integrated economic, hydrologic, and institutional ...water-research.nmsu.edu › files › 2016 ›...

Integrated Economic, Hydrologic, and Institutional Analysisof Policy Responses to Mitigate Drought Impacts

in Rio Grande BasinFrank A. Ward1; James F. Booker2; and Ari M. Michelsen3

Abstract: In the Rio Grande Basin of North America, water is overappropriated and demand for water grows while supplies areconstrained by drought and climate change. The Basin is currently in its seventh year of drought, and reservoirs are at historically lowlevels. Thus agricultural and municipal river diversions have been sharply curtailed, and low flows threaten endangered species. A centralpolicy challenge is the design and implementation of plans that efficiently and fairly allocate the Basin’s water supplies. Such plans arecomplicated by the demands of existing water users, potential new users, three state governments, and two sovereign nations. To addressthese issues, an integrated basinwide nonlinear programming model was designed and constructed for the purpose of optimizing waterallocations and use levels for the Basin. The model tests whether institutional adjustments can limit damages caused by drought andidentifies changes in water uses and allocations that result from those adjustments. Compared to existing rules governing the riversystem’s water use, future drought damages could be reduced by one-fifth to one-third per year from intrastate and interstate watermarkets, respectively, that permit water transfers across jurisdictions. Results show hydrologic and economic trade-offs among water uses,regions, and drought control programs.

DOI: 10.1061/�ASCE�0733-9496�2006�132:6�488�

CE Database subject headings: Integrated systems; Hydrology; Economic factors; Droughts; Colorado; Texas; Water policy.

Background

A global challenge for 21st century water management is to sat-isfy growing human demands for water while protecting theaquatic ecosystems upon which economies and life depend. Es-pecially in dry places with rapidly growing populations andeconomies, this task is daunting. In the western United Statesmost water supply sources are fully claimed and used largely forirrigation, growing cities, and emerging industrial demands.Where water is unappropriated, it is protected for in-stream flows,endangered species, and environmental uses. Most economicallyaccessible groundwater is developed and connected to fully ap-propriated surface water. Throughout the West, drought and cli-mate change increase the growing competition for water suppliesand increase the importance of designing effective water alloca-tion laws and policies.

The upper Rio Grande Basin �Fig. 1� confronts challenges

1Professor, Dept. of Agricultural Economics and Business, NewMexico State Univ., Las Cruces, NM 88003. E-mail: [email protected]

2Associate Professor of Economics and Environmental StudiesProgram Director, Siena College, Loudonville, NY 12211-1462. E-mail:[email protected]

3Professor and Resident Director, Texas A&M Univ., El PasoAgricultural Research Center, El Paso, TX 79927-5020. E-mail:[email protected]

Note. Discussion open until April 1, 2007. Separate discussions mustbe submitted for individual papers. To extend the closing date by onemonth, a written request must be filed with the ASCE Managing Editor.The manuscript for this paper was submitted for review and possiblepublication on December 9, 2004; approved on May 4, 2005. This paperis part of the Journal of Water Resources Planning and Management,Vol. 132, No. 6, November 1, 2006. ©ASCE, ISSN 0733-9496/2006/6-
488–502/$25.00.
488 / JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT

faced by many of the world’s rivers that support economies andcultures in dry places. The upper basin extends from southernColorado to Fort Quitman, Texas. Hereafter we refer to that areaas the Rio Grande Basin, or simply the Basin. Previous researchhas described policy challenges in the Sacramento and Coloradoriver basins of the United States. These include climate changescenarios analyzed by Christensen et al. �2004�, analysis of theCentral Arizona Project by Holland and Moore �2003�, sensitivityof a basin’s outcomes to changes in policy objectives by Mah-moud and Garcia �2000�, and feasibility of water markets byNewlin et al. �2002�.

Challenges faced by the Yangtze and North China Plain,China, also have been examined. They include a comprehensiveintegrated hydrologic, economic, agricultural, environmental, andsocial/institutional decision support system used for evaluatingwater resources and the impacts of competing policies and devel-opment objectives for the North China Plain, �Michelsen and Bar-gur 1993; United Nations 1994�; the connection between forestecosystem processes and economic benefits examined by Guoet al. �2000�, displacement and resettlement issues �Li et al.2001�, land use problems �Liu et al. 2003�, wetland ecologicalprocesses �Liu et al. 2004�, ecosystem functions �Nakamura2003�, and environmental restoration �Yan and Qian 2004�.

Policy analyses of the Jordan River in the Middle East alsohave received attention. These include political and military con-trol challenges in regions with shared transboundary watersources, as described by Abu Zahra �2001�, Haddadin �2002�,Jagerskog �2003�, Mimi and Sawalhi �2003�, and Shuval �2000�.Several analyses of the Murray-Darling Basin in Australia havebeen published that analyze such important issues as flood plainecosystem protection �Arthington and Pusey 2003�, improvedconnections between research and extension �Keogh et al. 2004�,
environmental improvements �Quiggin 2001�, and improved eco-
© ASCE / NOVEMBER/DECEMBER 2006

fward

Typewritten Text

For non-commercial and educational use only. Please ask publisher for use permission

fward

Typewritten Text

system health �Reid and Brooks 2000�. Finally, several scholarshave examined opportunities and choices facing the Nile Basin inNorth Africa. Examples include analyses of water conservationchoices by El-Kady and El Shibini �2001�, connections betweengroundwater and surface water by Farah et al. �2000�, salinity byKotb et al. �2000�, and adaptation to climate change by Strzepek�2000�.

In most river systems in the world’s arid regions, water isoverallocated, supplies are threatened by seasonal or recurrentdrought and the effects of climate change, and growing demandsfor water are placing stress on existing water management insti-tutions. These factors continue to raise the interest of policymak-ers, scientists, farmers, and water managers for evaluations ofexisting water institutions and examinations of alternative de-signs. One challenge in all these regions is the development ofsystem operation rules and policies that complement existingwater management institutions and allocate water so as to mini-mize total economic damages, and to do so in an open and trans-parent environment. Achieving these aims requires approaches todecision support that encompass hydrologic boundaries and po-litical and institutional jurisdictions, and also account for eco-nomic impacts over all important sectors.

This paper focuses on the impacts of transfers resulting fromthe institution of water markets as one measure for reducing theeconomic damage from drought. Market-based transfers havebeen suggested for many years by many authors as one measureto cope with scarce supplies and limited storage, especially duringdrought �e.g., Vaux and Howitt �1984��.

In previous research, integrated hydrologic-economic modelsat the basin scale have focused on the economic impacts of trans-

Fig. 1. Upper Rio Grande Basin map

fers under typical supply conditions �e.g., Oamek �1990��. Other

JOURNAL OF WATER RESOURCES PLANNING

work developed extensions to managing water quality �Lee et al.1993� and incorporated additional nonconsumptive use values�Booker and Young 1994�. Under drought conditions, the impactsof several market institutions were estimated by Booker �1995�.At the subbasin scale, integrated modeling of economic impactsof water transfers for protecting instream flows was developed byHamilton et al. �1999� and Willis and Whittlesey �1998�; impactsof water markets for protecting water quality were examined byWeinberg �1993�.

The Journal of Water Resources Planning and Managementhas published a number of papers dealing with modeling and/orpolicy analysis at the river basin scale. Bouchart and Goulter�1998� developed risk analysis methods for improved reservoirreleases for irrigation. Elarabawy et al. �1998� developed a bas-inwide model for dealing with shortages in Egypt. Ravikumar andVenugopal �1998� developed a stochastic dynamic programmingmodel to analyze optimal reservoir operation for agricultural usein India. Levy and Baecher �1999� developed a policy simulationmodel of the Nile Basin. Lund and Guzman �1999� derived oper-ating rules for reservoirs in series or in parallel for optimizingtrade-offs between hydropower, water quality, and water supply.Philbrick and Kitanidis �1999� compared the use of stochastic anddeterministic reservoir control and operations models. Rangarajanet al. �1999� developed a reservoir operations model that opti-mized Canadian energy production in the face of uncertain futureenergy demands.

Wagner �1999� developed a decision framework for assessingthe value of groundwater sampling for improving groundwaterpumping programs. Watkins and McKinney �1999� developed ascreening model to support the choice of alternatives for meetingfuture water demands and protecting sensitive ecosystems in theEdwards aquifer region of central Texas. DeAzevedo et al. �2000�integrated surface water quality and quantity objectives for deci-sion support for the Piracicaba River Basin in Sao Paulo, Brazil.Loaiciga and Leipnik �2000� developed an economic model thatmaximizes the expected value of net income accruing from thesale of groundwater subject to a number of geologic, climatic, andenvironmental requirements. Shabman and Stephenson �2000�surveyed the literature on measuring values of water in alternativeuses. Eschenback et al. �2001� described the potential uses,power, and flexibility of River Ware’s basin optimization softwarefor application by the Tennessee Valley Authority. Newlin et al.�2002� described the potential use of an economic-engineeringnetwork flow optimization model for evaluating market watertransfers in southern California. Barros et al. �2003� analyzed anoptimization model based on a monthly time step for the manage-ment and operations of a Brazilian hydropower system. Cai et al.�2003� described the development and use of an integratedhydrology-agronomic-economic model in a river basin in CentralAsia for which irrigation is the dominant use.

In the Rio Grande Basin of the southwestern United States,despite the region’s conditions and the extensive contributions ofprevious work, little comprehensive analysis of the costs ofdrought and its application to effective policy response has beenconducted to date. With growing demands placed on the Basin’swater resources, future droughts will result in larger economicand environmental impacts. For this reason, the design ofdrought-coping policies will take on increasing importance to thisregion and to other drought-prone basins of the world faced bygrowing competition for water. This paper’s objective is to take afirst step toward comprehensive analysis of drought and its impactby contrasting existing institutions that govern the allocation and
use of water during drought periods with alternative designs. This
AND MANAGEMENT © ASCE / NOVEMBER/DECEMBER 2006 / 489

objective is accomplished through the development and use of acomprehensive basinwide nonlinear programming model of thehydrology, economics, and institutions involved. The model isapplied to evaluate the hydrologic and economic effectiveness ofselected potential drought-coping policies. The approach couldoffer considerable potential for more rational institutional designin other drought-prone basins of the world.

