Read JEMpaper8-23.doc text version

Water Transfers, Agriculture, and Groundwater Management: A Dynamic Economic Analysis

August 6, 2000

Keith C. Knapp Department of Environmental Science University of California, Riverside Phone: (909) 787-4195; FAX: (909) 787-3993;Email:[email protected] Marca Weinberg Resource Economics Division Economic Research Service United States Department of Agriculture Washington, D.C. 20036 Phone:(202) 694-5470; FAX:(202) 694-5776;Email:[email protected] Richard Howitt Department of Agricultural and Resource Economics University of California, Davis Phone: (530) 752-1521;FAX:(530) 752-5614;Email:[email protected] Judith F. Posnikoff Collins Associates Newport Beach, CA Phone:(714) 644-5771;FAX:(714) 720-1754; Email:[email protected]

JEL Codes: Q25, Q15

Water Transfers, Agriculture, and Groundwater Management: A Dynamic Economic Analysis


Water transfers from agriculture to urban and environmental uses are likely to become increasingly common in the American west and other parts of the world. A review of previous estimates for California suggests overall annual transfers of 5% to 15% from agriculture to urban and environmental uses over the forseeable future. A range of alternate transfer mechanisms is possible. Transfer effects on agricultural regions overlying groundwater aquifers will occur through a variety of mechanisms, are dynamic as the economic/hydrologic system responds, and will depend on the particular mechanism used. An empirical analysis is conducted for a representative region in California. Transfers via involuntary surface water cutbacks [email protected] the extraction schedule, and lower water table levels and net benefits over time. The effects are large for the water table level, but more modest for the other variables. Break-even prices are calculated for voluntary quantity contract transfers at the district level. These prices differ considerably from what might be calculated under a static analysis which ignores water table dynamics. For canal-lining, we find that districts may gain in the short-run but lose over time if all the reduction in conveyance losses is transferred outside the district. Water markets imply an evolving quantity of exported flows over time as the economic/hydrologic system responds. Overall we find that plausible levels of transfers are likely to have modest to moderate-at-worst effects on agriculture over the forseeable future given adjustment processes in that sector. The results support previous work that the overall social benefits of transfers can be quite large. However, depending on how the transfers are carried out, there can be adverse equity effects on the supplying regions. Groundwater usage is largely unregulated in California. Transfers do increase the benefits from management and management can eliminate the adverse equity effects associated with markets. On the other hand, overall benefits from management remain rather modest.

INTRODUCTION Water transfers and the associated mechanisms for achieving them are the subject of intense policy debate in the western region of the U.S. This is likely due to a combination of factors: water supplies that are generally fully allocated (if not over-allocated), growing water demand for urban and environmental uses, strong opposition to construction of new reservoir systems on cost and environmental grounds, and existing institutional structures that typically limit the ability and incentive of water users to transfer water. Since agriculture is the predominant water user, and since some existing uses within agriculture may have relatively low valuations of water at the margin, it=s likely that a significant majority of transfers will involve agriculture. These transfers might be between alternate uses within agriculture, or from agriculture to urban and environmental uses. They can also take a variety of forms ranging from involuntary cutbacks to voluntary water sales. In any event, knowledge of how water transfers potentially affect agriculture and the environment (groundwater in our case) is fundamental to the policy process. Many of the major agricultural areas in California and the West overlie groundwater basins. Significant amounts of extraction occur even in normal years, and they typically increase substantially in dry years. Return flows from agriculture and conveyance losses also heavily impact water table levels (and quality) of the underlying aquifer. One set of questions in evaluating water transfers is therefore the potential impact on groundwater use, how these transfers may affect the evolution of the agricultural production/groundwater aquifer system over time, and ultimately the consequences for agricultural producers. Furthermore, groundwater is a classic example of a common property resource that is currently unregulated in many areas. Thus the prospect of water transfers and markets also has implications for groundwater management - both how it is implemented and the economic consequences from instituting 1

