Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2 CAUSES AND OCCURRENCE OF DROUGHT John A. Dracup Civil Engineering Department University of California, Los Angeles Droughts are indiscriminate in terms of geography, climate, and political boundary. Each and every region of the United States has experienced conditions of below normal rainfall and streamflow runoff at some time. As noted by Buchanan and Gilbert (1977), "Hydrologists can pragmatically report on recorded data which show that almost every year, some area of our country experiences the conditions which constitute a drought." Many climatologists have noted a significant increase in the variability of the weather during the past two decades (CIA, 1974; Fritz, 1977; Schneider, 1977; Wallis, 1977~. It is now apparent that the years 1956 through 1971 constituted an abnormally stable period in terms of temperature and precipitation fluctuations, and that the disastrous worldwide weather conditions of 1972 heralded the end of that area. As Mitchell (1977) states, "It appears that we're returning to normalcy, and that means greater variability in the weather than This trend toward increased variability in the weather may be due either to a random fluctuation in the complex weather-generating processes or to a large-scale climate change. Unfortunately, the brevity of available hydrologic records makes it virtually impossible to distinguish between these two possible causes (Lettenmaier and Burges, 1978~. Nonetheless, the Present increase in meteorological variability should we're used to.t' ~ cause a corresponding increase in the occurrence of hydrologic droughts, at least in the immediate future. Hence from these considerations it is apparent that the study of droughts is an extremely relevant and essential aspect of water resources analysis. A surprisingly limited amount of attention has been given to hydrologic drought events in the literature -24-
(Whipple, 1966). A cursory survey of the Water Resources Abstracts substantiates this claim, it is found that the ratio of papers indexed under "Floods" to those indexed under "Droughts" exceeds 6 to 1 in all cases, and was nearly 10 to 1 in 1976. In addition, many of the "drought" entries actually refer to meteorologic droughts or low flows rather than to hydrologic drought events. DEFINITION OF DROUGHT What is a drought? It brings to mind extreme pictures of emaciated humans, hunger, and famine,- or simply empty reservoirs, parched fields, and dusty roads. However, a precise definition of drought is difficult to obtain (Dracup, 1980a). The difficulty in devising an objective definition of drought is fourfold. One source of confusion is the unavoidable diversity in the ways in which various fields of study view drought events. A water resource engineer views drought as a problem in supply and demand. He may state "there is no drought in the ocean," meaning, of course, that without a specified demand no drought can occur however severe the precipitation shortage may be. Thus the engineer views drought as a shortage in streamflow runoff and water storage. The geophysicist's view of drought will include climatology, meteorology, hydrology, limnology, and oil physics (Yevjevich, 196 7~. The farmer and agriculturalist view drought as a function of the specific crop under cultivation. Others may view drought as a function of the impact the drought has on institutional and human activity and on their response during the drought. A second factor militating against an objective drought definition is the variety of connotations given to the term "drought" in different parts of the world. For instance, in Bali any period of 6 days or more without rain is considered a drought, while in Libya droughts are only recognized after 2 years without rain; in Egypt before construction of the Aswan Dam, failure of the Nile River to flood constituted a drought, regardless of rainfall (Hudson and Hazen, 1964~. In Britain an "absolute drought" has been defined as a period of at least 15 consecutive days without 0.01 inch of rain on any one day, whereas a "partial drought" is