One important advantage of the model is its ability to integratethe essential hydrologic, economic, and institutional componentsof a river basin and to explore both hydrologic and economicconsequences of various policy choices. In a way similar to thatdescribed by Cai et al. �2003�, all elements are incorporated intoa single consistent model. Our model is solved in its entirety byGAMS-MINOS, a widely used, off-the-shelf nonlinear program-ming solver �GAMS 1998�. The present work is similar to that ofCai et al. �2003�, who solved a basinwide optimization model.However, the current model has less detail on irrigation choicesand salinity at the farm level, but greater detail on municipal andindustrial �M&I� activities and values. In addition, the currentmodel has more detail on institutional constraints governing thetransboundary allocation of water, namely interstate compacts andinternational treaties.

This paper provides an example of reducing economic dam-ages resulting from drought through market-based water transfersin the Rio Grande Basin �Fig. 1�, a region with a history of com-peting water uses, a complex set of institutions, and growing de-mands on available water supplies. We extend previous work bydeveloping an integrated hydrologic-economic model at the basinscale that incorporates surface- and groundwater interactions andtracks economic and hydrologic relationships over several timeperiods. We apply the model to the design of market-based watertransfers as a measure for coping with severe and sustaineddrought.

The remainder of the paper is organized in the following way:First, the physical and institutional environment of the Basin’swater use is summarized. Then, we summarize the methods usedto develop and use the basin model. Next, the results are pre-sented for which we compare the “Law of the River,” described inthe section on characterizing institutions, with two institutionsthat harness the power of market forces. These two drought-coping institutions are �1� water trading within existing politicalboundaries and �2� water trading across boundaries. Finally, wepresent our conclusions.

Physical and Institutional Context

High mountain snowpack is the Rio Grande Basin’s primary sur-face water source. The southern area of the Basin flows throughthe Chihuahuan desert, where annual average precipitation is 8 in.�20 cm�, most of which comes as widely scattered summer mon-soon thunderstorms �Schmandt 2002�. Total annual streamflow inthe Basin is variable from year to year, and history has recordedseveral periods of severe and sustained drought. The Basin hasbeen in drought conditions for the last 7 years and has facedsevere drought for the last 3 years. Basin inflows for the previous2 years were 11 and 10% of the 30-year average and were, re-spectively, the eighth and fifth lowest flows on record �Michelsenand Cortez 2003�. In fall 2004, water storage in Elephant Butte,the largest reservoir in the Basin, was less than 5% of capacity.After an unprecedented 25-year period of full water supplies,water allocations during 2003 were reduced to just one-third of
full supply conditions.

The Basin supports a rapidly growing population of more thanthree million people, extensive irrigated agriculture, and fish andwildlife habitat in Colorado, New Mexico, Texas, and the Mexi-can state of Chihuahua. An estimated 80 to 90% of the water inthe upper Basin is used for irrigated agriculture; the main cropsare forage, cotton, pecans, and vegetables. Only a portion of thewater is consumptively used compared to the total water appliedto crops. This portion varies from a low of about 30% in centralNew Mexico agriculture to a high of about 70% in southern NewMexico and west Texas agriculture. The remainder is an impor-tant source for groundwater recharge, supply for riparian habitat,and return flows to downstream users.

While M&I water demands in the basin have historically beenmet by groundwater, this pumping is not sustainable at currentuse rates, let alone with added pumping for growth. El Paso con-tinues to expand its use of surface water, and Albuquerque plansto withdraw and treat surface water. The largest basin city, CiudadJuárez, is projected to deplete its fresh groundwater reserves inless than a decade and is in need of new sustainable supplies�Paso del Norte 2001�.

Environmental demands for and values of water continue toincrease. The Rio Grande silvery minnow �Hybognathus amarus�was listed as an endangered species by the U.S. Fish and WildlifeService in 1994. Despite contentious debate and extensive andcostly litigation, minimum river flows have been mandated by thefederal courts �e.g., Parker �2002�� and to date have been carriedout by the federal agencies to sustain remaining minnow popula-tions, even in periods of drought. As a result, the federal govern-ment has temporarily acquired and reallocated water from exist-ing uses to provide adequate instream flows for the minnow in theSan Acacia reach of the river near Socorro, New Mexico �Fig. 1�.One cubic foot per second �cfs�=0.028321 m3/s.

Water managers in the Rio Grande face a complex set of in-ternational, federal, state, and local institutions regulating flowsand water allocation. These include an international treaty be-tween Mexico and the United States; a tristate compact involvingColorado, New Mexico, and Texas; involvement of multiple fed-eral agencies; state statutes; ongoing litigation; and numerouswater delivery contracts. Collectively, we shall later characterizethese as the Law of the River.

While existing institutions for allocating water have served thebasin well in recent decades, there is still considerable interest inthe design and evaluation of alternatives �e.g., Middle Rio GrandeWater Assembly �2001�, Tidwell et al. �2004�� that could bettermeet the needs of people who live there. The next section de-scribes economic and hydrologic principles underlying the devel-opment and application of an integrated model designed to testwhether alternative institutions for water management and alloca-tion could significantly reduce damages produced by the inevi-table drought.

Methods of Analysis

Integrated Basin Model

A framework for estimating future drought impacts and testingalternative policies was developed to account for the Basin’s criti-cal hydrologic relationships, institutions, and economic sectors.This integrated model is formulated as a mathematical optimiza-tion problem, using the sum of benefits from basin water diver-sions for off-stream uses and benefits of in-stream use �recreation�
as the objective. Constraints are used to characterize basin hydrol-

ogy and institutions. This analysis extends similar previous workby Vaux and Howitt �1984�, Booker �1995�, and Hurd et al.�2002, 2004�, all of whom developed integrated basinwide hydro-logic models for policy analysis containing an economic objec-tive. In this work, the hydrology identified by Cai et al. �2003� ascharacteristic of integrated models is expanded with the introduc-tion of multiple lagged ground- and surface-water interactions.

For the present work, the model is formulated and solved onan annual time step, with reservoir contents and other hydrologicand economic conditions carried forward to the next time period.The actual model code supports dynamic solutions as well, where,for example, reservoir contents could be managed over a prede-termined period to achieve the greatest benefits. Such a perfectforesight model could be useful in establishing a benchmark for“perfect” drought management. Additional details and the mod-el’s GAMS code are available online and from the writers onrequest at �http://wrri.NMSU.Edu/publish/techrpt/tr317/cdrom/�.While the model and its documentation were developed for theRio Grande Basin, it was designed to be easily adaptable to otherbasins, cultures, and contexts.

Characterizing Institutions

The allocation and use of water in the Rio Grande Basin isheavily constrained by limited water supplies and by existing in-stitutions. We characterize these institutions as the Law of theRiver. Among the institutions, the Rio Grande Compact and theU.S.-Mexico Treaty have the most influence on allocation and useand also are the least flexible to short-run change.

Law of the RiverIn our model, solutions representing water allocations under theLaw of the River are obtained through a single optimization bymaximizing basinwide economic benefits, subject to institutionaland hydrologic constraints. Our intent is to create a baseline ofexisting institutional allocations for comparison with institutionalchanges. The Law of the River is simulated by constraining waterallocations to produce outcomes consistent with historical wateruse patterns. These patterns are consistent with the Compact andMexican Treaty, as well as with patterns characterized by largeamounts of water used in agriculture that produce low marginaleconomic values compared to M&I use.

Intrastate BankingFor intrastate water banking, the objective function is the sum ofall net benefits in consumptive water uses; institutional con-straints previously specifying the allocation between users withineach state are removed from the constraint set. By finding waterallocations that maximize economic benefits of consumptive use,we approximate allocations resulting from an intrastate waterbank.

Within each of the three Compact states, the model maximizestotal regional economic benefits by equalizing marginal benefitsamong users. This maximization occurs to the extent permitted bythe constraints of hydrology, Compact allocations, and Treaty de-liveries. This regional objective is consistent with the institutionof an intrastate water market that facilitates trading water for cashwithin each state, but not across state boundaries. The recentlypublished New Mexico State Water Plan emphasizes markets asan institution to make a given amount of water perform moreeconomic work �increase economic efficiency�, as long as themarkets are implemented to protect third parties by accounting for
hydrologic impacts of water trades.

Interstate BankingA more geographically widespread market linking all NewMexico and Texas water users can be imagined, and Colorado andMexico could also be included in an interstate market of evenlarger geographic scope. This larger market is an important areafor future research. The potential for interstate trades betweenNew Mexico and Texas has received considerable recent attentionin the Basin. For this case, the objective function includes allmonetized benefits, including net benefits of consumptive uses inaddition to recreation benefits of reservoir recreation. Reservoir-based recreation benefits were estimated from the regional travelcost model described in Ward et al. �1997�. Compact deliveriesfrom Colorado are maintained, but we remove Compact con-straints allocating water among the potential water trading part-ners of New Mexico and Texas. The result is an estimate of allo-cations under interstate water banking between New Mexico andTexas. Total water supplies are unaffected by this institution, butinstead of water allocations among the states being governed bythe Compact, water flows to its highest-valued uses. Marginalbenefits of each use are equalized both within and across statelines of the trading partners subject only to hydrologic con-straints, such as gains and losses as water moves downstream.

Objective

The objective function follows the form

V = �v

��vt0 + �vt

1 Uvt + �vt2 Uvt

2 + �vtp Pvt� + �

u

�bu0 + bu

1Zut + bu2Zut

2 �

�1�

where the first summation term is a node’s total water-relatedbenefits as a quadratic function of consumptive use, Uvt, at the vthdiversion node and a negative linear function of that part of thatnode’s consumptive use consisting of pumping, Pvt. This last termin the first summation accounts for the higher cost of pumpingcompared to simple surface diversion ��vt

p �0�. The second sum-mation term defines recreation benefits at the uth reservoir, as aquadratic function of storage Zut.

Basin Hydrology

The basin hydrology is defined in both annual flows Xit and res-ervoir stocks Zut. The fundamental mass balance for flows can bewritten as

Xrt = �i

�irXit �2�

where Xrt is streamflow at the rth river nodes; Xit are flows of typei �Table 1�, and �r are the elements of an r� i matrix defining thespatial configuration of the basin that links i to river node r.Elements of �r can be either +1 �the inflow point is immediatelyabove river node r�, −1 �the off-stream diversion point is imme-diately above r�, or most commonly, 0 �flow r has no direct effectat river node r�. Flow types in Eq. �2� include headwaters inflows,off-stream diversions, current and lagged return flows, impactsof groundwater pumping, depletion from nonagricultural plants�phreatophytes�, net releases from storage, and conveyancefunctions.

