7 cc dein Near esc . 0 mala And +94 uc Lert ol: "g co sn Pis le oie 1 Mccain Comi re Bo van Si Pep eit fea fuks (A0 (eas do i Pag (ha Ao o alme ce pint an Bia Aan cs elle t d. mates aa cee a aora u: Aal EA Hag se ne Arris on Sd a hera oue ha ca lh se ne Beim s aler a 2d W W 10054 f W (tA i (erf con O elle cn allo oa te ctor food ne rae od Iban seo Onto aries P90 t coa conal te Pane aE alle 24.3 oa old along ARE d P finite iret Aure cer reader evel ade co, ane tas cr Sel. o ses i rio d woo Arod send h e ca del ees oe Smee 0 al wi n recipe sn +04 i Sct wok at W W } W Pad W $d Mind a ela o Me n defer ai an See ton SPAA OSL Facc cus [4D Scant ere re dou o m ian fe oleae a noe tat ce ccd o ao chote MCs rs No te $24.4 hed Tonee oa ar Poin a dasa s eles Lene mac wastes oh oh Lea os Dra o a . kone a ter nre te Arran a 9 Wee toa P se Huk int f prste dot W Pig hho Pcp 10, diode ete aie sine co h ende ap nos e ers opel Fou 73 ) Con aB Silos ches acon i gs 9) 00 Sols a Ranee ae am ert arre td 1 (9g CAn a $ W Post Cad : W pth i 193 Plc Biden ned ot o Acmea tie Ll are rele. . ol ie hie Ae do ha a a ma O30 2 0 0a eum ic eve ant nas coco dime s far nee ari rae Pola wd Fard 92 0 1g mc 1 Iole itd 4 W i (pe i Ped foo HL pio g B 5A pec e nt Sah onl el oth com o eu Loa ian en » ted C08 itl ne tio ae Ca ao S card ane ote eae AC 01 pi lag en's Mic Aida n 4 pot Ae det fendi i te (oal A rere a oo s ti wen so B ss at ny waite e o 1a red ma aa cena 957A 0 5 oo Lt 5200 c oa no ah ond p bel nve a on Led ann wily) tna cn MM 4 Bilge Macc W Lo y 4 W W a W : i Sie W W o hoo enfant sided oph me l one nece t 0 a Ma ei rr as E0 em OTe 39000010, 1.000 a od opm mar ho eeren an ry o tu £3 ca 200 0 cao alc -e c H _ 1 (es 100 aL Cde to ae ca ces 9C 20 0 ha c o Hc Ss oce Poe w ae on r teta Bida EEN » cara (ial cll ,u Aca a ach cag a a g lga e dg ba 00 0 ol Lely tan 1 Sl Coes. dr Wary -u i 14 i 09. Aoi too, a Cate oa N0 tf ade on Mach 2 al el nip eal dig o ag 94 Melee teas ram ce Tan ide t e armonia ras tt caree Poon nt ePas Pelo SE m ends a W W i ag page caf a t ease iad (eg aicas oad o cind aet cpt at d tt a Hove BiB ih ad id 9 1 Arm ote no rre pon node none ae a ones aie io Or 0 00 -o a o - Ooo si- Nco sf 10 < I~ 00 &) O moo @ SASBSBAREILIESARNNRA it Cms too ai o on a in to tt oo on og d Wind direction 5m 4m 2m 7m 4-5 August 1966-Continued 11m STUDIES OF EVAPOTRANSPIRATION beginning TaBLE A1-1.-Hourly averages of windspeeds and wind direction at the Buckeye Project during selected periods of 3 days each-Continued 1-hr. period F58 A - AMM p. ._ p >> ejefefefefe] mmm am amm m >Zmaz2Z2Z 22 2M maZ222>~ © T NT Oo O 00 s 29.7 9.1 29.8 10.0 30.0 10.1 30.1 11.0 30.7 121 31.3 8.0 32.0 9.4 31.7 10.4 31.6 11.0 33.0 14.4 12% 33.7 8.2 34.1 9.0 34.8 9.5 34.7 9.9 35.6 11.1 F64 STUDIES OF EVAPOTRANSPIRATION TABLE A2-1. Hourly averages of air temperature and vapor pressure at the Buckeye Project during selected periods of 3 days each. -Continued 1-hr period 11m 7m 5m 4m 2m beginning *C mb *C mb °C mb "C mb °C mb 5-6 June 1966-Continued 19 35.3 5.7 35.4 7.0 36.1 8.1 85.3 8.7 33.7 12.1 14 sol 36.1 4.0 36.1 5.6 36.4 6.9 36.4 7.1 37.4 10.8 1B Les 36.5 4.1 36.7 5.4 36.8 6.7 '-86.8 7.9 37.4 10.2 16 .:.....0.. 36.9 4.1 36.6 5.5 36.7 6.5 36.4 7.2 37.0 8.0 17}: 36.4 8.5 36.5 5.5 36.6 6.9 35.8 8.0 35.2 9.2 18 35.1 5.3 35.2 6.5 35.0 7.4 33.8 7.9 31.8 9.5 19 cen 32.5 7.5 31.9 8.5 31.6 9.0 30.6 9.6 26.6 11.4 20 _ 31.2 7.4 29.8 8.1 29.3 8.5 28.4 9.0 21.4 11.8 PI AL 28.7 6.9 27.2 7.6 26.0 8.1 25.1 8.5 18.6 11.1 22 unless cls 26.7 yA 25.1 7.7 24.1 8.1 283.4 8.3 16.6 10.9 BJ oie cleus 25.2 7.3 24.5 7.6 23.7 7.9 23.0 8.1 15.8 9.6 P1 "..on nlcs 25.1 7.0 24.0 71 21.8 7.3 20.7 7.4 13.9 9.5 I 2-2 23.5 6.5 22.7 6.8 22.3 6.9 20.4 P 13.0 8.7 e 22.8 5.2 21.1 5.8 20.3 6.1 18.9 6.4 13.5 Tsi de. 21.2 5.8 18.9 6.3 17.7 6.7 16.7 6.9 11.8 8.7 § ore 21.6 6.0 20.0 6.5 19.4 6.9 18.7 7.1 11.3 9.1 6-7 July 1966 D A 20.6 5.3 18.3 6.2 18.7 5.2 16.9 6.2 12.2 7.6 19.7 6.6 18.7 6.6 18.5 6.8 17.6 7.8 14.4 8.1 Ta 22.5 8.7 22.4 8.3 22.3 8.4 22.0 8.9 21.2 10.1 D 27.8 7A 27.6 1.9 27.8 7.9 27.9 9.3 26.8 10.0 9 32.0 5.9 6.4 32.3 6.5 31.0 8.3 31.5 9.6 10. -.....-..-- 35.5 7.3 35.5 8.0 35.8 8.2 35.5 9.0 35.5 10.2 1V 38.6 4.8 38.6 5.9 38.9 6.6 38.4 8.4 39.2 9.7 12 40.4 2.5 40.6 4.5 40.8 5.6 40.9 7.6 42.3 8.6 13 42.2 5.0 40.2 6.0 42.5 6.7 41.9 7.5 43.9 6.9 14s. 41.8 4.4 42.0 5.7 42.7 7.0 43.1 7.8 45.3 6.2 15 c:. 42.5 4.2 42.8 5.8 43.3 7.0 43.1 8.5 44.7 6.8 16; ....-....3s 42.6 6.0 42.8 7.2 43.2 8.0 42.7 9.0 42.2 8.0 T/ s. ring 42.2 6.4 42.5 7.4 42.4 7.5 42.3 8.7 41.1 8.6 TS" lessee 41.4 5.9 41.2 7.6 41.0 8.0 40.9 9.4 38.5 9.3 19 % 40.1 5.8 39.8 T1 39.4 8.0 38.9 8.7 36.0 8.3 0 cic recess 38.7 5.7 37.4 7.3 36.7 7.9 36.2 8.6 32.1 8.1 Ii 37.4 6.7 36.0 7.8 35.4 8.1 34.8 8.6 30.2 8.8 TA 37.1 6.6 35.8 7.5 35.2 7.9 34.7 8.6 30.8 9.1 23 30.6 8.0 29.6 9.0 28.5 9.4 271.9 10.1 25.0 10.4 AE 30.4 9.0 28.9 9.7 27.9 9.8 27.0 10.1 21.9 10.6 T 28.7 11.2 27.5 11.3 26.5 11.4 26.0 11.6 22.4 12.1 A Sell 28.0 13.4 26.9 18.0 26.0 12.8 25.71 12.4 20.8 12.0 G cl 27.1 13.5 26.6 12.7 25.4 12.0 23.6 11.6 19.5 12.1 A itis ence 26.9 12.8 25.3 12.5 23.4 12.3 22.9 12.1 18.8 12.8 7-8 July 1966 25.4 12.8 24.0 14.0 23.1 13.0 22.7 13.1 18.5 13.8 6:...z2. cass 26.0 13.3 24.5 14.1 24.2 14.9 23.5 15.3 21.0 16.4 ntt 29.2 16.1 28.6 16.8 28.3 17.1 7.0 31.4 17.0 31.4 $ 31.3 17.0 31.3 17.0 31.4 17.0 31.4 17.0 29.8 17.9 9 31.9 15.0 31.9 16.2 32.0 17.1 31.8 17.5 30.7 18.3 10 - 33.8 15.9 34.4 16.2 34.9 16.3 34.7 16.5 34.7 18.6 11: 35.0 10.5 35.5 11.7 35.8 18.5 36.5 14.4 38.2 15.6 12 cs 87.1 11.0 37.9 12.6 38.4 18.6 38.1 14.5 39.6 15.9 19 38.3 10.9 38.5 12.1 39.0 12.8 38.9 13.1 39.2 14.1 14}. 38.6 10.5 39.9 11.4 39.9 12.0 39.8 12.5 41.0 13.0 15. 38.1 11:2 38.2 12.5 38.3 12.9 38.7 13.7 40.9 14.0 16: 38.1 9.4 38.2 10.9 38.2 11.5 38.2 12.3 38.9 13.0 17 31.1 9.0 37.9 10.6 37.5 11.6 36.7 12.4 35.2 12.4 18 37.6 9.8 37.4 11.0 37.2 11.1 36.4 12.5 35.1 12.3 19». 36.0 10.5 36.0 11.5 35.8 11.9 35.0 12.8 32.8 13.0 Q 34.1 9.7 33.5 10.4 33.0 10.7 32.2 11.2 29.1 12.1 J so 34.9 10.0 34.0 10.8 33.5 112 33.0 11.6 29.0 11.4 Aa - oie 31.9 11.0 30.9 11.6 30.1 11:7 29.5 13:1 25.3 13.9 B esu 30.9 12:1 29.5 12.5 28.9 12.8 28.0 13.0 35.1 13.8 PE Clea 29.2 11.9 28.0 12.4 27.0 12.9 26.2 13.1 22.5 14.0 29.5 10.9 28.1 11.4 27.1 11.8 26.1 12.2 21.1 13.4 o 27.3 13.9 26.9 14.6 26.3 14.9 25.9 15.1 23.3 15.1 B 271.2 19.5 26.4 19.4 26.1 19.5 25.8 19.3 23.6 19.5 f 27.2 19.6 27.2 19.6 26.6 19.9 26.1 19.8 28.7 19.9 WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F65 TABLE A2-1. Hourly averages of air temperature and vapor pressure at the Buckeye Project during selected periods of 3 days each. -Continued 1-hr period 11m 7m 5m am Fw beginning °C mb °C mb *C mb °C mb °C mb 8-9 July 1966 D 26.9 20.4 26.4 20.2 25.9 20.1 25.5 19.9 28.1 20.0 6 css 26.4 20.5 25.7 20.4 25.2 20.3 24.7 20.0 22.6 19.4 T EJA... 26.9 20.6 26.9 20.7 27.1 21.0 26.6 21.6 26.0 283.4 $ 28.8 20.2 28.4 20.3 28.7 20.8 29.0 21.0 28.0 23.1 9 RL 30.2 19.0 30.5 19.6 30.7 20.4 30.7 21.0 30.1 22.6 10 31.7 18.1 31.8 18.9 32.2 19.2 32.1 19.5 32.5 21.4 11 32.7 18.5 32.8 18.8 33.4 19.0 33.9 19.5 35.4 19.5 12 34.0 18.2 34.4 18.3 35.5 18.5 35.2 18.7 37.0 20.9 1j scx: 35.3 14.8 35.6 15.6 36.0 15.9 35.5 16.5 38.7 18.7 14 NE Ri: ___ 36.7 14.5 36.9 14.7 87.3 14.8 37.1 15.3 39.0 16.8 15 RL 36.5 13.5 36.9 14.4 37.2 15.4 37.6 15.8 39.7 16.0 16 w.1.._.... 37.2 13.9 38.2 14.3 38.3 14.8 38.0 15.1 38.4 16.4 17 37. 4 13.4 38.0 14.1 38.1 14.9 37.6 15.2 37.3 16.1 18 371.2 14.2 37.6 14.6 37.8 14.7 37.2 14.8 35.8 16.4 19-s.i......> 36.5 14.0 31.4 14.9 36.1 15.1 35.2 15.7 33.1 16.6 0 s. 34.9 14.4 34.2 15.0 33.6 15.2 32.9 15.5 29.0 16.6 Y 2 36.0 13.8 34.5 14.1 33.5 14.6 32.7 14.9 27.5 16.1 33.5 14.7 33.1 15.0 32.6 15.6 31.6 15.9 28.3 16.6 eil... 32.4 14.5 32.6 15.7 32.1 16.1 31.3 17.0 29.0 16.6 24 31.4 16.2 31.3 16.9 30.8 17.1 30.3 17.3 28.2 17.6 31.6 17.3 30.9 17.8 30.7 18.0 30.1 18.1 21.2 18.4 a 31.6 17.2 31.2 18.0 31.1 18.2 30.6 18.6 28.3 18.4 S lr _....s 30.6 18.7 30.3 19.0 30.2 19.4 29.9 19.8 28.7 19.0 29.4 19.2 28.6 19.1 28.0 19.0 27.8 18.9 26.1 19.3 3-4 August 1966 1g) ie 31.2 20.0 30.7 20.5 30.3 21.1 30.0 21.7 26.5 21.4 31.8 20.5 31.7 21.1 31.6 21.5 31.3 21.8 28.3 22.1 T 33.4 22.1 33.4 38.7 23.3 33.1 23.5 31.9 24.4 Sal...... 34.8 21.3 35.2 21.8 35.4 22.3 35.3 22.5 34.5 25.0 35.4 22.1 35.7 22.7 36.0 23.0 36.7 23.3 35.8 25.2 10 37.1 21.0 37.8 21.8 38.5 22.1 38.6 22.9 38.9 24.0 11° 38.8 20.1 39.4 21.0 39.8 21.5 40.8 22.0 41.6 24.0 12: 39.4 20.4 39.7 20.9 40.0 21.4 40.6 21.7 41.4 23.3 19 it.... ..3 39.9 19.1 40.6 20.1 41.3 21.6 41.5 2L1 43.0 28.5 14}}..00...... 40.6 17.7 40.8 19.0 41.6 19.5 42.2 20.4 43.4 22.2 15: 39.7 20.4 40.1 20.5 40.3 20.8 40.6 21.0 41.4 22.0 16m --..... 38.3 23.5 38.5 28.7 38.9 23.9 39.0 24.2 39.9 25.0 37.3 22.5 37.9 22.6 37.8 23.0 38.0 23.2 38.1 24.2 SH. J...... 37.4 22.0 37.9 22.2 37.9 22.4 37.6 22.5 37.5 23.0 19 - 36.8 21.1 37.1 21.3 37.1 21.6 36.9 21.8 36.1 22.4 35.9 21.0 36.3 21.2 36.2 21.5 35.6 22.0 34.5 22.6 Blik ...l. .s 35.4 21.8 35.2 22.0 35.1 22.4 34.8 22.6 33.7 22.8 are ille ges-, 34.5 21.3 34.1 21.4 33.9 21.6 33.7 21.8 32.7 23.0 Tofel. 32.8 20.1 32.8 20.3 32.5 20.4 32.4 20.5 31.4 22.1 32.3 19.5 32.2 19.7 31.9 19.8 31.8 19.9 30.9 21.3 Hw::....:«> 31.5 20.1 31.3 20.2 31.1 20.3 30.8 20.4 29.7 22.1 28.8 20.7 28.7 21.1 28.3 21.7 28.1 21.8 28.0 22.4 28.1 20.9 27.7 21.1 27.6 21.3 27.3 21.6 26.8 22.8 Mies illi 271.8 21.4 27.0 21.7 26.6 21:7 26.5 21.8 25.1 22.9 4-5 August 1966 26.4 21.5 26.2 22.0 26.1 22.3 26.0 22.7 25.0 22.8 6g.:i:..._... 26.4 21.5 26.4 21.8 26.1 22.0 26.0 22.0 24.5 23.0 T 21.9 22.6 27.8 22.3 28.1 22.7 27.8 23.0 27.1 24.8 Si.}....~.~. 31.2 19.8 32.0 19.9 31.7 21.2 31.6 22.3 32.1 22.5 Dire ica 32.9 20.6 33.9 21:7 33.8 22.0 33.7 22.5 35.1 283.4 10.¢i...1...._ 34.4 20.5 35.8 20.8 35.9 21.2 36.3 21.5 37.1 24.0 37.2 20.2 37.7 21.0 37.8 21.4 38.1 21.5 39.9 23.8 38.5 20.0 38.9 20.3 39.5 20.6 40.1 20.8 41.9 22.7 39.6 19.0 40.2 19.8 41.3 20.3 41.9 20.6 43.9 22.1 14 etc 000 40.8 18.6 41.5 19.0 42.0 19.6 42.7 19.9 44.3 22.9 15 41.4 17.0 42.4 17.4 42.7 17.8 43.5 18.0 45.6 20.9 1G r.. ._._._.- 41.7 16.5 43.0 17.2 43.4 17.8 43.5 18.0 45.3 19.1 41.8 14.9 42.8 15.8 42.7 16.4 43.2 16.8 43.2 20.8 18} 41.5 15.4 42.0 15.6 42.3 16.0 42.1 16.5 40.4 20.3 19 :} 40.3 15.5 39.8 16.6 39.8 17.9 39.0 17.4 35.8 21.5 208.1 40.2 14.8 38.8 15.7 38.1 16.2 37.4 16.6 32.6 20.0 IEC 39.5 15.9 38.6 16.2 38.1 16.7 37.6 17.1 34.8 18.6 F66 STUDIES OF EVAPOTRANSPIRATION TABLE A2-1. Hourly averages of air temperature and vapor pressure at the Buckeye Project during selected periods of 3 days each. -Continued 1-hr period 11m Tm 5m 4m 2m beginning °C mb °C mb °C mb "C mb 'C mb 4-5 August 1966-Continued Pa 36.4 16.9 35.3 17.2 34.9 17.5 35.0 17.8 32.1 20.6 23 A.l.elc.. 34.9 17.8 34.8 18.3 34.6 18.5 34.5 18.8 32.8 20.4 24 Z2... ces 34.8 17.4 34.2 18.1 34.5 18.4 33.9 18.6 32.1 20.0 34.5 179 34.2 18.0 34.0 18.3 33.8 18.6 32.6 20.3 2 A einacece 31.7 19.2 31.3 19.5 31.2 19.5 30.9 19.5 29.0 21.5 S ce. 29.0 20.4 28.8 20.5 28.7 20.6 28.4 20.7 27.5 21.9 A 27.8 20.4 27.5 20.5 27.4 20.7 21A 21.0 25.4 22.4 5-6 August 1966 O rels 26.8 21.3 26.5 21.4 26.3 21.4 25.9 21.5 24.1 21.9 27.0 21.5 26.5 21.7 26.4 21.8 26.3 21.9 24.2 22.6 I 29.2 21.8 28.2 22.0 28.8 22.7 28.8 23.2 28.1 24.3 8 31.7 20.8 32.0 22.4 31.1 23.0 31.1 23.5 31.8 24.5 9 33.9 21.5 34.0 22.2 34.6 22.4 34.8 25.2 35.8 25.4 10 :c... :c. 36.9 19.6 37.1 20.7 37.4 21.5 37.5 21.8 38.8 24.1 11 :s--:szens 37.3 20.9 38.0 21.4 39.1 21.8 39.4 22.0 41.3 23.8 T2 38.7 18.4 39.4 18.9 40.2 20.0 40.6 20.1 43.3 22.0 13. 39.9 18.3 41.1 19.6 42.1 19.8 41.8 20.1 44.4 22.5 14 .e. 40.2 16.7 41.2 17A 41.9 17.8 42.6 18.6 44.7 21.4 40.8 15.4 41.7 16.0 42.1 16.8 42.6 17.4 43.7 21.5 TG *.._sc....l 41.8 15.8 42.0 16.4 42.5 16.7 42.8 17.3 43.5 19.9 t? 41.3 15.9 41.9 16.0 42.4 17.2 42.4 17.6 41.7 20.4 TS :.._s2....J 40.3 19.0 40.8 19.5 40.8 18.2 41.1 19.6 39.9 22.5 19 38.6 21.6 38.4 21.9 38.5 22.0 38.7 22.0 38.0 24.1 20 ic. 36.1 21.8 36.5 22.5 36.3 24.2 36.3 22.3 35.8 24.1 Pra malls 35.3 22.0 35.2 22.1 35.1 22.2 35.0 22.3 34.5 23.6 2a 33.7 22.3 33.8 22.4 33.9 22.4 33.6 22.5 33.2 24.3 20 32.6 22.3 32.9 22.6 32.6 22.6 32.5 22.9 32.1 24.1 ad 31.2 23.0 31.4 23.0 31.4 23.0 31.5 22.9 31.1 23.7 I. sss 30.5 23.9 30.6 23.5 30.4 23.6 30.2 23.7 29.7 25.0 A 30.1 24.4 30.1 24.3 30.2 24.4 20.9 24.5 29.1 25.4 G 29.1 24.8 29.3 25.3 29.3 25.1 29.1 24.7 28.0 25.6 A: -. 28.9 25.0 28.4 24.9 28.1 24.9 28.0 24.9 26.3 25.4 8-9 September 1966 bacrricccsle 24.2 14.3 23.3 14.4 22.6 14.5 22.0 14.6 18.2 14.8 O ecs 23.8 14.2 22.5 14.4 21.8 14.5 21.4 14.6 17.2 15.0 24.3 15.0 23.9 15.2 23.6 15.3 23.3 15.4 20.1 16.0 5 misc 27.5 15.8 27.4 16.0 27.5 16.2 27.3 16.3 26.3 16.5 P 30.2 16.3 30.4 16.4 30.6 16.5 30.7 16.7 30.8 16.8 10 32.2 17.1 32.6 17.2 83.1 17.3 33.4 17.4 34.2 17.6 11 35.0 15.0 35.4 15.5 35.4 15.8 36.3 16.5 37.4 15.3 12: 36.7 14.4 37.5 14.6 37.7 15.0 38.0 15.5 39.7 14.5 13 .s 37.8 14.5 38.1 15.0 38.5 15.4 38.4 15.5 40.7 14.9 38.3 12.8 38.9 13.2 |89.5 14.0 39.6 14.6 41.4 13.6 1h :: oue 37.6 14.2 38.1 14.4 38.9 14.6 39.6 15.0 40.7 13.9 to 38.2 13.4 38.5 13.6 38.9 13.8 39.1 14.1 39.8 13.3 17 38.1 13.8 38.5 14.0 38.7 14.2 38.1 14.5 38.2 13.3 18. 37.6 15.2 37.4 15.4 37.6 15.6 36.7 15.8 33.9 16.6 19. 35.3 15.8 33.8 16.8 33.2 17.6 32.5 18.2 27.1 20.0 0 32.8 15.6 31.9 16.6 31.5 17.3 31.1 17.8 24.7 18.6 lE. 32.8 14.4 31.4 14.7 |80.1 15.0 29.0 15.4 23.8 16.9 iface rare 31.2 13.6 29.6 13.7 28.5 13.9 27.6 14.3 22.6 16.8 ers 27.3 15.3 26.1 15.5 25.6 15.7 25.0 16.0 21.9 16.6 BL c:: 26.6 15.0 25.9 15.2 25.5 15.3 25.1 15.5 22.1 16.1 T 25.4 16.4 24.6 16.4 24.2 16.5 23.6 16.5 21.5 16.8 2 24.4 15.8 23.3 15.8 22.9 16.0 22.5 16.0 20.2 16.3 9 24.4 15.5 23.3 15.5 2a. T 15.5 22.1 15.4 19.1 15.8 eatin 24.1 14.6 22.7. 14.7 22.1 14.6 21.5 14.6 18.7 15.2 9-10 September 1966 23.1 13.8 22.1 14.0 21.8 14.2 21.5 14.1 18.7 14.8 Ca. -is? 23.3 14.0 22.4 14.0 22.2 14.1 21.7 14.0 19.2 14.6 Tver ure 23.3 14.8 22.9 14.9 23.0 15.2 22.8 15.2 21.5 15.4 8: 15.7 26.4 15.6 26.5 15.8 16.6 16.0 25.6 16.8 9 2 ¢ 30.5 15.5 30.7 15.5 30.8 15.8 31.0 15.9 31.6 15.3 10.2. :- 34.0 13.6 34.3 14.5 34.6 15.2 34.7 15.7 35.6 14.3 11 35.9 13.9 36.4 14.0 36.6 14.1 37.0 14.2 38.9 14.3 WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA F67 TABLE A2-1. Hourly averages of air temperature and vapor pressure at the Buckeye Project during selected periods of 3 days each. -Continued 1-hr period 11m 7m 5m 4m 2m beginning °C mb *C mb °C mb *C mb *C mb 9-10 September 1966-Continued 12 HELLO 37.5 13.3 37.8 13.8 38.0 14.1 38.4 14.3 40.0 13.6 19 m...... 38.6 13.0 39.0 13.0 39.3 13.2 39.8 13.4 41.4 13.5 14 39.2 13.5 39.3 13.5 39.9 13.6 40.0 13.6 41.7 13.2 15 38.5 13.5 38.6 13.8 39.2 14.0 39.0 14.5 59:7 14.7 160 J...... 37.5 14.1 37.6 14.8 37.8 15.2 38.2 15.5 38.6 15.0 f :A 37.7 14.9 38.0 15.0 38.3 15.2 38.4 15.5 38.6 15.0 ts Sev cl.... 37.3 15.1 36.7 15.4 37.0 15.6 36.5 15.8 35.0 16.0 19 34.8 15.5 34.1 15.6 33.6 15.8 33.2 16.0 28.0 17.6 20 [ 33.5 15.2 31.1 15.4 29.9 15.8 29.5 16.1 24.9 17.5 BF 30.6 15.0 28.7 15.2 28.1 15.4 27.5 15.5 23.1 16.6 P2 i 29.5 15.1 27.9 15.2 27.3 15.3 26.8 15.3 22.2 16.4 RJ 29.1 15.5 28.2 15.5 28.0 15.6 27.3 15.6 24.0 15.6 P4 26.3 15.8 25.9 15.7 25.7 15.7 25.3 15.6 23.5 16.0 I 24.7 15.6 24.0 15.8 28.17 15.8 283.2 15.8 21.2 16.0 YA: 26.2 13.7 25.4 14.0 24.3 14.6 283.0 15.0 19.2 15.6 o 26.1 12.6 24.2 13.0 23.6 13.6 22.8 13.8 19.0 14.4 # 25.1 12.3 24.0 12.4 28.4 12.6 22.6 12.8 18.8 13.4 10-11 September 1966 23.6 12.5 22.2 12.5 21.6 12.6 21.0 12.7 17.8 13.7 25.0 12.0 23.1 12.1 22.0 12.2 21.7 12.2 16.8 13.3 25.1 12.5 24.2 12.5 24.3 12.6 28.7 12.8 19.8 13.8 27.7 14.1 27.5 14.5 27.8 14.9 27.1 15.7 25.3 15.0 32.0 13.4 31.5 14.0 31.4 14.3 31.3 14.7 29.3 14.6 33.9 12.9 34.4 12.9 34.5 15.0 34.3 15.1 32.9 14.9 35.3 13.4 36.1 13.5 36.8 14.3 37.0 14.6 36.6 14.0 36.6 13.6 37.0 14.0 37.5 14.2 37.9 14.5 39.8 14.4 38.3 12.0 38.0 12.4 38.2 13.0 38.9 13.4 41.0 13.8 38.7 13.3 38.4 13.5 39.7 13.8 39.5 14.0 40.6 13.7 38.9 13.5 38.2 13.6 38.3 14.0 37.8 14.1 38.1 12.9 38.3 13.0 37.8 13.1 38.5 13.3 38.0 13.7 36.5 13.8 38.0 13.0 38.1 13.5 37.8 13.7 37.7 13.8 34.6 15.3 37.7 13.8 37.1 14.0 36.8 13.8 36.5 14.3 32.4 16.0 35.6 13.9 34.6 13.8 34.2 13.9 33.6 14.1 27.8 16.8 33.8 13.6 32.8 13.6 32.1 13.8 31.8 13.8 27.2 14.9 35.2 12.0 34.6 12.3 33.7 12.6 33.6 11.0 31.4 10.5 33.6 11.6 32.8 10.0 32.2 12.5 31.9 10.4 29.2 11.0 32.0 11.5 31.4 10.9 31.0 13.0 30.7 11.3 27.2 13.0 30.4 12.5 28.9 12.5 28.3 14.5 28.0 12.6 24.3 13.4 30.5 12.2 29.5 11.6 28.7 13.8 28.8 12.0 24.1 14.0 27.4 14.9 26.8 13.5 26.2 15.1 25.6 14.4 22.1 15.2 27.3 17.5 25.9 16.5 24.8 17.8 24.1 16.7 20.6 16.1 25.6 17.1 24.0 16.6 23.3 17.5 23.0 16.5 20.0 16.7 23-24 March 1967 Merl f yes p == fuks vek Fave ases Coxe Sets 8.5 4.5 7.8 5.0 6.5 5.0 5.8 4.8 2.8 5.3 9.5 4.5 8.5 4.8 8.2 4.4 8.0 4.5 4.8 5.6 BAP 12.0 5.5 12.0 5.6 12.0 5.5 12.0 5.4 10.5 6.0 16.5 6.1 16.8 6.4 16.8 5.9 16.8 6.4 17.0 6.5 10 20.0 6.0 20.8 6.2 21.0 6.1 21.2 7.1 22.5 6.3 .l ns... 24.0 5.3 24.5 6.6 24.5 6.2 25.2 6.6 26.5 6.2 27.0 5.9 27.5 6.2 28.0 5.9 27.8 7.0 30.2 6.2 19 it...... 29.0 5.8 30.0 6.8 30.0 5.9 30.2 5.8 32.0 5.0 14 :l... 30.0 4.3 30.5 5.6 30.8 5.8 30.5 5.3 32.0 4.6 29.5 6.3 29.8 6.9 30.0 6.8 30.8 7.6 31.5 7.0 16 ......... 30.0 5.1 29.8 6.1 30.0 6.0 30.2 6.7 30.5 7.0 17 %.. 29.5 6.0 29.8 6.5 29.8 6.0 29.8 6.5 29.2 6.9 18 t.. ...u. 29.0 5.0 29.0 6.3 28.8 5.8 28.5 6.0 25.8 8.2 19k. n...... 26.5 6.2 26.2 7.5 26.0 6.8 25.5 6.9 21.2 7.5 LO 26.5 5.8 25.2 7.0 24.5 6.6 24.2 6.4 18.5 8.5 26.0 5.0 24.2 5.6 23.5 5.7 23.2 5.9 17.5 7.6 f 25.8 5.1 25.0 5.2 24.8 5.3 24.5 5.5 22.2 5.4 24.2 4.6 23.8 4.8 23.2 5.2 23.0 5.3 20.5 5.4 I 22.0 5.1 21.0 5.6 20.5 5.7 21.0 5.1 17.2 5.3 02:2 18.2 5.4 15.8 7.0 15.2 6.5 15.0 6.6 13.2 6.5 P 18.5 5.9 17.0 6.3 15.5 6.4 14.5 6.6 12.5 6.6 Biel.... 19.2 5.8 18.0 6.2 16.2 6.3 16.0 6.0 12.8 6.4 fick 18.0 6.2 17.0 6.5 16.5 6.5 16.5 6.1 13.5 6.6 2m mb °C 4m mb °C 5m *C mb 24-25 March 1967 STUDIES OF EVAPOTRANSPIRATION 7m mb °C mb 11m °C TABLE A2-1. Hourly averages of air temperature and vapor pressure at the Buckeye Project during selected periods of 3 days each. -Continued 1-hr period F68 beginning sn o < op in Ninn oH t mn o into Gon oH n O oo Hn o inin u a w of n o o in o «tin «tor i o o u T 2 S o np o o m eny 16 to ~ 06 06 ~ NNN (G 16 in in 16 xh shoo hist w © o n- mN 00 06 of ~ <6 46 is in is 6 ~~~ 1G in in in t 16 to in im & & & as so 60 $0 os so ~ 9p in in & eq aq op & on in on an in an ir < an on on <> ey on <- 9p on itp & ao oy B Hg ats on (94 ire 4 of o l Co Mo CG O4 is is C4 00 04 C] og in coin to on on o th a th da & p SOT SOT SOT STT 88'+- 88 +- 88 +- 0 0 0 0 0 0 0 § c nino neden ¢ SOT 86°0 SUI STT 8o'g- seg- 88 +- 0 0 0 0 0 0 0 . 