PNAS-2010-MacDonald-21256-62

Water, climate change, and sustainability inthe southwestGlen M. MacDonaldInstitute of the Environment and Department of Geography, University of California, Los Angeles, CA 90095-1496

Edited by B. L. Turner, Arizona State University, Tempe, AZ, and approved October 26, 2010 (received for review August 29, 2010)

The current Southwest drought is exceptional for its high temperatures and arguably the most severe in history. Coincidentally, there hasbeen an increase in forest and woodland mortality due to res and pathogenic outbreaks. Although the high temperatures and aridity areconsistent with projected impacts of greenhouse warming, it is unclear whether the drought can be attributed to increased greenhousegasses or is a product of natural climatic variability. Climatemodels indicate that the 21st century will be increasingly arid and droughts moresevere and prolonged. Forest and woodland mortality due to res and pathogens will increase. Demography and food security dictate thatwater demand in the Southwest will remain appreciable. If projected population growth is twinned with suburb-centered development,domestic demandswill intensify.Meeting domestic demands through transference from agriculture presents concerns for rural sustainabilityand food security. Environmental concerns will limit additional transference from rivers. It is unlikely that traditional supply-side solutionssuch as more dams will securely meet demands at current per-capita levels. Signicant savings in domestic usage can be realized throughdecreased applications of potable water to landscaping, but this is a small fraction of total regional water use, which is dominated byagriculture. Technical innovations, policy measures, and market-based solutions that increase supply and decrease water demand are allneeded.Meeting 21st-century sustainability challenges in the Southwestwill also require planning, cooperation, and integration that surpass20th-century efforts in terms of geographic scope, jurisdictional breadth, multisectoral engagement, and the length of planning timelines.

From prehistoric pueblos to todaysburgeoning suburbs, water scarcityhas posed sustainability challengesfor the people of the Southwest. In

the 21st century, these challenges are be-coming acute. Since 2001, large portions ofthe arid Southwest (dened here as Cal-ifornia, Nevada, Utah, Arizona, and NewMexico) have experienced prolongeddrought. Particularly widespread droughtoccurred in 2002, 2003, 2007, and 2009(1). During these years, the regions pre-cipitation averaged as much as 2225% be-low the 20th-century mean, with localdecits being greater. In 2002 and 2009,annual precipitation in Arizona was 40%below normal (2). The effects of low pre-cipitation have been exacerbated by hightemperatures, increased evapotranspiration,and decreased runoff. The average annualtemperature for 20012009 was 0.8 Cwarmer than the 20th-century mean (2).The Colorado River is a critical conduit

of water in the Southwest and is appor-tioned to supply 20,400 million m3 (16.5million acre feet; MAF) of water to thebasin states and Mexico (3). Of that, about12,400 million m3 (10.0 MAF) are allo-cated to the arid Southwest. This repre-sents approximately one sixth of theannual water use for irrigation, domesticneeds, and industry (4). In Nevada, theriver mainly supports the domestic andindustrial demands of the Las Vegas re-gion, whereas in southern California about70% is used for agriculture. The allocationof Colorado River water was based uponan early 20th-century average annual owof around 20,970 million m3 (17.0 MAF)at Lees Ferry, Arizona. For 20012006,the estimated natural annual ow at LeesFerry averaged 13,814 million m3 (11.2

MAF), dropping as low as 7,647 millionm3 (6.2 MAF) in 2002 (3).Higher temperatures, earlier spring

warming, and decreased surface watercontribute to an increase in wildres. InCalifornia, the 2 largest wildres on recordand 11 of the 20 largest recorded resoccurred in the past decade (5). Outbreaksof forest pathogens such as bark beetlesare also promoted by higher temperaturesand drought. According to the US ForestService, the current outbreaks, occurringsimultaneously across western NorthAmerica, are the largest and most severein recorded history (6).The purpose of this special issue is to

assess current and future drought andchronic water-related challenges in theSouthwest and consider the problems andprescriptions for 21st-century sustainabil-ity. A particular focus is placed on thepotential impact of greenhouse warmingon current and future hydroclimatology.This issue cannot address all aspects ofthe water resource questions facing theSouthwest. Nor is it intended to presentexhaustive reviews of earlier work. In thispaper, I will set the spatial, temporal, andsustainability context for the Early 21st-Century Drought. I will draw upon theother papers in this issue to further explorethe nature of the current drought. I willexamine the possibility that arid conditionswill persist and intensify due to climatewarming and consider some of the sus-tainability challenges and solutions relatedto an arid 21st century.