-26- taken to be a period of 20 consecutive days for which the mean daily rainfall does not exceed 0.01 inch (Rodda, 1965~. Other examples exist in the literature (Tannehill, 1947) that serve to demonstrate that definitions of drought events are strongly related to the climatological and geological traits of a particular locale. The third problem in drought definition from the point of view of the hydrologist is that a drought event is manifested in terms of both a precipitation deficiency and a streamflow deficiency. Hence a thorough and complete definition of drought events requires consideration of both rainfall and runoff. However, due to constraints of time, economic resources, and professional expertise, most drought studies have focused on only one aspect of the drought event. Thus two drought definitions are often specified, one based on precipitation and the other based on runoff. Finally, there is a curious lack of uniformity in the conventional terminology relating to different hydrologic events. For instance, since the term low flow denotes an annually occurring minimum flow of short duration, one would expect the annually occurring maximum flow of short duration to be called a high flow; however, this is usually termed a flood event. In addition, because the term high flow may be used to refer to extended periods of above mean discharge, one would expect extended periods of below mean discharge to be called low flows; however, these are usually termed drought events. This lack of symmetry in hydrologic - terminology is shown schematically in Figure 2-1. It is in this context that general definitions of drought events have evolved. These definitions have, of necessity, been broad in scope so as to apply to as wide a variety of particular drought manifestations as possible. An extreme example of this is the drought definition offered by Matalas (1963~: "A drought is defined, in a broad sense, as an extended period of dryness." In addition, most definitions describe drought in relation to some locally determined water requirement. For instance, the U.S. Weather Bureau defines drought as a "lack of rainfall so great and long continued as to affect injuriously the plant and animal life of a place and to deplete water supplies both for domestic purposes and for the operation of power plants, especially in those regions where rainfall is normally sufficient for such purposes" (Havens, 1954~. A typical
2 7 Di sc ha rge mean o FLOOD ~ I GH FLOW LOW FLOW DROUGHT . . day month year FIGURE 2-1 Classification of hydrologic events. Duration
-28- textbook definition of drought is given by Linsley et al. (1982~: ''a period during which streamflows are inadequate to supply established uses under a given management system.'' A less subjective definition is offered by Whipple (1966~: "the term drought will refer to prolonged periods of runoff, averaging less than the long term mean." The generality of these definitions clearly leaves them open to subjective interpretation by individual researchers, a situation that has often provided a barrier to the advancement of the state of the art of drought analysis. In the midst of this ambiguity and complexity, one potentially satisfying drought definition has emerged in the past decade as being objective and yet flexible enough to be applicable to a wide variety of drought concepts. In effect the proposed approach merely systematizes the intuitive intentions of the various definitions mentioned above. The drought definition of interest has been proposed by Yevjevich (1967), and is based on the branch of statistical analysis known as the theory of runs. This is a method of analyzing a sequential time series of stochastic or deterministic variables, and hence is well suited to the study of hydrologic event. The fundamental parameters of the runs of an annual hydrologic series are shown in Figure 2-2. The parameters best suited to drought definition are tL (drought duration) and YL (drought severity). In the terminology of the theory of runs, t is referred to as a run-length and y as a run-sum. A third parameter may be identified by forming the ratio of y to t; this represents the average magnitude of the drought event. The selection of xo, the base value index by which all other values of a hydrologic variable x are described, is an important decision. Figure 2-2 indicates that the run-sum and run-length are determined by the choice of xo. Some candidates values for xo are the mean of the x series, the median of the x series, or an expression such as X0 = Xm + e Sx where Xm is the series mean, SX is the series standard deviation, and e is an elective scaling factor. It is noted that xo need not be a constant; it may be a stochastic variable, a dete'~,~inistic function, or any synthesis of the two.
29 Annual Average Flow X o Runs Parameters: -H '\ ITCH .~N _ _ V _ _~! . . . . l\ v . . 1 2 3 4 5 6 7 Ye = Run-sum above XO YL = Run-sum below XO (drought severity) YL/tL = Run-length above XO Run-length below XO (drought duration) Drought magnitude 8 9 10 11 Years FIGURE 2-2 Parameters of the runs of a hydrologic series.