2 SUI go'T go'T all 8a'g- sag- 88'F- 0 0 0 0 0 0 0 f ° T So'T 86.0 86°0 STT 8a9- 88 +- 82'9- M 0 0 0 0 0 0 Q n- Ma 16.0 16.0 So'T LLO 8T F- s29- 0 0 0 0 0 0 0 { nnn nnn -->> ga 16:0 16.0 86:0 16:0 88 +- 169- L169- 0 0 0 0 0 0 0 Q > 22 04.0 0L0 040 LLO e L9 L- Lg8- 0 0 0 0 0 0 0 Q n> 12 Cr 0 6r°0 9E°0 169- oro- 9L'6- 9L6- 0 0 0 0 0 0 0 Q > 0€ L0°0- 0 L0°0 L9 L- 8T'P- 90'6- 90 '6- 68 T- oL'o- 6€T- 68'T- 88°F STF 88'F gey :t 6T 9$8:°0- 99°0- 99°0- 0 60°C 0 6LC 6LC 6rg- 829- ZS 6T S6ET I8 ZZ s oon foton e 8T LLO- 16°0- LLO- £9°0- 2C O eava Ze 0C 9L'6- 9L'6- 9L6- 916- PT'TP P8 IP P8 TP tonto nono IT §00- OT- 86°0- 16:0- 9888 9g 88 ag 8g ag s8 TI- - S#7I- g8II- - gf11- 9g to 9029 1909 1909. ~~~ 91 66 I- 09 I- 6E T- egrs 6gre 69:ge eo'es s6EI- _ S6°gI- S6EL- S6ET- 6r6L 1091 89°g8 --- er SPB- ( £6°0- Cr OL 949 CIL a PESI- PEST - pPEST- pECT- €9°06 €9'06 $906 @@ T6 ~~~ PI 68 I- 66 I- 96 I- 98 T- CEL IBL n ZCEL rO'9T- ro'9T- ro'9T- rO 9T- 2066 TE 86 E86 Thee -o --> M 06 6- OE Z- frst 9Ts- 16°84 1681 16%L rO'9T- ro9T- po9T- TE 86 29 16 Tese Th BG =-- 21 PiI- Er.0L ZP 0L cr 0L Er 0L PESI - - PESI- PE ret- PLT po'ce pi'ze PT P6 ~~~ I1 L9 T- £21 L9 T- PL I- 19.09 Le 69 9819 1909 P9 FL- s6EL- F9 TI- P9 TI- goes 82 Ts g9gs RE pg n --- ot 09° T- IgT- I8 I- 9G T- 20 9P 20 IF 20 '9r 20 IF SCEIL- SCEI- 96ET- F8 89 +919 pg 89 Cbg toc 6 ero- evo- evo- 96:0- 86°65 86 66 66°60 86 6G 9T'TI- 9T'IL- 9T'II- 9T'TI- T8 8F T8°8F TT'SF oan g £00 8 0 PTO sagt SC BI sot S6ET 16°9- 169- 16°9- 19 L- ba ez 68 LZ 68 LZ 86 6g =- 1 r8.0 LEO 16.0 ¥8°0 0 0 0 oro- 68 T- 6€ T- 68 T- 60%- 906 19 L 906 . o-- 9 STT Sot SET SET seg- so'e- so'e- gog- 0 0 0 0 0 0 0 ( > ¢ 9961 2unf 9-p w 9g 9p Ls w arg ha pg w gre 9+ wE w 978 97+ p€ eq Sutuu18aq S "¥ "U poured go'9T- 86Lt- sPET- re ger Zr PEP 92 ILP gL'6PF - £9°0GI- _ - P8IT- _ ST60L zT'g69 09008 19 uo eop LLO £90 16°0 F8°0 6L%- Lo- 8TF- 8T F- 0 0 0 0 0 0 0 Fini * LLO 99°0 16:0 16°0 6rg- 6L%- 8TF- 8T'F- 0 0 0 0 0 0 0 Q --> g £9°0 Ct 0 99°0 F8°0 88 +- 88 +- 88 +- 8T'F- 0 0 0 0 0 0 0 Q - g 6°0 L0°0 €9°0 LLO sore- L9 L- 8TF- 88 +- 0 0 0 0 0 0 0 Q --- -- { 99°0 0 LLO 8o'g- LeB- 8T'F- 88+- 0 0 0 0 0 0 0 (p | ~~ pg SP0 12:0 F8°0 800 L169- STF L69- 0 0 0 0 0 0 0 Q --> ga Cr 0 82°0 T6°0 0 829- 8TF- 90°6- 0 0 0 0 0 0 0 Q nnn 22 99°0 ou F8°0 6r°0 81 +- 6rg- 6rg- 0 0 0 0 0 0 0 Q nmn 12 or 0 10.0 LLO 8o'g- 829- sog- seg- 0 0 0 0 0 0 0 ° 0Z L0°0 82.0 40:0 8T F- 6rg- 8T F- 88 +- oro- oro- 68 T- oro- ove 6T 8T F epg --- 61 6r.0- 99°0- sr0- 6E'T 6L'C GLC 68T- 6LT- 6€ T- 8T'FP- 60%- <9 I 9TTT PE CT TTL o-- 8T 16°0- SO'T- r8°0- 16:0- FO'9T ETSI SPLT 169- 829- 19 L- 19 L- gg'ge vee 96 98 ggiag ==-- It ui 8TI- 86°0- rO'T- L8'vE 86°62 95°98 98°88 9L'6- 90°6- 916- $16- og'cs Zo or 69 ge BTG o-- 91 60°6- $9 I- 192- 12:08 ECEF 09: TG 9T'TIL- 9r OT- 9TIT- I8 TI- 88° 1G gree ITIL Pigy et £00- TFC- PLT- £1.69 £069 ST F9 OL'QL 9T'TI- so Lgs- T0'9L 9818 PT iLg too" -~ pI PEI~ L9'T- gr I- £0.69 6LBL TO'9L SET - S6EL- SEEL- gg It- rLT6 PLT p06 paige --- g 08 6~ 88'T- 8I°C- LE C- 1699 6628 6L'8L £889 am SEET- eo LL'eg PLT o ooo 21 S8 E- il f- 00°€- 81.99 re :89 sr E9 g9e9 ICEL - IEEL- aces SET LP 98 gpipg onan sono 11 $TE- F2 8PE- 8rE- 9699 8T'L¢ SLZL- go SI LTLL OF LL OP LL OL BL o-- ot I9 8- I28- 8re- £8°6- LUCY ECEr ECEFr P8'TF 6E7TT- S8 IT- gL II- 9T TI- Esto 91'To 9g To 98 Tg ~ o-- 6 PUIT- I8'T- ILEC EO 00° IZ 098- 19 L- 90°6- 906- S0 68 Lr eg P8 TP PE Tp ~~~ g €6°0 8C 0- 86 0- IFO- 69°01 oral 9L6 916 ggg- gc'9- go'g- 88 +- 1082 og CZ I€22 B& OR on 4 $9°0 £9°0 16°0 Er 0 981 - 60%- oro- 6ET- 68 T- 6€ T- 68 T- IT's gag gore Vp . oooh ooo coisas 9 16°0 £9°0 LEO 88 T- 8I'F- 8T'P- 0 0 0 0 0 0 0 Q -> g 9961 4em 8-5 w 8/L L9 99 u 8/4 L9 978 w 8/L L9 998 w 8/L L9 978 'eq Buruurdaq poured 'ay-1 s "X "I [a2gowmu00 arenbs a0d ut [re ig 'mopgeey [ros pure '"; 'uorerped jou ("j 'opaqte :* 12]0g] mot f10s pun jou aanm-4ioys pun -8u0f 1n10 'opaq;n aanm-4i0oys anfos {mo r-T-gy a1gv J, F71 WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA CS ESE 96°00¢ 88 PPF £92086 STLIT- CV ECI 69901 - 26 689 FTOT9 60 +19 ep rer: ctc roto ana 199L £091 - IOLI- C6 PI - 8I'P- PTO 10°0 12:0 12:0 88 +- 8T'F- 8T F- 0 0 0 0 0 0 0 0 ( no- -- p L0°0 L0°0 40°0 FTO 88'P- 88 +- 88+: 8T F- 0 0 0 0 0 0 0 Q nm ¢ L0°0 L0°0 10°0 L0°0 88 +- 88 +- sog- 88 +- 0 0 0 0 0 0 0 ( nnn nnn neenee -= Z 0 ) LO°0O- L0°0 8g°¢- 88 +- 169- 88 +- 0 0 0 0 0 0 0 p nen I 0 LO°0- 0 0 869- 88 +- L6°9- de 0 0 0 0 0 0 0 FG L0°0- PTO- 0 8o'g- 88 +- eq:guu, 0 0 0 0 0 0 0 le gC pTO- 85°0- 0 pTO- - 8g¢- 169- 88+- 0 0 0 0 0 0 0 Q n- RZ 88°0- T20- - 829- 169- 829- 0 0 0 0 0 0 0 Q n- --- 12 98°0- 0L'0- PTO- - 8ST'FP- 8g'g- 16°9- 0 oro- 0 0 0 0 0 Q n- 0Z OL0o- 86°0- €9°0- 6ro- - 60%- 60°C 88 +- 8T F- 6€'T- 6rg- 0 on'o- 68 T 60% 6€T O onsen akin ook. 6T P8°0- 86°0- 86°0- €9°0- _ 88°F 809 STP 88°F 6L%- 6rg- oL'o- S6ET P9°PI P9PI OOT senso 8T SUI- 8UT- SE T- r80- - PEGI PE SI 2C 0G 96 88 +- 9T F- 8T'F- Maca IPPC BLBL M SE T- CE T- $9 T- 8SI'T- - 89°06 09° 99'9€ 86'6C 19 L- L9 L- L9 1~ Les- PEF 9698 06 0¢ G@L@§ 91 $9I- E9 T- I8'T- SEI- 9916 69° 8C 16 08 LYEE 906- 9F'OL- LeB- TSGF P8'IP ecreq Popp oI L9°T- 96'T- L9'T- _ IG°6P 9P E9 PS CF S8 IL- 98'IT- SS CL- 9T'IT- gge9 sr g9 PT PaT~- 88'T - L9'T- - 98:19 99:99 rE 89 120g S6EL- P9 MI- S6ET- ss L- 681 10°91 Lr 98 B& BL --- gT 09 I- O9 T- L9'T- - 8T'L9 96189 +299 92°98 SS PEST- 1971- 19 4- ean LLeg 6To8 O9 TQ nono 21 $9 I- LO'T- 9rI- - 1989 99:39 98° E9°PP 9g 19° MI- s6EL- 906- Z9 EL secs BOQ ton- TT L9°T- O9 T- 88 I- 09'T- - O88 OE CS 99 I8°8F S8'IL- I9 IL- SEgL- 90°6- ergo pg 89 gr0L OT mtaal S8 I- oP I- 9918 pTIP rSCF 65°62 g8TI- _ I8TT- S8 TI- 9T TI- TTSF grog gpg 88 Tg noon --- 6 99°0- EP LLUO- - 160 29 TZ OT'SZ 26 0 90'6- 91'6- Le B- L8ve 9298 CB 88 9g dongs g 1800- LO°0- 0 6¥'0- _ 9F'0T 916 90°6 SSCL 8g'g- 88 +- 88 +- 169- ETBI L 180 85.0 82.0 0 6E T- oro- 6€'T- 60%- 60%- oro- gorg- 61% 61% 607% BLE --- 9 120 120 L100 __ 8T'F- 6rg- 8T F- ( a 0 0 0 0 0 0 0 f° H 9961 isnSny 9-p w 979 Ip w€ w 978 Ip wE w 98 Ii¢ c w 978 Ii¢ s 'Aeq Buruu18aq - poued "H "q 80'IT- L9 TI- 8L8- 69°ET- 08F EC°E8F 99668 68 FTF g6GHI- - - Z8PEI- - I€IGI- _ 90'%gL SL'09L 68 ~> amor. 9E°0 FTO 120 £9°0 88 T- 8I'P- 89°¢- 88 +- 0 0 ) 0 0 0 0 Oon v coc tt 820 0 FTO €9°0 8§8- E 88 P- 0 0 0 0 0 0 0 $o t ¢ 82 0 L0°0 86.0 88 T- 88 +- 89°¢- 88 +- 0 0 0 0 0 0 0 $e oa oo Fa 96 0 L0°0 6r°0 SP0 88 T- 88 +- 8I F- 88 +- 0 0 0 0 0 0 0 pree e ett Heron I 960 L0°0 6r°0 89°¢- 869- 88 +- 88 +- 0 0 0 0 0 0 0 0 Are FTO 0 820 FTO sare- 169- 6rg- L69- 0 0 0 0 0 0 0 " a ae gz 10.0 40:0 82:0 PTO- 88 +- 8T F- 9L'6- 0 o 0 0 0 0 g ( 2C 10:0 FTO 10:0 10°0- so- sag- 8T F- 91'6- 0 0 0 0 0 0 0 § 0 nnn nnn enne 12 L0°0- 0 820 - L9 L- L69- 8a9- 9p OT- 0 0 0 0 0 0 0 § 0 onne hta 96°0- 82.0- 6¥°0- 8T F- 6LC- 6L%- L9 L- 6ET- 60%- 68 T- gee L6°9 88¢F C... vides 61 99°0- 99°0- 85°0- r8'0- 8T F LG'8 64% 60°C 88'+- 60%- 8a'g- ET8T 10°82 SS CT 8T OLO- £9°0- Sr- STT- Go°CI 29°61 6LC €0°91 L69- 10 6rg- 90°6- 82°08 92°98 PEST Cont hona ek: * L 16°0- LLO- P8°0- SFI 69°82 6T'LG Z9°6I 9L6- 9TTL- 9F OL - Le B- 68 pS L669 secs (pipe (Aon once 91 L9°T- 9P I- ET- 9T C- OECS 810g 6628 CP LP S6EL- 96EL- SS ZC EL 19¥L T6'€L BFL "Ants: crane Fh £00- 608- OE C- LE @- 98 F9 16 99 +6 99 9619 PLL - rEST- PL9L- ET8T- LL'S8 2668 itn PI STI- 86°0- li 09° T- 9129 Z8 TL 18'8F rE 89 P9°PI- rO'9L- 99 L- 868 SET6 srg9 H | ah r ~ gt SLT- 09° T- PLI- go't- 96°68 S8 TL Er £1.69 PESI- SS IL- PLIL- SE T6 2906 Ses Frie. ZI tg t= 19T- +699 €069 1699 PEgI- - vESI- - PO9I- - 9918 Lr98 9T L8 Ba'es -- II S0 T- 16°0- 66° T- wa 88°18 8p 98 8P98 g6'8T- 96 EL - 96 EL- P9 PI- 6r6L TO°9L 8C C8 6108. ": OTI OT L9 T- PLI- £9°0- 99¢- 99°L€ £6°Er OT IZ E6°6r 96ET- 98 TI- 829- 09 te 1909 rara 6 evo- 86.0- £9°0- IFFE 68 LC EP LL 68° LC 9L'6- 9L'6- 90°6- 9T TI- PE ZP PEP 18 g 0 LO°0- $10- FTO 98 TT 98TI 99° CI 9T'TT 829- 8o'g- 82'9- L69- O8 sz ILE 09°92 (gag --t gief L 85.0 9€6°0 16°0 6T - oLo- oLo- 6ET- oLo- 6€7T- 68 T- 829 6LT 191 G1 se ess "thg €9°0 tao £9°0 STT sa'g- 6rg- 88 +- 0 0 0 0 0 0 0 OLG 0 ton ¢ 9961 4inf 6-4 w 6/8 g/L 19 w 6/8 g/L L9 wu. 6/8 g/L u9 w 6m g/L v9 eq Suruut8aq w poured S "I " "4 j10s pun jau pup -8u0f 1Dj0j 'opaqn aapm-440ys [D30 -'T-gy g1gy J, SLb- 9G I- L9°P- 06 LL06€ 0 908 6F 808 09 FOI SUSIT- 09 80T - pTFT6- 26969 26 689 ET FES AB BOF oor nner 199L TUDIES OF EVAPOTRANSPIRATION STT 86°0 99°0 88'F- 0 0 0 0 0 0 0 M a p STT 16°0 LLO L69- 90°6- 88'F- 0 0 0 0 0 0 0 i : agad siir on g 860 LLO P80 859- 916- sare- 0 0 0 0 0 0 0 g. 2 86:0 £90 16:0 90'6- 191- 9TIL- 19° 4- 0 0 0 0 0 0 0 eram tae aes - I 86°0 £90 £90 906- 104~ 9I'IT- 906- 0 0 0 0 0 0 0 gare are ieee hs ra STT 00 85.0 9L.6- 169- 9F'0OT- 9TIL- 0 0 0 0 0 0 0 uart esis EC STT Lio 12:0 906- L69- 90°6- I8 L- 0 0 0 0 0 0 0 p: enfin san 2C am 00 00 169- sarg- L69- 90'6- 0 0 0 0 0 0 0 ane ores 12 92 T 0L'0 00 LE8B- LgB- 9L6- 8a9- 0 0 0 0 0 0 0 QA arn Ma 9O'T 99:0 £9.0 906- 9L6- 169- 0 0 0 0 0 oL'o 0 ane nees re 6T 96°0 85.0 12.0 88 +- 8T'P- 6re- 6L%- oLo- 169 906 L9L gap A nan 8T FTO L0°0 8T 'P 89g 6P°€ 6LC 88°+- 6L%- 200 IFFPE IS CC COT neden Lt SPr0- LLO- OL'O- eave [Ex4a 68°LC 9F OI L9L- 906- 90°6- 8T P- 9918 P8 IP 20 TPWE --- 9T 66 T- SS T- SE T- L8're CE IP 99'9€ IFFE 9F'OL- am 9TTII- Lg8- 8T LG pz 99 9g To B& Bp --- ef 9Te- $- PLI- £€9TP Eras 9TIL- S8TI- <8 IT- 906- Lees 1919 9€ to BP Lp --- PI PPE- L9 1- PPC- 20°99 9029 98°98 Pr 0p 9T'TIL- S6EL- 90°6- 9T'II- v699 8988 gr9¢ €T IE d S6T- £8°6- 09 TS 19°09 95°96 L989 SS CL- 96 EI- Ec S6°ET- TESL 86°C8 L9°09 ‘ CI LUO- S8 I- 60°88 68 P8 66°28 S6 EL - 96 EL- S6ET- 96 EI- OL'OL TO°9L 6L8L II 96° P8 IFP eras F8Tr PI'IFP SS CL- SS CL- S8 IL- SS 9129 9129 9109 OT 160- £9°0- 99°0- 6TLC 08°SZ 6T'LE 60°62 9F OL - 9T'II- 9F'OIT- 9F OL - IT'8H £9 PP 09'TG Gp 6 PTO- 0 916 LE°8 96°ET L6°9 859- L6°9- IE2E ILEC 29 TE sobs er fa ea 8 86°0 €9°0 LLO 6LT- 6L%- 6ET- 6€T- ou'o- 6E7T- 8T'F 829 6rg opere nour then fe doco ranean s L 8T'T £9°0 STT L6°9- 19 L- L6°9- 0 0 0 0 0 0 0 Qe Aera rao "aome 9 S2'T $90 | . .l 82'9- L9 L- $eF_-~:.. ~~~ 0 0 0 0 0 0 0 p y ¢ L961 YEN 925-¥6 90/42 SCFC FEIS u 90/97 SCFC FEET w 90/42 SCFC POET u 90/87 SCFC FEIEZ £eq MAwWM-WMM S "u 'H "H $9 ¢- 29 8- 881 - v8 662 CL vOE T9 6LG SP PIE SP III- SO'SIL- C8 YEeL- 99 T8 £9°098 ok a +o baleen a 19°L 16°0 16°0 seo. "~~~ 169- 869- 0 0 0 0 0 0 0 16°0 86°0 L69- 809- 809 - 0 0 0 0 0 0 0 LLO 86°0 869- 809- 809- L69- 0 0 0 0 0 0 0 99°0 LUO OL'0 809- 88 T- L69- 80 0 0 0 0 0 0 0 99°0 99°0 LLO L69- 88° L8 8- L69- 0 0 0 0 0 0 0 96°0 £9°0 16°0 L6°9- L9 L9 L- t 0 0 0 0 0 0 12:0 F8°0 16°0 L9 L- 9F OI - L69- 88 %- 0 0 0 0 0 0 0 FTO F8°0 L9 L- SUI 0 0 0 0 0 0 0 €9°0 LLO LUO L69- L6B- L69- 0 0 0 0 0 0 0 SU CVO L9 L- L9 L- LeB- L69- 0 0 0 0 0 0 0 L0°0 PTO- PTO- 8T'F- 88 +- 8I'F- 8TF- 6L%- 66° C- L69 88°F WB MATA 8T £9°0- 99°0- L9° L 6P°€ 99CI L9°L 869 - LO L- L69- 29 TZ PEST IFFE L Noo cone LT 16°0- r80- 16:0- PLT 89°9 08° IZ L69- 6rg- 1L69- 9L6- LP EE 18 22 BLG noon --- 9T 86°0- STI- 8I'T- LLCE SE TE d P8 CY 90 6- L9 L- LGB- S8'IL- CV LP 99'L€ F8'TFP 9009. contes fen terone ne ST 09 I- Ig1- I8 I- 12:08 I9°6P CP LP 6608 S8 IL- QI IL- S8 II- SC EI- EF0L FE 89 iid) -- gl e PI £0G- 91 C- 20 G- 9029 v6 89 L9Bs Lt 69 S6°EL- S6 EL- 96 ELI- 6108 8C CS8 TESL i ir alty ane dret areal tr tat os ta ET 09 T- $9 1- 191- 19°09 tg 89 60°49 19°89 S6 EL - P9 I- S6EL- S6 EL- 89°I8 86°28 OT'8L SO Tea o: fetta tnt rat 9G I- 9P I- SE T- 88°18 9029 60°88 S66L- 96 ELI- S6ET- S6 EL- OT'8L OPLL OT'8L OMSL (s ol nene rea II Og I- E6 G- IT8F OECS 2O'9F 2O'9¢ SCEL- SC EL- SCEL- SC EL- v6 99 v9 L9 v2 99 $o Le osa tate ema t OT 99°0- OLO- r8'0- LP EE 96-98 89°06 80°CE 98 II- S8'IL- S8 IL- S8TL- 09 TG 09 TS T2'0¢ 08 es Aint as fronts 6 98°0- 6°0 PLOT ET8I P rO'9T 90 6- LEB- 906- 80°2E LLTE 89°08 Wipe Shn cope nn onn t 8 99°0 CV 0 ZP 0 60 C- 60°C 66 T- G- 8T F- 8T F- 9F OT S8 L6°9 |M g a tos tct w L SOT OL 0 160 89°¢- 899- 88 +- 0 0 0 0 0 oL'0 OL'O 6B m teo s ht ant 9 86 0 LLO 16°0 LG9- L69- 80 89°¢- 0 0 0 0 0 0 0 OA aoa mena ones g 9961 soqwoardag 11-6 w TL/OI O16 6/8 u TL/OT O16 6/8 w TL/OT O16 6/8 W TL/OT 01/6 6/8 eq Sutuu18aq poured S "H 'd " F72 10s pun jau aanm-440ys pun -8u0f D30; 'uouymppi aapm-40ys 1»f0s [po a19v J, F783 WEATHER AND EVAPOTRANSPIRATION STUDIES IN A SALTCEDAR THICKET, ARIZONA FTO STI- 20°C 20 EEF TG°96F L6IFF LPIOTI- - gg'GOT- _- Og'9sT- _ ag'gor- _ 29 s6L z6 69L BQ g- ~- n= no- mop 6ET 0L0 8T'T 829- so'g- L9 L- 0 0 0 0 0 0 0 Q n en- p 6ET €9°0 SCT L69- L9 L- 829- 0 0 0 0 0 0 0 Mle £ SET SOT SCT 82'9- 8a9- 829- go'g- 0 0 0 0 0 0 0 le 2 SOT 16:0 STT L69- 19 4- 16°9- 8a9- 0 0 0 0 0 0 0 Q nnn ne- T LLO OL 0 an LEB- 90'6- L6°9- 0 0 0 0 0 0 0 Q n- --- pg OL'0 16:0 16:0 90°6- 9F'OL- L9 L- 0 0 0 0 0 0 0 Q nn gz 99°0 STT E°0 9L'6- 9I'IL- L9 L- 9L'6- M 0 0 0 0 0 0 Q n- a 0L'0 82.0 9r'OL- §III~ LEB S8'IL- 0 0 0 0 0 0 0 Q --- I2 820 12:0 98 TI- SS CL- 9F OI - 0 0 0 0 0 0 0 Q ne- --- 0€ 12°0 LO°O- FTO 9L.6- A 9F OL - 681- 6E7T- 60%- 6€'7T- 6r'g 68g 88'P 60g . 6T vio- PLO- oro- oL'0o- 0 60%- oL0 see- 16°9- 6rg- Ec PEST E8°8T SSTI PQIPL o-- 8I E8°8T ET'BI E8°8T 9TTL- 9T'TI- 9p OT- 9T'TI- 2918 Greg 18°P€ ooo 8T 6P°0- £9°0- 6r°0- 96°96 92°98 96°98 99'L6 S6ET- S6ET- S6ET- Lzeg Lees Lees MBB coon HT F8°0- SUL- €9°0- 6688 09 TS 8T1LG 12:08 19 TI- PESI- P9 TI- P9°PI- TESL OPLL 10°94 eGL * CT 96 1- 60%- S6T- STF TFO v2 99 20 29 PLOL- pO9T- PL9L- S2 68 2668 sc 6s I& L8 o-- --- PI SE T- 66 T- 90'T- C9 CL 16°84 CCEL EF'LL- EF LL- ESFLL- EP 2066 T8 86 28°86 TGG --- gI 66 T- 9P I- 66 T- 16°81 28 TL 19°FL TO'9L EFLL- EFLL- EFrLL- EP 2066 2g 86 ZE 86 TPOOT === === 21 OT- S0'T- LLO- £069 £1.69 28'TL pO'9T- FO9T- pO'9T- p0T6 SE T6 SE T6 PTF C-- 11 9O'T- 9O'T- OL'O- 19.09 L9°B9 L6'6S 91 29 PESI- PEST- PEST- PE ST- 858 68°08 68°08 ggg fck OT STUI- STI- SO'T- CB CP £6°6F €9¢+P LOF P9 TI- P9 +I- 19 +I- eg'eg gg p9 P299 $q 6 f10- 1800- kld 08°CZ OT IZ s8TT- 9T TI- So L- 9T IT- E6 Tr g6gr ger BO'Op ~~~ g PTO 99°0 aly L69 906 6E7T 916 88 +- 60%- g29- ZS6T TL EZ IP OT TPPE O- --- I £9°0 16°0 ass 6LC- 8T'F- 60%- 687T- 6E7T- o'o- 6gT- 88 P 88+ 0 conn ---> 9 LLO SOT SUI L69- 19 4 - L69- 0 0 0 0 0 0 o § nnn --> ¢ L961 4eW pI-2I FPUET EUZT BVI w PUET EUVZI adm w PUVET EUZI SUIT wu PUVET EUVZT EVI Aeq Sutuut8aq poued y-; S "3 'd "H ponuruo;y-mof/;nay f10s pun 'uormpo4 jau aanm-440y48s pun -8u0f [D303 aavm-440ys 1nfos {p30 angy I, STUDIES OF EVAPOTRANSPIRATION F74 elevations above ground during seven periods of 3 days each at the Buckeye Project TABLE A4-1.-Windspeed, temperature, and vapor pressure at five [Windspeed, u in m s~', temperature, T in °C; vapor pressure, e in kPa; and elevations, z in m] Mean Day 3 Day 2 Day 1 Hrs 5-8 May 1966 H Pad < ON I I0 I~ < b- O - 00 09 NY

sh 09 in - 09 be sf b= i $9 ao op ap 0D O SD - 09 r CX 90 on t C in 10 tp & NP- N 1 09 O I~ 00 < O (D OY xf N iQ - I~ < +- \D N O) O - 5-8 9-12 13-16 17-20 21-24 1-4 3-6 June 1966 - id < N ~ bid < N I0 ¥ N in oo © OO N N m 0 0 0) N N OO ONO OOO AH HH rH 5-8 9-12 13-16 17-20 21-24 jJ-4 11-14 May 1967 H b- 10 CN H IIO ¥ N

q M M- q' do r T 1; A do T; do r O t T u L T- its do w do w' do z L Za do 26 do A M T* {or H LC T-'} B,, M T= C M L= D,; 1 E L E ML * T- £; L E# € E1 L F ML-*T- F,; do G HL- T- (or M T-) G, do H HL" T- (or M T-) H, do STUDIES OF EVAPOTRANSPIRATION Description e in the air above the surface (8) e at the water surface (8) saturation vapor pressure, See dd von Karman constant (16) atmospheric pressure (18) humidity of the air (3) deviation from mean humidity (4) evapotranspiration resis- tance (24) external r (24) stomatal r (25) correlation coefficient (fig. 43) time (34) horizontal windspeed (8) friction velocity (30) vertical speed of air movement (3) eddy component of w (4) vertical height or depth (12, (385) z in the air above the surface (23) roughness length (23) flux of sensible heat (13) the advective energy part in equation 22 (23) carbon dioxide content (15) damping depth (42) evaporation (1) evaporative flux (16) actual or measured evapo- transpiration (25) flu) (e, - ea) (21) evaporation from open water (8) flux (83) flux of carbon dioxide (15) heat flow into or out of the ground (17) heat flow at depth z (35) sum of energy input at the surface, except A and LE (22) net gain of radiation at the water surface (21) Symbol Dimensions Description K L° T-' eddy transfer coefficient K, do K for sensible heat (13) K.; do K for carbon dioxide (15) ;, do K for momentum (12) K,, do K for water vapor (13) L LC T> latent heat of vaporization (or H M-) (17) M L moisture in the atmos- phere (2) V -M L divergence or convergence of M (2) M, M L= moisture content of the soil by volume (38) M., M M- moisture content of the soil by weight (88) N O empirical constant (11 c) P L precipitation (1) P T. period (40) R, L2 T* - Richardson number (45), sta- bility ratio in equation 45 R,, M T* net radiation (17) (or M L-" T-!) RO L runoff (1) S L soil moisture (1) A S L change in S (1) SR O LA T* _ stability ratio (27) T 0 temperature (13) V M L- T-* _ water-vapor flux (14) W L precipitable water in the at- mosphere (2) AW L change in W (2) a angle phase angle (page 20) B O Bowen ratio (18) y M L-' T- O0~' psychrometer constant (19) A M L-' T-* 0~' slope of the saturation vapor ® Pa pa pressure versus the temper- ature curve (de,/d°C) (21) O water-air molecular ratio (23) 1" P" thermal diffusivity of soils (384) H L-' 0-* T-* Thermal conductivity of soils (or (384) ML density do density of air (3) do deviations from mean density (4) M L-' T* shearing stress or momentum transfer (12, 20) 1 10s I Water Use by Saltcedar and by Replacement Vegetation in the Pecos River Floodplain Between Acme and Artesia, New Mexico By EDWIN P. WEEKS, HAROLD L. WEAVER, GAYLON S. CAMPBELL, and BERT D. TANNER s LC- DI ES OF _E-V A P O TCR AN SP I R A [ 1 N U.S. GEOLOGICAL SURVEY PROFESSIONAL PAPER 491-G Prepared in cooperation with the U.S. Bureau of Reclamation UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1987 DEPARTMENT OF THE INTERIOR DONALD PAUL HODEL, Secretary U.S. GEOLOGICAL SURVEY Dallas L. Peck, Director Library of Congress Cataloging in Publication Data Water use by saltcedar and by replacement vegetation in the Pecos River floodplain between Acme and Artesia, New Mexico. (U.S. Geological Survey professional paper ; 491-G) Bibliography: p. Supt. of Docs. no.: I 19.16:491-G 1. Tamarix chinensis-Water requirements. 2. Floodplain flora-Pecos River Valley (N.M. and Tex.)-Water require- ments. 3. Evapotranspiration-Pecos River Valley (N.M. and Tex.) 4. Plant-water relationships-Pecos River Valley (N.M. and Tex.) I. Weeks, Edwin P., 1936- . II. United States. Bureau of Reclamation. III. Title. IV. Title: Replacement vegetation in the Pecos River floodplain between Acme and Artesia, New Mexico. V. Series. QE75.P9 no. 491-G _ 557.3 s _ 85-600364 [QK495.T35] [583'.158] For sale by the Books and Open-File Reports Section, U.S. Geological Survey, Federal Center, Box 25425, Denver, CO 80225 CONTENTS Page Page ADSIFACL 40s G1 | Instrumentation-Continued Introduction c6s 1 Other instrumentation --------------~-~-----~--~------- G13 PurpOSe and 8COPé 1 | Field measurements 14 Acknowledgments 3 Discussion 15 Historical background -------------------------------- 3 Energy-budget closure 17 PreviOUs WOTK. 4 | Estimates Of Water US@ 21 Consumptive use by saltcedar ------------------------- 4 Water use by saltcedar -:----..-...«--..s..««.«._c.._.-s<<» 22 Consumptive use by replacement vegetation ----------- 5 Water use by replacement vegetation ------------------ 24 Discussion of previous results ------------------------- 7 Comparison with potential evapotranspiration ---------- 24 'ThGOfy 8 Annual Water Use 27 Boundary-layer limitations ---«.--.--.-..-...-..-<..cl._.. 8 Comparison with previous results --------------------- 28 'The Bowen-ratio method ------:----.--¢--s-*«=%s.....~ 8 Estimates of evapotranspiration salvage --------------- 30 The eddy-correlation method -------~------------------ 9 | Summary and conclusions 31 Insiriumentalion sos 5s i ae 10 | Recommendations for future studies of evapotranspiration -- - 32 Sonic @N@MOMeter 11, | References 32 Lyman-alpha hygrometer--------------~--------------- 12 ILLUSTRATIONS Page FiGurE 1. Map showing location of stream gages and evapotranspiration measurement sites used in this study ---------------- G2 2, 3. Graphs showing: 2. Annual rates of bare-soil evaporation of ground water versus depth to water for various evapotranspirometer @XDEriIMeNtS 6 3. Relationship of evapotranspiration by saltgrass versus depth to water for various evapotranspirometer Studies 7 4. Diagrams showing the physical characteristics of the original sonic anemometer, the revised sonic anemometer, and the Lyman-alpha hygrOMeter 80808080 eee cess 11 5, 6. Graphs showing daily cycle of energy fluxes 5. Measured at the wet old-growth site on September 1, 1982 15 6. Measured at the grass and forbs site on June 26, 1982 15 7. Graphs showing summary of estimates of daily evapotranspiration by saltcedar by the energy-budget-residual method and the eddy-correlation MethOd 23 8. Sample logs showing the nature of the materials from land surface to the water table for each saltcedar MEASUr@MNt Site 88000808 eee ence ee mense eee eee 25 9-12. Graphs showing: 9. Summary of estimates of daily evapotranspiration by replacement vegetation by the energy-budget- residual method and the eddy-correlation MethOd 26 10. Estimated annual water use by saltcedar in the Pecos valley, New Mexico, as determined by the energy- budget-residual method and the eddy-correlation MethOd 27 11. Estimated annual water use by replacement vegetation in the Pecos valley, New Mexico, as determined by the energy-budget-residual Method see 28 12. Graph showing monthly precipitation at Roswell, N. Mex., during the period 1980-82 ----------------- ---> 29 TABLES Page TABLE 1. Locations and descriptions of the sites at which evapotranspiration measurements were made ------------------------ G14 2. Schedule of eddy-correlation measurements made during the Study 16 3. Summary of daily sums of energy-density measurements at the various sites for which eddy-correlation data were oa I E o e 18 4. Summary of daily water-use estimates at the Vv@rIOUS Sites 22 5. Estimates of the range in annual water use by saltcedar and by replacement vegetation in the Pecos River floodplain and of projected water salvage due to Saltcedar @r@dication 28 III Iv CONVERSION FACTORS CONVERSION FACTORS For readers who prefer to use inch-pound or other non-SI units, conversion factors for terms used in this report are listed below: Multiply By To obtain Area hectare (ha) 2.471 acre square meter (m*) 10.76 square foot (ft?) Energy Joule (J) 9.4787 10-4 British thermal unit (mean) (Btu) 0.2388 calories (cal) Energy/Area Megajoule/square meter (MJm~2) 88.05 Btu/ft? Energy/Area + Time watt/square meter (Wm~2) 5.2895x 10-3 Btu/ft? < minute 1.434x10~3 cal/centimeter? - minute (cal cm~* min-) Energy/Mass Joule/gram (Jg~!) 