management. In fact, at least under some circumstances, California water law currently prohibits water transfers unless a groundwater management plan is in place (California Water Code, sections 1220 and 1745.10). Finally, water districts facing the prospect of potential sale of existing water supplies need guidance on how to price the water taking into account future availability and costs of groundwater supplies. A number of economic studies have evaluated surface water transfers in general as well as potential effects on agriculture and the environment. These include Vaux and Howitt (1984), Howe, Schurmeier, and Shaw (1986), Saliba (1987), and Weinberg, Kling, and Wilen (1993), as just a few examples among many. However, this literature is typically static, and either ignores groundwater or treats it as a static resource. A moderate-sized literature has also developed on the economics of groundwater use and management. These studies explicitly account for groundwater dynamics over time, and also address the common property versus economically efficient usage issue. Classic studies include Burt (1964), Brown and Deacon (1972), and Gisser and Sanchez (1980), and there have been a number of extensions and empirical applications since then. However, these studies invariably focus on overlying uses and do not address how water transfers might affect groundwater use and management. This paper evaluates the effect of water transfers on agricultural production, groundwater management, and water pricing, explicitly taking into account aquifer dynamics over time. We first discuss alternate transfer mechanisms, summarize the variety of ways in which water transfers can affect groundwater systems, and then review previous estimates of the magnitude of transfers from agriculture to other sectors. Next we turn to an empirical analysis using Kern county, California as a representative example. Four alternate transfer mechansims are considered: involuntary cutbacks such as those imposed by the Central Valley Project Improvement Act (CVPIA) of 1992, voluntary water sales for a contracted 2

amount, transfers associated with canal-lining schemes, and an endogenous spot market for water. These mechanisms are analyzed under both common property (unregulated) usage to show how transfers might affect groundwater stocks and usage in the current institutional setting, as well as under economically efficient management. Issues addressed include how transfers alter the dynamics of the agricultural production/groundwater aquifer system, the welfare economics of transfers including efficiency gains as well as equity effects on agriculture, and the implications for groundwater management. WATER TRANSFERS AND GROUNDWATER As the dominant existing water user in most arid environments, agriculture is the most likely source for water transfers to meet new demands for urban and residential uses, and for instream and environmental uses, given the unlikely possibility of significant new water supply projects. In particular, we consider an irrigated agricultural producing region with surface water supplies overlying a groundwater aquifer. As depicted in figure 1, irrigation water may come from either or both of ground and surface water. Some irrigation water is not used by the crop and percolates back to the aquifer. This may arise from either leaching requirements or nonuniform irrigation. The water table level in the aquifer evolves over time in response to groundwater extractions and percolation flows back to the aquifer. Water transfer programs can and do take a variety of forms. The specifics of how the transfer is carried out will determine the economic and hydrologic effects on the region. We distinguish here four types of transfer programs: C Surface water cutbacks. By this we mean that the volume of surface water delivered to the region is cut back involutarily. Such cutbacks have been imposed in the CVPIA Act for several areas in California (dependent on flow conditions) to meet environmental goals [Loomis (1994), Weinberg 3

(1997a)]. The region does not received financial compensation for the cutback. C Quantity contract exports. Here we refer to voluntary water exports from a region. We assume that the region establishes a contract with the importing region such that a given amount of water is delivered on an annual basis. The region receives financial renumeration for the exports. C Canal-lining. Water deliveries within districts are subject to conveyance losses through the canals. Here the outside water purveyor pays for lining and other improvements in the canals. In return the purveyor gets the water [email protected] from the canal lining, that is, the reduction in conveyance losses. Although at first glance this may appear to be a scheme which leaves the district as well off as before, we=ll show that the effects are somewhat more involved and that, on net, the district may actually need to be compensated or have less water transferred in order to maintain its same economic status. C Water markets. Here we assume that water users in the district are free to sell water outside the district on annual basis. In this instance the volume of water exported will vary from year to year as the aquifer and hence groundwater withdrawals respond over time to the induced hydrologic stress. Other types of transfers are possible, but this illustrates a range of possibilities and the transfer programs to be analyzed in this paper. The potential impact of water transfers on groundwater systems depends critically on the type of transfer considered; as conceptualized in figure 1 a wide array of effects is possible. In its most direct form, a water market could, in theory, allow for the transfer of groundwater that would not otherwise have been used. The impact of that type of transfer clearly would be to increase the rate of withdrawal from the 4