-30- Once the desired value of xo has been stipulated, the time series is divided into upper (above xo) and lower (below xo) sections. A primary advantage of using runs to define droughts is that the run-sum and run-length series of the upper and lower sections are amendable to statistical analysis in order to determine properties such as time dependence, probability distribution, or serial correlation. These properties may be determined analytically or by a suitable data generation approach (e.g., the Monte Carlo method). In the case of drought events this implies that the statistical properties of drought duration, magnitude, and severity may be assessed (Dracup et al., 1980b). The objectivity of this method of drought definition is one of its strongest advantages; if a number of analysts select the same value of JO to be applied to a particular hydrologic record, then each should derive identical statistical properties for duration, severity, and magnitude. The method's flexibility is due to the freedom allowed in the choice of xo; the hydrologist may use mean annual flow, while the agriculturalist may prefer seasonal soil moisture. The combination of these two characteristics in an approach to drought definition and description is particularly attractive because of the wide variety of drought concepts held throughout various scientific disciplines. CAUSES OF DROUGHT IN TEMPERATE LATITUDES The material covered in this section concerns the statistical, synoptic, and physical aspects of drought on time scales of a month to several years. The current literature reveals that the causes of drought are complex and are not yet completely understood by meteorologists and climatologists. Namias (1985) states . . . it should be made clear that there are many unsolved "mysteries" of drought. While some physical understanding has been achieved for droughts that last for a month to a season, spells of years characterized by drought are poorly understood, and thus remain on the agenda for research climatologists. When studying the factors responsible for drought, one is struck by the "chicken and the egg" analogy, that is,
-31- which of the processes actually occurs first? Perhaps a gradient of sea surface temperatures (SST) first is formed between the warm eastern Pacific and the cold central and western Pacific. This process may cause high-pressure cells in the mid-troposphere to emerge and persist. This in turn causes subsidence (sinking) or warm, dry air in the middle troposphere, which then causes adiabatic heating, low relative humidity, and a reduction in the growth of cumulus clouds. The result is a reduction of precipitation. The reduced cloud cover and precipitation in turn increases insolation, drying out the soil and increasing its albedo, thus further aggravating the process. Let us investigate each of these phenomena in turn. Recently, it has been suggested that there are important teleconnections (relationships between two phenomena that are physically hundreds or thousands of kilometers apart) within the atmosphere-ocean system (Namias, 1978a). As shown in Figure 2-3, there appears . to be a teleconnection between a warm SST anomaly and the continental high-pressure cell. It is well known that high-pressure cells exist over drought-affected regions and that companion oceanic high-pressure areas exist simultaneously. As shown in Figure 2-4, strong high-pressure areas exist over the Atlantic and the Pacific as well as the continental United States. These high-pressure areas help shape the upper-level long-wave westerlies. In turn, the positioning of these long-waves determine to a large extent climatic variations. During drought periods there is a general northward shifting of pressure zones. Therefore the westerlies are displaced northward over the continental United States during droughts, bringing stronger than normal winds to the high latitudes as shown in Figure 2-5. The result of these atmosphere anomalies is the presence of (relatively) warm, dry air in the middle troposphere. In drought regions, the warm air aloft subsides at the rate of several hundred meters per day. The sinking of the dry air and the attendant adiabatic heating of it inhibits precipitation through the suppression of the growth of cumulus clouds. The subsidence of the warm, dry air aloft is caused by air flowing out of the bottom boundary layers of the high-pressure cells. This air is replaced by the further sinking of air masses aloft. The warm air aloft resulting from the subsidence is shown in Figure 2-6.
32 tom' ._ .h A, ~ i] ,... - c i~ 1~1 u to 5- . - o o lo. , CO o ~ ~ o ~ so o c) of ~ ~ o en U ~o~ = - so o ~ o cn Carl ~ . Us o o cr u' a' o ~ V .,, U ~ ~ ^= a' 0 U ~ ~ , cn V C In so en Ct a) U] - 1 cq Cal ~ Cal o H C) 00 o · ~ o o a) U a, a, U ~0 U' o ~rl cn CO .e S~ Ln CD ~ U. ~' a S~ . o ~: o o .. C' S~ O O . ~ ~ o C~ U) ~ Z; ~ O .,1 N ~ri 00 S~ o
33 AUGUST I 936 I,: , ,~Z 1~ 700 mb . . CINear Normal Abme Normal TEMPDN FIGURE 2-4 (Top) Average contours of the 700-mb surface for August 1936, a drought month. (Bottom) Average temperature departures from normal (OF) for August 1936. (DN: departures from normal). SOURCE: Beran and Rodier (1985). Permissions granted World Meteorological Organization and UNESCO Press.
