Drought in the Southwest, 1942—56 GEOLOGICAL SURVEY PROFESSIONAL PAPER 372 Tflz's professz'ma/paper was printed as cflapters A—IJ, inclusive UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1963 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY Thomas B. Nolan, Director v5“ 22*”? \fi 3 V' QB); Fe 4, v. .33 Z23- CONTENTS sflms LIBRARY [Letters designate the separate chapters published or in press] (A) The meteorologic phenomenon of drdught in the Southwest, by H. E. Thomas. (B) General efiects of drought upon water resources, by J. S. Gatewood, Alfonso Wilson, H. E. Thomas, and L. R. Kister. (C) Eflects of drought in central and south Texas, by H. E. Thomas and others. (D) Effects of drought in the Rio Grande basin, by H. E. Thomas. (E) Efl’ects of drought in basins of interior drainage, by H. E. Thomas and others. (F) Efi'ects of drought in the Colorado River basin, by H. E. Thomas and others. (G) Effects of drought along Pacific Coast in California, by H. E. Thomas and others. (H) General summary of effects of the drought in the Southwest, by H. E. Thomas. 781 we W aflufi /‘L£{/@f [. o The Meteorologie Phenomenon of Drought in the Southwest GEOLOGICAL SURVEY PROFESSIONAL: PAPER 372—A (”£0415 kiwi ‘ KAI/<1 Mcgflm @575 flé 1/. 372. EARTH SCIENCES LIBRARY 1,3223! FEB 111964 EFFECTS OF WATER SURPLUS AND OF DROUGHT, ELEPHANT BUTTE RESERVOIR, NEW MEXICO Upper view, reservoir spilling for the first and only time, in the spring of 1942. Lower View, reservoir on September 4, 1951. High-water mark of 1942 is faint line above most conspicuous white watermark. The Meteorologic Phenomenon of Drought in the Southwest By H. E. THOMAS DROUGHT IN THE SOUTHWEST, 1942—56 GEOLOGICAL SURVEY PROFESSIONAL PAPER 372—A UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1962 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY Thomas B. Nolan, Director For sale by the Superintendent of Documents, US. Government Printing Office Washington 25, DC. ‘ PREFACE One prerequisite for this study is a definition of “drought” that is applicable to condi- tions in the Southwest, and yet not out of line with the general though varied usage of the term in other regions. Chapter A defines drought as a meteorologic phenomenon, and presents some of the published and recorded conclusions and ideas concerning the basic meteorologic factors that influence the patterns of precipitation in the Southwest. It also considers the characteristics of drought as indicated by meteorologic records. Another chapter will give a general discussion of the effects of the Southwest drought as shown by hydrologic data. Subsequent chapters will provide more detailed evaluations of the effects of drought in individual river basins and specific localities. The effects of drought are discriminated from water shortages due to other causes wherever possible. Because “drought” is a comparative term, a study of drought requires analysis of the longest records available, whether of precipitation, temperature, wind, streamflow, reservoir storage, use of water, or fluctuations of water levels in wells—with special attention to the drought period to see how it differs from other periods of record. Despite this special emphasis, a study of drought must be almost as broad in scope as general studies of the interrelations of' the several phases of the hydrologic cycle, such as the studies of rainfall and runofl" in the United States by W. G. Hoyt and others (1936). In many comprehensive studies of water resources, an important product has been a hydrologic equation, in which the volume of mean precipitation has been calculated, and its disposal accounted for in ground—water recharge, discharge, and changes in storage; in surface runoff and changes in storage; and in evapotranspiration or other consumptive use. In all such studies, emphasis has been on average conditions, and standard procedures have been developed, or at least suggested, for calculation of such items as mean or median runoff, duty of water, consumptive use, safe yield, and the like. By contrast, in a study of drought the emphasis has necessarily been on nature’s failure to achieve the average, and the evaluation of that deviation. The methods of study for this report have been of the old-fashioned qualitative type~peering at graphs and noting similarities and contrasts—and include very little statistical analysis of the rather meager data available. The wet year 1941 provided an excellent starting point for the drought in most of the Southwest, but as year followed year it became evident that drought visited some part of the Southwest practically every year, the only exceptions being wet years like 1941—and history indicates that very few of those occur in a century. The studies leading to this report cover especially the years 1942 to 1956, inclusive. The report includes maps showing the boundary of the area where drought was recorded in the 15 years 1949—56, but the boundary is pri- marily for the purpose of defining the limits of study and has been determined empirically. Areas outside that boundary have recorded severe droughts during some of these years. Personnel of the Geological Survey who were assigned specifically to analysis of data and preparation of this report on “Drought in the Southwest” include Joseph S. Gatewood and Alfonso Wilson of the Surface Water Branch; John D. Hem and subsequently Lester R. Kister of the Quality of Water Branch; Harold E. Thomas and George D. Scudder of the Ground Water Branch. Mr. Thomas was designated coordinator of the studies for the Water Resources Division. Other projects of the Water Resources Division throughout the South- west have been a major source of data and analyses pertinent to the drought. Credit for much of this work is shown in the body of the report, by quotations and references, and by authorship of certain sections. The supervisors of the Geological Survey district offices in the seven southwestern States have also contributed much to this report, in technical assistance III IV PREFACE whenever it was requested, and in basic data in files of those offices. These district supervisors include Francis M. Bell, Revoe C. Briggs, John H. Gardiner, Berkeley Johnson, Douglas Lewis, Stanley Lord, Wallace Miller, Trigg Twichell, Jack M. Terry, and Milton T. Wilson of the Surface Water Branch; Clyde S. Conover, William E. Hale, John W. Harshbarger, Omar J. Loeltz, Thad G. McLaughlin, Joseph F. Poland, Ray W. Sundstrom, Herbert A. Waite, and George F. Worts of the Ground Water Branch; Eugene F. Brown, John G. Connor, Burdge Irelan, Jay Stow, and Ishmael W. Walling of the Quality of Water Branch. Other important sources of data pertaining to drought have included the offices of the following State officials: Robert Ernst, and subsequently Obed Lassen, State Land and Water Commissioner of Arizona; Harvey Banks, Director of the California Department of Water Resources; Jean S. VVhitten, State Engineer of Colorado, and Ivan C. Crawford, Director of Colorado Water Conservation Board; Hugh A. Shamberger, State Engineer of Nevada and since 1958 Director of the Nevada Department of Conservation and Natural Re- sources; John Erickson, and subsequently Stephen Reynolds, State Engineer of New Mexico; R. M. Dixon, and subsequently Durwood Manford, Chairman of the Texas Board of Water Engineers; Joseph M. Tracy, and subsequently Wayne D. Griddle, State Engineer of Utah. CONTENTS Page Preface ____________________________________________ 111 Characteristics of recorded droughts in the Southwest—- Abstract ___________________________________________ A—1 Continued General introduction ________________________________ 1 Climatic fluctuations shown by weather records—— Droughts in general _________________________________ 2 Continued Drought in a humid region _______________________ 3 Analysis of records—Continued Page Drought in an arid region ________________________ 5 Graphic presentation of data _____________ A—18 Precipitation as a basic factor in drought __________ 5 Annual precipitation ________________________ 22 “Want of water” as a factor in drought ____________ 6 Drought in the Pacific Border zone _______ 23 Definitions used in this report ____________________ 7 Drought in the Great Plains zone _________ 23 Patterns of precipitation in the Southwest _____________ 8 Cyclic fluctuations __________________________ 23 Major airmass movements _______________________ 8 Drought in the Great Basin and Sonoran Airmass types in the Southwest ___________________ 9 Border zones _________________________ 26 Orographic influence ____________________________ 11 Seasonal precipitation _______________________ 26 Subdivisions of the Southwest ____________________ 11 Temperature ________________________________ 29 Deviations from average precipitation ______________ 14 Orographic effect upon precipitation ___________ 29 Characteristics of recorded droughts in the Southwest--- 14 Climatic fluctuations shown by other data _________ 33 Area of drought in 1942—56 ---------------------- 14 Historic data _______________________________ 36 Climatic fluctuations shown by weather records ----- 17 Lake levels --------------------------------- 36 Analysis of records -------------------------- 17 Fluctuations in tree growth _____________________ 38 Consistency of observed data _____________ 17 Summary of drought characteristics --------------- 40 Central tendency in precipitation _________ 18 References cited ------------------------------------ 41 ILLUSTRATIONS FRONTISPIECE. Effects of water surplus and of drought, Elephant Butte Reservoir. PLATE FIGURE Page Areas where annual precipitation was appreciably less than the long-term mean, 1942—56 ---------------- In pocket . Areas of dry and moist climate in the United States, excluding Alaska and Hawaii --------------------------- A—4 . Average trajectories of storm centers in the United States, excluding Alaska and Hawaii ---------------------- 10 Mean annual precipitation in the Southwest ----------------------------------------------------------- 12 Meteorologic zones in the Southwest ------------------------------------------------------------------ 13 Frequency of annual precipitation at six cities ---------------------------------------------------------- 16 Area of Southwest drought, 1942—56 ------------------------------------------------------------------ 17 Water-year precipitation at five cities in the Southwest -------------------------------------------------- 19 Cumulative monthly precipitation at Socorro and Santa Rosa, N. Mex., 1942—56 --------------------------- 20 Comparative graphs of water-year precipitation at four cities in the Southwest _____________________________ ‘ 21 Fluctuations of annual precipitation at five cities in the Pacific Border zone -------------------------------- 24 . Fluctuations of annual precipitation at eight cities in the Great Plains zone -------------------------------- 25 . Fluctuations of annual precipitation at six cities in the Great Basin and Sonoran Border zones --------------- 27 . Fluctuations of seasonal precipitation at six cities in the Southwest --------------------------------------- 28 . Progressive 10-year average October—March temperature at seven cities in the Southwest -------------------- 30 . Progressive 10-year average April-September temperature at seven cities in the Southwest ------------------- 31 . Snow accumulation in relation to altitude in selected years, southern Sierra Nevada _________________________ 32 . Cumulative departure of winter precipitation from 1930—49 mean, southern Sierra Nevada ------------------- 34 . Summary of findings pertaining to post-Pleistocene climates ______________________________________________ 35 . Fluctuations in precipitation and runoff, deduced from historic records _____________________________________ 37 . Fluctuations of level-in several lakes of the Southwest -------------- . ------------------------------------- 39 V DROUGHT IN THE SOUTHWEST, 1942—56 THE METEOROLOGIC PHENOMENON OF DROUGHT IN THE SOUTHWEST By H. E. THOMAS ABSTRACT The recent drought is one of sever'al which have been re- corded in the arid Southwest in the past century. In regions where precipitation comes chiefly from a single source, as in California and the Great Plains, prevailingly dry periods have alternated with wetter periods, each lasting 10 to 15 years. In the intervening area that includes the basins of the Colorado River and Rio Grande and numerous basins of interior drain- age, a deficiency from one source may be ameliorated in some localities by precipitation from another source; in other areas all sources have failed and the drought has continued for a quarter of a century. Earlier droughts of similar duration, and some of longer period, are indicated by historic records and by studies of tree rings, lake levels, and archeologic data. GENERAL INTRODUCTION In one of the earliest and most outstanding reports on the arid region of the United States, John Wesley Powell (1879, p. 3—5) points out both the prevailing moisture deficiency and the marked variations in water supply from time to time: 1 The Arid Region is the great Rocky Mountain Region of the United States, and it embraces something more than four-tenths of the whole country, excluding Alaska. In all this region the mean annual rainfall is insuflicient for agriculture, but in certain seasons some localities, now here, now there, receive more than their average supply. Under such conditions crops will mature without irrigation. As such seasons are more or less infrequent even in the more favored localities, and as the agriculturist cannot determine in advance when such seasons may occur, the opportunities afforded by excessive rainfall cannot be improved. The limit of successful agriculture without irrigation has been set at 20 inches, that the extent of the Arid Region should 1Concerning Powell’s report, Stegner (1954, p. 212) makes the fol- lowing analysis: “Embodied in the scant two hundred pages of his man- uscript—actually in the first two chapters of it—was a complete revolu- tion in the system of land surveys, land policy, land tenure, and farming methods in the West, and a denial of almost every cherished fantasy and myth associated with the Westward migration and the American dream ’ of the Garden of the World. Powell was not only challenging political forces who used popular myths for a screen, he was challenging the myths themselves, and they were, as rooted as the beliefs of religion. He was using bear language in a bull market, ‘deficiency terminology’ in the midst of a chronic national optimism well recovered from the panic of 1873.” 642666 0—62—2 by no means be exaggerated; but at 20 inches agriculture will not be uniformly successful from season to season. Many droughts will occur; many seasons in a long series will be fruit- less; and it may be doubted whether, on the whole, agriculture will prove remunerative. On this point it is impossible to speak with certainty. A larger experience than the history of agri- culture in the western portion of the United States affords is necessary to a final determination of the question. In fact, a broad belt separates the Arid Region of the west from the Humid Region of the east. Extending from the one hundredth meridian eastward to about the isohyetal line of 28 inches, the district of country thus embraced will be subject more or less to disastrous droughts, the frequency of which will diminish from west to east. For convenience let this be called the Sub-Humid Region. Statistics and. statisticians have been multiplying in the decades since Powell made this observation, and they confirm his judgment in all essentials. In the 80— year sample of climate that has now been recorded in the West, there has been a central tendency toward semiaridity in the Great Plains and extensive areas farther west, and toward even greater dryness in many lowland areas of the West. In most of the western half of the United States the mean precipitation is less than the potential evapotranspiration. (See p. A—3.) Countless plants and animals of the West are adapted to the varying degrees of dryness: for example, grasses thrive in the Great Plains where average precipitation is less than 25 inches; and cactus, creosote bush, yucca, and other xerophytes persist in deserts where the average annual rainfall is less than 10 inches. So far as the people are concerned, whether comparisons are made with the more humid eastern half of the United States or considering the water needs of ourselves and the ag— gregations of plants and animals that serve us, the western half of the country is recognized as one of pre- vailing water deficiency. It would probably not be too difficult to adapt ourselves to these climatic condi- tions if only the “average” climate at each locality were the climate actually experienced. But in most of the West there is a marked variation in precipitation from year to year. For example in the Great Plains, there A—l A—2 is a central tendency toward semiaridity but the rain- fall in one year may be as great as the average for Iowa, and in another as little as the average for southern Arizona. It is the deviation from average conditions that leads to difficulty in many parts of the West. This report is concerned primarily with these deviations from the averages as developed from available records, rather than with the averages. Specifically, we are concerned with the deviations in the southern part of the area west of the 100th meridian that have been be- low the average in recent years. Drought was a matter of great concern in the South- west in 1951. Although there were several localities Where the total precipitation that year was less than 50 percent of the long-term average, the concern was caused chiefly by the fact that so many of the years pre- ceding 1951 had also passed with less than average precipitation. The cumulative deficiencies of these years over extensive areas of the Southwest evidently had a pronounced eflect on the water supply in many localities. “The Drought in Southwestern United States as of October 1951” was the subject of a report by the US. Department of the Interior (1951), based on data col— lected and compiled by those bureaus within the de- partment which are concerned in one way or another with the water resources of the region: the Geological Survey, Bureau of Reclamation, Bureau of Land Man- agement, Bureau of Indian Affairs, and Fish and Wild— life Service. The report pointed out that the drought was considered to have begun in 1942 in Arizona, 1943 in New Mexico, 1945 in southern California, and 1947 or later in Texas; also that the drought as of 1951 was one of the eight most severe droughts that had occurred in the region since the 13th century. The con— tribution by the Geological Survey to that 1951 report was the basic hydrologic data on the drought (p. 11—29). Although those data were hurriedly assembled and analyzed, they gave some insight into the hydrologic phenomenon of drought; more important, they gave promise that a more comprehensive analysis of avail- able data might provide valuable information concern- ing drought, its causes and effects. After 1951, drought and the study of drought in the Southwest continued with varying diligence, both as to time and place. Southern California enjoyed a relatively wet year in 1952, sufficient to terminate the drought of the preceding 7 years; that drought has been described by Troxell (1957). However, 1953 was a year of somewhat less than average precipitation in southern California, and so were 1955 and 1956. For the remainder of the Southwest, each of the years 1952 to 1956 brought subnormal precipitation over extensive DROUGHT IN THE SOUTHWEST, 1942—56 areas, and the drought shifted from one area to another within the 7 southwestern States. Furthermore, it extended beyond the borders of those States in several years, notably into the northern Great Plains, where serious drought began in 1952 or earlier. This report summarizes existing information con- cerning drought in the Southwest, and especially the effects of drought upon the ground- and surface-water resources, based on studies chiefly during 1955 through 1957. Although the report is primarily concerned with drought subsequent to 1941 in the States of California, Nevada, Arizona, Utah, New Mexico, Colorado, and Texas, the characteristics of drought during these years are developed largely by comparison with other “dry” periods, and by contrast with intervening “wet” periods. DROUGHTS IN GENERAL “Drought,” and its alternative spelling “drouth,” trace their etymologies to the Anglo-Saxon “drfigath,” meaning “to dry.” The words are characteristically associated with the undesirable aspects of being with- out water. Dictionaries indicate a widespread and somewhat varied use of the word “drought”: (a) Dry weather, especially when so long continued as to cause vegetation to wither; want of rain or water; aridity. (b) Dryness of the throat and mouth for want of water; thirst. (c) Figuratively, scarcity of any necessity; dearth. (d) Dryness; also a desert. Drought is also discussed, or at least mentioned, in numerous scientific and technical writings pertaining to climate or to aspects of water supply. In these refer— ences, of course, drought is always indicative of dryness rather than wetness, but that is about as far as one can go toward a universal definition. Reflecting the inexactitude of popular definitions of the word, drought may indicate a moisture deficiency at a certain place over a period of time, or it may describe the condi— tions in one region as contrasted with another. The US. Weather Bureau (1953), after pointing out that drought conditions occur in one or more sections of the country nearly every year, states: A drought is usually defined as a “period of dry weather sufficient in length and severity to cause at least partial crop failure.” But rainfall, while an important criterion, does not give the complete picture, and a period of scanty rainfall that would be fatal to crops in one region might be sufficient for growth in another. Factors besides rainfall that intensify or mitigate drought effects are temperature, wind, evaporation, sunshine, character and conditions of soil, stage of crop de- velopment, etc. Although the term “drought” is used most widely with respect to precipitation and soil moisture, it is applied also to deficiencies in other water supplies. And because drought also means “want of rain or THE METEOROLOGIC PHENOMENON OF DROUGHT water,” many writers consider adverse effects upon man or his activities as a necessary condition of drought. Thus technical definitions of drought vary rather con- siderably, depending upon the technologist’s hydrologic environment and special field of interest. In particular, the concepts of drought developed in humid regions differ in many respects from those developed in arid regions. As stated by Thornthwaite (1948, p. 55—56, 75) : We cannot tell whether a climate is moist or dry by knowing the precipitation alone. We must know whether precipitation is greater or less than the water needed for evaporation and transpiration * * *. The vegetation of the desert is sparse and uses little water because water is deficient. If more water were available, the vegetation would be less sparse and would use more water. There is a distinction, then, between the amount of water that actually transpires and evaporates and that which would transpire a‘nd evaporate if it were available. When water sup- ply increases, as in a desert irrigation project, evapotranspira— tion rises to a maximum that depends only on the climate. This we may call “potential evapotranspiration,” as distinct from actual evapotranspiration * * *. Where precipitation is exactly the same as potential evapo- transpiration all the time and water is available just as needed, there is neither water deficiency nor water excess, and the climate is neither moist nor dry. As water deficiency becomes larger with respect to potential evapotranspiration, the climate becomes arid; as water surplus becomes larger, the climate becomes more humid. Thornthwaite’s areas of moist and dry climates are delineated on figure 1, which shows also the areas of moist climate in which there is a seasonal (summer) Water deficiency; the map is based on long-term aver- age precipitation and temperature at more than 3,000 Weather Bureau stations. DROUGHT IN A HUMID REGION Several scientists define drought in a humid region on the basis of deficiencies in soil moisture. As stated by J. C. Hoyt (1938, p. 1) : When in an area that is ordinarily classed as humid, natural vegetation becomes desiccated or defoliates unseasonably and crops fail to mature owing to lack of precipitation, or when precipitation is insufficient to meet the needs of established human activities, drought conditions may be said to prevail. This is elaborated by Thornthwaite (1947, p. 88) : Drought is most accurately described as a condition in which the amount of water needed for transpiration and direct evapo- ration exceeds the amount available in the soil. It results from too little rain. Soil moisture is used up, and plants then suffer from lack of water *1 *. It is evident that we cannot define drought as a shortage in rainfall alone. Such a definition would fail to take into account the amount of water needed. Furthermore, the efiect of a shortage of rainfall depends on Whether the soil is moist or dry at the beginning of the period. Shantz (1927) explained that drought in its proper sense is A—3 related to soil moisture and that it begins when the available soil moisture is diminished so that the vegetation can no longer absorb water from the soil rapidly enough to replace that lost to the air by transpiration. Drought does not begin when rain ceases but rather only when plant roots can no longer obtain soil moisture. As early as 1906, Henry pointed out that the intensity of drought could not be measured as a departure of rainfall from the normal, “since a deficiency of 50 percent in a region of abundant rainfall is not so serious as the same deficit in a region where the average precipitation is barely sufficient for the needs of staple crops.” Accurate identification of drought conditions under these definitions requires a detailed water budget such as Thornthwaite (1953) maintained at Seabrook, N .J . This detailed work shows that hidden drought, which brings about a borderline soil-moisture deficiency, of- tentimes holds crop yields to as little as a third of the potential. Because of the difficulty of similarly keeping water accounts for extensive areas, drought is usually defined as a period of consecutive days without rainfall. Ac- cording to Henry ( 1930) , a drought exists whenever the rainfall for a period of 21 days or longer is but 30 percent of the average for the time and place, and this is approximately the definition used by the Tennessee Valley Authority. The British Rainfall Organization defines a “partial drought” as a period of more than 28 days with a very small rainfall per day, and an “absolute drought” as a period of at least 15 consecutive days to none of which is credited as much as 0.01 inch of rain. In previous studies of droughts in the United States by the Geological Survey, it has been customary to identify the areas of drought on the basis of meteoro- logic records, as stated by J. C. Hoyt (1938, p. 2) : Although deficiency in precipitation is the prime cause of drought, it is not possible to set for any region an exact limit of the total annual precipitation above which a drought does not exist and below which a drought may prevail. In general, however, in the humid and semiarid States there are no serious drought effects unless the annual precipitation is as low as 85 percent of the mean—that is, unless there is an annual de— ficiency of 15 percent or more. This limit is used in the present report as a measure of a drought year and may serve in many drought studies. However, annual precipitation is not a sufficient cri- terion for all droughts. A severe drought may develop because of deficient precipitation during the growing season, even though precipitation in other months may be enough to bring the annual total up to average. Conversely, even though the annual precipitation is well below average, drought may not be recognized if rainfall in the growing season is average or above, because the evapotranspiration is so little in the rest of the year that precipitation less than average may still exceed the need. Thus the distribution of precipi- DROUGHT IN THE SOUTHWEST, 1942—56 A—4 mi: >mfgfisaoafl S65 333.4 .fifiBam add dawn: “33305 $35 USED Q5 5 SEE? «22: 65“ >5 H6 334 A 556E THE METEOROLOGIC PHENOMENON 0F DROUGHT tation in the year may also be a determining factor in drought. Most reported droughts in humid regions affect man- kind chiefly during the growing season of a single year; the resulting soil—moisture deficiencies are replenished by the precipitation of the following winter, and the drought receives a designation such as “drought of 1936,” or “drought of 1930.” A drought in the north- eastern States in 1949 was exceptional among droughts in humid regions because it extended beyond the grow- ing season. No rain was recorded in New York City in the 22 days beginning May 27, and the total rain during June was only 0.16 inch, or 5 percent of the long-term average for the month. This deficiency affected the crops that depended upon rainfall for maintenance of their soil-moisture requirements, a typical feature of drought in the region. But the precipitation continued to be less than usual: in the 8 months ended January 1950 it was less than 65 percent of the long-term average for those months. Discharge from springs and stream- flow diminished, with marked effect upon numerous municipal and industrial water supplies, including es- pecially New York City’s Croton and Catskill reser- voirs. Although this drought was of lesser intensity than several earlier droughts, it caused more trouble to the people of the metropolitan area. (See p. A—7.) One may surmise at this point that the emphasis on soil-moisture deficiency in the growing season, as a prime measure of drought in a humid region, is to be expected wherever surface— and ground—water resources are habitually more than sufficient for man’s domestic and industrial needs. But as man increases his devel— opment and utilization of water resources, he will be- come progressively more concerned with deviations below the average, regardless of the season or the source of supply. DROUGHT IN AN ARID REGION An arid region year after year receives precipitation insufficient for the requirements of most crops, and those crops therefore require irrigation. From the viewpoint of a humid region, this is “perennial” Or “permanent” drought. From the Viewpoint of the arid region, how- ever, it may be questioned whether the term “drought” is applicable when the dearth of precipitation is usual, or “normal,” for this dearth is embraced in the defini- tion of aridity that characterizes the region. The na- tive vegetation is well adapted to the climatic norm— the plants in the driest areas subsist on very little water and may remain dormant for months and even years without water. Man, too, can adapt himself and his culture to the climatic norm by irrigation of crops to supply their water requirements throughout the grow- A—5 ing season. If “want of water” were the basic criterion of drought, irrigated lands might well be drought-free, even though the entire growing season were rainless. Nevertheless, the inhabitants of the arid regions are cognizant of drought. At first glance it would appear that their concept of drought must be vastly different from that of the inhabitants of humid regions, for in recognition of the characteristic inadequacy of precipi— tation upon their lands, they have made themselves in— dependent of it by irrigation. But the water needed for irrigation, as well as that required for domestic, municipal, and industrial use, is also derived ultimately from precipitation somewhere, and variations from the average in precipitation and temperature are of critical importance to all water users. Such variations are as evident in arid as in humid regions, and may be far greater percentagewise. PRECIPITATION AS A BASIC FACTOR IN DROUGHT The arid and humid regions depicted on figure 1 have much in common. In both regions, precipitation constitutes the ultimate source of the fresh—water sup- plies that are useful to man. Except for the water that evaporates at the surface, the soil has top priority upon the water that falls as precipitation. Overland runoff does not occur unless or until the precipitation exceeds the capacity of that surface layer to absorb water. The soil holds water until its field capacity is reached, and any additional water absorbed by the soil moves downward by gravity. In the intervals between storms soil moisture may be depleted by evaporation and transpiration, and this depletion must be made up during subsequent storms before there can be additional downward percolation. The water that percolates downward through the soil enters ground- water reservoirs, and water moves in them by grav- ity to be discharged ultimately into lakes or streams or oceans, or at the land surface by springs or by evapo- transpiration. Streams have the lowest priority on the water that falls as precipitation, for water enters a stream only if it falls directly into the channel, or if it cannot get into the ground by infiltraton, or if it is subsequently discharged from subsurface sources. Even after water has reached a stream it may be lost by evapotranspiration or it may disappear by seepage into underlying ground-water reservoirs. The great differences in soil—moisture resources, ground-water resources, and streamflow characteristics in various parts of the country are traceable not only to variations in precipitation and other climatic factors, but also to differences in the soil and underlying mate- rials through which water may pass. From any single source—whether it be stream, reservoir, lake, well, ’A—6 spring, seep, infiltration gallery, cistern, or soil—varia- tions in available supply are generally caused by variations in precipitation, unless the natural storage facilities for water are changed by man. Droughts re— sulting from less than average precipitation, if con- tinued long enough, will ultimately affect all these sources of water, and as Thornthwaite has pointed out, there is no need to classify droughts as climatic, biologic (or agricultural) and hydrologic (or water-supply). In the eastern half of the country (fig. 1) precipi— tation generally exceeds evapotranspiration and thus provides a surplus that moves toward the ocean in streams or underground. By contrast, the western half of the country is prevailingly a region (unshaded part of fig. 1) where there is little or no water surplus in any season. This region, of course, includes a wide variation in moisture conditions—from hardly any precipitation at all to sufficient precipitation for certain grasses and grains, for native vegetation suitable for forage, and for dry farming. Of major importance to the region, however, are the small isolated areas (shaded areas within the unshaded part of fig. 1) that produce water surpluses seasonally or throughout the year. These are the humid islands—or highlands—to which the inhabitants of the arid region look for water to meet their requirements. The western United States, then, is not a uniformly arid region, but a composite of arid and semiarid low- lands and humid and subhumid mountains and plateaus. The mountains and plateaus are sparsely inhabited because of ruggedness, poverty of soil, or short grow- ing season, but the precipitation upon them may be enough for a forest cover and a thick snow cover in the winter. The inhabitants of the fertile but arid low- lands look to these highlands for most of the water they use for irrigation as well as for municipal and industrial supplies. Because of the water flowing from these highlands both in streams and underground, the lowlands may have a relatively large perennial water supply, even though the lowland precipitation may be nil during the growing season, and in places very little at any time of the year. “WANT OF WATER” AS A FACTOR IN DROUGHT The definitions of drought quoted on page A-2 sug- gest that the water needs for established human activi- ties are an essential criterion for drought conditions. W. G. Hoyt (1942, p. 580) states: In connection with present-day activities of a highly-organized civilization, it is therefore increasingly difficult to define and delineate droughts on the basis of a study of meteorological and hydrological conditions alone. DROUGHT IN THE SOUTHWEST, 1942—56 In descriptions of droughts as thus defined, the meteorologic phenomenon may play only a secondary role, and the existence of drought depends chiefly upon man’s activities. Thus Troxell (1957, p. 7, 29) says: To the city dweller, as well as to the agriculturist, a drought exists whenever he is required to reduce his water uses. In light of the foregoing definitions, the frequency of a drought will depend largely on the amount of the water reserves in relation to the region’s annual water requirements. If these reserves are large in terms of the water requirements, then many a period of deficient precipitation may pass without any curtailment of the water requirements, and the droughts may go unnoticed as meteorological phenomena. However, there are many less fortunate areas where the reserves are small in terms of the water requirements. In these areas each such period of deficient precipitation is recognized as a drought. * * * as man‘s water requirements increase, the adverse effect of the drought on his economy also increases. And fre- quently these economic features form the principal index of the drought’s magnitude in the mind of the public. Some droughts are described as resulting entirely from causes other than meteorologic, as is brought out by Brooks (1949, p. 169) in reporting the desiccation of South Africa: In the past fifty years the country has been suffering increas- ingly from drought, but the conclusion from expert evidence is that this is not due to an actual decrease in the amount of rain- fall, but to a change in the nature of the soil and vegetation. When South Africa was first settled, the country was covered by a rich vegetation, the rainfall was steady and persistent, and a large proportion of it was absorbed. The eflect of over-pas- turage has been to destroy much of the protective vegetation, and the soil has been washed away or trampled hard. The temperature contrasts have been increased owing to the heating effect of the sun on the patches of bare ground, and the rain now falls largely in heavy “instability” showers, including destructive thunderstorms. The run—off is proportionally greater, owing to the more torrential nature of the fall and the loss of the vegetation, so that with nearly the same rainfall the amount of water available for use has decreased. Drought would be of little interest to a man if it had no bearing upon his activities, and it becomes of progressively greater concern as it cuts into the water supplies needed by him. Obviously “want of water”— or water requirement—is overwhelmingly important, for it is a basic reason for man’s interest in any aspect of water resources. However, from a brief outline of the water situation in the arid parts of the United States (p. A—5) the difficulty of defining drought in terms of water “deficiency” is apparent. Such a defini— tion might be possible in terms of a single crop covering a small area, but in any sizable area, even within a radius of a few miles, water deficiencies might be very pronounced at one place, not apparent at another, and overcome at a third. The chief difficulty in using water requirement as a factor in the measure of drought is that it easily becomes THE METEOROLOGIC PHENOMENON OF DROUGHT paramount in the minds of men, and a water shortage may be labeled a “drought” regardless of its cause and, in particular, regardless of the meteorologic back- ground. In terms of water requirements, precipitation alone is never satisfactory for man, beast, or plant: precipitation everywhere is intermittent and irregular and the water must be stored somehow if it is to meet the needs of any form of life. Therefore, any definition of drought based on water requirements must embrace an evaluation of these storage facilities—in soils, aqui- fers, lakes, reservoirs, and streams—facilities which may be modified by man’s activities and in any case are problems separate and distinct from the basic meteoro- logic factors in drought. Thus, the city of New York would not have been affected by drought in 1949 if the storage facilities then under construction in the head- waters of the Delaware River had been in operation. Lesser demand or greater storage facilities would not have altered the basic meteorologic factors responsible for the drought, but would have made a considerable difference in the effect of the drought upon the city of New York. As another example, the Imperial Valley of southern California is within the southwestern drought area as described in subsequent sections of this report, a broad area in which precipitation has been substantially below the long-term mean for several years. However, Imperial Valley is dependent not upon local precipitation but upon the Colorado River for its water, and since Hoover Dam was completed the storage facilities have been ample to meet the valley’s require- ments. Thus Imperial Valley feels no effects of the drought that engulfs it. DEFINITIONS USED IN THIS REPORT Summarizing the preceding sections, the term “drought” has been commonly applied, rather incon- sistently, to three major forms of dryness: (a) a natural condition caused by less than average precipitation over a certain period of time; (b) a natural condition where the average precipitation is low; and (c) Nature’s fail— ure to fulfill the wants or to meet the developed require- ments of man. The first of these definitions is used in this report— that is, drought is a meteorologic phenomenon and oc- curs during a period when precipitation is less than the long-term average, and when this deficiency is great enough and continues long enough to hurt mankind. Drought is thus measured in terms of the duration and magnitude of the departure from the average climate in the area under consideration. Human bias cannot be eliminated from the physical phenomenon, for the records that delineate the physical conditions are avail- able because of the interest of mankind—as shown for A—7 instance by the fact that we have no basis for evaluating or studying drought in the Antarctic, where there is no one to be affected. Studies of drought are justified chiefly because of man’s interest in the phenomenon in relation to his “want of water.” The dryness of a region with a very low average pre- cipitation is termed “aridity.” Drought occurs in arid regions, as it does in regions with a higher average pre- cipitation, only when the precipitation is significantly less than the long-term average. “Water shortage” de- scribes the conditions where the water requirements of man or his crops or industries are greater than the avail- able supplies. Drought may be a contributing factor, or even a predominant factor, in shortage of water sup- plies, but some shortages can be credited entirely to man. The effects of drought are measured in the various sources of water—soils, lakes, streams, and surface and underground reservoirs—upon which man depends for his supplies. Depending upon the extent to which a drought reduces these supplies with respect to the de- veloped demand, the effects of drought may be mild, moderate, or severe. By first considering the meteoro— logic phenomenon without regard to the effects of drought, it is possible to evaluate and compare climatic trends without introducing complications that reflect the varying degree of water-resources development in various localities. Inasmuch as full development of the water resources in a drainage basin is planned on the basis of the long—term average precipitation and stream— flow and ground-water recharge, comprehensive analy- sis of climatic fluctuations, independent of factors re- lated to the stage of development, is an important ele- ment in planning. On the basis of a great variety of data it is pos— sible to identify, or at least infer, climatic fluctuations with great ranges in both duration and magnitude (p. A—17 to A—40). Of the drier periods that'are shown by these records, some are measured in days, others in months, seasons, years, decades, centuries, millennia, and even geologic epochs. Not all these dry periods, however, can be called droughts, if the term “drought” is limited to the periods when the precipitation defi- ciency is great enough and continues long enough to hurt mankind. In humid regions, a rainless period of several con- secutive days is significant for crops dependent entirely upon soil moisture, and is logically termed a drought (p. A—3 to A—5) ; in a study of such droughts, therefore, the focus is upon climatic fluctuations measured in days or weeks. Throughout the West, however, the develop- ment of ground- and surface-water resources has created independence from day-to-day fluctuations in precipitation; instead the focus is especially upon the A—8 annual and longer climatic fluctuations that affect those ground- and surface-water resources. There is a clue from prevailing usage that the term “drought” reflects the relative insecurity of mankind in the face of a natural phenomenon that he does not understand thoroughly and for which, therefore, he has not devised adequate protective measures. A West- erner does not call a rainless month a “drought,” and a Californian does not use the term even for an entire growing season that is devoid of rain, because these are usual occurrences and the developed water economy is well bolstered against them. Similarly, a dry period lasting several years, or even several decades, would not qualify as a drought if it caused no hardship among water users. At present, however, we do not under— stand these long—term fluctuations or their causes sufli- ciently to anticipate them, and our corrective measures have generally not yet accomplished a perennially stable water economy. Dry periods of several years’ duration, therefore, still create a feeling of insecurity, and are properly designated “droughts.” This report is concerned chiefly with dry periods of several years’ duration, which are the ones that affect markedly the resources of surface water and ground water. Dry periods measured in centuries are also of scientific interest in the Southwest, but most of the available records are not long enough to cover periods of such length. Here it is necessary to make some ground rules for this report, because the term “drought” as defined is dependent not only upon climatic factors but also upon the population, its requirements for water, and its capabilities in regulating and utilizing the water resources. In the discussions of specific regions, and of the numerous components of those regions (sub- sequent chapters) the effects of climatic fluctuations upon the established civilization can be recognized, and drought can then be described quantitatively. How— ever, regional comparisons throughout the Southwest, and comparisons of the most recent drought with ear- lier droughts, are based upon the evidence of climatic fluctuations and their effect upon the water resources. PATTERNS OF PRECIPITATION IN THE SOUTHWEST If drought is defined as a meteorologic phenomenon, some understanding of basic meteorology is necessary before droughts can be discussed intelligently, although adequate treatment of this subject is quite beyond the scope of this report. The following sections provide merely a brief summary, with numerous references to more detailed discussions. In particular, the reader is referred to Tannehill (1947), for a comprehensive and very readable exposition of the causes of drought. DROUGHT IN THE SOUTHWEST, 1942—56 MAJOR AIR MASS MOVEMENTS As pointed out by Tannehill (1947, p. 54) : The sun heats the continents more in summer than in winter. The oceans change temperature more slowly than the con- tinents. There is nearly always a powerful temperature contrast between the hemispheres, between the oceans and continents, and between the tropics and poles. This keeps the atmosphere in motion and brings moisture to the interior of the continents. The contrast between solar heating in the tropics and at the poles has a global effect upon atmospheric circulation, to which we can attribute some of the aridity that characterizes southwestern United States. The heating in the tropics causes expansion, decreasing density, and rising of air—most pronounced north of the equator in July, and south of the equator in J anu- ary. The air tends to pile up at the horse latitudes (30° N. and S.) with resulting high pressure and gen- erally descending winds, but temperature contrasts and air circulation between continents and oceans break up these high-pressure belts in many places. From the highvpressure areas along the belt at 30° N. latitude, part of the air flows back toward the equator, deflected to the right by the earth’s rotation to form the north- east trade winds; part continues north, also deflected to the right, and constitutes the prevailing westerlies that are dominant in most of the United States. The contrasting temperatures of oceans and con— tinents are a dominant factor in the circulation of air over the United States. The Pacific Ocean in particu- lar, because of its size and the prevailing west-to-east direction of air movement identified by Tannehill (1947, p. 78—87) as “The Monster in the Backyard,” controls especially the weather throughout western United States but probably exerts considerable influence also farther east. The temperature contrast between ocean and continent results from the fact that the water tem- peratures change very little from winter to summer in comparison with the land temperatures. The con- trast either in summer or in winter is greater in Can- ada than in the United States. In winter, because of the cooling of the northern part of the continent and the resulting greater density of air in comparison with that over the ocean, there is rapid eastward move- ment of air from the ocean. The air moving on east— ward over the Atlantic becomes warmer and piles up. at high levels, so that the eastward outflow of upper atmosphere from the continent is less than the inflow from the Pacific. As the pressure builds up over Canada, cold air is forced southward into the United States, obstructing the flow of warm air from the south. In summer, when the continent is warmer, conditions are reversed: air moves slowly into the continent from the Pacific and more rapidly out of the continent to THE METEOROLOGIC PHENOMENON OF DROUGHT the northeast. The resulting decrease in pressure over Canada induces northward movement of warm air across the United States from the Gulf of Mexico. The temperature contrasts between equator and poles, and between oceans and continents, would cause move- ments in the atmosphere regardless of its composition, so long as it is a gaseous envelope around the earth. So far as precipitation is concerned, however, the im— portant item in the atmosphere is the water vapor, even though the other gases make up 96to 99.98 per- cent of the total. The degree of saturation of the air near the land surface indicates whether the atmosphere can absorb more water (if available) by evaporation, or drop some by condensation or precipitation. The requirements for precipitation include the pres- ence of condensation nuclei and the cooling of an air mass sufficiently to change some of the water vapor into liquid or solid. Clouds may form as a result of the cooling of rising air, and local showers may result if there is sufficient condensation. In the larger air- mass movements mentioned earlier, the cold air that descends and becomes warmer at latitudes about 34° N. and S. does not cause precipitation; most of the world’s deserts, including those in southwestern United States, occur in these belts. In winter the cold air flowing southward from Canada obstructs the warm moist air flowing northward from the Gulf of Mexico; the resulting movement of the warm air up and over the cold air causes rain or snow in the United States. In summer, also, the warm moist air moving from the Gulf of Mexico is an important source of precipitation as it moves northward toward cooler latitudes. Mountains form barriers to air circulation; they force the air to rise and become colder, and thus they may be major factors in producing precipitation. In partic- ular, the mountain chains, which extend north-south in the western part of North America, wring the mois- ture from the air moving eastward from the Pacific Ocean, and thus they receive heavy precipitation on their western slopes. For this reason the air accumulat- ing over Canada in winter is generally dry (Tannehill, 1947, p. 89—96). The Great Basin and Colorado Plateaus, which in- clude Nevada and Utah and parts of Colorado, New Mexico, Arizona, California, Oregon, and Idaho. are arid or semiarid because of mountain barriers formed by the Sierra Nevada and the Cascade Mountains on the west, the Mogollon Rim on the south, and the Rocky Mountains on the east. Particularly in the autumn and winter when the land is cooling, cold dry air accumu- lates in this basin to form a persistent high-pressure area, which feeds dry air into the Southwest (Tanne- hill, 1947, p. 97—106). 