Table 2 shows the geographical and functional relation amongthe basin’s major water sources and uses. Depending on the loca-tion patterns of headwaters, river reaches, cities, agricultural
users, reservoirs, and endangered species requiring minimum

flows, a table such as Table 2 would be adapted to the uniqueconditions surrounding any basin’s hydrology, settlement pat-terns, and demand conditions. The table characterizes the basin’sgeometry, summarizing locations of water supply and use points,and shows how each of the major functions that influence stream-flow at a point are affected by various upstream activities. Thesefunctions include use, seepage, pumping, net seepage, net ground-water return flow to the river, surface-water return flow to theriver, evaporation, and reservoir net release.

Brief explanations of each function are described in the table’sbottom row. Each column is a river basin function described by ablock of similar equations. Coefficients are typically equal to 1

Table 1. Classification of Nodes in Rio Grande Basin Model

Set hierarchy and labels

DescriSet Type of node

i Set of all flows, occurring over

r River nodes

h Headwater inflows, above all m

v Diversions for off-stream use

n Return flows in current period

g Lagged return flows net of impfrom pumping

b Depletion by phreatophytes

c Conveyance functions for unme

w Set of flows linked to stocks �re

e Evaporation from basin reservo

s Net contribution to flows from

u Set of reservoir stocks, measure

Note: NA�not applicable.

Table 2. Hydrologic Balance in Rio Grande Basin

Item Inflowa Diversionsb Usec Seepaged Pump

Flows

Inflow 1 1 — — —

Diversions — — 1 �ds —

Use — — — — —

Seepage — — −1 — —

Pumping — — 1 �ps —

Net seepage — — — — —

Net 1

Groundwater flow to river — — — — —

Surface return flow to river — 1 −1 — —

Evaporation — — — — —

Reservoir net release — — — — —

River flow — 1 — — —

Reservoirs — — — — —aInflow=source flows.bDiversions�upstream contributing flows �wet river�.cUse=diversions−surface returns+pumping−seepage.dSeepage increases with �diversions+pumping�.ePumping makes up for surface flow shortages.f
Net seepage=seepage−pumping. R

�contributing�, −1 �reducing�, or 0 �no effect�. For some equationblocks, coefficients are not equal to 1. An example is the seepageblock. Seepage into an aquifer depends positively on total waterdiverted from the river and applied to beneficial use, as well as ontotal water pumped and applied. This relationship is shown by thetwo sets of positive coefficients Bds �effect of diversions on seep-age� and Bps �effect of pumping on seepage�. Each cell in Table 2is an intersection of a row defined as a set of nodes in the basinwith a common function and a column defined as a set of nodeswith another common function. For example, the boldface vari-able 1 in Table 2 shows the intersection of all river flow nodeswith all other river flow nodes.

Sign of � Number of elements

ar period NA 48

NA 10

diversions � 6

— 8

� 8

groundwater depletion � 8

— 2

gains and losses � 5

r storage� — 3

NA 3

ir releases � 3

ch year’s beginning NA 4

Netseepagef

Netground-water

flow toriverg

Surfacereturnflow

to riverh Evaporationi

Reservoirnet

releasejRiverflowk Reservoirsl

Stocks

— — — — — 1 —

— — �dr — — −1 —

— — — — — — —

1 — — — — — —

−1 — �pr — — — —

— �ng — — — — —

1

— — — — — — —

— — — — — 1 —

— — — — — — −1

— — — — — 1 −1

— — — — — 1 —

— — — �re — — 1

Groundwater flow to river increases with net seepage.

Surface return flows to river increase with �diversions+pumping�.vaporation increases with reservoir volume.

eservoir net releases for use.

Gauged flow=sum of upstream contributing flows.

ption

a 1-ye

odeled

acts of

asured

servoi

irs

reservo

d at ea

inge

g

h

iEjRk

l
eservoir level=previous level−net release.

Table 3 shows the kind of detail required to fill out the boldface cell of Table 2. The coefficients in Table 3 show how flows ateach river node depend on flows from one or more upstream nodethroughout the basin. Source flows at the gauge at each columnoriginate from flows at the row for which there are one or morecoefficients equal to 1. Each cell in Table 2 has this kind of detail:node-by-node connections between column headings and rowheadings.

The mass balance for reservoir stocks Zut is given by

Zut = Zu,t−1 + �w

�iwXwt �3�

where �iw are matrix elements linking annual changes in storageto the flow types directly causing storage changes. These flowsconsist of evaporation, calculated as a quadratic function of stor-age, and net releases, calculated as reservoir outflow from storageminus inflow; all nonzero elements of �iw equal −1. Mass balancefor off-stream consumptive use Uvt and diversions Xvt is given by

Uvt = Xvt + Pvt − Svt − �n

�nwXnt �4�

where Pvt is withdrawal of pumped groundwater, Svt is seepageinto groundwater from canal leakage and field infiltration exceed-ing use by plants, and Xnt is the return flow to the river within thecurrent time period. The return flows are given as a proportion �n

of diversions and pumping.

Xnt = �n��v

�vn�Xvt + Pvt�� 5�

where �nv is the matrix of 1 s and 0 s that configure the basin bylinking diversion and return flow at relevant locations. The inter-action of flow elements shown in Fig. 2 shows that infiltration togroundwater from the river itself is treated as a negative returnflow �lagged or current period�. Where such infiltration is notdirectly modeled, conveyance functions, as shown in the figure,allow for river losses.

Total seepage into groundwater from diversion v, equal to Svt,is given by

Svt = �v�Xvt + Pvt� �6�

where �v is the proportion of diversion v infiltrating into ground-water rather than contributing directly to return flows. The effectof this infiltration is to contribute to the net seepage Nvt to

Table 3. Geographic Configuration of River-to-River Nodes in Rio Gran

Node Lobatos Embudo Chamita Otowi CO

Lobatos — 1 — —

Embudo — — — 1

Chamita — — — 1

Otowi — — — —

CO gauge — — — —

Acacia — — — —

Marcial — — — —

EB gauge — — — —

El Paso — — — —

Fort Quitman — — — —

groundwater, defined as


Nvt = Svt − Pvt �7�

where Pvt is the direct withdrawal of groundwater by pumping.Eq. �8� can then be used to find return flows from groundwater Xgt

calculated as a lagged function of net seepage

Xgt = Gg + �t�=0

t

g,t−t��v

�vgSv,t−t�� 8�

where g,t−t� and Gg is the contribution to groundwater flows ineach period independent of net seepage.

The model has a yearly time step, but lagged seepage pro-duced in any given year is carried forward to the following year.The parameters g,t−t� are unique to each groundwater basin andrepresent our best judgment from the literature and from localexpertise on the magnitude of and lag between ground- andsurface-water impacts. There is no explicit representation ofgroundwater storage; rather, changes in any period’s stock ofgroundwater are represented through past seepage and pumpingactivities.

Groundwater pumping limits are specified that reflect bothavailable infrastructure and the short-term possibility of signifi-

sin

e Acacia Marcial EB gauge El Paso Fort Quitman

— — — — —

— — — — —

— — — — —

— — — — —

1 — — — —

— 1 — — —

— — 1 — —

— — — 1 —

— — — — 1

— — — — —

Fig. 2. Interaction of flow elements for Rio Grande Basin model

de Ba

gaug

—

—

—

1

—

—

—

—

—

—


cant drawdown or depletion of shallow aquifers during drought.The latter effect is captured through a pumping limit that is adecreasing function of lagged river flows. The purpose of thefunctional form is to capture decreasing ability to pump fromshallow, river-flow-dependent groundwater

Pvt � Pbase + �P0 − Pbase��1 − exp�vXrt�� 9�

where Pbase defines the pumping capacity from flow-independent,large deep groundwater reserves, and the second term, P0− Pbase,defines the pumping capacity from the shallow aquifer. Lagged

flows Xrt are a simple average of recent �typically 3 years� riverflows. Pumping depth �lift� increases with reduced lagged riverflows for aquifers connected to the river, so long periods of lowflow indirectly reduce pumping capacity by increasing the cost ofpumping through greater lift required to access groundwater. Im-portant future research needs to connect current and lagged riverflows, pumping lift, the change in cost of pumping groundwater,and the reduced economic attractiveness of that pumping.

Depletions by phreatophytes are similarly defined by a laggedfunction of past river flows. Conveyance functions representingnonmodeled gains and losses Xct �e.g., ungauged inflows� are lin-ear functions of headwaters flows.

Institutional Constraints

Rio Grande CompactThe Compact—signed in 1938 by Colorado, New Mexico, andTexas—divides the annual flow of the Rio Grande. Under theCompact each state receives more water in wetter years. ArticlesIII and IV of the Compact oblige Colorado to deliver water at theColorado-New Mexico state line. These flows, measured at theLobatos stream gauge �Fig. 1�, must be at least

XLobatos,t � �h

�0h + �1hXch,t + �2hXch,t2 �10�

This additive quadratic equation approximates Colorado’s totaldelivery requirements to New Mexico defined by Compact ar-ticles III and IV and is based on annual source runoff measured atColorado’s relevant headwater gauges Xch,t. The � terms are con-stant coefficients for the intercept, linear, and quadratic terms,respectively, for headwater gauge flows. The delivery scheduledefines Colorado’s water rights to be senior to those of NewMexico, hence

�2XLobatos,t

�XCh,t2 0 �11�

which says that Colorado’s delivery obligation to New Mexicoincreases at an increasing rate with increases in Colorado’sgauged runoff, Xch,t. The effective seniority of Colorado’s waterright compared to New Mexico comes from the historical fact thatprior to 1929, Colorado water users took a higher percentage ofColorado’s native runoff in dry years. Each state’s right to usewater and its requirements to deliver to the downstream stateunder the Compact are based on historical use patterns and re-sponses to drought that occurred prior to 1929. So the desirableposition occupied by water users located further upstream, stillcommon throughout the world, is reaffirmed by the terms of theCompact.