aquifer. Potential implications of increased withdrawals include overdraft and land subsidence, reduced water quality due to either reduced dilution capacity if the overlying area is used for production activities that leach nutrients or other contaminants into the aquifer, and seawater intrusion in coastal basins. In addition, reduced discharge of groundwater into rivers and springs results in lower flows for those surface water resources, and poorer habitat conditions for species dependent on them. Indirect effects will occur if transfers involve surface water in a basin in which groundwater is also used. Those effects may be positive or negative, depending on whether the surface water is transferred in to or out of the recharge basin. Water transfers that move surface water out of the basin but allow that water to be replaced with groundwater are analytically identical to a direct groundwater transfer in terms of the environmental effects of the transfer, except insofar as conveyance losses associated with surface water delivery systems are also reduced. However, if surface water is transferred out of the basin but is not replaced, or is only partially replaced, two offsetting effects could occur with increased groundwater use. First, groundwater recharge could be reduced, with an accompanying reduction in the volume of water stored in the aquifer. Second, groundwater quality could improve if the mass of contaminants leached into the aquifer declines. Both effects could arise if decreases in applied irrigation water reduce the volume of water percolating below the root zone. The first effect could also occur if the water to be transferred is made available by reducing conveyance losses, e.g., by lining with concrete earthen ditches used to convey surface water to and within irrigation districts. In addition, changes in cropping patterns in response to the opportunity to transfer water could increase or decrease the quality of the underlying aquifer. Each crop has different input needs, including 5

fertilizers, pesticides, fungicides and herbicides, and irrigation management practices, all of which influence both the volume and quality of leachate for a given volume of applied water. The net effect will depend on whether crops associated with higher nutrient or chemical use and leaching fractions replace or are replaced by those with lower values for each. Conversely, water marketing opportunities could result in transfers of water into the basin. In that case, the impact of water transfers would be the opposite of those described above. Increased access to surface water could reduce pressures on groundwater resources; withdrawals could decline and the rate of recharge could increase. In some cases, water could be transferred into the basin for the sole purpose of recharging the aquifer. In particular, aquifers could be used to store water in wet years for use in times of drought (see the description of the Kern Water Bank, below). In other cases, the surface water transfers could be used to increase application rates, thus increasing deep percolation (recharge) from crop production and possibly also increasing the amount of chemicals and nutrients leached into the aquifer. MAGNITUDE OF POTENTIAL WATER TRANSFERS What is the potential magnitude of transfers out of agriculture to other sectors of the economy? This answer depends of course on a range of circumstances. It depends on the magnitude of the agricultural sector relative to the other sectors, growth in those sectors, the cost of transferring water, geographic location and many other factors. Even within a given area such as California, individual districts will likely exhibit a range of water transfers depending on their particular circumstances. Here we review some previous studies to get an idea of the magnitude of water transfers which could be achieved for agriculture as a whole in California on average and under normal year water conditions. (This amount could of course vary for wet and/or dry years.) 6

An early study by Vaux and Howitt (1984) developed a spatial equilibrium model for interregional water trading in California. Their results imply (after aggregrating), that the magnitude of water trades from agricultural to urban uses ranges from 6.3% of total agricultural supplies in 1980 to 11.5% of total agricultural supplies in 2020. This is the level of transfers that equalizes marginal values net of transport and other costs across alternate uses in California. A Programmatic Environmental Impact Statement of the CVPIA - a requirement of the act evaluated the potential water transfers occurring pursuant to the act. Using a model of Central Valley agriculture, including federal, state, and local water supplies, and of water demand for all major urban districts in the state, the authors estimate that the vast majority of all water transfers would be from Tulare Basin agriculture (including Kern County) to Southern California urban users (U.S. Bureau of Reclamation, 1998). Estimates from a simulation that modeled surface water transfers, allowing for the substitution of groundwater, suggest that four percent of the surface water in that region would be transferred, or a total of 97,000 acre-feet. Similarly, in a separate study of the CVPIA, Weinberg(1997b) estimated that water transfers from CVP water users would be four percent of federal water use for a scenario in which voluntary water transfers were the only provision considered. The CVPIA of 1992 mandates that some 1.2 million acre-feet of water annually must be reallocated to environmental uses. Under average water supply conditions, this requirement implies that approximately 20% of current CVP deliveries would be transferred from agricultural water users to the environment (Loomis, 1994; Weinberg, 1997a). However, it should be noted that this reallocation is only from federal water supplies; it does not include reallocations from state or local supplies. Also, some of this water may be met in other ways, such as use of so-called [email protected] water resulting from improved operating 7

