Volume 40, Issue 4
Water Policy, Economics, and Systems Analysis
Free Access

Economic efficiency and cost implications of habitat conservation: An example in the context of the Edwards Aquifer region

Dhazn Gillig

Dhazn Gillig

Department of Agricultural Economics, Texas A&M University, College Station, Texas, USA

Search for more papers by this author
Bruce A. McCarl

Bruce A. McCarl

Department of Agricultural Economics, Texas A&M University, College Station, Texas, USA

Search for more papers by this author
Lonnie L. Jones

Lonnie L. Jones

Department of Agricultural Economics, Texas A&M University, College Station, Texas, USA

Search for more papers by this author
Frederick Boadu

Frederick Boadu

Department of Agricultural Economics, Texas A&M University, College Station, Texas, USA

Search for more papers by this author
First published: 30 April 2004
Citations: 6

Abstract

[1] Groundwater management in the Edwards Aquifer in Texas is in the process of moving away from a traditional right of capture economic regime toward a more environmentally sensitive scheme designed to preserve endangered species habitats. This study explores economic and environmental implications of proposed groundwater management and water development strategies under a proposed regional Habitat Conservation Plan. Results show that enhancing the habitat by augmenting water flow costs $109–1427 per acre-foot and that regional water development would be accelerated by the more extreme possibilities under the Habitat Conservation Plan. The findings also indicate that a water market would improve regional welfare and lower water development but worsen environmental attributes.

1. Introduction

[2] The Edwards Aquifer (EA) underlies the San Antonio area of Texas and is the source of water for more than two million people. Edwards waters are used by agricultural, municipal, industrial, recreational, and environmental interests. The EA naturally discharges through springs (mainly at Comal and San Marcos Springs) that in turn support habitat for endangered species [Longley, 1992]. These springs also provide a substantial proportion of the Guadalupe River base flow, and important freshwater inflows to Calhoun Bay and the Gulf of Mexico.

[3] EA use has proven to be a source of environmental concern. Groundwater pumping has grown on the order of 1% a year, and as usage has increased, aquifer elevation fluctuation has increased while springflows has declined [Collinge et al., 1993]. Diminishing springflows worsen the suitability of spring fed habitat for endangered species (these species include the Comal Springs Dryopid Beetle, Comal Springs Riffle Beetle, Fountain Darter, Peck's Cave Amphipod, San Marcos Gambusia, Texas Wild Rice, Texas Blind Salamander, San Marcos Salamander, Cagle's Map Turtle, and Terresstrial Karst Invertebrates [Bio-West Inc., 2001]) along with downstream ecosystems in the Guadalupe River and Calhoun Bay.

[4] As a result of a lawsuit filed by the Sierra Club and the Guadalupe-Blanco River Authority, different levels of pumping limits were placed on EA water usage to protect springflows [Sierra Club et al. versus Bruce Babbitt et al., 1996] (see also Senate Bill 1477). Senate Bill 1477, passed by Texas Legislature in 1993, limits current pumping for five counties in the EA region (Uvalde, Medina, Bexar, Comal, and Hays) to 450,000 acre feet (acft) annually and requires restriction to 400,000 acft by 2008 to insure that minimum springflows of the Comal and San Marcos Springs are maintained at desirable levels to protect endangered and threatened species. The provisions of SB 1477 are just coming into effect in 2003–2004. Besides pumping limits, temporary programs such as the irrigation suspension program [Keplinger and McCarl, 2000], aquifer elevation dependent water usage restrictions, and a drought management plan have been implemented to protect springflows.

[5] Simultaneously, Texas created the Edwards Aquifer Authority (EAA) to manage EA water use. The EAA is legislatively charged with springflow protection and is developing a Habitat Conservation Plan (HCP) designed to protect federally listed threatened or endangered species inhabiting the EA, along with Comal and San Marcos Springs. The hydrological component of HCP that was used to assess biological risks found that it might be necessary to restrict EA water use to as low as 175,000 acft to protect the species [Bio-West Inc., 2001].

[6] Furthermore, the EAA has begun replacing the traditional rule of capture (where landowners have basically unlimited access to waters under their land) with a permit system coupled with a water market. Only agricultural, municipal, and industrial users are being assigned permits. Recreational and environmental interest groups will not be eligible. Initial agricultural permits are expected to amount to 230,000 acft, exceeding 1998 agricultural consumption while initial municipal and industrial (M&I) annual permits are expected to amount to 220,000 acft (100,000 acft below 1998 consumption). Seniority will not be assigned. Theoretically, a well established permit system will encourage the emergence of water markets that lead to efficient allocation of water resources by allowing water transfers from lower-valued use to higher-valued use [Anderson and Hill, 1997; Characklist et al., 1999]. Water market activity transferring water intersect orally has begun even in advance of the permit issuance.

[7] Drought conditions and population pressures in the Western region of the United States have historically induced changes in water allocation and management. Rights changed from Riparian to appropriative in the 1800s [Hundley, 2001]. In the last two decades, water marketing has become common with water reallocation using market mechanisms being employed extensively in Colorado, Nevada, California, New Mexico, and Texas [Saliba and Bush, 1987]. The goal in resorting to market-based allocation of water has been to increase efficiency, that is, by transferring water from lower to higher-valued uses, and also to encourage conservation. However, transfers through marketing have been known to have several negative unintended consequences. For example, in California, a major drought in 1991 led to institutional changes that declared water transfers to be a beneficial use. During this drought period, California started trading water with an offer price of $175 per acre foot made by municipalities to irrigators [Howitt, 1998]. At this offered price, response by irrigators was unexpectantly high leading to an excess of water of 264,000 acft. Water marketing activities that ensued yielded over $100 million in net benefits [Griffin, 2004]. However, it was also found that idling lands to enable farmers to lease water to other uses also had adverse impacts on both the farm input and output industries, and in effect, hurt the local economy [Griffin, 2004]. Most of the water marketing activity to date has involved surface water.

[8] Property rights in groundwater and the transferability of these rights is another item of concern today. Unlike surface water transfers, groundwater transfers present their own unique difficulties. One of the well-known groundwater marketing activities is found in Southern Arizona where groundwater use constitutes about 60% of all water use. Historically, irrigated agriculture used over 85 to 90% of the annual water supply, but today the trend is water transfers out of agriculture to urban users. Also, there is considerable groundwater marketing activity in the Gila-San Francisco Basin in New Mexico. Here also, what one finds is a transfer of water from irrigation uses to urban uses, with Silver City purchasing most of irrigation water rights [Saliba and Bush, 1987]. These trends are also observed within the Edwards Aquifer region of Texas.