Geography and Trajectory of theSouthwest Sustainability ChallengeThe spatial and temporal contexts of theEarly 21st-Century Drought can be clearly

demarcated relative to the climate of thelast century (18952000 mean values fromref. 2). From 2001 through 2009, manyregions of the conterminous United Statesexperienced elevated annual temperatures(Fig. 1A), but temperatures in the South-west have been exceptionally high (>1to >2 SD above 20th-century means). Thedifference in annual precipitation betweenthe early 21st century and the 20th centuryshows a strong geographic contrast be-tween West and East. Many areas ofeastern North America experienced pre-cipitation >0.15 SD above the 18952000mean. In contrast, much of the Westexperienced lower than average pre-cipitation (Fig. 1B). The net result of theenhanced temperatures and decreasedprecipitation has been the developmentof persistent aridity (measured in termsof the Palmer Drought Severity Index;PDSI) in the Southwest and adjacent in-termountain Westincluding the head-waters of the Colorado River (Fig. 1C).Although much of the conterminousUnited States experienced increased tem-peratures in the early 21st century, weare a nation divided in terms of changesin precipitation and resulting waterresource challenges.Annual values and 5-y running means

for temperature (Fig. 2A) indicate thatin the late 20th and early 21st centuries theSouthwest has experienced an unpre-cedented period of sustained high tem-

Author contributions: G.M.M. designed research, per-formed research, analyzed data, and wrote the paper.

The author declares no conict of interest.

This article is a PNAS Direct Submission.

E-mail: [email protected].

2125621262 | PNAS | December 14, 2010 | vol. 107 | no. 50 www.pnas.org/cgi/doi/10.1073/pnas.0909651107

peratures relative to the 20th century.There has been a general, but episodic,decline in regional precipitation (Fig. 2B).During the 21st century, the regional PDSIfor the Southwest reached its lowest levelduring the period of record (Fig. 2C). Itis the high temperatures, rather thanunprecedentedly low precipitation, thatappear largely responsible for the ex-ceptionally low regional PDSI values(Fig. 2 AC). The Colorado River has alsoexperienced the lowest 5-y mean ows onrecord (Fig. 2D). Other periods of region-wide aridity and coincidental declines inColorado ow have occurred over the 20thcentury (19001904, 19241936, 19531964, and 19881991). These perfectdroughts of widespread persistent aridityhave also been associated with warmerregional temperatures (Fig. 2 AC).However, the amount of warmingduring the Early 21st-Century Droughtis exceptional.Although meteoric and extraregional

supplies of water may have diminished, thehuman demand for water remains consid-erable. Over the 20th century, the pop-ulation of the Southwest has increasedfrom about 2,100,000 to over 50,000,000people (7) (Fig. 2E). More than 36 millionof those people live in California. Initially,the amount of irrigated acreage increasedin tandem with population and reachedover 4.8 million ha (12 million acres) inthe 1970s (Fig. 2F) (8, 9). The vast ma-jority of that land is in California. Sincethen, there have been a attening anddecreases in irrigated farm acreage (Fig.2F). Factors at play include the full de-velopment of most practically farmablelands by the 1970s, fallowing of land dur-ing the 19871991 drought and the con-version of some farms to suburbs andcities. Between 1990 and 2004, more than200,000 ha (500,000 acres) of Californiafarmland were converted to urban andsuburban land uses (10). This trend iswidespread, with the conversion of809,000 ha (1,999,082 acres) in the sevenstates of the Colorado River Basin be-tween 1997 and 2007 (8).To support the growing population,

water withdrawals for domestic use in theSouthwest increased to over 12,334 millionm3 (10 MAF) annually (4) and continueon an upward trajectory (Fig. 3C). How-ever, the largest use of water is for agri-culture. Industrial uses are relativelynegligible in comparison with agricultureand have declined in recent decades (Fig.3C). Roughly 80% of all water withdrawalsare used for agricultural purposes. Agri-cultural water use in the Southwest rose toover 700,000 million m3 by the 1970s andthen attened and declined. This is con-temporaneous with, but not wholly attrib-utable to, the accelerated withdrawalof irrigated farm lands for other uses

Fig. 1. (A) Composite standardized temperature anomalies for 20012009 relative to 18952000. (B)Composite standardized precipitation anomalies for 20012009 relative to 18952000. (C) Mean PDSIvalues for the period 20012009. Data are from ref. 2 and mapped by state climate divisions.

MacDonald PNAS | December 14, 2010 | vol. 107 | no. 50 | 21257

SPECIA

LFEATURE:PERSP

ECTIV

E

(Fig. 3C). There were also declines inwater use during the 19871991 drought.Agricultural water use stood at about61,859 million m3 (50 MAF) by the endof the 20th century. Although domesticuse has steadily increased, declines in ag-

ricultural and industrial withdrawals pro-duced a decrease in overall water use inthe 1980s followed by a gradual increaseover the 1990s in which increasing do-mestic consumption has played a signi-cant (33%) role (Fig. 3C).

The net result of increasing population,agriculture, and industry over the 20thcentury is water use in the Southwest es-timated to have totaled 77,425 million m3

(62.7 MAF) in 2000 (4). This is a declinefrom a peak of 88,218 million m3 in 1980.However, through this period, net do-mestic consumption continued to rise.