procedures (Loomis, 1994). Thus the actual percentage reduction in total surface water diversions to agriculture for environmental uses could be significantly lower. Although inter-basin water transfers associated with water markets in California are rare, two such projects already exist or are being planned for Kern county. These include one water storage and transfer project currently in operation, the Kern Water Bank, and another conjunctive surface water storage and transfer proposal, the Arvin-Edison exchange project, under intensive negotiation. Kern county is located in the southern portion of California=s Central Valley. Its proximity to the Los Angeles metropolitan region, location south of the Delta, and access to both surface and groundwater supplies, position it well for participation in such projects. Additional economic pressure for transfers from Kern county in dry years will be induced by: (i) Increasing restrictions on the conveyance of water through the Delta under dry conditions; (ii) The high quality of the State Water Project water delivered to Kern county contractors, relative, for example, to Colorado River water; and (iii) Population growth in Southern California. The Kern Water Bank was initiated by an agreement between the California Department of Water Resources (DWR) and the Kern County Water Agency (KCWA) in 1987. In 1994 DWR transferred most of the operational control of the Kern Water Bank to the agricultural contractors in the county under the Monterey agreement. Urban contractors are also given access to the Bank. The Bank consists of 20,546 acres southwest of Bakersfield which are used for storage of water that is recharged during wet years. Planned storage for the current stage is 350,000 acre feet with an expansion to one million acre feet in a second stage. Extraction facilities have not yet been developed, but annual dry year yields from the Bank could be up to one third of its capacity, starting at 100,000 acre feet and increasing to 300,000 acre feet. 8

The Arvin Edison - Metropolitan Water District ground-water exchange proposal was initiated when the Arvin Edison Water Storage District filed a contingent water exchange program with Metropolitan Water District and with the California Water Resources Control Board. The Arvin Edison district uses the groundwater aquifer underlying the district to distribute water to many of its contractors. The district has surface water entitlements of 128,300 acre feet per year from the federal Central Valley Project. Metropolitan has substantial entitlements from the California State Water Project (SWP) which flow through the Arvin district before being pumped over the Tehachapi mountains into the Los Angeles basin. The energy required to pump the water adds substantially to its cost. In wet years Metropolitan has calculated that other supplies with lower delivery costs can substitute for some of their SWP entitlement. However, in dry years Metropolitan requires additional surface water supplies. The proposed exchange with Arvin would allocate 135,000 acre feet of Metropolitan=s SWP entitlement to ground-water recharge in wet years, and in exchange Arvin would allocate its surface water to Metropolitan in drought years. The maximum amount that could be extracted from the aquifer in any year is 75,000 acre feet. Thus the yield of the project in dry years is the sum of the Arvin entitlements and the withdrawals, which is approximately 200,000 acre feet. Despite the apparent advantages of this exchange, as of this writing the two parties are still negotiating. These two storage and transfer projects already in operation or negotiation could eventually yield up to 300,000 acre feet in dry years. The level of surface water flows in the quantitative model developed in this paper is 1.97 million acre feet; the yield from the proposed transfer and storage projects therefore comprises 15% of this amount. However, because the proposed projects involve offsetting increases of surface water recharged in wet years, they do not provide a clear indication of the net volume of water that 9

may eventually be transferred out of the region. For that reason, and to account for the possibility of other transfers, we bracket the 15% figure and consider two levels of surface water transfers C 10% and 20% C in our analysis. These figures are also comparable to transfer levels from other studies and the CVPIA noted earlier. EMPIRICAL ANALYSIS The remainder of the paper provides a simulation analysis of how surface water transfers might impact groundwater usage and management in a basin. The analysis is for that portion of Kern County, California which overlies the major aquifer. This constitutes an area with approximately 900,000 acres of farmland. Monetary values are in 1992 dollars with an interest rate of 4% where necessary. Other parameter values are in Table 1 with data sources as noted below. This section sets out the structural components of the model and empirical specification; subsequent sections specify the particular behavioral assumptions for the regimes being considered. The primary variables in the model are ht = hydraulic head of the groundwater aquifer at the beginning of period t, and wts and wtg which respectively denote surface diversions and groundwater withdrawals used for agricultural production during year t. In the model, ht is the state variable, while wts and wtg are control variables. Water transfers under voluntary water sales are expressed as wte; in some instances this will be exogenously specified while in other instances it will be a control variable. Average annual surface water availability to the district is denoted by s which is assumed to be constant over time. Annual social net benefits are


e t = p(q)dq p ( w + w )c( ht , w )+ p (q)dq s s t e t g t 0 0





where wt is total irrigation water during period t, p is the irrigation water demand curve, ps is surface water cost, c is the groundwater pumping cost function, and pe is the export demand curve. The first term on the right-hand side of (1) gives the net returns to agriculture as a function of water use, the next two terms are the cost of water supplied to the region, and the last is benefits from external water sales. Regional returns to land and management in agricultural production are (1) minus consumer surplus accruing outside the region; that is, (1) with the last integral replaced by revenue from water sales.