[9] Unlike the southwestern states, those in the northeast (North Carolina and New York, for example) have relied more on state planners in water management. Florida's effort at reforming its water sector presents a contrasting situation to what one finds in the Southwestern states. In response to severe droughts in 2000 and 2001, policy makers in Florida and the public have been engaged in debate on options for institutional response to growing water needs against a backdrop of dwindling supplies. The approaches pursued in the Southwestern region that have been based largely on water marketing and transfers have been argued by one author to be unsuitable for Florida, the water marketing and banking efforts initiated in western states such as Texas, California, and Kansas are not easily adaptable to Florida's water use permitting system. While the prior appropriation system in most western states is very conducive to reallocating water rights through water marketing and water banking, implementing similar reallocation systems in Florida would require reform of Florida's administrative water permitting system [Fletcher, 2002]. The approach advocated for Florida retains central control of the system, while encouraging cooperative water transfer agreements between local authorities similar to the approach followed in New York [Fletcher, 2002]. This brief review has been presented to point out that water marketing, and especially groundwater is not available in all regions of the United States. Furthermore, states' responses to water demand and supply pressures differ based on the existing institutional matrix within which water policies are made.

[10] Besides water marketing, an active regional planning effort is ongoing. The South Central Texas Regional Water Planning Group (SCTRWPG), created under the auspices of a water planning initiative authorized by Texas Senate Bill 1, has defined a number of water development alternatives to meet a projected regional shortage [see Gillig et al., 2001].

[11] This paper investigates how HCP implementation affects water markets, demand for water development, regional resource allocation, economic stability, and environmental benefit. Specifically, the analysis will focus on (1) the economic and environmental consequences of a set of pumping limits suggested under the HCP; (2) the implications of the emergence of a regional water market; (3) the interrelationship between HCP and likely regional water development; (4) tradeoffs between efficiency in water allocation and environmental costs; and (5) a model deriving “optimal” ways to achieve HCP goals.

2. Methodology

[12] This study uses an expansion of an integrated EA groundwater and river system, hydrological and economic simulation model (EDSIMR) described by Gillig et al. [2001] as its principal analytical tool. This study introduces a detailed depiction of water marketing and water development alternatives.

2.1. General Scope of EDSIMR

[13] EDSIMR was developed to analyze the economic and hydrological implications of cooperative water management alternatives for regional ground and surface waters in the South Central Texas Region [McCarl et al., 1999; Gillig et al., 2001]. It depicts supply and use of regional groundwater (in the EA and other Aquifers along the river systems) and surface water (in the Guadalupe, Blanco, San Antonio, Nueces, and Frio Rivers). Users include cities, industries, and agriculture. The model also depicts price dependent demand, return flow, recharge augmentation, reuse, water market rights, springflows, river flows, bay/estuary inflows, EA elevation levels, and major impoundment management (Corpus Christi, Choke Canyon, Medina, Diversion, and Canyon Lakes). It also depicts construction of additional water development projects as discussed below.

[14] EDSIMR incorporates uncertainty using a two-stage stochastic programming with recourse [Dantzig, 1955] or discrete stochastic programming [Cocks, 1968] formulation considering variability in recharge and crop yields. EDSIMR operates across a 10-state weather and recharge empirical probability distribution based on the EA history for the period 1934 to 1996. To form the probability distribution, the data was first ordered from most dry to most wet recharge years. These historical recharge years were organized into 10 groups. The frequency for each group was counted and divided by the total number of recharge years (63 years) yielding a probability for each state of nature. A marker year was then selected to represent each state of nature. The marker years are 1956 (annual recharge at 43,758 acft), 1951, 1963, 1989, 1952, 1996, 1974, 1976, 1958, and 1987 (annual recharge at 2,003,643 acft). Crop yields also vary by state of nature. The Erosion Productivity Impact Calculator (EPIC) simulation model was used to develop associated crop yield data as influenced by state of nature [Williams et al., 1989]. EDSIMR is a price-endogenous, mathematical program implemented using the General Algebraic Modeling System (GAMS) [Brooke et al., 1992]. The model maximizes expected net benefits (benefits minus costs) of water use by municipal, industrial, and agricultural interests. Municipal and industrial benefits from water use are estimates of consumer surplus using constant elasticity municipal and industrial demand curves developed by Griffin and Chang [1991] and Renzetti [1988], respectively. Agricultural sector benefits are the net agricultural income derived from irrigated and dryland crop production (embodying the assumption of perfectly elastic demand for agricultural products). This study assumes that the agricultural products in the EA region have no influence on market prices; in other words, market prices are independent of production fluctuations by farmers in the EA region and thus farmers sell whatever they produce at a fixed market price, simply a horizontal line. Farm welfare therefore is equal to the revenues from selling agricultural products minus the production costs (see Appendix A). Because we did not have data on recreational and environmental benefits from springflows, these benefits are excluded from the consumer surplus, but they are the motivation behind the usage caps, as argued by McCarl et al. [1999], and need to return benefits at least as large as the cost to pumping users for overall welfare to have the potential to increase. As noted by a reviewer, the economic welfare associated with the recreational and environmental uses (the ecological benefits derived from springflows) are not zero. Incorporating the ecological benefits would involve a nonmarket valuation method such as travel cost or contingent valuation methods as discussed by Hanley et al. [1997], Loomis [1998], Barrens et al. [2000], and Eiswerth et al. [2000]. This omission leads to an underestimation of the consumer surplus derived from water uses that in turn underestimates the total welfare. However, as discussed by McCarl et al. [1999], the welfare measures from this study could be viewed as a lower bound on what the recreational and springflow values would need to be in order to have a potential welfare gain from the pumping restrictions.