34 ~ i > - CY ! l 1 A) 1 1 0 J ~ \ 1 1 l l ' ,' Yi, 1\ / i' \ ~ _ . ~ / 1~ _/ _' ., A) Lo) . _ CY o 7 1 V; . ~ 1 ma_ ' \\ \\ L ~ Cal 'l~1 S~0153Q ,LD UP US an O U) Q. Cal Hi O it- I _ ~ 1 : ~ a, a, CO curl a, sol cn a) ~3 a: o P4 a, ~ U~ U] . U] ¢ e - | o N 1 o o O O O S~ 1 C~ U' C) O ~ O . - o o S~ o S~ o a S~ ~0 00 o ~rl CO CO ~rl S~ O .,1 o U] U] a o U) z . a, 0 ·- N ~ ~1 C' O S~ U. O
35 / ~ 6 - ~ 5- at: 4 3 2 10 9 8 OAKLAND, CALI PORN IA U.S.A. \~ \\1976 None\ ( 190-~\\ Rainfall 1976 9 mm Normal Ill mm O . -50 -so JANUaRY 'I 39~\ \:Re~t, - Humidity \\\ -- - 1 , r 3 \.\\s~. \\\ 46 \~ 426% \I' 84~ol \.7S% -30 - -20 -10 0 - 10 20 TEMPERAT1JRE (°C) FIGURE 2-6 Upper-air temperatures in the core (Oakland) of the California drought for January 1976 and the normal temperatures. Numbers beside temperature plots give relative humidities. Rainfall amounts are given in the lower left. SOURCE: Beran and Rodier (1985~. Permissions granted World Meteorological Organization and UNESCO Press.
-36- Here the increase in temperature at any given height is evident. Once the high-pressure cells are formed and the drought is initiated, there is a self-perpetuating mechanism that continues the process. One theory holds that the land, rendered hot and dry during drought, further heats the air above it and thereby enhances the regional high pressure zone. A second theory states that drought causes an increase in fine dust particles in the air, which lead to high concentrations of very small cloud droplets whenever cumulus clouds form and thus make it more difficult for precipitation to form (Twomey and Squires, 1959~. A third theory suggests that high albedo in dry areas creates a mechanism that produces warm air aloft (Charney, 1975~. Whatever the mechanism, there is statistical evidence that hot, warm springs over the plains of the United States are followed by hot, dry summers and these tend to persist from one year to the next (Namias, 1960~. In reviewing the literature in meteorology on the ; causes of drought, one is struck by the many different mechanisms that in themselves are well understood. That is, SST anomalies, high-pressure cells, long-wave westerlies, warm air subsidence, increased albedo, and so on, are all understood by atmospheric scientists. However, the interrelationships between these variables and how their interrelationships cause a persistent drought, are not well understood. POTENTIAL FOR DROUGHT FORECASTING A scientific forecast of a geophysical event such as a drought must include four vital components. These are the event's time of occurrence and its magnitude, duration, and location. For example, the essential elements required for the prediction of an earthquake must include the expected time that the earthquake will occur, along with its expected magnitude (i.e., Richter scale value), duration, and location. In the case of a drought, forecast must include location (i.e., areal extent), duration, and magnitude. The product of duration and magnitude result in the determination of drought severity. The forecasting of any catastrophic event is fraught with difficulties, and droughts are no exception. This section summarizes the current state of the art of
-37- long-range forecasting. For a more extensive review the reader is referred to Namias (1953, 1968, and 1978b). The forecasting procedures can be divided into meteorological methods and hydrological methods. The meteorological methods include analogues, linear regression models, teleconnections, statistical and kinematic models, air-sea interactions, statistical time series forecasts and extrapolation in time using cyclicities (Beran and Rodier, 1985~. Only teleconnections and air-sea interactions will be discussed here. Teleconnections are shown on charts that specify conditions remote from the area being forecasted. As indicated on Figure 2-7, the correlation coefficients for all grid points are specified at the 700-mb level. Other examples of teleconnections include the following: 1. Links between SST and continental weather (Namias, 1963; Shulka et al., 1977~. 2. Wind in East Africa and monsoon in India (Findlater, 1977; Raghaven et al., 1975 and 1978~. 