Article V of the Compact and the February 1948 resolution ofthe Compact oblige New Mexico to deliver water to Texas mea-
sured at the outflow of Elephant Butte Reservoir �Fig. 1�. New

Mexico’s delivery requirement to Texas is based on New Mexi-co’s annual supply, defined as total flows at the Otowi streamgauge, north of Santa Fe, New Mexico. It is approximated in themodel by

XEB,t − XEBR−r,t � �0 + �1XOtowi,t + �2�XOtowi,t�2 �12�

where the � terms are constant coefficients, XEB,t is annual out-flow from Elephant Butte Reservoir, the location at which NewMexico must deliver to Texas under the Compact, and XOtowi,t

indicates annual flow past the Otowi gauge. Under the Compact,annual flow at that gauge is defined as New Mexico’s total supplyfrom which a proportion must be delivered to Texas. The twoleft-hand side terms are Elephant Butte Reservoir outflows �XEB,t�and net releases XEBR−r,t, respectively. The algebraic difference ofthose two terms, when added to reservoir evaporation, is math-ematically equivalent to the Elephant Butte Reservoir outflowplus the net change in reservoir storage, which is New Mexico’sdelivery obligation to Texas. The original Rio Grande Compactdefined New Mexico’s delivery obligation to Texas as inflows toElephant Butte. However, that measurement proved unreliable, soa more accurate measurement was moved to the gauge immedi-ately downstream of the reservoir. Measuring what New Mexico’sdelivery into the reservoir with no reliable inflow measurement isaccomplished by measuring outflow, change in reservoir storage,and evaporation.

U.S.-Mexico Treaty of 1906A constant 60,000 acre-feet annual delivery to Mexico is assumedin the absence of severe drought, as required by the U.S.-MexicoTreaty of 1906. Historically, in times of severe drought, U.S. de-liveries to Mexico have been reduced considerably below60,000 acre-feet in approximate proportion to the reduction ofdeliveries from Elephant Butte Reservoir to U.S. lands, consistentwith the language of the Treaty. Thus, following that language,U.S. deliveries to Mexico are described as

XMX,t = 60,000��LB

XLB,t

�LB

XLB,t �13�

where XMX,t=annual deliveries to Mexico, XLB,t=actual deliveries

to U.S. lands from the Rio Grande Project, and XLB,t is the deliv-ery to U.S. lands in a full supply condition, defined as wateroutflows from project storage, equal to 790,000 acre-feet.

Minimum Flows to Protect Endangered SpeciesA biological opinion issued by the U.S. Fish and Wildlife Service�U.S. Department of Interior 2001� estimates that the endangeredRio Grande silvery minnow requires at least 50 cfs of year-roundstreamflow between the river at approximately Socorro, NewMexico, and the inflow to Elephant Butte Reservoir �Fig. 1�. Aregression analysis showed that total annual deficits, defined asthe total additional acre-feet of water needed to overcome allshortages in streamflow below 50 cfs, takes the following form:

XMt � �0v �14�

That is, annual flows at the San Acacia gauge �Fig. 1�, XMt, mustexceed a critical level required for the minnow’s survival, about
240,000 acre-feet per year under current operating conditions.

Demand, Infrastructure, and CustomCurrent water users in the Basin have typical maximum levels ofsurface diversions and pumping. In the case of agricultural uses,these may be limited by land and irrigation infrastructure and arefairly stable over time, as long as the Basin’s reservoirs are suit-ably full. For regional municipal uses with rapidly growing popu-lations, the maximum use increases as population increases. Thecombination of existing and future water demand, limited infra-structure, and both customary and legal restraints puts a practicalmaximum on consumptive use at each off-stream location. Thuswhile economic demand functions with known elasticities areused for approximating market allocations, the demand functionis capped at an upper limit. The institutional constraint limitsactual consumptive use, Uvt, to be no more than anticipated future

demands in the case of municipalities, Uvt,

Uvt � Uvt �15�

Water Allocation Efficiency

Water allocations or water policy designs can be judged from theview of economic efficiency, where water allocations are seen asa series of investment projects. Water is the limiting resource, andthe economic sectors use the resource and produce economic re-turns. In an economically efficient water allocation, the marginalbenefit from the use of the water is equal across uses in order tomaximize total benefits produced by the water. Economically ef-ficient water use in a river basin requires that the equimarginalcondition be satisfied to the extent possible consistent with hydro-logic and institutional constraints. That is, the marginal benefit�MB� of one additional acre-foot of water to one user is equal tothe incremental benefit from an acre-foot for every other user.Where this condition fails to occur, society benefits economicallyby allocating more water to the sector where the incremental ben-efits are highest. The economic efficiency challenge posed bywater scarcity is to allocate a known supply of water among allcompeting users to maximize the economically beneficial use ofthat water. Economic efficiency is achieved if the followingequimarginal rule is followed:

MB1 = MB2 = . . . = MBn �16�

Any policy or institution that effectively transfers the water tousers who value the marginal water more also increases total netbenefits produced by the water’s use. People who lose use of thewater lose a smaller economic benefit than water gainers gain inbenefit. Finally, when enough water is transferred so that MBs areequalized among all users within each state, it is not possible toincrease overall total net benefits by more transfers.

Valuing Water in Alternative Uses

The economic value of water comes from the many uses to whichwater can be put in satisfying human needs. The economic valueof water is defined as the amount that a rational user of a publiclyor privately supplied water resource is willing to pay. Willingnessto pay for water reflects the water user’s willingness to forgoother consumption and is measured by a demand schedule relat-ing the quantity of water used at each of a series of differentprices. For any potential quantity that could be supplied, demandis limited. So the economic value of an added unit of water sup-plied decreases as greater quantities are offered to water users.
For example, most people will use water only for irrigating their

lawns or for low-valued crops if the price of water is suitablycheap. At a high price, those economic values are too low tojustify purchasing the water.

Quantity supplied is only one dimension of water’s use. Timeutility of water use can be improved by building dams and devel-oping groundwater reserves, while location utility can be im-proved by building water transport systems such as aqueducts tomove water to places far from the natural watercourse. Moreover,location utility itself is measured in three dimensions, sincegroundwater is increasingly expensive to put to beneficial usewith increasing depth and increased energy prices. The modeldescribed in this paper does not consider quality effects on wa-ter’s value nor does it consider increasing pumping costs withdepth. Finally, water may be of varying qualities depending uponthe soils through which it moves or on how people affect thewater by supplying or using it.

Water Values in AgricultureFor the Colorado region, the economic value of water is deter-mined using an optimization model that maximizes annual agri-cultural income in the San Luis Valley �SLV� for various possibleannual water supply conditions. Water supply conditions are de-fined by �1� the quantity of water in the aquifer and �2� totalannual streamflow in the Rio Grande available for use in Colo-rado. The allocation of water by water right priority distributesstreamflow from the Rio Grande to irrigation ditches and canalsholding the highest priorities. Cropping patterns are dependentupon the amount of surface water available and whether ground-water pumping rights are owned by the producer. Those patternsand the associated net returns from irrigation water are then op-timized based on crop production functions and costs of produc-tion for the major crops grown in the study area �Dalsted et al.1996; Sperow 1998�.

Downstream in New Mexico and west Texas, the agriculturalanalysis uses methods similar to those applied in the Coloradoregion, but with less detailed accounting of the explicit interactionbetween the economics and hydrology of surface water andgroundwater. The analysis is based upon estimating how croppingpractices under full water supply conditions adapt to variousdegrees of drought severity. All three of the Basin’s major agri-cultural regions in New Mexico and west Texas were chosen foranalysis: �1� Middle Rio Grande Conservancy District �MRGCD�near Albuquerque, New Mexico; �2� Elephant Butte IrrigationDistrict �EBID� near Las Cruces, New Mexico; and �3� El PasoCounty Water Improvement District No. 1 �EP#1� near El Paso,Texas. For each of these three farming areas, agricultural prices,yields, and production costs are incorporated for the area’s mostimportant crops.

The analysis is based on farm cost and return enterprise bud-gets published by New Mexico State University and Texas A&MUniversity. For each area, linear programming models were de-veloped and applied to represent behavior of commercial produc-ers that maximize net returns, using standard methods for valuingwater in agriculture �e.g., Ward and Michelsen �2002��. Income-maximizing farm behavior models are calibrated to produceoptimized cropping patterns consistent with historical croppingpatterns, in the spirit of the positive mathematical programmingapproach described by Howitt �1995�, Martinez et al. �1999�, andHeckelei and Britz �2000�. Results of these income-maximizingmodels are based on constraints on available land in each majorcropping area and by cropwater production technologies �Ward
et al. 2001�.

Table 4 summarizes the benefit functions for these four irriga-tion regions. These agricultural benefit functions were developedin a previous analysis using the methods described above. Afterthat previous analysis was completed, the functional forms andcoefficients summarizing the benefit functions were used for thecurrent model.

Groundwater and surface water are typically good substitutesfor each other, but pumping costs make groundwater more expen-sive to deliver to farm fields than surface water. Groundwater andsurface water are substitutes when both can be applied to thesame use at the same time. If the two are available in two differ-ent regions without available conveyance between the regions,then they are not substitutes. In that case, separate benefits func-tions should be used for each source: once the supply of one isgone, there is no substitute backup. However, if conveyance canbe built between the two sources at a cost less than the incremen-tal value produced by each serving as a supplemental backup forthe other, then that construction passes the test of economic effi-ciency. In this study region, many of those efficiency gains havealready been developed: groundwater is available for the sameuse and time, and conveyance systems are in place and are used tosubstitute for surface water supplies. For this reason, the Table 4shows groundwater entering the benefits function with a negativecoefficient for all regions in which groundwater pumping occurs.That is, total benefits depend on the sum of surface water andgroundwater used, but benefits are reduced on that part of thetotal supplied by groundwater because of its additional pumpingdelivery cost.

Water Values for Municipal and Industrial UsesThe use of water produces considerable economic value in a mod-ern household. Beyond satisfying basic human requirements,water has been extensively analyzed as an economic resource forwhich there is considerable urban demand, particularly in thedesert Southwest. Similarly, water shortages resulting fromdrought cause economic damages for which people are willing topay considerable amounts to avoid. Besides cooking, washing,cleaning, and sanitation, the typical Rio Grande Basin householdin the United States uses water for outdoor cleaning and to sustaina domestic landscape environment.