[15] The model defined total water delivery costs to include pumping, diversion and distribution costs, along with new water development costs. The annual per acre foot of developed water project is derived from dividing the 20-year amortized water development costs with the expected capacity generated by that water project. The water development costs are obtained from the South Central Texas Regional Water Plan (1999 and 2000). These costs include fixed costs of constructing facilities and other nonstructural costs relating to construction activities, variable costs of operation and maintenance, and other costs. (See The 2001 Water Regional Plan, Appendix A: Cost estimating procedures, at http://www.watershedexperience.com/2001_download.html for details on the water development costs). The per acre foot of pumping costs refer to a cost to pump EA groundwater. These costs are specified as a function of pumping lift determined by J17 and Sabinal wells ending elevation levels. To compute the water market price for the permanent transfer of a pumping right, a change in the regional welfare under water market operation from the BASE is divided by the amount of water transfers.

[16] The new water development costs consist of fixed and variable operating costs that are incurred if the alternative is adopted. A transactions cost for water transfers is added when water markets are active. The objective function probabilistically weights the annual benefits that vary by state of nature (see Appendix A for a mathematical description).

[17] Water supply comes from regional groundwater and/or surface water depending upon availability at a particular site. The EA is subdivided into east and west while 52 associated river reaches are modeled. Municipal use is further divided into discretionary and nondiscretionary uses in Bexar County (where San Antonio is located). Agricultural use and demand is developed using regional linear programming models in 65 river or aquifer reaches. Each agricultural use alternative can pursue irrigation or revert to dryland farming depending upon water prices and availability.

[18] EDSIMR estimates water use on a monthly basis with river flows incorporated in a network flow model component depicting 52 river/reservoir reaches in the Nueces, San Antonio, and Guadalupe River Basins. The nodes represent points of inflow/outflow, diversion, and river confluence as well as reservoirs and aquifers. The links represent surface water flows along river reaches, into reservoirs or out of aquifers. Links are characterized by levels of instream flows, channel loss, return flow, reservoir leakage and evaporation. The network and hydrological characteristics were specified using data from U.S. Geological Survey (USGS); HDR Engineering, Inc. [HDR Engineering, Inc., and Geraghty & Miller, Inc., 1991; HDR Engineering, Inc., and Espey, Huston and Associates, Inc., 1993; HDR Engineering, Inc., 1999]; Texas Natural Resource Conservation Commission; Texas Water Development Board [1996]; Nueces River Authority, San Antonio Water System; Guadalupe Blanco River Authority; Bexar-Medina-Atascosa River Authority; and Zavala-Dimmit Water Improvement District 1. The data used involved projected groundwater and surface water demands; naturalized and gauged streamflow; river or reservoir evaporation; return flow; reservoir storage capacity; EA recharge and discharge; and Carrizo Aquifer recharge and discharge. Municipal water demand dependence on state of nature was developed based on the Griffin and Chang [1991] estimate of the climate elasticity.

[19] The network model depicts river system and EA hydrological interconnections, including (1) aquifer recharge from surface water where a given amount of surface flow over certain locations enters the aquifer largely through a fractured limestone outcropping, (2) springflow discharge where waters in the aquifer flow into rivers mainly at San Marcos and Comal springs, the level of which depends on aquifer level and in turn recharge, and pumping water use, (3) treated water return flow which is a function of municipal and industrial usage, and (4) agricultural water return flow which is a function of usage and irrigation system. EDSIMR predicts springflow and ending EA elevation using four linear functions as developed by McCarl et al. [1999] which express these items as a function of recharge, agricultural pumping, and M&I pumping:
equation image
equation image
equation image
equation image
where
  • r
  • 10-state probability distribution of aquifer recharge;
  • Comalrm
  • Comal springflow under state of nature r during month m;
  • San Marcosrm
  • San Marcos springflow under state of nature r during month m;
  • End J17r
  • ending water level elevation of the eastern reference well named J17 under state of nature r;
  • End Sabinalr
  • ending water level elevation of western reference well (named Sabinal) under state of nature r;
  • J17r
  • beginning water level elevation of well J17 under state of nature r;
  • Sabinalr
  • beginning water level elevation of Sabinal well under state of nature r;
  • Rechargerm
  • total recharge into the EA under state of nature r up until month m;
  • AgPumprm
  • total agricultural pumping from the EA under state of nature r up until month m;
  • M&Ipumprm
  • total municipal and industrial pumping from the EA under a state of nature r up until month m.
  • [20] The model contains constraints on ground/surface water demand and supply availability (both existing and newly developed supplies), agricultural production activities, pumping lifts, springflow minimum limits, water use maximum limits, nonnegativity conditions for decision variables, and binary restrictions for water development decisions (see Appendix A).

    2.2. Incorporation of Water Marketing

    [21] Water marketing mechanisms are depicted in EDSIMR allowing EA groundwater to be transferred between Agricultural and M&I parties. However, water transfers (purchase or sale of water) are limited by water rights (permits) on a county basis and the rule imposed by the SB 1477 that at least one acre-foot of irrigation water rights must be retained on each land parcel.

    [22] EDSIMR simulates water market operation annually. Two type of water transfers are traded: single year leases and permanent water rights transfers. The permanent transfers involve permanent conversion of land back to dryland or land retirement, while the one year lease involves single year land use change from irrigated to dryland. The decision among user groups whether to buy, lease, or sell water rights depends on economic benefits including opportunity costs and transactions cost. If for a municipality the benefits derived from an acre foot of non agricultural water consumption are greater than costs of buying water rights which equal the farm value of water plus transactions costs, then water rights are acquired. If for a farm the benefits of acre foot of water from leasing or selling water rights exceeds the revenues from crop production by more than the transactions cost of conveyance, then leasing or selling water rights makes economic sense. Similarly, if the benefits derived from leasing water just when conditions create a need are greater than those derived from a permanent acquisition, then leasing water rights makes more economic sense. Economic welfare of water market is presented in Figure 1.

    Details are in the caption following the image
    Economic welfare of a water market. A water market shifts the EA water supply curve in the M&I sector outward to EA1 by the amount of water transfers, Q1 − Q0, and vice versa in the agricultural sector. The shift in the EA water supply increases the M&I sector surplus by area M4 + M5 but decreases the agricultural sector income by area A5 + A6. However, the agricultural sector earns additional income from trade by area A4 + A5 + A6, which in turn brings this sector to experience a net gain of area A4. On the other hand, the M&I sector faces an additional selling and purchasing cost of area M5, yet it still captures a net gain of area M4. The total net effect of water market would be a gain of area A4 + M4.