642666 0—62—3 A—9 Water vapor is itself a dynamic factor in air circula- tion. Because of its lower density, it has less weight than an equivalent volume of air. Thus, the greatest concentrations of water vapor form areas of low pres- sure. Other air moves toward these low-pressure areas, for pressure gradients are important factors in air movements. AIRMASS TYPES IN THE SOUTHWEST From the brief discussion above of major air mass movements, it is possible to deduce the principal source regions of air contributing to the climate of the South— west. Thornthwaite and others (1942, p. 4) list the following types of airmasses: (a) Cool moist Polar Pacific, from the northern Pacific Ocean; (b) warm moist Tropical Pacific, from the southern Pacific Ocean; (0) warm moist Tropical Gulf, from the Gulf of Mexico; ((1) cold dry Polar Continental, from Can— ada; (e) hot dry Tropical Continental, from Mexico. The weather at any given place may be influenced in many ways by these masses, but some types of move- ment predominate in certain seasons. The following descriptions of the predominant movements of these masses into the Southwest are based largely upon stud- ies by the California Institute of Technology, Depart- ment of Meteorology (1943). The Polar Pacific type (fig. 2) brings moisture to the southern half of the United States chiefly in win- ter, when there is strong southward movement of cold dry air in western Canada. Typically, low-pressure areas are generated along the coast of British Columbia, move southward either along the coast or inland across Oregon, Nevada, and Utah, or perhaps as far south as northern Arizona and New Mexico, before trending eastward across the Great Plains and northeastward over the Great Lakes. By summer the average trajec- tories of low centers migrate northward to the northern tier of States. The Tropical Pacific type invades the Southwest infrequently, but when it does it may result in very heavy precipitation. The type is generated between Hawaii and southern California, commonly moves east- ward across southern California, Arizona, New Mexico, ‘and north Texas, and continues northeastward. In most winters not more than one or two major storms are identified as of the Tropical Pacific type, but there were eight of them in 1941, which was one of the wettest years on record in the Southwest. The Tropical Gulf type brings moist air from the Gulf of Mexico into the Southwest, especially into the Great Plains east of the Rocky Mountains and also into western New Mexico and Arizona. These masses are the principal source of precipitation throughout the DROUGHT IN THE SOUTHWEST, 1942—56 A—10 .mwma SMBPSREE no 325.3an .hwomodaowfi we 353mg 598535 80.5 “33°de .fiuBam SE «€24» mica—~83 .moufiw USED 35 5 23:3 :2on «o moioaownsb wwagwiwltm BEER .%//// THE METEOROLOGIC PHENOMENON OF DROUGHT year in much of the Great Plains, as well as in other parts of New Mexico. In the summer these airmasses are the chief scurces of precipitation in most of the Southwest. Dry Polar Continental air may invade the South— west from the north or northeast during the winter. Such invasions produce a persistent high-pressure area and sustained periods of cold weather in the Great Basin. The remaining airmass type, the Tropical Continental, is also dry, but hot; it is particularly prev- alent in Arizona and New Mexico during late spring and early summer. A study by Horn, Bryson, and Lowry (1957), which utilizes the Fourier analysis for the description of the annual march of precipitation over the United States, outlines the situation in the Southwest as one of tran- sition betWeen broad regimes. The Pacific winter rains extend eastward in diminishing relative importance across Arizona and New Mexico. The Gulf summer rains sweep northwestward rather suddenly around July 1 to cover Arizona and New Mexico, but diminish northward in relative importance and fail to reach west- ward to coastal California. Over Arizona, southern Utah, and southern Nevada those two regimes are of similar importance and result in a dominant semiannual variation of rainfall, somewhat more in winter in most of Utah, Nevada, and western Colorado, somewhat more in summer south of a line through Elgin, Nev., Cedar City, Utah, and Gunnison, Colo. Subsequently, Bryson (1957) made a more detailed analysis of the annual march of precipitation in Arizona, New Mexico, and northwestern Mexico, and concluded that: an area consisting largely of the Sierra Madre Occidental in northwestern Mexico, and the portion of Arizona southeast of Tucson constitute a single rainfall province with a strong sum- mer maximum of rainfall. This province also has a winter max- imum but only in Arizona does the semiannual term exceed the annual in amplitude. Within the United States the Gila and Rio Grande valleys constitute rainfall provinces of inter- nally similar annual march, while the upland areas tend to resemble the Pacific coastal pattern to the west. Recent studies by Eugene Peck of the US. Weather Bureau show that in southern Utah also the annual march of precipitation in upland areas is difl’erent from that in adjacent lowlands. OROGRAPHIC INFLUENCE The average annual precipitation at Weather Bureau stations has been the basis for various maps showing mean precipitation by isohyets (lines connecting places havmg equal precipitation) orby zones such as 0—10 inches and 10—20 inches. For the country as a whole the most recent map of mean precipitation was pub— lished by the US. Weather Bureau (1957), based on the A—11 30—year period 1921—50. Figure 3, adapted from this map, shows the close relation of precipitation to topog- raphy in the Southwest: the highest average precipi- tation occurs on the high mountain ranges, and precip- itation is least in the lowlands. ’ In recent years the precipitation pattern has been studied in greater detail in several areas in the seven southwestern States, notably by Spreen (1947) in west- ern Colorado, Hiatt (1953) in northern Arizona, Trox- ell and others (1954) in southern California, and Peck (1956) in northern Utah. Troxell and others found that the lines of equal average precipitation increase in altitude with increasing distance from the ocean and are notably higher on the lee slopes than on the wind- ward (western) slopes of mountain ranges. Hiatt found altitude to be the most important topographic factor in precipitation in northern Arizona, although the degree of rise and direction of rise were also significant. The mean annual precipitation necessarily includes precipitation from both Pacific and Gulf sources. It is well recognized that the prevalence, directional move— ment, and characteristics of airmasses follow no single defined standard pattern during winter or summer: Polar Pacific air may sweep across the Southwest at any time, and Tropical Gulf masses may cause precipitation during winter as well as summer. Nevertheless there is a seasonal predominance of each that may influence the average weather, or climate, at any given location. The relative proportion from Gulf and Pacific sources is therefore suggested by comparison of precipitation during the summer (April—September) and winter (October—March). Dorroh (1946, fig. 2) shows that the average summer precipitation is more than 75 per- cent of the annual total in the Great Plains, and slightly less than 50 percent in western Arizona and Utah. By contrast, the April—September precipitation in southern California is a very small proportion of the annual total. SUBDIVISIONS OF THE SOUTHWEST In his analysis of precipitation data for Arizona, Hiatt (1953, p. 191) found it necessary to use the param- eter of “Zone”— * * * an area contained within logical topographic boundaries which separate the stations into sizable groups exhibiting a consistent relationship between station precipitation and the parameters used. The Zone parameter, in particular, is also a ' catch-all which serves to separate the data into groups contain- ing unaccountable but consistent variations in precipitation. Early in the studies for the present report it became apparent that the broad region of the Southwest is not a homogeneous meteorologic unit, and that it includes several subdivisions of contrasting precipitation pat- ~ A—12 DROUGHT IN THE SOUTHWEST, 1942—56 / \ \\\\\\:\‘§\\\§\ \\\\\ _ \\, x ./ \\\ \ 63‘ \\‘&\\\\ \‘ ’9 §Q§§§\\\\\: \2 E3 ‘ \ , ‘ U ........... 2% I}! <<<<<< (k EXPLANATION a E ipitation '\\\\\\\\\‘\;\\\s\ 9 \\ rmals, 1921—50, supplemented by long—term averages In mount ' clpitatlon in the Southwest. Based on 30-year no a1 pre mun: 3.——Mean annu THE METEOROLOGIC PHENOMENON OF DROUGHT terns within each of which there is some degree of con— sistency among the records from individual localities. This is to be expected from the various airmasses that cause the precipitation, and from the indicated effects of topographic barriers upon that precipitation. In part the subdivisions suggested below and outlined on figure 4 are extensions of the zones outlined by Hiatt in Arizona. The Pacific Border includes the part of California west and south of the Sierra Nevada and the high ranges of southern California—~that is, the coastal val- leys and ranges and the San Joaquin Valley—where most of the precipitation originates from Polar Pacific or Tropical Pacific airmasses, and falls during the six months October through March. The Sonoran Border adjoins the Pacific Border to the southeast, and lies generally south of the Mogollon Rim and other highlands as far east as the middle Rio Grande Valley in New Mexico. It includes the Mojave and Colorado deserts in California and the Sonoran desert of Arizona, which border the Mexican State of L Isa: Lak 571/ REjAT B'AsmE—cb oRA ,. \ - . . [I l g . , McGig ,' i a i I \ EXPLANATION 9 Station having records exceeding 50 years, used in studies of drought intensity and duration . Station having records exceeding 25 years, used in determining area of drought Boundary between meteorologic zones A—13 Sonora. Here too the Pacific Ocean is the source of precipitation during the winter, but the amount is considerably less than that received along the Pacific Border, because the deserts of southern California are in the rain shadow of high ranges, and southern Arizona and New Mexico are far inland. The California des- erts receive no more precipitation in the summer (April through September) than does the Pacific Bor- der subdivision, but summer precipitation constitutes a larger proportion of the annual total. Farther east, in Arizona and New Mexico, most of the annual precipi— tation occurs during the summer, during the “Sonoran summer monsoon” (Bryson, 1957). This Sonoran Bor- der includes the most arid region in North America, which receives very little precipitation from any source. A feature of the zone is the hot dry Tropical Continen- tal airmasses originating in Mexico, especially in late spring and early summer. The Great Basin—Colorado Plateaus subdivision is east of the Sierra Nevada, north of the Mogollon Rim, and west of the Rocky Mountains, and thus includes the .'\,r-o \ \ ,3, .4: . cmyrsrr \ \ \ggownsvill FIGURE 4.—Meteorologic zones in the Southwest. A—14 Great Basin of Nevada and western Utah and the C010- rado Plateaus farther east. The western part of this area is in the rain shadow of the Sierra Nevada. The Ruby and Snake Ranges in eastern Nevada and the Wasatch Range in northern Utah are the most sig- nificant orographic features within this subdivision. The area receives precipitation from both Polar Pacific and Tropical Gulf airmasses, and in Winter it also receives dry Polar Continental air from the north, which may develop persistent high-pressure areas within the Great Basin. The Great Plains subdivision includes the part of the drought area lying east of the Rocky Mountains in eastern Colorado, eastern New Mexico, and Texas. Much of this subdivision is in the rain shadow of the ROCkies insofar as moisture of Pacific origin is con- cerned, and a large part of its precipitation occurs during the six months April through September. DEVIATIONS FROM AVERAGE PRECIPITATION The precipitation patterns of the Southwest, and the airmass movements involved, have been summa- rized in preceding sections in terms of average condi- tions: in the case of precipitation, the computed mean at each locality, based on the period of record; in the case of airmass movements and storm tracks, the av- erage of observations of many individual storms as they progressed into and across the region. These averages are important elements in the descrip— tion of the climate of a region; in some parts of the earth the climate, at least seasonally, is sufficiently uniform that such averages indicate the normal or usual climate. This is not true of the Southwest, where in most places the averages of annual precipitation are based upon a great range in yearly totals; as an out— standing example, the maximum annual precipitation at Indio, Calif., (in 1939) was 89 times the minimum (in 1894). Thus, although in common parlance and in Weather Bureau publications prior to 1955 2 “nor- mal” is used as synonymous with “mean,” the mean precipitation in southwestern localities is not normal in the sense of being usual or most common. At six cities having precipitation records covering 63 to 103 years (p. A—15), the yearly precipitation during more than half of the years has been either less than 85 per- cent of the record mean or more than 120 percent of that mean. 21h January 1955 the Weather Bureau issued the following defini- tions: Normal: 3. mean based on the 30-year period, 1921—50, if neces- sary adjusted to the most recent location, and revised at the close of every decade by dropping the first 10 years of data and adding the 10 most recent years (according to the World Meteorological Organization regulations). Record mean : mean for the entire period of record without adjustments. Mean: mean for any specific period of record, other than normal or record mean. DROUGHT IN THE SOUTHWEST, 1942—56 Many of the irregularities of individual storms are obviously caused by detailed physical aspects of the earth’s surface as the airmasses pass over it. As Namias (1955, p. 205) points out: To name only a few pertinent factors, one might cite the differing effect of various surfaces (snow cover, open water, bare land, etc.) upon incoming radiation from the sun, the impact upon various air flows of mountain chains, and the net result of the latent heat of condensation released when precipi- tation falls. These factors presumably operate differently de- pending upon the initial state of the general circulation. In this sense subsequent anomalous patterns are viewed as a natural evolution from preceding patterns which are them- selves anomalous. In other words, the circulation changes, whether treated on the scale of a day, a week, a month, a season or more, are self-evolving and are guided by an ever-changing radiational balance incident to changing season. CHARACTERISTICS OF RECORDED DROUGHTS IN THE SOUTHWEST This section considers drought in the Southwest solely as a meteorologic phenomenon—a deficiency of precipitation in comparison with the long-term aver- age—and disregards the question of its significance or seriousness to mankind. Drought during the years 1942—56 is considered with respect to geographic area and the annual variations, frequency of dry years, and persistence of drought within that area. Comparisons of this drought with earlier droughts and with inter- vening periods of more abundant water supply are based upon data from a variety of sources which docu- ment the climatic fluctuations in the Southwest. The most abundant and detailed records pertaining to climatic fluctuations are those from the US. Weather Bureau’s network of precipitation stations, which pro- vide the only direct and quantitative measurements of the gross water supply that falls as precipitation. Although each station provides only a sample of the rainfall in a broad area, a suflicient number of records reasonably consistent with each other provides a sound basis for conclusions as to regional climate and fluctua- tions in that climate. Hence these records are given especial attention. In the Southwest hundreds of these records cover more than 50 years, and about half a dozen extend back for a century. Climatic fluctuations prior to 1850 are indicated by data from tree rings, lakes in closed basins, historic records, and archeologic and geologicstudies. Based on these data, inferences may be drawn concerning climatic fluctuations in the past several thousand years. AREA OF DROUGHT IN 1942—56 For a first approximation in delineation of the area of the recent drought in the Southwest, it was assumed that there are no serious drought effects unless the THE METEOROLOGIC PHENOMENON OF DROUGHT annual precipitation is less than 85 percent of the record mean as computed by the Weather Bureau. J. C. Hoyt uses this criterion for State averages of precipitation in identifying drought in humid and semiarid States, but points out also that in arid regions it might be misleading because of the wide range of precipitation; he shows (1936, p. 65—66) that in the Southwest the State average precipitation was less than 85 percent of the long-term mean in 12 to 19 out of 54 years, and less than 70 percent of that mean in 3 to 7 of those years. It has long been recognized that relative variability of precipitation varies inversely with mean total precipi- tation; a recent confirmation of this principle is con- tained in a report by McDonald (1956), who points out that the coefficient of variability (the dimensionless quotient of the standard deviation of precipitation di- vided by the mean) is generally higher for Arizona localities than for more humid regions. Plate 1 shows the areas in the Southwest where pre- cipitation was less than 85 percent of the record mean as of 1941 in each of the calendar years 1942 to 1956, inclusive. These areas are outlined on the basis of data from more than 300 precipitation stations (locations shown by dots on the maps) where records cover more than 25 years. The area of precipitation deficiency changed markedly in size and position from year to year but every year it was large enough to extend into at least four States, and in most years it covered parts of all seven States, California, Nevada, Arizona, Utah, Colorado, New Mexico, and Texas. In some years, no- tably in 1956 and to a lesser extent in 1950 and 1952 to 1954, inclusive, drought conditions enveloped a large part of the United States (US. Weather Bureau, 1953, 1957). In these years the area of precipitation deficiency extended beyond the arid and semiarid Southwest and into more humid regions farther east, where the chief effects of drought are deficiencies in soil moisture during the growing season. The dashed line forming the eastern boundary on each of the maps of plate 1 (extending from central A—15 Kansas south to the Gulf of Mexico) represents the boundary between areas of normally moist and dry climates, as delineated by Thornthwaite (see fig. 1). East of this line the mean precipitation exceeds the po— tential evapotranspiration and provides a surplus for surface or underground flow. West of the line (except in numerous but isolated mountainous areas) the aver— age precipitation is insufficient for the evapotranspira- tion demand, and the climate is arid or semiarid. It is a general rule that dry years—When precipita- tion is less than average—are more frequent that wet years. In other words, the median rainfall—the amount that is exceeded in 50 percent of the years—is signifi- cantly less than the mean, or long-term average. The graphs of figure 5 show the frequency of annual pre- cipitation at six cities in the Southwest. In this group the annual precipitation has been less than the mean during 52 to 65 percent of the time, and has been less than 85 percent of the mean in one-third to one-half of the years of record. The “scatter” of individual years from the central tendency is indicated by the standard deviation. Statistics of these six records are summarized in the table below. Precipitation in every part of the Southwest has been less than 85 percent of the record mean in at least 3 of the years 1942—56, and in some areas as much as 13 of those years. If annual precipitation less than 85 percent of the mean is to be expected at least one-third of the time, as suggested by the records from several stations (fig. 5), then drought or exceptional dryness occurs only when and where such a precipitation defi- ciency exceeds this frequency. The areas where precipitation was less than 85 per- cent of the mean in more than half of the years 1942 to 1956, inclusive, are bounded by heavy lines on figure 6. This boundary defines the southwestern drought area as discussed in this report; it includes nearly all of Arizona and New Mexico, substantial parts of Cali- fornia and Texas, and some of Nevada, Utah, and Colorado. Annual precipitation, in inches, at six cities in the Southwest Length of 1 2 3 4 5 6 \ record in 1954 Station (see fig. 5) (years) Median Record Observed Observed Standard Coefficient mean maximum minimum deviation of variability 1 Indie, Calif __________________ ‘ 75 3. 07 3. 56 12. 47 0. 36 1. 92 0. 54 El Paso, Tex _________________ 74 8. 26 8. 67 17. 46 2. 4O 3. 43 . 40 St. George, Utah _____________ 63 8. 29 8. 45 20. 11 3. 01 2 81 33 San Diego, Calif ______________ 103 9. 63 10. 05 26. 09 3. 63 3 48 35 Tucson, Ariz _________________ 85 10 80 11.19 20. 90 4. 73 3 26 29 Santa Fe, N. Mex ____________ 103 13. 73 14. 18 26. 75 7. 31 4 25 30 1 Ratio of standard deviation to mean. A—16 PRECIPITATION. IN INCH ES DROUGHT IN THE SOUTHWEST, 1 942—56 3° I I EXPLANATION 25 — M / Mean, ”normal," or average for period of record . C 20 _ m ,_ 2 Median, or the precipitation «I <5 as often exceeded as not SANTA FE’ ; NEW T; MEXICO 15 II E 1/ E X, / E TUCSON. 5' g ARIZONA o . 15 0°. :1 2 II , u 2 / E —-—+"‘/ 10 I 1 / 35 percent M 5 SAN DIEGO, 5 CALIFORNIA 15 =' “-3 3 2 / a; II II 2 ___/ /— 10 ”4,27 I / 85 percent M fr 5 .5' EL PASO, g as TEXAS / 00' II 10 II 2 / // 85 percent M 5 / c F. .E' ‘3 ST. GEORGE. g UTAH “I: 10 ll ’— E /_—+_———' I 5 85 percent M c' I 10 5 S INDIO, h CALIFORNIA m. o m «5 II 5 II 2 I E 4/ ' I I 85 percent M 00 10 20 30 40 50 40 30 20 10 0 PERCENTAGE OF YEARS WHEN PRECIPITATION IS LESS THAN THAT INDICATED PERCENTAGE OF YEARS WHEN PRECIPITATION IS GREATER THAN THAT INDICATED FIGURE 5.—J§‘requency of annual precipitation at six cities. THE METEOROLOGIC PHENOMENON 0F DROUGHT -: . \: "I/ in. I] s / \ ‘7 EXPLANATION Precipitation 1&3 than 85 percent ‘ of the mean In 8 or more years InlOormoreyears . Station having records exceeding 25 yeaxs tss ‘ Q \ ’4 4 u\ //3'o. /, //////// \‘s, x & \\\\ A—17 0 a . // // - D: ‘3‘ 3 120 “67)» '3 1‘ 3° \ . E e-‘q’ E %9& DJ 34°“ 2 o 3 ’3 1° We“ w D \‘3’?’ E 2 o g; 80 210 I 2 2 flange _. ce“\° ' g 60 69 oe‘ 200 c: . o‘ 0 9m- 9 8 e90 y” 12 5.) 40 180 z o yw \ 9; ,: as < H : ._: Q. —- > 0 c: Lu 20 160 < E A W W 2 < —. E o 140 E ‘Average in Z i— < _l 3 E 120 O (0' NJ I o E A \y 100 E 5% ~ g“ $ Y” o 09} I! 3 ,5 80 z Vx y P < Z 0 60 I: < r— E 0 DJ (I f n. / 40 9 / a/ (‘8 20 Average 0 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 FIGURE 8.-——Cumulative monthly precipitation at Socorro and Santa Rosa, N. Mex, 1942456. cided. On the other hand, some climatic phenomena accumulated actual and normal precipitation during a extend over most of the Southwest; for example, 1941 designated period. Figure 8 shows the cumulative total was one of the three wettest years at all five stations. of precipitation by months from January 1942 to De- Another form of presentation is by graphs of the cember 1956 at two localities in New Mexico. At both A—21 THE METEOROLOGIC PHENOMENON OF DROUGHT SEHONI NI ‘NOIlVlIdIOBHd ‘IVnNNV .«wwtgzcm of 5‘93? .53 “a coflafifiowa “waives? «o 935% eZuEwanolld ammo—rm m_< Hwzzfin. .5310 Wwfi 82 83 omfl 82 22 82 83 82 SE 32 on”: 88 SS 82 82 , x/ 08 \l\ r/ 02 a / >> >> / /\/(\/\ < (R 0 \K < 02 mw com .2 85.98 .oozéao ”852: oo 8H 8 ( >\( o /\|\/\/ \/|\\ 02 mm (2 com 8.5.2 52 dmammomg H$9.0m 25528 0 O8 >\J /)\L\/ 2: 0H ?ZK<;\|//\/I\\ / oom 7/ 8m 2 . 08 55835 .mmfioz,‘ m9 ”528 0:65 2 cow 1 7 2 \A 02 / \/\/ /\/ > 2 v o \ < x < ,\ /\. < m.2 m>_mwm~.oomn_l .m Z_ OmIHomH $.03“. mmDHm’EmD m>F<43$50I .1 NVBW HVEA-OS 50 iNEOHEId NI ‘NOIiVlIdIOEHd 'anNNV :10 aumavaaG 3Auv1nwno A—22 stations the seasonal cycle, with heaviest rainfall in summer, is noticeable even in dry years, although sum- mer precipitation was below average in most years. In order to show major trends in precipitation, and thus compare the intensity and duration of recorded dry periods, it is necessary to suppress, or smooth, the great fluctuations in precipitation at individual locali- ties. Graphs of cumulative departure from mean pre- cipitation, and graphs of progressive or moving averages of precipitation have both been widely used to indicate general trends; each form of graph has its advantages and disadvantages. Both types of graphs, showing precipitation at four localities, are presented in figure 9. For the graphs of cumulative departure, the mean water-year precipitation has been computed for the base period 1901—50, and yearly deficiencies or surpluses with respect to this mean have been added progressively. Each graph crosses the line of origin in 1900 and 1950, and is extended back to the beginning of record for each station, and also forward to include water—year 1958. For better comparison of localities with markedly dif- ferent mean precipitation, the cumulative departure in figure 9 is expressed in percentage of the 1901—50 mean. On a cumulative-departure graph the wettest years are shown not by a peak, but by the most steeply rising parts of the graph, and conversely the driest years are indi- cated by the steepest descents. Progressively decreasing precipitation throughout the base period is shown by a convex-upward curve on a cumulative-departure graph, and progressively increasing precipitation is indicated by a concave—upward curve. Alternating wet and dry periods are faithfully depicted on a cumulative-depar- ture graph, but if the record begins with a wet period the entire graph may be above the line of origin and if it begins with a dry period the graph is generally below the base line. Progressive averages have been used by several stu— dents of long-term trends, notably by Kincer (1933) for temperature, by W. G. Hoyt and others (1936) for pre- cipitation and runoff, and by McDonald (1956) to show secular trends in seasonal precipitation. All three used averages for 10-year periods: in their graphs each plotted point represents the “mean” that would have been computed had the record been initiated ten years earlier. Progressive-average curves have several ad— vantages over cumulative-departure curves: they de- pict not only long-term cyclic fluctuations, but also any long—term trends that persist throughout the period of record; and they permit direct comparison of stations having great differences in length of record. However, there are the disadvantages that cyclic fluctuations may be obscured or amplified, depending upon the method DROUGHT IN THE SOUTHWEST, 1942—56 and time interval chosen for “smoothing”; and a single year of outstandingly excessive or deficient precipita- tion creates a plateau or trough whose significance is to be discounted. ANNUAL PRECIPITATION Figure 9, which provides comparisons of graphs showing cumulative departures and progressive aver- ages of annual precipitation, also presents graphs for a city in each of the four meteorologic subdivisions of the Southwest described on pages A—ll to A—14. The pairs of graphs for each city are similar in that they show upward trends and downward trends during the same general periods. Each graph trends downward and thus indicates a preponderance of dry years during the period 1942—56, but the dry years within that period vary from one city to another. Earlier droughts are indicated by similar downward trends in each graph— and these also vary considerably in the four cities rep— resented. There are very few years in the past 75 when at least one of the four cities represented on figure 9 was not undergoing extended drought. However, although many droughts occurred in the Southwest in the past century, these droughts characteristically did not engulf the entire region. The graphs for Los Angeles, Calif., show an alterna- tion of wet and dry periods, generally of 10 to 13 years’ duration. Dry periods occurred in 1894—1904, 1917—32, and 1945—56; the intervening periods were compara- tively wet. This alternation is typical of the Pacific Border zone, Where precipitation originates almost en- tirely in moist Pacific airmasses, and occurs during the winter. The graphs for Amarillo, Tex., also exhibit an alternation of wet and dry periods, similarly of 10 to 13 years’ duration, but the dry periods do not cor— respond to those in the Pacific Border zone. Instead, the dry periods occurred in 1907—18, 1929—40, and 1950— 56. These dry periods have been recorded at many other places in the Great Plains, where most of the precipitation comes from the Gulf of Mexico and occurs during the summer. The other graphs on figure 9‘ represent localities in the Great Basin—Colorado Plateaus zone—which re- ceives precipitation from both Pacific and Gulf air masses—and the Sonoran Border zone, which similarly depends upon both sources but receives little from either. The droughts shown at these cities occur in part during the dry years recorded in the Pacific Bor- der, and in part during the dry years recorded in the Great Plains. Because the annual precipitation in this broad region is derived from two major types of air masses, each producing cyclic alternations but in op- posite phases, the trends in precipitation show an in- THE METEOROLOGIC PHENOMENON 0F DROUGHT teresting variety. Notable examples are a quarter century of predominantly wet years at Durango, 0010., and a dry period of corresponding length at Lords- burg, N. Mex. Because of the contrasting precipitation trends at these four cities representing the principal meteorologic zones in the Southwest, it is desirable to explore in greater detail the trends in precipitation within each- zone. DROUGHT IN THE PACIFIC BORDER ZONE The graphs assembled in figure 10 represent three cities near the Pacific coast and two in the Central Valley of California. All graphs show declining trends during the years 1945—56, similar to that in Los Angeles (fig. 9). All graphs likewise indicate the dry periods shown at Los Angeles in 1894—1904‘and 1917—32, although in several cities those dry periods extended somewhat beyond the periods indicated. The longest records also indicate a drought in the years 1870—79. The four drought periods are shaded on figure 10. In San Diego and San Luis Obispo the steepest de- clines, indicative of the most severe droughts, were in 1894—1904 and 1945—56. In the Central Valley (Fresno and Sacramento) the wettest period (1935—44) of the past century was followed by a drought (1945—51) that was short and intense in comparison with earlier pe- riods when precipitation was prevailingly less than the long—term average; in both localities the precipitation deficiency in 1894—1904 was less than in the cities farther south. The graph based on measured rainfall at San Fran- cisco shows trends similar to those indicated by the records for San Luis Obispo, Los Angeles and San Diego. However, when this record is adjusted on the basis of contemporaneous records for nearby Oakland and San Jose (Piper, 1959), the indication of drought during the decade 1891—1900 disappears. Instead, the adjusted graph has trends similar to those shown for Sacramento, which is less than 100 miles northeast of San Francisco. The official San Francisco rain gage has been moved 10 times during the past century, and the largest indicated correction was during the period 1892—1906. San Francisco is not within the area of Southwest droughts as defined in this report; it is brought in to suggest that sampling errors may be of considerable magnitude, particularly in the longest records. The uppermost graph of figure 10 shows the an— nual relative sunspot number, as developed at Zurich Observatory by Wolf; as shown by Ellison (1955, p. 31—54) this graph is similar in form to a graph of mean area of sunspots based on daily photographs, as A—23 recorded by the Greenwich Royal Observatory. Most of the dry periods in the Pacific Border apparently began during years of maximum sunspot activity, and have continued through several years of declining ac- tivity, and sometimes until a subsequent and lower maximum was recorded. DROUGHT IN THE GREAT PLAINS ZONE Precipitation in the Great Plains, reaching a maxi- mum in the summer, commonly occurs during convec- tional storms that may be quite localized. The graphs of figure 11 represent eight cities in the Great Plains area, from Denver, Colo., which is farthest north and farthest from the Gulf of Mexico, to Brownsville, Tex., which is farthest south and borders the Gulf. On this graph, single years when precipitation was more than 50 percent above the average for 1901—50 are shown by heavy lines. Commonly these years include storms of high intensity but local or irregular distribution. The alternating wet and dry periods shown so clearly by the record for Amarillo, Tex., (fig. 9) appear also on most of the graphs of figure 11, but not in the graph for Brownsville, Tex. The graphs for Denver and Pueblo, 0010., indicate some alternation of wet and dry periods, but the periods are not synchronous with those recorded at cities farther south. In the last 50 years the major trends shown on the Denver graph are counter to those shown at Brownsville, and suggest that when Brownsville is receiving more than average rain, there is less available for the remoter parts of the Great Plains as represented by Denver. Tannehill (1947, p. 109, 119) points out that such difi'erences in precipitation trends are evident not only between the northern and southern Great Plains, but between the northern and southern parts of the continental interior. Comparison of these graphs with the graph of rela— tive sunspot numbers indicates that many of the droughts in the Great Plains, like those in the Pacific Border, begin in years of maximum sunspot activity; these droughts, however (as distinguished from those in the Pacific Border), begin during the lower maximum of the double sunspot cycle. CYCLIC FLUCTUATIONS To summarize the fluctuations that appear on figures 10 and 11: the wet and dry periods noted in the Pacific Border and Great Plains are generally of 10 to 13 years’ duration, but are almost opposite in phase, so that when one zone is drier than average the other is wetter than average; northern localities in each zone exhibit different trends from those farther south; cyclic fluctuations are more pronounced at some localities than at others in each zone, and are not clearly discernible A—24 WOLF NUMBER(R) 8 8 8 O 25 20 15 10 15 10 30 25 20 ANNUAL PRECIPITATION, IN INCHES 15 3O 25 20 10 DROUGHT IN THE SOUTHWEST, 1942-56 1870—79 drought 1894-1904 RELATIVE SUNSPOT NUMBER ,v—/\/ /v 1860 1870 1880 1890 Fromm 10.—F1uctuations of annual precipitation at five cities in Pacific Border zone. drought SACRAMENTO, CALIFORNIA FRESNO. CALIFORNIA 'usted (Piper, 1 agree with Oak and San Jose) x,” \ Measured SAN FRANCISCO, CALIFORNIA SAN LUIS OBISPO. CALIFORNIA SAN DIEGO, CALIFORNIA 1900 1910 1920 10-year average. 1917—34 drought 1930 1940 1945-56 drought 1950 1956 Progressiv'e WOLF NUMBER(R) CUMULATIVE DEPARTURE OF ANNUAL PRECIPITATION. IN PERCENT OF 50-YEAR MEAN THE METEOROLOGIC PHENOMENON 0F DROUGHT 1887—98 1907—18 1930-40 1950—56 drought . drought drought drou ; ht 150 100 50 RELATIVE S 200 100 100 100 100 20" PUEBLO. COLO RADO 100 100 20° DODGE CITY, K \NSAS 100 100 20° ' KLAHOMA 100 100 200 100 100 200 100 100 20° EAGLE PASS. TEXAS 1 200 300 BROWNSVILLE, TEXAS 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 FIGURE 11.—Fluctuations of annual precipitation at eight cities in the Great Plains zone. /\F A—25 A—26 in the records for some localities. As to possible causes of these fluctuations, Tannehill (1947, p. 135—144) states: During periods of deficient rainfall in the United States we frequently hear that droughts come in cycles. The most fre- quent claim is that they come in some fraction or multiple of the sunspot period of about 11 years. There is much evidence that rainfall varies in a cycle of about 11 years, and also there are some indications of variations in periods of 22 or 23 years, and in the so-called Bruckner cycle which is roughly three times the sunspot period * * *. In the long run, the sun’s heat is responsible for all changes in the weather, as the sun rises and sets and goes north and south with the seasons. The diurnal effects, and especially the seasonal efiects, will afford some clues as to what happens when the sun’s radiation changes in a period of eleven years, or whatever it may be. In spring and summer a high percentage of the year’s rain- fall is carried into the interior of the continent. The sun is high and the continent warm, and the oceans are relatively cool. In autumn and winter a greater percentage of the year’s rainfall occurs in coastal areas or nearer to the coasts than in spring and early summer. When the sun is low, the continent is cold and the oceans are relatively warm. Therefore, if the sun gets progressively colder from year to year there should be an increasingly lower percentage of rainfall in the interior, and vice versa. There will be many local variations, of course, and the lag of ocean temperature changes will have to be taken into account; but the broad effects should be fairly clear if the sun’s radiation changes in an eleven-year, as we believe it does * * *. [From study of the difierences between the national rainfall in the first and second halves of the calendar year] we see that in general when there were numerous sunspots there was more rain in the first half of the year and when there were few sun- spots there was more rain in the second half of the year * * *. The temperature differences which carry rainfall into the interior of the United States in late winter, spring, and early summer are owing to (1) the sun’s heat and (2) the lag of ocean temperatures. The temperatures of the Pacific rise in the spring and fall in the autumn at a much slower rate than the tempera- tures of the continent in the same latitudes. In the same way, the temperatures of the Atlantic and Gulf also rise and fall at a slower rate than the continent. The Atlantic and Gulf, es- pecially the latter, are smaller than the Pacific and consequently their temperatures change more rapidly than the Pacific but less rapidly than the continent. Furthermore, there is a varia- tion in temperature between the eastern and western parts of the oceans. These varying temperature differences between the two oceans (and the Gulf of Mexico) introduce a secondary effect which is evident chiefly in changes in the seasonal distri- bution of the rainfall in the area east of the Rockies * * *. When sunspots are numerous, more of the rainfall tends to go to the interior, including the Great Plains * * *. At high sunspottedness, the Southern Great Plains get relatively more rain‘ [than the Gulf Coast]. The difference is very important, with a range of nearly 10 inches per year. There are definite indications that when the increase in the-sun’s heat is unusually great, or when the oceans are relatively colder than usual at time of maximum sunspots, the rainfall is diverted stil-l farther into the interior * * ‘. The varying rates of response of the continents and oceans give us rainfall variations that do not follow exactly the sun- spot cycles; but the rainfall curves of the United States, if DROUGHT IN THE SOUTHWEST, 1942—56 analyzed properly, seem to fit the broad pattern. Perhaps we can watch the sunspots and predict what our rainfall will be. It may be possible when we know more about it. In the past we have been as badly confused about sunspots as on the drought question * * *. Galileo published his paper on sunspots in 1613. Later it was found that the numbers of these sunspots vary in a cycle of about eleven years. After more observations were made, the period was found to average slightly more than eleven years. Actually, the period varies, going as long as seven years or as high as fifteen years. The discovery of the sunspot cycle brought forth the idea that the temperatures on the earth should vary with the number of spots, and many scientists en— thusiastically assembled temperature records to prove it. There was great hope that it would be possible to predict the weather far in advance. The investigators expected to find that at max- imum sunspots, when the sun is hottest, the atmosphere would be warmed. With few spots and a cool sun, the atmosphere was expected to be cooler. To their astonishment, they found that the opposite was true. At the surface of the earth, where we have all our long weather records, more sunspots bring lower temperatures and fewer spots bring higher temperatures. 