p(w) = a 0 + a 1w + a 2 w2


A quadratic demand curve for applied irrigation water is specified as

with the parameters estimated from data in KCWA (1998) as well as Feinerman and Knapp (1983) and Knapp and Olson (1995). As estimated, the demand curve has a vertical intercept at approximately $146/a-f, water use at a zero price equal to 5.3 maf/yr, and the curve is convex to the origin. Specification of the export demand curve is described in a later section.

c(h, w g ) = (k+e )wg + e(hh)w g +

Groundwater pumping costs are defined as

e ( wg ) 2 As y 2



where k is average cost per acre-foot of groundwater extractions related to equipment use, e denotes pumping costs per unit of lift per unit of water, s is drawdown, and _ is height of the land surface. The first h term captures costs associated with maintaining and repairing capital (well and pump) equipment resulting from use of the well and pumps for withdrawals, plus the energy costs associated with drawdown below the water table surface as the pumps run.1 The second term captures the energy costs of lifting water from the existing water table surface to the land surface, and the third term captures the effect that, as water is withdrawn over the course of the irrigation season, the water table level is declining. Well and pump related costs and drawdown were estimated from data in Dixon (1988) adjusted for inflation. Energy costs were estimated from data in Feinerman and Knapp (1983) and Dinar (1994) also adjusted for inflation.

wt = ws + w tg t Total applied irrigation water is defined as


ws le (1 ) (s- wte ) t


or just the sum of applied surface and ground water. Surface water usage is constrained by


where ? is the fraction of surface flows lost to deep percolation during conveyance through the water district. Current California law is that groundwater can generally be used only on overlying land. This formulation restricts water exports to come only from surface water. This formulation also implies that water transfers are occuring before the water reaches the region boundary, thus reducing the amount of conveyance losses to the aquifer. Surface water for this region comes from three major sources: the California State Water Project, the federal Central Valley Project, and the Kern River. Annual average surface flows are 1.97 maf. Of this amount, 70% is assumed available for irrigation and the remainder goes to the aquifer as conveyance losses. Deep percolation flows from agricultural production are assumed to be 20% of the amount applied. These parameters, along with the aquifer parameters, are from Feinerman and Knapp (1983). Initial hydraulic head was estimated from data in KCWA (1998).

h t+1 = ht +

[ (s- we )+ wt + wtg ] t A sy


The equation of motion for the groundwater aquifer is given by

where is net natural recharge to the aquifer, ß is the fraction of irrigation water that percolates through the rootzone to the groundwater aquifer, A is the aquifer area, and sy is the specific yield of the aquifer. Net natural recharge includes recharge from natural sources and lateral flows, as well as possible leakage through the bottom of the aquifer. Natural recharge is quite small in the empirical model in comparison to other sources of recharge and is assumed here to be a constant. The reference point for computing hydraulic 13

heads is mean sea level (MSL). Hydraulic head is bounded by h # ht # h where h and h are the lower and _ _ upper bounds respectively as determined by the aquifer geometry. We will consider four types of water transfers. The first type of transfer is involuntary cutbacks as in the CVPIA Act. This is modelled by reducing the value of average annual surface flows s but with no exports. The second type of transfer is voluntary water sales on an extended quantity contract basis. Here we will specify a constant (over time) value for wte, thus this will be a parameter in the analysis. In general the export price pe will depend on a variety of factors including bargaining power of the region. Here we will sidestep this issue, and instead calculate pe as a break-even price. Such prices can be of value to districts trying to decide how much to accept for future water sales. The third type of transfer is canal-lining. As explained later, this is implemented via a reduction in the conveyance loss parameter and a reduction in surface flows s. The fourth transfer program is an endogenous spot market for water. ECONOMICS OF GROUNDWATER USE AND MANAGEMENT The basic concepts of groundwater economics are set out in the seminal articles by Burt (1964), Brown and Deacon (1972), and Gisser and Sanchez (1980) and the ensuing literature. In this section we illustrate the basic economic concepts in the context of the empirical model. This will set the stage for a dynamic analysis of the effects of water transfers and water markets on agricultural producers and the groundwater resource. Most of the groundwater basins in California underlying the major agricultural regions are currently unregulated with regard to withdrawal volumes, thus we will begin with this assumption. Following that, we sketch the reasons why unregulated (common property) usage is inefficient and then identify an economically efficient allocation over time. Under common property usage there is no regulation of the resource system; individual growers are