3. General circulation and sea surface "signature antecedent to deficit seasons in Brazil, Central America, and Africa" (Hastenrath, 1976 and 1978~. The use of air-sea interactions first involves measuring the macroscale thermal character of the sea surface. The second step is to determine the atmospheric circulations that are compatible with the SST anomalies. Correlations are then made and teleconnections are used to translate these data into circulation patterns, temperature, and precipitation over continental areas (Namias, 1976~. The hydrological methods of forecasting drought include streamflow-regression-based methods, regression methods, the analysis of cycles in annual streamflow, and other statistical analyses of historical streamflow events. Probably the most widely used method of hydrologic forecasts is the regression method. 'A linear regression equation of the form Y = ho + blX1 + b2X2 + bnXn may be used for drought and/or streamflow forecasting where Y may be the pressure, temperature, or streamflow
38 700 rnb SUMMER MONTHS TELECO~CTIONS ( CROSS -CORRELATIONS ) ''I ~- ~, ~ .\_\ ~% ~ ~ it. _. _% I, V'~'-~ Mali l l . , ~'~ 1 t~ I: ~ i. ~7r in FIGURE 2-7 (Top) Teleconnections between 700-mb heights over the field as correlated with a point in the North Pacific (label' 1.00~. (Bottom) Same with a point in the North Atlantic (labeled 1.00~. Note in each case the positive correlation with 700-mb heights over the central United States. SOURCE: Beran and Rodier (1985~. Permissions Meteorological Organization and UNESCO Press. granted World
-39- at a point. The variable Y may take the form of an index summarizing the total weather situation derived using the principal components method. The forecasting X variables must be observable at the time the forecast is to be made and like the Y variable can include prior values of pressure and temperature, summarizing indices sea surface temperature, wind, ice extent, and even sunspot number or some other cyclic variable. The estimation of the coefficients be, bl, etc., is by least squares and makes use of a run of back data, typically 30 years for seasonal or annual forecasts. In the case of streamflow, soil moisture and climatic factors such as precipitation and temperature are used as the independent variables. V ~ 1 ~ ~ ~ CURRENT OPERATIONAL DROUGHT FORECASTING TECHNIQUES IN CAL IFORNIA AND THE WESTERN UNITED STATES The current operation drought forecasting techniques as presented here are divided into meteorological methods and hydrological methods. Meteorological Forecasts The Scripps Institute of Oceanography (SIO) provides the California State Department of Water Resources (CDWR) with seasonal temperature and precipitation forecasts as shown in Figure 2-8. Namias (1984) states that these forecasts are physically based using statistical and synoptic methods which employ large scale fields of Northern Hemisphere atmospheric geopotential height and Pacific Ocean sea surface temperatures. The forecasts are made in equally likely tercile classes, light (L), moderate (M), and heavy (H). The forecasts are called "experimental" and are part of ongoing research at SIO aimed at improving long-range-forecasting techniques. A total of four forecasts is made each year starting in September. The fall forecast is for September, October, and November In December, the winter forecast is for December, January, and February. Similarly, three-month forecasts are made for the spring and summer months. In addition
-4~ PREDICTED WINTER G82-83 tOEC. 82, JAN., FEB. 83 ) j.~ TEM,PERATURE ~` ~ ~.·~-.·.~ .~\ BBELOW NORM i 4 ~ ~ i NNORMAL ! rV./ AABCVE NORM i 1 ~1 ~1 ~.. 1 _ ~ _-;~ · · a · ~ ~~ -~.~5 A ~ ;, _ · _ · ·~a · ~ . · · · ·t ~ . ~ . .] ~ . .~. )~ IV / ~ ~ ~ ~ ;i- ~ - ~ - e>. ~- a - ~; ~ i ~ H · ~ ~ . ~ ~ .-2`. ~ -e ~ ~ a ;~. ~ _. 3~\ r" _, ;~ '~~ ·-a!- ~ ~ · ~e~" '~ · '- · ·~< ~ \ , ~' · ~ ~ ~t · ~ ~ ~ 's ~ te ~ ~ ~ , ~ ~ a ~a a a ~ ,j - e . ~ \ _ ~ {~e ~ ~ i · - ~ -'~--~i~ ~-~ - ^~;';-~~ / \ ~ \/ _ , LLIGHT ~ | ' 'ce · · . . . . · ., . · ~ · ... . ~ - . - . -1. - . ,. -~ , ~ ~L ~ [~~ a ~al a ·~a~-~a~~'~.-~ ' 1t A ~ \ -/~\ ~ ~-- - \~a-~-~;~;~;~;~;-\ -7~~ ~ 1Vl_ \ I PRECIPITATION l - ` · e~ · ~ ~ ~ ~ ~ - ~ ~ - ~ ~ [a · /~ ~ \ I MMODERATE i / UFA\/v I I Completed Nov. 29,1982 from data ending Nov. 23,1982 J.lMamias FIGURE 2-8 Forecasts of temperature and precipitation for winter 1982-1983. SOURCE: Namias (1984~.