'This fact is fully supported by records in the tropics, where there are fewer violent changes in the weather than in the latitude of the United States. A hot sun makes a cool earth, and vice versa * * *. Sunspots represent only one activity in the sun. There are other evidences of solar activity, such as prominences,‘ magnetic activity, faculae, and the corona. Records of sunspots are avail- able for a long period of time. The other evidences of solar vari- ation seem. to follow the same course, but the records are by no means so extensive. DROUGHT IN THE GREAT BASIN AND SONORAN BORDER ZONES The cumulative-departure graphs of annual precipi- tation at several cities in the Great Basin and Sonoran Border zones (fig. 12) show a considerable variety in precipitation trends, and a significant lack of uniform- ity in the region during the period 1942—56, as Well as . during earlier droughts. In this broad region the an- nual precipitation is contributed from both Pacific and Gulf sources, one source predominating in the Winter, the other in the summer. For evidence of cyclic fluctu- ations, or of fluctuations synchronous with those ob- served in the Pacific Border and Great Plains, it is therefore desirable to analyze the precipitation trends in this intermediate region by seasons. SEASONAL PRECIPITATION A comprehensive analysis of precipitation in the Southwest on the basis of major sources of that precipi- tation would require the study of individual storms, but for the purpose of this report it is considered sufficient to work with periods during which the precipitation is predominantly from one source. On this basis, annual totals have been used for analysis of precipitation in the Pacific Border, where precipitation comes predomi- nantly from Pacific airmasses, and in the Great Plains, where precipitation comes chiefly from the Gulf of Mexico. THE METEOROLOGIC PHENOMENON 0F DROUGHT 100 /\ /\ 0 m wok J V WWW k / \/ 2°C WINNEVIUCCA, NEVADA \I‘VJV 200 100 \ c /‘ c \ 10° SALT LAKE CIT ,UTAH 200 100 o f /\ \hAf 100 MANTI, UTAH CUMULATIVE DEPARTURE OF ANNUAL PRECIPITATION, IN PERCENT OF 50—YEAR MEAN 1 200 1950 1870 1880 1890 1900 1910 1920 1930 1940 1960 Fromm L2.—Fluctuations of annual precipitation at six cities in the Great Basin and Sonoran Border zones. In the Great Basin, Colorado Plateaus, and Sonoran Border the precipitation during the winter (starting in October or November and continuing until April or May) comes chiefly from the Pacific Ocean. Gulf sources are dominant in the summer, notably in July, August, and September, and to a lesser extent in earlier and later months. On the basis of storm genesis no reg— ular dividing dates can be set between the “winter” and “summer” seasons. Indeed there may be a considerable overlap in some years when the region is the recipient of storms from one source and then another; and in other years there may be an interval when very little precipitation is received from either source (June is ordinarily the driest month-of the year in many com- munities, evidently too late for general winter storms and too early for the summer convectional storms). For convenience, however, the water year (ending Sep- tember 30) can be subdivided into halves: a winter season from October through March, and a summer A—27 season from April through September. These are the seasonal subdivisions considered in the following discussion. Graphs showing progressive average seasonal precip- itation at six cities are presented in figure 13. All these places except Salt Lake City are in the area of Southwest drought (fig. 6). Salt Lake City and St. George, Utah, and Durango, Colo., usually receive most of their precipitation during the six winter months, October through March. At the other three cities, lo— cated farther south, most of the yearly precipitation occurs during the months April through September. On the chart for winter (October—March), the periods that have been prevailingly dry in the Pacific Border (fig. 10) are shown by shading. Winter precip- itation is less than average during these periods at several of the cities represented in figure 13, notably at Tucson and Prescott, Ariz., and St. George, Utah, but to some extent even at ROSWell, N. Mex., which is in the Great Plains zone. On the chart indicating summer (April—September) precipitation, shading indicates the periods that have been prevailingly dry in the Great Plains. Precipita- tion has trended downward during these periods at sev- eral of the cities represented by the graphs of figure 13, including Roswell as well as localities in the Sonoran Border and the Great Basin-Colorado Plateaus zones. A graph of the yearly sunspot numbers (fig. 13) for the years 187 0—1955 shows a cyclic fluctuation, with a minimum (ordinarily less than 10) every 10 to 13 years. The maximums alternate between “higher highs” that approach or exceed 100 (in 1870, 1893, 1917, and 1938) and “lower highs” in the range of 60 to 75 (in 1883, 1907, and 1928). The graphs of figure 13 indicate that the dry periods in Pacific winter precipitation commonly begin within a year or two of the higher maximums of the sunspot double cycle; they may continue until the neXt (lower) maximum, and even longer in the instance of the 1917— 34 dry period. The dry periods in Gulf summer precip- itation appear to begin at about the time of the lower sunspot maximums, but they too ordinarily continue for about a decade. This suggestion of correlation between drought pe- riods and sunspot cycles may merit considerable skep- ticism, because the precipitation trends indicated on figure 13 may well be created by other factors, including for instance sampling errors at individual precipita- tion stations, the vagaries of distribution during in- dividual storms, the method of smoothing the observed data, and the empirical though convenient subdivision of the year into two seasons of equal length. On the other hand, these same factors might also explain why A—28 WOLF NUMBER(R) ANNUAL PRECIPITATION. IN INCHES p.- O o‘ Q ~— 01 O 01 O 1893— 1904 drought 0 0 9 a ; o 7 DROUGHT IN THE SOUTHWEST, 1942—56 1916-34 drought : ,- . . o RELATIVE SUNSPOT NUMBE 1944-57 drought 1886—98 drought 15C 100 50 L R. ZURICH OBSERVATORY 13 12 11 10 PRESCOTT, ' 5 SALT LAKE CTY. UTAH 1906—18 drought 1929-40 drought 1949—56 drought TUCSON, 1890 1900 1910 1920 WINTER (OCTOBER-MARCH) 1930 ROSWE LL, NEW MEXICO 1950 1880 1890 1900 1910 1920 1930 1 SUMMER (APRIL—SEPTEMBER) FIGURE 13.——Fluctuations of seasonal precipitation at six cities in the Southwest. Progressive 10-year average. 940 1950 THE METEOROLOGIC PHENOMENON OF DROUGHT all the graphs on figure 13 do not show regular and uniform cyclic fluctuations. TEMPERATURE In the preparation of this report, studies of long-term fluctuations of temperature have been limited to the records of San Francisco and San Diego, as represent— ative of the northern and southern parts of the Pacific Border zone; Salt Lake City and Phoenix, in the Great Basin and Sonoran Border zones; and Dodge City, Kans., El Paso and Abilene, Tex., in the Great Plains zone. Graphs of progressive 10-year average tempera- tures at these seven cities in October to March are shown in figure 14, and in April to September in figure 15. Both Winter and summer temperatures at San Fran- cisco show a downward trend in the 1890’s, an upward trend for about the first four decades of this century, and downward since the mid-1940’s. These trends may reflect changes in temperatures of the Pacific Ocean in midlatitudes. The temperature graphs for San Diego are generally similar in trend to those for San Francisco, except that winter temperatures rose in com- parison with San Francisco during the dry periods 1894-1904 and 1945—51. The trends observed in the Pacific Border are also evident as far inland as Salt Lake City and Phoenix in the summer, but not in the winter. The temperature trends in the Great Plains appear to be quite independent of those observed far- ther west, and in many instances independent of each other. There is some similarity among the graphs of winter temperature at Dodge City, El Paso, and Abi- lene, but marked divergences in summer trends at these three places. At practically all seven cities both summer and winter temperatures reached the maximum for the period of record in the early 1940’s; thus one common charac— teristic of the graphs is an increase in temperature during the first four decades of the century and at some places ever since the beginning of the record, and a decline in subsequent years. Similar trends have been observed in many parts of the world (p. A—35) . No general rule can be stated concerning the trends of temperature during dry periods. The steepest rise in summer temperature shown on any graph is that at Dodge City during the drought of the 1930’s, and sum- mer temperatures trended upward at all cities during : those years. But temperatures did not rise at Dodge City during an earlier dry period (1907—18). Of the seven localities, only Abilene has recorded a rising trend in temperature during each of the dry periods 1907—18, 1930—40, and 1950—56. Winter temperatures declined during the dry period 1945—51 in the Pacific Border, but rose at most places during the dry period A—29 1917 —34. Thus if there is any relation between drought, and temperature, it is obscure and not consistent. GEOGRAPHIC EFFECT UPON PRECIPITATION The mountains in the Southwest have a marked effect upon precipitation, and precipitation in the mountain- ous areas as a rule is greater than that in adjacent lowlands. The distribution of precipitation during individual storms may vary markedly from the pat— ern shown by long-term averages, and such variations may characterize most of the storms in an entire sea- son. For example, along the Wasatch Range in Utah at altitudes lower than 8,000 feet, the snowfall in 1952 was so far above the usual proportion at those altitudes that damaging floods developed in several creeks having low headwaters. Thus, precipitation-altitude relations during individual storms or seasons may range widely from the established long-term mean. In regions where the dominant source of precipita— tion varies seasonally, it is to be expected that the precipitation—altitude relation also will vary from sea— son to season. Along the Wasatch Range, where winter storms originate chiefly in the North Pacific, and sum- mer storms may come from South Pacific or Gulf sources, Peck (1956) has found significant difiz'erences from winter to summer in the precipitation—altitude relation. A question for which we ourself could find no pub- lished answer is whether the variations from the long- term mean relation of precipitation to altitude are entirely random in character, or whether there are evi- dences of time trends in those variations, as there ap- pear to be in the records of precipitation at individual IOcalities. The problem could well be important in a study of the effects of drought, because precipitation records are collected chiefly in lowland areas, and unless the fluctuations shown by them are representative also of fluctuations in the highlands, the conclusions con— cerning the water supplies contributed by the high mountain ranges must be inconclusive. For preliminary study of this question we selected the records of snow-survey courses in the Kings River basin in California, extending from Fresno northeast- ward to the crest of the Sierra Nevada. This area re- ceives precipitation chiefly from Pacific sources and chiefly during the winter. Because many of the snow— survey courses were not established until 1930, and because it is standard practice on them to measure the water content as of April 1 each year, the study was limited to the October to March precipitation in the period 1930-56, inclusive. This 27 -year period includes several years of the prevailingly dry period 1924—34, the entire wet period 1935—44, and the prevailingly dry period 1945—51. A—so TEMPERATURE, IN DEGREES FAHRENHEIT 56 55 54 \V/x/ 53 40 39 38 37 41 40 39 38 53 52 51 1880 1890 FIGURE 14.—Progresslve 10-year average October—March temperature at seven cities in the Southwest. DROUGHT IN THE SOUTHWEST, 1900 DRY CISCO, CALIFORNIA SAN DODGE 1910 DRY 1942—56 TAH 1940 DRY . TEMPERATURE, IN DEGREES FAHRENHEIT TEMPERATURE, IN DEGREES FAHRENHEIT .67 60 59 58‘ 57 66 65 64 71 70 69 78 77 76 75 1880 THE METEOROLOGIC PHENOMENON OF DROUGHT 1890 1900 1910 DRY DRY BILENE, TEXAS 66 65 64 NEW MEXICO 60 KANSAS V— 77 76 75 1920 1930 1940 1950 1956 DRY DRY FIGURE 15.——P‘rogressive 10-year average April—September temperature at seven cities in the Southwest. TEMPERATURE, IN DEGREES FAHRENHEIT A—3 1 A—32 At each of eight snow courses, ranging in altitude from 5,500 to 10,800 feet above sea level, the snow ac- cumulation in each year was computed in percentage of the average for the course in the 20-year period 1930—49. These percentages for each year were then plotted against altitude of the courses. The resulting graphs indicate that in most years the snow accumula- tion, in percent of the mean, does not vary greatly with DROUGHT IN THE SOUTHWEST, 1942—56 altitude; but in some years there is a far greater pro- portion of snow at low altitudes than at high, and in other years the reverse is true. Graphs representing several selected years are assembled in figure 16. Al- though there are obvious variations from year to year, these variations appear to be random, whether compari- sons are made between wet and dry years or between years in prevailingly wet or dry periods. Thus in 1952 350 300 \ Z < Lu 2 3 ’9, g e m 250 2 LL 0 —_ \ /’ \ E / \‘ '3 \ n: \ E 200 \ x \ \ z \ _. \\19 \ .—a ' \fi :—1 \ \ Q: \‘~ 0. /— § ~ _ _ __ _— —— < x Z l 194 ____‘ O 150 / l, v ‘ —::.______ o .z" ’ Z // -w ’ ’ O / 636/ ’2 ,/ 39/ f5 / / z 100 8 / / E / / // 3 .— 1955 M \\ ’( ¥\ 1953 A£———-—— so \ \\ / / /<. _— ———:-:=* h / \ \V ,/’ f 4 \~ / 4/ / 1931 \\\7L / / 0 \l-/ 5000 6000 7000 8000 9000 10,000 1 1,000 ALTITUDE, IN FEET ABOVE SEA LEVEL FIGURE 16.—Snow accumulation in relation to altitude in selected years, southern Sierra Nevada. THE METEOROLOGIC PHENOMENON OF DROUGHT and 1932—Wet years in prevailingly dry periods—there was a greater proportion of snow at low altitudes. The same is true of 1938—a wetlyear in a prevailingly wet period (see fig. 10)——but not of 1937 and 1941, which were other wet years in the same period. As a rule the variations in the relation oflsnow accumulation to alti- tude have a less regular pattern below 8,000 feet alti— tude, which are caused at least in part by melting and therefore reflect differences precipitation. This was the example, when rains meltec snow and caused floods in s graph for April 1956 shows accumulation above 8,000 fee at lower altitudes. It is con inconsistencies in relation c are not random, at least we to them. The graphs of figure 17 re ascertain whether there are of precipitation to altitud dashed lines are based upon in percent of the 20-year me tudes 10,800 and 8,500 feet and the solid lines are basec March precipitation, in pe mean, at Weather Bureau 4,900, 2,100, and 300 feet, wit represent moving 5-year a deviations from the mean, a mon scale, although the orde are more or less in accordan respective snow courses and ] in temperature rather than case in December 1955, for 1 some of the accumulated everal Sierra streams; the greater than average snow xt, but far less than average cluded that if the observed f precipitation to altitude can see no regular pattern present another attempt to time trends in the relation 3. The graphs shown by the April 1 accumulation, an, at snow courses at alti- in the Kings River basin; l on the recorded October— rcent of the same 20-year stations at altitudes 7,000, .hin that basin. All graphs verages of the percentage nd all are plotted to a com- r and spacing of the graphs ce with the altitudes of the precipitation stations. The curves for the four Weather Bureau stations are re— markably similar to each 0 ther and to the curves for the snow courses, suggesting that there is no significant difference in the climatic fluc and low altitudes. tuations as recorded at high From this very brief analysis of records from an area which appears to be an excel lent one for study of varia- tions in orographic effects upon precipitation, the ten- tative conclusions are that there are indeed variations from year to year in the long-term mean relation of pre- cipitation to altitude, but these variations appear to be random in character and have no apparent relation to the major climatic fluctuat served at all altitudes. CLIMATIC FLUCTUATIONS ions which have been ob- SHOWN BY OTHER DATA According to Willett (1953, p. 55)—~ There is continually in progrejs an entire spectrum of cyclical fluctuations of climate, cycles of plitude being superposed on th shorter period and smaller am- se of longer period and larger A—33 amplitude. These cycles include one whose half period, at least in Europe, extends from the Climatic Optimum at about 3000 B.C. to the peak glaciation from A.D. 1600 to 1900, a second cycle of smaller amplitude and a period of some 2000 years, cool- - wet from 500 BC. to AD. 100, warm-dry from A.D. 400 to 1000, and cool-wet from the thirteenth century to the present, and shorter and smaller cycles, from a few centuries in period to the 80-year, the double sunspot, and the single sunspot cycles ob- served during the past two centuries. The longest records of the U.S. Weather Bureau, ex- tending back about a century, are not long enough to give evidence for or against the longer cycles mentioned by Willett. For this evidence we must look to clima- tologic records from other parts of the world, and to other types of data in the Southwest and elsewhere. Evidence of climatic fluctuations of assorted periods and amplitudes, such as those mentioned by Willett, comes from climatologic, historic, archeologic, biologic, glaciologic, limnologic, oceanographic, and geologic data. As we probe farther into the past, all these forms of data become less complete and less definitive, and ' require more shoring up by inferences. Major climatic fluctuations—or more properly climatic variations, since they represent changes maintained over periods meas- ured in tens of thousands of years—are clearly shown by the geologic record of the Pleistocene continental glaci- ations and interglacial epochs. Numerous climatic fluctuations have been identified in the period of 10,000 years since the last advance of continental glaciers (about 8000 B.C.). Generally these reported fluctua- tions are reckoned in millennia for the first 8,000 years (BC), in centuries during most of the Christian era, and in years only since about A.D. 1700—because of the increasing detail of the available records in recent years. Summaries by Sears (1958) and Deevey and Flint (1957) form the basis for the graphs assembled in fig- ure 18, which suggest the following broad postglacial sequence: 8000—7000 B.C.: Recession of continental glaciers, leaving tundra; preponderance of spruce in pollen record. Cli- mate cold and moist. 7000—5000 B.C.: Beginning with climate “similar to today,” then progressive increase in temperature and decrease in precipitation. Abundance of oak, beech, and hemlock in pollen record. Also includes Coch- rane glacial advance. 5000—2500B.C. : Climate generally warmer and drier than today, so that period has been designated “climatic op- timum,” “altitherm‘al,” “megathermal,” “thermal maximum,” and “xerothermic.” Abundance of oak and hickory in pollen record. Drying up A—34 DROUGHT IN THE SOUTHWEST, 1942—56 \\ A / \ /\ ,4 +50 \ v -v’ > ‘35 'aemmmmemrmnwr . , 443a, , EXPLANATION a 1m Snow-course record Rain-gage record — + 50— Cumulative departure from 1930-49 mean. in percent 1930 1935 1940 1945 1950 1955 FIGURE 17.——Cumu1ative departure of winter precipitation from 1930—49 mean, southern Sierra Nevada. of lakes and glaciers, 30-foot rise in sea level the period 7000 to 600 BC. Deevey and Flint during period. have proposed the term “hypsithermal.” ’ 2500—600 3.0.: .600 B.C.—A.D. 1300: Gradual decrease in temperature, with probable A period providing gradually increasing detail rebirth of some lakes and glaciers. Some marked concerning climatic fluctuations, as summarized glacial advances. For sedimentary sequence in by Brooks (1950, p. 113) for countries bordering THE METEOROLOGIC PHENOMENON OF DROUGHT A—35 l l GLACIATION: EUROPE AND I\ (Miller, 1958; Karlstrom, 1956—57; I l ORTH AMERICA )eevey and Flint, 1957) PEAT BOGS: I SWEDEN E I I "'— POLLEN : (Sears, 1958) g g Boreal) E§ Birch $3,. -' Spruce .- """" , ,- ‘ Willovg ‘ , (Tundra) -- sausages“ Pipe Elm | (Xerothe rmic) Oak .. ........... Beech ASh El'n'i Hemlock Hickory (Ariz,) LAKES: SOUTHWESTERN UNITED STATES (Antevs,1948) r TREE RINGS: SOUTHWESTERN (Schulman, 1956) (Douglass, 1947) FLAGSTAFF W curve (subjective) W UNITED STATES Regional deficiency TEM PERATU RE TRENDS (subj - PRECIPITATION TRENDS (subj 10,000 A.D. 1300—1850: A.D. 1850—1940: 9000 8000 I 7000 I 6000 V V I I the North Atlantic and Arctic Oceans, and by Co-Ching Chu (1926) for China. Generally the rainfall was near or below present averages, with some periods of marked drought; the tempera- ture was probably somewhat higher than today, judging by the history of Iceland and Green- land. However, some glacial advances in Alaska. “Little Ice Age,” marked by increasing rainfall and storminess, advancing glaciers, abandonment of northern outposts of civilization, rising lake levels toward end. ”.40 The lines of eviden climatic fluctuation mann (1949) includ several places in the ce concerning the “present ’ as summarized by A111- 9 climatologic records from Northern Hemisphere, the 2000 00' 30 2000 1000 BC 1000 AD FIGURE 18.—Summary of findings pertaining to post-Pleistocene climates. longest of which began in Holland in A.D. 1706. These records indicate generally increasing tem- peratures beginning about 1850 in many parts of the world, but as Ahlmann points out they are not distributed quite widely enough to allow us to characterize the temperature fluctuations as world wide, because of the dearth of informa- tion concerning the Southern Hemisphere, and particularly the Antarctic. The fluctuation in the past century is indicated also by'the recession of glaciers in practically every glacial region in the world, by warming of ocean currents in far northern latitudes, by rising ocean levels with the melting of ice stored on the continents, and, in the Northern Hemisphere, by northward migration of the zones most suitable for certain species of fish and of vegetation. Ahlmann concludes (1949, p. 190) : A—36 If we find in the Antarctic similar evidence of the present climaticfluctuation as has been found in other parts of the world, we shall be justified in concluding that the present fluctuation is a world-wide phenomenon and probably the re- sult of variations in solar activity which, slow as they may be to take effect, are actually resulting in- an improvement‘ in the climate of our world. The accumulated evidence concerning climatic fluctu- ations of various lengths is far from complete, and be- comes more sketchy as one probes further into the remote past. However, there is considerable evidence that the rising temperature trends in the last century have not persisted since the ice recessions of the last glacial epoch some 9,000 years ago. Instead, those 9,000 years have included centuries that were warmer, cen- turies that were colder, centuries that were wetter, and centuries that were drier than the one we know best. The general world conditions as outlined above are based in part upon studies in the Southwest. Many of the findings in that region fit in nicely with the broad summary, and some conclusions serve to amplify or modify the general picture. The results of specific studies pertaining to long-term climatic fluctuations in the Southwest are summarized in the following sections. HISTORIC DATA The diaries and notes of the Spanish missionaries in southern California have been the basis for estimates by Lynch (1931, 1948) of the annual rainfall in the Los Angeles area since 1769, a full century before the beginning of official precipitation records. His cumula- tive-departure graph (fig. 19) indicates dry periods in 1793—1809, 1822—32, and 1843—59 comparable in magni- tude and duration to those shown (fig. 9) by Weather Bureau records in 1870—83, 1894—1904, 1917—34, and 1945—57. The apparent overall downward trend in this graph results partly from the fact that the graph begins with a wet period. LAKE LEVELS Fluctuations in level of the enclosed lakes of the Southwest reflect the water-supply conditions on their drainage areas, for the water accumulated represents the difference between the sum of all elements of inflow and the total discharge. With sufficient records and interpretive study it is possible to discriminate the effects of such factors as precipitation upon the lake, 4 In using the word “improvement,” Ahlmann speaks especially from the viewpoint of those in Sweden, England, and far northern latitudes generally. In the Southwest, which already includes the hottest and driest parts of the country, an upward trend in temperature and result- ant increase in evapotranspiration would not be called an improvement, because it would reduce the net supplies of surface and ground water, unless there were also a general increase in precipitation that might tend to offset the increased rates of evapotranspiration. DROUGHT IN THE SOUTHWEST, 1942—56 evaporation from the lake, ground-water and surface- water inflow; and more indirectly, the effects of precipitation and of evapotranspiration or other con- sumptive use of water in the part of the drainage basin not occupied by the lake. Without such detailed records it is still possible to draw broad conclusions concerning climatic fluctuations: the lake level will ordinarily rise if precipitation increases or evapotranspiration de- creases, or both, within the drainage basin; and it will decline if precipitation decreases or evapotranspiration increases. The rate of evapotranspiration commonly rises with increasing temperature, decreasing humidity, and increasing wind movement. Thus lake levels indi- cate the resultant effects of several meteorologic factors that determine climate. Several enclosed basins in the Southwest were filled to overflowing during parts of the Pleistocene, and several have been evaporated to dryness in subsequent warmer or drier times. Antevs (1948, p. 178) uses data from Abert and Sum- mer Lakes in Oregon and Owens Lake in California as evidence of a warmer and drier climate in the South- west during the period of the “Climatic Optimum”: These lakes lacked outlets in postpluvial times, but never- theless have only a low salinity, a salinity so low in fact that they cannot be remains of the pluvial lakes in the same basins. The pluvial lakes must have dried, and the accumulated salts must have been removed by wind or have become buried, before the modern‘lakes came into existence. The amount of salts in the waters of these lakes in 1887 to 1912, the salt contents of their main feeder streams, and the rate of evaporation suggest that the accumulation of the salts may have required some 4,000 years. This means that the modern lakes were reborn 4,000 years ago and that their basins were dry for long ages before 2000 B.C. Actual records of lake levels in the Southwest began in 1840 or later, but indirect evidence of low stages prior to 1850 has been obtained from trees growing around the margins of some of the lakes. Harding (1935, p. 90) has summarized this evidence at Eagle Lake, Mono Lake, and Lake Tahoe in California, and at Great Salt Lake in Utah: “(1) For 100 years prior to 1850 these lakes were continuously lower than they have been at any time since 1850; and (2) the high stages reached about 1915 are higher than any reached for from 250 to 300 years.” Thus for one or more cen— turies prior to 1850, the Southwest appears to have had a hotter, drier, or hotter-drier climate than that since 1850. In some drainage basins the hydrologic records are sufficient to permit computation of the natural runoff (lake inflow) of the drainage basin. This is the case in the Truckee River drainage basin, which includes Lake Tahoe in its headwaters and Pyramid and Win- nemucca Lakes in its lowest parts. From records of THE METEOROLOGIC PHENOMENON OF DROUGHT A—37 ' 9300 KJV\’\r'\ r’\I\ / V, M A wJ V MIA Ha AA IV. v , W .J w a W RAINFALL IN LOS ANGELES AREA, CALIFORNIA (Lynch, 1948) 150 -___ / E I\ IV [IV “71 [l / ”IV v ’/ \\ I \\\ // ‘ /’ E100 [/h (I \\ ”\ / /\ 2 / /\ I \ \\ I, l /\ f l \ \ ”’n\\\ /// A \\ I/ / \ I \ D ,g \\ / z 50\ /\ A I\ 3 \j \] \J 3 J 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960 RELATIVE SUNSPOT NUMBER, ZURICH OBSERVATORY 1000 a Z Lu 0 a s :1: CL 2 Z 500 /\/"\\ E ;- J D. .. s o 2 E 8 01 < .-1 5‘ 5 0“ g s WJV o H E O 1: LL 500 1840 1860 1880 1900 1920 1940 1960 RUNOFF INTO PYRAMID LAKE, NEVADA (Hardman and Venstrom, 1941) FIGURE 19.—Fluctuations in precipitation and runoff deduced from historic records. fluctuations of levels in these lakes, runoff and diver- sions of the Truckee River, and consumptive use within the basin, Hardman and Venstrom (1941) estimated the annual natural runofi of the Truckee River in percent- age of the 1840—1939 mean. The trends in cumulative departures from this mean are similar to those shown by Lynch’s estimates of rainfall in Los Angeles, especially after 1900 (fig. 19). Hardman and Ven- strom’s general conclusions as to the 19th century history of the drainage bas in are similar to those by Harding concerning Eagle and Mono Lakes: drought conditions prevailed for many years prior to 1840; a period of generally increasel precipitation began about 1860 which, although broken by minor drought periods, lasted until about 1917. Graphs showing the fluctuations in level of Mono Lake, Eagle Lake, and Tulare Lake in California, and of Great Salt Lake, in Utah, are presented in figure 20. The levels of all these lakes have been affected by man- made diversions—Mono Lake since 1940, Eagle Lake since 1924, and Tulare Lake since 1879: Great Salt Lake levels have doubtless been affected by progressively increasing diversions from its tributaries during the last century and according to Harding (1935, p. 89), “general computations of the effect of this diversion indicate that the lake would have been as high from 1923 to 1927 as it was from 1868 to 1878 if irrigation had not increased between these two‘ periods. It is also probable that the lake would not have been as low in 1934 [to 1940] as it was [in 1904—1907] had it not been A—38 for the same effect.” Harding also notes that Great Salt Lake does not fluctuate in harmony with all the variations of the lakes in the western part of the Great Basin, but does tend to follow their major movements. FLUCTUATIONS IN TREE GROWTH Tree-ring analysis has provided information con- cerning climatic fluctuations in the Southwest far bet- ter than any other source of information for the period prior to the beginning of Weather Bureau records. Douglass (1914), reporting on studies begun in 1901, noted that in properly located trees in the Southwest the relation between annual tree growth and rainfall is close. For climatologic studies it is desirable to obtain tree-ring sequences free from the efl'ects of pests, fire, and other environmental or structural factors that are more or less nonclimatic, and therefore it has been nec— essary to develop a considerable set of principles of se- lection, reduction, and interpretation of specimens. Then, to show climatic history of a region, concurrent and homogeneous sequences of annual ring-widths of trees are merged, and may be plotted as a mean growth ' curve. Most of the local variations are eliminated in averages of 5 to 10 well-selected cores. The results of dendrochronologic studies in the semi- arid regions of North America during the past half cen- tury have recently been summarized by Schulman (1956, p. 68—69) : Based on about one-third million annual rings in selected, drought-sensitive trees from semiarid sites, regional indices have been derived which are believed to represent, to a fair approximation, fluctuations in rainfall in the upper basins of all the major streams of the western United States. In most areas, where the tree growth may be closely correlated with runoff, the indices provide information about that variable also. Almost all indices are statistically well-based for about 500 years; the Colorado and Missouri indices are well documented for about 800 years. Three series of maximum length are: Colorado, 2,009 years, including the extension in archaeological beams; Snake, 1,494 years; Missouri, 973 years. A major element in the construction of these indices was the discovery and extensive sampling of a category of drought- 'recording, stunted conifers growing with extreme slowness on the most adverse sites and attaining ages twice or more the normal for the species on optimum growth sites. The oldest tree thus far discovered in each of the principal species is: limber pine, 1,700 years; bristlecone pine, 1,500 years [recent col- lections have extended these maximum ages as follows: limber pine, 2,000 years; bristlecone pine, 4,100 years]; pifion pine, 980 years; Rocky Mountain Douglas fir, 890 years; ponderosa pine 860 years. The 3,200-year Sequoia chronologies developed by Douglass and Huntington for the relatively moist southern Sierra Nevada have been reexamined in light of the highly sensitive southern California and Colorado River Basin chronologies. Some of the Sequoia series show a fairly good relation to these chronologies and thus perhaps can be taken to provide a fair first approximation to southern Sierra rainfall for the past 1,500 to 2,000 years. Too few good records in earlier DROUGHT IN THE SOUTHWEST, 1942—56 years of the Douglass series and cumulative errors in the Hunt— ington series lead to the conclusion that at least the 1,300 years of Sequoia chronology in BC. have at present very limited cli- matic value. Despite limitations in amount of material and uncertainties inherent in the data of this report, some conclusions of fair reliability may be drawn regarding long-term rainfall varia- tions in the West. Evidence is strong for the existence of a great ZOO-year wave in rainfall and runoff in the Colorado River Basin, the 1200’s extraordinarily dry, the 1300’s extraordinarily wet (more pre- cisely, perhaps, 1215—1299 and 1300—1396). The droughts of the first interval and the floods of the second appear to have far exceeded in duration and intensity those recorded by mod- ern gages. Noteworthy among other climatic events is the pronounced and extensive drought of 1573—1593, which seems to have been as severe in southwestern Montana as it was in the southern Rocky Mountains and southern California, though it apparently did not extend into Alberta. The total flow of the Colorado River during the two years 1584—85 may not have exceeded the record for low runoff recorded in 1934. Some of the indices suggest the possibility of a peculiar change in the characteristic march of rainfall some three cen- turies ago. In southern California the tendency for long-term swings to one side of the growth-mean or the other, which had been typical for some centuries preceding the mid-1600’s, gave way to swings of much shorter average duration. This tendency is very pronounced in the Colorado River Basin. 0n the other hand a reverse tendency seems to be indicated in the Missouri River Basin indices, in which shorter variations are more characteristic preceding 1650. It is perhaps no more than a coincidence that about 1645 to 1715 occurred the well-known great dearth in sunspots, for some of the representative growth indices in this report, from regions other than those just noted, apparently do not record such a change. Comparison of growth fluctuations in recent decades with those for the past several centuries suggests that in many areas of the West the interval since 1870 or so has been one of de- cidely abnormal climate. The present climatic fluctuation has taken the form of a major drought in recent decades over much of the West; it is pronounced in the Colorado River basin and particularly in southern Arizona. In the latter region the drought began in 1921, has been broken in very few years, and appears to be the most severe one since the late 1200’s. The evidence is very strong that the present fluctuation rep- resents, in terms of centuries-long dendroclimatic data, a major disturbance in the general circulation, at least over western North America. It remains to be determined whether this presages a fundamental change in type of climatic fluctuation such as seems to have occurred in some areas in the 1600’s. More light on this will probably come when more extensive, significant dendroclimatic histories have been developed. In an appendix Schulman (1956, p. 7 0—123) presents graphs showing growth curves for groups of trees in several western drainage basins. Climatic fluctuations with lengths ranging from a few years to several dec- ades are indicated in all these records, as well as in other areas where tree—ring chronologies have been de- veloped. Seeking evidence of cyclic recurrence of climatic phenomena, Schulman analyzed the data for THE METEOROLOGIC PHENOMENON OF DROUGHT A—39 110 _\ ’- ~-— ‘- ”-\-_—” ‘\ ”J-‘~/"-"\\’4"’ V . \\ \ 90 ~\ ‘\ \ +=beginning of manmade diversions \ d 70 L l l E EAGLE LAKE, CALIFORNIA .1 a 6430 “J r" _\ S —” V a: I’- < 6410 ,— b- ,’ uJ ” uJ ” LL ’/ z 6390 ’ d MONO LAKE, CALIFORNIA > LU " 220 U x 5 \/\/\/\’ \ .1 < 200 D \ Z Z “ WM m/VM 1\ IL 5 180 . , ._ g TULARE LAKE, CALIFORNIA Dry at 179 ft i < 2 4215 4205 ,—\ I; \A‘ _>‘, \V \k/w /\ 4195 \"\'_/ GREAT SALT LAKE, UTAH 150 \\ A m \’ 50 I \//\_\ \ ‘ RELATIVE SUNSPOT NUMBER, ZURICH OBSERVATORY 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 WOLF NUMBER(R) FIGURE 20.—-—F1uctuatlons of level in several lakes of the Southwest. Dashed lines are inferred. cycle—lengths in the range 15 tO 100 years. Cycles Of identified in some records. However, Schulman (1956, 20 to 24 years (approximately the length of the double p. 63) points out that “in no extensive data were cyclic sunspot cycle) were noted in data from several regions tendencies found of such regularity and strength as to in the West, and cycles up to a century in length were suggest physical reality,” and concludes that none Of A—40 the apparent cycles are systematic enough to permit long-range synoptic forecasting after the fashion of the daily forecasting of weather. SUMMARY OF DROUGHT CHARACTERISTICS In an arid or semiarid region, because it is a “have- not” region so far as precipitation is concerned, every storm is important and every dry period of whatever length is significant to someone. Dry farmers, from Hopi Indians to wheat ranchers, greet each successive rainless day with a sadder shake of the head, and wel- come each rain with a smile. To them, a drought is measured in days and weeks, for their needs are similar to those of farmers in more humid regions, although they have generally learned to get along with less. Stock ranchers are similarly dependent upon soil mois- ture, but they may avoid the pinch of short-term droughts if they are allotted suflicient rangeland—par- ticularly if the rangeland extends into highlands with their habitually greater precipitation. The majority of people of the Southwest can ignore a deficiency in precipitation extending over several months, and many are not affected adversely by a com- plete lack of precipitation throughout the growing sea- son. To these people, dry periods measured in days, months, or whole seasons are not serious. When the precipitation deficiency extends over several years or decades, however, with resulting curtailment in stream or ground-water supplies, that dry period is serious, and is properly termed a drought. This report is con- cerned especially with the effects of such dry periods upon the net supplies of surface and ground water, and the emphasis therefore is upon droughts lasting at least several years. These droughts cover periods of various length and various magnitudes of precipitation defi- ciency; each of them aggravates the water deficiency that characterizes the average climate. A considerable variety of data shows that the cli- matic history of the Southwest is comprised of a long succession of alternating wetter and drier periods; other data indicate that there have also been alternat- ing warmer and cooler periods which have influenced thewater—supply conditions in the region. The re- corded fluctuations include many that may be cyclic and of fairly regular recurrence interval. Indeed a cycle enthusiast can find so many cycles, ranging widely in period and amplitude, that he ends up with the same conclusion as the cycle skeptic: the causative factors are too little understood and the resultant precipitation too irregular to serve as an adequate basis for long— range forecasting of the climate. Nevertheless, some climatic fluctuations appear to have recurrence inter— vals sufficiently regular, and amplitudes sufficiently DROUGHT IN THE SOUTHWEST, 1942—56 great, that they may be important considerations in long-range planning for water-resource development and utilization. The droughts best shown by the available data are those having average durations of 10 to 13 years. The evidence for these droughts is especially in the Weather Bureau precipitation records covering most of the past century, in tree-ring records covering as much as 2,000 years, and in other less specific supporting data. These droughts are products of climatic fluctuations that have durations comparable to the double sunspot cycle, which varies considerably in length but has a recurrence in- terval averaging 22 to 23 years. The droughts with average durations of 10 to 13 years and recurrence intervals of slightly less than a quarter of a century do not by any means span the entire Southwest. Characteristically they are most prominent in regions where the precipitation is chiefly from a single source: in the Pacific Border where mois- ture comes from the Pacific Ocean, and in the Great Plains where moist airmasses come chiefly from the Gulf of Mexico. In both these regions the precipitation deficiency during a 10- to 13-year drought period may be equivalent to as much as three times the mean annual precipitation. Within these regions recurrent drought periods are very pronounced in some localities, fairly clear at others, and not clearly recognized at still other places. In the intervening region, which includes the Great Basin, Colorado Plateaus, and Sonoran Border zones, droughts are less regular in duration, recurrence inter- val, and magnitude. This intervening region receives its annual precipitation from several sources, but in various parts of the region one source is likely to be dominant in certain seasons, and another source domi- nant in other seasons. By analysis of precipitation by seasons, therefore, it is possible, at least in some locali- ties, to discriminate drought periods with recurrence intervals similar to those observed in the Pacific and Great Plains zones. Long-term trends are shown in records of tempera- ture, and for several decades prior to 1940 these trends were generally upward throughout the Southwest. Records from other parts of the Northern Hemisphere indicate that similar temperature trends are very wide- spread and possibly worldwide in distribution. The trend toward increasing warmth throughout the North- ern Hemisphere—and in the Southern Hemisphere wherever records are available—continued for about a century, and seems now to be changed. Records of tree growth indicate that during the past 2,000 years there have been droughts of exceptional magnitude in various parts of the Southwest, such THE METEOROLOGIC PHENOMENON 0F DROUGHT as those that are shown by analysis of tree rings in the central Pueblo area of northern New Mexico and Ari- zona to have occurred in about AD. 7 00—720, 1070—1100, 1275—1300, and 1570—1600. Exceptional droughts at various times are also inferred from historic, archeo- logic, and other data, but those data provide no basis for comparison with the conditions of recent years. These exceptional droughts may be the products of centuries-long climatic fluctuations, augmented by the shorter-period fluctuations noted in the records for the last century. However, Schulman points out that even in highly sensitive trees the sequences may contain cen- turies-long swings which are not necessarily represent- ative of a general climatic fluctuation, and he concludes (1956, p. 56) that no decisive conclusion may yet be drawn regarding the existence of a long-term trend in Southwestern rainfall in postglacial times on the basis of ring growth. Individual trees having ages approach- ing 4,000 years, of which several have recently been found in California, are likely to provide much addi- tional information concerning climatic fluctuations hav- ing lengths of several centuries. The area of Southwest drought in 1942—56, depicted on figure 6, achieves insignificance in the light of these conclusions concerning the climatic fluctuations that are responsible for drought. The map is correct for the specific 15-year period, but it does not by any means delineate an area that is especially vulnerable to drought. The 15-year period is significant because it is slightly longer than the average duration of the droughts of greatest magnitude in the past century. Because these droughts are caused by fluctuations that are almost opposite in phase in the Pacific and Great Plains subdivisions of the Southwest, a map for any 15-year period is likely to show extensive areas in the Southwest where precipitation has been significantly less than average. But, as suggested by plate 1, these areas shift from year to year. REFERENCES CITED Ahlmann, H. W., 1949, The present climatic fluctuation: London, Geog. J our., v. 112, p. 165—195. Antevs, Ernst, 1948, Climatic changes and pre—white man in The Great Basin: Utah Univ. Bull., v. 38, no. 20, p. 168—184. Brooks, 0. E. P., 1949, Climate through the ages: New York, McGraw-Hill Book Co., Inc., 386 p. 1950, Climatic fluctuations and the circulation of the atmosphere: London, Weather, v. 5, p. 113—119. Bryson, R. A., 1957, The annual march of precipitation in Ari- zona, New Mexico, and northwestern Mexico: Arizona Univ. Inst. Atmospheric Physics Tech. Rept, no. 6, 24 p. California Institute of Technology, Department of Meteorology, 1943, Synoptic weather types of North America: Pasadena, multillthed rept., 218 p. A—41 Co-Ching Chu, 1926, Climatic pulsations during historic time in China : Georg. Rev., v. 16, p. 274—282. Deevey, E. S., and Flint, R. F., 1957, Postglacial hypsithermal interval: Science, v. 125, p. 182—184. Dorroh, J. H., J r., 1946, Certain hydrologic and climatic charac- teristics of the Southwest: New Mexico Univ. Pub. Eng. no. 1, 64 p. Douglass, A. E., 1914, A method of estimating rainfall by the growth of trees: Am. Geog. Soc. Bull., v. 46, p. 321—335; see also Monthly Weather Rev., 1909, v. 37, p. 225—237. 