0.2390 cal/gram (cal g~') 0.4303 Btw/lbm Heat conductance watt/meter - Kelvin (Wm 9.629 10-3 Btu/ft min degree Fahrenheit 2.387x 10-3 cal/em + second - Kelvin Length nanometer (nm) 0.3937 10-4 mil micrometer (jm) 397% 10-1 mil millimeter (mm) 0.03937 inch (in) centimeter (cm) 0.3937 in meter (m) 3.281 foot (ft) kilometer (km) 0.621 mile (mi) Mass gram (g) 2.205x%10-3 pounds mass (Ibm) Mass flux 17.71 pounds mass/ft*/day Mass/Volume kilogram/meter (kgm~5) 10-2 pound mass/foot© 0.001 gram/cm} (gem gram/meter® (gm 6.24 10-5 pound 10-6 gem ~* Power watt (W) 3.412 Btwhour 1.340 10-3 horsepower 0.2388 calories/second (cal s~') Pressure kilopascal (kPa) 0.2953 inches of mercury (in Hg) 0.1450 pound/inch (psi) 10 millibar (mb) Temperature degrees Celsius (°C) 1.8°C+32 degrees Fahrenheit (°F) Kelvin (K) (K-273.15)1.8+32°F Velocity or Rate meter/second (ms ~*) 2.237 mile/hour (mph) Volume cubic meter (m?) 8.107 acre-foot (acre-ft) Volume/Area cubic meter per hectare (m*/ha) 3.281x10~4 acre-ft/acre STUDIES OF EVAPOTRANSPIRATION WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION IN THE PECOS RIVER FLOODPLAIN BETWEEN ACME AND ARTESIA, NEW MEXICO By EpwIN P. WEEKS, HAROLD L. WEAVER, GAYLON $. CAMPBELL,! and BERT D. TANNER ABSTRACT Water-use estimates for saltcedar and for replacement plant com- munities following root plowing in the Pecos River floodplain be- tween Acme and Artesia, New Mexico, were made by the eddy- correlation technique and a combined eddy-correlation energy-budget technique during 1980-82. Twenty-seven measure- ments of daily water use were obtained for various periods during the growing season-17 measurements from four saltcedar thickets and 10 from three stands of replacement vegetation. Large uncertainties exist in these estimates because of problems in extrapolating the data seasonally and areally, and because of discrepancies between the two methods. Nonetheless, the measurements indicate that an- nual water use by saltcedar probably is about 0.3 meters greater than that by the replacement vegetation. Such reductions in water use should have resulted in increased base flow of the Pecos River of 1.2-2.5%x 10" cubic meters per year (10,000 to 20,000 acre-feet per year). The fact that such gains have not been identified from stream- gage records may arise from masking of short-term gains by varia- tions in climate and in ground-water pumpage and from a continuing decline in the ground-water contribution to base flow from the shal- low aquifer. INTRODUCTION Saltcedar was cleared, initially (1967) by bulldozing and mowing and later (1974) by root plowing, from 8,700 hectares (21,500 acres) of the floodplain of the Pecos River in the reach between Acme and Artesia, N. Mex., in anticipation that substantial quantities of water would be salvaged from evapotranspiration. An unpublished analysis of base flow in the reach (G.E. Welder, U.S. Geological Survey (USGS), written com- mun., 1978) indicated that no readily identifiable gains in base flow occurred following either mowing or root 1Professor of Soils, Washington State University. 2Campbell Scientific, Inc. plowing, suggesting that the actual salvage of water by evapotranspiration was considerably less than antici- pated. PURPOSE AND SCOPE The current study was begun in 1979 to determine evapotranspiration from representative plots of saltcedar and of replacement vegetation, in order to determine whether initial estimates of consumptive use by saltcedar were erroneously high and (or) whether estimates of use by replacement vegetation were erroneously low to the extent that actual salvage from evapotranspiration was quite small. This report contains a preliminary evaluation of the evapotranspi- ration measurements and of the suggested magnitude of ground-water salvage from evapotranspiration. Measurements of evapotranspiration were made pe- riodically during the growing season in 1980-82 using an eddy-correlation approach. During 1980 and 1981, the measurements were made over three plots (fig. 1)- one with old-growth saltcedar, one with saltcedar re- growth following mowing, and one with replacement vegetation. In 1982, two other saltcedar sites and two other replacement vegetation sites were added to gain areal coverage. Weather data, including solar radiation, air temper- ature, relative humidity, precipitation, windspeed, and soil temperatures were collected on a continuous basis during the growing season of each year from 1980 to 1982 at the New Mexico State Wildlife Refuge at Arte- sia. Other data, including stomatal resistances, plant- water potentials, soil-moisture contents, root densities, moisture-characteristic data, and other plant- and soil- related parameters, also were collected, in order to model the evapotranspiration process. G1 G2 STUDIES OF EVAPOTRANSPIRATION 105° o 34° 1°14 103°30' EXPLANATION A Stream gage ha Evapotranspiration measuring site Study area (Mower and others, 1964) Wet old-growth saltcedar 0 G % Acme Acme gaging station Bitter Lake National Wildlife Refuge 7 Grass 0 Greenfield Hagermann 33° |- Lake Arthur _ C Mowed saltcedar Old-growth saltcedar Grass and forbs ArtesiaO Artesia gaging station Burned saltcedar ' Lake McMillan Study area C3 Lake Avalon Carlsbad ° 0 10 20 30 MILES | | | ] | I I I 0 10 20 30 KILOMETERS Base from U.S. Geological Survey State of New Mexico Scale 1:1,000,000 32° ve my umes ims as mmumure he ins sin as cnames us us emai» nie wee intents 'on o jus: mn names sio os Servier kp mss. uss ute FIGURE 1.-Location of stream gages and evapotranspiration measurement sites used in this study. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. ACKNOWLEDGMENTS This study was partially financed by the U.S. Bureau of Reclamation (USBR). In addition, USBR personnel, including Leon Marcell, Kirby Lykins, and Joseph Odell, obtained landowner permission for the studies, made backhoe excavations for soil sampling, and pro- vided materials and equipment. Personnel of the USGS New Mexico District office, including G.E. Welder, Paul Davis, and Doug McAda, provided logistical sup- port and assisted at various times. Tinco van Hylckama (USGS) provided assistance by identifying the various vegetation species at the replacement vegetation site. Other USGS personnel who assisted in data collection at various times include Rita Carmen, Dwight Hoxie, Eric Lappala, David Redinger, and David Stannard. Keith Bristow of Washington State University also provided advice and assistance in making plant-water status and stomatal-resistance measurements. HISTORICAL BACKGROUND The Roswell basin, which basically includes the study reach of the Pecos River, has been extensively developed for irrigation. The first irrigation develop- ment, which started in the 1890's, involved the use of flowing wells that tapped the artesian aquifer that underlies the basin. Development of the shallow allu- vial aquifer took place primarily during the period 1935-37. State law ended growth of irrigation develop- ment in both aquifers, and in 1967, flow meters were placed on all irrigation wells in the basin by order of the State Engineer's office. Since that time, pumpage from existing wells has been limited to a 1.1% 10* m per ha (8.5 acre-ft per acre) annual allocation, and no new wells have been permitted. These restrictions have resulted in stabilization of water levels in the artesian aquifer, but water levels in the shallow aquifer were still declining in 1975 (Welder, 1983, fig. 21). Saltcedar (Tamarix chinensis) was first noted in the basin in about 1912 (Eakin and Brown, 1939); it rapidly became the dominant species in the Pecos River floodplain, occupying 11,200 ha (28,000 acres) of the 16,400-ha (41,000-acre) floodplain in 1958 (Mower and others, 1964). Saltcedar mainly replaced two phreato- phytic grass species, saltgrass (Distichlis stricta) and alkali sacaton (Sporobolus airoides). Saltcedar has a substantial reputation as a plant that consumes large quantities of water (Robinson, 1965), and the observed declines in streamflow in the middle basin of the Pecos River (the reach from Alamagordo to Red Bluff Reser- voirs) was attributed in part to the saltcedar invasion (Thomas, 1963). Because of the concern that saltcedar could be a sub- stantial agent in the depletion of the flow of the Pecos River, the New Mexico State Engineer requested in G3 1956 that the USGS investigate the use of water by "nonbeneficial vegetation" in the 80-km (48-mi) reach of the river from Acme (now a ghost town) to Artesia, N. Mex. This reach was chosen for study because long- term streamflow records were available. In that study (Mower and others, 1964), saltcedar was found to cover about 11,200 ha (28,000 acres) of the floodplain within the study reach, although only 7,200 ha (18,000 acres) supported an areal saltcedar density of more than 10 percent. Also, on the basis of the extrapolation of water-use data for saltcedar developed by Gatewood and others (1950) for the Safford Valley, Mower and others (1964) estimated that the eradication of saltcedar in the reach would result in a reduction in consumptive use in the floodplain of 3.5% 10" m} (28,000 acre-ft) per year. Blaney (1961) made a separate estimate of water sal- vage by saltcedar eradication in the reach of the Pecos River from Acme to Artesia. He estimated annual sal- vage of about 2.5% 10" m* (20,000 acre-ft). In 1967, the U.S. Bureau of Reclamation began a program to eradicate saltcedar from about 8,700 ha (21,500 acres) in the Acme-Artesia reach. Initially, the trees were felled by bulldozers and burned. Thereafter, an attempt was made to control saltcedar regrowth by mowing. This technique was not successful, because the mown saltcedar regrowth grew back with great vigor. Consequently, a program of root plowing salt- cedar from the previously bulldozed and mowed areas was initiated in 1974, followed by a maintenance pro- gram. These clearing operations included nearly all the denser thickets of saltcedar except those in a 15-m-wide (50-ft-wide) strip along each bank of the river, which meanders for 132 river kilometers (82 mi) within the 80-km (48-mi) reach. Saltcedar thickets also were re- tained on the Artesia State Wildlife Refuge and on the Bitter Lakes National Wildlife Refuge. About 600 ha (1,500 acres) of dense to moderately dense saltcedar and about 2,000 ha (5,000 acres) of saltcedar with a density of less than 10 percent remain of the saltcedar plots mapped by Mower. In 1971, G.E. Welder, of the USGS New Mexico Dis- trict office, began a study of base flow of the Pecos River in the Acme-Artesia reach to determine whether the expected gains due to saltcedar eradication could be detected. His unpublished study indicated that the base-flow pickup in the reach declined from about 10" m (32,000 acre-ft) in 1957 to about 1.810" m* (15,000 acre-ft) in 1964 (3 years before initial clear- ing) but then varied between 2.0% 10" and 3.0% 10" m* (16,000 and 25,000 acre-ft) per year without a particu- lar time trend until the present time (1982). Lack of any apparent gain in base flow that could be attributed to saltcedar eradication, either following the period of mowing in 1967-69 or following root plowing in 1974- G4 76, raised concern that the amount of water that could be salvaged by saltcedar eradication had been substan- tially overestimated. This concern prompted the present study. In order to select a micrometeorological technique to be used in the study, grants were made to Lloyd Gay of the University of Arizona and to Leo Fritschen of the University of Washington to demonstrate such tech- niques in the summer of 1979. Gay (1980) made meas- urements necessary to apply the Bowen-ratio method, and Fritschen and others (1980) made measurements necessary to apply the eddy-correlation technique. Both methods are described in the section on "Theory." PREVIOUS WORK CONSUMPTIVE USE BY SALTCEDAR Perhaps the earliest measurements of water use by saltcedar were from tank experiments performed in 1940 at Carlsbad, N. Mex., by Blaney and others (1942). They reported water use of 1.67 m/yr (meters per year) from a tank with a 0.6-m depth to water and of 1.43 m/yr for a tank with a 1.2-m depth to water. An extensive study of water use by saltcedar was made in the floodplain of the Gila River in the Safford Valley in 1943-44 by Gatewood and others (1950). They measured evapotranspiration by saltcedar in nine tanks, two of which were 3.05 m in diameter and seven of which were 1.8 m in diameter, with a range in depth to water of 1.2 to 2.1 m. However, these tanks were located in a natural clearing (Gatewood and others, 1950, p. 37) and were undoubtedly subject to "oasis effects" (explained in "Discussion of Previous Re- sults"). Water use ranged from about 1.2 m/yr to about 3 m/yr. Gatewood and others also introduced the con- cept that consumptive use of water by saltcedar is pro- portional to its volume density. Based on this concept, their consumptive use adjusted for 100-percent volume density ranged from about 2.1 to 3 m/yr, excluding pre- cipitation. Gatewood and others (1950) also measured water use from 3,765 ha (9,300 acres) of the Gila River floodplain by a water-budget (inflow-outflow) method. These measurements indicated an annual water use of about 0.73 m over that area, about half of which had a cover of saltcedar and the remaining half a cover of seepwil- low (Baccharis), cottonwood (Populus fremontis), mesquite (Prosopis velutina), undifferentiated brush, and bare ground (predominantly sandbars in the river channel). The U.S. Bureau of Reclamation (1973) measured evapotranspiration by saltcedar at a site in the flood- plain of the Rio Grande near Bernardo, N. Mex. Meas- STUDIES OF EVAPOTRANSPIRATION urements were made in six large tanks 93 m* in area by 3.7 m deep that were planted with saltcedar in 1961 and 1962. The tanks were operated as constant-water- level evapotranspirometers from 1962 to 1968. In 1968, saltcedar in two tanks was replaced by Russian olive (Elaeagnus augustifolia). Salt was added at a rate of about 5,000 mg/L to the feed water for two other tanks. The water level in the remaining two tanks was dropped from about 0.90 m to about 2.75 m below land surface by cutting off the feed water or by pumping water from the tank. Water use by saltcedar in the various tanks showed substantial variation, ranging from 0.7 m to 1.4 m, including rainfall. Average water use during the period of operation with fresh water and constant water levels was 1.0 m (including precipita- tion), with water levels maintained 0.9 to 1.5 m below land surface. Lowering the water table from 0.9 to 2.7 m initially reduced annual consumptive use by 0.30-0.45 m, but usage nearly recovered in about 2 yr. Hughes (1972) analyzed the U.S. Bureau of Reclama- tion (1973) data to show that evapotranspiration by saltcedar was not linearly related to volume density, but instead was almost as large for 50-percent volume density as for 100-percent volume density. He also demonstrated that water use by saltcedar was less than potential evapotranspiration as computed by the Pen- man (1956) equation, unless corrected downward by use of an empirical stomatal-resistance term. Van Hylckama (1974) measured water use by saltcedar during the period 1962-67 in the floodplain of the Gila River near Buckeye, Ariz., using six large (9 mx9 mx4.25 m deep) tanks located within a salt- cedar thicket and two smaller (6 mx6 mx2 m deep) tanks within a large bare area. In 1965 (the year dur- ing which vegetative cover and lysimeter operations were most representative of the natural environment), water use by saltcedar in three large tanks using fresh water ranged from 2.6 m/yr, with a depth to water of 2.7 m, to 3.4 m/yr, with a depth to water of 1.5 m. Water use in one small "oasis" tank previously flushed with fresh water was 3.6 m/yr, with the water table main- tained at a depth of 1.2 m. Van Hylckama also per- formed experiments that indicated that a decrease in volume density of saltcedar from 100 percent to 50 per- cent reduced consumptive use by only about 10 percent. Gay and Fritschen (1979) determined evapotranspi- ration from a saltcedar thicket at the Bernardo, N. Mex. lysimeter site during a 5-day (d) period in June 1977 using an energy-balance (Bowen-ratio) technique. They measured rates of evapotranspiration of 7.4 and 9.0 mm/d (millimeters per day) at two locations about 75 m apart, which compared favorably with measure- ments in four lysimeters that varied from 6.4 to 9.2 mm/d during the same period. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. Fritschen and others (1980) measured evapotranspi- ration from two thickets of saltcedar in the Lake McMillan delta area of the Pecos River near Artesia, N. Mex., by the eddy-correlation method. One thicket constituted regrowth from a fairly recent burn, and the other constituted old-growth saltcedar. Water use dur- ing an 8.5-h (hour) period in July was 2.3 mm for the regrowth and 1.9 mm for the old growth. The day was cloudy, and hence the results probably are not repre- sentative of water use by saltcedar. Gay (1980) meas- ured water use of 1.7 mm during an 8-h period on the same day by the same regrowth plot measured by Fritschen and others (1980), using a Bowen-ratio tech- nique. Van Hylckama (1980) made micrometeorological measurements at the Buckeye site in Arizona. The measurements were made during 3-d periods about one month apart from May to September in 1966 and in March and May 1967. The measurements included dry- bulb and wet-bulb temperature measurements and windspeed at five levels, net radiation, and soil-heat flux. These measurements should be adequate to esti- mate evapotranspiration by the energy-budget (Bowen-ratio) method. However, Bowen ratios com- puted using various pairs of wet- and dry-bulb temper- ature sensors indicated significant scatter in the com- puted Bowen ratio. These variations, and the fact that bias among the various sensors was not eliminated, as is generally recommended for Bowen-ratio studies (Suomi and Tanner, 1958, for example), caused van Hylckama to reject the energy-budget approach. Our examination of van Hylckama's data suggests that the results are better than he allowed, particularly when examined in the context of the results of the study by Culler and others (1982) and of our study. For example, a cursory examination of the hourly average temperature and humidity data at various heights above land surface tabulated by van Hylckama (1980, p. F62-F69) indicate that, for the measurements at 4-5-m and 5-7-m heights, the Bowen ratio is typically near zero during the hours of peak radiation. Based on the net-radiation and soil-heat-flux measurements made by van Hylckama (1980, p. F70-F 73), evapotran- spiration would have been about 8 mm/d, compared with values of as much as 20 mm/d measured from van Hylckama's tanks. The Bowen-ratio estimates are in relatively good agreement with ours, and with those of Culler and others (1982) described below, whereas the evapotranspiration-tank data are not. Leppanen (1981) applied the Bowen-ratio method to the measurement of evapotranspiration from rapidly growing young saltcedar near the upstream end of San Carlos Reservoir on the Gila River in Arizona. His measurements covered a 48-d period extending from G5 August 17 to October 3, 1971. Evapotranspiration av- eraged 5.8 mm/d during this period, and 7.0 mm/d dur- ing the last 15 days of August. Depth to water at the site varied greatly because of reservoir-level fluctua- tions, but soil water content was high throughout the period of measurement. Culler and others (1982) measured evapotranspira- tion by the water-budget method from a floodplain area of 2,190 ha along the Gila River between the town of Bylas and the backwater of San Carlos Reservoir in Arizona. The dominant riparian vegetation is saltcedar (1,540 ha) and mesquite (600 ha), with about 40 ha mapped as not having phreatophyte cover. Before clearing, consumptive use (excluding precipitation) from this area averaged about 0.83 m annually. Water use from reach 3, which had a cover of about 97 percent saltcedar, was about 1.0 m/yr. CONSUMPTIVE USE BY REPLACEMENT VEGETATION Water use by the various plant communities that have replaced saltcedar in the Pecos River floodplain has been less intensively studied than water use by saltcedar. Nonetheless, some studies are available. The most comparable data were collected by Culler and others (1982). They measured evapotranspiration from vegetation that replaced saltcedar following root plowing in the floodplain of the Gila River. The saltcedar in the floodplains of some reaches was cleared by root plowing in 1967, and other reaches were cleared during 1968-71. Water-use data on replacement vege- tation were collected from 1968 to 1971. These meas- urements indicate that evapotranspiration from the patches of bermuda grass, annual weeds, and bare ground that replaced the saltcedar averaged 0.64 m/yr, including 0.28 m of annual precipitation. Patches of bare soil remain in some of the cleared areas of the Pecos River floodplain, and a possible water-management scheme is to maintain the soil bare over large areas of the floodplain. Thus, an estimate of bare-soil evaporation would be desirable. Evaporation from bare soil underlain by a shallow water table has been studied extensively, using theory, laboratory columns, and field evapotranspirometer tanks. Only the evaporation from bare soil measure- ments derived from evapotranspirometer studies will be reviewed here. Blaney and others (1930) and Blaney (1933) determined evaporation at Santa Ana, Calif., from nine 0.60-m-diameter tanks, three containing repacked Hanford fine sandy loam and six containing "undisturbed" soil of the same composition. Results of their studies, as summarized by Blaney (1933), in 150 STUDIES OF EVAPOTRkANSPIRATION 120 90 60g DEPTH TO WATER, IN CENTIMETERS fine sandy loam I I I I Blaney (1933) Undisturbed McDonald and ~ Hughes (1968) Sandy Silt < -~ ~- A I I | van HyIckama (1974) X«~ Coarse-textured alluvium Gatewood E] E? [F ®\ and others .S. Bureau of Blaney (1933) (1950) Reclamation (1973) Disturbed Gravel Sandy loam fine sandy loam Robinson (1970) U.S. Bureau of 290 (~ Sandy silt Reclamation (1973) [*] - Sand R | | | 1. ear- | | | o 10 20 30 40 | so 60 70 so 90 CONSUMPTIVE USE, IN CENTIMETERS PER YEAR FIGURE 2.