Some Sustainability ChallengesIs the increasing aridity in the Southwestcapable of posing signicant challenges tosocioeconomic and environmental sus-tainability as we move further into the 21stcentury? The paper by Sabo et al. (11)tackles the current water sustainabilitychallenges in the broader West by focusingupon the concerns raised by the late MarcReisner in his book Cadillac DesertTheAmerican West and Its Disappearing Water.Sabo et al. calculate that humans nowappropriate the equivalent of 76% of theWests streamow for agriculture, domes-tic use, and other purposes.It is not anticipated that population

growth in the Southwest will abate over thelong term. The US Census estimates thatby 2030 over 67 million people will live inthe region (12). California would add thegreatest number and reach a populationsize of over 46 million. Nevada, Arizona,and Utah would be among the top 5 statesin the nation in terms of percentage ofpopulation increase. Arizona is projectedto add over 5 million people to becomeone of the 10 most populous states in theUnited States. Not only are populationsincreasing but the geographic distributionof the population is changing in an im-portant fashion. Since 1950, there hasbeen a strong increase in the proportionalgrowth of suburban populations. In 2000,suburbanites accounted for 50% of thepopulation (7). Southwestern suburbandevelopments, in which 70% or more ofthe water is often used for landscaping(13), amplify the water demands exertedby the increasing population. Sabo et al.estimate that per-capita virtual waterfootprints are seven times higher for citiesin the arid West than in the East. Theysuggest that with a doubling of population,the West would require the equivalentof more than 86% of its total stream-ow to meet human use at currentper-capita levels.Agriculture remains an important sector

of the Southwests economy. Californiasfarm receipts totaled $36.1 billion in 2008(14). Aside from a fundamental role indomestic consumption and food security,its exports contributed some $13.6 billion(14) to the nations international tradebalance. Changes in the agricultural pro-ductivity of the Southwest in response towater shortages and/or reallocation willhave direct implications for food supplyand security. Aside from the negative

A

B

C

F

E

D

Year

Millionm

3o C

cmPD

SIMillionPe

ople

MillionHectares

Fig. 2. (A) Southwest (California, Nevada, Arizona, New Mexico, Utah) average annual temperature (2).(B) Southwest average annual precipitation (2). (C) Southwest average annual Palmer Drought SeverityIndex (2). (D) Naturalized discharge of the Colorado River at Lees Ferry, AZ (3). (E) Southwest populationsize (7). (F) Southwest irrigated agricultural land area (8, 9).

21258 | www.pnas.org/cgi/doi/10.1073/pnas.0909651107 MacDonald

impacts of discrete droughts, chronic saltaccumulation in soils promoted by hot andarid climate can also produce agriculturallosses. In the West today these losses arealready on the order of $2.5 billon/y (11).Avoiding salt accumulation places addi-tional restrictions on agricultural watermanagement in the Southwest.The reservoir system on the Colorado

River is one of the most important buffersagainst drought in the Southwest. Althoughsignicant loss of reservoir capacity due tosedimentation may not be imminent (11),water supply and demand challenges forthe reservoir system are clearly acute to-day. The Colorado system has seen storagelevels decline precipitously, and they stoodat 40,766 million m3 (33.05 MAF) or55.6% of capacity as of October 1, 2010(3). The level of Lake Mead has nowfallen more than 40 m below capacitylevel. A further decline of a few meterswill trigger a level 1 water shortage dec-laration. At the extreme end of the spec-trum, a recent study suggests that LakeMead and Lake Powell have a 50% chanceof receding to inoperable status by the2020s (15). Loss of reservoir storage alsoproduces loss of hydroelectric productionand decreases energy supplies.The recent drought has prompted

emergency restrictions on outdoor wateruse by residents in cities such as Las Vegas,

Tucson, Albuquerque, Los Angeles, andSan Diego. The Metropolitan Water Dis-trict of Southern California serves 17million people and in April 2009 voted tocut deliveries in its largely urban andsuburban service area by 10%. Althoughurban water restrictions may be incon-venient, drought conditions have an ap-preciable nancial impact on agriculture.In 2008, California alone experienced atleast $308 million in lost agricultural rev-enue due to drought (16).The increasing temperatures and aridity

of the early 21st century also pose chal-lenges for wildlands and land management.For example, experimental studies havefound that a 4 C warming produces a 30%increase in pion pine mortality amongdrought-stressed trees (17). Earlier springwarming and decreased surface water ap-pear to contribute to a recent increase inres (18). The annual cost of wildland resuppression in California alone now typi-cally exceeds $200 million (5). In 2007,over 3,000 structures were destroyed andtotal suppression costs plus damages wasalmost $780 million (5). Total costs ofbark beetle damage are difcult to calcu-late, but during the 5-y period of 20052009, over $75 million of federal, state,and local funds were spent on prevention,suppression, and restoration. That pro-duced treatment of only about 200,000 ha

(500,000 acres) throughout the Westa fraction of the more than 8 millionhectares (22 million acres) of forest andwoodland area under threat (19).