-41- to the precipitation forecasts, the movement of various upper level winds and storm tracks also are predicted shown in Figure 2-9. In addition to pictorial forecasts, an annual report is furnished (Namias, 1984~. The report contains a discussion of results of the project forecasts to date and an analysis using a "skill score." The CDWR translates these forecasts into California Water Supply Outlook report (1985) and a State Water Project Water Delivery Rule Curve and Criteria for 1985 report. These reports are mainly used to operate reservoirs throughout California mainly for irrigation supply deliveries (agricultural irrigation accounts for 87 percent of water supply deliveries in California). In addition to DiC tonal Hydrologic Forecasts The National Weather Service (USDC) and the Soil Conservation Service (USDA) jointly publish monthly reports on Water Sunplv Outlook for the Western United States for f year. Similarly, the CDWR publishes a report entitled Water Conditions in California. These reports contain streamflow forecasts for the entire water year (October through September), streamflow forecasts for specific on irrigation periods (April to September), and data current reservoir storages. It is important to note that these forecasts are for unimpaired flows. Similar reports are published throughout the United States. The meteorologic and hydrologic forecasts are principally used to allocate irrigation water supplies to senior and junior irrigation districts with appropriative rights in California and throughout the West. The only "drought index" used in California is the 'iFour River Index," which comprises the sum flows of the Sacramento River, the American River at Folsom, the Yuba River at Smartville, and the Feather River inflow to Oroville Dam. This index is published in Water Conditions in California (Bulletin 120-85) and is - pr~mar~ly used as a means to meet salinity standards in the Sacramento Delta. Other input to the Water Conditions in California report include snowpack measurements, precipitation, reservoir storages, and streamflow runoff.
42 ran m LL] Lee Z A: of an: Car ~ k: he En: o 11 llJ C) LLJ En: Q a' 1 Cat oo set o W~ I A u of o U' set o o 0 3 o - 0 oo 0 ~ C) set o oe Z 1 Cat ·e IN P Pa Cat = o U)
43 FORECAST ACCURACY In Thomas Heggens classic play Mister Roberts (1946), he describes life aboard a cargo ship during WWII: Now in the waning days of the second World War, this ship lies at anchor in the glassy bay of one of the back islands of the Pacific. It is a Navy cargo ship. You know it is a cargo ship by the five yawning hatches, by the house amidships, by the booms that bristle from the masts like mechanical arms. You know it is a Navy ship by the color (dark, dull, blue), by the white numbers painted on the bow, and unfailingly by the thin ribbon of the commission pennant flying from the mainmast. Heggens goes on to state: It has shot down no enemy planes, nor has it fired upon any, nor has it seen any. It has sunk with its guns no enemy subs, but there was this once that It rlreo. Tnls periscope, the lOOKOUt sighted it way off on the port beam, and the Captain, who was scared almost out of this mind, gave the order: "Commence firing "' The five-inch and the two port three-inch guns fired for perhaps ten minutes, and the showing was really rather embarrassing. The closest shell was three hundred yards off . . . Hopefully, meteorologic and hydrologic forecasts are currently doing better than the guns of the USS Reluctant. Namias evaluates his meteorological forecasts for California and the western United States using the following skill score equation: Skill = Correct Forecasts - Correct Forecasts Expected by Chance Total Forecasts - Correct Forecasts Expected by Chance For the time period 1975-1976 through 1981-1982, skill scores of approximately 0.67 were experienced. However, during the 1982-1983 water year a skill score of only +0.25 (6 out of 12 correct forecasts) was achieved, and during the 1983-1984 water year a skill score of 0.00 (4