-Annual rates of bare-soil evaporation of ground water versus depth to water for various evapo- transpirometer experiments. terms of annual water use versus depth to water, are shown in figure 2. Gatewood and others (1950) con- ducted experiments on 12 tanks, each 1.2 m in diame- ter, to determine evaporation from bare soil in the Saf- ford Valley near Glendale, Ariz. Six of the tanks were filled with sand and gravel, two with gravel, and four with clay loam. Results for those tanks for which full- year data are available are shown in figure 2. Large (3x3 to 12x12 m) plastic-lined tanks have been used near Yuma, Ariz. (McDonald and Hughes, 1968), near Buckeye, Ariz. (van Hylckama, 1974), near Win- nemucca, Nev. (Robinson, 1970), and near Bernardo, N. Mex. (U.S. Bureau of Reclamation, 1973), to meas- ure evaporation from bare soil underlain by a constant shallow water table. The tanks near Yuma and Win- nemucca were filled with sandy silt, while those near Buckeye contained a coarse-textured alluvial soil but with more clay and silt than the Yuma soil (van Hyl- ckama, 1974). Three tanks at Bernardo were filled with sand and silt with clay lenses of various thickness, and one was filled with sand. The results are summarized in figure 2. Veihmeyer and Brooks (1954) measured evaporation from 1935 to 1939 from bare soil near Davis, Calif., in 38 tanks-13 tanks 0.64 m in diameter filled with Yolo fine sandy loam and 25 tanks 0.7 m in diameter filled with Yolo silt loam. They show an aver- age monthly water use in August of about 2.5 cm. The Pecos River has developed a natural levee, and the depth to water landward of the levee is about 1.5 to 3 m, with an average depth to water of about 2 m. The materials above the water table are typically sandy loam to clay loam in texture. None of the tank experi- ments involved depth to water greater than 1.2 m, but the data shown in figure 2 suggest that evaporation discharge of ground water should be no more than about 0.1-0.2 m/yr from bare soil in the Pecos River floodplain. Young and Blaney (1942) cite measurements of water use by various species of weeds grown in tanks having a shallow water table. Of the weeds grown, only Russian thistle (Salsola kali) was common to the Pecos River floodplain. The Russian thistle used about 0.60- 0.65 m of water during the period May 3 to Septem- ber 17, 1932, with a depth to water of 0.30-0.45 m. Blaney and Hanson (1965) quote water use for alkali sacaton, a species prevalent in the Pecos River flood- WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. 189, I I I I I + ISLETA, N. MEX. @ BERNARDO, N. MEX. (U.S. Bureau __| of Reclamation, 1973) [+] ESCALANTE, UTAH s A MESSILA, N. MEX. * OWENS VALLEY, CALIF. X SANTA ANA, CALIF. ® LOS GRIEGOS, N. MEX. - YOUNG AND BLANEY INTERPRETATION 1.50 [- DEPTH TO WATER, IN CENTIMETERS o to [+] I 0.30 |-- 0 | | | I | 0 0.30 0.60 0.90 1.20 CONSUMPTIVE USE, IN CENTIMETERS PER YEAR FIGURE 3.-Relationship of evapotranspiration by saltgrass versus depth to water for various evapotranspirometer studies (modified from Young and Blaney, 1942). plain, that was grown in tanks at Carlsbad, N. Mex. Annual water use by sacaton was 1.2 m with a 0.60-m depth to water, and 1.05 m with a 1.2-m depth to water. Average daily use by sacaton in the tank with the 1.2- m depth to water was 5.6 mm in June and 6.6 mm in August 1940. Evapotranspiration by saltgrass has also been meas- ured at several locations using tanks. Results of most of these studies have been summarized by Young and Blaney (1942). In addition, the U.S. Bureau of Recla- mation (1973) has measured water use by saltgrass at Bernardo, N. Mex. The results (fig. 3) show substantial scatter, and annual consumptive use ranged from about 0.25 m to 1.20 m. However, for water-table depths of a meter or more, the range in water use by saltgrass was 0.3 to 0.8 m, with climate a major factor. Although the authors observed little saltgrass in the study area, Mower and others (1964) report that salt- grass was the dominant species in about half of the Pecos River floodplain prior to the saltcedar infesta- tion. Hence, saltgrass may again become prevalent in the floodplain as a regrowth species as time progresses. Fritschen and others (1980) measured water use by kochia (Kochia scoparia) at the forbs site described be- low in July 1979, by an eddy-correlation technique. (The listed common name for Kochia scoparia is GT summer-cypress, but the weed is locally known as "kochia" in the Roswell basin area and kochia will be used as its common name in this report.) Gay (1980) made simultaneous measurements of water use by the same stand at an adjacent site using a Bowen-ratio technique. Both methods indicated water use of about 3 mm/d. DISCUSSION OF PREVIOUS RESULTS Most rates of evapotranspiration from saltcedar have been made in the arid Southwest in areas having rela- tively similar climates. Nonetheless, the measured val- ues vary by a factor of about six, with summertime latent heat flux, defined as the evapotranspiration rate, E, multiplied by the latent heat of vaporization of water, \, or AE, ranging from about 0.7 times to more than 4 times the energy available from solar and ther- mal radiation, R,,. Conditions that produce values of AE/R,, greater than unity are commonly called "oasis effects" and indicate a locally greater supply of water compared with closely adjacent areas. The most ex- treme values of AE/R,, have been obtained from evapo- transpirometer tanks, suggesting that these tanks were strongly influenced by oasis effects. The tanks used by Gatewood and others (1950) were located at the edge of a saltcedar thicket, and the potential for an oasis effect at that site is apparent. The six large tanks used by van Hylckama (1974) were located within a saltcedar thicket, however, and the occurrence of very localized oasis conditions is less easily inferred. The temperature profile data collected by van Hylckama (1980) indicate that oasis conditions did not generally prevail for the thicket as a whole, so that local oasis effects must have occurred within the vegetation in the tanks. Possibly the shallow water table in the tanks and the greater vigor of the relatively newly planted saltcedar combined to allow those saltcedar to use water more freely and to thus create oases within the thicket. In fact, the saltcedar outside the tanks appears to have been deprived of a significant source of water by a dike constructed in November 1964 to divert floods in Waterman Wash from the area. Following construc- tion of the dike, saltcedar in the vicinity of the tanks began to die during periods of prolonged drought (van Hylckama, 1974, p. E29), indicating that the saltcedar underwent substantial moisture stress when deprived of soil moisture. The occurrence of localized oases is also suggested by the work of Gay and Fritschen (1979). One of their Bowen-ratio towers demonstrated the occurrence of a significant oasis effect, while the other, 75 m distant, did not. However, the saltcedar cover in that study area reportedly is relatively homogeneous. Thus the indi- cated oasis effect is not readily explainable. G8 Results of the other cited studies, including the tank studies at Bernardo (U.S. Bureau of Reclamation, 1973) as well as our own results, indicate that water use by saltcedar is approximately equal to or less than net radiation, suggesting that oasis conditions gener- ally do not prevail over large saltcedar thickets or within the floodplain environment as a whole. Thus, extrapolation of the tank experiment data on water use to natural floodplain environments may lead to erro- neous conclusions concerning water use in the natural environment. Historically, tank-derived evapotranspiration values have been extrapolated using the volume-density tech- nique developed by Gatewood and others (1950). This method relies on the assumption that evapotranspira- tion is proportional to the volume density of plant fo- liage, an assumption that probably is valid when strongly oasitic conditions prevail. However, when such conditions do not prevail, water use may be rela- tively constant for a variety of situations, from full areal vegetative cover down to some fraction of full cover, as noted, for example, by Hughes (1972, fig. 1) and by van Hylckama (1974, p. E13-E16). Culler and others (1982) developed an empirical rela- tionship indicating that evapotranspiration from plots with less than 100-percent volume density was propor- tional to the fractional volume density raised to the 0.75 power. This relationship is more linear than that determined by Hughes (1972) or van Hylckama (1974). However, the average volume densities of their reaches 1, 2, 2a, and 3 were 0.41, 0.54, 0.43, and 0.81, respec- tively (Culler and others, 1982, table 10). Data from reach 3, the only one showing substantially different cover, were obtained for only two periods in the sum- mer of 1965. During those periods, water use per unit area from this reach was about the same as that from reach 2 but was significantly larger than that from reach 1. It is possible that the differences in water use among reaches were due to saltcedar vigor or other factors rather than to volume density. The above data suggest that, under conditions of varying cover, net radiation will be the main energy source for evapotranspiration, with local advection of energy from areas between plants increasing evapo- transpiration by individual plants in less dense stands. Thus, the volume-density approach may result in an underestimate of evapotranspiration under conditions of less than full cover. However, estimates of floodplain water use based on too-high water-use rates with over- compensating plant-density-effect estimates may re- sult in overall projections that are nearly correct for an entire floodplain area. Nonetheless, grossly erroneous estimates would result from the volume-density tech- nique for individual thickets, and use of these values to STUDIES OF EVAPOTRANSPIRATION estimate total water consumption resulting from in- creases in saltcedar density in recently colonized areas would also be overemphasized. THEORY Several micrometeorological techniques have been developed to determine evapotranspiration from mete- orological measurements made in the turbulent boundary layer above a plant canopy or other surface. Of these, the Bowen-ratio technique has been widely applied to measure evapotranspiration from floodplain vegetation in previous studies, and the eddy- correlation technique was applied in this study. The theory of both techniques is briefly described below, following a discussion of boundary-layer limitations that are applicable to both methods. BOUNDARY-LAYER LIMITATIONS Application of both the Bowen-ratio and eddy- correlation methods is based on the assumption that horizontal gradients in the vertical fluxes of heat, vapor, and momentum are zero. However, as air moves horizontally from a surface of a given roughness, wet- ness, and temperature to another surface having sig- nificantly different properties, the air-velocity, vapor- density, and air-temperature profiles (and the associated fluxes) must be in equilibrium with the new surface. This adjustment is completed within a boundary layer over the new surface that grows in thickness with distance from the upwind edge of the surface. A rule of thumb, based on review of wind- tunnel data and on various theoretical studies, indi- cates that the micrometeorological measurements should be made at a height above the canopy of no more than 0.01 times the "fetch," or distance from the up- wind edge of the canopy (Tanner, 1967, p. 547). Local inhomogeneities in the canopy surface also may cause horizontal gradients in the vertical fluxes. Effects of such local inhomogeneities can be minimized by increasing the height of the instruments above the canopy. However, the use of this approach is limited by the above requirement that the instruments be at a height no greater than 0.01 times the fetch. THE BOWEN-RATIO METHOD The Bowen-ratio method (Tanner, 1967) relies on the measurement of the energy budget (net radiation and soil-heat flux) and of temperature and vapor density or vapor pressure at two or more heights above an evapo- transpiring surface. Heat and vapor are assumed to be transported by an eddy-diffusion process, and the eddy diffusivities are assumed to be equal for heat and WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. vapor. Under these assumptions, the ratio of sensible to latent heat flux is given by the Bowen ratio: B=H/)\E=-y-AK§ § (1) where B=Bowen ratio, dimensionless; H=sensible heat flux, Wm~*; A=latent heat of vaporization, Jg~'; E =evapotranspiration rate or vapor flux, gm y=psychrometer constant, gm 5°C; AT =difference in temperature at two vertically dis- placed measuring points; °C; and Ap=vapor-density difference measured across the same vertical interval, gm -*. Evapotranspiration is then computed by the equa- tion E=(R,-@OM1+B) , (2) where R,, =net radiation, Wm~*; and G =soil-heat flux, Wm-*. Both R,, and G are readily measured using commer- cially available instrumentation, as described by Fritschen and Gay (1979). This method also requires the measurement of temperature and of vapor pressure or vapor density at two heights above the canopy or exchange surface. Because of the instrument height-to- fetch requirement and the need to make measurements far enough above the exchange surface to avoid the effects of surface heterogeneities, the temperature and vapor sensors in the Bowen-ratio setup must be quite closely spaced, commonly about 1 m apart. Under these conditions, the temperature and vapor differences are quite small, and slight biases in the temperature and humidity sensors, easily tolerated for most point meas- urements, can cause a significant error in the meas- ured Bowen ratio. One common approach to avoiding this bias is to switch sensor positions periodically, so that each sensor is in the upper position half the time and in the lower position the other half. Bowen ratios are calculated for intervals including a full sequencing of sensor positions. This method was used by Gay (1980) and by Gay and Fritschen (1979). Switching sen- sor positions is a practical and proven method for avoid- ing bias-type errors in sensors if data are collected at only two heights. However, the Bowen-ratio technique can be made more reliable by collecting data at three or more heights. A ploy used in these circumstances to avoid bias in sensor readings is to sequentially bring air from different levels past the same sensor and then to average several readings to avoid the effects of not G9 having simultaneous readings. Leppanen (1981) used this approach. THE EDDY-CORRELATION METHOD The eddy-correlation method for measuring the flux of heat and water vapor from a surface is based on the concept (van Hylckama, 1980, p. F13) that a moving parcel of air carries with it, on its journey up, down, or sideways, the heat and water vapor that it contained at the start. Thus, if heat and water vapor are being car- ried upward in the atmosphere from a warm, moist surface or vegetation canopy, the updrafts will be warmer and wetter than the downdrafts. Hence, if the vertical wind vector, air temperature, and water-vapor density can be measured at high frequency within the turbulent boundary layer, both sensible and latent heat fluxes can be measured directly (Priestley, 1959). The vertical wind in turbulent eddies that transport the heat and vapor may be separated into two compo- nents, a mean velocity, w, and a fluctuating instanta- neous deviation from the mean velocity, w', such that w=w +w' , where w is vertical windspeed, ms~!. Similarly, temperature, T, and water vapor density, p, can be separated into mean and fluctuating compo- nents: T=T +T' , and p=p+p' , where T =temperature, °C; and p=vapor density, gm 5. In this study, air temperature was measured with a fine-wire thermocouple as the difference between the junction temperature and that of a reference junction in contact with a large thermal mass. Thus, the tem- perature equation becomes T=T ~Te¥T' . where T, is the reference-junction temperature. Assuming that the heat and vapor fluxes are indeed due mainly to convection by the turbulent eddies, as opposed to thermal conduction or molecular diffusion, and that fluctuations of air density with temperature are small, the heat-flux equation can be written H=Cpp,Tw-Tw+T 'w+wT -Tw' T') . (3) G10 Products of a mean and a fluctuating quantity become nearly zero when averaged over an appropriate time interval, and the mean vertical wind is assumed equal to zero. Hence, the equation reduces to H=Cpp,w'T' , (4) where H=sensible heat flux, Wm~*; p., -mean density of moist air, gm *; Cp =heat capacity of air at constant pressure, dge °C ': and the overbar signifies the time average of the quan- tity. Webb and others (1980) show that the assumption that p,, is constant is adequate in computing sensible heat flux. By a similar process, the vapor flux can be shown to be given by the equation (Webb and others, 1980) E=(1+M)w'p' , (5) where E =evapotranspiration, gm~*s~'; and M =ratio of moles of water vapor to moles of dry air in a unit volume of moist air, dimensionless. However, Webb and others (1980) show that, for vapor- flux measurements, the effects of temperature on air density cannot be ignored and the vapor flux must be corrected by using the equation = (1+M)[w'p+@/T) w'T'l , (6) where T.=temperature, °K. This correction term can add up to 10 percent of the vapor flux when sensible heat flux is large. The vapor flux can be converted to a latent heat flux by multiply- ing the evapotranspiration rate by the latent heat of vaporization of water, \, where ) is in Jg~' The eddy-correlation fluxes represent one measure of the energy budget: H+AE=R,,-G-AsS -P (7) where R,, =net radiation, Wm; G =soil-heat flux, Wm; AS =rate change in heat storage in the vegetation canopy during the time of interest, Wm ~*; and P =energy flux used in photosynthesis, Wm~* In general, AS and P are only a few percent of the energy budget and are ignored. Hence, HA+AE*~R;=G . (8) STUDIES OF EVAPOTRANSPIRATION Measurements of R,, and G thus provide a means of checking fluxes measured by eddy correlation. In practice, vertical wind, temperature, and vapor density are sampled a finite number of times during a given sampling period, and the sensible heat flux is computed as N - N 2 X PHN" I : "to) {2 where N is the total number of samples taken during the sampling period. Likewise, the latent heat flux, L, is computed as N L=A(1+M)| Y, wip;/N {=1 x - N ; -S m |. A10) i=1 i=1 Pa INSTRUMENTATION The instrumentation used to measure eddy- correlation fluxes included CSI (Campbell Scientific, Incorporated) CA 27-T sonic anemometers equipped with fine-wire thermocouples, CSI Lyman-alpha hy- grometers, Fritschen (1965) net radiometers, home- made soil-heat flux plates (Tanner, 1963), and thermo- couples for measuring soil temperature. Mean temperatures and mean vapor densities were measured using a Delta T aspirated wet-bulb dry-bulb psychome- ter, a model 201 Physchem probe that measures tem- perature with a thermistor and relative humidity with a polystyrene resistance chip, and (or) a manually read Assman wet-bulb dry-bulb psychrometer. Windspeed and direction were measured using a Met-One anemometer and a Met-One wind-direction indicator. The eddy-correlation data were obtained, processed, and recorded on a CSI CR-5 data logger that contained special hardware for 10-Hz sampling and special soft- ware to output sensible and latent heat fluxes. The net radiometers, soil-heat flux plates, and soil-temperature thermocouples were read by integrator modules in the CR-5 data logger, and other instrument responses were obtained using a CSI CR-21 data logger. 3The use of trade names in this report is for identification purposes only and does not imply endorsement by the U.S. Geological Survey. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. G11 Sonic transponder I I C ] Source | tube 10 emf—4 cm Fine-wire [3 0.8 cm (typical) thermocouple 1 Detector tube C-] Electronics case A Level bubble Electronics case j C Fine-wire thermocouple ; Socket \/ --- 5—4cmq‘ 10 cm I B FIGURE 4.