    2.3. Incorporation of Water Development

    [23] Restrictions on water use from the EA along with demand growth are likely to stimulate regional water development. The recent SCTRWPG report forecasts development needs for an additional 450,000 acft by 2050. Adoption of the more severe suggested HCP measures could accelerate the need for these alternatives. Thus, following Gillig et al. [2001], EDSIMR was expanded to allow development of eleven water management options identified by the SCTRWPG as viable water development options to meet demand by 2012 (see South Central Texas Regional Water Planning Group, 2001, http://www.watershedexperience.com). These options involve interbasin water transfers, EA recharge and storage enhancement, and water reuse. EDSIMR depicts these as integer variables with accompanying amortized fixed development costs, per acre-foot operating costs, and capacities. These integer variables depict a binary decision as to whether or not to develop each potential source in response to water scarcity. The resultant model in effect compares the marginal benefits of the developed water with the annualized fixed costs of development (see Appendix A).

    2.4. Base Model Specification

    [24] The EDSIMR version used here simulates water supply and use for the year 2012 (the year HCP conditions must be met). Municipal and industrial water demands were projected based on Texas Water Development Board estimates. Agricultural price projections were drawn from Food and Agricultural Policy Research Institute (FAPRI) documents using the 2010 data [FAPRI, 2001]. Permit allocations were specified based on the proposed EAA permit allocation as of June 2000 [J. L. Wilson and Associate, 2000]. The rest of the model was specified as discussed by Gillig et al. [2001], McCarl et al. [1999], and Appendix A. EDSIMR was validated by comparing the simulated EA groundwater and surface water demand (consumption) against the 1996 observed data under the unrestricted pumping scenario. EDSIMR results matched the 1996 observed pumping use data as reported by Gillig et al. [2001].

    3. Water Management Simulation and Results

    [25] Eight scenarios were simulated as summarized in Table 1. These scenarios involve combinations of (1) presence of water market, (2) level of water pumping restrictions with no limit, usage limitation to SB 1477 400,000 acre foot limit, usage limit to HCP biologically suggested limits of (350,000 or 175,000 acre foot limits), or springflow based limitation restricting usage by insuring springflow at or above U.S. Fish and Wildlife Service (USFWS) jeopardy level, and (3) availability or non availability of the new water development strategies identified by the SCTRWPG. All scenarios operate under the EAA permit allocation which currently exceeds 400,000 acre feet. The base model had unlimited total water use, no new water development policies and no water market (the pre SB 1477 condition that persisted until at least 2000).

    Table 1. Water Management Scenarios Examined
    Management Policy Policy Description
    Pumping Limitation Water Market Operation New Water Development
    Base none none none
    LIM400 400,000 acft. (SB 1477) none none
    LIM340 340,000 acft. (HCP high) none none
    LIM175 175,000 acft. (HCP low) none none
    MK400 400,000 acft. (SB 1477) yes none
    MK340 340,000 acft. (HCP high) yes none
    MK175 175,000 acft. (HCP low) yes none
    MK400WD 400,000 acft. (SB 1477) yes yes
    MK340WD 340,000 acft. (HCP high) yes yes
    MK175WD 175,000 acft. (HCP low) yes yes
    MKSPRWD Springflow 200 acft.a yes yes
    • a Springflow limits set at the U.S. Fish and Wildlife jeopardy level.

    [26] The simulation is carried out as follows. First, only the reduced total water pumping limits are considered to determine the impact of water pumping limits relative to the historic EAA permit constrained BASE. Second, both water pumping limits and water market operation are jointly examined to determine the incremental impact of EAA induced water market operations along with water pumping limits. Third, reduced total water pumping limits, water market operation, and water development are jointly considered to determine the incremental impact of water development. Fourth, water usage limits that are not a fixed usage quantity are examined based on USFWS springflow limits that allow water use to vary with recharge quantity. These are examined in conjunction with water market operation, and water development and are used to determine the impact of springflow limits relative to water pumping limits.

    3.1. Water Pumping Limits

    [27] Table 2 shows the impacts of the water pumping limits (LIM400, LIM340, and LIM175, columns 4–6) relative to the BASE (column 3). The results indicate that the adoption of the reduced total pumping management regimes cause substantial income or welfare losses for regional water users, in particular, M&I users. M&I users lose $247 million per year under LIM400 with the loss rising to $664 million per year under LIM175 (subtracting M&I welfare in columns 4 and 6 from the BASE). Agricultural users marginally gain because of the size of their EAA granted permit allocation and because of a reduction in pumping costs due to an increase in EA elevation resulting from the imposition of pumping limits.

    Table 2. Economic and Environmental Simulation Results With and Without a Water Market in Operation Under Various Pumping Limits Under Projected 2012 Water Demand Without New Water Development
    Unit Base Water Management Scenarios Examined
    LIM400 LIM340 LIM175 MK400 MK340 MK175
    Economic welfare
      Agricultural income $106 23.11 24.42 24.61 24.26 22.07 23.12 23.81
      Agricultural income+a $106 23.11 24.42 24.61 24.26 28.54 28.09 24.42
      M&I welfare $106 953.56 706.32 622.36 289.78 894.53 795.32 289.87
      Total welfare $106 976.67 730.74 646.96 314.04 923.07 823.41 314.28
    Environmental attributesb
      J-17 elevation Feet 552.73 656.76 671.22 735.96 629.45 657.95 735.78
      Sabinal elevation Feet 656.07 772.01 788.44 861.39 740.44 772.75 861.28
      Comal flow cfs 0.00 237.32 321.54 690.06 68.93 234.73 689.07
      San Marcos flow cfs 20.51 85.92 95.10 134.98 67.13 85.24 134.86
      Calhoun Bay flow 103 acft 292.42 458.11 503.86 701.67 344.72 369.64 442.76
      Corpus Bay flow 103 acft 175.38 102.50 102.50 102.44 99.15 99.12 99.12
    Water consumption
      EA agricultural pump 103 acft 83.49 86.96 86.93 75.74 75.74 75.74 75.74
      EA M&I pump 103 acft 468.98 252.59 222.21 99.26 324.26 264.26 99.26
      Water transfers 103 acft 119.42 90.15 9.64
      Value of water market $106 192.33 176.44 0.24
    • a Agricultural income + refers to farm income plus payments from water transfers.
    • b The elevations, springflows, and bay flows reported are those arising observed under the EA recharge drought conditions.