The 21st CenturyThe remaining papers in this issue look atthe Early 21st-Century Drought and theremainder of the century. The studiestackle various aspects of water sustain-ability with an array of approaches thatinclude analysis of current meteorological,socioeconomic, and ecological data, pale-oenvironmental data analysis, model sim-ulations, and policy analysis. From thepapers presented here, several importantinsights emergeoften possessing partic-ular weight because they arise from morethan one research approach. These in-sights can be organized around fourcritical questions.

1. Is the Early 21st-Century Drought ExceptionalCompared with Earlier Droughts and Is ThisAttributable to Increasing Greenhouse Gasses?Cayan et al. (20) examine the Early 21st-Century Drought relative to historicaldroughts of the 20th century. They concludethat for the Colorado River Basin, the Early21st-Century Drought has been the mostextreme in over a century, and might occurin any given century with a probability ofabout 60%. They point out that 3 of the 11driest years experienced over the past 100 yhave occurred in the past decade (2002,2007, and 2008). Only the 1930s experi-enced a comparable run of dry years. Sim-ilarly, Seager and Vecchi (21) conclude thatthe Southwest has been experiencing a gen-eral drought that is at least as severe as anyin the past 100 y. They also note that thedrought appears to be part of a longer-termtrend of strong drying that began around1979 (Figs. 2 and 3). Woodhouse et al. (22)use paleohydrological reconstructionsto show that although the 21st-CenturyDrought is severe by standards of the past100200 y, it pales in terms of spatial ex-tent and duration compared with the pre-historic drought of the 12th century. As badas things might seem, they have the dem-onstrated potential to become worse.Both Cayan et al. and Woodhouse et al.

point out that warmer temperaturesare typically associated with prolongeddroughts in the Southwest. Cayan et al. ndthat summer temperature anomalies dur-ing past Southwest droughts have rangedfrom +0.5 C to +1 C. Similar to thepresent drought, this warming in theSouthwest occurred in concert with wide-spread warming over the conterminousUnited States. Although the currentdrought is consistent with the observedrelationship between extreme droughtsand high temperatures, the magnitude andprolonged nature of the high temperaturesof the Early 21st-Century Drought have no

Year

o CIS

DPsno illi

Mm

3

B

A

C

Fig. 3. (A) Southwest average annual temperature (2). (B) Southwest average annual PDSI (2). (C)Southwest water withdrawals by usage sector (4).


analog in the 20th century (Fig. 2A).Woodhouse et al. use paleoclimatic re-cords to show that the current warmingin the Southwest may exceed any otherwarming episode experienced over thepast 1,200 y. Cayan et al. and Seager andVecchi also note the inuence of warmtemperatures in the impact of the currentdrought on decreased snowpack, earliertiming of snowmelt, and greater evapora-tion rates and transpiration demands. Thecurrent drought is therefore exceptional interms of the magnitude of warming andadditional evapotranspiration stresses.The studies by Cayan et al. and Seager

andVecchi suggest that the recent warmingis consistent with the IntergovernmentalPanel on Climate Change AssessmentReport 4 (IPCC AR4) projections of an-thropogenic climate change (23). How-ever, both studies conclude that it is notpossible to denitively attribute the Early21st-Century Drought to increased green-house gasses. Cayan et al. conclude thatthe Early 21st-Century Drought, althoughsevere, is not outside the realm of naturalvariability in the Southwest. Seager andVecchi argue that the great North Amer-ican droughts of the past 200 y werecaused by very small sea surface temper-ature (SST) anomalies. They note thatthere has been a general cooling trend inthe eastern Pacic following 1979 and thatsuch cooling typically is associated withdrought in the Southwest. The drivers ofsuch SST anomalies remain poorly un-derstood, as does the potential impact ofincreasing greenhouse gasses on PacicSSTs. Seager and Vecchi conclude that thegeneral drying in recent decades and the21st-Century Drought could be a result ofnatural decadal variability in Pacic SSTs.

2. Is It Likely That the Southwest Will Experi-ence a More Arid Climate Due to Global Cli-mate Change Driven by Increasing GreenhouseGasses? The climate model estimates ana-lyzed by Cayan et al. and Seager and Vecchiall indicate that continued warming couldproduce increased aridity, overprinted bymore severe droughts. Analysis of the resultsof 15 and 24 different general circulationmodels lead Seager and Vecchi to argue thatincreasing aridity in the Southwest would bean expected outcome that results froma poleward expansion of the subtropical dryzones as the planet warms. Southwest dryingis mainly being driven by a decline in winterprecipitation associated with increasedmoisture divergence due to changes in meanatmospheric ow and reduced moistureconvergence via transient eddies. The dryingof the Southwest and similar subtropicalregions is a highly robust result from themodel simulations. Seager and Vecchi an-ticipate that anthropogenic aridity will be aslarge in magnitude as the droughts caused bynatural decadal variability in climate by

around 2050. They also conclude that it isunlikely that the Southwest will see a returnof any prolonged periods of moist conditionssimilar to the long wet spells experiencedin the 20th century. The analysis by Cayanet al. similarly indicates that the Southwest islikely to become drier and experiencemore severe droughts than witnessed overthe 20th century. Drought activity is likely toincrease toward the end of the 21st century,particularly in the Colorado River Basin.Drought episodes typied by continuoussoil moisture depletion will increase from410 y to periods of 12 y or more.The paleohydrological analysis of