-Physical characteristics of (A) the original sonic anemometer, (B) the revised sonic anemometer, and (C) the Lyman-alpha hygrometer. SONIC ANEMOMETER The sonic anemometers used in this study determine windspeed by measuring the phase shift between a transponder and a receiver resulting from the upward or downward displacement of the emitted sound waves as they are transported by the vertical wind. This model is an improved version of that described in detail by Campbell and Unsworth (1979). The original model was updated in 1980 to increase its operational temper- ature range and sensitivity and again in 1982 to im- prove its cosine response. The transducers in this sonic anemometer (figs. 4A, 4B) have 10-cm separation. Hence, the velocity of eddies smaller than 10 cm in size will be underestimated owing to space-averaging. The CA-27T sonic anemometer is factory calibrated to gen- erate 1.00 V (volt) in a 1-ms~' wind, and in the anemometer-data logger system resolution is 0.005 ms~*. A wind-tunnel calibration in 1982 indicated that the factory calibration is accurate to about +7 percent for the units used in this study. High-frequency temperature fluctuations can also be measured with a sonic anemometer. However, the Campbell-Unsworth design did not function well as a G12 thermometer, and, instead, the CA-27T anemometer has a built-in 12.5-wm chromel-constantan thermocou- ple mounted 4 cm laterally from the center of the sonic path. This thermocouple has a time constant of about 0.01 s. The reference junction for the thermocouple is located in the metallic mount for the anemometer and has a time constant of about 20 min. The thermocouple signal is amplified to provide an output of 0.25 V/C. Output from the sonic anemometer, V,, and from the fine-wire thermocouple, Vp, is in volts. The fluctuating component of the vertical wind, w', is w -w, so wi=C; (Vy -V;) ; (11) where w/=the fluctuating component of wind for sample i, ms ~'; C, =sonic anemometer calibration, ms ~- 'V ~'; V,,; =sonic anemometer volt reading for sample i, V; and Vs =mean sonic anemometer reading for sampling N period, 3: V,,.ANV. L: Likewise, Ti=Cp (Vi;-V¥1) , (12) where T/=fluctuating temperature component, °C; Cp=thermocouple signal calibration, °C V-; V;, =-amplified thermocouple output for sample i, V; and V;=mean amplified thermocouple output, V. For the instruments used here, C,=1 ms~'V~' and V~-'. Substituting these expressions into the eddy-correlation equation (eq. 4) for sensible heat, N H=CrpiC.Cr | >, Va¥rIN i=1 N N > 3. Vs J. ) . (18) [<1 . {31 LYMAN-ALPHA HYGROMETER High-frequency humidity measurements were ob- tained with a Lyman-alpha hygrometer (fig. 4C), which measures the attenuation by water vapor of ul- STUDIES OF EVAPOTRANSPIRATION traviolet radiation having a wavelength of about 121.56 nm (the Lyman-alpha spectral line). Water va- por has a conveniently high attenuation coefficient at this wavelength, while the coefficient for oxygen is about an order of magnitude smaller. Radiation at this wavelength is generated by a hydrogen-glow discharge lamp and is detected by a nitric oxide ion chamber (Buck, 1976). Magnesium fluoride windows set the lower wavelength limit of detected radiation at about 115 nm. The ion chamber itself imposes the upper wavelength limit at about 135 nm. Buck (1976) has poineered the development of Lyman-alpha hygrome- ters and provides further description of their design and operation. For monochromatic radiation, the relationship be- tween transmitted radiation and water-vapor density is provided by a form of Beer's law (Campbell, 1977, p. 48): V;z;=V, exp (-£kpx) (14) where V;;=hygrometer voltage, which is proportional to the transmitted light intensity; V,=voltage proportional to the emitted light inten- sity; Ak =attenuation coefficient, m*g~'; p=water-vapor density, gm ~*; and x=path length, m. The hydrogen-ion glow tubes in the Lyman-alpha hygrometers emit a low-level radiation continuum in addition to the spectral line at 121.6 nm, which causes departures from the simple Beer's law attenuation. These departures are satisfactorily accounted for by assuming a modified form of Beer's law that includes a vapor-density-dependent attenuation coefficient. The attenuation coefficient was found by laboratory cali- bration assuming kx =a +bp (15) so that V1; =V, exp [-(a +bp)p] . (16) The values for a among the different instruments, as determined by calibration in an environmental cham- ber or by pumping air of known vapor density through the light path, ranged from 0.200 to 0.235 m*g~'; b ranged from -0.0027 to -0.0043 m'g~* for different hygrometers but did not shift greatly for a given hy- grometer with time. Noting that p{=p;-p, substituting p; and p into equa- tion 16, taking logarithms of both sides, and subtract- ing, yields WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. Vim e is In=-=-(a+bp)(p;-p) ; (17) H where N 171.1: E VHt/N i=1 If fluctuations of V;, about EH are small, In may be approximated by V;;,/V;, and equation 17 can be rearranged V 'u pus ini Na - (18) -Vy(a +2bp) Recalling that w '=C,(V,;- V,), a-, C; sr agre wipi = (VsiVHi_VsVH) s (19) -V;y4(a +2bp) Substitution of equation 16 into the eddy-correlation equation for evapotranspiration and multiplying by A to obtain latent heat flux yields N M1+M)C psf MCLS I[ Ev.vaw -Vy(a+2bp) N ., N - SVs S Vi +(x5/f)fl7 (20) The voltage output from the sonic anemometer, the fine-wire thermocouple, and the Lyman-alpha hygrom- eter are integrated over 0.1-s intervals, and the inte- grated voltages for alternate 0.1-s intervals are accu- mulated in the microprocessor to produce means and covariances. The following block averages over 2.5-min (before June 1981) or 5.0-min (during and after June 1981) intervals were accumulated: N N N ZVsi/N’ E V&V r;/N , sziVHi/N > N SV Y, V4 , and YV Y, Vs >, ¥milN G13 These were stored by the microprocessor, accumulated, and averaged for a longer (typically 30-min) user- selected interval in the data logger. At the end of the interval, the longer term averages were printed on paper tape and recorded on magnetic cassette tape. Computation of sensible heat flux from the recorded data-logger counts requires that the counts be multi- plied by the density of air, which is a function of air temperature and of vapor density. However, the actual air temperature is not provided by the thermocouple output. Also, latent heat flux computations need to be corrected using the actual mean vapor density. The Lyman-alpha hygrometers show substantial slow diur- nal drift about a persistent mean, the drift mainly caused by variations in the intensity of the emitted Lyman-alpha radiation. Consequently, the hygrometer provides a relatively reliable measure of water-vapor- density fluctuations but not of absolute vapor density. Because of these factors, temperature and water-vapor densities were measured independently using the in- strumentation described above to provide 30-min aver- ages of air temperature and vapor density simulta- neously with the eddy-correlation data on other cassette tapes. The 30-min average values of the eddy-correlation sensor outputs of temperature and vapor density, and of net radiation and soil heat flux, were read from the cassette tapes into a mainframe or microcomputer, where the appropriate multipliers were computed to provide sensible and latent heat flux in Wm~". The data were also used to compute the energy-budget clo- sure, (H+A\E)/(R,,-G) . (21) The energy-budget closure provides a measure of the accuracy of the flux measurements, as described below. OTHER INSTRUMENTATION Net radiation was measured using commercially available miniature net radiometers (Fritschen, 1965) manufactured by Micromet Instruments, Bothell, Wash. These net radiometers have an output of about 5 VW (microvolts per watt per square meter) of net radiation. Output from these net radiometers was read out directly in terms of Wm~* through a millivolt- integrating module in the data logger. Soil-heat flux was measured using soil-heat flux plates and (or) thermocouple pairs. The soil-heat flux plates were constructed and calibrated by one of the authors (Weaver) according to the method of Tanner (1963). They were constructed by winding copper- constantan thermopiles on glass plates covered by elec- trically insulated aluminum. The thermal conductivity G14 of the plates is about 0.8 Wm~'K-', which is about the same as that for fairly dry soil. The data loggers, sonic anemometers, and Lyman- alpha hygrometers all operate with very low current drain and were powered in the field using eight D-cell batteries for each data logger and a 12-V lantern bat- tery for each sonic anemometer-Lyman alpha hygrom- eter pair. The ability to operate the systems with small DC power sources, rather than AC power, greatly en- hanced the portability of the systems and minimized power supply maintenance. FIELD MEASUREMENTS Three sites were initially chosen for eddy-correlation measurements of evapotranspiration: a thicket of old- growth saltcedar at the New Mexico State Wildlife Refuge at Artesia; a thicket of saltcedar regrowth after STUDIES OF EVAPOTRANSPIRATION mowing located about 1 km to the south; and a site with replacement vegetation, including annual weeds and alkali sacaton grass, located about 5 km south- southeast of the old-growth site. Evapotranspiration from these sites was sampled repeatedly during the growing seasons in 1980 and 1981. In 1982, measure- ments were made at these and at four other sites within the study area in order to determine the areal represen- tativeness of the three sites that had intensive meas- urements. The type of vegetative cover for all seven sites is given in table 1, and the locations are shown in figure 1. Typically, evapotranspiration measurements were made over saltcedar by mounting the eddy-correlation sensors and net radiometer on scaffolding or on a tele- vision antenna tower at a height of about 1 m above the average height of the cover. Such a height left the in- struments about level with an occasional protruding TABLE 1.-Locations and descriptions of the sites at which evapotranspiration measurements were made Saltcedar Water Approximate Site Location 52:1: £33222 Remarks (meters) (meters) Old growth 300 m NNW of SE corner of 3.4 + 0.3 8 000 Artesia Wildlife Refuge, NW!/4 of SE*/4 of sec. 35, T. 16 S., R. 26 E. Mowed regrowth --------------------- 300 m WSW by W of SE cor- 3.3 + 0.3 2.5 Mowed in 1977. ner of Artesia Wildlife Refuge, SW*!/4 of SE*/4 of sec. 35, T. 16 S., R. 26 E. Wet old growth ---------------------- 150 m east of Pecos River 0.6-1.0 §o 000 bank, about 900 m NNE of El Paso Pipeline river crossing; NW!/4 of SE*/4 of sec. 1, T. 8 S., R. 25 E. Burned regrowth --------------------- SE*/4 of NE!/4 of sec. 13, T. 17 1.5-2.0 2.5 Burned in about 1974. S., R. 26 E. Replacement vegetation Water Approximate Site Location Dominant table height of Remarks species depth vegetation (meters) (meters) Grass and forbs ------------------- Near the highway crossing of Alkali sacaton 1.7 + 0.6 0.2 Root plowed in the Pecos River east of grass, kochia, 1974. Artesia, SE*/4 of SE*/4 of and seepweed sec. 12, T. 17 S., R. 26 E. FOPb8 --- About 1.5 km west of Lea Kochia 1.5 0.1 Root plowed in Lake in the SE*/4 of 1974. NW!/4 of sec. 36, T. 11 S., R. 25 E. Gragg On Bitter Lakes National Alkali sacaton 1.6-2.1 0.3 Root plowed in Wildlife Refuge, SE*/4 of SW!/4 of sec. 34, T. 9 S., R. 25 E. grass 1974 or 1975. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. G15 dead branch or extra-tall plant. However, the fetch re- quirement that the instruments be no higher above the canopy than .01 times the horizontal distance to the upwind limit of the thicket precluded placing the in- struments at greater heights above the vegetation. Soil-heat flux was measured using a soil-heat flux plate inserted about 1 cm below land surface and by one or two sets of thermocouples placed 2 and 10 cm into the soil. The soil-heat-flux measurements are necessarily point measurements and are subject to substantial variation owing to alternate sunning and shading of the soil at the sampling point and to spatial variability in the thermal properties of the soil. At the replacement vegetation sites, the instrumen- tation was mounted on tripods or on guyed 1.9-cm (3%/4- in) pipes at heights ranging from 1.6 to 2.0 m above the land surface and at least 1 m above the taller plants in the vicinity of the setup. Soil-heat flux was measured in a manner similar to that for saltcedar. Measurements were made for approximate 5-d peri- ods in June and October 1980 and in May, June, and September 1981, and for approximate 10-d periods in June and August 1982. For most of the measurements, at least two, and as many as four, sets of instruments were in operation. However, in August 1980 and again in August 1981, rain-associated problems caused both of the available sets of instruments to fail early in the tests, and insufficient data were collected at either time to allow a valid analysis. Schedules of successful runs at each site, along with a key identifying the instru- ments used for the measurements and the approximate depth to water during the run, are listed in table 2. 700 T ST T T T 600 - NET RADIATION --- -- SENSIBLE HEAT FLUX —————— LATENT HEAT FLUX 500|- _ -: --- SoiL HEAT FLUX 400 |- 300 [- 200 |- 100 |- FLUX DENSITY, IN WATTS PER SQUARE METER ol <> Nd crime ~ reg --are M -100 | | | | L2 0 4 8 12 16 20 24 TIME (MOUNTAIN DAYLIGHT TIME) FIGURE 5.-Daily cycle of energy fluxes measured at the wet old- growth site on September 1, 1982. DISCUSSION Typical sets of daily measurements are shown for saltcedar in figure 5 and for replacement vegetation in figure 6. Radiation flux is considered positive toward the canopy or soil surface, and the other fluxes are considered positive away from the soil or canopy sur- face. The net-radiation curves are typical of those meas- ured on sunny summer days at this latitude. Time is shown as mountain daylight time, and the locations are slightly east of the 105th meridian, so that peak radia- tion occurs at about 1300 hours. Maximum net radia- tion is greater for the saltcedar than for the replace- ment vegetation because of slightly lower shortwave reflectance and because of substantially lower surface temperatures, which reduce emitted longwave radia- tion. Net radiation becomes negative at night as short- wave radiation goes to zero and the outgoing longwave radiation exceeds that incoming. Sensible heat flux typically is negative at night as heat moves from the air to the cool radiating surface, but becomes positive during the day, when the surface becomes warmer than the air. This behavior was fol- lowed in all the measurements made over saltcedar and over replacement vegetation. However, a day's run over an alfalfa field located near the burn site showed a negative sensible heat flux throughout the day as heat moved from the air to the evaporatively cooled alfalfa-an example of the oasis effect. Latent heat flux generally is near zero at night for the vegetative communities studied in this investiga- 600 1 T NET RADIATION §00]- __ --- -- SENSIBLE HEAT FLUX —————— LATENT HEAT FLUX -.--.- SOIL HEAT FLUX 400 300 200 100 FLUX DENSITY, IN WATTS PER SQUARE METER ~1090 | | I | | 0 TIME (MOUNTAIN DAYLIGHT TIME) FiGURE 6.-Daily cycle of energy fluxes measured at the grass and forbs site on June 26, 1982. G16 STUDIES OF EVAPOTRANSPIRATION TABLE 2.-Schedule of eddy-correlation measurements made during the study [Lw, Lyman-alpha hygrometer-Washington State Univ.; L¢, Lyman-alpha hygrometer-property of Campbell Scientific, Inc. (CSI); L(n), Lyman-alpha hygrometer no. n-property of USGS; He, Sonic anemometer-property of CSI; H(n), Sonic anemometer no. n-property of USGS; Hw, Sonic anemometer, property of Washington State Univ.; G, Soil heat flux plate; Gt, Soil thermocouple pair; NR, Net radiometer; sc, saltcedar; dashes indicate no data] Site Beating Beginning of? 3:21] Instrument hNumbeII' of 133:er time integrations (hours) (meters) June 1980 Grass & forbs 24 1940 34 H1, L2, 2 NR's 2 1.9 H3, L1, H2, L3, G Old-growth se 25 1000 52 H3, L2, 2 NR's 2 3.7 Saltcedar H1, L1 H2, L3, G Mowed se 26 1830 19 H1, L1 1 3.5 October 1980 Grass & forbs 21 1030 27 H1, L2, Gt, NR 4 1." H2, L1, Gt Old-growth se 22 1700 43 H1, L1, Gt, NR 2 3.3 Mowed se 22 1700 43 H2, L2, Gt 2 3.1 May 1981 Grass & forbs 11 1100 26 H2, L2, Gt, NR 1 1.7 Old-growth se 12 1600 25 H2, L2, Gt, NR 1 -- Mowed se 13 1700 24 H2, L2, Gt, NR 1 =~ June 1981 Grass & forbs 22 1600 BH H2, L2, G, NR 2 1.4 H1, L1, G, NR 2 Old-growth se 24 1030 45 H2, L2, Gt, NR 12 3.3 Mowed sc 24 1030 45 H1, L1, Gt, NR 2 3.0 September 1981 Old-growth se 14 0830 57 H1, L1, Gt, NR 13 3.2 Mowed se 14 0830 80 H2, L2, G, NR 123 -- He, Le 123 Grass & forbs 16 2030 20 H1, L1, G, NR 11 He, Le 1 1.2 June 1982 Grass & forbs 6-22-82 1430 112 H3, L1, G, NR 7 2.1 H1, L2 Old-growth se 6-24-82 0800 35 H4, L2, G, NR 2 3.3 Burned se 6-26-82 0700 55 H4, L2, G, NR 3 21 Wet old-growth se 6-28-82 1300 53 H3, L1, G, NR 1 1.0 Grass 6-29-82 0900 15 H4, L2, G, NR 31 2.1 August 1982 Grass & forbs 8-24-82 0830 478 H4, Lw, G, NR5S 1 2.1 Old-growth sc 8-25-82 1000 122 H3, L2, G, NR 2 3.3 Burned se 8-26-82 1230 54 He, Le, NR" 2 zw Mowed se 8-28-82 1000 56 Hw, Lw, G, NR 2 _- Wet old-growth se 8-29-82 92 H3, L2, G, NR 3 0.6 Grass 8-29-82 1900 88 H4, L4, G8 3 2.5 He, L2, G, NR 2 Forbs 8-31-82 1100 51 Hw, Lw, G, NR 2 1 120-hr period; 4-hr shortfall at night. 2Two 16-hr periods; 8-hr shortfall at night. 39-hr shortfall, mainly at night. not continuous. 5After 0800, 8-26-82, R, was reconstructed from R,, T,, T,,, and p,. (Abbreviations explained in text.) replaced Lw at 1330, 8-25-82. No G measured; assumed G=0.07 R,,. SUntil 1400, 8-31-82, R,, was reconstructed from R,, T,, T,, and py. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. tion, as the plant stomates tend to close and the vapor deficit in the canopy becomes small. Typically, the la- tent heat flux starts positively upward about sunrise and returns to near zero about sunset. Soil-heat flux is generally positive (heat moving from the surface into the soil) during the day and is negative at night. Because the measurement is made at a point beneath partial vegetative cover, the measured soil-heat flux fluctuates during the day as the plate is alternately in sun and shade. Soil-heat flux is quite large at the replacement-vegetation sites, which typi- cally have sparse vegetative cover, and maximum day- time soil-heat flux is typically about one-third of maxi- mum net radiation. For the saltcedar sites, on the other hand, maximum soil-heat flux is commonly about 10 percent of maximum net radiation. ENERGY-BUDGET CLOSURE Throughout the measurements, the eddy-correlation fluxes have been qualitatively correct. Positive sensi- ble heat flux was consistently indicated whenever up- ward air temperature gradients were measured or whenever surface temperatures, as measured by an infrared thermometer, exceeded air temperature. Like- wise, negative sensible heat fluxes always coincided with downward air-temperature gradients or with sur- face temperatures cooler than air temperatures. Meas- ured vapor fluxes also always appeared to be qualita- tively consistent, showing expected patterns in almost all cases. Despite the good qualitative behavior of the eddy- correlation fluxes, it is obvious from figures 5 and 6 that the magnitude of the sum of the sensible and la- tent heat fluxes is substantially less than the net radi- ation minus soil-heat flux. For example, the daily sums of the energy-flux densities at the wet old-growth saltcedar site for September 1, 1982, included 14.9 MJm~ of net radiation minus 1.1 MJm~ of soil-heat flux, or 13.8 MJm~". The daily sum of sensible and latent heat fluxes measured by the eddy-correlation system was 8.4 MJm~*, or 0.61 of the energy budget. Similarly, for the grass and forbs site, net radiation minus soil-heat flux was 12.8 MJm~2 on June 26, 1982, whereas the sum of sensible and latent heat fluxes was 8.6 MJm~, or 0.67 of the energy budget. Thus, a sub- stantial portion of the energy budget is missed by the eddy-correlation system, as will be discussed in greater detail below. On a half-hourly basis, the energy-budget error is largest in the morning and becomes progressively smaller in the afternoon. This probably occurs because of heat storage in the soil above the soil-heat flux plate G17 and in the biomass, and can be minimized by integrat- ing all flux measurements over a 24-h period. The need for such integration is particularly important because of the point-measurement character of the soil-heat flux. Collection of data for short-term analysis would have required the measurement of soil-heat flux by an array of sensors, as recommended by Tanner (1963) and by Fritschen and Gay (1979). During each run, eddy-correlation data generally were collected at each site for at least 24 h, and some- times for more than 100 h. Results of these runs were integrated over 24-h periods to minimize diurnal varia- tions in soil and biomass heat storage. If data were collected for other than 24 h, the data commonly were integrated twice, once starting with the beginning time and again starting with a later time that was an even multiple of 24 h previous to the end time. This proce- dure was followed to ensure full use of the collected data. Fluxes measured by eddy correlation and the energy-budget components measured at the various sites are presented in table 3, along with the fraction of the energy budget measured by the eddy-correlation system. Although this fraction ranges from 0.6 to 1.0, it frequently is in the 0.6 to 0.7 range for all sites except the old-growth saltcedar site, where closure of the energy budget is consistently better. Failure to close the energy budget has been a matter of persistent concern since the start of the study, and several attempts have been made to identify the cause and correct the problem. Duplicate systems were oper- ated side by side repeatedly throughout the project, and the measured fluxes almost always agree within 10 percent, suggesting that individual sensors and data loggers are consistent. Careful calibration of the Lyman-alpha hygrometers in an environmental cham- ber in September 1980 and by pumping air containing a known vapor content in January 1983 led to nearly identical calibrations. The good agreement achieved by the careful calibrations suggests that no systematic error due to calibration shifts occurred. The calibrations indicated that the voltage output of the anemometers deviated by as much as +7 percent from the factory calibration, suggesting that little bias in the estimates due to the use of the factory calibra- tions should have occurred. The shortwave response of the net radiometers was determined periodically by simultaneously shading and exposing both the net radiometer and a precision Eppley pyranometer, and calibrating the net radiome- ter accordingly. These calibrations showed some deteri- oration in the response of the net radiometers as the polyethylene shields became scratched, but the origi- nal factory sensitivity was restored when new shields G18 STUDIES OF EVAPOTRANSPIRATION TABLE 3.