    [28] On the other hand, the reduced total water pumping limits improves the regional environment through the period 2012. Comal springflow rises to 237 cubic feet per second (cfs) under the LIM400 and goes up even further under tighter pumping limits. These model results show that EA pumping limits as low as 175,000 acft are not needed to assure habitat protection but suffer from a model characteristic. Namely, EDSIMR sets initial water levels to the average of ending water levels so cannot simulate the effect of a multiyear drought (simulations under a repeat of a severe multiyear drought as observed in the 1950s gave rise to the 175,000 acft pumping limit under the HCP studies). See R. L. Williams et al. (Elevation dependent management of the Edwards Aquifer: A linked mathematical and dynamic programming approach, submitted to Department of Agricultural Economics, Texas A&M University, 2003) for an alternative approach. If the current EAA permitted regime (the base) is maintained at the level substantially above 400,000 acre feet, Comal springflow falls to zero under many rainfall/recharge states of nature. This result implies that pumping restrictions are needed to avoid compromising the endangered species habitat in the face of anticipated water demand growth.

    [29] Other environmental attributes also show gains. Historically, the EAA Emergency Drought Management Rules were put in place because spring flows fell below sustainable levels as exhibited under the BASE. The model results show that water pumping limits mitigate such adverse effects. There exists a tradeoff between the welfare of pumping users and environmental attributes as depicted graphically in Figure 2 (dashed lines).

    Details are in the caption following the image
    A trade-off between regional welfare and environmental quality under water management strategies.

    [30] The environmental cost to preserve springflows (dividing changes in the total welfare by changes in springflows) averaged across water pumping restrictions is approximate $100,000 per additional cubic foot per second of springflow across a year or $138 per acft. This may be an overestimate as offsetting economic benefits to recreational and environmental users associated with changes in springflows are omitted.

    3.2. Water Market Operation

    [31] The specter of pumping limits has stimulated formation of a fledgling water market that will certainly take wing upon final definition of water property rights. To date the market has largely transferred water from agricultural to M&I users. The consequences or incremental impacts of the water market operation (MK400, MK340, and MK175) relative to the water pumping limits (columns 4–6) are presented in Table 2 (columns 7–9).

    [32] Results from our study are consistent with some of the outcomes predicted in recent literature. Under water market trading, income from agricultural commodities falls since irrigated production is curtailed as water is sold and less profitable dryland production is expanded. This result also motivates adoption of more efficient irrigation systems, deficit irrigation and less water intensive crops but the difference in water values is so large that much of the regional agricultural water is purchased. From an income perspective, the agricultural sector is better off overall considering as the additional income from water market sales more than makes up for the lost production income (Figure 1).

    [33] Medina and Uvalde Counties are primary agricultural sellers. The predominant M&I buyer is Bexar County, which is home to San Antonio, the largest municipal water consumer. The average prices for acquiring additional EA water from the agricultural sector through a water market operation are $54, $55, and $63 per acft under MK400, MK340, and MK175, respectively (dividing incomes from water transfer by the amount of water transfers). The estimated total values of the water market operation are $192 million, $176 million and $0.24 million under MK400, MK340, and MK175, respectively.

    [34] While the water market operation leads to efficiency gains, it reduces the environmental gains. For example, the Comal springflow with the water market (MK400) is about 168 cfs lower than LIM400. To achieve the jeopardy level when the water market operation is inplace mandates lower pumping limits and is achieved at a 340,000 acft limit as shown by the results under MK340. This is in contrast to the achievement of the jeopardy level at the 400,000 acft limit without water market operations (scenario LIM400).

    [35] The above result is due to the geology of the aquifer and the undifferentiated nature of the water in the market. There is a flow constricting geological feature called the “Knippa Gap” (see Kuniansky et al. [2001] for a hydrologic discussion) which impedes water flows from the west to the east. This causes withdrawal of water in the pool west of the gap to have substantially less springflow implications than water withdrawn east of the gap. For example, in a regression study, McCarl et al. [1997] find springflow effect of water use in the eastern counties is seven times that of use in western counties (McCarl et al. [1997] provide additional details, as do the EAA and Eckhardt web pages). Thus water use rights transfers from west (Uvalde county) to east negatively affect springflow and subsequent downstream environmental attributes even though the same quantity of water is pumped. The ramifications for trading would be assigning differentiated property rights and transfer rules (e.g., seven to one) for west to east pool trading.

    3.3. Water Development

    [36] The HCP and water marketing provisions increase water scarcity and value and thus suggest increased desirability for new regional water development alternatives. This study explored this issue by running all water management jointly (MK400WD, MK340WD, and MK175 WD shown in Table 3, columns 4–6) to determine the incremental impact of water development relative to the water market operation described in 8.

    Table 3. Economic and Environmental Implications With a Water Market and With/Without Water Development Under the Projected 2012 Water Demand
    Units Base Water Management Scenarios Examined
    MK400 MK340 MK175 MKSPRWD MK400WD MK340WD MK175WD
      Economic welfare
      Agricultural income $106 23.11 22.07 23.12 23.81 22.84 23.46 23.47 23.79
      Agricultural income+a $106 23.11 28.54 28.09 24.42 28.01 30.10 28.98 25.00
      M&I welfare $106 953.56 894.53 795.32 289.87 933.32 937.00 927.38 821.34
      Total welfare $106 976.67 923.07 823.41 314.28 961.33 967.09 956.36 846.34
    Environment attributesb
      J-17 elevation feet 552.73 629.45 657.95 735.78 663.99 636.34 662.37 741.03
      Sabinal elevation feet 656.07 740.44 772.75 861.28 778.93 751.30 781.54 870.56
      Comal flow cfs 0.00 68.93 234.73 689.07 258.09 105.38 256.38 715.06
      San Marcos flow cfs 20.51 67.13 85.24 134.86 87.39 70.91 87.29 137.43
      Calhoun Bay flow cfs 292.42 344.72 369.64 442.76 535.32 403.03 475.48 661.64
      Corpus Bay flow 103 acft 175.38 99.15 99.12 99.12 75.68 99.17 99.16 91.20
    Water consumption
      EA agricultural pump 103 acft 83.49 75.74 75.74 75.74 49.93 68.92 63.02 62.71
      EA M&I pump 103 acft 468.98 324.26 264.26 99.26 300.73 328.45 276.98 112.29
      Future developed supply 103 acft - - - - 123.54 105.12 124.20 208.74
      Water transfers 103 acft - 119.42 90.15 9.64 95.93 123.61 102.87 22.67
      Value of water development $106 - - - - 37.43 44.02 132.95 532.06
    • a Agricultural income+ refers to farm income plus payments from water transfers.
    • b The elevations, springflows, and bay flows reported are observed under the EA recharge drought conditions.