Woodhouse et al. provides evidence insupport of the potential for prolongedaridity and greater droughts if warm tem-peratures persist in the 21st century. Thedriest and most widespread interval ofdrought documented in the paleorecordsoccurred in the mid-12th century and iscoincidental with the period of greatestprolonged temperature increase. The 12thcentury, typied by warm temperaturesproduced by increased insolation, de-creased volcanic activity, a coincidentalcooling of the eastern Pacic Ocean, andwidespread, prolonged, and intensedrought in southwestern North America,has been used as a comparison for thecurrent drought (24). During the decade of11461155, the ow of the Colorado Riveraveraged about 78% of its 20th-centurymean. The portion of the Southwest ex-periencing drought in any given year av-eraged 65.5% of the total land area. Thepaleorecords also show a general consis-tency between warmer temperatures andprolonged drought, and indicate that theobserved droughts of the 20th century donot capture the full potential severity andduration of droughts exacerbated by warmconditions. Even in the absence of man-made climate change, the region is proneto periods of prolonged warming and ex-ceptional drought that should be consid-ered in planning efforts for a sustainableSouthwest.

3. What Are the Potential Impacts of IncreasingAridity on Wildland and Urban/Suburban Sys-tems? Williams et al. (25) examine correla-tions between climate and the radial growthof trees across North America. They showthat conifer trees in the Southwest are par-ticularly sensitive to temperature and aridityrelative to other regions. They use climatetree growth relations calculated for the past100 y, combined with IPCC climate modelestimates for the 21st century, to predict thelikely fate of important Southwest tree spe-cies such as pion pine (Pinus edulis), pon-derosa pine (Pinus ponderosa), and Douglasr (Pseudotsuga menziesii). Williams et al.conclude that woodlands and forests willexperience substantially reduced growth

rates and increased mortality at manySouthwest sites as the century progresses.Based on analysis of satellite data and

aerial photographs, Williams et al. dem-onstrate that Southwest forests andwoodlands have been experiencing signif-icant impacts from wildres and barkbeetles in recent decades. Climate warmingand drought promote forest ammabil-ity and can increase lightning ignition.Southwest forests are hosts to three im-portant species of bark beetlesprucebeetle (Dendroctonus rupennis), pinebeetle (Dendroctonus ponderosae), andpion ips beetle (Ips spp.), the last beingthe most widespread in the region (26).Climate warming allows for greater beetlereproduction and expansion of beetlesranges to higher and cooler elevations.Drought weakens the resistance of trees tobeetle infestations and promotes greatersusceptibility and mortality when in-festations occur. Dying trees can increaseforest fuel loads and promote res. Wil-liams et al. estimate that from 1984 to2006 some 2.73.0% (6,420 km2) of thetotal area of southwestern forest andwoodland has experienced mortality dueto stand-clearing wildres. A staggering7.611.3% (18,177 km2) of woodland andforest has experienced mortality due tobark beetles between 1997 and 2008. Whatis most disturbing is the high rate of forestloss. They estimate that 1418% of theSouthwests forests have been impacted byres or bark beetles in the period 19842008. There has also been a steady rise inthe annual area burned by severe res. It islikely that bark-beetle- and re-relatedmortality will increase should 21st-centuryclimate warming continue, and this willpose a signicant challenge for conserva-tion and resource management acrossthe Southwest.Gober and Kirkwood (27) look at

Phoenix as an example of the water chal-lenges facing cities in the Southwest.Phoenix displays and amplies many of theattributes typical of the Southwest in-cluding increasing population, large sub-urban development, limited water supply,and shifting agricultural to urban land andwater use. They use the WaterSim modelto simulate conditions in the year 2030.Future climate scenarios from IPCC AR4were used to develop a range of scenariosfor the ows of the Salt/Verde and Colo-rado River systems. These two systemssupply much of the citys water. Ground-water conditions were also estimated. Thewater demand estimates were based uponextrapolated population-size and land-useprojections. Many of the scenarios in-dicated that achieving sustainability wouldrequire decreases in per-capita urban wa-ter use to slightly less than current indooruse. This suggests a dramatic curtailmentof almost all usage for landscaping and


pools in Phoenix. Even restricting pop-ulation growth by 50% would not allowcurrent per-capita water usage to be sus-tained under many water-supply scenarios.Increased groundwater reliance does noteffectively mitigate the concerns. Underworse-case simulations, groundwaterdrawdowns range from 6 billion to 14 bil-lion m3. Gober and Kirkwood concludethat policy action to limit groundwater usewill be necessary even without climatechange to contend with. Limiting growthto 50% of projected levels and eliminatingmost irrigated outdoor landscaping andprivate backyard pools may be needed toachieve groundwater sustainability evenunder normal river ow conditions. Thesimulations suggest that with or withoutclimate change, the Phoenix area facesclear sustainability challenges in theopening decades of the 21st century.