-Summary of daily sums of energy-density measurements at the various sites for which eddy-correlation data were collected [Energy density in megajoules per square meter. H, sensible heat flux; Lec, latent energy flux by the eddy-correlation technique; Leb, latent energy flux as the residual of R, -G-H; G, soil heat flux; R,,, net radiation; Rec, recovery ratio, or fraction of energy budget measured by the eddy-correlation system] Date H Lec Leb G Rn Rec Old-growth saltcedar site 1980 June 25 10.2 2.1 4.5 1.7 16.4 .88 11.2 4.1 5.5 1.7 18.4 .92 10.1 2.5 4.8 1.9 16.8 .85 hss 9.7 2.4 4.1 3.0 16.8 .88 10.0 2.4 4.2 2.7 16.9 .87 Oct, 22 3.7 3.9 4.2 -0.8 7.6 .96 BJ 4.7 3.9 4.8 -1.0 8.5 .91 1981 May 5.3 7.5 11.8 0.6 17.7 15 June 24 7.6 1.7 3.9 0.6 12.1 .81 3D 13.4 1.9 2.2 0.5 116.1 .98 Sep 14 A lel ect an 3.2 10.8 10.8 0.4 114.4 1.00 15 ress. seed nene. ore 2.2 9.2 12.3 -0.4 14.1 19 1982 June 24 10.9 3.2 7.9 0.8 19.6 15 JG 12.7 2.9 5.2 1.4 119.3 .87 Aug. 28 6.2 3.2 7.6 1.8 15.6 .68 BQ - ick Aras s scan bles sens 7.9 S1 6.1 2.0 16.0 19 Mowed saltcedar site 1980 June 26 ssn mee. 4.4 -- 2.0 210.0 -- Oct. 22 2.4 3.8 4.6 0.6 7.6 .89 0J aes 3.8 3.0 5.5 -0.8 8.5 13 1981 May 18 4.8 5.4 11.7 0.0 16.5 .62 June 24 1.1 8.8 14.6 0.9 16.6 .63 gn Ar.. dell MU oci 2.3 8.4 15.0 0.9 18.2 .62 Sep.: 14 3.4 5.8 10.5 1.1 15.0 .66 sn =s ss 2.3 5.7 10.7 0.8 13.8 62 16 ess 3.2 7.8 11.0 1.4 115.6 T4 1982 ut AP ee 3.0 4.8 10.2 1.7 14.9 .59 OQ .no. Aire 3.0 5.9 10.8 1.7 15.5 64 Burn site 1982 June 26 6.9 5.0 11.4 1.5 19.8 .65 0] ss 6.0 4.1 8.3 1:1 15.4 T1 7.7 4.7 10.6 1.4 19.7 .68 Aug. 27 84.2 36.6 8.9 1.4 14.5 .82 2 34.6 37.1 6.8 1.3 12.7 1.03 Wet old-growth site 1982 June 27 1.0 7-1 12.3 1.9 15.2 .61 Ip ste 48 d ans 0.3 12.4 15.6 41.8 17.7 .80 ges .r : AEL. NITES es- 0.9 12.0 16.1 41.9 18.9 16 Aug. 830 2.1 5.8 10.3 0.9 13.3 64 $1. ns rs seus 2.8 5.5 11.1 0.5 14.4 .60 yep "In=. o lr t_ roll c_ _ CL Toy 3.2 5.2 10.6 1.1 14.9 .61 WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. G19 TABLE 3.-Summary of daily sums of energy-density measurements at the various sites for which eddy-correlation data were collected- Continued Date H Lec Leb G Rn Rec Grass and forbs replacement-vegetation site 1980 June 24 6.0 2.5 4.7 4.0 14.7 19 Lo H A Hei 3.1 2.3 2.2 0.3 5.6 1.02 3.4 2.5 2.9 0.5 6.8 0.94 May 11. 6.0 3.5 7.2 0.9 14.1 12 June 22 8.1 1.7 1.6 0.7 10.4 1.01 6.9 1.0 3.3 0.6 10.8 117 03 7.9 1.5 4.3 0.7 12.9 +17 6.4 0.9 1.1 0.9 8.4 .97 Sep: 106 1.5 2.6 5.4 -0.8 16.1 59 1982 June 22 4.8 3.2 8.5 1.8 15.1 .60 4.6 3.3 7:4. 11 13.4 64 $3 on 4.7 3.1 7.5 1.3 13.5 64 B4 5.2 4.1 8.1 2.2 15.5 70 B5 5.7 3.6 8.7 1.9 16.3 .65 06 sn 6s o 5.1 8.5 7.7 2.6 15.4 .67 Aug. 24 2.9 3.9 7.2 2.1 512.2 .67 BT 2.6 5.5 7.3 3.1 613 15 OB on 4.4 3.5 5.3 8.93 $13 59 Grass site 1982 June 27 6.6 1.4 7.6 0.6 714.8 .56 Aug. 30 4.1 4.5 -- 2.2 -- f BL 2.9 3.4 -- 3.0 8 __ -- 3.8 3.5 7.2 2.1 13.1 .66 Sep. = 1 3.1 4.2 -- 4.4 813.1 .84 4.2 4.2 7.2 2.7 14.1 14 Forbs site 1982 Aug. 81 3.6 -- 6.4 1.2 11.2 -- Sep. 1 4.0 3.0 7.6 1.9 13.5 60 'Less than 24-hr period; shortfall at night. CEstimates based on ts of solar radiation, albedo, surface temperature, and 2Less than 24-hr period; Lec three times that at old-growth site. Includes Lec of about 3 MJ/m2 and H of -0.9 that are night time anomalies. Data not included in evapotranspiration or other analyses. 4G not measured, but estimated as 0.1 R,,. 5Integration from 0800 to 1800. were installed. Field calibrations were made fre- quently enough so that no systematic problem due to calibration drift should have arisen. The total response of the net radiometers was also checked in 1982 by independently measuring the com- ponents making up net radiation, as given by the equa- tion R, =R,q -R,, +Lq-Ly, ; (22) where R,; =shortwave incoming radiation; air temperature resulted in computed net radiation of 19.5 MJ on Aug. 27, and of 16.8 MJ on Aug. 28. These values are implausibly large and were replaced by the rounded values for net radiation measured at the grass and the forbs sites a few days hence. 7Integration for 19-hr period; Lec for 16-hr period, including entire daylight interval. Integration for 21-hr period; R,, not available or indirectly estimated. R.,, =shortwave outgoing radiation; L; =longwave incoming radiation; and L,, =longwave outgoing radiation. R,,; was measured using a Kipp pyranometer pointed upward. The same pyranometer was periodically in- verted to obtain the shortwave reflectance of the canopy. Longwave outgoing radiation was computed from the surface temperature of the canopy measured with the infrared thermometer using the equation (23) L, =ec T,* , G20 where e=longwave emissivity of the surface, dimension- less; a=Stefan-Boltzman constant, 5.67 10~° Wm-~°K~+; and T, =surface temperature, K. e was assumed to be 0.98 for the surface in each case, as almost all vegetative and bare-soil surfaces have an emissivity ranging from 0.95 to 1.00. One estimate of L,, was obtained using air temperature in the equation (Montieth, 1973, p. 34) Lg=e,oT* , (24) where e, =longwave emissivity of the atmosphere, com- puted by the empirical equation e, =1.2-171V/0T,4 (25) where T,, =air temperature, K. A second estimate of e, was obtained using the vapor density of air at a height of approximately 1 m above land surface in the equation (Brutsaert, 1975) e=0.58 p'" , (26) where p=water vapor density, in gm ~*. These checks consistently indicated a net radiation that is about 10 percent less than that measured by the Fritschen (1965) net radiometer. These estimates sug- gest that the longwave calibration for the Fritschen net radiometer may be in error, and that net radiation is slightly overestimated on an algebraic basis. However, the error would appear to be substantially smaller than the energy-budget-closure error. Another potential source of error involves the fre- quency cutoff for high-frequency eddies that arises be- cause of limits on the rate of sampling imposed by the response time of the sensors or by the processing speed of the data logger. The response times for the fine-wire thermocouple and the infrared hygrometer are on the order of 0.01 s. Response time of the anemometer is difficult to determine, because it can result from aver- aging the velocity of a small eddy over the 10-cm sonic path length, as well as the speed of response of the electronics. Nonetheless, it almost certainly is substan- tially less than 0.1 s. The data logger samples by simul- taneously integrating each signal for 0.1 s. This inte- gration smoothes and attenuates high-frequency signals, so that 98 percent of a sinusoidal signal with a period of 1 s, 76 percent of a sinusoidal signal with a period of 0.25 s, and none of a sinusoidal signal with a period of 0.1 s is measured. Moreover, the covariances of water-vapor density versus vertical wind and air STUDIES OF EVAPOTRANSPIRATION temperature versus vertical wind are attenuated by the product of the attenuations of the two signals. Hence, sampling by integration results in an effective sampling frequency cutoff of about 3-4 Hz. Tests were conducted to determine whether a significant portion of the fluxes were not being sampled because of this high- frequency cutoff. In one test, the sensors were instanta- neously sampled at about 100 Hz using a microcom- puter as a data logger. In other tests, data were recorded at heights varying by a factor of 2 to make use of the tendency of the eddy frequency to decrease with height above land surface according to the relationship U fa; > where f=frequency of eddy, s'; %=mean horizontal windspeed, ms~'; and z=height above land surface or zero-displacement plane, m. In no case was energy-budget closure improved. Other tests were made to determine whether some of the flux was not being measured because of the low- frequency cutoff imposed by averaging the fluxes for time blocks of 2.5 or 5.0 min. However, the change from 2.5- to 5.0-min averages showed no effect on budget closure. Another test in 1979 included side-by-side comparison of monitoring systems with 2.5-min and variable 5- to 30-min block averages. In general, the two systems behaved identically. Occasionally, how- ever, the system for which 15-min or 30-min averages were obtained showed a digression in the measured fluxes to give combined flux values that substantially exceeded the energy budget. These excursions were as- sumed to represent correlations in wind, temperature, and vapor density that do not represent eddy flux. Fi- nally, the fact that systems operating simultaneously at different heights gave flux values that agreed as closely as side-by-side systems suggests that the meas- ured fluxes were not being attenuated by the low- frequency cutoff either. Another hypothesis concerned possible interference with the vertical wind component by the underlying base of the original sonic anemometer. It was felt that the base could attenuate updrafts but have a smaller effect on downdrafts. One experiment performed in a strong wind during the winter of 1981 at the Denver Federal Center, Colo. suggested that inverted and up- right sonic anemometers gave somewhat different sen- sible heat fluxes. Consequently, the sonic anemometer was redesigned (fig. 4B) to eliminate interference by the base. No difference in budget closure was detected during the subsequent 1982 measurements, which were made under much calmer conditions. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. Another consideration was the relative spacing of the fine-wire thermocouple and of the position of the Lyman-alpha hygrometer in relation to the sonic path of the anemometer. Both temperatures and vapor den- sity were measured at points spatially removed from the point at which wind was measured. This separation results in some spatial averaging of the correlated wind-temperature and wind-vapor measurements, which in turn would result in at least minor attenua- tion of the measured fluxes. The fine-wire thermocou- ple is physically part of the anemometer and is located 4 em from the sonic path on a perpendicular through the midpoint between the sonic transducers. The Lyman-alpha hygrometer is mounted separately, and it usually was positioned so that the vapor-density measurements were made about 15 ecm from the sonic path. Thus, the vertical wind-vapor density correlation should be at least somewhat more attenuated by space- averaging than the vertical wind-temperature correla- tion. However, an experiment was conducted at the Nevada Test Site in 1983 in which one Lyman-alpha hygrometer was positioned to make vapor-density measurements about 5 cm from the sonic path, while another was positioned to make measurements 15 cm from the same anemometer. The flux measurements determined from measurements from the two hygrome- ters were almost identical. The inherent patchiness of the wild land vegetation was also considered a possible cause of the lack of energy-budget closure. Theoretically, this patchiness could cause horizontal gradients in the vertical flux of heat and vapor, at variance with the assumptions made in making the measurements. Deviations in the meas- ured flux due to such surface heterogeniety should de- crease as instrument height above the surface in- creases. However, as mentioned above, instruments operated at different heights produced nearly identical results. Also, experiments were performed over a bare field at a lysimeter installation at Davis, Calif.; in al- falfa fields near Artesia, N. Mex., Logan, Utah, and Kimberley, Idaho; a field with a dense wheat-stubble cover at Kimberley; and over parking lots at Washing- ton State University in Pullman and at the Denver Federal Center, Colo. In all of these cases, the surface for exchange of heat and vapor areally was quite uni- form, but the magnitude of departures in the energy- budget closures were similar to those obtained in the Pecos River floodplain. ESTIMATES OF WATER USE Although the lack of energy-budget closure casts doubt on the validity of the eddy-correlation measure- ment of vapor flux, the data can be used to establish G21 lower and upper limits on evapotranspiration. The lower limit is established by assuming that the latent energy flux, \E, measured by eddy correlation (eq. 20) is correct. Such an assumption implies that all the eddy-correlation error occurs in the sensible heat-flux measurement (H, eq. 13). The upper limit of evapotran- spiration is obtained by assuming that the sensible heat-flux measurements are correct and that all of the error occurs in the vapor-flux measurement (QE in eq. 20). Under this assumption, the latent heat flux is computed as the residual of the energy budget minus sensible heat flux: Leb=R,-G-H , (27) where H is as measured by equation 13. The contention that Leb represents a maximum esti- mate of evapotranspiration is based on the assumption that sensible heat flux is not overestimated by the eddy-correlation measurements. This is a reasonable assumption, in that most of the recognized potential sources of error discussed earlier result in underesti- mates of flux. Also, during the tests performed over parking lots, latent heat flux was nearly zero yet the eddy-correlation-measured sensible heat flux was less than the measured energy budget. These tests thus show that sensible heat flux is at least somewhat underestimated by the eddy-correlation measure- ments. The eddy-correlation values of latent heat flux, sym- bolized as Lec, differ greatly from the energy-budget residual (Leb) values. Several arguments can be made in favor of the energy-budget values: 1. The output of the fine-wire thermocouple is sta- ble, and the theoretical calibration is well known, sug- gesting that the sensible heat flux is well determined. On the other hand, the Lyman-alpha hygrometer must be calibrated experimentally. Although careful cali- brations performed in June and September 1980 and in January 1983 are in good agreement, the recom- mended practice of frequent Lyman-alpha calibration (Redford and others, 1980) was not followed. Moreover, the output of the Lyman-alpha hygrometer shows sub- stantial diurnal drift. It has been assumed that this drift affects the offset, but not the slope, of the voltage- vapor density curve, and that the flux measurements are little affected by the drift. Nonetheless, additional uncertainty is added to the vapor-flux determinations. Finally, when relative humidity is high, current leak- age sometimes occurs along the outer surface of the detector tube in the Lyman-alpha hygrometer, creating an additional voltage offset that affects the mean volt- age, and possibly the calibration of the hygrometer. This mean voltage is used as a divisor in computing G22 vapor flux. Also, as discussed above, the vapor-density measurements are made at a point farther away from the wind measurements than are the temperature measurements, potentially resulting in greater error in measuring vapor fluxes. Hence, the vapor-flux compu- tations are subject to many more instrumentation un- certainties than are the sensible heat-flux determina- tions. 2. The Leb estimates for the different saltcedar sites for May, September, and October (described below) are in close agreement, whereas the Lec estimates show substantial variance. Soil moisture storage was high at those times owing to recent rains, and it is plausible that the rates should be similar. 3. A multiple regression of the energy budget (R,,-G) against sensible heat flux () and latent heat flux (Lec) for the data in table 3 yielded the following equation (forced through the origin): (R,,-G)=1.13H +1.61Lec . (28) This relationship suggests statistical confirmation that most of the underestimation occurs in the latent heat determination. 4. Estimates of evapotranspiration by the Leb method are in good agreement with those determined by some other studies, as described below, whereas esti- mates determined by the eddy-correlation technique are much lower than most other measurements. Because of the preponderance of data favoring the use of the Leb values, those values are emphasized. WATER USE BY SALTCEDAR Two sets of water-use estimates for saltcedar, derived by multiplying Lec and Leb in table 3 by a conversion of 0.41 mm water per MJm~*d~', are sum- marized in table 4 and in figures 7A and 7B. The con- version factor is based on a value for the latent heat of vaporization of water of 2,450 Jg~*. These figures show the water-use measurements for each saltcedar thicket for the 3-yr period of measurements. Water use varies substantially among sites. Depth to water, plant den- sity, and plant age also vary substantially among sites, but there is no apparent correlation between measured water use and these factors. For example, the old- growth site provides about 80-percent ground cover, although it was mapped as dense saltcedar by Mower in 1958 (Mower and others, 1964). The saltcedar at the old-growth site was burned sometime prior to 1974, but the growth seems to be again quite mature. The mowed saltcedar thicket, located about 1 km southwest of the old-growth site (fig. 1), currently provides about 50- percent ground cover, although it was mapped as mod- erately dense (70-90-percent cover) by Mower in 1958. STUDIES OF EVAPOTRANSPIRATION TABLE 4.-Summary of daily water-use estimates at the various sites Water use (Evapotranspiration) (millimeters per day) Site Date Eddy Energy budget correlation residual Saltcedar 1980 Old growth -------- June 25-26 1.2 1.9 Oct. 22-23 1.6 1.8 1981 May 12 8.1 4.8 June 24-25 0.7 1.3 Sep. 14-15 4.1 4.7 1982 June 24-25 1.2 2.17 Aug. 28-29 1.3 2.8 1980 Mowed ------------ June 26 1.8 |= 00 Oct. 22-23 1.4 2.1 1981 May 13 2.2 4.8 June 24-25 3.5 6.0 Sep. 14-16 2.6 4.4 1982 Aug. 28-29 2.2 4.8 Burned ------------ June 26-27 1.9 4.1 Aug. 27-28 2.8 3.2 Wet old growth ---- June 27-29 4.3 5.4 Aug. 30-Sep. 1 2.2 4.4 Replacement vegetation 1980 Grass and forbs --- - June 24 1.1 1.5 Oct. 21 1.0 1.1 1981 May 11 1.4 2.9 June 22-23 0.5 1.1 Sep. 16 11 2.2 1982 June 22-26 1.4 8.3 Aug. 24-28 1.8 2.1 Grass -------------- June 27 0.6 2.9 Aug. 30-Sep. 1 1.6 8.1 Forbs -------------- Aug. 31-Sep. 1 1.2 2.9 The site was mowed in the fall or winter of 1977, and no shoots were visible in March 1978. This site has a sub- stantial understory of alkali sacaton grass. Depth to water at both sites ranges from 3 to 3.5 m. The burn site is located about '/2 km south of U.S. Highway 83 and about 5 km south-southeast of the old-growth site. The site consists of dense saltcedar and is regrowth from a burn that occurred in 1974. The depth to water is about 2 m, and the vegetation is about 4 m high. The site is south of the area mapped by Mower. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. 10 G23 A BURNED WET OLD GROWTH MOWED OLD GROWTH WATER USE (EVAPOTRANSPIRATION), IN MILLIMETERS PER DAY 0 MAY 12-13 JUNE 24-29 AUG. 27 - SEPT. 1 SEPT. 14-16 OCT. 22-23 FiGURE 7.