The empirical analysis for the current study for estimatingdrought’s economic impact is based on earlier work by Michelsenet al. �1998�. That impact is measured as the willing-ness to pay to avoid drought damages. Seven study areas wereselected for that study. With the cooperation of water utilities in

Table 4. Total Benefits of Consumptive Use for Agricultural and MuniGrande Basin, Colorado, New Mexico, and Texasa

Location Label State Sector�0

�$��1

�$/acre-foot�

San Luis Valley SLV CO Ag 195 145

Albuquerque ALB NM M&I 0 10,843

Middle Valley MRGCD NM Ag −30 67

Mesilla Valley EBID NM Ag 137 94

El Paso EP TX M&I 0 9,507

El Paso EP#1 TX Ag 0 193

Note: NA�not applicable.aFunctional form for economic benefits at these nodes are total benefits=�bFor EBID, incremental benefits from added water use fall slowly o-feet/acre, so 18.80 is a considerable extrapolation beyond observed use

California, Colorado, and New Mexico, information was col-


lected on residential water use, rate structures, and revenue fromwater sold and nonprice conservation programs covering the pe-riod from 1980 through mid-1994. Across seven cities, water’sdemand was found to be price inelastic, which means that a largepercentage of increases in price are required to induce small per-centage decreases in water use. The highest price elasticity esti-mate was for summer landscape use �approximately −0.20�.

The current study applied the empirical demand schedule find-ings to the climatic and demographic conditions of the basin’stwo major U.S. cities, Albuquerque �ALB� and El Paso �EP�. Foreach city, a linear demand schedule was defined to pass throughthe water use and price combination for 2003 �Ward et al. 2001�.The slope of each city’s demand was defined to produce theknown price elasticity and the 2003 combination of price and use.For a known price elasticity, the slope of a linear demand curvecan be determined once the price and quantity are known. Wechose the integral of the marginal benefits of water use to measuretotal benefits of that use. A linear demand function produces aquadratic total benefits function, of which those total benefitspeak at the level of water consumption produced by a zero price.For higher consumption levels, marginal benefits of additionalwater are negative. When water is scarce, a model that optimizestotal benefits will assign water only to use for which marginalbenefits are positive. Table 4 shows that for each city the marginalbenefits of groundwater use are lower than for surface water, be-cause of higher pumping costs for obtaining groundwater.

Interpretation of Benefit FunctionsThe last column of Table 4 provides a key to interpreting thebenefit functions by showing the quantity of water at which in-cremental benefits of additional water use are zero. That quantityof water represents the point at which the quadratic total benefitsfunction reaches a maximum for both agricultural and M&I uses.Note that for the typical household, maximum total benefits occurat a point of about 0.5 acre-feet per household per year. For agri-culture, maximum total benefits per acre occur at a much largeramount, typically larger than 4 acre-feet per acre. This largerquantity of water reflects the fact that in the face of plentifulwater supplies, irrigators have shown that they substitute largeamounts of comparatively low-valued water for land, labor, andother resources.

Results

Results show drought impacts starting from the severely depleted

nd Industrial Uses of Water Per Acre �or Per Household� Per Year-Rio

�2

cre-footb��p

�$/acre-foot pumped�

Use at which MB=0�acre-feet per acre

or per household, surface water�

−14.0 NA 5.18

9,627 −10 0.56

−5.9 NA 5.68

−2.5 −76 18.80b

9,392 −100 0.51

−21.5 NA 4.49

�acre-feet consumed� +�2�acre- feet consumed�2+�p �acre-feet pumped�.

e observed data range; however, maximum observed use was 6 acres.

cipal a

�$/a

−

−

0+�1

ver thpattern

reservoir conditions existing at the end of 2003. We focus on


seven scenarios representing a combination of three water supplyconditions and two drought-coping institutions for dealing withshortage situations other than the existing Law of the River. Thehydrologic conditions produce three constant inflow levels to thebasin for six consecutive future years, 2004–2009: 100, 75, and50% of long-term mean annual flows. For comparison, the lowesthistorical annual inflow occurred in 2002 and produced only 37%of the annual mean.

Basin supply and demand for the years 1998–2003 are usedfor model calibration. Historic Elephant Butte reservoir volumesand actual deliveries to users downstream from the reservoir weretargeted, with unmeasured groundwater return flows in the reachbetween Cochiti Lake and Elephant Butte Reservoir as the pri-mary calibration factor. Typical model dimensions are 112 vari-ables, 153 equations, and 380 nonzero elements for the first yearof what is typically a multiple-year, sequentially solved problem.Solving for the Law of the River with a 12-year time period usingMINOS solver requires approximately 15 s of actual clock timeon a 1.8 GHZ PC with 512 MB memory running GAMS version20.7; solutions estimating intrastate and interstate market alloca-tions require similar total run times.

A benchmark from which to gauge future drought damages isprovided by the same 1998–2003 time period. Basin supply anddemands for the years 1998–2003 are used for model calibrationand to provide a benchmark from which to gauge future droughtdamages. Results are presented to show the kinds of impacts as-sociated with various levels of drought and different institutionsfor coping with drought. If basin inflows for years 2004–2009equal levels typical of the years 1998–2003, about 75% of thehistorical mean, then there will be no recovery of reservoir stor-age and little resiliency to cope with short-term droughts. Whileactual future basin inflows will certainly vary from one year to thenext, results suggest that sustainable recovery is unlikely over the2004–2009 forecast period with average inflows equal at the 75%level.

Total basin consumptive use in New Mexico and Texas ini-tially sustains levels similar to those in 1998–2001, but then fallsas the City of Albuquerque replaces its current groundwaterpumping with contracted surface withdrawals from the RioGrande as it develops its surface treatment capacity. The tempo-rary partial recovery of consumptive use levels comes at the con-

Table 5. Average Annual Consumptive Use by Institution, Drought Con

Baseline Institutionand policy adjustmentsto drought

Drought condition,percentage of average

annual surface water supply�%�

Colorado

SLV

Ag

Law of the riverc 100 �Baseline� 758 �147�

75 625 �118�

50 445 �73�

Intrastate water marketinge 75 625 �118�

50 445 �73�

Interstate water marketingf 75 625 �118�

50 445 �73�aAverage annual surface supply produced by the headwater gauges is 1.5bAnnual growth in M&I demand functions are based on historical populacCurrent system operation rules maintained in the face of drought. No tradThe distinction between water consumption from pumping versus surfaceTrading water for cash is permitted within each state.fWater can be traded for cash both within a single state and across state

tinuing cost of groundwater depletion by Albuquerque, El Paso,


and Juárez. Surface water deliveries to Mexico are disproportion-ately impacted compared to water use in the basin as a whole. Forthe 2004–2009 forecast period, Mexican deliveries are only 73%of the full supply level of 60,000 acre-feet. Consistent with the1906 U.S.-Mexico Treaty, total surface deliveries to all users �in-cluding Mexico� below Elephant Butte Reservoir average only73% of full levels during the 6-year period.

Impacts of Alternative Water Supplies

Table 5 shows hydrologic impacts of the Law of the River and thetwo alternative drought control institutions applied to each watersupply scenario. The basic structure of each is identical in pre-senting impacts on the major agricultural and municipal sectorsunder the seven combined flow and policy scenarios. For 2004–2006, Albuquerque relies entirely on groundwater for its M&Iuse. For 2007–2009, Albuquerque is presumed to have finished itssurface-water facility, diverting up to 97,000 acre-feet of surfacewater annually. Lagged impacts of previous pumping on riverflows persist during this period. Table 5 shows total consumptiveuse in each sector, including use derived from surface water andpumped groundwater. Growth in M&I water use from growingdemand occurs each year from 2004 to 2009 for both Albuquer-que and El Paso.

Law of the RiverUnder existing water allocation institutions, drought impacts areconcentrated among Colorado agriculture, MRGCD, and usesbelow Elephant Butte Reservoir, including Mexican irrigators.Municipal users are largely insulated from drought impacts forthe 6-year model projections here, as they rely more heavily thanagriculture on nontributary groundwater.

Colorado agriculture has little reservoir storage, relying in-stead on groundwater storage, injecting surface water into theiraquifer in wet years and withdrawing it in dry years. Reducedbasin inflows produce surface-water shortfalls of up to about 40%at the 50% basin supply level, compared to surface-water useunder 100% of normal supplies. Compensating for the low sur-face supply is considerable pumping capacity. With drought per-sisting over many years, shallow groundwater reserves resultingfrom irrigation recharge are largely exhausted, producing consid-

State, and Sector in Rio Grande Basin �1,000 s a-f /year�,a 2004–2009

New Mexico Texas Mexico

Total

MRGCD EBID EPb EP 1 Ciudad Juarez

I Ag Ag M&I Ag Ag

5� 195 �0� 277 �68� 89 �59� 133 �0� 56 �0� 1,609

5� 195 �0� 234 �69� 89 �66� 105 �0� 44 �0� 1,393d

5� 177 �0� 146 �69� 88 �77� 49 �0� 20 �0� 1,026

5� 195 �0� 235 �69� 89 �39� 77 �0� 44 �0� 1,366

5� 153 �0� 81 �69� 89 �41� 55 �0� 20 �0� 944

4� 96 �0� 279 �67� 89 �44� 113 �0� 44 �0� 1,347

01� 39 �0� 166 �67� 89 �45� 66 �0� 20 �0� 926

on acre feet. Bracketed numbers are consumption from pumping.

rowth of 2.4% per year for Albuquerque and 3.7% for El Paso.

f water for cash is permitted.

r use is shown in the table.

dition,

ALB

M&

101 �7

101 �7

101 �7

101 �7

101 �7

101 �9

101 �1

7 milli

tion g

ding o

e wate

lines.

erable economic damage. The resulting estimated benefits from


water in Colorado agriculture are highly sensitive to impacts ofreduced recharge on groundwater availability, but suggest thatgroundwater is a critical buffer at moderate surface supply reduc-tions, while substantial damages occur under the largest supplyreductions we consider.

For MRGCD, Table 6 shows that economic impacts of droughtare typically modest. With incremental economic damages of$26 to $28 per acre-foot of consumptive use in forage crop pro-duction and MRGCD drought shortfalls of about 4% of typicaluse, drought damages are less than $1 million for both droughtsituations. Downstream of Elephant Butte Reservoir, water use islimited by reservoir storage and by New Mexico’s reduced Com-pact deliveries to Texas produced by the ongoing severe drought.The Rio Grande Compact treats all use downstream of the reser-voir as Texas, even though there are more than 100 mil of NewMexico left before the river reaches the Texas state line. Withreduced deliveries under the Compact, Texas water users sufferconsiderable drought impacts. For the 2004–2009 period, surfacewater deliveries fall to 30% of normal under the 50% basin sup-ply scenario. Elephant Butte Irrigation District �EBID� uses exist-ing supplemental groundwater pumping to offset much of theshortfall, but still is able to maintain only just over 50% of fullconsumptive use under this lowest-flow scenario.