4. What Policy Prescriptions and OtherStrategies Might Help Us to Develop Water-Use Sustainability in the Southwest? Watersustainability can be maintained throughtwo basic variables: (i) increased supply or(ii) decreased demand. As is pointed outby Gleick (28), the preferred response towater challenges over the 20th century wasbased on engineered solutions on thesupply side: Build large-scale, centralized,federally subsidized infrastructure to movewater in both space and time to meetcurrent and projected demands. Suchdependence upon extralocal water andengineering approaches predate the pastcentury (29). Archeological evidence andearly historical accounts tell us that peo-ples such as the Hopi, Zuni, Rio GrandePueblo, and Pecos Pueblo built large vil-lages and practiced irrigated agriculturealong rivers including the Little Colorado,the Rio Grande, and the Pecos. Indeed,native peoples engineered small checkdams and irrigation canals beginningabout 2,000 y ago. In the 18th century,Spanish missions and settlements in Cal-ifornia were typically established near riv-ers and developed masonry dams >3 mthick and stone aqueducts that ran formore than 10 km. In the 20th century,large infrastructure projects such as theHoover Dam, which incorporates2,600,000 m3 of concrete, and the Cal-ifornia Aqueduct, which is 1,151 km long,were built. After 2,000 y of application,there is certainly still a role for additionalstorage and transference capability; how-ever, the engineering of water reservoirand transference systems as a comprehen-sive solution to Southwest water sustain-ability has run its course. Some of thelimiting factors include the huge size ofthe current population, the importanceand water demands of its agriculture, thelimitations of meteoric and groundwatersupply, the potential for decadal-length

drought, and the challenges of globalwarming in terms of decreased pre-cipitation and increased evapotranspira-tion. To these limitations should be addedenvironmental concerns over the preser-vation of the ecosystems and species inplaces such as the Sacramento Delta, theriparian systems along the Colorado River,and the waterfowl habitat of the SaltonSea. In cases such as the SacramentoDelta, transference has already been sig-nicantly curtailed due to environmentalconcerns and resulting judicial restrictions.If the now-desiccated Colorado RiverDelta lay in the United States rather thana few kilometers over the border in Mex-ico, similar environmental concerns wouldlikely be placing additional constraints onthe usage of Colorado River water.Aquifer drawdown and saltwater in-

trusion limit further extraction of ground-water (27, 29). Enhanced water harvesting,particularly stormwater capture, can aug-ment supplies. In California, the Storm-water Resource Planning Act allowsmunicipalities to access funds for projectsthat capture stormwater for reuse or torecharge groundwater. The City of LosAngeles estimates that during rainy daysas much as 37,854,117 m3 (10 billiongallons) may ow through the stormwatersystem. However, such water still requiresconsiderable treatment depending uponintended use (30). Sabo et al. call for in-creased urban desalination plants. Large-scale desalination, although technicallyfeasible, requires signicant energy andremains expensive (31). Treating brackishwater is less expensive, and Gleick outlineshow desalination of brackish groundwateris signicantly augmenting municipal wa-ter supplies in El Paso. Improvements intechnology and particularly the use of so-lar energy could help offset the energy andcost restrictions of desalination (32). Graywater recycling and use in landscaping isalready being applied (33). This holdsmuch promise given the prominent roleof suburban lawns and gardens in South-west water demand. Potable reuse ofsome recycled wastewater is possible, butfaces economic and community accep-tance challenges.Despite the innovations outlined above,

increased supply will likely not provide thecomplete answer for the Southwest in thenear future. Although there remains rela-tively greater uncertainties in projectionsof precipitation than temperature, theconsensus is that global climate change dueto increased greenhouse gasses will exac-erbate surface water deciencies in theSouthwest (20, 21, 23). Much of the in-crease in demand is and will be driven bycities and their suburbs. Water could betransferred from farms to maintain urbangrowth. For example, a recent modelingstudy by Tanaka et al. (34) concludes