-Summary of estimates of daily evapotranspiration by saltcedar by (A) the energy-budget- residual method, and (B) the eddy-correlation method. The wet old-growth site is a dense saltcedar thicket about 3 km north of the U.S. Highway 70 crossing. There are no signs that the saltcedar ever burned, and the site has a very shallow (less than 1 m) depth to water. The saltcedar is about 5 m tall. Water use by the different thickets showed great variation, particularly during June and August. For example, in June 1981 evapotranspiration from the mowed saltcedar was about five times that from the old growth, based on either method. Other supporting data suggest that these differences are qualitatively correct. Stomatal diffusion resistances of fronds of the old growth were about three to five times greater than those of fronds for the mowed saltcedar in June of each G24 year. Also, in June 1982, the old-growth saltcedar had midafternoon frond temperatures, as measured by an infrared thermometer, of 43°-45°C, whereas the mowed saltcedar frond temperatures were 36°-38°C. These frond temperatures suggest that much less evap- orative cooling was occurring from the old growth than from the mowed saltcedar. Finally, the old-growth saltcedar showed visible signs of moisture stress and had not flowered, whereas the mowed saltcedar was vigorous and in full bloom. Measurements made during each June followed a long sequence of hot, dry, sunny days, when soil moisture was substantially depleted. Finally, examination of hand-auger samples obtained at the old-growth site indicated that live roots were prevalent in the top 1.8 m of material above the clay layers (fig. 8) but were sparse in and beneath the clay layers. At the mowed site, on the other hand, live roots were prevalent throughout the section down to the water table. These observations suggest that the old- growth saltcedar derives most of its water from soil moisture, while the mowed saltcedar relies heavily on ground water during periods of extreme moisture stress. Because of the alkali sacaton in the understory, water use at the mowed site represents a combination of water use by the saltcedar and the grass. It is as- sumed that the saltcedar accounted for most of the evapotranspiration. Evapotranspiration determined as the energy- budget residual (Feb) for wet old growth (4.4 mm/d) slightly exceeded that for the mowed site (4.3 mm/d) in August 1982. No measurements of evapotranspiration from the mowed saltcedar were made in June 1982, but the June 1981 evapotranspiration from the mowed site (6.0 mm/d) exceeded the June 1982 evapotranspiration from the wet old-growth site (5.4 mm/d) by about 10 percent. Both measurements of Eeb from the burn site are intermediate between the old growth and the mowed site. WATER USE BY REPLACEMENT VEGETATION Estimates of water use by the replacement vegeta- tion are listed in table 4 and shown in figures 9A and 9B. Water use, as measured by the residual-energy- budget method, ranged from as low as 1.1 mm/d to as much as 3.3 mm/d and typically averaged about 3 mm/ d. The value of 2.2 mm/d measured on September 16, 1981, was measured on a cloudy day and may be atyp- ically low. The forbs at the grass-and-forbs site varied greatly between measurements, although alkali sacaton and saltbush (Atriplex patula) are perennials and were al- ways present. In June 1980, desert seepweed (Suaeda suffractescens) was the dominant species, with some STUDIES OF EVAPOTRANSPIRATION Russian thistle present. Large bare patches were present, and young kochia had died. By the ill-fated visit in August 1980, when rain caused instrumenta- tion failures, kochia was the dominant annual and desert seepweed had nearly disappeared. Kochia re- mained dominant in October 1980. In May and June 1981, desert seepweed, horsenettle (Solanum elaeagnifolium ), and an unidentified weed were prevalent. By August and September, kochia, a mustard plant (Brassica nigra), and Russian thistle were again common. In June 1982, kochia, the mus- tard, and desert seepweed were prevalent (the first June in which kochia was dominant). The kochia and mustard were still prevalent in August 1982. Vegetative cover at the forbs site is dominantly kochia. The dominant cover at the grass site is alkali sacaton, a grass that apparently was one of the common types of cover before the saltcedar infestation began in the Pecos valley. COMPARISON WITH POTENTIAL EVAPOTRANSPIRATION Potential evapotranspiration is defined as the evapo- transpiration rate from a well-watered reference crop that completely shades the ground, such as short grass (Penman, 1956) or alfalfa (Jensen, 1973). Potential evapotranspiration is dependent on climatic factors, in- cluding solar radiation, air temperature, vapor pres- sure or vapor density, and windspeed, and on crop prop- erties, including the aerodynamic roughness of the vegetation, reflectance of the canopy surface to short- wave radiation, the canopy temperature, and the resis- tance to vapor loss imposed by the leaf stomates. The vegetative properties of short grass or alfalfa can be estimated, so that potential evapotranspiration can be defined by climate alone. The concept is widely used to estimate evapotranspiration from phreatophytic vege- tation. Hence, a comparison of values measured in this study with potential evapotranspiration is useful to evaluate the suitability of the potential evapotranspi- ration concept for such estimates. Many different equations exist for computing poten- tial evapotranspiration; the selection depends on the climatic data available and on usually site-specific cal- ibration constants (Jensen, 1973). The Jensen-Haise equation (Jensen, 1973, p. 73, 74) was used to compute potential evapotranspiration for this study. The equa- tion is E,2041(T,-T;) R,;iWCi+Cz CH) ; (29) where E,, =potential evapotranspiration, in mm/d; T., =mean monthly air temperature, in °C; WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. G25 Sand, fine to medium Silt, sandy Sand, silty Clay, silty Sand, silty, clayey Clay, silty, with gypsum crystals Sand, clayey, silty Clay, silty, with gypsum crystals above 3.5 m. Water table Sand, clayey A DEPTH BELOW LAND SURFACE, IN METERS Sand, fine to medium Clay Sand, clayey, to clayey sand Clay, silty C Silt, sandy, with gypsum crystals Clay, silty, with gypsum crystals Silt, sandy to clayey Sand, silty to clayey, with gypsum crystals Clay, sandy, with caliche nodules Sand, fine, silty Clay, sandy, sitly plastic to sand, clayey Sand, fine, with silt and clay D FiGURE 8.-Sample logs showing the nature of the materials from land surface to the water table for each saltcedar measurement site: (A) old-growth site, (B) regrowth site, (C) burn site, and (D) wet old-growth site. G26 STUDIES OF EVAPOTRANSPIRATION 8 A GRASS AND GRASS FoRss Forss g ~- C® C0 SS m wal o ir T ise [C] ® CC u G s = ; s: 3 ~f w oa I _c > a m o r- O ye ou I> a O - 2- l k.— < C = a. «- ® C e at < CC § 0 S @a 4 w ¢ ! B CC u mi g 2 c S. < (£- = 0 SB ‘ | C MAY 11 JUN. 22-27 AUG. 24 - SEPT. 1 SEPT. 16 OCT. 21 FIGURE 9.-Summary of estimates of daily evapotranspiration by replacement vegetation by (A) the energy-budget-residual method, and (B) the eddy-correlation method. T,=-2.5°C-1.4(e,-e;)°C/kPa-(altitude in m/ 550 m)°C; R, =incoming solar radiation, in MJm~*d~; C, +(-2°C x altitude in m/305 m); 02:7.600; CH=5O kPa/(ez—el); and ei, ea=saturation vapor pressures, in kPa, at the maximum and minimum average tempera- tures for the warmest month of the year. Solar radiation records from El Paso, Tex., and mean monthly temperatures at Roswell, N. Mex., as provided by the National Weather Service, were used for the computations. El Paso is located about 205 km south- west of Artesia, and Roswell is located about 60 km north of Artesia. The climates of the two cities are relatively similar to the climate of Artesia, although solar radiation at El Paso might be 10 percent or so greater than at Artesia on an annual basis. Results of these computations are shown by the potential evapo- ration curves in figures 10 and 11. Annual potential evapotranspiration was computed to be 1.75 m. The energy-budget-residual estimates of water use by saltcedar ranged from 13 percent of potential evapo- transpiration for the lowest value measured at the old- growth site to about 65-70 percent for the largest val- ues measured during each month. The fact that the estimated maximum water use by saltcedar is less than Jensen-Haise potential evapo- transpiration is consistent with observations made during the study. The empirical constants used in the Jensen-Haise method were developed to provide esti- mates that agree with the measured water use by al- WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. falfa. The stomatal resistance to vapor loss from alfalfa is very low, so that the resistance to vapor transport from the canopy is comparable to the resistance to heat transport. Hence, the alfalfa behaves as a freely tran- spiring surface and is usually cooler during daytime in summer in the Artesia area than the air, owing to evaporative cooling. All of our observations indicated that saltcedar is less freely transpiring, and that the resistance to vapor transport is greater than the resis- tance to heat transport. This is evidenced by the fact that saltcedar surface temperatures during sunny sum- mer days, as measured by an infrared thermometer, were always higher than the air temperature, both at our eddy-correlation measurement sites and at various other saltcedar thickets in the floodplain, as spot- measured in June and August 1982. Thus, the Jensen- Haise equation will overestimate water use by saltcedar, at least in the Roswell basin area, unless a crop coefficient is used to adjust the values downward. A similar conclusion regarding water use by saltcedar in the Gila River floodplain was reached by Culler and others (1982, p. P39) on the basis of their water-budget measurements of saltcedar evapotranspiration. ANNUAL WATER USE The potential for salvage of ground water by saltcedar eradication is dependent on annual water use by both the saltcedar and the replacement vegetation. However, only spot measurements were made, and these were made only during the growing season. Hence, it was necessary to interpolate between meas- 10 ANNUAL WATER USE (m) (From areas under curves) & I Maximum 1.07 Median 0.90 Minimum 0.77 o I PS I y ‘y‘w/ WATER USE, IN MILLIMETERS PER DAY | SEPT. 1 JAN. 1 MAR. 1 MAY 1 JULY 1 NOV. 1 DEC. 31 G27 urements, and to extrapolate the measurements through the nongrowing season. For the growing season, maximum, median, and minimum estimates of water use were made by draw- ing straight-line segments through selected data. (More formal interpolation schemes were attempted initially, but these schemes inferred low water use after the start of the wetter "monsoon" season in July. Hence they were deemed less realistic than the use of straight-line segments.) For the energy-budget- residual estimates, minimum water use by saltcedar was obtained by drawing line segments through the measured values for the old-growth site, including the average of only the two lower June measurements. The median estimates are based on average values for all available saltcedar measurements for each month. The maximum water use was determined by the data from the mowed site, which gave the largest water-use val- ues in midsummer. Evapotranspiration during the nongrowing season (approximately November 15 to April 1), both for saltcedar and for replacement vegeta- tion, was arbitrarily assumed to be 40 percent of poten- tial evapotranspiration. Saltcedar in the Roswell area drops its fronds after a killing frost and leafs out again in mid-April. Hence, it is reasonable to assume that evapotranspiration for both cover types, consisting mainly of evaporation from the soil surface, is about equal during the nongrowing season. The estimation procedure is illustrated in fig- ure 10A . The daily water-use rates for saltcedar meas- ured by the energy-budget-residual method are plotted ANNUAL WATER USE 8 (From area under curve) 0.6 m WATER USE, IN MILLIMETERS PER DAY o \ JULY 1 __ SEPT. 1 nov. 1 DEC. 31 FIGURE 10.-Estimated annual water use by saltcedar in the Pecos valley, New Mexico, as determined by (A) the energy-budget-residual method, and (B) the eddy-correlation method. G28 as bar graphs. Line graphs through the selected values or means show the estimated minimum, median, and maximum rates. The areas under these line graphs were integrated to determine the annual water-use rates given in table 5. This procedure indicates that water use by saltcedar at the old-growth site (the min- imum value) may be about 0.77 m/yr, whereas that for the mowed site (the maximum value) is about 1.07 m/ yr. The average for the four sites tested is about 0.9 m/yr. Estimates of annual water use by saltcedar based on the eddy-correlation measurements (fig. 10B) were made only for the mean of the values for all sites for each month. These measurements indicated that water use at the mowed site in May and September was less than that at the old-growth site, although the mowed- site measurements were greater in June and August (table 4). Hence, the approach described above for ana- lyzing the energy-budget-residual measurements would have resulted in maximum, mean, and mini- mum values that are all of about the same magnitude. The analysis of the eddy-correlation measurements in- dicates that the minimum value for average water use by saltcedar is 0.6 m/yr. Estimates of annual water use by the replacement vegetation (fig. 11) were obtained in a similar manner. For the energy-budget-residual measurements, mini- mum water use was determined as that for the grass- and forbs-site, using only the seasonally low values for June 1980 and June 1981, the minimum value meas- ured in August 1982, and measured values for May, September, and October (table 4). The median value was obtained using the averages of all the monthly measurements. The maximum estimate is based on the highest measurement each month, and, in addition, it was assumed that the September measurement, made on a cloudy day, was equal to the average of the maxi- mum August measurement and the September meas- urement. These data indicate that water use by re- placement vegetation ranged from 0.57 to 0.67 m/yr (table 5). TABLE 5.-Estimates of the range in annual water use by saltcedar and by replacement vegetation in the Pecos River floodplain and of projected water salvage due to saltcedar eradication [In meters per year} Replacement Saltcedar vegetation Salvage Based on the energy-budget residual Maximum 1.07 0.67 0.40 Median .90 .62 .28 Minimum vy l .57 .20 Based on eddy correlation Median 0.60 0.40 0.20 STUDIES OF EVAPOTRANSPIRATION 10 I § ANNUAL WATER USE (m) (From areas under curves) Maximum 0.67 Median 0.62 Minimum 0.57 WATER USE, IN MILLIMETERS PER DAY JAN. 1 MAR.1 MAY 1 JULY 1 SEPT. 1 NOV. 1 DEC. 31 FIGURE 11.-Estimated annual water use by replacement vegetation in the Pecos valley, New Mexico, as determined by the energy- budget- residual method. Estimates of annual water use based on the eddy- correlation measurements of evapotranspiration (not shown in fig. 11) were made in a manner similar to those for saltcedar. This analysis resulted in a water- use estimate for the replacement vegetation of 0.4 m/ yr, as listed in table 5. COMPARISON WITH PREVIOUS RESULTS In this study, the best estimates indicate that sum- mertime daily water use by saltcedar is as little as 1.0 mm/d or as much as 6 mm/d, but estimated rates in the range 4 to 5 mm/d are common (table 4). Annual water use is estimated to range from 0.77 to 1.07 m. These values are at the low end of the range of those measured by other investigators. For example, the minimum annual value coincides with the estimate of consumptive use by floodplain vegetation determined by Gatewood and others (1950) using the inflow- outflow technique. However, only about half the area for which they determined consumptive use contained saltcedar or other brush. Moreover, their analysis was based on 1 yr of record, and changes in soil-moisture and ground-water storage were not measured. Hence, their estimates could be in error, being either too large or too small. The median and maximum estimates of annual evapotranspiration compare closely with those Culler and others (1982) obtained using a water-budget tech- nique. They measured annual water use, including pre- cipitation, of about 1.0 m as an average for three WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. G29 A# I | I I | I | | | l | (9k. 123 cm‘ al >> sop 1, A | s --- NORMAL - \ ay 4 \ t o ang it * \ ~ PRECIPITATION, IN CENTIMETERS PER MONTH MAY JUNE JULY AUG SEPT OCT FIGURE 12.-Monthly precipitation at Roswell, N. Mex., during the period 1980-82. reaches that include some mesquite and bare ground. The climate of the Gila River floodplain in the vicinity of Safford, Ariz., where both of the above studies were conducted, is similar to that of the Pecos River flood- plain near Artesia. The latitudes are the same (33° N.), and rainfall is similar (292 mm at the Gila site versus about 320 mm at Roswell, N. Mex.). The altitude of the Gila site (about 770 m) is somewhat lower than that for the Artesia, N. Mex., site (about 1,000 m). On the basis of these comparisons, evapotranspiration rates might be anticipated to be slightly lower in our study area than in the Gila River study site. Our estimates of annual evapotranspiration may be somewhat low because most of the measurements were made during dry weather, partly by design and partly because equipment malfunctions induced by rain pre- vented measurements immediately afterward. Evapo- ration rates from wet canopies and wet soil are quite high but may be about equal for the saltcedar and the replacement vegetation. If the periods following rain had been sampled more representatively, our estimates of annual evapotranspiration would be at least a little higher. Monthly rainfall at Roswell, N. Mex., for the study period is shown in figure 12. Measurements of daily evapotranspiration using the Bowen-ratio method made by Gay and Fritschen (1979) and by Leppanen (1981) resulted in typical evapotran- spiration rates of 7 mm/d. These values are about 50 percent larger than those measured during our study. Leppanen's measurements were made over rapidly growing young saltcedar, which may not be compara- ble to mature stands. On the other hand, there is no obvious explanation for the difference between our re- sults and those of Gay and Fritschen. The Bowen-ratio measurements by Gay (1980) and the eddy-correlation measurements by Fritschen and others (1980) both indicate evapotranspiration rates of about 3 mm/d. However, their measurements were made on a cloudy day, and the results are not fully comparable to most of those in this study. The portion of the energy budget used in evapotranspiration during their studies was about the same as that which would result in an evapotranspiration rate of about 5 mm/d on a sunny day. Estimates of evapotranspiration based on G30 the temperature and humidity profiles, net radiation, and soil-heat-flux data presented by van Hylckama (1980) for a saltcedar thicket near Gila Bend indicate that daily evapotranspiration rates at that site are on the order of 8 mm/d, which is in better agreement with the results of the other Bowen-ratio studies than with our results. On the other hand, the analyses of van Hylckama's data are uncertain enough to preclude definitive comparisons. The estimates of daily evapotranspiration from saltcedar made during this study are one-fourth or one- fifth those measured in tanks in Arizona by Gatewood and others (1950) or by van Hylckama (1980), but are nearly the same as those measured in tanks at Bernardo, N. Mex. (U.S. Bureau of Reclamation, 1973). The discrepancies with the Arizona tank studies proba- bly arise from oasis effects on those tanks, as described in the section "Discussion of Previous Results." Energy-budget estimates of water use by replace- ment vegetation (fig. 11) are in the general range of those determined by other workers. Culler and others (1982, fig. 10) found that replacement vegetation in the Gila floodplain generally was 1-1.5 mm/d in April and May and about 2-2.5 mm/d during June through Au- gust. Their values are slightly lower than those found in this study. However, their vegetative cover was mainly poorly established bermuda grass, annual weeds, and bare ground. Also, many of their meas- urements were made immediately after clearing and before the replacement vegetation could become well established. Hence, the differences are easily accepted. Measurements by Fritschen and others (1980) and by Gay (1980) at the forbs site indicated that evapotran- spiration was about 3 mm/d in July 1979, which agrees with our results. Although several evapotranspirometer-tank studies of evaporation from bare soil and evapotranspiration from saltgrass, sacaton, and Russian thistle have been made, the depths to