    [37] The findings indicate that water development can substantially increase regional welfare but not contribute a great deal in terms of the environmental problems. This is because newly developed water supplies decrease the regional water demand deficit by 70% and add $44 to $653 million to regional welfare relative to the water market operation (subtracting the total welfare in columns 4–6 from that in columns 8–10, respectively).

    [38] The results also indicate that water development has small environmental implications. The reason for the small environmental influence is largely economic. Newly developed supplies cost more in both an operating and a development sense than EA water. Hence, even when new water is developed, EA water is used first. Subsequently, the additional water supplies do not greatly offset the EA usage induced environmental effects. However the availability of new water does lower the cost of the pumping restrictions and thus the cost of preserving springflows.

    [39] Table 4 presents the water supply volume resulting from by water development alternatives under the various pumping limits. The results indicate that HCP implementation is likely to cause considerable acceleration in regional water development compared to the pattern envisioned by the SCTRWPG. For example, under the 400,000 acft pumping limit, a total of 143,420 acft are developed from seven water management projects that are currently underway and three water development alternatives – municipal water conservation, irrigation water conservation, and transfer of irrigation rights. However, if the tighter HCP limits are imposed (under 175,000 acft pumping limits), virtually all of the water development alternatives that the SCTRWPG proposed as a way to support the region until 2050 are in the near term to meet the HCP induced water scarcity. This result implies a need to develop a broader regional suite of development strategies if the tight HCP strategies are adopted.

    Table 4. Amount of Future Water Management Strategies Employed Under Alternative Pumping Limitsa
    Water Management Strategy Pumping Limit Springflow Limit
    400,000 340,000 175,000
    Current Strategyb
      Schertz-Seguin water supply 12.4 12.4 12.4 12.4
      Western Canyon regional water supply 4.5 4.5 4.5 4.5
      Lake Dunlap expansion 5.2 5.2 5.2 5.2
      Carrizo aquifer 4.0 4.0 4.0 4.0
      Trinity aquifer 1.0 1.0 1.0 1.0
      GBRA Canyon reservoir 6.7 6.7 6.7 6.7
      Hays and IH35 water supply 4.5 4.5 4.5 4.5
      Total 38.3 38.3 38.3 38.3
    Future strategyc
      Municipal water conservation 24.81 43.6 44.49 42.74
      Irrigation water conservation 39.91 40.2 40.34 40.22
      Transfer of EA irrigation rights 40.41 40.4 40.39 40.41
      Recharge from natural drainage
      Canyon Reservoir 2.08
      Lower Guadalupe River diversion 0.2 18.99 0.18
      Carrizo at Wilson and Gonzales 15.43
      Carrizo at Gonzales and Bastrop
      Simsboro aquifer
      SAWS recycled water program 19.03
      Recirculation 28.00
      Total 105.12 124.20 208.74 123.54
    • a Specific details of each water management option can be found at http://www.watershedexperience.com/2001_rwp.html. Values are in 1000 acre-feet.
    • b Current strategies refer to water management projects that are currently underway. Their yields will be available by 2010.
    • c Future strategies refer to water management projects that are proposed by the South Central Texas Regional Water Planning Group and are also acceptable under the Habitat Conservative Plan. This set of water strategies will be available only if they are constructed or funded.

    3.4. Springflow Driven Pumping Limits

    [40] Pumping limits are institutionally simple to implement but are only indirect instruments for controlling springflow and are not sensitive to regional weather conditions. An alternative approach would be to monitor springflow and curtail pumping when needed. We investigated this issue by running a scenario imposing the USFWS determined species jeopardy level as the minimum monthly springflow limit that must be achieved in conjunction with water market operation and water development (MKSPRWD). Subsequently, EDSIMR determined the pumping levels. Under that scenario, water use generally depended on recharge with the higher use observed under the higher recharge events although at the highest recharge levels water use declined somewhat since the wet conditions moderated irrigation landscape based demand.

    [41] Not surprisingly the impact of the minimum springflow limits was cheaper than the impact of the pumping limits when the model pumping was allowed to vary with water abundance. Although the springflow take level can be achieved under MK340WD or MKSPRWD, the achievement costs (dividing changes in the total welfare by changes in springflow) are different (Table 3). The achievement cost falls by 25% from $110 per acre-foot of the Comal springflow under MK340WD to $82 per acre-foot of the Comal springflow under MKSPRWD. This result implies that perhaps a water rights seniority system is needed to allow water use adjustments to water abundance.

    4. Conclusion

    [42] There is a distinct tradeoff in the EA region between the economic well-being of pumping users and regional environmental attributes. Leaving behind the rule of capture to take on any of the springflow preservation motivated pumping limits reduces regional pumping user economic welfare. The most extreme pumping limit examined (175,000 acft) under the emerging HCP raises the welfare loss to $664 million per year but would be offset by the value of springflow to recreational and environmental users. The emergence of the EA water market improves regional welfare to pumping users but worsens environmental attributes particularly if water exchanges are freely allowed between the East and West pools. The results also indicate that perhaps a water rights seniority system is needed to allow water use adjustments to water abundance. The EAA has recently set out “senior water rights” and “junior water rights”. All permits under the 450,000 acft pumping limits falls into the “senior water right” whereas permits above the 450,000 pumping limits are considered the “junior water right.” The “senior water right” permit holders can pump EA groundwater as long as the J-17 index well is above 650 feet mean sea level which refers to the critical period management plan. The “junior water right” permit holders can pump EA groundwater only when the J-17 index well is above 665 feet mean sea level (see “GBRA, SAWS to study New EAA Pumping Limits Interim Management Plan, Seguin, San Antonio,” available at http://www.saws.org/latest_news/NewsDrill.cfm?news_id=171, 24 October 2003).