This model outcome reflects what actually occurred in the2003 irrigation season. Resulting damages from loss of cropsworth just over $100 per acre-foot are about $16 million annu-ally. For El Paso-area agriculture, little installed groundwater ca-pacity existed before 2003. Lacking such capacity, consumptiveuse could fall as low as 31% full when inflows are at the 50%level, with marginal consumptive use damages above $100peracre-foot, reaching totals of about $10 million annually. As withEl Paso municipal use, further damages and water managementstresses will likely result from elevated salinity at low-flow levels.

Deliveries to Mexico under severe drought are presumed to be

Table 6. Total Annual Drought Damages of Consumptive Use RelativCondition, State and Sector in Rio Grande Basin

Baseline institutionand policy adjustmentsto drought

Drought condition,percentage of average

annual surface water supplyDistribution o

damages

Law of the river 75 Drought damage

50 Drought damage

Intrastate marketing 75 Drought damage

Cash compensati

Net damages

50 Drought damage

Cash compensati

Net damages

Interstate marketing 75 Drought damage

Cash compensati

Net damages

50 Drought damage

Cash compensati

Net damagesaBaseline means 100% of long-term average basin inflows, 1.57 million acbenefits from reduced water availability compared to that baseline. A negdifficulty of assigning recreation benefits to state of visitor orgin, recreatbPositive �negative� number means cash received �paid�. Cash compensatithe River times proportion of basin-wide damages avoided by the transfe

limited to the surface-water proportions received by Texas water


users. Under the Compact, all use below Elephant Butte Reser-voir, New Mexico, is considered Texas. While surface water de-livered to Mexico is currently used for agricultural purposes, thatsupply is important for groundwater-dependent Ciudad Juárez.From a treaty allocation of 60,000 acre-feet, Mexico can expectsupplies to fall to about 45,000 and 20,000 acre-feet under meanwater inflow at the 75 and 50% basin supply levels, respectively.

Albuquerque and El Paso M&I water users are unlikely toexperience the severe damage suffered by agriculture. Albuquer-que is well insulated from drought impacts and currently satisfies100% of water use through largely nontributary groundwaterpumping. Further, the city will likely supplement those supplieswith surface-water diversions beginning in about 2007. Our re-sults show that under existing institutions, Albuquerque increasesits contracted consumptive use of the river to meet growing de-mand. Moreover, the city is able to reduce gross groundwaterpumping from about 185,000 acre-feet in 2004–2006 to 100,000acre-feet after beginning surface-water diversions, except whenbasin inflows fall to 50%. When that happens, high levels ofgroundwater pumping are used to maintain consumptive use.

Our results underestimate drought vulnerability in El Paso:water quality from elevated salinity declines considerably withboth lower river flows and higher pumping levels of marginalwells. El Paso is more likely to suffer short-run drought damagefrom poorer-quality supplies than from an absolute shortage ofavailable supplies. Still, results show that with continued popula-tion growth in El Paso, pumping will soon be insufficient to meetdemand at current water prices under very low river flows �e.g.,the 50% scenario shows a minor shortage in 2006–2009�. How-ever, greater than normal short-run pumping by El Paso also re-duces the length of time to depletion of its fresh groundwater

ormal Supplies and Existing Institutions, by Institution, Water Supply

Colorado New Mexico Texas

Total amongtrading partners

SLVAg

ALBM&I

MRGCDAg

EBIDAg

EPM&I

EP#1Ag

�Dollars, millions per year�

0.7 0.1 0.0 4.9 2.3 2.1 9.4

18.2 2.4 0.6 16.0 7.5 9.9 36.4

0.7 −0.2 0.0 4.9 −7.2 5.3 2.8

NA −0.2 0.0 3.4 −7.9 4.7 0.0

0.7 0.0 0.0 1.5 0.7 0.6 2.8

18.2 −0.2 1.5 25.1 −6.3 8.8 28.9

NA −2.1 1.0 12.4 −12.3 0.9 0.0

18.2 1.9 0.5 12.7 6.0 7.9 28.9

0.7 2.4 4.1 −0.1 −7.2 1.3 0.5

NA 2.4 4.1 −0.4 −7.3 1.2 0.0

0.7 0.0 0.0 0.3 0.1 0.1 0.5

18.2 3.5 7.8 13.4 −7.2 6.9 24.4

NA 1.9 7.4 2.7 −12.2 0.3 0.0

18.2 1.6 0.4 10.7 5.0 6.6 24.4

, for which damages are defined as zero. Damages are losses in economicumber refers to lower damages than under that baseline. Because of theefits are not included in drought damages.

istributed to each user by multiplying its drought damages under Law oftical avoided drought damages can produced different cash payments.

e to N

f

sa

s

s

onb

s

on.

s

on

s

on.

re-feetative nion ben

on is dr. Iden

resources, an impact not quantified here.


Intrastate MarketingWhile existing water policy in the Rio Grande Basin is basedlargely on maintaining historical patterns of water use, there areincreased calls for water allocation based more on current needsand demands. One alternative within the basin would allow ex-panded water markets to operate within each state. Such marketswould allow water transfer between willing buyers and sellers.Two market areas are considered: �1� New Mexico, betweenMRGCD agriculture and Albuquerque, and �2� Texas, among ElPaso M&I, El Paso agriculture, and EBID agriculture. Deliveriesto Mexico are assumed unaffected by either market. Economicimpacts, summarized in Table 6, include the sum of changes indrought damage from the value of water and offsetting cash trans-fers in which water is traded for cash. All parties are shown togain under this institution in comparison to losses produced bydrought under the Law of the River. Whenever a water user re-duces use �sells water� under this institution, cash income is largerthan the reduction in value of water used. Water users who in-crease their usage gain a value of water in use larger than theamount of cash paid for the right to use the additional water. Theadditional water is not an absolute gain, but only a smaller lossthan is produced by a drought under the Law of the River.

Introducing intrastate marketing has a modest effect insideNew Mexico at higher basin inflow levels, but considerably in-creases Albuquerque’s ability to use surface water under the low-est inflow levels of 50%. This is accomplished through transfersof 28% of the consumptive use otherwise applied to MRGCDagriculture. The net benefit of the transfer is shown in Table 6 bythe reduced drought damages incurred by agricultural users aboveand below Albuquerque, as well as reduced damages incurred byAlbuquerque M&I use. For the Texas intrastate market at the 75and 100% inflow levels, reductions in El Paso area agriculturaluse support reduced groundwater pumping by El Paso. EBID useis for the most part slightly affected compared to the Law of theRiver. Table 5 shows that total use throughout the basin varies byinstitution because of compensating groundwater pumping result-ing from transferred water.

Interstate MarketingA more geographically widespread market linking all NewMexico and Texas water users can be imagined. Colorado andMexico could also be included in an interstate market of evenlarger geographic scope. This larger market is an important areafor future research. New Mexico and Texas are currently the mostlikely trading partners. We term this an interstate marketing op-tion and assume that such a market arrangement will leave long-term Compact allocations and long-term property rights unaf-fected, but actual water use temporarily changes as water is tradedfor cash both within and across state lines. That is, interstatemarketing will not affect long-term Compact delivery obligations,but will create a short-term spot market that moves water tohigher-valued uses on a short-term basis.

If an interstate market is established in which water is tradedfor cash, water buyers will be users who pay out cash and receivewater put to a higher-valued incremental economic use; sellersreceive cash and forgo lower incremental economic benefits pro-duced by some existing uses. The buyer gains a value of watergreater than the cash paid, and the price received by the seller ishigher than the economic value of water forgone. Establishing aninterstate market will cause the difference to narrow �sometimesto zero� in incremental economic benefits of water use betweenbuyers and sellers. Water buyers will see falling incremental ben-
efits of water use compared to use patterns under the Law of the

River and therefore pay sellers to forgo water use. Water sellerslose some benefits of existing water use, keeping only enoughwater for higher marginal valued uses. Compensation by buyersfor sellers to forgo use results in a net benefit to both the buyerand seller �Table 6�.

Impacts of interstate marketing in Texas are more complex. Atthe 100% inflow level �full supply of water�, there is no transferfrom MRGCD agriculture, but El Paso agriculture transfers nearly50,000 acre-feet of consumptive use to reduce El Paso’s munici-pal pumping. At the 75% inflow level, 45% reductions inMRGCD consumptive use support reduced pumping by both Al-buquerque and El Paso. That is, it is cheaper for both cities to rentsurface water from MRGCD agriculture than to deal with short-ages through constant or increased pumping. Under this market-ing arrangement, EBID use is increased through increasedsurface-water use, compared to its loss of 43,000 acre-feet underthe Law of the River. EBID’s absolute use does not increase whenthe basin’s supplies fall from 100 to 75% of normal. However,when the basin’s supplies are 75% of normal, EBID’s use doesincrease under an interstate market compared to the Law of theRiver. This occurs because EBID’s high-valued agriculture canafford to buy imported surface water from lower-valued MRGCDagriculture. All parties gain from the trade. At the 50% inflowlevel, MRGCD consumptive use reductions are limited only by amodel constraint: even the highest-valued MRGCD use producesa lower economic value than other surface-water uses in thebasin. Use reductions by MRGCD support consumptive use in-creases by EBID and El Paso area agriculture and groundwaterpumping reductions by El Paso. At all inflow levels there is anincrease in surface-water diversions below Elephant Butte Reser-voir under the interstate marketing institution compared to theLaw of the River because of high-valued agricultural and M&Iuses in that part of the basin.

Discussion

Modeling the hydrology, economics, and policy for a complexriver basin is still in the experimental stage and we will need tosee considerable future development before there is in widespreaduse to support the design of improved policies. Model structure,simplifying assumptions, and limited data all introduce the poten-tial for errors.

This analysis measures only primary economic damage towater users resulting from drought and from institutional alterna-tives for coping with drought. Secondary impacts, such as re-duced trading with local business by water users who lose waterin a drought, are not counted. As described by Howe �1997�, notcounting these impacts amounts to assuming there are no realcosts of displaced resources faced by local economies when wateris reallocated at a significant scale, either by drought itself or bymitigating actions taken to cope with drought. In fact, these ad-ditional losses are often politically and economically important.To the extent these indirect impacts can be measured and valuedin a way that is consistent with direct impacts, they should beaccounted for in any comprehensive analysis of drought policy.