1947, Photographic tree-ring chronologies and the Flag- staff sequence: Tree-Ring Bull., v. 14, no. 2, p. 10—16. Ellison, M. A., 1955, The sun and its influence: London, Rout- ledge and Kegan Paul Ltd., 224 p. Harding, S. T., 1935, Changes in lake levels in Great Basin area: Civil Eng., v. 5, p. 87—90. Hardman, George, and Venstrom, Cruz, 1941, A 100-year record of Truckee River runofi estimated from changes in levels and volumes of Pyramid and Winnemucca Lakes: Am. Geophys. Union Trans, v. 22, p. 71—90. Henry, A. J ., 1930, The great drought of 1930 in the United States: Monthly Weather Rev., v. 58, p. 351—354. Hiatt, W. E., 1953, The analysis of precipitation data, in Subsur- face facilities of water management and patterns of sup- ply—type area studies: U.S. 83d Cong., House Comm. Interior and Insular Affairs, Phys. Econ. Found. Nat. Re- sources, v. 4, p. 186—206. Horn, L. H., Bryson, R. A., and Lowry, W. P., 1957, An objective precipitation climatology of the United States: Wisconsin Univ. Dept. Meteorology Sci. Rept., 6, Air Force Contract 19(604)992. Hoyt, J. C., 1936, Droughts of 1930—34: U.S. Geol. Survey Water- Supply Paper 680, 103 p. , 1938, Drought of 1936 with discussion on the significance of drought in relation to climate: U.S. Geol. Survey Water- Supply Paper, 820, 60 p. Hoyt, W. G., 1942, Droughts, in Meinzer, 0. E., ed., Hydrology: New York, McGraW-Hill Book 00., 703 p. Hoyt, W. G., and others, 1936, Studies of relations of rainfall and runoff in the United States: U.S. Geol. Survey Water- Supply Paper 772, 282 p. Karlstrom, '1‘. N. V., 1956, Problem of the Cochrane in late Pleistocene chronology: U.S. Geol. Survey Bull. 1021—J, p. 303—328. 1957, Tentative correlation of Alaskan glacial sequences, 1956: Science, v. 125, no. 3237, p. 73—74. Kincer, J. B., 1933, Is our climate changing? A study of long— time temperature trends: Monthly Weather Rev., v. 61, p. 251—260. Kohler, M. A., 1949, On the use of double-mass analysis for test- ing the consistency of meteorological records and for mak- ing required adjustments: Am. Meteorol. Soc. Bull., v. 30, p. 188—189. Lynch, H. B., 1931, Rainfall and stream runoff in southern Cali- fornia since 1769: Metropolitan Water Dist, S. Calif, Los Angeles. 1948, Pacific Coast rainfall—Wide fluctuations in hun- dred years: Western Construction News, July 1948, p. 76—80. McDonald, J. E., 1956, Variability of precipitation in an arid region; a survey of characteristics for Arizona: Arizona Univ. Inst. Atmospheric Physics Tech. Rept. 1, 88 p. A—42 ,Merriam, C. F., 1937, A comprehensive study of the rainfall on the Susquehanna Valley: Am. Geophys. Union Trans, v. 18, p. 471—476. Miller, J. P., 1958, Problems of the Pleistocene in cordilleran North America as related to reconstruction of environ- mental changes that affected early man, in Smiley, T. L., ed., Climate and man in the Southwest—a symposium: Arizona Univ. Bull., v. 28, no. 4, p. 19—49. Namias, Jerome, 1955, Some meteorological aspects of drought, with special reference to the summers of 1942—54 over the United States: Monthly Weather Rev., v. 83, p. 199—205. Peck, E. L., 1956, Wasatch Front precipitation study: U.S. Weather Bur. unpub. mimeo. rept., 11 p. Piper, A. M., 1953, The nationwide water situation, in Subsur- face facilities of water management and patterns of sup- ply—Type area studies: U.S. 83d Cong, House Comm. In- terior and Insular Alfairs, Phys. Econ. Found. Nat. Resources, v. 4, p. 1—20. 1959, Derivation of a master record of rainfall: Paper before American Geophysical Union, Pacific Southwest Region, February 1959. Powell, J. W., 1879, Report on the lands of the arid region of the United States: U.S. Geog. and Geol. Survey of Rocky Mtn. Region, Washington, U.S. Govt. Printing Ofiice, 182 p. Schulman, Edmund, 1956, Dendroclimatic changes in semiarid America: Tucson, Arizona Univ. Press, 138 p. Sears, P. B., 1958, Environment and culture in retrospect, in Smiley, T. L., ed., Climate and man in the Southwest—a symposium: Arizona Univ. Bull., v. 28, no. 4, p. 77—84. Shantz, H. L., 1927, Drought resistance and soil moisture: Ecol- ogy, v. 8, p. 145—157. Spreen, William C., 1947, A determination of the efi‘ect of topog- raphy upon precipitation: Am. Geophys. Union Trans, v. 28, p. 285—290. DROUGHT IN THE SOUTHWEST, 1942—56 Stafford, H. M., and Troxell, H. C., 1953, Coastal basins near Los Angeles, Calif., in Subsurface facilities of water man- agement and patterns of supply—Type area studies: U.S. 83d Cong, House Comm. Interior and Insular Affairs, Phys. Econ. Found. Nat. Resources, v. 4, p. 21—50. Stegner, Wallace, 1954, Beyond the hundredth meridian: Bos- ton, Houghton Mifflin Co., 419 p. Tannehill, I. R., 1947, Drought, its causes and effects: Prince- ton, N.J., Princeton Univ. Press, 229 p. Thornthwaite, C. W., 1947, Climate and moisture conservation: Assoc. Am. Geographers Annals, v. 37, p. 87—100. , 1948, An approach toward a rational classification of cli- mate: Geog. Rev., v. 38, p. 55—94. , 1953, Climate and scientific irrigation in New Jersey: Johns Hopkins Univ. Climatology, mimeo. rept., 15 p. , and others, 1942, Climate and accelerated erosion in the arid and semiarid Southwest: U.S. Dept. Agriculture Bull. 808. Troxell, H. C., 1957, Water resources of southern California with special reference to the drought of 1944—51: U.S. Geol. Survey Water-Supply Paper 1366, 136 p. , and others, 1954, Hydrology of the San Bernardino and eastern San Gabriel Mountain, Ca1if.: U.S. Gieol. Survey Hydrol. Inv. Atlas HA—l, 13 pls. U.S. Department of the Interior, 1951, The drought in south- western United States as of October 1951: Washington, U.S. Govt. Printing Office, 65 p. U.S. Weather Bureau, 1953, National summary, special drought issue: Weekly Weather and Crop Bull., v. 40, no. 45, No- vember, 12 p. , 1957, National summary for drought: Weekly Weather and Crop Bull., v. 44, no. la, January, 12 p. Willett, H. 0., 1953, Atmospheric and oceanic circulation as fac- tors in glacial-interglacial changes in climate, in Shapley, Harlow, ed., Climatic change: Cambridge, Mass, Harvard Univ. Press, 318 p. Page Abilene, Tex., temperatures at....._..._.._.- 29 Abstrm‘f 1 Ahlmann, H. W., quoted ____________ - 35—36 Airmass movements _____________ . 3-9 Airmass types ________________________________ 9-11 Altitude, precipitation aflected by. See Moun- tains, precipitation aflectei by. Amarillo, Tex., precipitation at ............... 22,23 Antevs, Ernst, quoted. . 36 Aridity, defined. . ----- . 7 Atmospheric circulation ...................... 8—9 Bibliography _________________________________ 41—42 Brooks, C: E. P., quoted ___________ 6 Brownsville, Tex., precipitation at. . . 23 Bryson, R. A., quoted ________________________ 11 Canada, airmass movements from ............ 8—9 air-mass types from _________________ 9, 11 Climatic fluctuations, cyclic fluctuations ..... 23—26 historic data indicating ................... 36 lake-level fluctuations indicating. 36-38 postglacial sequence of ........ . ............ 33-36 tree-growth fluctuations indicating ....... 38-40 weather records showing .......... 17—33 Clouds, formation of .......................... 9 Colorado Plateaus. See Great Basin-Colorado Plateaus zone. Denver, 0010., precipitation at ............... 23 Dodge City, Kans., temperatures at.. - 29 Double—mass analysis, defined ................ 17—18 Drought, in arid regions ...................... 5 definitions of ..... ---- 2—3, 7—8 extent. .' .................................. 14—15 in humid regions ......................... 3—5, 7 precipitation as a factor in ................ 5—6 summary of characteristics.. .. .---......_ 40—41 “want of water” as a factor in ............ 6-7 See also particular zone. Durango, 0010., precipitation at .............. 23, 27 El Paso, Tex., precipitation at ................ 15 Evapotranspiration ..................... 1, 3, 5, 15, 36 Gleciation, continental ....................... 33 Great Basin-Colorado Plateaus zone, defined. 13-14 drought in ................................ 26 precipitation in ........................... 22, 27 INDEX Page Great Plains zone, climatic trend in .......... 1, 2 defined ................................... 14 drought in ................................ 23 precipitation in ........................ 22,23, 27 Gulf of Mexico, airmass movements from ..... 9 airmass types from. .- ................... '.. 9—11 precipitation from ..................... 22, 26—27 Harding, S. T., quoted .................... 36, 37-38 Hiatt, W. E., quoted.... ................. 11,18 Hoyt, J. C., quoted.. . 3 Hoyt, W. G., quoted ......................... 6 Imperial Valley, Calif., effect of drought in._.. 7 Indie, Calif., annual precipitation at _____ ... 14, 15 Introduction ................................. 1—2 Isohyets, defined ............................. 11 Kings River basin, Calif., variation of precipi- tation in .......................... 29—33 Lakes, fluctuating levels in ................... 36—38 Lordsburg, N. Mex., precipitation at. . 23 Los Angeles, Calif., precipitation at_. . . 18, 22, 23,36 McDonald, J. E., quoted ..................... 15,18 Mexico, airmass types from ......... .. 9, 11, 13 Mountains, precipitation aflected by.. .. 9, 11, 29—33 Namias, Jerome, quoted ...................... 14 New York City, drought in .................. 5, 7 Orographic influence. See Mountains, efiects precipitation afiected by. Pacific Border zone, annual precipitation in.. 22,27 defined ................................... 13 drought in ................................ 23 Pacific Ocean, airmass types from ............ 9, 22 air movement aflected by ................. 8 precipitation from ..................... 13, 26—27 Polar Continental airmass .................... 11, 14 Polar Pacific airmass ................. 9, 11, 13, 14, 22 Powell, J. W., quoted ........................ 1 Precipitation, annual ......................... 22—26 definitions of ............................. 14 deviations from average .................. 14, 15 requirements for .......................... 9 seasonal .................................. 26—29 weather records showing tendency of ..... 18 ‘See also individual locality. Page Prescott, Ariz., precipitation at ............... 27 Pueblo, 0010., precipitation at- .............. 23 Roswell, N. Mex., precipitation at ........... 27 Salt Lake City, Utah, precipitation at ........ 27 St. George, Utah, precipitation at ............ 15,27 San Diego, Calif., precipitation at ............ 15, 23 temperature fluctuations at ............... 29 San Francisco bay area, temperature fluctua- tions in ........................... 29 weather records from ..................... 18, 23 San Luis Obispo, Ca1if., precipitation at. . 23 Santa Fe, N. Mex., precipitation at .......... 15 Schulman, Edmund, quoted ................. 38, 39 Sonoran Border zone, defined. 13 drought in- .............................. 26 precipitation in ........................... 22, 27 Stegner, Wallace, quoted ...... 1 Summary of drought characteristics ......... 40-41 Sunspots .................................. 23, 27—29 Tannehill, I. R., quoted ...................... 8,26 Temperature, fluctuations of. . 29 Thornthwaite, C. W., quoted.. 3 Tree growth, tree-ring analysis 01. 38—40 Tropical Continental airmass.. ..- 11,13 Tropical Gulf airmass . . 9-11, 14 Tropical Pacific airmass.. 9, 13, 22 Troxell, H. C., quoted ....... . 6 Tucson, Ariz., precipitation at- .............. 15,27 U.S. Weather Bureau, quoted ................ 2 Variability coefficient, defined ................ 15 Wasatch Range, Utah, precipitation aflected by. .............................. 29 Water shortage, defined ...................... 7 Water vapor, atmospheric circulation aflected by ............................... 9 Water year ................................... 18, 27 Weather records, consistency of data of ....... 17—18 graphic presentation of data of ............ 18-22 precipitation tendency shown by ......... 18 Wiliett, H. C., quoted ....................... 33 Zone, defined ________________________________ 11 A—43 U.S. GOVERNMENT PRINTING OFFICE: 1982 0—642688 ............................. 4 . fl .. “\ . 4 .. 4, . c . .. \ . 4. .. 4 . . 4 . . . _ . 44 . . . . .4 . 44.4 .4 . . .\ . ... . .4 .4 4-0 wt; ..4 . , , . .4 4 .. . . 4 . , .4. .4. 4. \ .. , . . .. .4. o 4. . 4 . . 4 n . . . 4.l..|,..l....r,|..|...!.|...\ . .. 4.4 ...... n .u . . ... . . . . .. . . . ... .... ...“. .rmn. ......r .../.4... . . . 4. . .. ,4 I! \\\I 4. . .4. .. . . i . o ..41 .4.|.Ii.nl. .l s. . JrJlesr “4.1%. . .....v. . ...... .....w......1r..u.4..........,.... .. . ... . . . .. _. ... . . ., . 4.1: \I v . . . . . . o .4. . .. m4 .. > a ...... . .. ... ... 7 ..L I . 4 ., . . . ... ... .n -. ... ..4....... 4.5.5:... .. .. , . . 0— o . . ...4 . 4. o, . . . a .. v... . .... . u .. . x... .. .... .. .... u... . . . 4.1;... . .. " ... o ..4 .. 4. . n. .. . ...... . . . ... ....,.. . .q. .. .4. ... ,. . ..4 / ~ , 4 . . 4 . . .... .... u .4 .. .1.\...,‘ x .1 ... . . . . _ . . . 4. .. , .44 . ..N . . . . 4 Ex .. _ . ... l . _ .I u v. 4 “X .... Lin... .. . .. ... 4 n 4 ‘ 4 4 .~ . x. 1 r .. 4 .. .... 3 O . . . ., . .. .. ... x . 4 — a n o 9.4”. \. ,. . 3.. . 4 .4 . . 4 1”/ 4 . . . - . . , . 4. , . . . . ..i . . . . , . A x ,. 4x .. ... ._ . . _ ._ , .4 \+ll/ ....... ' 4 Wm %,M% ¢ .4. ,/ (MI/.70 g //////MWM/ fl, //// x/fl/ . . . . 4. . . ...... . . . . _ ). . . .... . .. .. l . .4 . ... .... . . .. . A . . , .. _. 1 .. 44 , . . _ . ._ , _ _ .. . 4'..I:I:L I, _ . _. v .4 4 r . . . ; .4 A. .. i .. .. . 4 . .. , , g. 4 «x . . . 2!. . . . . 4 1.3!... \ . . . . .. .4 . r \ c . 4. . . . ..- ..;,. . . . 4 ..afiabove R1ver Otow1 Sauble near R1ver Se - near at at at near near near near River at near at I‘urners near Budge, Ranch, Shoe- at tern er No- Charles- Do- Du— Du— Bay- Sag- Del at Edith, Mogote, Park Ranch, Ques- near near maker. Con- 30 of gales, ton, lores, rango, tango, fleld, uache, Norte, Rosa, Colo. 0010.18 View, near ta, N. San 01m- N. d1yo, year Ariz. Atiz. 0010. Colo. 0010. 0010.17 Colo. 0010. N. Mex. N. Mex. Fort Mex. Ilde- arron, Mex. N. indi- Gar- fonso, N. Mex. cated land, N. Mex. Mex. Colo 1“ ’0 , (52) 53) (54) (55) (56) (57) (58) (59) (60) (61) (62) (63) (64) . (65) (66) (67) (68) (69) 21 806. l _____________________________________________________________________________________ 834.1 58.5 60.4 69.7 47.6 77.6 70.0 :1 995. 7 1 658 224. 0 79. 5 n 1, 021 1,057 145.0 56.4 983. 7 1,007 155 o 61.0 21780 8 908.3 118 0 88. 7 833 5 983.1 130 0 1925..-. ........ 48.9 1652 9 626.8 95 7 1926..-- 122.8 533.0 61.0 710 7 787.2 128 6 44.09 517.4 . _. -- 58.6 898 1 1,290 149 5 16.98 336. 5 566.1 61.18 221.7 77.2 740 5 656. 91 15 54.08 393.3 770.8 105.4 369.6 69.2 857 4 1,123 148 0 53.51 305.1 541.6 69.24 229.1 47.3 597 5 651.6 93 00 1931._._ 39.34 64.94 130. 3 291.3 41.08 168.9 32.6 357 8 451 71 00 . . > 1932..-- 31.64 45.94 453. 0 741.7 111.1 372.6 55.6 891 2 1,400, 183 0 . . - - 6 88 28 15 213. 3 431 0 51 61 194 6 37.0 499 5 528 78 0 . . . 7.70 33.24 101.9 249.7 27.97 124. 7 24.33 338. 2 320.8 57.0 109.8 92. 49 9.1 19.85 414 0 3.88 5 87 8 78 1935..-- 21.00 44.04 305.9 567.3 99.60 317. 2 42.99 679. 0 1,143 155.0 295. 5 333 5 18.5 44.43 1,156 11.17 42 36 31 20 1936_._- 14.65 44.69 290.8 522.4 72.15 255.0 34.16 475.5 741 115.3 222.1 263.2 13.5 32. 37 981 3 4.70 9 80 18 98 1937..-. 16.05 55.96 366.6 540.6 80.44 284.1 *43. 57 572.8 1,149 169.8 320.8 424.0 25.1 65.60 1,685 16.19 32 70 28 04 1938...- 8.25 34.60 426.2 709.6 98.34 351.4 *60. 46 791.9 1,096 142. 5 314.3 332.4 23.2 51.88 1,287 5.81 27 45 15 82 1939--.- 18.41 49.80 191.9 426.1 45. 52 207.6 *46. 88 561.7 577.9 86. 21 172 0 173.1 14.8 34 01 846 0 10.03 35 79 21 97 1940.--. 9.66 58.48 216.4 360.6 39.88 149.4 *20. 35 299.5 424.9 70. 45 153. 3 171.3 13.9 25. 99 565 0 4.21 13 02 20 63 40 73 521 8 948 9 142 5 464 5 70.70 948.3 1,777 218 5 385 7 451 0 28 9 71.22 2,416 28.21 218 6 54 42 23 74 572 1 831 6 105 3 349 9 80.46 925.6 1,335 191 4 283 2 384 5 37 2 76.90 2,323 26.26 173 7 53 46 47 62 324 7 538 3 61 72 216 3 *34. 25 564.7 621 8 91 29 195 8 187 3 9 8 28.07 703 1 8.12 29 95 17 17 24. 30 448 2 768. 0 87 14 347 6 63. 53 849. 7 923 5 116 0 294 4 252.1 22 8 47. 62 1,337 13.26 38 68 24.21 37. 82 327 5 547. 4 67 85 223 0 42.39 544. 9 757 8 119 3 240 1 261 4 19 5 56.78 1,152 9.93 28 60 28 54 33 50 215 6 421 8 39 09 168 7 34.82 417.2 342 2 54 3 143 6 92 9 9 8 17.83 453 7 3.87 19 96 10 53 32 28 315 6 625 8 66 82 259 2 50.89 621.7 545 9 77 5 219 4 177 6 20 7 39. 97 750 3 7.89 13 06 11 67 . 33.18 388.8 769.0 102.4 375.1 71.29 933.8 925.8 103.2 288. 8 243.4 19 05 43.52 1,389 10.21 51 29 25 10 1949-.-. 13.97 47.18 378.2 774.9 104.1 359.4 72.41 912.6 1,064 121.9 287.1 335.3 21 15 48.71 1,327 7. 59 37 80 23 07 1950..-- 20. 06 31.42 233.0 410.2 42.56 169.8 27.49 488.5 477.3 66.03 169.7 188.0 7 23 22.41 527 4 3.21 9 96 6 47 1951..-. 4 81 19 66 138 5 324 4 30 06 143.2 22 54 310.4 327 9 52 18 122 7 98 14 7 3 20.10 365.2 3 9.63 7.06 1952-.-- 5 53 26 14 492 8 813 0 98 94 361.9 51 01 810.2 1,235 156 4 370 5 384 2 20 73 51. 71 1, 386 12 34 16.59 24.60 195 5.24 28.40 195.1 391.9 37.29 148.6 41.65 415.7 459.7 65.21 163.3 127.9 12 13 26.90 540 2 5 84 22.78 16 23 1954...- 29.81 86.73 155.6 364. 0 43.79 194. 5 26. 26 370.8 433.4 62. 51 135.2 122.5 10 6 23.80 45Q6 3 o 3 24 12.0 1955..-- 56.03 86.91 203.8 409. 7 42.38 198. 7 24.43 382.8 434.5 56.31 135. 5 117. 3 21 2 30 90 432. 0 7 42 13. 67 15 90 1956.... 6.92 20. 49 197.2 378.6 35. 89 167.3 24.31 340.7 464.7 63.41 168.4 21 124 10.32 17.51 377.1 2 06 3.62 6. 78 1957--.. 5.24 22.43 _ 504.3 798.8 102.6 394.6 80.7 801.2 1,153 162.4 325.6 22 375 25.87 44.80 1.297 9 61 23.95 25. 62 See footnotes at end of table. B12 DROUGHT IN THE SOUTHWEST, 1942—56 TABLE 1.—Yearly runofi’, in thousands of acre—feet, adjusted for storage and diversion to show approximate natural flow at selected gaging stations in the Southwest—Continued Water year ending September 30 of year indicated Pecos River near Pecos, N. Mex. (70) Gallinas River at Mon- tezuma. N. Mex. (71) Pecos River near Puerto de Luna, N. Mex.23 (72) Brazos River at Sey— mour, Tex. (73) C olorado River at Bal— linger, Tex.“ (74) Middle Concho River near Tank- ersly, Tex. (75) San Saba River at Menard, Tex.25 (76) Devils River near Del Rio, Telxfifi (77) Llano near J unc- tion, Tex. River Nueces River at La- guna, Tex. (79) Brazos River at Waco, Tex.27 (80) North Bosque River near Clifton, Tex. (81) Leon River near Belton, Tex.28 (82) Little Lampasas River at Cameron, Tex.” River at Youngs- port, Tex. (84) Guada- lupe River near Spring Branch, Tex. (85) 59. 75 22. 91 83. 02 133. 4 115. 8 119. 4 236. 4 155. 5 87. 4 128. 5 103: 1 173. 1 82. 79 174. 1 90. 19 38. 14 51.64 27. 88 17. 64 28. 61 58. 22 166. 2 See footnotes next page. GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES Divide in the United States. Although there is no storage upstream, there are diversions for about 5,000 acres of irrigated land, an acreage that has probably changed but little during the period of record. The measured runoff was not adjusted, and a considerable part of the record prior to 1928 was estimated. The station was discontinued in 1955; runoff for 1956—58 was estimated on the basis of records for the station below Blue Creek, near Virden, N. Mex. . The Rio Grande near Del Norte, Colo., drains the high mountains of southern Colorado. The runoff was not adjusted, although it is affected by storage in 6 or 7 reservoirs, by diversions for irrigation, and by 5 transmountain diversions that bring water into the basin above the gaging station. The reservoirs are small and there is usually little carryover storage from season to season; yearly runoff, therefore, is not affected greatly. The diversions for irrigation are small and are partly offset by the importations. 6. The Brazos River at Waco, TeX., drains the northern plains of Texas, south and southeast of the Pan— handle, although the western third of the drainage area is noncontributing. The runofi was adjusted for change in contents in Possum Kingdom and Phantom Hill Reservoirs since 1941, in Whitney Reservoir since 1952, and in Lake Stamford since 1954. No adjustment was made for the effect of storage in several other small reservoirs or in a large number of conservation, stock, and flood—control ponds or for relatively small diversions for irriga- tion, municipal supply, and oil—field operations. The adjustments are considered adequate, except for the unknowu effect of the several thousand ponds con- structed as(a conservation measure, mostly since 1943. Other conservation practices put into effect during the same period, such as contouring, terracing, strip cropping, and clearing of brush may also have caused OX B13 some modification of the natural runoff. As no ad- j ustment was made for these ponds and practices, the recorded runoff since the early 1940’s may have become progressively smaller than natural runoff. Bar graphs of yearly runoff (fig. 1) show some of the following similarities in the flow at all six stations: There are no long unbroken periods of either drought or high water; drought periods of several years dura- tion almost always include some years of average or even high flow; the recent drought period was preceded on all streams by high runoff in 1941, and in all streams except the Brazos River, the runoff in 1952 was above the long-term mean. The years 1951 and 1956 stand out as the years in which the recent drought was per- haps greatest in severity and geographical area affected. The graphs in figure 1 indicate that for the Southwest as a whole the recent drought was the worst within the period of record, except for the drought of 1892—1904. Graphs of the ratio of 5—year progressive average runoff to long—term runoff (fig. 2) indicate no trends or cycles common to all streams. There is an apparent trend toward lower runoff since 1910, although for the San Gabriel and Brazos Rivers this is true only for the period of the recent drought. For the Gila and Virgin Rivers this apparent downward trend may result from the fact that records for both of these streams started during a period when runoff was high everywhere in the Southwest. Frequency-distribution curves (fig. 3) are plotted as dimensionless ratios to permit comparisons between streams, despite the various lengths of record and ranges of runoff. The most noticeable difference be- tween the curves is in their slopes, which range from relatively flat for the Rio Grande and the Virgin River to steep for the San Gabriel River. These differences in slope have important implications in study of the drought and are one of the reasons that realistic com— TABLE l.-—Yearly runofi, in thousands of acre-feet, adjusted for storage and diversion to show approximate natural flow at selected gagings stations in the Southwest—Continued 1 Adjusted for diversion by Kern River No. 3 Canal. 2 Most of the estimated figures prior to 1921 from report “Utilization of water resources in southern portion of Santa Barbara County” by Quinton, Code and Hill- Leeds and Barnard, Feb. 27, 1939; since 1921, records computed on basis of changes in cogtents of Gibraltar Reservoir and other hydrologic items necessary to compute in ow. 3 Adjusted for diversions by Doheny ditch and Piru Water Co. for 1928—34. Regu- lfitfid l1>3y Lake Piru after October 1955; figures for 1956—1957 based on Piru Creek above a e iru. 4 Adjusted for diversion by Santa Paula Water Works since 1928. 5 Adjusted for diversion, storage, evaporation from reservoirs, and imported water. 6 Adjusted for diversion by Southern California Edison Co.’s canal. . 7 Adjusted for diversions by Southern California Edison Co.’s conduit and Fontana Union Water Co.’s infiltration line. 8 Runoff for 1951 was recorded as zero; adjusted runoff estimated as 500 acre-feet because of diversions in dry year. 1‘Published as Strawberry Creek prior to 1953. Adjusted for diversion by De! Rosa Water Co. for 1932—43. 10 Adjusted for diversion by Southern California Edison Co.’s canal and for storage in Big Bear Lake. 11 Published as ”near Mesa Grande" prior to 1954. Adjusted for 1954, 1955 for storage in Sutherland Reservoir and {or other hydrologic items necessary to compute inflow. Estimated record for 1929-36 based on discharge measurements made by city of San Diego. 12 Adjusted for diversion by Springdale Canal. 13 Adjusted for change in contents in Bartlett Reservoir since 1939 and in Horseshoe Reservou' smce 1945. 690-212 0—63—3 1‘ Includin flow on Tonto Creek. '5 Adjuste for diversion by Brown Canal. 13 Runoff from 5,140 square miles downstream from Coolidge Dam computed as runoff of Gila River at Kelvin less runoff of Gila River below Coolidge Dam. 17 Adjusted for change in contents in Vallecito Reservoir since 1941. ‘8 Runoff for water years 1891—1903 tram report of Rio Grande Joint Investigation. fl{lifiMissing winter periods for years prior to 1948 estimated to complete yearly run- 0 gures. 20 Records for water years 1891—95, 1906-12, 1915, 1917—19 from reports of Rio Grande Joint Investigation. Adjusted for change in contents in El Vado Reservoir since 1935. 2| Preliminary revision. 22 Based on Rio Chama near La Puente reduced by 4 percent. 23 Records for water years 1906—13, 1918, 1926 from reports of Pecos River Joint Investigation as given in Senate Document No. 109. _ 24 Adjusted for change in contents in, and diversions from, Lake Colorado City since 1949, Oak Creek Reservoir since 1953, and Lake Thomas since 1954. 25 Adjusted for diversion by Noyes Canal since 1925. 16 Records since 1932 furnished by International Boundary and Water Commission. Station discontinued August 1957. ’ 2" Adjusted for change in contents in Possum Kingdom and Fort Phantom Hill Reservoirs since 1941, in Whitney Reservoir since 1951, and in Lake Stamford since ’5 Adjusted for change in contents in Belton Reservoir since 1954. . . 2° Adjusted for change in contents in Belton Reservoir since 1954 and for diversion by Aluminum 00. of America since 1956. ’ B14 YEARLY RUNOFF, IN ACRE-FEET YEARLY RUNOFF, IN ACRE-FEET YEARLY RUNOFF, IN ACRE-FEET 400,000 300,000 200,000 100,000 0 2,000,000 1,600,000 1,200,000 800,000 400,000 0 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0 DROUGHT IN THE SOUTHWEST, 1942—56 Mean, X1896—1954 — — — 1 ——, _ J 1890 1900 1910 1920 1930 1940 1950 1960 SAN GABRIEL RIVER NEAR AZUSA, CALIF. I - Mean, —- /l889-1954 1890 1900 1910 1920 1930 1940 1950 1960 VERDE RIVER BELOW BARTLETT DAM, ARIZ. I _ Mean, 1890—1954 1890 1900 1910 1920 1930 1940 1950 1960 RIO GRANDE NEAR DEL N ORTE, COLO. YEARLY RUNOFF, IN ACRE-FEET YEARLY RUNOFF, IN ACRE-FEET YEARLY RUNOFF, IN ACRE-FEET 400,000 300,000 200,000 100,000 0 500,000 400,000 300,000 200,000 100,000 0 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 0 Mean, 1910—54 _ I _ _I 1900 1910 1920 1930 1940 1950 1960 VIRGIN RIVER AT VIRGIN, UTAH Mean, 1913—54 _ 1900 1910 1920 1930 1940 1950 1960 GILA RIVER NEAR RED ROCK, N. MEX. Mean, K1899—1954 _ ._ I- —_ — 4 _ .1 1900 1910 1920 1930 1940 1950 1960 FIG van l.—-Year1y runofl of six selected streams. BRAZOS RIVER AT WACO, TEXAS GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES B15 [\f Sun Gabrel River near Azusu, Calif. l A /\ W J acre-H GI|o River near Red Rock, N.Mex. I40. 400 acre—fr LC -z0ar meanI I 42-31”, AdeII 9/3-54, WW a Rio Grande near Del Norte. Colo. \/\\/\VirTin River at Virgin, Unlih 45- our mean, BIO-54, [56,000 can ff RATIO OF 5-YEAR PROGRESSIVE AVERAGE RUNOFF TO LONG-TERM MEAN 65—00! mum, /890-/954, 63/,400\_acre-ff A/\ , \/\\/ V V \J Verde Rwev below Bar'len Dom, Ariz. r/\ A Brazos Rwer at Waco Texas 1= \ | A A A 667:; .E n f\ /,84.9\000 amen/fr 10 man, n/HBQ-l954. A5/41200 acrwff 56-m n /599- 954, I in v T . RM ”4%?“ J V V s ‘v A! V “\J 1900 1910 1920 1930 1940 1950 |960 1890 1900 1910 1920 1930 1940 1950 1960 FIGURE 2.—Ratio of 5-year progressive average runoff to long-term mean runofl of six selected streams. parison of discharge between streams is difficult to ‘ make. For example, as read from the frequency curves, the San Gabriel River during a third of the period of record has had yearly runofi' less than 42 percent of its mean runoff. On the other hand, the Virgin River has never had annual runofi’ as low as 42 percent of its mean runoff and the Rio Grande has had but 1 year that low. Yearly runoff, so low in relation to the average \that it never or rarely occurs on the Rio Grande or the Virgin River, is common on the San Gabriel River and occurs about a third of the time. These differing slopes are evidence of streamflow variability (p. B17). ANALYSIS OF REGIONAL RUNOFF The preceding discussion and accompanying illus- trations indicate the effect of drought on six selected streams. But what happened on the thousands of other streams in the Southwest, or even at other points on the six streams? Fewer than a hundred records in the entire Southwest are of suitable length and quality for similar analysis. If these were presented in tables, diagrams, and text, as was done for the six records, the additional information would detail only the particu- lar streams included. Little would be added to knowl~ edge about runofl' in ungaged streams or about broad trends unless general conclusions could be drawn about B16 DROUGHT IN 4.0 3.0 2.0 San Gabriel River near Azusa, Calif. 1.0 0.5 0.3 0.2 0.1 RATIO OF YEARLY RUNOFF T0 1896—1954 MEAN RUNOFF p o 01 0.01 0.1 1 5 10 30 50 70 90 95 99 99.9 99.99 PERCENT OF YEARS RATIOS EQUALED OR EXCEEDED THAT INDICATED 4.0 3.0 Verde River below Bartlett Dam, Ariz. 1.0 0.5 0.4 0.3 0.2 0.1 RATIO OF YEARLY RUNOFF TO 1889—1954 MEAN RUNOFF 0.05 0.01 0.1 1 5 10 30 50 70 90 95 99 99.9 99.99 PERCENT OF YEARS RATIOS EQUALED OR EXCEEDED THAT INDICATED 4.0 3.0 20 Gila River near Red Rock, N.Mex. 99.0 who! p N 0.1 RATIO OF YEARLY RUNOFF T0 1913—54 MEAN RUNOFF 0.05 0.01 0,1 1 5 10 30 50 70 90 99 PERCENT OF YEARS RATIOS EQUALED OR EXCEEDED THAT INDICATED 99.9 9999 THE SOUTHWEST, l 9 4 2—5 6 4.0 3.0 at Virgin, Utah 2.0 1.0 0.5 0.4 0.3 0.2 0.1 RATIO OF YEARLY RUNOFF TO 1910—54 MEAN RUNOFF 0.05 0.01 0.1 l 5 10 30 50 70 90 95 99 99.9 99.99 PERCENT OF YEARS RATIOS EQUALED 0R EXCEEDED THAT INDICATED 4.0 3.0 Rio Grande near Del Norte, Colo. 2.0 1.0 0.5 0.4 0.3 0.2 0.1 RATIO OF YEARLY RUNOFF TO 1890-1954 MEAN RUNOFF 0.05 0.01 0.1 1 S 10 30 50 70 90 95 99 PERCENT OF YEARS RATIOS EQUALED 0R EXCEEDED THAT INDICATED 99.9 99.99 4.0 3.0 Brazos River at Texas 2.0 1.0 0.5 0.4 0.3 0.2 p ,— RATIO OF YEARLY RUNOFF T0 1899-1954 MEAN RUNOFF .05 0.01 0.1 1 5 10 30 50 70 PERCENT OF YEARS RATIOS EQUALED 0R EXCEEDED THAT INDICATED 90 95 99 99.9 9999 FIGURE 3.—Frequency of yearly runofl of six selected streams. 74.. v v GENERAL EFFECTS or DROUGHT 0N WATER RESOURCES B17 runoff from large areas. Such conclusions are the pur- pose of a study of fluctuations in regional runoff. Streamflow records cannot be compared or averaged by any simple means. Stream basins differ characteris- tically in their drainage areas, runoff per unit of area, and streamfiow variability. Although many methods have been utilized to reconcile these characteristic dif— ferences, none has been fully successful. Use of dis- charge in terms of cubic feet per second per square mile (csm) offsets the difference in size of drainage area and is useful in humid regions, but in the South- west it has little value because of the rapid, large, and erratic changes in discharge per square mile, often within a relatively small area. The \Vater Resources Review 1 compares streamflow records in terms of the percentage variation of the flow of each stream from its own median. Difference in streamfiow variability, brought out in the discussion of the frequency-distribution curves for the Virgin and San (iabriel Rivers (p. B15), is shown in the maps of the Water Resources Review by dotted areas, having discharges in the lower quartile of the range of flows, and by crosshatched areas, having discharges in the upper quartile. Thus the area of the Virgin River would have been shown as dotted, whereas that of the San Gabriel would not have been designated by a spe- cial pattern. The method used in this report to compare stream- flow records is roughly that used in the Water Resources Review, but allowance is made for the streamflow vari— ability of each basin. By measuring the variability of each basin and adjusting for it, runoff from all basins is put on the same basis. STREAMFLOW VARIABILITY Streamflow variability has been defined and com- puted in several ways by hydrologists, notably by Gal- ton in England in 1875, Gherardelli in Italy in 1934, Contagne in France in 1935 as cited by Wing (1950), and Lane and Lei (1950) in the United States. The com— putation of variability was made by Lane and Lei directly from flow-duration curves similar to the frequency-distribution curves shown on figure 3. Each natural stream has its own individual curve, the slope of wliich'is determined by characteristics of the basin and the time pattern of precipitation on it. Although such a curve can be drawn on the basis of a relatively short record, the frequency distribution, like the mean flow, is likely to vary within limits that depend on the period for which it is computed. Thus, although the frequen— cy distribution is as fixed a characteristic of a stream as is the mean flow, its determination requires records | Published monthly, with annual water-year summaries, by the Geological Survey of the U.S. Department of Interior, in collaboration with the Water Resou roes Branch of the Canada Department of Northern Affairs and National Resources. sufficiently long to represent an adequate sampling of runoff events. Streamflow variability is the slope of the frequency- distribution curve. Lane (1950) has shown how the slope of the flow-duration curve can be drawn if the variability is known: The shape of the duration curve can be obtained by drawing a straight-line duration curve on logarithmic probability paper with a slope such that the ratio of the discharge exceeded 15.87 percent of the time to the discharge exceeded 50 percent of the time is equal to the antilogarithm of the variability index selected. This plotting is possible because the point at 15.87 per- cent is one standard deviation from the median flow. Lane and Lei computed their variability index by deriving the standard deviation of 10 uniformly spaced points picked from a flow-duration curve. In this re- port, the streamflow—variability index for each basin has been computed as the standard deviation of the logarithms of yearly runoff. As thus computed, the variability index (standard deviation) is in logarithmic units. Logarithms of yearly runoff are used so that yearly runoff will approximate a normal distribution— one in which positive and negative deviations from the mean of the logarithms occur with equal frequency. STREAMFLOW EXPRESSED IN UNITS OF STANDARD DEVIATION The studies of regional runoff are based upon the records of 85 stream—gaging stations distributed throughout the Southwest. Of these, the records of 25 cover the period 1904—53 and 60 cover the period 1930—53. The locations of the 85 gaging stations, and the areas tributary to them, are shown in figure 4. The observed annual runoff in the period of record, adjusted for this report where necessary (and possible) for stor- age and diversions, is given in table 1. The variability index was computed for each station. As the index based on one period usually differs some- what from that based on a different period, the indexes for the 25 long-record stations were computed both for the period 1904—53 and for the period 1930—53. Table 2 gives these indexes; those based on the period 1930—53 are listed in ascending order of magnitude. Although there are notable exceptions to the rule, streamflow variability tends to be of similar magnitude for streams in the same region. In general, the streams in Utah and Colorado have the lowest indexes, those of Arizona, New Mexico, and Texas are somewhat higher, and those in California are highest. Of the 85 stations listed in table 1, 12 have records continuous for the period 1904—53. For 13 others it was possible to complete records for 1904—53 by estimating runoff for some years. The yearly runoff in logarithmic standard-deviation units for the 25 stations was com- dos: 3:2me 3 365m 5 won: 395 $803 323526 2855 S“ 2:93 mmwflfiflld HEEHW /, z :8. E x \\\/J/ tom: .2: 2 3332.3 53.. .o .3532 62.2353 :05: / // 3:23. 5 van: mus Ego. 30:53.5 52: .3. Econ 3335 \I} ‘\ / ~\ / 71 J ® “ , /. toucaohcofiom . 2:2. 2:23 :38 // \ J 2 )/ {I\ M no.6... // zo_.r / n / / mm , om082 // T _ \ Nl/uzo 22:2 /. U — t» I! || 0 _ Va 2 v‘ t O J QM 08.8 23 S _ 25 325* f/ o z o W v 10mg m ’ / Emoza \ 258:8 52:33 _ T f x _ _ \ 82 t w. N k/s) _ o _ x m z 3 m z _ x .0 $385. 8.. I r 1\ II T lr\r\/f muxoh .2250 .33 ' \\. J _ / /n mud H . o _ J. /n «Nu/Row 92 G o . / «85:03.2 _ // U m _ 3.53.2 _ NA / :/a’ ton 2 ‘ < J 1 _ \0 mm m z /, ? fl1 on w a: q a! p-IH — 584 —. 463 - 303 .729 1 — 970 . 950 1 1 533 1.252 2 053 . 828 1 —1 105 —l. 106 — — 886 —. 294 — .' .' .‘ r. — 187 —.465 — 685 1 085 1. 173 1 049 —1 249 ——.697 — 875 —1 758 —1. 696 -1 631 207 . 062 005 — 780 —.398 — 321 1 051 . 849 978 — 051 .041 — 265 — 860 —1.003 — 795 —1 273 —1. 428 —-1 499 l 664 1.390 1 463 — 823 —.221 — 387 - 788 —.494 — 278 — 057 . 197 055 — 224 . 246 119 —1 355 ~2.262 — 979 —l 465 —.980 958 —1 198 —.285 — 493 247 . 685 558 — 739 —l. 124 — 968 —1 333 —1. 721 —1 754 1956 ....... —1.301 —1. 215 —1.574 -1. 574 —1.019 —.972 -l. 160 —-1.509 —1. 375 —2.179 —1. 418 ~2. 292 —1.675 —1. 854 —l.974 —2. 116 —2. 576 —2. 321 1957 _______ —1. 483 —l. 384 —.906 —l.224 -.847 —.963 —.431 —1.050 —1.038 —2. 099 .752 —1. 797 —l. 141 —-.940 —l.628 —.490 .159 .571 690—212 0—63—4 B22 DROUGHT IN THE SOUTHWEST, 1942—56 TABLE 3.—Yearly runofl', in standard-deviation units, on basis of 1904~53 median runofi' at selected gaging stations in the Southwest—Con. Salt Salt San Fran— Gila Gila. Gila San Rillito Sonoita Santa San Florida Los Sa- Water River River cisco River River River Mimbres Carlos Gila Creek Creek Cruz Pedro Dolores Animas River Pinos guache year at near River near near near River River River near near River River River River near River Creek ending Roose- Chryso- near Gila, Red 8010- near near at Kel- Tuc- Pata- near at at Do— at Du- Duran- near near September velt, tile, Glen- N. Rock, mon, Mimbres, Peri- vin, son, gonia, N 0- Charles— lores, rango, go, Bay- Sa- 30 of year Ariz. Ariz. wood, Mex. N. Ariz. N. Mex. dot, Ariz. Ariz. Ariz. gales, ton, 0010. Colo. 0010. field, guache, indicated N. Mex. Mex. Ariz. Ariz. Ariz. 0010. Colo. (40) (41) (43) (44) (45) (46) (47) (48) (49) (50) (51) (52) (53) (54) (55) (56) (57) (58) 1889 _______ 0 .756 ____________________________________________________________________________________________ _ . _ . . 1890 _______ .921 ____________________________________________________________________________ _ ________ 0 .437 1891 _______ 1 .802 1892 _______ —2 .057 1893 _______ — .731 1894 _______ — 1 .496 1895 _______ .275 — .201 — .254 — .710 —1 .411 —2.351 — .390 ~1 .896 —1.348 —1 .360 .831 .733 .581 :877 1 .241 GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES B23 TABLE 3.—Yearly runofi‘, in standard-deviation units, on basis of 1904—53 median runofi at selected gaging stations in the Southwest—Con. San Rio Trinchera Rio Rayado Rio Juan Navajo Conejos Chama Creek Red Grande Creek at Water Grande River River River at above River at Otowi Sauble year near at at near Park Turners near Bridge, Ranch, ending Del Rosa, Edith, Mogote, View, Ranch, Questa, near San near September Norte, N. Colo. Colo. N. near Fort N. Ildefonso, Cimarron, 30 of year 0010. Mex. Mex. Garland, Mex. N. Mex. N. Mex. indicated Colo. (59) (60) (61) (62) (63) (64) (65) (56) (67) Mora River near Shoe- maker, (68) Santa Cruz River at Cun- diyo, Mex. N. Mex. (69) Gallinas Pecos 0010- Middle Pecos River River Brazos rado Concho Devils River at near River River River River near Monte- Puerto at at near near Pecos, zuma, de Sey- Ballin- Tank- Del N. N. Luna, mour, ger, ersly, Rio, Mex. Mex. N. Tex. Tex. Tex. Tex. Mex. (70) (71) (72) (73) (74) (75) (77) B24 DROUGHT IN THE SOUTHWEST, 1942—56 TABLE 3.—Yearly runofi, in standard-deviation units, on basis of 1904~53 median runofl at selected gaging stations in the Southwest—Con. Water year ending September 30 of year indicated Llano River near Junction, Tex. (78) Nueces River at Laguna, Tex. (79) Brazos River at Waoo, Tex. (80) North Bosque River near Clifton, Tex. (81) Leon River near Belton, Tex. (82) Little River at Cameron, Tex. (83) Lampasas River at Youngsport, ex. Guadalupe River near Spring Branch, Tex. (89) lllll 8 .— l .100 .575 99 -l .408 - .375 -—l .047 —2 .253 ' —1 .452 — .987 — .870 —2.173 1.106 GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES TABLE 4,—Adjustments for converting yearly logarithmic standard- deviation units computed on basis of the period 1980—53 to those computed on basis of the base period 1904—53 Station Adjustment in logarith- mig (sitand- ar -' evia- No" Name tion units Pacific border 1 Kings River at Piedra, Calif ___________________________________ —0.072 2 Arroyo Seco near Soledad, Calif __________ —. 158 3 Kaweah River near Three Rivers, Calif._ +.028 4 Tule River near Porterville, Calif .............................. —. 010 Central California, mean used for group ................... —.053 8 Santa Ynez River above Gibraltar Dam near Santa Barbara, .- Calif ___________________________________ ._.. ._ —.297 ll Arroyo Seco near Pasadena, Calif._._ — 111 14 San Gabriel River near Az“sa, Calif. — 170 26 Santa Ana River near Mentone, Calif —.299 28 Santa Ysabcl Creek near Mesa Grands, C .. —. 263 Southern California, mean used for group _________________ —. 22S Sonoran border 37 Verde River below Bartlett Dam, Ariz _________________________ 40 Salt River at Roosevelt, Ariz __________________________________ Mogollon Rim, mean used for group ...................... 45 Gila River near Red Rock, N. Mex ______ Upper Gila River, mean used for group Middle Gila River, mean used for group1 __________________ Great Basin-Colorado Plateau 29 Beaver River near Bevver, Utah _______________________________ —0. 340 30 Sevier River at Kingston, Utah__ _ —.499 31 Virgin River at Virgin, Utah ____________________________ - —.44l Southwestern Utah, mean used for group ________ _ —— 427 55 Animas River at Durango, Colo ____________________ _ — 428 59 Rio Grande near Del Norte, 0010...... _ — 453 60 San Juan River at Rosa, N. Mex... . 62 Conejos River near Mogote, Colo ........................ .. —.324 66 Rio Grande at Otowi Bridge, near San Illdet'onso, N. Mex ______ San J u‘an Mountains, mean used for group ................ 71 Gallinas River at Montezuma, N. Mex ......................... 72 Pecos River near Puerta de Luna, N. Mex _____________________ Upper Pecos, mean used for group ........................ Great Plains 74 Colorado River at Ballinger, Tex ............................... —0. 157 West-central Texas, mean used for group .................. — 157 South-central Texas, mean used for group 2 ________________ — — 157 80 Brazos River at Waco, Tex ..................................... -——()(E Central Texas, mean used for group ....................... —. 