    [43] Regional plans for new water development are substantially stimulated greatly by imposition of the tighter HCP related EA use restrictions. The EA region will have to develop an expanded set of water development alternatives if the severe Habitat Conservation Plan based restrictions are imposed.

    [44] Besides water management examined in this study, other water management such as pumping taxes or financial incentives (as suggested by an anonymous reviewer) may be alternatives. However, given the charter of and initial action of the EAA it appears water markets are the way the region will go and this avoids the complicated determination of the seasonally and recharge dependent optimal tax or subsidy that ensures the springflow levels at the desirable level which is the jeopardy level in this case.

    [45] Water pricing is another alternative management that can be applied to encourage water conservation. However, how a (right) price should be established or how price should be applied is debatable. Theoretically, an efficient use of water can be achieved if water is priced at its marginal cost; yet, in practice water is priced too low at its average cost which leads to overutilization of water. Also an issue regarding whether water price should be a flat rate or price discrimination is questionable.

    [46] A limitation of this study is that the recreational and environmental benefits from springflows were omitted. This omission may lead to an underestimation of the regional total welfare. To overcome this problem, future studies would involve applying the nonmarket valuation methods as discussed by Loomis [1998], Barrens et al. [2000], and Eiswerth et al. [2000] to estimate economic values of springflows and instream flows.

    Acknowledgments

    [64] Senior authorship is shared between the first two authors. This work was funded by the Texas Higher Educational Coordinating Board through the Advanced Research/Advanced Technology Program and by the Texas Agricultural Experiment Station. The authors thank Steven Raabe at the San Antonio Water Authority; Agatha C. Wade, John Waugh, and Darren Thompson at the San Antonio Water System; Steven Walthour at the Edwards Aquifer Authority; and Sam Vaugh at HDR Engineering, Inc. for information and useful discussions.

      Appendix A:: Modeling Framework

      [47] EDSIMR chooses water consumption and new supply development options so as to maximize net regional economic value subject to initial EA elevation, ending EA elevation, springflow regression, crop mixes, irrigated land and dryland, irrigation water balance, EA pumping balance, river system, regional water market, water management development, and other features and constraints.

      A1. Mathematical Term Definitions

      [48] Mathematically, the algebraic representation of these functions is presented below where all endogenous variables are typed in upper case and all exogenous parameters are in lower case. Subscripts refer to set indexes, with one letter per index. See Table A1.

      Table A1. Definitions of Exogenous Parameters and Endogenous Variables
      Parameter Definition
      Subscript
        a, a′ region
        c crop
        d water development plan
        k crop mix alternatives
        m month of the year
        n, n′ river/upper river node and is a subset of county, n ∈ p
        p county where activities occur and is a subset of region, p ∈ a
        r state of recharge nature
        s springs
        y crop mix alternative
      Exogenous parameters
        agpermitp initial agricultural sector water rights or permits in county p
        annualcostd annual cost associated with water development d
        capacityd water supplied by water development alternative d
        drylandp initial dryland in county p
        dryprofitcr dryland production profit for crop c under SON r
        endinitialaa′ regression parameter in the ending water level equation on the effect of initial elevation in region a′ on the ending elevation in region a
        endintercepta intercept for the ending water level regression in region a
        endrechargea regression parameter in the ending water level equation on the effect of recharge in region a
        enduseaa regression parameter in the ending water level equation on the effect of total water use in region a′ on the ending elevation in region a
        gagpcostpr cost per unit of EA agricultural pumping in county p under SON r
        ginddempmr inverse demand curve for EA industrial water in county p for month m under SON r
        gmundempmr inverse demand curve for EA municipal water in county p for month m under SON r
        indpermitp initial industrial sector water rights or permits in county p,
        mixdatapcy proportion of crop c in irrigated/dryland mix alternative y grown within county p where equation image
        munpermitp initial municipal sector water rights or permits in county p,
        gmipcostpr cost per unit of municipal and industrial EA pumping in county p under SON r
        irrlandp initial irrigated land in county p
        irrprofitcr irrigated production profit for crop c under SON r
        probr probability of state of nature r
        pump total amount of water pumped from the EA
        rechargemr amount of EA recharge received during month m under SON r
        sagpcostnr cost per unit of agricultural diversion at river node n under SON r
        sinddemnmr inverse demand curve for industrial surface water at river node n for month m under SON r,
        smipcostnr cost per unit of municipal and industrial surface water cost at river node n under SON r,
        smundemnmr inverse demand curve for municipal surface water at river node n for month m under SON r
        springdevdmr springflow created in month m under SON r by development of water alternative d, for spring s
        sprinitialsma regression parameter in the springflow equation on the effect of initial water level in region a for spring s in month m
        sprinterceptsm intercept of the springflow regression for spring s in month m
        sprrechargesmm* regression parameter in the springflow equation on the effect of EA recharge for spring s in month m
        sprusesmm*a Regression parameter in the springflow equation on the effect of total water use on spring s in month m in region a
        transaction transaction cost per unit of water transfer
        watrequirepcmr water requirement in county p for irrigating crop c for month m under SON r
      Endogenous variables
        AGSELLpr agricultural water sales by county p under SON r
        BUILDWATERd integer variable indicating whether water development alternative d is undertaken
        DRYCROPMIXpy usage in county p of dryland mix alternative y
        DRYCROPPRODpcr acres of dryland production for crop c in county p under SON r
        ENDLEVELar ending EA water elevation of a reference well index in region a under SON r
        FLOWnmr river flows at river node n for month m under SON r
        GAGBUYpr additional agricultural water acquired by water rights purchases in county p under SON r
        GAGOWNpr agricultural water use of permitted water in county p under SON r
        GAGWATERpmr total EA originating agricultural water use in county p for month m under SON r
        GINDpmr total EA originating industrial water use in county p for month m under SON r
        GMUNpmr total EA originating municipal water use in county p for month m under SON r
        INDBUYpr purchases of water rights by industry in county p under SON r
        INDOWNpr industrial use of permitted supply in county p under SON r
        INDSELLpr industrial sales of water rights in county p under SON r
        INFLOWnmr instream flows at river node n for month m under SON r
        INITIALLEVELa initial water elevation at the EA reference well in region a
        IRRCROPMIXpy usage in county p of irrigated mix alternative y
        IRRCROPPRODpcr acres irrigated of crop c in county p under SON r
        LOSSnmr system channel loss at river node n for month m under SON r
        MUNBUYpr purchases of water rights by municipalities in county p under SON r
        MUNOWNpr municipal use of permitted water in county p under SON r
        MUNSELLpr municipal water sales in county p under SON r
        NEWATERmdr use of water from water development alternative d in month m under SON r
        RECHARGEnmr aquifer recharge occurring at river node n for month m under SON r
        RELEASEnmr reservoirs releases and spills at river node n for month m under SON r
        RETURNFLOWnmr return flow at river node n for month m under SON r
        SAGWATERnmr agricultural surface water use at river node n in month m under SON r
        SINDnmr industrial surface water use at river node n for month m under SON r
        SMUNnmr municipal surface water use at river node n for month m under SON r
        SPRINGsmr springflow in spring s in month m under SON r
        TODRYpr irrigated land converted to dryland in county p under SON r
        TRANSFERSpr net water transfers through the water market in county p under SON r