Our results depend on the choice of drought scenarios andexisting water supply conditions; other drought scenarios andwater supply conditions are possible. For this study, an importantassumption is the initial condition of a nearly depleted ElephantButte Reservoir, which actually occurred in 2004. If we designeddrought-coping measures beginning with a full reservoir, an im-portant result of this study would be reversed: central and north-
ern New Mexico would be immediately more vulnerable to

drought flows through Compact limits on use, while Texas useswould be protected for a number of years by carryover storage.

Our modeling approach only approximates the operation ofwater markets. We do not explicitly include transaction costs as-sociated with the design and operation of a water market. Impor-tant transactions costs of markets include the fixed costs of estab-lishing a water marketing system and the variable costs ofexpanding the scope of the market in the face of increasingdrought severity. Market institutions are modeled using an objec-tive maximizing the total benefits of water use. As such, themodel internalizes externalities, defined as impacts on third par-ties resulting from cash-for-water trades �Hanak 2003�. Designinga water market institution to protect third parties to water tradesmay require additional policy instruments. Conditioning watertrades on a lack of objections after the seller makes a publicannouncement in a local newspaper may be one way to protectthird parties while promoting economic efficiency in water mar-kets. Indeed, many existing prohibitions or strict limitations onmarket transfers are motivated by concern for return flows thatmight be lost �an externality� when water is transferred. Whileoptimization models such as ours easily include such reducedreturn flows in designing efficient allocations, implementing ac-tual markets to appropriately internalize externalities is a majorchallenge facing water researchers and policy analysts.

Another important limit of our approach is that only long-termaverage flows were considered in the model. The use of averageflows sharpens our ability to simplify impacts of complex watersupply patterns, but it also masks important impacts of year-to-year fluctuations around those long-term averages. Where stream-flows fluctuate widely from year to year, the presence of watermarkets produces a larger benefit in the driest of years comparedto benefits in more normal years. So the market institution takeson the role of a powerful institution for coping with risks pro-duced by unpredictable and uncertain year-to-year supply fluctua-tions. The impact of lagged flows on the effectiveness of themarket trading institutions is more pronounced with an increase inthe number of sequential dry years.

Finally, minimum flow levels in reaches of the river impactedby MRGCD diversions are required to support the endangeredsilvery minnow. While these minimum flow levels are also incor-porated in the model, biological research has not yet establishedthe timing or quantity of flow levels in the river required for theminnow’s survival. Because MRGCD’s consumptive use allowedunder the Compact depends on a number of large and potentiallypoorly understood depletions and interactions, model results of itsuse levels must be interpreted with caution.

Conclusions

In light of the emerging trends of population growth, agriculturaluse levels, and institutions in the Rio Grande Basin, we concludethat multimillion-dollar direct economic damage to water users islikely to occur when inflows to the Basin fall to between 50 and75% of their long-run averages. Under these conditions, stream-flows and reservoir storage volumes are dramatically reduced.The average economic damage can be as high as $100 per acre-foot of reduced inflows to the basin, but is more often in the rangeof $25 to $50 per acre-foot, reflecting damage to agricultural,municipal, and recreational users.

The development and use of an integrated hydrologic, eco-nomic, and institutional model to examine a range of alternative
policy responses finds considerable mitigation in drought damage

resulting from policies such as intrastate market water transfers.Impressive reductions in drought damage compared to whatwould occur under the existing Law of the River are possible if aninterstate market program is enacted. This market would effec-tively reallocate drought-induced water losses from higher-valueduses of water to lower-valued uses. These larger reductions indrought damage under an interstate market program would occuras agricultural water users discover they can reduce their droughtlosses by trading water for money whenever the income fromrented water is larger than the income produced by using the samewater in agriculture. In the face of intrastate markets, the cities ofAlbuquerque, New Mexico, and El Paso, Texas, would lose lesswater to drought than they could expect to lose under the Law ofthe River. In much the same way as occurred with the CaliforniaWater Bank established in the early 1990s, the presence of anintrastate water market in the Rio Grande Basin would consider-ably soften the economic blow to the region produced by drought.The institution of a water market has the effect of making themarginal benefits of water use more nearly equal among tradingpartners than would occur under the Law of the River.

Even larger reductions in drought damages are possible if theinstitution of an interstate water market can be established. How-ever, the Rio Grande Compact precisely allocates the share ofdrought-induced water shortages among the three states in law, inwater managers’ minds, and in popular culture. Combined with along-standing mutual suspicion among water users in the Basin,the difficulty of trading drought shortages across state lines willmake an interstate water market more difficult to establish andharder to implement than an intrastate water market. That is, be-cause of the importance of the Compact, considerably more insti-tutional flexibility will called for to allow water trades across statelines than within any single state. In this Basin, as in most othersin dry places, public policy changes rarely do not involve costs,and any introduction of water marketing across state lines will besubject to considerable debate and scrutiny.

The model and corresponding results assume sophisticated,perfectly competitive markets that internalize or include third-party effects such as changes in return flows from water transfers.However, real markets are not perfect and rarely incorporate thethird-party effects of transactions, either hydrologic or economic.Real markets are also influenced by other factors such as transac-tion costs, market power, and negotiating party personalities. Be-cause of this, real markets are not able to achieve the completebenefits of transfers assumed under perfect markets, and the ben-efits reported above will require adjustments to reflect imperfector real market conditions. This is an important area for futurestudy and policy implementation.

Despite the institutional challenges of introducing water mar-kets as a drought-coping measure, their potential economic gainsare considerable. Compared to existing water allocation institu-tions, we find that future drought damage could be reduced byone-fifth to one-third per year under the most serious droughtthrough intrastate and interstate water markets, respectively, thatwould extend across current water management jurisdictions.

Acknowledgments

The writers are grateful for generous and sustained financial sup-port by the Rio Grande Basin Initiative Project, the U.S. Geologi-cal Survey, and the agricultural experiment stations and water
resources research institutes of Colorado, New Mexico, and

Texas. Reviews by two anonymous referees and a coeditor im-proved this paper considerably.

References

Abu Zahra, B. A. A. �2001�. “Water crisis in Palestine.” Desalination,136�1–3�, 93–99.

Arthington, A. H., and Pusey, B. J. �2003�. “Flow restoration and protec-tion in Australian rivers.” River Research and Applications, 19�5–6�,377–395.

Barros, M. T., Tsai, F. T., Yang, S. L., Lopes, J. E., and Yeh, W. W.�2003�. “Optimization of large-scale hydropower system operations.”J. Water Resour. Plann. Manage., 129�3�, 178–188.

Booker, J. F. �1995�. “Hydrologic and economic impacts of drought underalternative policy responses.” Water Resour. Bull., 31�5�, 889–906.

Booker, J. F., and Young, R. A. �1994�. “Modeling intrastate and inter-state markets for Colorado River water resources.” J. Envir. Econom.Manage., 26�1�, 66–87.

Bouchart, F. J. C., and Goulter, I. C. �1998�. “Is rational decision makingappropriate for management of irrigation reservoirs?” J. Water Resour.Plann. Manage., 124�6�, 301–309.

Cai, X., McKinney, D. C., and Lasdon, L. S. �2003�. “Integratedhydrologic-agronomic-economic model for river basin management.”J. Water Resour. Plann. Manage., 129�1�, 4–17.

Christensen, N. S., Wood, A. W., Voisin, N., Lettenmaier, D. P., andPalmer, R. N. �2004�. “The effects of climate change on the hydrologyand water resources of the Colorado River basin.” Clim. Change,62�1–3�, 337–363.

Dalsted, N. L., Gutierrez, P. H., Nitchie, D., Sharp, R. L., and Tranel, J.�1996�. Selected 1995 crop enterprise budgets for Colorado, ColoradoState Univ. Cooperative Extension Service, Ft. Collins, Colo.

de Azevedo, L. G. T., Gates, T. K., Fontane, D. G., Labadie, J. W., andPorto, R. L. �2000�. “Integration of water quantity and quality instrategic river basin planning.” J. Water Resour. Plann. Manage.,126�2�, 85–97.

Elarabawy, M., Attia, B., and Tosswell, P. �1998�. “Water resources inEgypt: Strategies for next century.” J. Water Resour. Plann. Manage.,124�6�, 310–319.

El-Kady, M., and El-Shibini, F. �2001�. “Desalination in Egypt and thefuture application in supplementary irrigation.” Desalination, 136�1–3�, 63–72.

Eschenbach, E. A., Magee, T., Zagona, E., Goranflo, M., and Shane, R.�2001�. Goal programming decision support system for multiobjectiveoperation of reservoir systems.” J. Water Resour. Plann. Manage.,127�2�, 108–120.

Farah, E. A., Mustafa, E. M. A., and Kumai, H. �2000�. “Sources ofgroundwater recharge at the confluence of the Nile, Sudan.” Environ.Geol., 39�6�, 667–672.

GAMS Development Corp. �1998�. GAMS, A user’s guide, �http://www.gams.com/� �March 16, 2005�.

Guo, Z. W., Xiao, X. M., and Li, D. M. �2000�. “An assessment ofecosystem services: Water flow regulation and hydroelectric powerproduction.” Ecol. Appl., 10�3�, 925–936.

Haddadin, M. J. �2002�. “Water in the Middle East peace process.” Geo-graphical Journal, 168, 324–340.

Hamilton, J. R., Green, G. P., and Holland, D. �1999�. “Modeling thereallocation of Snake River water for endangered salmon.” Am. J.Agric. Econom., 81�5�, 1252–1256.

Hanak, E. �2003�. Who should be allowed to sell water in California?Third party issues and the water market, Public Policy Institute ofCalifornia, San Francisco.

Heckellei, T., and Britz, W. �2000�. “Concept and explorative applicationof an EU-wide, regional agricultural sector model �CAPRI project�.”Paper read at 65th EAAE-Seminar, Univ. of Bonn, March 29-31, atBonn, Germany.

Holland, S. P., and Moore, M. R. �2003�. “Cadillac Desert revisited:


Property rights, public policy, and water-resource depletion.” J. Envir.Econom. Manage., 46�1�, 131–155.