007 Arliiio iii/215th); in group with record for 190443. Mean for Salt River near Roosevelt, T 3x130 vggtig; an group with record for 1904—53. Mean for Colorado River at Ballinger, By expressing the year’s runoff in terms of logarith- mic standard-deviation units, the effects of differences in size of drainage area, in yield per square mile and in percentage of average yearly runoff, are offset or greatly reduced. These units also tend to offset the effects of the variation between basins caused by differences in nat- ural storage that result from different topography, ge- ology, vegetation, and effects of man’s occupancy. Thus comparison between two streams is possible in spite of these differences, because the logarithmic standard- B25 deviation units reflect primarily the effect of the varia- tion of precipitation on the year-to—year runoff from the basins. Figure 5 illustrates runoff expressed in loga- rithmic standard-deviation units in graphs for the six stations whose records form the basis for figures 1 and 2. The same deductions drawn previously (p. B13) re- garding severity and length of drought can be drawn again, but in addition figure 5 shows the degree to which year-to—year variations of flow in the six streams are in unison. If there is sufficient unison of variation, 0r homo- geneity, among the records from the 85 gaging stations given in table 2, they can be the basis for evaluating the drought for extensive areas in which the drainage basins whose runoff was measured are representative. If ho- mogeneity is lacking, the drought can be described only by specific areas, and overall conclusions arerunwar- ranted. To determine whether there is unison of vari- ation, or homogeneity, among the streams, statistical correlative techniques were used. The yearly runoffs, in logarithmic standard-deviation units, of pairs of streams were correlated and the degree of homogeneity was determined from the coefficient of correlation—a. coefficient of unity indicating perfect homogeneity and zero indicating no unison of variation whatsoever. A correlation coefficient of 0.7 was adopted as the mini- mum required to indicate homogeneity between two sta- tions. The 24-year period 1930—53 was used, because it is the longest period for which concurrent records are available for all 85 stations used in the study. The min- imum accepted coefficient of 0.7 is well above the sig- nificance level of 0.41 for 24 independent events and indicates that about half the variance between stations has been accounted for. Because runoff is persistent and not truly random, the number of independent run- off events in 24 years is undoubtedly less than 24, but even if the number is as low as 10 a correlation coeffi- cient of 0.62 is still significant at the 5-percent level. As a rule, the coefficients of correlation are highest between nearby streams and decreaSe with increasing distance. Among 22 stations in California, high coeffi- cients of correlation indicate that the yearly deviations of the streams from their medians are in remarkable accord. The correlation of these streams with more dis- tant ones resulted in lower coefficients; correlations with streams in Arizona sometimes gave coefficients lower than would be expected by chance. Poor correlations made it evident that not all streams in the Southwest fluctuate in any semblance of unison, and the Southwest cannot, therefore, be considered as a homogeneous unit. However, by grouping those stations whosa correla— tions indicated a high degree of unison of variation, regions were delineated such that the regional mean DROUGHT IN THE SOUTHWEST, 1942—56 B26 .3323 382% 5m 3 .muE: nosaFoUéauuSm Bantams: E fies: Eamowld ES 2m coma mmmfi ommfi mvmfi owmfi mmmH 0mm: mNmH ONmL , mama oflofi momfi .oomH mmwfi omwa mwwa . cm“: . _> .N:< .Emo tfltmm 26.8 32m wEw> .Emo .mm:~< :3: 52m .238 :mm ZOE.E mONSm le Z) .62 .z 28m 8m :3: 52m 25 m.m| >5}; _.__ _.__... ._.,.. _._. zo_.r MM 2 a: A. ....> > 4 or 22;: \m 3.. 1.7 _/._.. _._._.. : . W 3; E < c 6.7.. 5/... 5. 4....“ >, N _._; 0 Z... < :31»... _.__. g a z. Z a.» r. :N Z \ r 2 T Z £2 .4. . Z \ 1/ 4 ON Ow SllNfl NOIiVIABG-GHVGNVLS OlWHiIHVSO'! NI ‘JJONHH GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES runoff expressed in logarithmic standard-deviation units is truly representative of the regional character- istics that affect runoff. The total area of all regions so identified represents only a small proportion of the Southwest drought area, because the 85 drainage basins whose runoff records are suitable for analysis represent only about a quarter of the total area affected by drought. The remainder of the drought area is either desert with little or no runoff or areas where natural runoff has been greatly modified by man’s water devel- opments. The 85 gaging stations include many that measure runoff from high headwater areas, and the ag— gregate area embraces several of the principal water- producing areas of the Southwest. The principal regions delinated on the basis of these correlative techniques are four hydrologic zones that correspond to the meteorologic zones (Thomas, 1962) already outlined independently on the basis of studies of meteorologic factors. Thus the statistical studies of streamflow records confirm the four meteorologic zones as principal hydrologic subdivisions in the Southwest. Wherever the streams in an extensive area have similar runoff characteristics—that is, where the runoffs of all streams in a region fluctuate from their median flows with considerable consistency from year to year—these fluctuations of runoff reflect, in large part, the meteoro- logical forces acting over the region. Within each of the four meteorological zones, further subdivisions were indicated by the correlations. Table 5 shows the eventual groupings, the minimum and maxi- mum correlation coeflicients in each group, and the dis- tribution of'the ranges of the coefficients within each group. A summary of the 390 correlations involving a total of 85 gaging-station records shows that the cor- relation coefficients for 6 were between 0.60 and 0.69, 29 were between 0.70 and 0.79, 152 between 0.80 and 0.89, and 203 were more than 0.90. Thus only 11/{2 percent of the correlations resulted in coefficients less than the adopted acceptable lower limit of 0.7. REGIONAL RUNOFF CHARACTERISTICS Several illustrations Show the yearly runoff, in log- arithmic standard—deviation units, of each hydrologic zone as computed from the runoff of the streams com— prising the zone. Figure 6 shows runoff' for the two groups of stations in the Pacific border zone; the dashed line shows the mean for southern California, and the solid line shows the mean for a group in Santa Barbara County and farther to the north. Both graphs indicate a predominance of years with flow greater than median in periods centering about 1910 and 1940, and a predom- inance of years with less than median flow in periods centering about 1900, 1930, and 1950. These major B27 trends reflect the fluctuations in precipitation noted by Thomas (1962)—dry periods in 18924904, 1924—34, and 1946—56, and intervening wetter periods. There is a distinction between the two groups in the magnitude of deviation below the median during the three dry peri-/ ods. In the southern group the annual flow was gen- erally farther below the median during the 1946—56 drought and also during the 1892—1904 drought; in con- trast, the flow at stations in the northern group was farther below the median during the dry years 1924—34. Thus the records of streamflow confirm the evidence from precipitation records that the northern group of stations is near the boundary of the area affected by the recent Southwest drought. TABLE 5.—Sammary of product-moment correlations of yearly runofi‘ for stations in each hydrologic region of the Southwest, water years 19330-53 Number of correlations giving coefficients Stations comprising zones and regions (see table 1) 0.90-0.99 0.70—0.89 0.62—0.69 Pacific border meteorological zone 1—6 _____________________________________ 10 5 0 7—28 ____________________________________ 137 93 1 Sonoran border meteorological zone 37—41 ___________ 5 5 0 43—47 ___________ 9 1 0 48—50 ___________ 0 2 1 Great Basin~Colorado Plateau meteorological zone 1 5 0 35 41 2 3 12 0 Great Plains meteorological zone 73—75 ___________________________________ 1 1 1 77—79 ___________________________________ 0 3 0 80—85 ___________________________________ 3 11 1 Total _____________________________ 204 179 6 In the southern group of stations the 2—, 3-, and 5-year periods of minimum runoff occurred during the 1892—1904 drought, but the year and decade of minimum runoff occurred during the 1946—56 drought. Thus, on the Whole, these two droughts were of about equal mag- nitude in that region. On the other hand, among the group of stations farther north, the runoff in 1-, 2-, 3—, and 10-year periods was least during the drought of 1924—34. Although the driest 5 years of record in this group were 1946—50, the effect of the drought of 1924—34 upon streamflow was generally more intense than the recent drought in the central part of California. The yearly runoff in the Sonoran border zone is shown in figure 7. There are three groups of stations in this zone that include stations along the headwaters of B28 DROUGHT IN THE SOUTHWEST, 1942—56 Shaded areas indicate periods when deviation below the median(during 'drought) is greater in the southern group of stations _=__4 EXPLANATION Mean runoff of a group of stations in central California (stations I-6) Mean TJn-ot-f of a group of stations in southern California U) ’: z 3, 2.0— 2 9 q,— S > 3 l I o Lo— ,A .'\,’ m I [I \ I <2 , Il ' a \II ' 2 l ‘I < ‘, I | 5 l .’ ' i I 1. 22> o I. l \ .2 \' ‘|l\i a: \V' .;v 5 ,. o \l .1 I" I z r _ ~|.o— I: \l LI. 0 z D x >- E' —2.0— < [Ll )— (stations 7‘28) ESO 1895 I900 [905 |9|O I915 I920 |925 (930 I935 |940 l945 I950 |955 FIGURE 6.—Yearly runoff in the Pacific border zone. Mogollon Rim in Arizona—which provide water to the Salt, Verde, and Little Colorado Rivers; along the headwaters of the Gila River in New Mexico; and along the middle Gila River and some of its tributaries in Arizona. The total range of fluctuations in the period of record (41/2 to 5 logarithmic standard-deviation units) is slightly greater than that shown by the group of stations in the Pacific border zone; and the year-to- year fluctuations are considerably greater, for many of the changes in consecutive years are 3 to 4 units. The resulting graphs are less regular and offer far less evi- dence of alternate wet and dry periods than do those shown on figure 6. This may reflect, in part, the dif- ference in the precipitation pattern in the Sonoran bor- der zone, where most of the annual total originates in the Gulf of Mexico and occurs in thunder storms during the summer. In the Mogollon Rim region the 1-, 2-, 3-, 5-, and 10- year periods of least runoff all occurred during the 1892—1904 drought. Comparison of the two drought periods suggests that the average runoff in 1892—1904 was 63 percent of the 50-year (1904—53) median and in 1943—56, 75 percent. In the groups of stations in the Gila River drainage basin, whose records do not ex— tend back to the 1892—1904 drought, the 1—, 2-, 3-, 5-, and 10-year periods of least runoff have all occurred within the recent drought period. By comparison with either 1892—1904 or 1943—56, the drought of the 1930’s was of minor significance, although the runoff in certain years, such as 1934, was relatively low. Figure 8 shows the regional runoff trends for groups of streams draining the High Plateaus of southwestern Utah and the San Juan Mountain region of southwest- ern Colorado, both within the Colorado Plateau hydro— logic zone. Despite the geographic separation of these two groups of stations, there is remarkable similarity in the graphs. Generally the years of high runoff and years of low runoff coincide in the two regions. There are several indications that the stations in southwestern Utah, which are closer to the Pacific Ocean, are also more greatly influenced by it. Thus the stations in southwestern Utah recorded a dry period in 1924—34, concurrent with one recorded by stations along the Pa- cific border (fig. 6) and distinct from the 1931—40 drought recorded in southwestern Colorado and in the Great Plains. The recent Southwest drought is also more clearly marked in southwestern Utah than in southwestern Colorado and began in 1945 as it did in California. In the San Juan Mountain region, which includes the headwaters of the Rio Grande and the Arkansas, San Juan, Dolores, and Gunnison Rivers, the 10-year period of least runoff occurred during the 1892—1904 drought, but the 2—, 3-, and 5-year periods of minimum runoff occurred during the recent drought. In southwestern Utah, where records do not extend‘back to the 1892—1904 drought, the 2-, 3-, and 10—year periods of minimum run- off occurred during the 1943—56 drought. In both areas, runoff in 1934 was less than in any year of the recent drought, although the deviation below median flow dur- YEARLY RUNOFF, IN LOGARITHMIC STANDARD-DEVIAT|ON UNITS GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES Upper Gila River basin (stations 43—47) Middle Gila River basin (stations 48—50) \ Mogollon Rim (stations 37—41) | 1 l EXPLANATION Deviation below median runoff 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 FIGURE 7.—Yearly runoff in the Sonoran border zone. B29 YEARLY RUNOFF, IN LOGARITHMIC STANDARD-DEVIATION UNITS B30 IN LOGARITHMIC STANDARD-DEVIATION UNITS YEARLY RUNOFF, DROUGHT IN THE SOUTHWEST, 1942—56 Southwest Utah (stations 29—32) 1 San Juan Mountains (stations 54—66) \ 2 l Colorado River at Lee's Ferry, Ariz. (reconstructed virgin runoff) 1 A 0 —1 EXPLANATION —2 Deviation below median runoff I V 1890 1895 1900 1905 1910 1915 1920 1925 1930 9 FIGURE 8.—Yeatly runofl in the Colorado Plateau. 1940 1945 1950 1955 YEARLY RUNOFF, IN LOGARITHMIC STANDARD—DEVIATION UNIT5 GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES ing the 1930’s was generally less than during either the 1892—1904 or the 1943—56 drought. Figure 8 also shows that the trends in yearly runoff of Colorado River at Lees Ferry, Ariz., have been similar to those of streams in the San Juan Mountains. This station was not in- cluded in the statistical studies for this report, because the basin extends beyond the Southwest drought area 'as previously defined and because the runoff from the 110,000-square—mile basin has been considerably modi- fied by man. The stations in the GreatPlains include one group along the flanks of the southern Rocky Mountains in central New Mexico and, several hundred miles to the southeast, three contiguous groups deep in the heart of Texas. The subdivision into three groups in Texas has been necessary, because there is a wider range in runoff characteristics and lesser correlation between nearby stations than is common in the other meteorological zones. The yearly runoff of four groups is shown on figure 9. The Upper Pecos River basin group shows effects of the droughts of 1909—18, 1930—40, and 1946—56, which have been recorded in the Great Plains farther eastward—in Kansas, Oklahoma, and the Texas Pan- handle (Thomas, 1962). The 2—, 3-, 5-, and 10-year periods of minimum runoff were all recorded in the last of these three drought periods. In central Texas the three groups (west-central, central, and south-central) show several similarities and also several divergences in the effects upon runoff of these several periods of drought in the Great Plains. The effect of‘the 1909—18 drought upon all three groups is shown to be marked. By contrast, the decade 1930—40 was not a period of marked reduction in runoff, although 1934 was general- ly a year of very low flow. In all three groups the re- cent drought has been more intense than the earlier droughts and resulted in the 1—, 2-, 5-, and 10-year periods of minimum runoff in the period of record. The effect of this most recent drought was variable among the three groups of stations, however; two groups re- corded flow above the median in 1947—49, although the third group registered less than median flow in each of the years 1948—56. Several of the streams in central Texas drain areas that are underlain by the ground—water reservoir in the Edwards limestone (Thomas and others, 1963b), which effects the amount of runoff variously. Storage and delayed release of water may account for the trend of annual runoff shown in the graphs representing the three groups in central Texas. To an extent greater than is indicated by the graphs in figures 6—9, the years of high runoff are commonly followed by a second year of relatively high runoff, and there are indications of a long-continued declining trend in runoff after years of maximum runoff, such as 1919 and 1935 (Thomas and B31 others, 1963b). Plate 1 shows maps of the Southwest that delineate the areas where the annual runoff was less than the median for each of the years 1942—57. These areas are similar in major outline to the areas of precipitation deficiency in corresponding years (Thom- as, 1962). The inadequate coverage of the Southwest by the 85 selected gaging stations is evident from in- spection of the resulting maps, but more adequate data concerning this period of drought are not likely to be obtained. STREAMFLOW AS INDICATED BY TREE RINGS The derivation of the median runoff in the base period 1904—53 and the conclusions about the effects of drought in terms of deviation from this median lead logically to the following question: Where does this 50-year base period stand in relation to the period of occupancy of the Southwest by man?» This question is analogous to that posed about climate generally in that we have a century of record that can be projected back to earlier centuries and millenniums. Studies of tree rings have provided some of the most definitive conclusions con— cerning the climate of the Southwest during these earlier centuries. Studies of tree rings also enable us to make intelligent deductions concerning streamflow in bygone centuries. Tree-ring studies for the Southwest drought area covered by this report, except for Texas, indicate that mean runoff during the base period 1904—53 was closely representative of mean runoff for the 154-year period 1800—1953 and also for periods of 850 years or more. Thus the 50—year period was representative, despite the fact that runoff in the first part of the period was far above the mean and in the last part far below the mean; or perhaps it is more accurate to say that the 50-year runoff was representative, because the period included both high and low runoff periods, which happened to occur at about the proper time to give a representative average. The conclusion is that the figures presented in this report showing runoff deficiencies during the re- cent drought in terms of the 50—year (1904—53) median runoff show reasonably well the runoff deficiencies in terms of median runoff for much longer periods. In other words, the recent drought is a real drought; it is not, as has sometimes been conjectured, a period of nor- mal runoff following a great excess of runoff in the early part of this century. Schulman (1956, p. 65) has drawn several conclusions regarding specific areas of the Southwest from his tree- ring studies. He regards the present drought in the Colorado River basin above Lees Ferry as having started about 1930, and he says of that basin: We may conclude that (1) the average departure during the interval 1300—1396, the wettest during the past eight centuries B32 or more, equalled, and in some decades greatly exceeded the much shorter major recent maximum in rainfall and runoff of the Colorado River basin, 1905-1930; * * * (2) the current growth deficit, since 1930, does not as yet seem as severe in the upper basin as that during the interval 1871—1904, and especially after 1892; (3) the average annual growth and, perhaps to a great approximation, rainfall and runoff, during the 85-year drought of 1215—1299 seems to have been about half that during the drought since 1930 in this basin. The general conclusion would be. then, that the current drought is severe but has been exceeded in the not—too-distant past, and wet years and periods far exceeding those noted since gage measurements began are a not unreasonable possibility. Schulman’s conclusion that the current drought is not as severe in the Colorado River basin above Lees Ferry as was the drought after 1892 is in agreement with the relative severity of the two droughts as computed from runofi' records. The statement that runoff during the drought of 1215—99 seems to have been about half that during the drought since 1930 emphasizes the facts that the current drought in the Colorado River basin above Lees Ferry is not record breaking over the centuries and that, although the amount of runoff in the future is ex- pected to average about the same as that in 1904—53, there is always the possibility of a drought much worse than any known since the coming of the white man. Schulman (1956, p. 66—67) has also drawn conclu- sions regarding the drought in southern Arizona and the Gila River basin, as follows: A major difference [from the upper Colorado River basin] in recent decades is the relatively more pronounced nature of the current drought. Beginning in 1921 in the southernmost DROUGHT IN THE SOUTHVVEST, 1942—56 areas of the State [of Arizona] and particularly after 1933 in the entire Gila Basin, wet years have been very rare. It is clear that the deficiency beginning in 1934 has been the most severe since at least 1800 * * *. Reference to the 350 year series for southern Arizona and for the Gila headwaters area, suggests that no drought during that entire interval is as seri- ous as the current one * * "‘ It appears highly likely, in View of the general parallelism with the chronologies in Colorado and Utah, that this is the most severe drought since the late 1200’s * * *. It should be of significance in statistical forecasting that the total deficiency during the current drought is now (1955) greater than the total excess during the interval 1905—1920 in southern Arizona, 3 period which was probably one of the wettest in many centuries. PROBLEMS OF CORRELATION Many of the problems in correlating the fluctuations in streamflow with those in tree-ring widths are the same as the problems in the studies of tree rings as indi- cators of climatic fluctuations. Among these problems, Schulman (1956, p. 29—31) mentions the “standardiza- tion” of the growth rate by eliminating the “age trend” that is characteristic of tree growth, the inhomogeneity introduced in computation of regional tree-ring indices by reason of the various number of trees and localities for which records are available, and the growt “re- leases” in numerous trees as a result of occupancy of the region by white man. The correlation of tree rings with streamflow during the period of contemporaneous records also brings to light special problems, limitations, and requirements, some of which have been summarized by Schulman (1945, p. 36—37) : m 1: z :> 2 g South-central Texas 1: (stations 77—79) 1‘ > LIJ 9 A g 1 < \‘ o 2 <2: .— a) 9 0 2 I V |: o: < 0 O _J g —1 b“: EXPLANATION O z 3 n: >. _2 Deviation below median runoff 0-; I < i i i :3 4 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 FIGURE 9.—Graphs showing GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES B33 2 Upper Pecos‘ River (stations 67-72) 1 /\ O _ 1 E g E ‘3 2 2 3 z _ z 9 West-central Texas ' E 2 (stations 73—75) 5 S A a la] a ,5 l 1 1 1 o m n: < < o \ o z z 5 E (D (I) o ‘5’ 0 s I E / a E < w 8 3 .1 a — 1 z “: u: .5 a z z a a >- 2 _2 5 .J In: Central Texas E E. (stations 80—85) m >- >- 1 O — 1 ~— 2 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 Yearly runofl in the Great Plains. B34 All streams in the Southwest are subject to summertime flash floods, but these are usually highly local and ephemeral phe- nomena, which normally do not play an important part in the total annual runoff of the Gila, Salt and Verde Rivers. For, though the summer rainfall amounts to about half the annual total, its translation into usable runoff is very inefficient as compared with winter rainfall. Thus, the fluctuations in the water~year data * * * are largely governed by the winter storms, and so may be safely compared with fluctuations in tree growth. Both the Gila and the Salt are very nearly in the class of ephemeral streams. Draining regions where great extremes in seasonal weather are the rule, they can show fluctations of even greater amplitude in the annual runoff in successive years. In very wet years the runoff reaches relatively great extremes, for when the characteristically large losses in evaporation, tran- spiration, and other processes are met, every inch of excess rain- fall is increasingly effective. Of course, many details affect this general relation, such as the distribution of storms during the year and the frequency of various intensities and duration of storms. Thus, wet winters, such as in 1905, lead to an exaggeration in runoff * * *. Lag effects also appear for these years, so that the trees fail to indicate extremes in runoff such as occurred in 1905 and 1915; there is evident, however, a general corre— spondence in maxima and minima of growth and runoff, as well as much agreement in the details of fluctuation for most of the years. When the trees show a persistent maximum in growth, as in the late 1860’s, it is probable that one or more years of extremely heavy runoff occurred. On the whole, how- ever, as elsewhere in the Southwest, the tree curves give the drought years with greatest fidelity. Schulman has emphasized the necessity of using only ring series from living trees for comparison with pre— cipitation or runoff records, because the validity of such comparisons depends entirely on the relative width of the rings. The indices used for hydrologic comparisons are, therefore, baSed almost entirely on ring records de- rived from living trees, thus avoiding the uncertainties of indices derived by combining records overlapping in time. RESULTS OF CORRELATION STUDIES Of the large number of tree-ring records available in the Southwest, some are for geographic areas for which records of natural streamflow are short, meager, or completely lacking. On the other hand, tree-ring indices are not available in some areas where long records of streamflow are available, as for example in Texas and eastern New Mexico. Tree-ring indices are available to represent several of the regions where streamfiow records have been found to have a high de- gree of homogeneity—as for example southern and central California, the high plateaus of southwestern Utah, the San Juan Mountains of southwestern Colo— rado, the headwaters of the Gila River in New Mexico, and the middle Gila River basin and Mogollon Rim of Arizona. In addition, two indices appear to be excel— lent representatives of the upper Colorado River basin. DROUGHT IN THE SOUTHWEST, 1942—56 One test of the agreement between the tree—ring index and the runoff from an area is the product—moment correlation of the two. Schulman (1956, p. 47) shows the correlation between tree—ring indices and yearly runoff for 9 streams, of which 3 are in the Southwest. Comparing Colorado River at Lees Ferry, Ariz., with the Douglas fir series in the period 1895—1950, he found coefficients of correlation of 0.73 on a yearly basis and 0.84 on a 3-year smoothed basis; using the pinyon series for the period 1895—1948, he found that the correspond- ing coefficients were 0.51 and 0.69. For Rio Grande near Del Norte, Colo., in the period 1890—1950, he found coefficients of 0.60 on the yearly basis and 0.43 on the 3—year smoothed basis. For San Gabriel River near Azusa, Calif, in the period 1896—1950, the coeflicient was 0.67 on the yearly basis and 0.88 on the 3-year smoothed basis. Correlations giving coefficients within the same range were found for streams studied during this investigation. For example, correlation on a. yearly basis, for 1891—1951, of the upper Rio Grande tree- ring index with the runoff of the following streams gave these coefficients of correlation: Animas River at Durango, C010,, 0.62; Conejos River at Mogote, 0010., 0.81; and Rio Grande at Otowi Bridge near San Ilde— fonso, N. Mex., 0.71. For this report correlations were made between yearly regional runoff and the yearly tree-ring index con- sidered to represent the region best. The correlations were made graphically on logarithmic paper by plot- ting tree-ring index figures against runoff in standard deviation units. (See table 3.) The regional correla— tions, as shown in table 6, gave coefficients that range from 0.35 to 0.88. TABLE 6.—_Correlation of tree-ring indices with yearly runofl' for hydrologzc regions and the standard error of estimate of yearly and mean runofi based on tree-ring indices .z: a) ’a 1: u I 'U' A I a“: '6 “’73 E5 a as Essfimsw E§°§ £55 3%,} as §Egesafsgflat a‘o‘g Hydrologic region 8‘35 :13 5.23355 «0.53.332 u§3 “€25 's‘é'A sogefl °Eswa g3 °ss Ea“; was s§=°aevaesss$ m {5va es: 288%:sasssm: sea :14 e 9.. m #1 ed (1) (2) (3) (4) (5) (6) (7) Central California ______ 1896-1941 77 0. 56 ~49-95 11 —15—29 Southern California _____ 1896—1950 78 .82 —46—86 11 —14—26 Southwest Utah ________ 1904—50 (1) . 78 —22—29 8 —8-10 Mogollon Rim___- . 1889-1953 67 . 79 —42—74 17 -10-18 . ' 65 . 35 ~ 2 12 — 11-18 66 . 45 —52-109 10 ——16—34 70 . 83 —22—29 16 -6—7 51 . 71 —15-18 12 —4—5 basin. 1897-1945 (3) . 88 —l9—23 10 —6-7 l Yearly mean ring widths from Schulman (1950, table 4—D, p. 14). 2 Yearly mean ring widths from Schulman (1945, table 5, p. 38). The highest coefficient, 0.88, was between runoff of Colorado River at Lees Ferry and the 1945 tree—ring series. A major reason for this high correlation a1- GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES most certainly is the fact that this tree-ring record is the only one weighted with regard to the percentage of runoff from different parts of the drainage basin. The three lowest coefficients, ranging from 0.35 to 0.45, pertained to southern Arizona, where a large part of the yearly runoff is in the form of frequent flash floods each summer. These floods cause the yearly runoff to vary from year to year, but the variation is not fully re- flected in the tree rings because of the short duration of the floods and because they occur at a time of year when the trees normally grow but little. The low coefficient of 0.56 in central California is attributed to the insensitivity of the tree rings and to the fact that most of the trees were located north and east of the river basins used in this study. Two other significant measures of the relation be- tween runoff and tree-ring indices are shown in table 6. The standard error of estimate gives the percent- age range (column 5) within which runoff for an indi- vidual year may be estimated from tree rings. The standard error of the mean gives the percentage range (column 7) within which the mean runofl for a period covered by tree rings may be estimated. There are serious statistical limitations upon the long-term mean estimated from tree rings——for example, there is a high degree of autocorrelation within any one series of tree rings, the percentage range is too high, the tree rings may not be truly representative of the mean, tree rings do not reflect slow secular trends in climate—but no other method is known by which a better estimate can be made. Table 7 shows the relation of the mean tree-ring indices for 1904—53 to those for longer periods. The first eight columns show how estimates for 1904—53 B35 were made for the tree—ring indices: column 3 gives the runoff, in percent of median runoff, for 1904—53; col- umn 5 gives the runoff for the shorter period 1904 to the last water year for which the indices are available; column 6 gives the ratio of the two runofl’s, which is the figure by which the short-term runoff should be multiplied to calculate the 50-year runoff; column 7 _ gives the mean of the short-term tree—ring index; and column 8 gives the mean of the tree-ring index for the 50-year period 1904—53, computed on the assumption that the ratio between short-term and 50-year means of runoff is the same for the short-term and 50-year means of the tree-ring indices. All the tree-ring indices ex— tend back at least to 1800 (column 12) and, therefore, the relation of tree-rings for the 50-year base period 1904—53 to those for the 154-year period 1800—1953 was computed for all regions (column 11) on the basis of data given in columns 8 to 10. The range of the ratios given in column 11 is seen to be rather limited, from 0.95 for the upper Gila River basin to 1.08 for southwestern Utah. There is a pos- sibility that the real difference between the ratio 0.95 and the ratio 1.08 is negligible and that either may apply equally well to the Gila Basin and southwestern Utah. However, the ratios given in column 11 are the most probable ratios of mean runoff for the period 1904— 53 to that for the period 1800—1953, despite the large possible range in standard error. These ratios indicate that the mean runofi for the base period 1904—53 was close to the mean runoff in the 154-year period 1800— 1953. The average ratio for the 9 indices in table 7 is 102 percent, and if the-Colorado River basin above Lees Ferry is excluded, it is 100 percent. It is, therefore, concluded that the deviations of runoff below the 50- TABLE 7.——Relatz'ori of means of tree-ring indices for 1904—53 to those for 1800-1953‘ and for longer periods for several hydrologtc regions of the Southwest Mean of tree-ring Mean of index for Mean of Mean of tree-ring period from Runoff, in Ratio Mean of tree-ring tree-ring index for Ratio first year Ratio Table i Runoff, PM yeafi’ pegginggof column 3 trey-ring indJex index {pr 1890—1953, column 8 Firsft 1p Erhich column 8 11 11 per- or w 1c — in ex for for ase perio a juste year 0r 11 ex is _ . report by ' cent of tree-ring median, column 5 period period 1800 to since 1904 column 10 which used to column 13 Hydrologic region Schulman median, index is for period 1904 to 1904—53 year on basis tree-ring 1953, (1956) for base available 1904 to year (column given in of mean index adjusted period year given in column given in is used since 1904 1904—53 given in column column 4 column 8 on basis column 4 4 7) of mean given in column 8 (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Central California__._ 77 130.3 1941 134 7 0 97 101 l 98 1 101 4 100 4 0 98 1353 99 7 0 98 Southern Callfornla__ 78 144. 3 1950 148. 0 98 106. 3 104. 2 100. 2 99. 7 1.05 1414 98 8 1 05 Southwestern Utah--. (1) 105.8 1950 107. 2 99 2 41. 7 2 41. 3 2 39. 5 a 38. 4 1. 08 1800 ........................ gogollon Rm} ........ 67 119. 2 1953 119. 2 1 00 98. 6 99. e 99. 9 99.9 .99 1800 ........................ S DD? Gila RIVER...“ 65 125. 2 1939 134.0 93 98.6 91. 7 98. 5 96. 3 .95 1603 101.0 91 03111 ufim Mountains, 70 108. 1 1951 108.8 99 96. 5 95. 5 100. 1 99.8 .96 1375 99. 4 96 0 tone 0 River basm 49 103. 9 1950 104. 6 107.8 106. 7 99.9 99. 7 1.07 1099 98.8 1 08 a we Lees Ferry. 51 103.9 1950 104.6 99 107.6 106.5 99.4 99.2 1.07 1800 ........................ (a) 103.9 1945 105.0 99 103.9 102.9 99.7 99.6 1.03 1288 100.1 1 03 1 Yearly mean ring widths from Schulman (1950, table 4—D, p. 14). 2 Figures represent mean ring widths; indices are not available. 3 Yearly mean ring widths from Schulman (1945, table 5, p. 38). B36 year mean during the recent drought do represent devi- ations of about the same order below the mean for the 154-year period. Six of the nine tree-ring indices in table 7 cover periods beginning prior to 1800 (column 12). Column 14 shows the ratio of the mean of tree-ring indices for the base period 1904—53 to the mean for periods that range from 305 to 855 years. To the extent that tree rings reflect runofl', this is the ratio of mean runoff for the period 1904—53 to that for the 305- to 855—year pe- riods. For all the regions, except the headwaters of the Gila River, the ratios are within 1 percent of those given in column 11 for a 154-year period; this small difference suggests that it matters but little whether the mean runofl' for 1904—53 is compared to the mean for 154 years or for longer periods. QUALITY OF WATER By L. R. KISTER Wherever “water” has been mentioned heretofore in this report, the common usage of the word has been in- tended: something wet and, in fact, the most abundant of wet substances. In discussing the processes of the hydrologic cycle (p. Bl) “water” includes rainwater, river water, soil water, lake water, well water, spring water, and even sea water. Now, however, we want to emphasize the great variety that is embraced by the all- inclusive term “water,” and to discriminate waters on the basis of their physical and chemical qualities. To the chemist, water is the chemical compound H20. Long ago, when the chemistry of water was considered to be relatively simple, it was recognized that pure H2O is not found in nature but must be obtained by such processes as artificial distillation. However, as pointed out recently by Buswell and Rodebush (1956), the for— mula of water is not simply H20, and water is not a single substance. The purest water that can be pre- pared in the laboratory contains three isotopes of hy- drogen and three of oxygen, which can be combined in 18 different ways. With the various kinds of ions that can be formed from water’s atoms, pure water contains no fewer than 33 substances. Thus the formula H2O for pure water is a group designation. Natural waters are solutions, suspensions, and mix- tures of a great variety of chemical compounds and ele- ments in H20. Water in each phase of the hydrologic cycle is likely to contain measurable amounts of such impurities. Even the water precipitated as rain, snow, fog, frost, or dew commonly contains soluble and sus- pended substances. Analyses show greater concentra- tions of chloride in coastal than in inland areas; so at least part of these soluble substances evidently come from the oceans. In interior areas of the Southwest, DROUGHT IN THE SOUTHWEST, 1942—56 torrential storms wash significant quantities of dust and soluble salts from the atmosphere—materials that were picked up from the land by wind prior to the storm. Surface water and ground water may be only slightly more mineralized, or they may be far more mineral- ized, than the precipitation from which they were de- rived. - Wherever we find these waters, their quality is a product of the environment through which the water has passed since it fell as rain or snow. Because en- vironmental changes over the years are relatively slight, it is likely that the variations in quality of water at any specific point will be less than the great variations that are noted in waters from different geographic locations. Generally we assume a fair degree of uniformity in the quality of water from any individual well or spring. Although the range in quality of surface waters is greater, the water in some streams is characteristically clear; in others it is muddy; in some it is relatively pure; and in others it is charged with mineral matter. Nevertheless, there is abundant evidence of signifi- cant changes with time in the quality of water from specific sources. The quality of water in a flowing stream changes as the discharge changes. During pe- riods of low discharge, most stream waters are more mineralized than when flood flows occur. The quality of water in lakes and reservoirs changes in response to changes in quality of the inflow and also to the effect of evaporation from the reservoir. The quality of wa- ter from some wells and springs also has changed with time; many of these changes have been traced to the effects of increasing development and use of ground water, but some are clearly the result of fluctuations in the rate of natural recharge. If changes in quality result from changes in quantity of surface water, as indicated previously, climatic fluc- tuations must affect the quality as well as the quantity of water. Thus drought affects directly the quality of water in streams and lakes and aISO in some ground- water reservoirs. Drought may also have indirect ef- fects upon the quality of both surface and ground wa- ters by changing the environment through which the water moves. In any specific environment, when there is less water to carry sediment or soluble mineral mat- ter, the total amount of sediment or soluble matter car- ried must be less; and as tributaries or springs cease flowing, their contributions of mineral matter to major streams must be nil. On the other hand, the concentra- tion of mineral matter in streams may increase greatly as the volume of water is reduced. Thus the general effect of drought would be to increase the proportion of impurities in water, and yet reduce the total quantity of those impurities, because of the reduced quantity of water. GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES Continuous records of the sediment or dissolved matter in streams are obtained at few places in the Southwest—far fewer than records of the quantity of surface water. Most of the records of quality were begun during the recent drought, and very few begin as early as 1942; we have, therefore, little basis for com- paring the quality of surface water in the drought years with that in earlier and wetter years. Data that show the effect of drought upon the quality of ground water are even more meager, and all inferences drawn from the data must be tentative. The meagerness of quality data imposes still another handicap upon this report. Practically the only data suitablefor evaluation of the trends in quality with time are from a few stations on the Colorado River, Rio Grande, and Pecos River, and such evaluations belong properly in the detailed discussions of the effects of drought in individual river basins. But in order to draw any conclusions at all on the general effects of drought upon water quality, it is necessary to cite some of those details to elucidate the following discussion. QUALITY OF SURFACE WATER The longest record of the quality of surface water in the Southwest is that for the Colorado River at Grand Canyon, Ariz., which indicates both the phyical and EXPLANATION Runoff 20 Dissolved solids sediment 18 1931-40 drought 16 \ Runoff, 1926—57 mean, 14 12.4 million acre-feet RUNOFF, IN MILLIONS OF ACRE—FEET 12 DISSOLVED SOLIDS, IN MILLIONS OF TONS 1926 1930 1935 1940 B37 chemical quality of the water and also provides essen— tial data for the interpretation of changes in the quality of water in Lake Mead. Beginning in 1925, this record spans periods of greater than average runoff and drought periods when runoff was significantly less than the long-term mean. Graphs showing the annual run- off, in millions of acre-feet, and the total annual sedi- ment load and the total dissolved minerals, in millions of tons, are presented in figure 10. These graphs indi- cate that the total load transported by the river is generally greatest in years of high runoff and less in years of low runoff. Runoff was less than the long-term mean in 9 of the 14 years 1943—56, and during those 9 years the suspended-sediment load and the dissolved load also were less than average. Except in the 4 years 1953—56, the effect of this drought was less intense than that of 1931—40 (p. B28). Although the sediment load, like the streamflow, is far less during drought years than during years of normal precipitation, the proportion of sediment to water is commonly increased during drought years. This is best shown by records of the monthly suspended sediment at Grand Canyon, which has ranged from 156,000 tons in 298,000 acre—feet of water in January 1944 to 134 million tons in 2 million acre-feet of water 1953—56 , drought 520 440 360 280 Dissolved solids, 1926— mean, 10.0 million tons Suspended sediment. 1926—57 200 mean, 149 million tons 120 SUSPENDED SEDIMENT, IN MILLIONS OF TONS 4O 1945 1950 1955 1957 FIGURE 10.—A1mual runofi, dissolved solids, and suspended-sediment load of the Colorado River near Grand Canyon, Arm, 1926-57. B38 in August 1929. A plot of the monthly runoff against sediment load (Thomas, Gould, and Langbein, 1959) shows a rather consistent straight—line (exponential) relationship for the months of August through March in the period 1926 through 1950. During the months of maximum runoff (April through July) the sediment concentration was generally less than would be expected from the sediment-runoff relation defined by the months March through August. In 7 years of low runoff (1931, 1933—35, 1939—40, 1946) the sediment concentration gen- erally was greater than in corresponding months of other years. Thus in these drought years the sediment concentration was higher during the annual freshet and during other months of the year than in years of more abundant precipitation. In each of the years since 1942, the annual sediment load at Grand Canyon has been 50 to 100 million tons less than would be expected on the basis of a curve established by data for the. period 1926 to 1941, although the annual runoff and the seasonal distribution of run- off since 1941 have not been significantly and consist- ently different from those for earlier years. ‘On the basis of data to 1950 it was tentatively concluded by Thomas, Gould, and Langbein (1960) that this change in relationship was an effect of the Southwest drought. Lake Mead is within the drought area, as is most of the sediment-producing area of the Colorado River basin—the basins of the Virgin, Little Colorado, San Juan, and smaller tributaries that enter the Colorado below the mouth of the Green River. On the other hand, the principal sources of water flowing through the Grand Canyon are far to the north, in a region where precipitation was generally average or above, at least until 1952. Thus the reduced proportion of sediment ' to runoff since 1942 is attributed to reduced streamflow in sediment-producing tributaries and a corresponding reduction in their contribution of sediment to the main stem. A double-mass plot of cumulative annual runoff against sediment load at Grand Canyon (fig. 11) indi- cates a fairly consistent relation during the years of high discharge 1926—30 and a new relation, with less sediment in proportion to runoff, during the drought years of the 1930’s. In the period 1943—54 there was still less sediment in proportion to the runoff, although the runoff for several years was greater than the long- term mean. During the succeeding years of pronounced drought in the headwaters (1955—56), the sediment— runoff ratio was comparable to that during drought years of the 1930’s. DISSOLVED SOLIDS m STREAM WATER The dissolved matter in surface water is derived from the soluble minerals in rocks and soils with which the water comes in contact. When this contact is brief, as DROUGHT IN THE SOUTHWEST, 1942—56 in the case of direct runoff from rainfall or melting snow, the resulting surface water is generally low in dissolved solids. Water that enters ground—water res- ervoirs generally is subject to prolonged intimate con- tact with solid mineral matter and attains a higher dis- solved-solids concentration than it would in overland flow across the same materials. As a result, the water in a stream normally has the greatest concentration of dissolved solids when the stream is receiving all its water from effluent ground-water seepage, as during rainless periods. , The base flow of some streams includes a component from one or more sources of highly mineralized water, and the difference in concentration at high and low stages may be marked. Streams whose qualities are strongly affected by drought are those that have large and relatively constant inflows from saline springs. Outstanding examples of such streams are the Salt River in Arizona (Thomas and others, 1963c) and the Pecos River in New Mexico (Thomas and others, 1962). The salt content of the Pecos River increases consid- erably at Malaga Bend in southeastern New Mexico, where springs discharge brines (chiefly of the sodium chloride type) into the river channel (Thomas and others, 1962), as shown by the records from gaging stations upstream and downstream from the springs (fig. 12). In 20 years (1938—57), the mean discharge at the downstream station (Pecos River near Red Bluff) has been about 21/2 percent greater than that at the up- stream station (Pecos River east of Malaga), corre- sponding to a 2-percent increase in drainage area. In only 4 years (1941—43, 1955) has the mean annual dis- charge exceeded the 20-year average, and in each of those years the quantity of inflow between the gaging stations was less than 2 percent of the total streamflow. In years of less than average streamflow, the inflow between the stations is a larger proportion of the flow measured at the downstream station; in several years this inflow ex- ceeded 10 percent, and in 1954 it was more than 30 percent of the total. The increase in dissolved solids in the Pecos River between the Malaga and Red Bluff stations averaged about 500 tons per day in the drought years 1943—57, and more than 1,000 tons per day in the wet years 1941— 42. Storm runoff from a drainage area of 350 square miles was doubtless responsible for some of this increase in dissolved solids, particularly in wet years such as 1941 and 1942, and return flow of irrigation water also contributed some; but the saline springs and seeps in the bed of the river at Malaga Bend probably are re- sponsible for most of the increased mineralization during periods of low flow, or during droughts. This water rises under artesian pressure from underlying beds of halite (Robinson and Lang, 1938), presumably B39 GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES //////777// /4 III Inn- m m. . ,‘77/ 777/7/%%%u m I-llll m §fl7////77///////2 m m III] a o 7777 + d s m _II- m m II 0 0 W wo/NI/ I .wm M 7/¢/// - m m r. m c, 777/ - m T, m I. , 2m. m m, 77/// 7 m m Wax _ m Mm. 7/x7//7 5 1| a. M fivl. mm m a 3mm Ill .6 2 2,9 _ 2L n u 763 II n m on“) + w m m To 77775 w M m 004/ o w Mm ”mm _““- m m m m m mm $1 2/777 l N :5. L , m M N my ill m a w /_ m m mmm , I- p m m / + m H Um .2777 a w 0/ m mm II- m. My 0 mfl mm 772 I m m av, m L E me, . .III E m s m bd m” film 4/ + F m m n I “ow m m. Dmmm x77/VV « m 5 m m lfil / +1 mm.