-45- These and other problems raise such questions as who is using these forecasts, what are their actions as a result of the forecasts, and do the long-term benefits exceed the potential short-term costs? That is, can many years of benefits be countered by a single lawsuit from a single faulty forecast? Or should these forecasts be made by the private, rather than the public, sector, which would assume liability via a corporate structure? - CONCLUDING REMARKS This chapter, which focuses on the causes and occurrence of droughts, deals with the technical aspects of this subject. However, in considering this topic and the broader issue of drought management and its impact on public water systems not only the engineering and technological aspects of the problem should be analyzed, but also the areas of economics, finances, legalities, politics, society, and the environment. For example, suppose an elaborate international network for drought prediction were developed, which included ground sensors, report collection stations, telemetering satellites, simulation models, and information dissemination centers. Such a system would be required to answer such questions as do the benefits exceed the costs, who will finance such an enterprise, is such a system completely legal under present laws, is it supported politically, will society accept the end results, and is it environmentally feasible? Only if such a drought prediction system can successfully pass each of these feasibility tests will it actually be established. REFERENCES Beran, M. A., and J. A. Rodier. 1985. Hydrologic Aspects of Drought, Studies and Reports in Hydrology Series, UNESCO/WMO Panel, Rep. 39. Buchanan, T. J., and B. K. Gilbert. 1977. The drought: A pervasive problem. Water Spectrum 9~3~6-12. California Department of Water Resources. 1985. Water Conditions in California. Bull. 120-85. Sacramento Calif.
46 California Department of Water Resources. 1985. State Water Project Water Delivery Rule Curve and Criteria for 1985. Sacramento, Calif. California Department of Water Resources. 1985. California Water Supply Outlook. Sacramento, Calif. Central Intelligence Agency. 1974. A Study of Climatological Research as it Pertains to Intelligence Prob lems. Office of Research and Development, CIA, Washington, D.C. Charney, J. G. 1975. Droughts in the Sahara. Science 187:435-436. Dracup, J. A., K. S. Lee, and E. G. Paulson, Jr. 1980a. On the definition of droughts. Water Resour. Res. 16~2~:297-302. Dracup, J. A., K. S. Lee, and E. G. Paulson, Jr. 1980b. On the statistical characteristics of drought events. Water Resour. Res. 16~23:289-296. Dracup, J. A., D. L. Haynes, and S. D. Abrams on. 1985. Accuracy of Hydrologic Forecasts, Proceedings of the 53rd Annual Meeting, Western Snow Conference, Boulder, Colo. Findlater, J. 1977. A Numerical Index to Monitor the Afro-Asian Monsoon During the Northern Summer. Meteorol. Mag. 105:134-143. Fritz, H. C. 1977. Cited in Drought--Is Stable Climate at an End?, Los Angeles Times. Pp. 1, 20, 21. Glantz, M. H. 1982. Consequences and responsibilities in drought forecasting: the case of Yakima, 1977. Water Resour. Res., 17~1~:3-13. Hastenrath, S. 1976. Variations in low latitude circulation and extreme climatic events in the tropical Americas. J. Atmos. Sci. 33:202-215. Hastenrath, S. 1978. On Modes of Tropical Circulation and Climate Anomalies. J. Atmos. Sci. 35:2222-2231. Havens, A. V. 1954. Drought and agriculture. Weatherwise 7:5 ~51, 68. Heggens, T. 1946. Mister Roberts. Houghton Mifflin, New York. Hudson, H. E., and R. Hazen. 1964. Droughts and Low Streamflow, Handbook OIC Applied Hydrology, Section 18, V.T. Chow (ed.~. McGra~Hill, New York. Lettenmaier, D. P., and S. J. Burges. 1978. Climatic change: detection and its i~pact on hydrologic design. Water Resour. Res. 14~4~:679-687. Linsley, R.K., Jr., M. A. Kohler, J. L. H. Paulus. 1982. Hydrology for Engineers. 3rd edition. McGra~Hill, New York.