Californias water system can adapt to thefairly severe representations of populationgrowth and climate warming. This adap-tation will be costly in absolute terms andinclude transaction, institutional, and xedcosts not quantied in the model, but, ifproperly managed, should not threaten thefundamental prosperity of Californiaseconomy or society, although it can havemajor effects on the agricultural and en-vironmental sectors. As discussed above,environmental concerns are already cur-tailing water transference and it is unlikelysuch policies will be signicantly reversed.However, signicant transfers of waterfrom agriculture to satisfy the domesticdemands of a growing suburban pop-ulation also raise a plethora of importantconcerns including loss of agriculturalsector sustainability, rural socioeconomicdecline, increased food prices, decreasedfood choice, decreased food security, in-creased carbon footprint for importedfood, and decreased foreign trade balance.Thus, as pointed out by Gleick, Gober andKirkwood, and Sabo et al., innovationsand policies that decrease overall demandmust gure prominently in planning forwater sustainability. Sabo et al. estimatethat to completely eliminate freshwaterstress would require decreased water useto an appropriation of only 60% of thetotal streamow in the region. They arguefor a compromise target of a 15% de-crease. Gleick suggests that increased ef-ciencies are to be found in domestic andindustrial water uses. He notes that some50% of agricultural water use in Californiais for inefcient ood-irrigation. Saboet al. also suggest greater water-use ef-ciencies can be implemented in the agri-cultural sector. Gober and Kirkwoodarticulate a three-pronged strategy of im-plementing policies in urban areas thatwill slow population growth, focus re-maining growth in high-density develop-ments, and alter outdoor consumption byencouraging xerophytic landscaping anddecreasing private swimming pools. It isencouraging that even more modest policyprescriptions, such as public informationcampaigns, water-efcient building re-quirements, and limited restrictions, canhave signicant results. Water deliveriesby the Metropolitan Water District ofSouthern California peaked in 1990 atabout 3,207 million m3 (2.6 MAF) andby 2008 had fallen to about 2,466 millionm3 (2 MAF), despite a population increaseof 2 million people. In response todrought, the City of Los Angeles was ableto reduce total water usage by 17% overthe 1-y period of 20082009. Sabo et al.suggest that market-based pricing of waterand the restriction of government sub-sidies to only those uses that fulll basichuman needs should also be used. In a re-gion where the majority of water use is


often for exterior landscaping, decreasingper-capita demand does not have to meanfundamental hardships in terms of drink-ing water and cleanliness. Efcienciesclearly remain to be realized in the South-west in urban and suburban water use.For example, per-capita water use in Tuc-son is half that in Phoenix despite similar-ities in climate for the two cities (27).In viewof thebroad scopeof theproblem,

Gleick and Sabo et al. highlight the need forcomprehensive, multisectoral, and trans-regional policies to formulate water strat-egies forSouthwest sustainability.AsGleickdemonstrates, these efforts must fostercommunication, planning, and imple-mentation among a plethora of agenciesand jurisdictions. In addition, as the cli-mate models and paleoclimatic studies in-dicate, the region could become more aridand droughts could extend over decades.Typical 3- to 5-y drought plans are in-sufcient to address climate change anddecadal-to-multidecadal droughts.Discussion of sustainability must also

incorporate consideration of ecosystemservices and protection of endangeredspecies. Williams et al. point out that the

vegetation of the Southwest is likely un-dergoing profound changes. Managementof forests, woodlands, streams, deltas, andother habitat to preserve ecosystem func-tioning and conserve biodiversity will beextremely challenging and at times come atan appreciable cost in terms of water-supply options for other demands. Saboet al. point out the threats posed to nativesh species should care not be taken inwater-infrastructure projects.Cooperation and strategic integration

that surpass 20th-century efforts in terms ofgeographic scope, jurisdictional breadth,multisectoral engagement, and planningtimelines are required to develop South-west water sustainability. Given theimpacts of the current drought on watersupplies and infrastructure, those effortsshould be undertaken with expediency.However, with greenhouse gas concen-

trations at their current levels, we likely willnot escape signicant warming and result-ing increased aridity over the 21st century(20, 21, 23). Coupled with the demographicprojections, the climatic estimates for thenext decades compel us to develop waterresource strategies that adapt to these

changing conditions and promote sustain-ability in the face of increasing generalaridity as well as more severe episodicdroughts. Finally, the proximal economiccosts of reducing greenhouse gas emissionsare often cited as a rationale for inactionon emissions reduction. Because climatewarming will exacerbate water sustainabil-ity problems, the Southwest is likely toexperience some of the highest economicexpenses and environmental lossesrelated to climate change. As the papersin this issue illustrate, the ultimate costs ofinaction in curbing greenhouse gas emis-sions will be particularly high forthe Southwest.

ACKNOWLEDGMENTS. I thank B. L. Turner, Jr. forthe opportunity to prepare this special issue. I alsothank him, David Stopak, and Josiah Armour fortheir help and advice along the way. I thank all theauthors for their thoughtful papers and the ref-erees for their careful reviews. Recent research onthe topic of Southwest drought and the formula-tion of some of the ideas used in this paper and inorganizing this special issue was supported byNational Science Foundation grants and a 20082009 Guggenheim Fellowship.

1. National Drought Mitigation Center US Drought Moni-tor. http://www.drought.unl.edu/monitor/monitor.htm.

2. National Oceanic and Atmospheric Administration,Earth System Research Laboratory, Physical SciencesDivision http://www.esrl.noaa.gov/psd/data/usclimdivs.

3. US Bureau of Reclamation Upper Colorado Region.http://www.usbr.gov/uc/water.

4. Konieczki AD, Heilman JA (2004) Water-use trends inthe desert Southwest19502000. US GeologicalSurvey Scientic Investigations Report 20045148.