“ ”77/77., / $1 III- 0 o m o o o o o oo _ _ _ 7777 m w a a m w m 5 m m m 7 5 w w m m hmmeMO< “—0 $2054.5— Zq .n—LOZDK m>F_DON Z_ .mD_mOA_IU FIGURE 12.——Weighted average chloride in water from Pecos River near Malaga, N. Mex., and near Red Bluff, N. Mex., 1938—55. B40. at a relatively constant rate. In the reach between Mal- aga and Red Bluff the weighted average concentration of dissolved solids increased by more than 1 ton per acre— foot in each of the water years 1944—51 and by more than 3 tons per acre-foot in the years 1952—54 inclusive. Figure 12 provides a graphic comparison of the weighted average chloride concentration, in equivalents per million, at each station for the 20 years: small dif- ferences in the wet years 1941—43, moderate differences in most other years, and large differences in 1952—54 and 1957 when average streamflow was least. In other words, this comparison suggests that the chloride con- centration is inversely related to streamflow. The relation of mean annual chloride concentration to discharge of Pecos River near Red Blufi is shown in figure 13. On the basis of this chloride rating curve, a reasonably accurate estimate can be made of the weighted average chloride content during any year in which the mean annual discharge is known. Such a rating curve could be developed for any gaging station where a substantial part of the load of dissolved solids is contributed at a relatively constant rate, as for ex- ample from the ground-water sources near Malaga DROUGHT IN THE SOUTHWEST, 1942—56 Bend. Less than half the sodium chloride in the river near Red Bluff has come from the Malaga Bend area; the rest has come from the 19,200-square- mile drainage area upstream from Malaga. The small dispersion of points from the curve in figure 13 sug- gests that the dissolved solids throughout this drainage area may be contributed to the river at relatively con- stant rates. The variations in streamflow cause varia- tions in dilution of this saline contribution. Variations in streamflow similarly cause variations in dilution where streams enter salt-water bodies. Coastal streams characteristically enter the ocean over a “wedge” of salt water. In times of minimum flow this wedge may invade the stream channel for several miles, and in floods it is driven out to sea. At any point along the lower reach of the channel, therefore, varia- tions in streamflow may cause variations in chemical quality. Within the Southwest drought area, as de- limited in this report, few streams flow perennially into the ocean, and for these few we do not have records of the effect of the ocean upon the quality of the river water. However, such effects are measured in Trinity River at Liberty, Tex., where tides create backwater [20 Z Q j E E n. IOO ‘3 a) ,53 as E ,57 8 5 ° N ‘2‘ a D I 00 8 so 8 g 54\\ Z 5 52- 8 E E M D: a) E 60 S § .,56 N o o (46 40 39 5 -//5I g 40 47\. -/ (38 I 48’45 o\ o ,". 55 (“5| 49 44 o / 3E \ LIJ ‘ z 43 x ‘ ~ 42 20 ‘ ‘ . a ‘ s if ~ I g R LIJ 3i 0 IO 50 ICC 500 | 000 5000 MEAN ANNUAL DISCHARGE, IN CUBIC FEET PER SECOND FIGURE 13.—Discharge-chloride relationship, Pecos River near Red Bluff, N. Mex. (water years 1938-55). GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES effects when the river discharge is less than 4,000 cfs. In 1950 the stage-discharge relation was affected by tides about 20 percent of the time, and the dissolved solids exceeded 300 ppm about 30 percent of the time. By contrast, in the drought year 1952 the river at the station was affected by tides about 80 percent of the time, and the dissolved solids exceeded 300 ppm about 75 percent of the time. QUALITY OF GROUND WATER Drought alone has little effect upon the quality of water in most ground-water reservoirs. The chief ef— fect of drought is to reduce the quantity of natural dis- charge, without any marked change in quality, par- ticularly if the ground-water reservoir is recharged by precipitation and if the water is still usable at the point of natural discharge. In ground—water reservoirs that are recharged by streamflow, the quality of water is necessarily depend- ent upon that in the stream at the time of recharge, and if recharge occurs during drought the water may be more mineralized than usual. Also, if fresh ground water has a hydraulic connection with saline water—— as for example in some coastal aquifers and in fresh- water aquifers that are in contact with brine aquifers— a decrease in amount of fresh water may result in increased concentration of the water discharged. The effects of drought upon the quality of ground water may be enchanced markedly where man has been involved. By pumping from wells he may induce flow of saline water into a fresh-water aquifer that is near the ocean or saline lakes or that overlies or underlies saline—water aquifers. This pumping may not be re— lated to drought in any way or it may be indirectly re— lated in that it is necessitated by deficiencies in pre- cipitation and (or) in surface water. On the other hand, a deficiency of surface water may result in inter— ruption of irrigation, which retards leaching of saline soils and therefore the transport of the soluble salts into underlying ground-water reservoirs. The effects of drought and development upon the quality of ground— water vary from one locality to another; many types of effects are described in subsequent detailed discus— sions of specific areas. ECONOMIC AND RELATED EFFECTS By J. S. GATEWOOD and ALFONSO WILSON This section concerns the significance of drought to society. For comparison with the effects of wars, pesti— lence, and other factors that affect the welfare of man- kind, it might be desirable to express the effect of drought in dollars and cents. But the drought in the Southwest occurred during a period of increasing popu- lation, increasing industrialization, and inflation of the B41 dollar, and isolation of the effects of drought upon an expanding economy is a difficult and controversial op- eration. The following discussion, therefore, includes costs of the drought to some extent, but in many in— stances it is limited to comparisons of the production of commodities during the drought with the production in earlier years of greater water supply. In the Southwest as a whole, where water is at all times a scarce and valuable commodity, shortage of water presumably must cause some loss of income. The effect of drought upon water supplies varies greatly from place to place; and the effect upon production may range as widely, although not necessarily according to the same pattern. For example, of two streams equally affected by drought, the flow in one may be ample to serve those who depend on it, whereas the other pro— vides far less than the demand. The economic effect of drought thus depends not only upon the magnitude of the deviation below average water supplies but also upon the effect of that deficiency upon people. As an example of the reaction of people to drought, consider the water-supply situations of cities in Texas, as shown in the following table : Municipal water-supply situations in Texas during drought years [Data from Texas State Board of Health] i 1950 i 1951 ‘ 1952 1953 I 1954 i 1955 Number of cities: 1 Having less than 90—day supply__ . ._._ 12 30 40 ____________ Rationing water _________________ _ . 40 65 77 ____________ Using emergency supplies ___________________ 10 28 ............ Hauling water __________________________ 12 5 9 11 ____________ Permanent improvements: Drilling wells _________________________________ 113 79 114 96 69 Building reservoirs ___________________________ 6 8 I Number of cities diminished after 1953 because of permanent lmnrovements. All the given activities could have been caused by drought, but not necessarily so. Wells, reservoirs, dis- tribution systems, and other facilities can wear out or become clogged so that they need replacement, and they can also become inadequate with increased demand for water. Thus the figures do not provide a measure of the effects of drought but merely an indication of the increased attention given to water-supply facilities dur- ing the years of drought. The public reactions indi- cated by this table include: awareness of impending crisis (“less than 90—day supply”), reduction in use to balance available supply (“rationing water”), emer- gency operations to counteract the current shortage (“hauling water” and “using emergency supplies”), and operations that may also provide some insurance against recurrence of shortages under similar conditions in the future (“drilling wells” and “building reser- voirs”) . These reactions are characteristic of cities and B42 also of all those who obtain economic benefit from use of water. i The belt-tightening operation of reducing the use of water to the reduced supply is necessary for those who are largely at the mercy of climatic fluctuations: farm- ers whose crops, pasture, rangelands, and woodlands depend upon soil water that is replenished only by pre- cipitation. Wildlife reacts also to reduce the use of water because of its dependence upon the rangelands and woodlands. The “drought-disaster areas” desig- nated by the US. Department of Agriculture are pri- marily those where soil-moisture deficiencies have caused a substantial loss in agricultural income. Be- cause streamflow represents the residual or surplus water from precipitation after evapotranspiration, water users who depend upon unregulated streamflow are also at the mercy of climatic fluctuations. Reduc- tions in streamflow, and especially periods of no flow, obviously affect fish, run-of-the-river hydroelectric powerplants, and all those who divert water from an unregulated stream for any purpose. Temporary relief can be had by hauling water or food and feed into areas most severely stricken by drought and by the migration of wildlife from drought— stricken areas. Some cities and industries use water of high cost or inferior quality until sufficient quantities again become available from the normal sources of sup- ply. Emergency release of water from some reservoirs has been necessary for sanitary reasons, when the nat- ural flow has been insufficient for adequate dilution of the polluting wastes consigned to the stream. There is some cost to the individuals who benefit from these emergency operations, but the money, like the water, is merely transferred from one area to another. Such op- erations do not necessarily involve a loss of income re- gionally, although they change the circulation pattern of money. . The development of a perennial water supply that is adequate for use during a drought, in addition to reliev- ing water shortages of the moment, ensures a compa- rable supply during future droughts of similar magni- tude. Such developments are achieved at some cost but do not represent an economic loss if a precarious supply of water is replaced by an adequate one; furthermore, the money paid for the benefit provides income for those who construct wells and reservoirs and funds for pur- chase and maintenance of their equipment. Thus drought may spur development of adequate water supplies. The methods used during the drought to provide an adequate supply of water are no different from those required at any other time. The use of water for irri- gation frees the farmer from dependence on the vagaries DROUGHT IN THE SOUTHWEST, 1942—56 of precipitation. Irrigation water must be obtained from some type of reservoir where water accumulates during periods of natural surplus and can be withdrawn during periods of drought. Surface reservoirs or ground-water reservoirs may similarly provide adequate supplies for cities, industries, and other users. The con- struction of reservoirs and drilling of wells are thus indications of efforts to develop an adequate water sup— ply for the future. There are also several methods of obtaining greater economic benefit from the water sup- plies already developed and in use, and these pertain es— pecially to the water used nonconsumptively. They in- clude treatment and dilution of municipal sewage, recycling of water used in industry, and increased effi- ciency in irrigation. TEXAS Of the 254 counties in Texas, 245 were in the drought disaster area as of January 1, 1957, and had received Federal aid totaling $223 million. TheSe counties have a rural population of more than 1 million who live on about 282,000 farms and ranches. One of the counties most severaly affected by drought was Karnes, in south Texas, where the population was reduced 10 percent during the drought years, and 60 percent of those re- maining were on the county’s free-food program during 1956. The State’s annual production of wheat provides an indication of the effect of drought upon crops which depend directly upon precipitation, because wheat is rarely irrigated: in wet 1946, 6,835,000 acres was planted in wheat, of which 5,992,000 was harvested; in dry 1955, 4,308,000 acres was planted and only 1,508,000 acres was harvested. Almost two-thirds of Texas is rangeland—including pasture, woodland pasture, and forest—where grasses are the dominant vegetation of economic value. These grasses were severely reduced during the drought .years 1951—56, with a consequent reduction in wildlife popu- lation; according to the Texas Fish and Game Commis- sion, turkeys and quail were fewer than ever before and the deer population was reduced by malnutrition, par- ticularly in 1954. There was also a reduction in domes- tic livestock population because of the diminished feed and forage. From 1945 to 1955 the population of sheep decreased 46 percent and of hogs 52 percent; the num- ber of cattle in the western two-thirds of Texas was re- duced 17 percent during the decade, partly by shifting herds to the more humid eastern third of the State. Diminution of streamflow during drought resulted in reduction of hydroelectric-power generation. At the Devils River powerplants near Del Rio, operated by the Central Power and Light Co. of Corpus Christi, the average production in the 5 years 1951~55 was only about half the average for the 17 years prior to the GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES drought, and production in the driest year was less than 25 percent of the maximum annual production. The hydroelectric plants operated by the Lower Colorado River Authority on the Guadalupe River generated 36.5 million kwh during 1949 but less than 3 million kwh in 1956, because of decreased flow in the river. Drying of some streams, lakes, and reservoirs elimi- nated the fish population, but this destruction included numerous coarse fish too. On the other hand, some large reservoirs (for example, the Buchanan Reservoir, fig. 14) held a high proportion of their capacity throughout nearly every year of the drought. Emergency activities during the drought have in- cluded various methods of water rationing, such as lawn watering at even-numbered houses today and odd-num- bered houses tomorrow, no car washing, as well as ar- rangements for hauling water or otherwise obtaining temporary supplies during the period of shortage. In some places emergency measures were necessary despite considerable progress toward development of assured water supplies for the future. For example, the city of Dallas depended for many years upon Lake Dallas, but the storage was reduced by drought so seriously as to be inadequate by 1953. The city was also allocated 188,000 acre-feet in Grapevine Reservoir, completed in July 1952, and 415,000 acre-feet in Garza-Little Elm Reservoir, completed in November 1954, but these reser- voirs did not fill during the drought. Hence, Dallas, facing a critical shortage of water in February 1954, began pumping from the Red River, which was saline enough to give Dallas the temporary and dubious dis- tinction of using water that was more highly mineral- ized than that in any other large city in the Nation. Comparison of cotton production in 1945 and 1955 illustrates the effect of drought upon the irrigation economy, although production was influenced also by acreage allotments and other factors independent of water supply. In 1945 the total State production was 1,794,000 bales from 6 million acres, or an average yield of 143 pounds per acre; about 40 percent of the total was produced in the western part of the State. The total production in 1955 was 4,039,000 bales from 6 mil- lion acres, an average of 281 pounds per acre and an in- crease of 225 percent in total production over that in 1945; about 70 percent of the total was produced in the western part of Texas. The combination of success on irrigated farms and crop failures on farms that de- pended on precipitation produced a marked trend in the agricultural economy from small dryland farms to large irrigated farms. In many instances this has meant a geographic shift, involving abandonment of small farms and development of new acreage Where ample water supplies are available for irrigation. One B43 result of this trend was a reduction from about 330,000 farms and ranches in Texas in 1950 to 290,000 in 1956. The drought encouraged conservation practices such as the use or reuse of water that once went to waste. Sewage-treatment plants have converted municipal waste water for use, and the sewage effluent from San Antonio and Lubbock, for example, is used for irriga- tion. Industrial plants, by recycling cooling water, have been able to operate with less makeup water or to expand operations with no increase in intake. Increas- ing numbers of canals and ditches carrying irrigation water have been lined with concrete or replaced by underground conduits since 1945; this practice is of especial value where the water lost from the canals by seepage could not possibly be recovered from ground- water reservoirs. In many areas consumptive waste of water has been reduced by eradication of such native vegetation as saltcedar and water hyacinth. Conserva- tion is practiced by some people all the time, but there is more universal attention to it during periods of water deficiency. NEW MEXICO The drought of 1942—56 was longer and more severe in New Mexico than in any other State in the Southwest. The entire State has been recognized as a drought- disaster area. In most places even the average precip- itation is insuflicient for cultivation of crops without irrigation, but there is some dry farming, principally of wheat, in the eastern part of the State; as much as 60 percent of these crops failed during the dry year 1956. The range in all parts of New Mexico deteriorated because of deficient precipitation year after year. From 1951 to 1956 the number of cattle was reduced each year owing to scarcity of forage. The seriousness of the de— pletion of the range by drought has been pointed up in a report by the Rocky Mountain Forest and Range Ex- periment Station (1956, p. 61) : “Over the 40—year period 1915 to 1954 the average ba- sal density of black grama grass (Boulteloua eriopoda), the most important forage plant on the. range, on quad— rats protected from grazing has varied greatly. Density was reduced to 0.3 percent of the surface area in 1923 as the result of the dry period starting in 1916. With the return of rains, density of black grama again in- creased until in 1933 it was 9.5 percent, the maximum for the period of study. As the result of the current drought, black grama has disappeared from the quad- rats. The density each year is correlated with the amount of rainfall received during the preceding 15 months * * *. Such variations in plant cover affect the livestock production. Over the 40—year period, stocking has been virtually eliminated twice as the re- sult of drought, while in most favorable years stocking DROUGHT IN THE SOUTHWEST, 1 9 4 2—5 6 B44 Anne—Em ha 55% 8a “5986 S 93th Bunch. uni—6 26:82 @3029. .8 £333.82“ backwli auburn m mm._.<>> H OX3 mmmm 0mg mvm mmmH __ __ _ _ 5.6 .521 xwm £3 cam .3220: 9.5.. a, $2-98 832 2588 I'll-l1 w, vl om OOH omH .N_.< cw>E 35> .m._oEomwz 8:320: +tw_tmm OOH $2.28 .200 :35 351 m3 _ 7 _ $3.28 832 5388 __ __ 533me S _8_ _m> 00m mN om mm 00H mNH 380V :10 SGNVSflOHi NI ‘EIOVHOLS 133:! m mmh<>> mmmfl 0mm" magma ovmm mmmfi A.» .Slbaloomefl 368.3 .xos. .z .mEmco 2m £0233. 93> E r: . .35. .2 52¢ $5950 :63va 35:00 I. l- 68.28 0855 €88 am .308 3% ‘5 «a; _ wmxfl. cuzm 32200 . B ucm um $9.2m _ :oicmmm cmcmzoam Lam» mctau mmSBm E:E_xms_ 33.28 000 Nam 56.3 mo ‘om OOH omH mNH 0mm mum omN 00m 0mm 000 H omNH lBEId'EIHOV JO SONVSHOHJ. NI ‘EIOVHOLS GENERAL EFFECTS OF DROUGHT ON'VVATER RESOURCES as high as 20 acres per animal unit has seemed satisfac— tory.” Many irrigated lands that have a reasonably depend— able supply in most years received reduced supplies in some drought years because of the diminished base flow of the streams or because of inadequate reservoir stor- age. Many reservoirs in northern New Mexico were emptied in the summer of 1956 for the first time since they were built. This happened to Costilla Reservoir on Costilla Creek (a tributary of the Rio Grande) , built in 1920 with a capacity of 15,000 acre-feet; to Eagle Nest Reservoir on the Cimarron River (a tributary of the Canadian River), built in 1918 with a capacity of 70,000 acre-feet; and to Santa. Cruz Reservoir on the Santa Cruz River (a tributary of the Rio Grande), built in 1929 with a capacity of 4,500 acre-feet. El Vado Reservoir on Rio Chama (a tributary of the Rio Grande), built in 1934 with a capacity of 197,500 acre- feet, was emptied for the first time in August 1951 and has held little water since (fig. 14). Conchas Reservoir an the Canadian River did not drop below 85 percent of capacity until 1951 but was drawn upon heavily in subsequent dry years (fig. 14). The effect of drought on irrigated cropland was felt most severely in 10 counties in northern and central New Mexico where irrigation is carried on mostly by individuals and small irrigation districts whose water is taken directly from streams on which there is little or no storage. The drought had a disastrous effect upon small subsistence farms along tributary streams and narrow creek bottoms, for it forced many marginal farmers out of agriculture. The New Mexico Economic Development Commission estimates that in the 10 counties from 1940 to 1955 the population decreased 10 percent, agricultural employment decreased 30 percent, and the number of farms in operation decreased 30 to 40 percent. Although the drought was mainly respon- sible for these changes, migration from drought-stricken farms was encouraged by relatively high wages in other areas. Because of idleness or abandonment of large numbers of small irrigated farms and dry farms, the total acreage under cultivation in New Mexico de— creased during the drought years. The drying of streams and reservoirs caused the loss of many tons of fish. N0 figures are available as to the effect of drought upon wildlife, although some depletion is presumed. The high mountains in some parts of New Mexico, however, provide an advantage not available in Texas, for by climbing a few thousand feet an animal can reach a zone of higher rainfall; in the plains an animal might have to travel hundreds of miles for the same advantage. Emergency measures to relieve water shortages dur— ing the drought included hauling of water for domestic B45 and stock use and cloud—seeding operations in various parts of the State. In 1951 weather—modification con- tracts were in force for more than three-fourths of the State. Despite cloud seeding, it was one of the driest of all years throughout New Mexico, indicating a de- ficiency in atmospheric vapor as well as in precipita- tion. An unusual form of emergency measure was a court injunction to prevent draining of Elephant Butte Reservoir, on the ground that it would drydrock thou— sands of fish and thereby create a. health hazard for the city of Truth or Consequences, 3 miles downstream and sometimes downwind from the reservoir. The capital city, Santa Fe, developed additional sur— face storage during the early years of the drought by tripling the capacity of its McClure Reservoir in 1947. Nevertheless, water rationing was necessary in 1951 and six wells were drilled soon thereafter to provide supple- mentary water. High runofl' caused McClure Reservoir to fill and spill in June 1952, but it was necessary to pump ground water as a supplementary supply most of the time after that year. The story of Santa Fe is typical of many places where storage diminished and shortage increased during series of dry years, even though there was some replenishment of water during occasional wetter years. In many 10- calities supplemental ground-water supplies were de- veloped, and one result has been increasing dependence upon ground water especially for irrigation, which is the principal use of water in the State. According to rough estimates, the area irrigated exclusively by sur- face water decreased from about 385,000 acres in 1950 to 293,000 in 1955; the area irrigated from wells when stream supplies are not available increased from 80,000 acres in 1950 to 131,000 in 1955; and the area irrigated exclusively from wells increased from 300,000 to 445,000 acres in the same period. Chiefly because of the increased use of ground water, the overall irrigated acreage in the State increased 14 percent in the 5—year period despite the abandonment of some acreage be- cause of lack of water. COLORADO As of the end of 1956, 33 of Colorado’s 63 counties had been declared drought-disaster areas, and the Fed— eral Government had spent nearly $31 million for drought aid, mostly in counties other than the 7 in the southwestern part of the State. The economic loss of the southwestern counties resulting from the drought was more than offset by the tremendous expansion of uranium mining during the period. Thus the towns and general economy have grown and been prosperous despite drought. The part of southern Colorado considered in this re— port is limited to these 7 counties, in the San Juan and Dolores River basins and the Rio Grande basin, where runoff was above the long-term mean in only 4 of the 14; years 1943—56. The forage in these 7 counties suf- fered severely, even at high altitudes. The mountain meadows, an important source of feed for livestock, were frequently dry and the stock forage inadequate during the drought. Drought caused some reduction in power generation at the several small hydroelectric plants in the area, and some municipalities were short of water. UTAH The southern half of Utah was affected by drought for as long as any part of the Southwest, and the southern two-thirds was included in the Federal Government’s drought-aid program as of the end of 1956. The livestock industry probably felt the drought more severely than any other part of the State’s econ- omy. In some areas, the carrying capacity of the range in 1956 was reduced about 40 percent. In other areas, as early as 1951, parts of the range were unused because of failing springs and dry waterholes. In 1951 and in some years since, water was hauled to sheep and cattle in Sevier, San Juan, Beaver, and Kane Counties. Effect of the drought on hydroelectric-power production is illustrated by data from the Southern Utah Power Co., serving Washington and Iron Counties. The annual generation of hydroelectric power exceeded 13 million kwh in 1952 and was greater than 11 million in each of the wet years 1941—45 and 1947—50. In the drier years the generation was generally less than 10 million kwh and reached a low of 7.7 million in 1956. NEVADA It may appear anomalous that southern Nevada, one of the most arid parts of the United States, has recently experienced a drought more intense than any since 1904, and yet the economic effect has been less than in any other of the seven Southwestern States. The explana- tion is that because of the prevailing aridity, there are fewer people to be affected; and those few utilize chiefly the water supplies of greatest dependability. Exclud- ing the Las Vegas metropolitan area, southern Nevada has fewer than 25,000 people in an area larger than the State of Indiana. Many of these people obtain water from springs of fairly uniform discharge or from ground-water reservoirs of large storage capacity. Lo- cal runofl' has long been recognized as undependable, because it occurs only as a result of exceptional storms; thus the only developed surface-water supplies are those brought from other States by the Virgin and Colorado Rivers, and the flow in the Colorado River, as regulated by Hoover Dam, was adequate for the small requirements in Nevada. Thus the drought af- ’ B46 DROUGHT IN THE SOUTHWEST, 1942—56 fected chiefly the domestic and wild products of the rangeland, which deteoriorated during the years of less than average precipitation. The effect of drought on the rangeland was severe enough that the four southern counties were declared a drought-emergency area by the Federal Government and eligible for aid to cattle- men. This aid was given mainly as subsidies to buy hay. ARIZONA All Arizona in 1956 was within the drought-disaster area eligible for Federal aid for livestock feeding. Spe- cial relief was given to the Navajo Reservation by dis- tribution of 28,000 tons of feed grain to tribal stock owners in late 1956. As suggested by the type of relief, those parts of the economy that depend upon soil mois- ture obtained directly from precipitation were severely affected by the drought: livestock and wildlife depend- ent upon the range, and dry farming. Dr. Robert Humphrey, range-management specialist of the Uni- versity of Arizona, after an inspection of the ranges of southern Arizona in December 1956 said that condi- tions varied from a 10 percent kill of perennial grasses in some areas to as high as 80 or 90 percent in others. The Flagstaff office of the Forest Service reported that after 1947 the grazing period was progressively reduced until in 1956 it was only about 55 percent of that per- mitted in 1947. Few figures are available to show the adverse effect of drought on wildlife, but studies by the Arizona Game and Fish Department of the antelope, deer, and elk population indicate lower population, lower vitality of the survivors, and lower reproduction. Drought may afl’ect the quality of wildlife food, as well as the quan- tity, as indicated by a study of quail by the Arizona Cooperative Wildlife Research Unit of the University of Arizona in 1956; it was found that the quail are now taking less preferred foods and covering more territory to fill their crops. The lack of specific nutrients is perhaps more serious than a general food shortage. Water may be the “nutrient” that is lacking during seasons when there is no succulence in the form of green growing plants or juicy fruits. Certain vitamins may also be unavailable when the birds are on a diet of dry seeds without green leaves or suc- culent fruits. L Only about 50,000 acres is farmed without irrigation in Arizona, mostly in the northern part of the State. In the dry-farming area in the vicinity of Flagstaff, the Agricultural Stabilization Conservation office re- ported a reduction in acreage of pinto beans and small grains from 20,000 in 1949 to 14,000 in 1951, an increase to 18,000 after the wet year 1952, and reduction to 10,000 acres by 1956. Crop production in 1956 was poorer than the number of acres would indicate, for beans produced only about 30 percent of normal yield per acre. GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES Many activities that depend upon surface water also have been adversely affected by drought. Recreational facilities have been affected by drying up of several natural lakes in the northern part of the State. For example, Mormon Lake, south of Flagstaff, has 3. nor- mal surface area of 81/2 square miles and is fed by the natural runoff from 38.3 square miles, mostly a high pine-forested plateau. The lake has never been known to overflow and therefore is a measure of the residual runoff from its drainage area. The lake was dry .at several times during the drought that ended in 1904, but so far as known it contained water throughout the period 1905—46. Since then the lake has been dry in parts of the years 1947, 1948, 1951, and 1953—56. Sev- eral other lakes and springs in the vicinity of Flagstaff have had histories similar to that of Mormon Lake. Emergency measures during the drought included some rationing of water, hauling of water 40 miles by rail for the city of Williams, and contracts for cloud seeding to increase precipitation. The city of Yuma, obtaining water from the Colorado River, had con- siderable expense in maintaining a channel between its intake and the meandering river in periods when almost all water was diverted upstream at Imperial Dam. New sources of supply were developed, notably by pumping from underground reservoirs. In fact, de- velopment of new wells enabled municipal and irriga- tion uses to expand during the drought. New supplies have also been developed by surface storage, such as the modification of Horseshoe Dam in 1950, when the city of Phoenix increased the height of the dam by 4 feet and installed spillway gates to provide 76,100 acre— feet of additional storage, available for release down the Verde River as needed (fig. 14). CALIFORNIA In southern California the drought of 1945—56 was of comparable magnitude to that of 1894—1904, and more severe than any other dry period in 75 years of rec- ord (p. B27, 136). Nevertheless, the economic effects of this latest drought were minor, partly because the dry period was interrupted by the wet year 1952 but chiefly because the water shortages due to natural causes were overshadowed by thOSe created by expanding require- ments of a rapidly increasing population. In the boom created by urbanization and industrialization, the effect of the drought upon the State’s economy was practically negligible. The relative insignificance of the drought is indicated by the fact that no part of the State was designated a drought-disaster area eligible for Federal aid. Because most of California’s precipitation falls dur- ing the winter, relatively few agricultural products depend upon soil water derived directly from precipita- B47 tion. Those products—grass suitable for winter forage and hay, grains, and vegetables and other crops that can grow throughout the mild winters of several coastal valleys—grow during the season of least evapotranspi- ration and greatest water surplus, and their yield is dependent more upon the timing and intensity of the rain than upon the annual total. The deficiency in streamflow during the drought years had some effect upon hydroelectric power, and upon recreational facilities dependent upon the stream— flow. The production by hydroelectric plants on streams in the southern Sierra Nevada fluctuated some— what with the streamflow in the drought years; it ranged from 512 million kWh in dry 1949 to 809 million in wet 1952, and was greater than 710 million kwh in 8 of the 11 years 1945—55. The detrimental effects of drought upon recreational activities resulted chiefly from reduced inflow to lakes and reservoirs. For ex- ample, Elsinore Lake southeast of Los Angeles was dry in August 1951 for the first time since records began in 1915, and according to newspaper reports for the first time since 1859; in 1952 there was some recovery but the lake was dry about half the time during the 5 years 1952—56. Similarly, in Big Bear Lake and other reservoirs where storage is regulated for irrigation or municipal use, the water in storage in 1951 was less than it had been for many years; there was some in- crease in 1952, but the storage by the end of 1956 was again near the minimum of record. Although some rationing or other emergency meas- ures were necessary in various localities during the drought years, these were overshadowed by activities seeking more permanent alleviation of water shortages, either by conservation of supplies already developed or by development of additional supplies. Among the conservation measures might be noted the shift to agri- cultural products of lower water requirements, as for example in the Camarillo and Santa Paula districts of Ventura County where some lemon groves were re- placed by lima beans because of the scarcity of water. The city of San Diego has used a water—saving tech- nique since 1946 to reduce evaporation losses from its reservoir system: Water that could be stored either in El Capitan or in Cuyamaca Reservoir has been stored when possible in El Capitan Reservoir, because its stor— age capacity per acre of surface area is more than 6 times that of Cuyamaca Reservoir, and evaporation losses are therefore leSS per unit of stored water. Great strides in water conservation have been made by various industries in southern California, as pointed out by Pickett (1956) : * * * For example, when plans were first developed to es- tablish the Lever Brothers Soap and Edible Oils Plant in Ban- B48 dini, we were told that similar plants in other sections of the United States were producing as much as 8 million gallons :1 day of liquid industrial waste. Because of the hazard of water pollution and nuisance, we determined that these wastes could not be discharged to the ground or to the drainage channels in the vicinity of the proposed plant. Also, the limited capacity of the sanitary sewer system precluded the discharge of such quantities of waste from a single industry. Finally, the cost of water in the Los Angeles metropolitan area indicated that wastage of the quantities mentioned would be unfeasible. In this case the industry employed sanitary engineers who were able to develop methods for reclamation and reuse of water from various industrial operations to such an extent that the plant is now wasting only 0.236 million gallons a day to the sewer system * * *. * * * We have encouraged many industrial plants including industries manufacturing paints, synthetic rubber, automobiles, aircraft, and hundreds of other products to install facilities for the treatment and reuse of process wastes. Of great interest to us have been the changes at some of our major refineries where management and engineers have been actively engaged in development of comprehensive programs for such reclama- tion, reuse, and conservation of all available water * * *. In most of southern California the maintenance of an adequate water supply throughout the dry years has been achieved by importations, of which some have been going on for a long time. For example, an aqueduct has carried water from the Owens Valley to the city of Los Angeles since 1913, and practically at its capacity of 330,000 acre-feet per year since 1950. The Colorado River aqueduct of the Metropolitan Water District first delivered water to southern California in 1943; the imports increased from 50,000 acre-feet in 1945 to 430,000 in 1956. These figures include the water car- ried by the San Diego aqueduct, whose first barrel was completed in 1947 and second barrel in 1954. The Colorado River has been the source also of progressively increasing quantities of water for irrigation in Imperial and Coachella Valleys. In the southern part of the San Joaquin Valley, the Friant-Kern Canal began to deliver water southward from the San Joaquin River in 1949, and the deliveries increased progressively from 184,000 acre-feet in 1950 to 1,322,000 in 1956. The irri- gated area in the southern part of the San Joaquin Valley increased from 1% million acres in 1945 to nearly 2 million acres in 1955. In the south coastal area which includes the metro— politan areas of Los Angeles and San Diego, the irri- gated acreage decreased substantially during the drought, but there the agricultural land was a casualty in the increasing urbanization, as shown by the figures for the heavily urbanized and industrialized Los Angeles County, beginning as early as 1920. (Seefol- lowing table.) It is likely that, had there been no urbanization, some agricultural land would have been forced out of pro- duction during the drought because of increasing cost DROUGHT IN THE SOUTHWEST, 1942—56 of water. The prices of water have risen because of in— creasing pumping lift, increasing need for artificial recharge of ground-water reservoirs, and increasing costs of storage and importation of surface water. The municipal and industrial users of water pay more per acre-foot of water than most farmers can afford at present prices for crops. Irrigated area, in thousands of acres, in California’s south coastal basin [Figures for 1900—50 rounded from U.S. Census; for 1954 rounded from Agricultural Department of Los Angeles Chamber of Commerce County 1900 1910 1920 1930 1940 1950 1954 Los Angeles .............................. 86 146 248 206 185 184 147 San Bemardino._ _______ ___. 38 70 105 108 111 107 102 Orange __________ ___. 42 55 87 112 119 125 101 San Diego__ ______ 16 25 25 43 52 56 63 Total _______________________________ 183 296 465 469 467 472 413 The increasing water cost has been borne especially by the newcomers to the State, who have been respons- ible for expanding urbanization and industrialization and who upon arrival found the local supplies of sur- face and ground water already developed and in use. To many of theSe immigrants, the only water available for their requirements was relatively expensive im- ported water. For many long-established users of the local water supplies the cost of water increased during the drought, as ground-water storage was depleted. Some of the increased cost resulted from purchase of imparted water, which was used for artificial recharge of the depleted ground-water reservoirs (Thomas and others, 1963d). ALLOCATION OF WATER By H.’ E. THOMAS Regulation of the development and use of water has been influenced by drought in someplaces in the South— west. Legal and administrative controls are effected through some form of allocation or apportionment of the water wherever there has been competition and con— troversy over it. Thus allocation is made necessary by water shortage, whether the shortage is created by aridity, by drought, or by such artificial factors as con- centrated draft or excessive demand. Because of drought, it may be necessary to apportion supplies which under normal conditions would be ample for all requirements. Also, drought provides an excellent test of the efficacy of the systems of allocation that have been devised in various areas. The following discus- sion therefore includes some consideration of the strengths and weaknesses of various devices of alloca— tion as pointed up by the drought, as well as mention of modifications in systems of allocation during the drought. The bases for allocation of water in the Southwest are contained in statutes, court decisions, administrative GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES regulations, and compacts and treaties—all which have originated because of the fact that some water users were in a favored position to monopolize specific sources of water. The resulting systems of allocation may be relatively simple or highly complex, but generally they define the rights and prerogatives of and also the limi- tations on those who are thus favorably situated. Owners of land contiguous to a spring, lake or stream, or overlying a ground-water reservoir are in a favorable geographic position to use the water from those sources, and to deny access to the water to others. Similarly, many water users are favored by topographic position in that they are upgradient from other users and there- fore can divert and use water that would otherwise flow naturally to those downstream. In many arid States where from the earliest days of settlement it was recog— nized that water rather than land was the limiting fac- tor in development, the first users of water have been given a favorable position by the doctrine that “first in time is first in right.” And in many places some uses of water are recognized as of greater economic value to society and are therefore favored above other uses. Thus systems of allocation may define the rights and limitations to use of water on the basis of (1) land- ownership involving both geographic and topographic positions, (2) priority of beneficial use,(3) designated preferences as to type of use. on the basis of public benefits derived, or (4) a combination of these. In the allocation of water for use, it is important to recognize that some of the water resources are replenish- able by precipitation and thus constitute a perennial supply; but others are stored in quantities which once removed cannot be replaced under present climatic con- ditions and are thus available for one-time extraction only, as are our resources of petroleum, coal, and the various metallic ores. Surface waters, including those collected in natural or artificial reservoirs, are generally replenishable, although because of climatic variations full replenishment may not occur every year. Many of our ground waters are similarly replenishable at rates that vary with the climate. Many ground-water reservoirs also contain such large volumes of water in storage that it is possible for a time to pump water far in excess of the rate of replenishment. The precipitation deficiency of drought reduces the rate of replenishment of water in surface or subsurface reservoirs and encourages an increase in the with— drawals from those reservoirs for use. Thus the overall effect of drought is generally a depletion of reservoir storage. The question, how far is it safe to draw down the storage and yet be insured of future replenish— ment—~the question of “safe yield”—has been asked in many areas of intensive ground-water development. B49 And statutes in several States have the objective of pre— venting ground—water draft from exceeding the safe yield of the reservoir. A period of drought is obviously not the time upon which to base calculations of the safe yield from a hy— drologic unit, whether of surface water or ground water or both, unless one is seeking a very conservative esti— mate. It is, however, an excellent time to test the cal- culations that had been made on the basis of past ex— perience. In Utah, Nevada, and New Mexico—three States which specify appropriation as the exclusive method of obtaining a water right—there are several ground-water reservoirs which have been declared to be fully appropriated and which therefore have been closed to further development. In Las Vegas Valley, Nev., there is evidence of progressive depletion of ground—water storage, but this causes no concern be- cause water is available to the valley—physically, economically, and legally—from Lake Mead on the Colorado River (Thomas and others, 19630). In the Roswell Basin of New Mexico (Thomas and others, 1962) and Cedar City Valley of Utah (Thomas and others, 1963a) there has been a progressive depletion in storage during the numerous drought years preceding 1957, but this is to be expected during drought, for the recharge is less than average and the pumping is greater than average to make up for the deficiency in precipi— tation. Even with 30 years of record for each basin we cannot be certain that the present development ex— ceeds the safe yield, because there is some possibility that rainfall in the future may be sufficient to increase recharge and reduce the demand upon wells to the point where the recent withdrawals from storage may be re- placed, at least in large part. In both Roswell Basin and Cedar City Valley, the annual pumpage has increased significantly during the period of State control. Thus, if the use of water had reached equivalence with the safe yield at the time the basins were closed to further development, the present pumpage necessarily exceeds that quantity. For the most part this contravention of the spirit of the law has been accomplished legally: in New Mexico by drilling wells outside the declared area but tapping the same ground—water reservoir; in Utah by increasing the yield of existing wells to the claimed maximum that had been used beneficially, thus upsetting the equilibrium condi- tions extant when controls were imposed. In most of the Southwest the areas first settled and now most densely populated are the fertile but arid lowlands. From the first days of settlement the econ- omy of these areas has been sustained by streams whose headwaters are in relatively uninhabited—if not unin- habitable—mountainous areas. Security in the econ- B50 omy thus requires limitations in the natural advantages which might otherwise accrue to anyone who settled up- stream from the valley developments. In most States the limitation has been achieved as follows: By repudi- ation of the traditional privileges of landownership and by declaration that water rights are based on priority in the beneficial use, irrespective of the location of use; by declaration that use for hydroelectric power (which is one of the principal uses of water in headwater areas) is subordinate to domestic and agricultural uses (which are chiefly in downstream areas); by development of headwater storage expressly for the benefit of the low- land users; or by some combination of these. Within each of the States in the Southwest, except Texas and California, water rights to streamflow are based entirely on priority of beneficial use, so that up- stream position gives no advantage, and deliveries to downstream users may be required even though natural conditions are such that most of the water is lost in transport. Because the jurisdiction of each State is limited by its boundaries, however, each State neces- sarily has a separate and independent system of rights based on appropriation. For apportionment of the waters of interstate or international streams, it is nec- essary to rely either upon interstate compacts and in- ternational treaties or upon adjudications by Federal courts. These tend to provide security in water sup- plies for the downstream areas by limiting the natural advantages enjoyed by the users in upstream areas. Some interstate compacts have been completed for apportionment of the water in relatively small streams, as for example the La Plata River Compact of 1922 and the Costilla Creek Compact of 1944 between Colorado and New Mexico. Water rights in both States are based on priority, and the compacts provide integrated distribution and operation, which would otherwise be hampered by the conflicting jurisdictions of the two States. The compacts thus provide for apportionment comparable to that which could be provided by either State alone for a stream entirely within its boundaries. For both the Rio Grande and the Colorado River there have been long histories of controversy between upstream and downstream users. The instruments that have been negotiated for the purpose of achieving an equitable distribution of the water include the follow- ing: The Rio Grande Convention of 1906 between the United States and Mexico; the Rio Grande, Colorado, and Tijuana Treaty Of 1944, also between the United States and Mexico; the Colorado River Compact of 1922, between California, Colorado, Nevada, New Mex- ico, Utah, Wyoming, and eventually Arizona; the Rio Grande Compact of 1938 between Colorado, New Mex- ico, and Texas; and the Pecos River Compact of 1948 DROUGHT IN THE SOUTHWEST, 1942—56 between New Mexico and Texas. A common feature of these instruments is that they do not admit that they establish any principles or precedents of general ap- plicability. Nevertheless a few generalizations may be in order. Several compacts and treaties include guarantees of certain minimum quantities of water to downstream users. These minimum quantities are generally far below the average flow available to the downstream area, but have been greater than the actual flow during some years of the recent drought. The Rio Grande Convention of 1906 called for delivery to Mexican water users near Juarez of 60,000 acre-feet of water, which is less than half the long-term average flow that has been available to them. Annual deliveries exceeded 60,000 acre-feet until 1951, when only 51,000 acre-feet was available; deliveries were less than half of the specified 60,000 acre-feet in the dry years 1954 and 1955. The treaty, however, has a proviso that in event of extra- ordinary drought the amount delivered to Mexican users shall be reduced in the same proportion as the water delivered to lands of the Rio Grande project in the United States (Thomas and others, 1962). The Rio Grande, Colorado, and Tijuana Treaty of 1944, in its provision concerning the international reach of the Rio Grande below Fort Quitman, Tex. (Thomas and others, 1962), apportions to the United States one—third of the flow of certain named streams that enter the Rio Grande from Mexico, but guarantees to the United States not less than 350,000 acre-feet (annual average in consecutive 5-year cycles) from those streams. The treaty specifies also that, in the event of extraordinary drought which prevents Mexico from making available this quantity of water, the de- ficiency is to be made up in the next 5—year cycle. Dur- ing the recent drought, the inflow to the Rio Grande from the named tributaries reached a minimum annual average of 647,000 acre-feet in the 5-year period 1951— 55, and the guarantee to the United States would thus have required more than half the flow of those tribu- taries. Both the Colorado River Compact of 1922 and the Rio Grande, Colorado, and Tijuana Treaty of 1944 were negotiated prior to and in anticipation of complete de- velopment of the water resources of the Rio Grande and the Colorado River. Thus, although protection of water rights already established by use was one purpose of the negotiations, a prime objective was to apportion the water still unappropriated and unused. Apportion- ments made by these compacts serve not only to guaran- tee rights to water for future development, but also to set upper limits on development in some areas and thus protect the development potentials of other areas. In GENERAL EFFECTS OF DROUGHT 0N WATER RESOURCES this way a compact may limit both the advantages of topographic position of the upstream user and of pri- ority of actual beneficial use. The Colorado River Compact (Thomas and others, 1963c) guarantees a practically constant annual quanti- ty of water to downstream areas regarless of variations in streamflow. It specifies that the outflow from the Upper Basin, as computed at the compact point of Lee Ferry, Ariz. (below the mouth of the Paria River), shall not be depleted below an average of 7.5 million acre-feet per year, plus an additional 1.0 million acre- feet granted to the Lower Basin, plus half the 1.5 mil- lion acre-feet which was subsequently allotted to Mexico by the Treaty of 1944. In the drought period 1943—56 the computed average flow at Lee Ferry was 11.1 million acre—feet per year, and in the 4 years 1953—56 it was only 6.6 million acre-feet per year. The developed require- ments for domestic and agricultural use below Lee Ferry have exceeded 7 million acre-feet per year since 1952; these were fulfilled throughout the drought, in part by holdover storage in Lake Mead which has a usable ca- pacity more than three times as large as the annual allotment of water to the Lower Basin and Mexico. But the compact as modified by the treaty with Mexico also apportions “in perpetuity to the Upper Basin the exclusive use” of 6.75 million acre-feet per year, of which an estimated average of about 2.5 million acre-feet is consumptively used in the existing pattern of develop- ment. A simple subtraction of the quantity that must flow past Lee Ferry (9.25 million acre-feet) from the measured flow during the period 1943-56 (11.15 million acre-feet) indicates that the unappropriated water available for use in the Upper Basin would have been less than 2 million acre-feet per year during that 14-year drought period, and therefore considerably less than is allotted by the compact for use in the Upper Basin. Thus the compact has offset the Upper Basin’s natural advantage of topographic position. Actual measurement of the flow of the Colorado River at Lee Ferry (above the mouth of the Paria River) began in 1921, just before the compact was negotiated, but a record of historic flow has been extended back to 1897 on the basis of available records at various stations in the basin. This compiled record, plus the measure- ments beginning in 1921, indicate that the average an- nual outflow from the Upper Basin in the 33-year period 1897—1929 was 15.3 million acre-feet. It has been esti— mated that the depletion by consumptive use in the Upper Basin increased from about 0.7 million acre-feet in 1897 to about 2.5 million in the 1920’s and that the average depletion in the 33-year period was about 1.7 million acre—feet. Thus the calculated virgin flow at B51 Lee Ferry in the period 1897 —1929 would have averaged about 17 million acre-feet. . In the 27 -year period 1930-56 the calculated average annual flow at Lee Ferry was 11.2 million acre-feet, a reduction of 4.1 million acre-feet from the average dur- ing the period 1897—1929. A small part of this redne- tion—less than a million acre-feet per year—is ac- counted for by increased consumptive use and stream depletion within the Upper Basin during the later dry period. The record now available thus indicates a wet period of more than 30 years duration and a dry period almost as long which included both the drought of the 1930’s and the subsequent drought in the Southwest. The average virgin flow during the wet period exceeded that of the dry period by more than 3 million acre-feet. Theoretically, with this long dry period a matter of record, we now have the basis for a more accurate de- termination of the average water yield of the Colorado River. The comparisons of streamflow with data avail- able from tree-ring studies (p. B34) indicate that runoff in the 50-year period 1904—53 corresponds to the average in the past 8 centuries. If this is true, the average nat- ural yield of the Upper Basin is slightly less than the 16 million acre-feet that has now been allocated by com- pact and treaty. But this average yield is not a safe yield, in the sense that it can be guaranteed to water users every year, unless means can be found for storing without loss the surpluses of wet years for use in dry years. Storage space already available, plus that au- thorized by the upper Colorado storage project, will be equivalent to more than five times the average natural flow of the river at Lee Ferry, and this may be sufficient for the accumulation of all the surpluses in a prolonged wet period. But if the wet and dry periods have a du- ration exceeding 25 years, as indicated by available data, the quantities available during a dry period would be reduced in comparison with those available in the wet period, because of the progressive evaporation of the water that must be held over for many years. Unless such losses can be prevented, the natural flow at Lee Ferry in the dry period 1930—56 (11.2 million acre-feet, plus calculated depletions of about 2.5 million acre-feet) is perhaps the best measure of the quantity available for use at all times, including the most adverSe conditions. In View of the compact guarantees to the Lower Basin, the 6.75 million acre-feet allotted to the Upper Basin may be available only in the wet cycles and may be sub- ject to a reduction of more than 2 million acre-feet in prolonged dry climatic cycles. More recent interstate compacts in the Southwest in- dicate greater awareness of the effect of major climatic fluctuations, for they do not guarantee specific quanti- ties of water to anyone but instead attempt to apportion B52 water on the basis of established hydrologic relations at the time of the compact, Thus the Rio Grande Compact of 1938 provides for the apportionment of water among the three major divisions of the Upper Rio Grande Basin: the San Luis Valley in Colorado, the Middle Valley in New Mexico, and the lands served by Elephant Butte Reservoir in New Mexico and Texas (and Mexico, as specified in the Treaty of 1906). The obligation of each division with respect to the next downstream divi- sion is specified in tabulations of relationships for vari- ous rates of streamflow which were developed from rec- ords covering a period of several years. Thus, although the compact makes no guarantee as to the quantity of water that shall be available to downstream users, it attempts to insure that the stream depletions in an up- stream division shall not exceed those of the period when the specified relationships were observed. During the 1943—56 drought both of the upstream divisions failed to deliver water in accordance with compact obligations. This failure is attributed, in part, to increased pumping from wells and, in part, to losses by evapotranspiration of an increasing proportion of the available Supply dur- ing the drought years (Thomas and others, 1962). The Pecos River Compact of 1948 attempts to re- strict upstream developments by specifying that the flow of the river shall not be depleted “by man’s activ- ities” below an amount which will give downstream users (in Texas) a quantity of water equivalent to that available under “the 1947 condition.” Thus, although the Pecos River Compact lacks schedules of inflow—out- flow relationships, it is similar to the Rio Grande Com- pact in its objective of protecting downstream users by limiting upstream depletions; and the history since 1947 indicates similar difficulty in meeting the terms of the compact. With the marked deficiencies in precipitation and resulting reduction in streamflow, it is difficult to assess the stream depletions in New Mexico in relation to the “1947 condition,” and once the change is com- puted there remains the more difficult problem of assess- ing the proportion of the depletion that is due to man’s activity. Studies prior to the compact negotiations in- dicated that the principal causes of any increase in stream depletion after 1947 would be increased con— sumption of water by saltcedar and the delayed effects of pumping from wells in the Roswell Basin; to these may be added the effects of pumping from wells drilled since 1947, notably in the Carlsbad area (Thomas and others, 1962). The effects of drought upon negotiated apportion- ment of water and the problems raised thereby are sum- marized as follows: Apportionment of fixed quantities based on average streamflow falters during drought un— less the storage facilities are adequate to stabilize the DROUGHT IN THE SOUTHWEST, 1942—56 natural fluctuations in runoff; and although this is rec- ognized in the Colorado River basin and answered by plans for storage totaling more than five times the aver- age flow of the river, it is doubtful that a constant yield corresponding to the water already apportioned can be realized throughout the long cyclic climatic fluctuations. Apportionment on the basis of observed inflow—outflow relationships as exemplified in the Rio Grande basin provides flexibility to match the climatic fluctuations, and if the apportionment is equitable each user shares in the “ups and downs” of water supply that cannot be overcome by regulation. Here it would be necessary to know the hydrology in exhaustive detail in order to dis- criminate the natural from the artificial effects upon in- flow and outflow. The effects of the recent drought also serve to point up the situation in some interstate areas Where no com- pacts have been negotiated for apportionment of the wa- ter and where in the absence of overall jurisdiction by either State the upstream user of water can enjoy the full advantage of his position. As an example, the headwaters of the Gila River have produced far less than normal streamflow during the drought, but part of the deficiency has probably been caused by increased pumping from wells (Thomas and others, 19630). If the Gila basin were entirely in New Mexico, this new ground-water development could have been stopped by declaring the area; if the entire basin were in Arizona, however, it would be difficult to prove that the wells are not pumping “percolating” water and, therefore, diffi- cult to deny water to a landowner. Should New Mex- ico, for the benefit of water users in Arizona, deny wa— ter to its own citizens when the water would not be denied to them under the laws of Arizona? It is note- worthy that the Pecos River Compact gives an affirma- tive answer to this question in a very similar situation between New Mexico and Texas: in order to deliver water to Texas in accordance with the Compact, it is necessary to restrict ground—water development and use in the Roswell basin and Carlsbad area (Thomas and others, 1962) ; no such restriction would be possible if the Roswell and Carlsbad areas were in Texas. REFERENCES CITED Begemann, Friedrich, and Libby, W. F., 1957, Continental water balance, ground water inventory and storage times, surface ocean mixing rates, worldwide water circulation patterns from cosmic-ray and bomb tritium: Geochim. et Cosmochim. Acta, v. 12. p. 277—296. Bonnen, C. A., and Ward, J. M., 1956, Some economic effects of drouth on ranch resources: ’Dexas Agr. Expt. Sta. Bull. 801, 11 p. Buswell, A. M., and Rodebush, W. H., 1956, Water: Sci. Ameri- can, v. 194, no. 4, p. 77—89. GENERAL EFFECTS OF DROUGHT ON WATER RESOURCES Jacob, C. E., 1945, Correlation of ground-water levels and pre- cipitation on Long Island, N. Y., pt. 2, Correlation of data: Am. Geophys. Union Trans, v. 25, pt. 6, p. 929—939. Lane, E. W., 1950, Streamflow variability: Am. Soc. Civil Engi- neers Trans, v. 115, p. 1132—1133. Lane, E. W., and Lei, Kai, 1950, Streamflow variability: Am. Soc. Civil Engineers Trans, v. 115, p. 1084—1098. Petitt, B. M., Jr., and George, W. 0., 1956, Ground-water re- sources of the San Antonio area, Texas: Texas Board Water Engineers Bull. 5608, v. 1, 78 p.; v. 2, pt. 1, 252 p.; v. 2, pt. 2, 285 p.; v. 2, pt. 3, 126 p., 38 p., 26 p., 11 p., and 24 p. Pickett, Arthur, 1956, Conference on industrial uses of water, Los Angeles. Piper, A. M., 1948, Runofi from rain and snow: Am. Geophys. Union Trans, v. 29, p. 511—519. Robinson, T. W., and Lang, W. B., 1938, Geology and ground- water conditions of the Pecos River valley in the vicinity of Laguana de la Sal, New Mexico: New Mexico State Engineer 12th and 13th Bienn. Rept., p. 79—100. Rocky Mountain Forest and Range Experiment Station, 1956: US Dept. Agriculture, Forest Service. Schulman, Edmund, 1945, Tree-ring hydrology of the Colorado River basin: Arizona Univ. Tree-Ring Bull., v. 16, no. 4, p. 22—38. , 1950, Dendroclimatic histories in the Bryce Canyon area, Utah: Arizona Univ., Tree-Ring Bull., v. 17, no. 1/2, 14 p. B53 Schulman, Edmund, 1956, Dendroclimatic changes in semiarid America: Tucson, Univ. Arizona Press, 138 p. Thomas, H. E., 1951, The conservation of ground water: New York, McGraw-Hill, 321 p. , 1962, Meteorologic phenomenon of drought: U. S. Geol. Survey Prof. Paper 372—A, 43 p. (In press). Thomas, H. E., Gould, H. R., and Langbein, W. B., 1960, Life of the reservoir, in Sedimentation in Lake Mead: U.S. Geol. Survey Prof. Paper 295, p. 231—248. Thomas, H. E., and others, 1962, Effects of drought in the Rio Grande basin: U.S. Geol. Survey Prof. Paper 372—1). (In press). Thomas, H. E., and others, 1963a, Effects of drought in basins of interior drainage: U.S. Geol. Survey Prof. Paper 372—E. (In press.) , 1963b, Efiects of drought in central and south Texas: U.S. Geol. Survey Prof. Paper 372—0. (In press.) ———, 19630, Effects of drought in the Colorado River basin: U.S. Geol. Survey Prof. Paper 372—F. (In press.) ————, 1963d, Effects of drought along Pacific Coast in southern California: U.S. Geol. Survey Prof. Paper 372—G. (In press.) Troxell, H. C., and others, 1954, Hydrology of the San Bernar- dino and eastern San Gabriel Mountains: U.S. Geol. Sur— vey Hydrol. Inv. Atlas HA—l, 13 pl. Wing, S. P., 1950, Streamflow variability: Am. Soc. Civil En- gineers Trans, v. 115, p. 1122. Page Abstract ...................................... B1 Alfalfa ....................................... 6 Allocation of Water ___________________________ 48 Animas River at Durango, Colo ______________ 34 Aquicludes ___________________________________ 4 Aquifers ...................................... 4 Aquitards _____________________________ 4 Arizona ...................................... 46 Arizona Cooperative Wildlife Research Unit of the University of Arizona, quoted. 46 Artesian water _______________________________ 38 Artificial recharge of ground-water reservoirs. 48 Azua Canal, Calif ____________________________ 7 Bartlett Reservoir, Ariz ______________________ 7 Base flow _____________________________________ 38 Big Bear Lake, Calif _________________________ 47 Black grama grass ..................... . _ 43 Blue Creek, N. Mex __________________________ 13 Bonnen, C. A., and Ward, J, M., quoted"-.. 3 Brazos River at Waco, Tex ___________________ 13 California .................................... 3, 47 Climatic fluctuations, eflect on stream flow.. 7 Cloud seeding ________________________________ 45 Cogswell Reservoir, Calif _____________________ 7 Colorado _____________________________________ 45 Colorado River aqueduct of the Metropolitan Water District ___________________ 48 Colorado River at Lees Ferry, Ariz........_._ 31,51 Colorado River Compact of 1922.... _ _ _ Conchas Reservoir, N. Mex __________________ 45 Conejos River at Mogote, Colo _______________ 34 Conservation practices _______________________ 43 Costilla Creek Compact of 1944 ______________ 50 Costilla Reservoir, N. Mex ___________________ 45 Cotton production ____________________________ 43 Crop failure ................................ 42, 43, 46 Cuyamaca Reservoir, Cali! ___________________ 47 Dallas, Tex ___________________________________ 43 Dissolved solids in stream water. 38 Diversions 13 Drainage ____________ 7 Drought-disaster areas, _______________________ 42 Eagle Nest Reservoir, N. Mex........ 45 Economic and related effects of drought. . 41 El Capitan Reservoir, Calif ___________ 47 El Vado Reservoir, N. Mex _____ _ 45 Elephant Butte Reservoir, N. Mex.- . 45,52 Elsinore Lake, Calif .......................... 47 Fish life ...................................... 45 Flagstaff, Ariz ________________________________ 46 Floods-.-._-.._-_ . 4 Friant-Kern Canal __________________________ 48 INDEX r Page Garza-Little Elm Reservoir, Tex _____________ B43 Gila River near Red Rock, N. Mex .......... 7 Grapevine Reservoir, Tex ____________________ 43 Great Plains __________________________________ 31 Ground water, discharge _____________________ 5 general discussion ________________________ 4 pumping ................................. 47, 48 quality of ................................. 41 recharge __________________________________ 4 storage ___________________________________ 6 Halite, beds 01 ________________________________ 38 High Plateaus, Utah- . 28 Horseshoe Dam, Ariz _________________________ 47 Horseshoe Reservoir, Ariz .................... 7 Hydroelectric power ....................... 42, 46, 47 Hydrologic cycle ............................. 1 Introduction _________________________________ 1 Irrigation _ . .. . _____________________________ 45, 48 Karnes County, Tex ......................... 42 Lake Dallas, Tex _____________________________ 43 Lake Mead ............................... 38, 49, 51 Lane, E. W., quoted ......................... 17 La Plata River Compact of 1922 _______ . 50 Las Vegas, Nev .............................. 46 Livestock _________________________________ 42, 43, 46 Long Island, N.Y. _ 2 Los Angeles, Calif ____________________________ 48 Malaga Bend, N. Mex ....................... 38 McClure Reservoir, N. Mex _________________ 45 Meteorological zones _________________________ 27 Metolius River, 0reg.. 2 Mogollon Rim, Ariz... ............. 28 Mormon Lake, Ariz.-. _____________ 47 Morris Reservoir, Calif ....................... 7 Navajo Reservation _________________________ 46 Nevada ______________________________________ 46 New Mexico _________________________________ 43 Pecos River Compact of 1948.. 52 Pecos River, N. Mex .......... 38 Phantom Hill Reservoir, Tex. 13 Phreatophytes ........ 6 Pickett, Arthur. 47 Population ........................ 42 Possum Kingdom Reservoir, Tex. _ . _ 13 Precipitation .................... 3, 4, 47 Pumping ..................................... 41 Quality of wateri ............................ 36 Rangeland ________________________________ 42, 43, 46 Regional runoff ............................... 15, 27 O Page Rio Grande Compact of 1938 ................. B50 Rio Grande Convention of 1906 ............... 50 Rio Grande at Otowi Bridge near San Ilde- fonso, N. Mex .................... 34 near Del Norte, Colo ..................... 13 Rio Grande, 0010., and Tijuana Treaty of 1944 .............................. 50 Rocky Mountain Forest and Range Experi- ment Station, quoted._ ....... 43 Runoff ...................................... 7, 8, 19 See also Regional runofl. Salt River, Ariz .............................. 38 Salt Water .................................... 40 San Diego aqueduct .......................... 48 San Gabriel Reservoir, Calif-.__.__ 7 San Gabriel River near Azusa, Calif .......... 7 San Juan Mountains ......................... 28 Santa Cruz Reservoir, N. Mex.- 45 Santa Fe, N. Mex ................. ‘. .......... 45 Schulman, Edmund, quoted ................. 31, 32 Sediment load ............... 37 Sewage-treatment plants ..................... 43 Soil water .................................... 2 Soil zone .......... 2 Sonoran border zone .......................... 27, 28 Springs, ephemeral ___________________________ 5 perennial ...... 5 Storage, deletion _______________________ 49 See also Ground Water. Streamflow, as indicated by tree rings ........ 31 expressed in units of stande deviation. . 17 natural ................................... 7 variability.....-... 17 Surface water, quality of ............... 37 Texas ...................................... 3, 41, 42 Thomas, H. E quoted. _ 1 Treaties.- 50 Tree-ring studies ............................. 31, 51 Trinity River at Liberty, Tex. 40 Tritium analyses ............................. 2 Truth or Consequences, Tex .................. 45 Utah _________________________________________ 46 Verde River below Bartlett Dam, Ariz ....... 7 Virden, N. Mex .............................. 13 Virgin River at Virgin, Utah ................. 7 Water, defined ............................... 36 origin ..... 1 Water rights.-...-.....-..... 48 Water supply during drought. ..... 42 situations ...................... 41 Wheat, production of. 42 Whitney Reservoir, Tex ..... 13 Wildlife ___________________________________ 42, 45, 46 Yuma, Ariz .................................. 47 B55 UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY COLORADO _-_____________r- NEW MEXICO L._..-_.__._-____.._.‘___. _ PROFESSIONAL PAPER 372—3 PLATE Lr ‘ L._.._.._.._..-_ \ ff?“ /W ' _' ““""T ’/ I////// l948 \ COLORADO / .f““" \\\ a V "’/// W ////// , W MHEKMHOMA // // ‘_ ogLAHOMA \ 3. I. . I ,_ KANSAS 61EITAH0MA A .5 i ! _,_._..._.._ Amman" runoffpmhoblylmthmmednn‘ THAN THE MEDIAN, 1942—157 EXPLANATION Runoff, in percent of 1904—53 median Amwflh' runoffpmbublyl-thn qmtofmed'nn 690212 0 — 63 (In pocket) 6:“ , , y 5;, 3”}. w- flEffects of Drought in Central and South Texas GEOLOGICAL SURVEY PROFESSIONAL PAPER 372-C Thomas and others—EFFECTS 0F DROUGHT IN CENTRAL AND SOUTH TEXAS Geological Survey Professional Paper 372-0 Effects of Drought in Central and South Texas By H. E. THOMAS and others DROUGHT IN THE SOUTHWEST, 1942-56 GEOLOGICAL SURVEY PROFESSIONAEL PAPER 372—C UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1963 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 Oflice Washington, DC. 20402 CONTENTS Page Pan Abstract ___________________________________________ Cl Gulf coastal plain ___________________________________ 018 Introduction --------------------------------------- 1 Precipitation-runoff relations _______________________ 18 Brazos River basin ---------------------------------- 3 Efl’ects of ground-water development ________________ 20 Drought and conservation ......................... 4 Efl’ects of drought ________________________________ 21 Drought and the-“salt 0f the earth” """"""""" 4 Llano Estacado, Texas and New Mexico _______________ 22 Colorado River basm ________________________________ 5 , , Upper Colorado River _____________________________ 5 Clunate asia controlling factor ______________________ 23 Concho River ____________________________________ 6 Control in geologic history _______________________ 23 . . Control in present economy ______________________ 24 Middle Colorado River ____________________________ . 8 _ , , Reservoir in Edwards limestone near Austin __________ 8 Inadequacy Of preclpitation ““““““““““““ 24 Lower Colorado River _____________________________ 11 Supplementary sources of water ________________ 24 Ground-water reservoir in the Edwards limestone, by Development Of ground water ---------------------- 25 B. M. Petitt, Jr _______________________________ 11 The ground-water reservoir ______________________ 25 Hydrology _______________________________________ 12 History of development and use __________________ 27 Hydrologic equation for the fault zone _______________ 14 Effects Of development and use ————————————————————— 27 Effects of water development and use _______________ 14 Efl’ects 0f drought -------------------------------- 30 Effects of drought ________________________________ 16 References cited ____________________________________ 31 ILLUSTRATIONS {Plates are in pocket] PLATE 1. Fluctuations in quantity and quality of water, Possum Kingdom Reservoir. 2. Hydrographs for the Austin area, 1933—57. 3. Fluctuations in runofl’ from the Edwards Plateau, 1933—57. 4. Hydrographs of wells in the Balcones fault-zone aquifer, 1931—58. P age FIGURE 1. Index map of central and south Texas _________________________________________________________________ C2 2. Annual runoff of Colorado River, 1920—58 _____________________________________________________________ 7 3. Hydrographs for the San Angelo area, 1933—57 _________________________________________________________ 9 4. Fluctuations in spring flow, ground-water level, and precipitation in the San Antonio area, 1932—57 ___________ 13 5. Recharge to and discharge from the Balcones fault-zone aquifer, 1934—57 __________________________________ 15 6. Correlation of water level in well 26, Bexar County, with changes in reservoir storage _______________________ 16 7. Annual discharge of water from the Balcones fault-zone aquifer ___________________________________________ 17 8. Fluctuations of precipitation and runoff in the Gulf Coastal Plain, 1937—58 ________________________________ 19 9. Four dimensions of ground-water development at Kingsville, Texas _______________________________________ 22 10. Frequency of precipitation at Muleshoe, Tex., 1922—52 __________________________________________________ 25 11. Section southeastward across High Plains ______________________________________________________________ 26 12. Development and use of ground water in the Llano Estacado, 1934—57"-.. ________________________________ 28 13. Hydrographs of selected wells in the Llano Estacado ____________________________________________________ 29 III DROUGHT IN THE SOUTHWEST, 1942—56 EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS By H. E. THOMAS and others ABSTRACT The effects of drought upon ground-water storage and dis- charge, and upon streamflow, vary tremendously in the central third of Texas (the area from the Panhandle to the Gulf of Mexico). Extremes are represented by (a) the Llano Estacado, Where the drought had negligible effect upon groundwater re— sources, which are being progressively depleted by pumping for irrigation; and (b) the Balcones fault-zone aquifer west of San Antonio, whose storage and natural discharge declined sub- stantially during the drought, but increased even more rapidly during succeeding years of more abundant precipitation. ‘ INTRODUCTION This chapter, the third in the report on drought in the Southwest, describes the effects of drought upon the ground-water storage and discharge and upon stream- flow in the central third of Texas—the area from the Panhandle to the Gulf of Mexico (fig. 1). Subsequent chapters describe effects of the drought in other areas in the Southwest. The effects of drought are discrimi- nated from water shortages due to other causes wher- ever possible. The patterns of land and water resources in central and south Texas have many features in common with those of other Plains States (for example, Oklahoma and Kansas) and different from those of the States to the west (New Mexico, Colorado, Arizona, Utah, Ne- vada, and California). In all the States mentioned the population is concentrated at the lower altitudes, and the higher lands have relatively low population density. The mountains and plateaus of the six States west of Texas are very sparsely inhabited and have remained for the most part in public ownership. By contrast, the higher plains of Texas are of agricultural importance comparable with those at lower altitudes. Superposed upon this contrasting pattern of land utilization is a comparable contrast in the water- , resource pattern. In the States west of Texas the scenic but relatively uninhabited mountains and pla- teaus receive the most precipitation and are the chief water—producing areas. The people in those States may differ in their ideas as to the most effective way of obtaining the greatest yield from these producing areas, but they are in practically unanimous agreement that this yield should be used in the water—deficient lowlands, where they live. By contrast, Texas’ pattern of precipitation is such that the lower landsto the southeast produce the greatest runoif per square mile, and the streams draining the higher lands to the north and west are more likely to be small and ephemeral. Needless to say, the inhabitants and landowners of these semiarid uplands would like to hold back this small runofl' for use as a supplement or alternative to inade- quate rainfall; and their interests thus come in conflict with those of people downstream, who also could utilize surplus runoff especially in time of drought. Thus there may be even greater controversy over specific pro- grams of water conservation and development in Texas than one would expect to find in the more arid regions farther west. _ Texas presents opportunities for hydrologic subdi- vision on the basis of precipitation, streamflow, or ground-water reservoirs. Precipitation is least in the west and progressively greater toward the east; it ranges from an average of less than 10 inches a year near El Paso to more than 50 inches annually in the southeast corner along the Gulf of Mexico. By con- trast with other drought—ridden States of the South- west, where precipitation is ordinarily greatest in the highlands and least in the lowlands, Texas receives its greatest rainfall on lands in the southeastern part of the State near sea level, and progressively less upon the higher plains to the west. The lines of equal rainfall (isohyets) run generally north and south across Texas, and the boundary between arid and humid ‘regions is also approximately a north-south line, through Fort Worth, Waco, and the mouth of the Guadalupe River. This report is concerned with drought in the region west of that line. Cl 02 7 i i i I I l I I l DROUGHT ’ IN THE SOUTHWEST, 1 942—5 6 .PORTALES' MULESHOE- \" 4...‘ ‘ \ -~ . I PLAIIIVIEw J‘- ' . “JN " I , “1"V~‘VI “W“ I : LLANIb LUBOBOCK W“ O o o . I ESTACADO s». r I, * \. l. I :I’UBLé‘ M r e 0 , \\ (0 \ 'o- HOBBS i .. i 6254,? Pfifiggw . \ 'l RESERVOIR ‘9» a I. ‘~ I s o .‘ II ' \\\\\ \ \ r-._--—--________J I new cb< . .‘ \‘s MIDLAND Q9400 ' \ \ “as“ BALLINGER ‘s ‘\ .‘ 5049- ' ' ‘4‘. |\ \ ° , SAN ANG -. ‘1 i a 9 ) “\ \ EDWARDS sau SABA' “‘13-“ %\ ° . ~ '\ PLATEAU aucumm " ' ‘\ RIvER RES. ( “ ‘ LLANO «, LAKE warm . I‘ \ 7’ AUSTIN . \ / .’\‘° 7 ) . ’ 4/ I '\ . .g %5MIfH- I??? K I __/~..__ _ . EV w N VILLE . ‘ ' ‘ k, 'h “Q o \‘\ I, \ . . Kai I, Igmif'lusfiml IIII i “‘3 SAN ANTONIO 906‘ I’ ; \~ . ‘\ ‘ GULF ‘ 544’ \n / \x '. u L-DE COASTAL 044,, a P \ PLAIN 4/ I 4/ ”05055 ’8 554/ x“ a ‘ / \ :\ . I \ " Cu PUS CHRISTI \ Kmesvaua l: 1 \‘- \_.. 100 0 100 MILES \ l l I I I I l , M \ \u-u - - \ - .4. FIGURE 1,—Index map of central and south Texas. The northern boundary of Texas, except for the Pan- handle, is formed by the Red River, which flows to the Mississippi. The State’s southern boundary 18 formed by the Rio Grande. Between these two rivers several other rivers rise within the State, flow southeastward, and empty into the Gulf of Mexico. About 60 percent of the area of Texas is drained by these intrastate streams. The principal intrastate rivers that drain the drought area in central and south Texas are the Brazos, Colorado, Guadalupe, and Nueces Rivers. The ground-water reservoirs of central and south Texas have still another hydrologic pattern. Several of them are in sandy materials that crop out in bands paralleling the Gulf coast, and each of these is thus crossed by several streams in their courses to the Gulf. A major ground-water reservoir occurs in limestones underlying the Edwards Plateau and the adjacent Bal- cones fault zone. Numerous tributaries of the Rio Grande, Nueces, Guadalupe, and Colorado Rivers rise in the Edwards Plateau, and depend upon ground wa- ter from this huge reservoir for part of their flow. The drainage divides in many places do not coincide with groundwater divides. The following discussion of central and south Texas EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS is organized by river basins insofar as possible. The Colorado River basin is considered a major unit because of the interdependence of developments up and down the stream, although the river traverses the eastern part of the ground—water reservoir in the Edwards lime- stone. Farther west, however, the limestone reservoir is considered as a unit, because it influences the flow of several major streams to the south. The streams that drain the Coastal Plain south of the Balcones escarp- ment—the Guadalupe, Nueces, and others between the drainage basins of the Colorado and the Rio Grande— are then considered as a group because of their com« munity of interests with the limestone reservoir and the ground—water reservoirs underlying the Coastal Plain. Only the effect of drought upon quantities of water is considered in the discussion of most of the individual river basins. Data concerning the effects of drought upon water quality are meager, and may be summarized for all Texas as follows: Many reports of changes in quality of ground water have been reported during the drought years, but available data do not indicate the causes of the changes. Increasing salinity has been reported in individual wells widely distributed throughout the western half of Texas; in some in— stances the cause may be related to drought or to in- creasing withdrawal of water, in others to corrosién of the casing or to other defects in the well, and in still others to surface disposal of brine from oil or gas wells. The changes in quality of many surface-water supplies appear to be related to the drought. According to Burdge Irelan (written communication), of the US. Geological Survey oflice in Austin: Drought has undoubtedly resulted in a general increase of dissolved solids, chlorides and sulfates, in water stored in and released from Denison Reservoir on the Red River and Possum Kingdom Reservoir on the Brazos River. Weighted average concentrations of water released from these reservoirs has trended upward from the beginning of sampling (1942 at Pos- sum Kingdom, 1944 at Denison) until this year’s [1957] ex- treme flood. Water sampled in May 1957 was much lower in mineral content at each reservoir than at any time since the first fillings. However, the evidence suggests that evaporation may not have been particularly significant in these changes. Salt springs and seeps located many miles upstream from the two reservoirs apparently contribute a more substantial por- tion of the salts in the water released from the reservoirs dur- ing dry years than during wet years. There is also evidence that a high rainfall on a salt-yielding tributary watershed sometimes results in more than normal loading of the stream waters. In contrast with the Brazos and Red Rivers, the Colorado River reservoir releases do not show increases in chlorides, sulfates, or dissolved solids. This is apparently be- cause the Colorado River has a much smaller portion of its drainage area in the region of Permian outcrops, and also be- cause only a small portion of its flow originates there. C3 BRAZOS RIVER BASIN The Brazos River basin is along the margin between the arid West and humid East. On the basis of average annual climate the upper part of the basin above Waco receives less precipitation than the potential evapotran- spiration, but there is some surplus of water in the lower part of the basin. The entire basin is marginal also in the sense that the boundary between semiarid and sub- humid areas shifts back and forth across it, according to the fluctuations in precipitation from year to year. In wet years the entire basin may qualify as humid, but in the recent dry years the efl'ects of drought have extended over the lower basin and an extensive area farther east. Only the upper part of the Brazos basin is considered here, because this report is limited to the region that is normally arid or semiarid, although it is recognized that the efl'ects of drought have extended farther eastward. In fact, much of the discussion of the upper Brazos basin is applicable also to the lower basin and to other drainage basins farther east in Texas. The drainage basin credited to the Brazos River above Possum Kingdom Reservoir is about 22,500 square miles, but more than 40 percent of this total is “noncontributingz” that is, it is in the High Plains (p. 022), where the tributaries to the Brazos drain only narrow strips along their channels and the rest of the area contributes no runoff. The contributing area east of the High Plains is in the Permian basin, where sandstone, limestone, and shale beds are interbedded with gypsum and other salines, and where, therefore, much of the ground water is not potable, and a good deal of the surface runoff is of poor quality. The contributing area is also within the North Central Plains soils belt, which is characterized by fairly per- meable sandy to clayey loams. Statistics concerning rainfall and runoff in recent years have been cause for alarm among the residents of the upper Brazos River basin. The average precipi- tation in the 10 years 1944—53 was about 90 percent of the 1924—53 average, but the runofl at Possum King— dom Reservoir was only about 50 percent of the 30- year average. The flow of streams characteristically fluctuates more widely than does the precipitation upon their drainage areas. When runoff of the upper Brazos is plotted against rainfall over the basin, it is readily apparent that an annual rainfall of more than 12 inches is required to produce substantial runoif. But comparisons between individual years or selected periods with equivalent rainfall are disquieting: In 1949, for instance, the basin precipitation was slightly greater than in 1935, but the streamflow was, only half as great. In the drainage basins of the Salt and Double 04 Mountain Forks, the average precipitation in the dec- ade 1941—50 was slightly greater than in the decade 1932—41, but the runoff was 17 percent less. DROUGHT AN'D CONSERVATION Several hydrologic studies have been made of the Brazos River basin, some of which include also the adjacent Colorado River basin, in order to analyze and explain the observed reduction in runoff. The reports of these studies are in agreement that the rainfall- runoff relation in recent years has been significantly different from that shown by the records for earlier years; this is clearly shown by double-mass analysis of the records of rainfall and