47 Matalas, N. C. 1963. Probability Distribution of Low Flows. USGS Professional Paper 434-A. U.S. Government Printing Office, Washington, D.C. Mitchell, J. M. 1977. Cited in Drought--Is Stable Climate at an End?, Los Angeles Times. Pp. 1, 20, 21. Namias, 3. 1934-1982. Short Period Climatic Variations, Collected Works, vols. I, II, and III. University of California, San Diego. Namias, J. 1953. Thirty-day forecasting: a review of a ten-year experiment. Meteorol. Monogr., Am. Meteorol. Soc. 2~6~:83. Namias, J. 1960. Factors in the Initiation, Perpetuation And Termination of Drought. IAHS Publ. 51. Symposium on Surface Waters, Helsinki. Pp. 81-94. Namias, J. 1963. Surface atmosphere interactions as fundamental causes of drought and other fluctuations, in Proceedings of the Rome Symposium on Changes of Climate. UNESCO (Arid Zone Research--XX), Paris. Pp. 345-359. Namias, J. 1968. Long-range weather forecasting-- history, current status and outlook. Bull. Amer. Meteorol. Soc. 49~5~:438-470. Namias, J. 1976. Seasonal forecasting experiments using North Pacific air-sea interactions, preprints of Sixth Conference on Weather Forecasting and Analysis, Albany, N.Y., American Meteorological Societies. Pp. 13-16. Namias, J. 1978a. The enigma of drought--a challenge for terrestrial and extra terrestrial research, in Proceedings of the Symposium on Solar Terrestrial Influence on Weather and Climate. Ohio State University, Columbus. Namias, J. 1978b. Recent drought in California and Western Europe. Rev. Geophys. Space Phys. 16:435-458. Namias, J. 1984. Seasonal Precipitation Forecasting Ninth Annual Report, Scripps Institut. of Oceanography, La Jolla, Calif. Namias, J. 1985. Factors responsible for the droughts, Chapter 3 In Hydrologic Aspects of Drought, M. A. Beran and J. A. Rodier, rapporteurs. Rep. 39. Studies and Reports in Hydrology. UNESCO-WMO, Paris. Raghaven, K., P. V. Puranik, V. R. Mujumdar, P. M. M. Ismail, and D. K. Paul. 1978. Interaction between the West Arabian Sea and the Indian Monsoon. Mon. Weather Rev. 106~5~:719-724. Raghaven, K., D. R. Sikka, and S. V. Gujar. 1975. The influence of cross equatorial flow over Kenya on the
-48- rainfall of western India. Quart. J. Roy. Meteorol. Soc. 101:1003-1005. Schneider, S. H., 1977. What climatologists can say to planners, in Proceedings of the Symposium on Living with Climatic Change, Phase II, Reston, Va. Pp. 45-56. Shulka, J., and B. M. Misra. 1972. Relationships between sea surface temperature and wind speed over the Central Arabian Sea and monsoon rainfall over India. Mon. Weather Rev. 105~8~998-1002. Siegel, B. 1985. Weather: lawsuit put accuracy of prediction to the test. Los Angeles Times, March 28. Tannehill, I. R. 1947. Drought: Its Causes and Effects. Princeton University Press, Princeton, N.J. Twomey, S., and P. Squires. 1959. The influence of cloud nucleus population on the microstructure and stability of convective clouds. Tellus 11:408-411. Wallis, He R. 1977. Climate, climatic change, and water , , ~ , supply. EOS, Trans. AGU 58~11~:1012-1024. Whipple, W., Jr. 1966. Regional drought frequency analysis. ASGE J. Irrig. Drainage Div. 92(IR2~:11-31. Yevjevich, V. M. 1967. An Objective Approach to - Definitions and Investigations of Continental Hydrologic Droughts. Colorado State University Hydrology Paper No. 23.