5. California Department of Forestry and Fire Protection(CAL FIRE) http://www.re.ca.gov.

6. Bentz B (2008) Western US bark beetles and climatechange. US Department of Agriculture, Forest Service,Climate Change Resource Center. http://www.fs.fed.us/ccrc/topics/bark-beetles.shtml.

7. Hobbs F, Stoops N (2002) Demographic trends in the20th century. Census 2000 Special Reports, US CensusBureau. http://www.census.gov/prod/2002pubs/censr-4.pdf.

8. US Department of Agriculture The Census of Agricul-ture. http://www.agcensus.usda.gov.

9. US Census Bureau (1953) Irrigation of agriculture lands:The United States. http://www2.census.gov/prod2/de-cennial/documents/38921062v3_TOC.pdf.

10. Thompson E., Jr (2007) Paving Paradise: A NewPerspective on California Farmland Conversion (AmericanFarmland Trust, Davis, CA).

11. Sabo JL, et al. (2010) Reclaiming freshwater sustain-ability in the Cadillac Desert. Proc Natl Acad Sci USA107:2126321270.

12. US Census Bureau (2005) US Population Projections.http://www.census.gov/population/www/projections/stproj.html.

13. Mayer PW, DeOreo WB (1999) Residential End Uses ofWater (American Water Works Association ResearchFoundation, Denver).

14. US Department of Agriculture (2009) Trade and agri-culture: Whats at stake for California? http://www.fas.usda.gov/info/factsheets/wto/states/ca.pdf.

15. Barnett TP, Pierce DW (2008) When will Lake Mead godry? Water Resour Res 44:W03201.

16. California Department of Food and Agriculture (2009)Drought tops agenda for State Board of Food andAgriculture. http://www.cdfa.ca.gov/egov/Press_Releases/Press_Release.asp?PRnum=09-009.

17. Adams HD, et al. (2009) Temperature sensitivity ofdrought-induced tree mortality portends increasedregional die-off under global change-type drought.Proc Nat Acad Sci USA 106:70637066.

18. Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW(2006) Warming and earlier spring increase western USforest wildre activity. Science 313:940943.

19. Western Forestry Leadership Coalition (2009) Westernbark beetle assessment: A framework for cooperativeforest stewardship. 2009 update. http://www.wccen-ter.org/news_pdf/325_pdf.pdf.

20. Cayan DR, et al. (2010) Future dryness in the south-west US and the hydrology of the early 21st centurydrought. Proc Natl Acad Sci USA 107:2127121276.

21. Seager R, Vecchi GA (2010) Greenhouse warmingand the 21st century hydroclimate of southwesternNorth America. Proc Natl Acad Sci USA 107:2127721282.

22. Woodhouse CA, Meko DM, MacDonald GM, Stahle DW,Cook ER (2010) A 1,200-year perspective of 21st centurydrought in southwestern North America. Proc Natl AcadSci USA 107:2128321288.

23. Intergovernmental Panel on Climate Change (2007)Climate Change 2007: The Physical Science Basis. Con-tribution of Working Group I to the Fourth AssessmentReport of the Intergovernmental Panel on ClimateChange, eds Solomon S, et al. (Cambridge Univ Press,Cambridge, UK and New York).

24. MacDonald GM, et al. (2008) Climate warming and21st-century drought in southwestern North America.Eos Trans AGU 89:82.

25. Williams AP, et al. (2010) Forest responses to increasingaridity and warmth in the southwestern United States.Proc Natl Acad Sci USA 107:2128921294.

26. Raffa KF, et al. (2008) Cross-scale drivers of naturaldisturbances prone to anthropogenic amplication:The dynamics of bark beetle eruptions. Bioscience 58:501517.

27. Gober PA, Kirkwood CW (2010) Vulnerability assessmentof climate-induced water shortage in Phoenix. Proc NatlAcad Sci USA 107:2129521299.

28. Gleick P (2010) Roadmap for sustainable waterresources in southwestern North America. Proc NatlAcad Sci USA 107:2130021305.

29. MacDonald GM (2007) Severe and sustained droughtin southern California and the West: Presentconditions and insights from the past on causes andimpacts. Quaternary Int 173-174:87100.

30. Begum S, Rasul MG, Brown RJ (2008) A comparativereview of stormwater treatment and reuse techniqueswith a new approach: Green Gully. Wseas TransEnviron Dev 4:10021013.

31. Karagiannis IC, Soldatos PG (2008) Water desalinationcost literature: Review and assessment. Desalination223:448456.

32. El-Sayed YM (2007) The rising potential of competitivesolar desalination. Desalination 216:314324.

33. Exall K, Jimenez B, Schaefer K, Marsalek J (2008) Waterreuse in the United States and Canada. Water Reuse:An International Survey of Current Practice, Issues andNeeds, eds Jimenez B, Asano T (International WaterAssociation, London), pp 68104.

34. Tanaka SK, et al. (2006) Climate warming and watermanagement adaptation for California. Clim Change76:361387.


Documents

PNAS-2010-MacDonald-21256-62