Howe, C. W. �1997�. “Increasing efficiency in water markets: Examplesfrom the Western United States.” Water Marketing—The Next Gen-

eration, T. L. Anderson and P. J. Hill, eds., Rowman and Littlefield,New York.

Howitt, R. E. �1995�. “Positive mathematical programming.” Am. J.Agric. Econom., 77, 329–342.

Hurd, B. H., et al. �2004�. “Climate change and U.S. water resources:From modeled watershed impacts to national estimates.” J. Am. WaterResour. Assoc., 40�1�, 129–148.

Hurd, B. H., Callaway, J. M., Smith, J. B., and Kirshen, P. �2002�. “Eco-nomic effects of climate change on U.S. water resources.” WaterResources and Climate Change, K. Frederick, ed., Edward Elgar,Cheltenham, U.K.

Jagerskog, A. �2003�. “The power of the ‘sanctioned discourse’ —A cru-cial factor in determining water policy.” Water Sci. Technol., 47�6�,161–166.

Keogh, D. U. �2004�. “Context evaluation: A profile of irrigator climateknowledge, needs and practices in the northern Murray-Darling Basinto aid development of climate-based decision support tools and infor-mation and dissemination of research.” Australian Journal of Experi-mental Agriculture, 44�3�, 247–257.

Kotb, T. H. S., Watanabe, T., Ogino, Y., and Tanji, K. K. �2000�. “Soilsalinization in the Nile Delta and related policy issues in Egypt.”Agric. Water Manage., 43�2�, 239–261.

Lee, D. J., Howitt, R. E., and Mariño, M. A. �1993�. “A stochastic modelof river water quality: Application to salinity in the Colorado River.”Water Resour. Res., 29�12�, 3917–3923.

Levy, B. S., and Baecher, G. B. �1999�. “NileSim: A Windows-basedhydrologic simulator of the Nile River basin.” J. Water Resour. Plann.Manage., 125�2�, 100–106.

Li, H. M., Waley, P., and Rees, P. �2001�. “Reservoir resettlement inChina: Past experience and the Three Gorges Dam.” GeographicalJournal, 167, 195–212.

Liu, J. Y., Liu, M. L., Zhuang, D. F., Zhang, Z. X., and Deng, X. Z.�2003�. “Study on spatial pattern of land-use change in China during1995–2000.” Sci. China, Ser. D: Earth Sci., 46�4�, 373.

Liu, X. P., Kelin, W., and Geli, Z. �2004�. “Perspectives and policies:Ecological industry substitutes in wetland restoration of the MiddleYangtze.” Wetlands, 24�3�, 633–641.

Loaiciga, H. A., and Leipnik, R. B. �2000�. “Closed-form solution forcoastal aquifer management.” J. Water Resour. Plann. Manage.,126�1�, 30–35.

Lund, J. R., and Guzman, J. �1999�. “Derived operating rules for reser-voirs in series or in parallel.” J. Water Resour. Plann. Manage.,125�3�, 143–153.

Mahmoud, M. R., and Garcia, L. A. �2000�. “Comparison of differentmulticriteria evaluation methods for the Red Bluff diversion dam.”Environ. Modell. Software, 15�5�, 471–478.

Martinez, V. S., Judez, J., and Chaya, C. �1999�. “Effects of the measuresenvisaged in ‘Agenda 2000’ on arable crop producers, beef and vealproducers and dairy farms: An application of PMP to representativefarms of a Spanish region.” Proc., 2nd Eurotools Seminar, Univer-sidad Politecnica, Madrid.

Michelsen, A. M., and Bargur, J. �1993�. “Macro-economic based mul-tiple objective analysis for regional water resources planning, Devel-opment and Management in North China.” Proc., United NationsInter-regional Seminar on Technologies for Applications to WaterResources Policy, Development and Management, United NationsDepartment for Development Support and Management and Environ-ment Program New York, and Ministry of Water Resources and StateScience and Technology Committee, Beijing.

Michelsen, A. M., and Cortez, F. �2003�. Drought watch on the RioGrande, Texas A&M Univ. System and U.S. Bureau of Reclamation,
El Paso, Tex.

Michelsen, A., McGuckin, T., and Stumpf, D. �1998�. Effectiveness ofresidential water conservation price and non-price programs in urbanareas in the western U.S., American Water Works Association Re-search Foundation, Denver.

Middle Rio Grande Water Assembly. �2001�. The Middle Rio Granderegional water plan, �http://www.waterassembly.org/5theplan/5.4theplan_The%20Water%20Plan.html� �March 16, 2005�.

Mimi, Z. A., and Sawalhi, B. I. �2003�. “A decision tool for allocating thewaters of the Jordan River basin between all riparian parties.” WaterResour. Manage., 17�6�, 447–461.

Nakamura, T. �2003�. “Ecosystem-based river basin management: Its ap-proach and policy-level application.” Hydrolog. Process., 17�14�,2711–2725.

Newlin, B. D., Jenkins, M. W., Lund, J. R., and Howitt, R. E. �2002�.“Southern California water markets: Potential and limitations.” J.Water Resour. Plann. Manage., 128�1�, 21–32.

Oamek, G. E. �1990�. “Economic and environmental impacts of interstatewater transfers in the Colorado River Basin.” Monograph 90-M3,Iowa State Univ., Center for Agricultural and Rural Development,Ames, Iowa.

Parker, J. A. �2002�. Memorandum Opinion and Findings of Fact andConclusions of Law, United States District Court for the District ofNew Mexico. CIV No. 99-1320 JP/RLP-ACE. September 23, 2002.�http://www.spa.usace.army.mil/hottopics/esa_hot_issues.htm� �No-vember 21, 2004�.

Paso del Norte Water Task Force. �2001�. Water planning in the Paso delNorte: Toward regional cooperation El Paso, �http://www.sharedwater.org/en/Projects/Reports/water_taskforce_ report_revl.pdf��March 16, 2005�.

Philbrick, C. R., and Kitanidis, P. K. �1999�. “Limitations of deterministicoptimization applied to reservoir operations.” J. Water Resour. Plann.Manage., 125�3�, 135–142.

Quiggin, J. �2001�. “Environmental economics and the Murray-Darlingriver system.” Australian Journal of Agricultural and Resource Eco-nomics, 45�1�, 67–94.

Rangarajan, S., Surminski, H., and Simonovic, S. P. �1999�. “Reliability-based loss functions applied to Manitoba hydro energy generationsystem.” J. Water Resour. Plann. Manage., 125�1�, 34–40.

Ravikumar, V., and Venugopal, K. �1998�. “Optimal operation of southIndian irrigation systems.” J. Water Resour. Plann. Manage., 124�5�,264–271.

Reid, M. A., and Brooks, J. J. �2000�. “Detecting effects of environmentalwater allocations in wetlands of the Murray-Darling Basin, Australia.”Regul. Rivers: Res. Manage., 16�5�, 479–496.

Schmandt, J. �2002�. “Bi-national water issues in the Rio Grande/RioBravo basin.” Water Policy, 4�2�, 137–155.

Shabman, L., and Stephenson, K. �2000�. “Environmental valuation andits economic critics.” J. Water Resour. Plann. Manage., 126�6�,
382–388.

Shuval, H. I. �2000�. “Are the conflicts between Israel and her neighborsover the waters of the Jordan River Basin an obstacle to peace? Israel-Syria as a case study.” Water, Air, Soil Pollut., 123�1–4�, 605–630.

Sperow, M. �1998�. “Three essays addressing production economics andirrigation: Managing for drought and pests transported in irrigationwater.” Ph.D. dissertation, Dept. of Agricultural and Resource Eco-nomics, Colorado State Univ., Fort Collins, Colo.

Strzepek, K. M. �2000�. “Responses and thresholds of the Egyptianeconomy to climate change impacts on the water resources of the NileRiver.” Clim. Change, 46�3�, 339–356.

Tidwell, V. C., Michelsen, A. M., Aparicio, J., and Passell, H. D. �2004�.“The Rio Grande/Rio Bravo—A river of change.” Water ResourcesImpact, 6�3�, 14–17.

U.S. Dept. of Interior. �2001�. Programmatic biological opinion on watermanagement in the Middle Rio Grande New Mexico, Fish and Wild-life Service, Washington, D.C.

United Nations Development Program. �1994�. “Water Resources Man-agement in North China.” CPR/88/068, Executive Summary and Vol.1-2 and Appendices A–M. Beijing, Peoples Republic of China. NorthChina Research Center, Institute of Water Resources and HydropowerResearch and State Science and Technology Commission.

Vaux, H. J., and Howitt, R. E. �1984�. “Managing water scarcity: Anevaluation of interregional transfers.” Water Resour. Res., 20�7�,785–792.

Wagner, B. J. �1999�. “Evaluating data worth for groundwater manage-ment under uncertainty.” J. Water Resour. Plann. Manage., 125�5�,281–288.

Ward, F. A., and Michelsen, A. �2002�. “The economic value of water inagriculture: Concepts and policy applications.” Water Policy, 4�De-cember�, 423–446.

Ward, F. A., Cole, R. A., Deitner, R., and Green-Hammond, K. A. �1997�.“Limiting environmental program contradictions: A demand systemsapplication to fishery management.” Am. J. Agric. Econom., 79�3�,803–813.

Ward, F. A., et al. �2001�. Institutional adjustments for coping with pro-longed and severe drought in the Rio Grande Basin, �http://wrri.nmsu.edu/publish/techrpt/tr317/� �November 22, 2004�.

Watkins Jr., D. W., and McKinney, D. C. �1999�. “Screening water supplyoptions for the Edwards aquifer region in central Texas.” J. WaterResour. Plann. Manage., 125�1�, 14–24.

Weinberg, M., Kling, C. L., and Wilen, J. E. �1993�. “Water markets andwater quality.” Am. J. Agric. Econom., 75�2�, 278–291.

Willis, D. B., and Whittlesey, N. K. �1998�. “Water management policiesfor streamflow augmentation in an irrigated river basin.” Journal ofAgricultural and Resource Economics, 23�1�, 170–190.

Yan, T., and Qian, W. Y. �2004�. “Environment migration and sustainabledevelopment in the upper reaches of the Yangtze River.” Popul. En-
viron., 25�6�, 613–636.

integrated economic, hydrologic, and institutional ...water-research.nmsu.edu › files › 2016 ›...

Documents