water are not comparable to those in this study and no comparisons are warranted. In summary, the energy-budget evapotranspiration measurements made during this study agree excep- tionally well with the results of the water-budget study conducted by Culler and others (1982), and with Bowen-ratio and eddy-correlation measurements made in the Pecos River floodplain by Gay (1980) and by Fritschen and others (1980). Agreement is less good with energy-budget studies conducted in other areas, and agreement with tank-measured evapotranspira- tion is very poor. The extensive water-budget measure- ments of Culler and others (1982) probably are the most representative of the entire floodplain environ- ment, and the good agreement of our results with those of that study give substantial credence to our results. STUDIES OF EVAPOTRANSPIRATION ESTIMATES OF EVAPOTRANSPIRATION SALVAGE Evapotranspiration salvage can be estimated as the difference in water use between saltcedar and the re- placement vegetation. Such estimates are, of course, fraught with uncertainty, because of the limited num- ber of measurements (17 for saltcedar and 10 for re- placement vegetation), the potential error in the meas- urements, and the difficulty in extrapolating the measurements in time and space. Nonetheless, the high, low, and median estimates for evapotranspira- tion provide probable bounds to the magnitude of evapotranspiration salvage. One set of salvage esti- mates can be obtained by comparing high, median, and low estimates for each cover type (table 5). These esti- mates suggest that the reduction in evapotranspiration might be as much as 0.4 m/yr or as little as 0.2 m/yr, with a median estimate for the energy-budget method of about 0.3 m/yr. These represent estimates of the maximum anticipated annual salvage. An estimate of the minimum anticipated annual water salvage can be obtained by subtracting annual water use by the replacement vegetation, as deter- mined by the eddy-correlation method, from the eddy- correlation-based estimate of annual water use by saltcedar. This analysis indicates that annual water salvage by saltcedar eradication is about 0.2 m. The above analyses indicate that annual water sal- vage by saltcedar eradication probably is more than 0.2 m but less than 0.4 m. Because the energy-budget residual or maximum values are considered more reli- able than the eddy-correlation values, a salvage esti- mate of about 0.3 m/yr is the best available. Alternate estimates of salvage can be obtained from the energy-budget values in table 5 by subtracting the minimum evapotranspiration for replacement vegeta- tion (0.57 m/yr) from the maximum water use estimate for saltcedar (1.07 m/yr) to determine an upper limit for water salvage of 0.5 m/yr. Similarly, the minimum esti- mate can be obtained by subtracting maximum water use by replacement vegetation (0.67 m/yr) from the minimum for saltcedar (0.77 m/yr). This minimum is 0.1 m/yr and tends to confirm studies of base flow gain. However, the validity of both these high and low esti- mates is suspect because the measured low saltcedar water use generally was coincident with measured low replacement-vegetation use. Hence, these alternate es- timates lack credence. A major assumption in the foregoing analyses of water salvage is that saltcedar at the measurement sites uses water at a rate typical of that formerly used by saltcedar in the now-cleared areas. Selection of sites for measurement of water use by saltcedar was limited, as most of the remaining saltcedar thickets in the flood- plain in the Acme-Artesia reach of the Pecos River WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. were too small to provide adequate fetch for micromete- orological measurements. Hence, it was necessary to select sites upstream and downstream from the cleared reach to obtain additional sample sites. Of all the sites, only the burn site had a depth to water typical of much of the cleared area, whereas all three replacement- vegetation sites had typical depths to water. However, the lack of any apparent relationship between depth to water and measured water use by saltcedar, coupled with the fact that the water-use rate at the burn site was intermediate among rates measured at the other sites, suggests that water use may have been ade- quately sampled. Use of water-salvage values to estimate the base- flow gain in the Pecos River between the Acme and Artesia gaging stations also is difficult. Although the USBR has cleared about 8,700 ha (21,500 acres) of saltcedar in the reach (Leon Marcell, written commun, 1983), Mower and others (1964, table 7) indicate that in 1958, 7,160 ha (17,900 acres) had an areal saltcedar density of greater than 10 percent and 5,280 ha (13,000 acres) had an areal saltcedar density of greater than 25 percent. Neither this study nor that of van Hylckama (1980) confirmed a linear relationship be- tween water use and area or volume density. However, it seems likely that water use by very sparse stands of saltcedar is substantially less than that from moderate to dense stands. The lack of transpiring material and the sparseness of the thickets suggest that the plants are not able to obtain large quantities of ground water. Hence, it may be reasonable to assume that water use by saltcedar of greater than 10-percent cover is equal to that measured, while that at less than 10-percent cover might be substantially smaller. However, the expected low use by the less than 10-percent cover might be compensated for by a large-scale oasis effect on the evapotranspiration from the remaining 15-m-wide strips of saltcedar along either side of the river. Also, the 30-m-wide strip of saltcedar along the 132-km-long river reach contains about 400 ha (1,000 acres) of saltcedar, all or most of which probably constituted dense stands at the time Mower and others (1964) pre- pared vegetation-density maps. Thus, about 6,800 ha (17,000 acres) of saltcedar of greater than 10 percent density may have been cleared. Such a clearing opera- tion would have resulted in roughly 1.2 x 10" to 2.5 10" m* (10,000-20,000 acre-ft) of salvage annually, based on the measurements described in this report. This value is less than the 3.5% 10" m (28,000 acre-ft) of salvage predicted by Mower and others (1964, table 14), but it is still sizable. The indicated 10,000-20,000 acre-ft of salvage is large enough that it should appear as a base-flow gain in the reach. The reason it has not been detected to date G31 may be that it is masked by large variations in the annual water budget. For example, saltcedar mowing appears to have been an ineffective means of control that had little impact on water use. The root plowing in 1974 coincided with an unusually wet year, resulting in large base-flow gains that year and some residual gains the next. Variations in ground-water pumpage and precipitation may have masked the instantaneous impact, and the overall base-flow gains that would be expected to occur for stable conditions actually may be compensated for by continuing declines resulting from ground-water pumpage in the shallow aquifer in the Roswell basin. SUMMARY AND CONCLUSIONS The eradication of saltcedar from 8,700 ha of the Pecos River floodplain between Acme and Artesia, N. Mex., did not result in base-flow gains in the reach that could be reliably related to saltcedar clearing. This suggests that consumptive use by the replacement veg- etation, mainly annual weeds and alkali sacaton, was about equal to that by saltcedar. To investigate that possibility, water use by selected saltcedar thickets and selected plots of replacement vegetation was periodi- cally measured using an eddy-correlation method. Measurements were made at approximately monthly intervals at two saltcedar sites and one replacement- vegetation site during the growing season in 1980 and 1981, and at four saltcedar sites and three replacement-vegetation sites in 1982. The eddy-correlation approach provides direct meas- ures of the sensible heat flux and the vapor flux, which may be converted to latent heat flux, from the thickets or plots. The sum of these fluxes should equal the en- ergy budget; however, the measurements frequently were only 60 to 70 percent of the energy budget on a daily basis. A second estimate of latent heat flux and evapotranspiration was obtained by subtracting sensi- ble heat flux from the energy budget. These estimates provided minimum and maximum values of water use from the two types of vegetative cover. Extrapolation of the measurements indicates that water use by the saltcedar thickets is between 0.6 and 1.1 m/yr, and by the replacement vegetation, 0.4 to 0.7 m/yr. The esti- mated probable range of water salvage by saltcedar eradication is 0.2 to 0.4 m/yr. Use of these estimates to compute the possible salvage for the entire reach is difficult because of the areal variability in hydrologic and soil conditions and in the density and vigor of the saltcedar prior to their eradication. However, salvage of 1.2-2.5x 10" m*/yr (10,000 to 20,000 acre-ft/yr) may have occurred, but is not recognized as base-flow gain because it is masked by declines in base flow related to ground-water development. G32 Measurements of evapotranspiration by saltcedar made during this study are in agreement with those of some other studies but are in wide disagreement with still others. In summary, measurements of water use by saltcedar for this study are in good agreement with those determined by the massive water-budget study by Culler and others (1982) and with measurements in the U.S. Bureau of Reclamation evapotranspirometers at Bernardo, N. Mex. (U.S. Bureau of Reclamation, 1973). Results of our study and that of Culler and oth- ers (1982, p. P39) indicate that evapotranspiration by saltcedar is substantially less than potential evapo- transpiration. Also, Hughes (1972) showed that the saltcedar evapotranspiration rates measured by the U.S. Bureau of Reclamation (1973) could be simulated by the Penman (1956) equation for potential evapo- transpiration only if a stomatal-resistance term was added. Results of this study indicate that evapotranspira- tion from different saltcedar thickets varied greatly, but no relationship among volume density of the plant cover, depth to water, and water use was apparent. The lack of dependency of water use on depth to water also was noted for the data from the evapotranspirometer tanks at Bernardo, N. Mex. (Hughes, 1972; U.S. Bu- reau of Reclamation, 1973). The lack of a linear rela- tionship between volume density and water use also was noted by Hughes (1972) and by van Hylckama (1974). RECOMMENDATIONS FOR FUTURE STUDIES OF EV APOTRANSPIRATION The main purpose of this study was to determine water use by saltcedar and by the vegetation that re- placed it in areas of the Pecos River floodplain from which it had been eradicated. A strong second motive among all of the authors was to test the eddy- correlation method for measuring evapotranspiration on the basis of newly developed equipment for making and recording the measurements. The method has many desirable attributes, including portability of the instruments, ability to make measurements at a single height above the canopy, and ability to check the fluxes by comparing them with the energy budget. In addi- tion, the conventional wisdom for many years has been that the eddy-correlation method would provide reli- able estimates of evapotranspiration once the instru- mentation problems were solved (Tanner, 1967, p. 545; van Hylckama, 1980, p. F13). Perhaps as a consequence of high expectations, our inability to close the energy budget, or even to improve closure in light of equipment modifications made dur- ing the study, has been very frustrating. Although we STUDIES OF EV APOTRANSPIRATION are confident that the eddy-correlation and energy- budget estimates of evapotranspiration flux do bracket actual water use, the range between the estimates is disappointingly large. Consequently, in future studies, estimates of evapotranspiration should also be obtained by the Bowen-ratio method. Battery-operated data loggers that provide user-programmable instru- ment control, such as the Campbell Scientific CR-21, CR-21X, or CR-7 data loggers, can be used to control a battery-operated device that switches the thermome- ter and psychrometer positions periodically (Simpson and Duell, 1984). Such a device is sufficiently portable to be used for the measurement of evapotranspiration at fairly remote wild land sites. Estimates of evapo- transpiration based on the Bowen-ratio technique would provide a useful adjunct to eddy-correlation- based estimates, or vice versa. A disadvantage exists in attempting to determine possible evapotranspiration salvage by saltcedar eradi- cation from a previously cleared area because of uncer- tainty about whether the remaining or offsite thickets are representative of those already cleared. Such un- certainty would be reduced by making measurements before and after clearing. However, climatic variability is large, even in the arid Southwest, so that it is desir- able to obtain measurements for a few years before clearing and for a few years after the replacement veg- etation is established. As it may take a few years to fully establish the replacement vegetation, a sequen- tial study might require several years to complete. Consequently, it is difficult to judge, even in hindsight, whether a sequential or a parallel approach is more desirable. Extrapolating measurements in time and transfer- ring them in space, even to nearby areas, requires that assumptions of uncertain reliability be made. The plant-soil-atmosphere modeling study now in progress, as preliminarily described by Weaver (1984), should provide insights into evaluating the effects of soil type, depth to water, ground-water and soil-water salinity, and plant vigor on evapotranspiration by saltcedar and replacement vegetation. These results in turn might be used to estimate evapotranspiration at other sites. REFERENCES Blaney, H.F., 1933, Water losses under natural conditions in wet areas in southern California, Part I, Consumptive use by native plants growing in moist areas in southern California: California Department of Public Works, Water Resources Division, Bul- letin 44, p. 19-139. 1961, Consumptive use and water waste by phreatophytes: American Society of Civil Engineers Proceedings, v. 87, no. 183, pt. 1, p. 37-46. WATER USE BY SALTCEDAR AND BY REPLACEMENT VEGETATION, N. MEX. Blaney, H.F., and Hanson, E.G., 1965, Consumptive use and water requirements in New Mexico: New Mexico State Engineer Tech- nical Report 32, 82 p. Blaney, H.F., Taylor, C.A., and Young, A.A., 1930, Rainfall penetra- tion and consumptive use of water, Part II, Evaporation and transpiration losses from moist areas: California Department of Public Works, Water Resources Division, Bulletin 33, p. 107- 146. Blaney, H.F., and others, 1942, Consumptive water use and require- ments, in [U.S.] National Resources Planning Board, Pecos River Joint Investigation-Reports of the participating agen- cies: Washington, U.S. Government Printing Office, p. 170-230. Brutsaert, Wilfried, 1975, On a derivable formula for long-wave radi- ation from clear skies: Water Resources Research, v. 11, p. 742- 744. Buck, Arden, 1976, The variable path Lyman-alpha hygrometer and its operating characteristics: Bulletin of the American Meteoro- logical Society, v. 57, p. 1113-1118. Campbell, G.S., 1977, An introduction to environmental biophysics: Heidelberg, Springer-Verlag, 159 p. Campbell, G.S., and Unsworth, M.H., 1979, An inexpensive sonic anemometer for eddy correlation: Journal of Applied Meteorol- ogy, v. 18, p. 1072-1077. Culler, R.C., Hanson, RL., Myrick, R.M., Turner, R.M., and Kipple, F.P., 1982, Evapotranspiration before and after clearing phreatophytes, Gila River flood plain, Graham County, Arizona: U.S. Geological Survey Professional Paper 655-P, 67 p. Eakin, M.E., and Brown, C.B., 1939 (rev. ed.), Silting of reservoirs: U.S. Department of Agriculture Technical Bulletin 524, p. 11- 18. Fritschen, L.J., 1965, Miniature net radiometer improvements: Jour- nal of Applied Meteorology, v. 2, p. 165-172. Fritschen, L.J., and Gay, L. W., 1979, Environmental instrumenta- tion: New York, Springer-Verlag, 216 p. Fritschen, L.J., Simpson, J.R., and Smith, M.O., 1980, Eddy- correlation measurements of evaporation from bare soil and of evapotranspiration from saltcedar groves in the Pecos River flood plain, New Mexico: Final Report to U.S. Geological Survey, Water Resources Division, Denver, Colo., 45 p. Gatewood, J.S., Robinson, TW., 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. Gay, L. W., 1980, Energy budget measurements of evaporation from bare ground and evapotranspiration from saltcedar groves in the Pecos River flood plain, New Mexico: Final Report to U.S. Geo- logical Survey, Water Resources Division, Denver, Colo., 18 p. Gay, L. W., and Fritschen, L.J., 1979, An energy balance budget anal- ysis of water use by saltcedar: Water Resources Research, v. 15, p. 1589-1592. Hughes, W.C., 1972, Simulation of saltcedar evapotranspiration: American Society of Civil Engineers Journal, v. 98, no. 1R4, p. 533-542. Jensen, M.E., 1973, Consumptive use of water and irrigation water requirements: New York, American Society of Civil Engineers, Irrigation and Drainage Division, 215 p. Leppanen, O.E., 1981, Evaporation from rapidly growing young saltcedar in the Gila River Valley of Arizona: U.S. Geological Survey Open-File Report 81-485, 31 p. McDonald, C.C., and Hughes, G.H., 1968, Studies of consumptive use of water by phreatophytes and hydrophytes near Yuma, Arizona: U.S. Geological Survey Professional Paper 486-F, 24 p. «U.S. G.P.0. 1987-181-407:60025 G33 Montieth, J.L., 1973, Principles of environmental physics: New York, American Elsevier, 236 p. Mower, R.W., Hood, J.W., Cushman, RL., Borton, RL., and Gal- loway, S.E., 1964, An appraisal of potential ground-water sal- vage along the Pecos River between Acme and Artesia, New Mexico: U.S. Geological Survey Water-Supply Paper 1659, 98 p. Penman, H.L., 1956, Evaporation: An introductory survey: Nether- lands Journal of Agricultural Sciences, v. 4, p. 9-29. Priestly, C.H.B., 1959, Turbulent transfer in the lower atmosphere: Chicago, University of Chicago Press, 130 p. Redford, T.G., Jr., Verma, S.B., and Rosenberg, N.J., 1980, Humidity fluctuations over a vegetated surface measured with a Lyman- alpha hygrometer and a fine wire thermocouple psychrometer: Journal of Applied Meteorology, v. 19, p. 860-867. Robinson, TW., 1965, Introduction, spread and areal extent of saltcedar (Tamarix) in the Western States: U.S. Geological Sur- vey Professional Paper 491-A, 12 p. 1970, Evapotranspiration by woody phreatophytes in the Humboldt River valley near Winnemucca, Nevada: U.S. Geolog- ical Survey Professional Paper 491-D, 41 p. Simpson, M.R., and Duell, L.F. W., Jr., 1984, Design and implementa- tion of evapotranspiration measuring equipment for Owens val- ley, California: Ground-Water Monitoring Review, v. 4, no. 4, p. 155-163. Suomi, VE., and Tanner, C.B., 1958, Evapotranspiration estimates from heat-budget measurements over a field crop: American Geophysical Union Transactions, v. 39, no. 2, p. 298-304. Tanner, C.B., 1963, Basic instrumentation and measurements for plant environment and micrometeorology: Madison, University of Wisconsin College of Agriculture, Department of Soil Science, Soils Bulletin 6, 216 p. 1967, Measurement of evapotranspiration, in Hagan, R.M., ed., Irrigation of agricultural lands: American Society Agron- omy Monograph 11, p. 534-574. Thomas, HE., 1963, Causes of depletion of the Pecos River in New Mexico: U.S. Geological Survey Water-Supply Paper 1619-G, 14 p. 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. van Hylckama, T.E.A., 1974, Water use by saltcedar as measured by the water budget method: U.S. Geological Survey Professional Paper 491-E, 30 p. 1980, Weather and evapotranspiration studies in a saltcedar thicket, Arizona: U.S. Geological Survey Professional Paper 491-F, 78 p. Veihmeyer, F.J., and Brooks, F.A., 1954, Measurement of cumula- tive evaporation from bare soil: American Geophysical Union Transactions, v. 35, no. 4, p. 601-607. # Weaver, H. A., 1984, A mechanistic model of evapotranspiration from saltcedar (Tamarix chinensis): Unpublished doctoral disserta- tion, Washington State University, Pullman, Wash. Webb, E.K., Pearman, G.I., and Leuning, R., 1980, Correction of flux measurements for density effects due to heat and water vapor transfer: Quarterly Journal of the Royal Meteorological Society, v. 106, p. 85-100. Welder, G.E., 1983, Geohydrologic framework of the Roswell ground- water basin, Chavez and Eddy Counties, New Mexico: New Mex- ico State Engineer Technical Report 42, 28 p., 28 figs. Young, A.A., and Blaney, HF., 1942, Use of water by native vegeta- tion: California Department of Public Works, Division of Water Resources Bulletin 50, 160 p. 2 U0 6 1 0 4.