runoff. But the authors of those reports are not in agreement as to the explana- tion for these changes. Freese (1954) cites a report by Dickson and others (1939) that runoff from individ- ual tracts of land may be reduced 30 to 100 percent during years of normal rainfall by soil—conservation measures, and quotes Stallings (1945) as to the measured effects upon runoff of terracing, contour cul- tivation, and strip cropping. He then compares rain- fall and runoff in the upper Brazos and Colorado River basins for the 10 years ending in 1950 with the rainfall and runoff during 10-year periods of like average rain- fall prior to 1943, and findsra decrease in annual runoff approximating 0.3 inch of water over the drainage basins. This he attributes to soil—conservation meas- ures, which began in west Texas in about 1943. Others, after studying the Brazos and Colorado Rivers by sub- basins in considerable detail, similarly have concluded that stock ponds and soil—conservation measures have caused a significant reduction in runoff during the pe- riod 1942—53. , In a study for the Brazos River Authority, Ca— rothers and Newnam (1956) agree that conservation practices, and particularly stock ponds, are factors in diminution of runoff, but they place them at a lower level of signifiCance. They estimate that, in the drain— age basin contributing to Possum Kingdom Reservoir, there are 18,000 stock ponds with average drainage area of 25 acres, surface area of 0.5 acre, and capacity of 3 acre—feet; and they estimate an average water loss of 13,500 acre—feet a year from these ponds. Diversions from the streams are estimated to have reduced run— off to Possum Kingdom by an average of 35,000 acre— feet a year. The total loss by diversions and stock ponds would thus be about 6 percent of the average annual runoff. As to other soil-conservation practices (contour plowing, terracing, strip cropping) Carothers and Newnam recognize that these would reduce the runoff, but consider that the reductions are offset by increased runoff resulting from clearing brush and DROUGHT IN THE SOUTHWEST, 1942~56 scrub timber, which consume more water than the grass that replaces them. Carothers and Newnam attribute most of the change in rainfall-runoff relation since 1942 to a change in rain- fall characteristics. Taking first the sample years 1936 and 1944, they note that the rainfall in 1944 was slightly greater, but the runoff was 75 percent less, than that in 1936. The rainfall in 1944 was evenly distributed in relatively small storms, with only two “storm events” producing more than 2 inches of rainfall, and these only slightly more. There were 6 storms of 2 inches or more in 1936, of which one in September was a “gully washer” of more than 5 inches; runoff after this storm reached a peak more than 4 times the maximum in 1944. Comparing then the drought period 1944—53 with the two preceding decades, Carothers and Newnam note a decided reduction in intensity of the rainfall events in the last decade, particularly in storms of 2 inches or more. Their conclusion, that this lower in- tensity of rainfall has been an important factor in the unseemly reduction in runoff, is in accord with findings ‘ in a recent study of the hydrology of the upper Chey— enne River basin of eastern Wyoming (Culler, 1961, p. 86—103). The correlation of runoff records in the central Texas region (Gatewood and others, 1963) does not offer any clue as to the comparative degree to which the natural rainfall pattern and the artificial effects of soil-conser- vation practices respectively have influenced runoff. There is good correlation among the records for Brazos River at Seymour, Colorado River at Ballinger, and Middle Concho River near Tankersly, but the drainage basins above all three stations have been the scenes of progressively increasing soil-conservation activities and stock-pond construction. DROUGHT AND THE “SALT OF THE EARTH” Possum Kingdom Reservoir, with total capacity of about 725,000 acre-feet, began storing water in March 1941; in the first 4 months of operation it received enough water to fill it three times over, and during 1942 it again received water equivalent to more than twice its capacity. Drought began in 1943, however, and in every one of the years 1943—56 the inflow to the reservoir was less than 750,000 acre-feet, and far below the 1935—42 annual average of 1,200,000 acre-feet. The uppermost graph in plate 1A shows the average annual discharge of Brazos River near Palo Pinto, Tex., which is approximately the outflow from Possum Kingdom. A second graph shows that the dissolved load has fluc- tuated with the runoff, greater in years of high runoff and less in years of low runoff, as is to be expected. Plate 13 shows a general increase in the concentration EFFECTS OF DROUGHT IN of dissolved mineral matter in the outflowing water dur- ing the local 1943—56 drought, although the upward trend was reversed in 1946, 1950, and 1953; the increases were chiefly in sodium, chloride, and sulfate. The abundant inflow during 1957 caused a sharp drop in the concentration of dissolved solids in the reservoir water. According to the US. Public Health Service recom— mended standards for drinking water (250 ppm chlo- ride, 250 ppm sulfate, 1,000 ppm—permissible, 500 pref- erable—total dissolved solids), the quality of water in Possum Kingdom Reservoir was marginal when it was first filled in 1941. The increase in salinity during the years of drought was therefore a matter of increasing concern to the Brazos River Authority, which distrib- utes some of the water for municipal and industrial use. The problem became most acute in the years 1954—56, when the quantity of runofl‘ was considerably greater than the average for the preceding 10 years but the qual- ity of the water deteriorated during the 3-year period. Was this a result of increasing ground-water contribu- tion from saline areas, reflecting increased recharge from precipitation? Or had the rains redissolved the soluble salts that had been left in the soils, in ponds or dry channels, during the worst years of drought? An- SWers to these and similar questions are being sought in current (1958) studies, and may be the basis for action to reduce the salt contribution to Possum Kingdom Reservoir during drought. One clue to the source of dissolved saltswthough not the detailed anSWer needed for an action program~is suggested by plate 10. which shows the inflow to the reservoir as represented by the record for Brazos River near South Bend, TeX. At this point the river has a contributing drainage area slightly greater than 12,000 square miles, of which about 46 percent is drained by Clear Fork (above Crystal Falls), 16 percent by Salt Fork (above Aspermont), and 12 percent by Double Mountain Fork (above Aspermont). The proportions of total reservoir inflow that come respectively from Salt and Double Mountain Forks, based on gaging records, are shown in the graph. These two tributar- ies, with 28 percent of the drainage area above the gag— ing station near South Bend, contributed 25 to 35 per- cent of the water in the years 1942—45, and also in 1950— 51. In most other years of the drought, Salt and Double Mountain Forks contributed a higher propor- tion of the water: 38 to 47 percent in 1946—49, and 50 to 62 percent in 1954—56. The upward trend in proportion of water coming from Salt Fork and Double Mountain Fork is sig- nificant in the study of quality of the reservoir water, because Salt Fork carries sodium and chloride and Double Mountain Fork carries calcium sulfate in far 690~1 76—46 3—~—2 CENTRAL AND SOUTH TEXAS C5 greater concentrations than are found in Clear Fork. This in turn reflects the differences in geology of the three drainage basins: gypsum and halite both common in Salt Fork basin, gypsum moderately so in Double Mountain Fork basin, and both rare in Clear Fork basin (pl. 1D). The fluctuations in mineral content of the reservoir water thus may be in response, at least in part, to geographic variations in precipitation—or drought-throughout the tributary drainage basin. The reservoir water becomes more mineralized when rainfall and runoff are greater in the Salt Fork basin and less in Clear Fork basin, as in 1952 and 1955; and it becomes less mineralized when Clear Fork has the bulk of the rain and the runoff, as in 1950 and 1954. COLORADO RIVER BASIN The Colorado River, like the Brazos, has tributaries that rise in the High Plains (although the High Plains contributes practically no runoff) , and flows southeast- Ward to the Gulf. According to long—term averages, the annual rainfall at the edge of the High Plains is less than 20 inches, and the rate increases progressively to more than 40 inches at the Gulf. The average (1924— 53) runoff accounts for less than an inch of the precipi— tation in the upper part of the basin—above Ballinger— and for more than 7 inches of the rainfall near the mouth, downstream from Columbus; thus the river originates in semiaridity and traverses bands of pro- gressively increasing humidity. The drainage basin can be subdivided into bands that are distinctive not only as to rainfall and runoff production, but also as to geology and soils, for the geologic “grain” of the region is generally parallel to the Gulf Coast and transverse to the axis of the river basin. UPPER COLORADO RIVER The drainage basin of the Colorado River above the gaging station at Ballinger is similar in many respects to the upper Brazos River basin. The similarity ap- plies to rainfall, runoff, geologic formations, soils, and ground-water conditions; and there has been similar controversy in both areas as to the degree of effect of stock ponds and soil-conservation measures upon the runoff since 1942. The drainage area above Ballinger is about 16,800 square miles, but 11,600 of the total is in the High Plains and probably noncontributing. The lowermost graph of figure 2 shows the annual runoff from the upper Colorado River basin as meas- ured at Ballinger. In the wet years 1941 and 1957 the runoff was about double the 49-year (19084—56) mean of 266,000 acre-feet. In four of the intervening years, 1942, 1947, 1948, and 1954, the annual total reached or slightly exceeded that mean, and in 1945 and 1949 06 the discharge was equivalent to the median (about 226,000 acre—feet) for the 49-year period. The other 9 years were definitely droughty, for the annual runoff ranged from 10 to 60 percent of the median. ‘Six of these dry years occurred in the period 1950—56, which included the 2 years of least runoff (1952, 1956) in the period of record. For purposes of regional correla- tions of runoff (Gatewood and others, 1963), the record at Ballinger has been accepted as representing natural streamflow. Thus no corrections have been attempted for the effects of soil conservation measures in recent years, concerning which there is no general agreement even among the experts (p. C4). Even if those meas- ures have modified the runoff significantly, there is the question whether they have caused changes from the natural runoff, or back toward the natural runofl’ under pristine (pre-white-man) conditions, which have been described (Central Colorado River Authority, written communication) as a “heavy growth of native grasses and vegetation which retarded the runoff of rainfall.” In comparison with the records from downstream sta- tions in the basin (fig. 2) the record at Ballinger shows no obvious and effective regulation of the natural runofl’. CONCHO RIVER The Concho River basin lies between the upper Colo- rado River basin and the Edwards Plateau, and has some of the geologic and hydrologic characteristics of each, as shown in a report by Willis (1954) which has been the chief basis for the following summary of hydrologic conditions. The upland areas, generally 2,000 to 2,500 feet above sea level, constitute a north- ward extension of the Edwards Plateau and are under- lain by the highly permeable Edwards and associated limestones (p. 011).. The broad valleys of the Concho River and its larger tributaries, however, have been out below the base of these limestones and into the westward-dipping rocks of Permian age, which include gypsum, sandstone, clay, limestone and dolo- mite. Alluvial terrace and flood-plain deposits in these valleys have a maximum thickness of 125 feet. ' Precipitation upon the drainage basin is believed to constitute the source of all potable water within the basin. In most of the area the Edwards and associated limestones constitute the surficial materials and thus the precipitation falls upon permeable materials; how- ever, the limestones do not form an important ground- water reservoir. Although domestic and stock wells obtain supplies of good quality from limestone in sev- eral places, most of the limestone is above the water table and therefore unsaturated. Water infiltrating from precipitation moves through the limestone and dis- charges at springs; this discharge takes place in a short DROUGHT IN THE SOUTHWEST, 1942—56 time, judging by the good correlation of the flow of the Middle Concho River near Tankersly with that of the Brazos River at Seymour and the Colorado River at Ballinger (Gatewood and others, 1963). Of these three streams, only the Middle Concho drains any extensive area of outcrop of the Edwards and associated lime- stones. Most of the rocks beneath the limestone yield im- potable water or no water at all. These rocks crop out in several places along the margins of the valley plains, and some sands and some dolomitic layers yield water of satisfactory quality and quantity for irrigation or for public supplies in the vicinity of their outcrops. However, in some areas of Permian rock outcrops, as for example in the hills just west of the city of ‘San Angelo, there is no ground water suitable for use. The alluvial sediments underlying the terraces and flood plains of the principal river valleys are the best ground-water reservoirs of the Concho River basin, and they are in the areas most suitable for cultivation of crops. Irrigation by means of wells began as early as 1924 near San Angelo; by 1940, about 800 acres was irrigated from wells, and by 1950 the acreage had in- creased to 3,700. This irrigation developed as a form of crop insurance in an area that is marginal as far as water supply for crops is concerned. The average an- nual precipitation at San Angelo is more than 20 inches, of which about 85 percent falls during the growing sea- son; this should be almost enough, provided the rains are well distributed, and in years of greater than aver- age rainfall crops may need no additional supply. But the annual rainfall has ranged from 8 inches in 1917 to nearly 41 inches in 1919, and has been less than the long—time average in all but one of the 14 years 1943—56. Supplemental irrigation is necessary in years of less than average rainfall, and particularly during extended dry periods within a single growing season. The de- velopment and use of ground water for irrigation has been of especial value in rainless periods, and has been indirectly a product of drought. Water levels have declined in wells in all areas where water has been pumped for irrigation, particularly during periods of increasing irrigated acreage and in- creasing rates of withdrawal. Since these declines have occurred during years of drought, there is a funda- mental question whether drought or development has been chiefly responsible. The most rapid declines have been recorded in wells tapping a dolomite in which several irrigation wells have been drilled since 1950 near the town of Veribest; near the center of pumping the static level in some wells was lowered as much as 19 feet in less than a year, and the decline is attributed directly to the pumping. EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS C7 5 4 E E At Smithvil'le, g drainage area 28,000 sq mi < 3 \\ a / .1 é \ {‘2 l E \ ll } E | l 1 '1 O | ‘ 1| 1 g 1 l l 1 I \ 11 “‘ 2 ' ‘ I i r "1 .1 1 1 § 1 ‘H | l‘\ I <2: . NearSan Saba, ./1 l H l I \ r I drainage area 18,700 sq n11, 1 I ll ‘ 1 \ l ' ‘ I " I r ‘ ' l ‘ 1 U 1 \ | ! 1L: 1 1 ' 4 II 1 ‘l 1 r‘ : r l r 1‘ \ [Ll ‘ I I 1 \ I l" \\l \ 1 k" \ A /\ / \l _ AtBallinger, _ A‘ I \ /\ / \ ‘1 drainage area'5240 sq ml l J h \ / 1‘ / A \/\/\/V ' \/\ v V ‘VN \ ‘ \/\ l " 1920 1925 1930 1935 1940 1945 1950 1955 FIGURE 2.—Annual runofl? of Colorado River, 1920—58, at three gaging stations. 1960 08 Most of the irrigation wells obtain water from the alluvium in the river valleys. Water levels have de— clined in these wells also, particularly during the years 1946—52. The hydrographs of figure 3 include one for an irrigation well (Gr—128) in the area just northwest of San Angelo, and one for an unused well (Gr—103) about 8 miles southeast of town. In both these wells the water levels in 1955 were higher than they had been 10 years earlier. From this it appears that the recharge during 1953 and 1954, when precipitation at San An- gelo was about average, was sufficient to replenish the storage that had been depleted by pumping during the preceding dry years. The flow of the Concho River downstream from the area of ground—water develop- ment was probably less than it would have been under natural conditions, because of this replenishment of ground-water storage. Water levels in wells rose mark- edly as a result of the abundant precipitation and run— off in 1957. Lake Nasworthy, which before 1954 was the sole sourCe of municipal water for San Angelo, is at the confluence of the Middle and South Concho Rivers, upstream from the present area of ground-water devel— opment for irrigation and therefore not subject to effects from that development. The lake had an initial ca- pacity of 14,000 acre—feet in 1930, but this capacity had been reduced to 12,400 acre-feet by 1953 because of sedi- mentation. As shown on figure 3, Lake Nasworthy achieved more holdover storage in some drought years (notably 1948—50 and 1953) than in most earlier years, and far more than in 1936, the year of maximum runoff. This apparent anomaly is attributed to changes in the pattern of reservoir operation: in 1936, with severe floods, the primary concern was flood control; in the next decade the lake was maintained at relatively constant level; since 1946, however, the reservoir has evidently stored water whenever available, and although it was practically drained on several occasions during the drought, it has also been filled in most years to higher levels than were reached prior to 1946. San Angelo Dam, on the North Concho River imme- diately west of San Angelo, forms a reservoir with total capacity of 396,400 acre-feet, of which 277,000 is for flood control and 38,800 acre-feet is dead storage. The conservation pool, with capacity of 80,400 acre-feet, is used in part as a supplementary supply for municipal and industrial use in San Angelo. The dam was com- pleted in May 1951 and storage began in February 1952, but the “conservation pool” received its first water in mid—April 1954. The discharge of the Middle Concho River near Tankersly is typical of the highly variable natural flow of streams in this part of Texas: the average annual DROUGHT IN THE SOUTHWEST, 1942—56 runoff in 1930—54 was about 30,000 acre-feet, but the channel has had a maximum discharge of 27,500 cfs, and is ordinarily dry during some seasons. The varia- tion in annual flow is likewise great—from 5 times the median to 1A; the median. During the drought period 1943—56 the annual flow was above the long-term mean in only 3 years, 1948, 1949, and 1954 (fig. 3). The municipal Lake Nasworthy, the conservation pool in the San Angelo Reservoir, and the alluvium that yields ground water to wells all serve to counteract this irregularity and provide a more stable water sup- ply for use—in other words, to reduce the efi'ects of drought. They do this by storing water when precipi- tation is sufficiently abundant, and making it available when the natural supply (of rain or of streamflow) is deficient. Declining levels in Lake Nasworthy and San Angelo Reservoir, and also in wells in the alluvium, are inevitable during periods when the withdrawals of water for use exceed the rate of natural inflow to the reservoirs. MIDDLE COLORADO RIVER The middle Colorado River basin is taken as the drainage area tributary to the river between the gaging stations at Ballinger and near San Saba. This drain- age area of about 7,200 square miles (not including the Concho River basin) is less than one-fourth of the total drainage basin above San Saba, but over a period of about 40 years it has contributed about half the flow measured at the San Saba gaging station (fig. 2). The mean runofl' in the middle Colorado section is equiva- lent to a depth of about an inch of water from the western part of the area and to nearly 2 inches from the eastern part. Like the Concho River, the San Saba River heads in the Edwards Plateau. The limestone ground-water reservoir in that basin has not been studied, and its effect upon streamflow is not known. Modifications by ground—water storage may well affect the flow of San Saba River at Menard, and thus account in part for the poor correlation of that record with records of nearby streams. However, lacking the requisite data concerning the entire ground-water reservoir in the Ed- wards and its efl’ect upon streamflow, all that can be said at this time is that the streams originating in the plateau do not all fluctuate in unison, and that the effects of the ground—water reservoir must be variable. RESERVOIR IN EDWARDS LIMESTONE NEAR AUSTIN Detailed studies of the ground—water reservoir in the Edwards limestone in the San Antonio area form the basis for rather specific conclusions concerning the hydrology and the effects of drought and ground—water 03 U! 0 O DEPTH TO WATER, IN FEET BELOW LAND SURFACE \l o 125 ANNUAL RUNOFF. IN THOUSANDS OF ACRE-FEET 0 EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS .: O I I I N-k‘j I I I I F I I 2 Well (3-103, unused \<>\\\ / . \0——Q /’ \ 9/ so \V 40 Well G—128, irrigaticn )\ \ \O\ ‘ ~ /‘ \\ / \\ o—-—o\ \ 60 Vertical bars show annual range in storage Lake Nasworthy, at con fluenc‘e of Smith Shaded areas represent holdover storage and Middle Concho Rivers E l ! 25 ReserVoir space lost through sedimentation AMiddle Concho River, San Angelo Reservoir, near Tankers y on North Concho River 75_ \ | l | \I \fl 1935 1940 1945 1950 1955 RESERVOIR STORAGE. IN THOUSANDS OF ACRE-FEET 09 FIGURE 3,—Hydrographs for the San Angelo area, Texas, showing annual fluctuations of water levels in wells, storage in surface relervolrs, and runofl of the Middle Concho River, 1933-57. ClO development upon the water resources in a broad area west of the Colorado River drainage basin (p. 016). The Edwards and associated limestones occur also within the Colorado River basin in the vicinity of Austin, but there no detailed hydrologic studies have been made. ‘ In contrast to the San Antonio area there has been very little development and use of water from wells in the Austin area. From, fragmentary information Petitt and George (1956, p. 61) state that there may be a ground-water divide in eastern Hays County, and that the ground-water reservoir in the Austin area may thus be a hydrologic unit separate from that in the San Antonio region. If this is true, the limestone reservoir in the Austin area is not affected by pump- ing in the San Antonio area, except perhaps by some shifting in the position of this divide. Thus it appears that the limestone reservoir near Austin has not been affected by development or use of water from wells. Even so, a variety of factors may affect the storage in and the discharge from the limestone reservoir near Austin. Graphs pertaining to several of these factors are assembled in plate 2. Among these graphs are some that show “inflow” to the area, which is either natural or regulated to only a slight degree; storage or changes of storage in surface and subsurface reser- voirs; spring discharge; and stream discharge from the Austin area. Alternating wet and dry periods of several years’ duration are indicated by the graph of annual precipi- tation at Austin: rainfall was generally less than the long—term mean in the 1930’s except in the wet year 1935; the years 1940—46 were relatively wet, and 1941 was the wettest of the 27 years 1931~57 ; the years 1947— 56 were years of drought, and rainfall was generally less than in the preceding drought of the 1930’s. The graph of monthly precipitation at Austin gives some indication of rainfall intensities in these alternating wet and dry periods. Monthly rainfall exceeded 7 inches in 15 of the 300 months 1931—55, and 5 of these months were in 1935 and 1936, although the period 1931—39 was generally dry. The wetter period 1940- 46 included 7 of these stormiest months, but monthly rainfall exceeded 7 inches only once in the drought of 1947—56. The runoff of Colorado River near San Saba varies widely from month to month, and in the early years of record the months of greatest runofl" frequently coin- cided with the months of greatest rainfall at Austin. However, the 2 months of greatest runoff (July 1938 and September 1936) were not months of extraordinary rainfall at Austin, and it is evident that flood—produc- ing storms may be restricted to only a part of the DROUGHT IN THE SOUTHWEST, 1942—56 drainage basin. Annual runoif near San Saba was greater in 1935 and 1938—2 years within a drought cycle—than in any other year since 1930. From 1939 through 1956 the annual runoff was less than the long- term average in all years except 1941, 1942, and 1955, and the trend was generally downward to all—time lows in 1950 and 1953. The contributing drainage area above the gaging sta- tion on Colorado River at Austin is almost 50 percent greater than that above San Saba, and the average annual runoff is almost twice as great. The runoff as measured at Austin is modified from that measured near San Saba by the storage and regulation afforded by Buchanan Reservoir and Lake Travis, and by the ground-water reservoir in the Edwards limestone. The runoff of Colorado River at Austin has trended generally downward since 1935, more or less like that of the river near San Saba, but there has been less fluctuation in annual discharge, particularly since 1942 when Marshall Ford Dam was completed. Several dissimilarities in the two graphs can be traced to regulation at Lake Travis: the flow at Austin was less than that near San Saba in 1942 and 1949, and almost as low in 1952, because of the increase in storage during those years in Lake Travis; and in 1943 and 1950 the flow at Austin was greater than it would have been under natural conditions, because of releases from Lake Travis. The graphs also suggest some effect of natural regulation by the Edwards ground-water reservoir, particularly in 1944—47 inclusive and in 1953, when the discharge from Barton Springs was fairly high and the river flow at Austin benefited accordingly. The hydrograph of well L—102 appears to be signifi- cant as an indicator of the circulation of water in the Edwards limestone. This well is a dozen miles south of Barton Springs, in the downdropped side of the fault zone. Water levels in this well fluctuate more or less in accordance with the discharge of Barton Springs, indicating that the limestone in the down- dropped side of the fault zone is hydraulically con- nected to the aquifer that feeds Barton Springs. There is also some similarity between these graphs and the graphs for two dug wells where the water level is generally within 20 feet of the land surface and some- times practically at the surface. One might surmise that this shallow ground water is recharged directly by precipitation, and in that case the similarity in hydrographs leads to the suggestion that rising water levels in the Edwards limestone represent a response to precipitation and recharge upon the outcrop area. In both shallow wells (H—136 and L—85) the water rose to a maximum in 1938 that was almost duplicated in several subsequent years (1941, 1944—47, 1957). This EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS common level suggests a reservoir so full that any additional increments of recharge are spilled immedi— ately into streams. From 1947 through 1956 the water levels‘in both shallow wells, and also in the deeper well L—102 that taps the Edwards limestone, were several feet lower than during the period ended in 1946. These lower levels during the drought years are indicators not only of reduced ground-water storage but also of lessened recharge and of lessened base flow in streams draining the area. LOWER COLORADO RIVER The Colorado River below San Saba generates hydro- electric power at six dams operated by the Lower ‘Colo- rado River Authority (LCRA) . The principal storage in this system is in Buchanan Reservoir, which has a capacity of 992,000 acre—feet, and Lake Travis, which has a capacity of 1,950,000 acre-feet. The combined capacity of these two reservoirs is slightly greater than the maximum annual runoff of the river recorded at the San Saba gaging station. Thus these reservoirs can provide regulation of the highly variable daily and seasonal flow in the Colorado River above Austin. The considerable variation in the outflow from the 10west reservoir commonly reflects fluctuations in power de— mand rather than in natural water supply. The graph showing runoff near Smithville, below the LCRA res- ervoir system (fig. 2), is similar in major aspects to ‘ that for the San Saba gaging station, although the run- off at the lower station is about twice as great. Storage began in Buchanan Reservoir in 1937, and in Lake Travis in 1940, and these two reservoirs held back some of the flow in each year from 1937 to 1942, but the total made little difference in the effects of the drought in subsequent years. The chief effect of regulation by these reservoirs has been to make some changes in the yearly pattern of runoff as compared to that near San Saba. Thus the runoff near ‘Smithville in 1943 was con- siderably higher than in 1942 because of releases of water ‘reservoired in the earlier year. And 1952 was the year of minimum runoff near Smithville because of the great increase in reservoir storage above the gaging station in that year. These reservoirs are evidently operated as a manof vision would operate the gas tank of his car—keeping the level in the upper half, so that crises are likely to be farther away. Actually this pattern of operation is dictated by LCRA’s objective of generating hydro- electric power from the water moving through its res- ervoir system. There is perhaps no reason why the lower Colorado River should be so completely regulated that its annual discharge would be at uniform rate in wet years and dry, because most of that water flows un- used into the Gulf of Mexico. However, the reservoirs 011 are not big enough to be capable of such regulation, even if the need for firm supply were to develop. Thus the natural cyclic fluctuatiOns of wet and dry periods of several years’ duration cannot yet be erased by present developmental facilities. GROUND-WATER RESERVOIR IN THE EDWARDS LIMESTONE By B. M. PETITT, JR. The Edwards Plateau is a conspicuous physiographic and geologic feature in the central part of Texas. Viewed from the south the Edwards Plateau is aptly named, because it is more than 1,500 feet above the coastal plain, having been elevated by faulting along the Balcones escarpment which forms its southern boundary. To the northwest the Edwards Plateau merges with the even higher High Plains (p. C22). Streams that rise in the Edwards Plateau have cut steep valleys and canyons below the upland surface, forming areas of pronounced relief. Much of the plateau has been cut into buttes and narrow ridges, and in many places the limestone that caps the highest hills is all that is left of the original plateau. The plateau is made up of limestones that are suf- ficiently vulnerable to weathering and erosion so that the present topography is; one of strong relief, but also sufficiently resistant so that the upland has not been eroded away entirely. They have been subjected to per- colation and solution by water until they now form one of the major ground—water reservoirs in the Southwest. The limestones include the Edwards limestone, the over- lying Georgetown limestone, and the underlying Comanche Peak limestone, which are all of Early Cretaceous age and have a total thickness ranging from about 450 to 1,000 feet. These limestones are here grouped as the Edwards and associated limestones. They are underlain by the relatively impermeable VV‘al- nut clay and Glen Rose limestone, and are overlain by the Grays0n shale, which is composed largely of clay. The Edwards and associated limestones form the sur— face of much of the Edwards Plateau north and north- west of the San Antonio area. Most of the stream valleys along the south edge of the plateau have been cut down to the underlying Glen Rose limestone, and in many places the Edwards and associated: limestones have been eroded from the interstream areas. South of the eroded belt, the Edwards and associated lime- stones reappear in the Balcones fault zone, which forms a curved strip about 20 miles long and 5 to 40 miles wide, and thence dip gulfward under younger forma- tions. Because of this erosion, the ground-Water reser- voir of the Edwards Plateau is not directly connected with the reservoir system in the fault zone except in a 012 few places. However, the two are hydraulically inter- connected because the streamflow from the plateau re- charges the reservoir in the fault zone. The south-draining slope of the Edwards Plateau, with its separation of the limestone reservoir into two parts, provides a unique opportunity to discriminate the effects of natural climatic fluctuations from the efl'ects of man’s development and use of water. The stream- flow originating in the plateau contains ground water that results from infiltration and deep percolation of rainfall on the plateau, as well as stormflow that feeds directly to the streams with little or no modification by man. This streamflow also provides most of the re— charge to the ground-water reservoir in the fault zone, although there is substantial recharge also by direct penetration of rainfall within that zone. The total dis- charge from the reservoir in the fault zone includes both natural discharge by springs and artificial dis- charge by wells, which are used chiefly for municipal supply and irrigation in the metropolitan area of San Antonio. The US. Geological Survey and the Texas Board of Water Engineers have collected water-resources data in the San Antonio area for many years, and most of this information has been published in reports of the two agencies, many of which are cited as references. The collection of data has been greatly intensified since 1949, and a progress report of investigations has been published (Petitt and George, 1956). HYDROLOGY The extensive‘net of precipitation stations maintained by the US. Weather Bureau in the San Antonio area is supplemented by several stations established by the US. Geological Survey. Precipitation is fairly well distributed throughout the year but is heaviest during April, May, and September; the rain falls principally in isolated thundershowers and only occasionally in widespread general rains. The average annual rainfall decreases from east to west. The recent Southwest drought is indicated by less than average rainfall in most of the years 1947—56. The precipitation defi- ciency has been greatest in the western part of the area, where most of the recharge to the ground-water reservoir normally occurs. Conditions are favorable on much of the Edwards Plateau for direct infiltration of rainwater. The water after entering the limestones moves laterally through them and issues as springs in the valleys along the south edge of the plateau where streams have cut down to the relatively impermeable Glen Rose limestone. These springs sustain the perennial flow of the streams in the eroded belt between the Edwards Plateau and the Balcones fault zone. As the streams cross the fault DROUGHT IN THE SOUTHWEST, 1942—56 zone most of the water again infiltrates into the Ed- wards and associated limestones. To a lesser extent there is additional recharge to the ground-water reser- voir by direct infiltration of precipitation falling on the outcrop 0f the limestones in the fault zone. Water levels in wells and the discharge of springs rise rapidly after heavy rains, and then recede more slowly. Figure 4 shows the general relation that exists among precipita- tion, spring flow, and water levels in wells in the Bal- cones fault zone. Several gaging stations have been maintained for 20 years or more on streams rising in the Edwards Plateau, and additional stations have been established since 1952 to aid in determination of recharge to the ground-water reservoir in the Balcones fault zone. A large propor- tion of the runoff occurs within the same month as the precipitation; thus, the runoff from the plateau is rather flashy and not regulated to any high degree by the plateau’s ground-water reservoir. Rainfall produces rapid runoff owing to the steep topography in the pla- teau and widespread exposures of the impermeable Glen Rose limestone in the stream valleys, and also produces an increase in the rate of ground-water discharge through the springs. The mean annual discharge is ordinarily several times as great as the mean discharge during the month of minimum streamflow, except dur- ing years such as 1950 and 1951, when storms and floods were inconsequential and the streams carried chiefly water discharged from the plateau’s ground—water res- ervoir. The eflects of the drought of 1947—56, as shown by the graphs in plate 3, include a reduction in the pro- portion of months of flood discharge, although some high discharges were recorded in 1949, 1952, and 1955; minimum monthly discharges that were as low as or lower than the previous record lows of 1934; and annual runofl' in 1956 that was only a small fraction of the aver- age during the decade prior to 1947. Available data indicate that the ground-water reser- voir in the fault zone consists of a highly permeable net- work of fractures, solution channels, faults, bedding planes, and other openings. Contour maps of the piezo- metric surface (Petitt and George, 1956, p. 62—63) show that water moves through this network generally from west to east; this has been confirmed by computations showing that two-thirds of the total recharge occurs within the Nueces River drainage basin, but more than 90 percent of the discharge occurs within the Guada- lupe River basin. The zones through which water moves are so permeable and so thoroughly intercon- nected 1 that the reservoir is like a surface reservoir in 1The overall permeability of the reservoir does not assure a large yield from a well at any particular place, because the well may miss all the conduits through which water moves, or at least the larger ones. EFFECTS OF DROUGHT IN CENTRAL AND SOUTH TEXAS through Comal Springs, Comal County DISCHARGE, lN BELOW LAND SURFACE CUBIC FEET PER SECONl) 013 ,_ N In “- Water level at Well 26. Bexar County 2 of DJ l— < 3 o i— I l- :1. Lu 9 Accumulated deviation from normal precipitation Boerne. Kendall County (I) LIJ I o E . E z’ 9. l— S > Lu :2 E, m Monthly precipitation, Boerne p— Lu < I i— o I E: E 8 z (z a. , 0 0 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 19471948 1949 1950 1951 1952 1953 1954 1955 1956 1957 Fromm 4,—Correlatlon of discharge from a spring and of water level in 8. Well with precipitation in the San Antonio area, Texas, 1932-57. Drought period shaded. several respects. Thus any changes in inflow or out- flow, though localized, are quickly reflected by changes in head throughout extensive parts of the reservoir. Localized pumping from the reservoir commonly has little efl'ect upon water levels—pumping 17,000 gpm for 17 hours caused a decline of less than a foot in a well 2,000 feet away—but regional pumping of several times that quantity in‘ daylight hours only has Caused a diur- nal fluctuation of more than 4 feet in wells around San Antonio. The ground-water reservoir is of course different from surface reservoirs in many respects. For example, water is confined under artesian pressure throughout 690—176—63—3 most of the reservoir. There may be considerable lag between recharge and the changes of water level in wells or of spring discharge that result from that recharge. Thus in May and June 1954 heavy rains caused floods in several tributaries of the N ueces River; most of the floodwater went into the underground reservoir and caused significant rises of water level in wells in Uvalde County. In other parts of the reservoir, however, water levels in wells continued to decline to an all-time low in 1956 or early 1957. The hydrographs for representative wells throughout the fault zone (pl. 4) Show the general uniformity of water-level fluctuations in the fault zone, which is con- 014 firmed by records from several dozen observation wells scattered throughout the area and by continuous records from key observation wells. The hydrographs all show a downward trend starting in 1947. HYDROLOGIC EQUATION FOR THE FAULT ZONE The hydrologic equation is basically a statement of the law of conservation of matter as applied to the hydrologic cycle. It states that all water entering an area during any period of time must either go into storage within its boundaries, be consumed therein or be discharged therefrom during the same period. As simplified for the ground- -water reservoir of the Bal- cones fault zone, the equation requires that the differ- ence between total recharge to the reservoir and total discharge therefrom be accounted for by changes in storage. For its solution, therefore, the equation re- quires adequate data on recharge, discharge, and stor- age changes. Recharge estimates have been made for the Edwards limestone in the fault zone by Petitt and George (1956, p. 21—40) , who utilized the runoff data of streams origi- nating above the fault zone. Those areas for which there is practically no runoff have been included either as parts of the basins for which estimates of recharge have been made or as separate areas having assumed runofl’ characteristics similar to other basins. The ab- sence of long—term records on all streams except the Nueces and Guadalupe Rivers made it necessary to ex— tend the records of flow by correlation, in order to com- plete the estimates for the 20-year period 1934—53. The recharge is distributed throughout the area but occurs chiefly in the western part. Most of the ground water has been discharged by springs, but the withdrawal of water by wells has been increasing and from 1954 through 1957 it exceeded the discharge by springs. Most of the water is discharged in the Eastern part of the area. The springs are gen- erally along faults that permit water from the lime- stone to escape into cracks and other channels and flow to the land surface. Wells were drilled beginning in 1885, and Bexar County (San Antonio) alone had about 100 in 1907' and nearly 2,000 in 1953, of which about 250 were of large capacity. The total recharge to and discharge from the fault- zone reservoir are compared by several different meth- ods in the graphs of figure 5. The differences between annual recharge and discharge (fig. 50) are the quanti— ties that are inferred from the hydrologic equation to have been added to storage, or if below (the line of origin, withdrawn from storage each yea . In the comparison of accumulated recharge and discharge (fig. 5B), the two curves are close enough to- DROUGHT IN THE SOUTHWEST, 1942—56 gether from 1934 to 1947 to suggest that the ground- water reservoir was in approximate equilibrium throughout the 14 years; however, from 1947 to 1956 the discharge exceeded the recharge, as shown by the divergence of the curves beginning in 1947. This di- vergence, is reflected in the steep downward trend of the graph of figure 51), showing the accumulated dif- ferences between estimated recharge and discharge. These cumulative differences coincide with the decline in water levels 1n wells since the beginning of drought in 1947. If the volume of rock that is saturated or unwatered each year throughout the reservoir were adequately measured, the average storage coeflicient for the reser- voir could be calculated, since the annualwater-storage changes have already been estimated from the hy- drologic equation. However, only rough estimates of the storage characteristics can be made because of the limitations of the data. Water—level changes in Bexar County well 26 (fig. 4) were used as an index of water- level changes 1n the reservoir for several reasons: (a) water-level records from this well appear to correlate with records from wells 1n remote parts of the reser- voir; (b) long—term continuous water-level records are available from this well; and (c) insufl‘icient data are available to determine the average change in water levels throughout the reservoir. Figure 6 shows a plot of year-end water-level altitudes versus the esti— mated difi'erence between annual recharge and dis- charge. During the period of record the range in water levels in the well exceeded 55 feet. Although the correlation appears to be a straight-line function indicating that the change in storage for each foot of change in water level in the index well through the range observed is about 50,000 acre-feet, the curve should not be extended beyond the limits of the data. EFFECTS OF WATER DEVELOPMENT AND USE Under natural conditions the ground—water reservoir of the Balcones fault zone was the source of some of the largest springs in the United States (Meinzer, 1927, p. 29—39). Recreational facilities, irrigated areas, in- dustries, and municipalities have long been established downstream from these springs and have utilized the water, which fluctuated in amount in accordance with the recharge, but had never been known to fail. Wells first tapped this artesian reservoir in 1885. During the 1930’s the average withdrawal from wells was of the order of 100 mgd, and this had increased to an esti- mated 140 mgd by 1946. The estimated average dis- charge from wells in 1955 was about 240 mgd, of which about 105 mgd was used for municipal supply and about 76 mgd was used for irrigation. The water used for 015 EFFECTS OF DROUGHH‘IN CENTRAL AND SOUTHITEXAS $2 4, 44, $2 7/ // /7// 47/77/774 7/// 7//// //7// /////// . D 32 ////////7///4/ . LE2 m 77/ 747W ////4 .mo 4 m 7777/4 m - u 47/74/777 we 1 M 82 7/7 06. -82 U ///,4.. DR c EA - C ”H l A Lm - w. UR - A 32 MN -32 EE as G CE . mR Aw . TE MH FEG mm 82 0mm 82 Rm DFEw. Acs . FNI “WED D. I «m: _ E: m m m o w m o m m m 1 1 5.3. “an; .5 82530.: . 85.185 25¢ moEzomm 25¢ 1 5E mmo< no 825205 Tmfifimw mom