      A2. Mathematical Equations

      A2.1. Objective Function

      [49] EDSIMR maximizes expected net benefits (benefits minus costs) of water use by municipal, industrial, and agricultural sectors. Benefits from using groundwater and surface water in municipal and industrial sectors are determined by the consumer surplus arising from the area under constant elasticity municipal and industrial demand curves less the costs of pumping and delivery. The agricultural sector benefits are represented by the net agricultural income derived from irrigated and dryland crop production. This study assumes that the agricultural producers are price takers as discussed in footnote 2 above. The farm income therefore is equal to the revenues from selling agricultural products minus the production costs. Crops considered are broccoli, cabbage, cantaloupe, carrot, corn, cotton, cucumber, guar, hay, honeydew, lettuce, oats, onion, peanuts, peppers, potato, sesame, sorghum, soybeans, spinach, sweet corn, tomato, watermelon, winter wheat, and rice.

      [50] New water development costs are also subtracted with consideration of both the fixed costs of development project construction and variable costs of consuming that new water. Transaction costs for water transfers are also added when the water market is active. The objective function is a probabilistically weighted of the above mentioned quantities across the states of nature to reflect stochastic weather.
      equation image

      A2.2. Initial Edwards Aquifer Elevation

      [51] The initial EA elevation of the reference well indices (J17 and Sabinal well) is an average of the ending EA elevation by region weighted by the probability associated with each SON in order to allow the EA level to fluctuate with the stochastic weather events.
      equation image

      A2.3. Ending Edwards Aquifer Elevation

      [52] The ending EA elevation by region and SON is a function of initial elevation, EA groundwater use and EA recharge. The endintercepta, endrechargea, endinitialaa, and enduseaa represent the regression parameters that were previously estimated in the EA model using the GWSIM-IV Edwards Aquifer Simulation Model as explained by Keplinger et al. [1997].
      equation image

      A2.4. Springflow Regression

      [53] Springflows are generated by a regression which is a function of the initial EA elevation levels in an eastern and western well (J17 and Sabinal wells), the EA monthly recharge, the aggregated EA water use in eastern (in Medina, Bexar, Comal and Hays Counties) and western (in Uvalde and Kinney Counties), and the water development associated with a springflows circulation. The sprinterceptsm, sprrechargesmm*, sprinitialsma, and sprusesmm*a represent estimated regression parameters as explained by Keplinger et al. [1997].
      equation image

      A2.5. Crop Mixes

      [54] A crop mix constraint implies that a farmers' crop production decision is a convex combination of crop mix in which all of the lands used (irrigated or dry land) for a crop c grown within a county follows the historical crop mixes observed on irrigated and dry land acres by crop c and county from the years 1975 to 1996 and from crop mixes that were collected from a 1994 farm program survey reflecting what would happen if the farm program was eliminated as explained by Schiable et al. [1999]
      equation image
      equation image
      equation image
      equation image

      A2.6. Irrigated Land and Dry Land

      [55] The constraint limits irrigated crop productions to irrigated land acres available by county but allows irrigated land acres to be converted to dryland acres under SON. The dryland acres are also limited to those available plus those converted from irrigated acres.
      equation image
      equation image

      A2.7. Irrigation Water Use

      [56] The irrigated crop water use across all crop is required to be less than or equal to monthly water available for irrigation for each county under SON plus or minus the water development when available.
      equation image

      [57] Actually the model also contains alternative irrigation strategies such as partial or full irrigation but this is not included in this exposition to keep down the complexity of the notation.

      A2.8. Edwards Aquifer Pumping Balance

      [58] The total water demand for agriculture, municipality, and industry using the groundwater pumped from the EA is limited to be equal to the total amount of the groundwater pumped from the EA. In general, the amount of water pumped from the EA (pump) is unlimited but is limited to 400,000, 340,000, and 175,000 acft per year when the pumping limit is imposed
      equation image

      A2.9. River System

      [59] The river system portrays a hydrological relationship of river nodes among upstream flow, downstream flow, as well as instream flows, reservoirs/lakes release and spill, diversions by agriculture, municipality, and industry, system channel loss, return flow, aquifer recharge, springflow, and a new water development.
      equation image

      A2.10. Regional Water Market

      [60] This constraint limits the EA agricultural water demand consumption to be less than or equal to the agricultural water supply plus additional water acquiring from purchasing water rights and the EA agricultural water supply plus the amount of water transfers via selling water rights to be less than or equal to water rights.
      equation image
      equation image
      [61] Likewise, the total EA water consumption for municipality and industry is less than or equal to the municipal water supply and industrial water supply plus additional water acquiring from purchasing water rights. The municipal and industry consumption and the amount of water transfers via selling water rights must be less than or equal to water rights.
      equation image
      equation image
      equation image
      equation image
      equation image

      A2.11. Water Management Development

      [62] The decision whether a water development d should be adapted is viewed as a binary choice, to adapt or not to adapt. If adapted, a water management development's variable cost is considered in the objective function. The amount of water that can be drawn from each water development is limited by the capacity of each water development:
      equation image
      equation image

      A2.12. Other Features and Constraints

      [63] There are a number of other features and constraints used in the study but not presented in the Appendix. For example, the pumping lift constraint is set as a function of J17 and Sabinal wells ending elevation level. Additional description of the model is given by McCarl et al. [1999].