assumed to make decisions that are strictly in their own best interest and ignore effects on others. With many users, each of whom is small compared to the resource, the effect of an individual's current decisions on future levels of the regional groundwater stock is perceived to be negligible. Therefore it is reasonable to suppose that decisions are made in each period to maximize net benefits (1) in each period without regard to the future level of the groundwater stock. This maximization is subject to the constraint that groundwater extractions are limited to the available supply, and the water constraints (4) and (5). Numerical simulations of this system were carried out using GAMS (Brooke et al 1992). The results are displayed in figure 2 for an initial hydraulic head of 168 feet above MSL, no transfers of any kind (e.g. a full allotment of surface water) and a 50 year time horizon. As demonstrated in the figure, hydraulic head declines over this time period by some 25 feet. As the hydraulic head declines, groundwater becomes more expensive and so withdrawals are somewhat reduced as well. Annual net benefits decline by approximately $5/acre during this period as a consequence of increased pumping costs as well as reduced withdrawals. Thus, the results suggest that, without regulation, the system is probably not too far from a steady-state situation. Of course this neglects the inevitable year-to-year variability in surface water supplies which translates into fluctuating withdrawals and water table levels over time. Groundwater extractions by an individual pumper lower future water tables and hence increase future pumping costs for everyone. As previously argued, under common property usage with a large number of extractors, it is reasonable to suppose that individuals only consider current pumping costs and ignore the additional costs imposed on everyone (including themselves) in the future. (Dixon [1988] provides extensive theoretical and empirical evidence on this.) This implies that current withdrawals are likely to be excessive and future water table levels too low in comparison to economic efficiency. We now


consider economically efficient use of the resource and contrast it to common property usage. We will then be in a position to identify the effects of water transfers on groundwater management, in terms of both allocation over time and the gains to be had from management. For economic efficiency, we choose a time path for groundwater extractions to maximize the

t t=1



present value of social net benefits (1) over an infinite horizon

where a is the discount factor. This optimization is subject to the equation of motion (6), the other constraints and definitions (2)-(5), and various bounds and non-negativity conditions. The problem is solved numerically using dynamic programming methods written in the GAMS language (Brooke et al 1992): backward recursions were used to calculate the infinite horizon value function, and this was then used in forward simulations to estimate time-series values over a 50-year horizon. Economically efficient use is contrasted with common property usage in figure 2 over the 50 year horizon and for the original surface water allocation. In contrast to common property usage, economic efficiency results in an increasing hydraulic head over the simulated time span. The difference is quite significant; after 50 years for example, there is almost an 80 foot difference in pumping lift. Although the system does not reach steady-state during the horizon considered, optimal management will eventually result in a steady-state, and obviously the water table will be at a substantially higher elevation than under common property usage. (Although natural recharge to the aquifer is small, conveyance losses and deep percolation 16

flows are relatively large. Therefore it is physically possible for the water table to rise over time provided extractions are less than these amounts. Whether it rises or falls depends on a variety of factors including institutional structure and parameter values.) As demonstrated in the middle panel of figure 2, optimal withdrawals are lower than withdrawals under common property usage. This results from the fact that the economically efficient solution accounts for the effect of current withdrawals on future pumping costs, whereas the common property solution does not. Note also that this difference can be significant. In particular, efficient withdrawals at the beginning of the horizon are some 38% less than those under common property usage. This difference does decline over time and, in fact, inspection of the equation of motion (6) indicates that, in this model, steady-state withdrawals will be identical under the two regimes, although the pumping lift will differ as noted above. Finally, figure 2 illustrates the main trade-off inherent in groundwater management. Social net benefits are reduced in the early years under efficient use due to reduced extractions and the consequent adjustments on the part of growers. After some point, however, management results in increased annual social net benefits relative to common property usage due to reduced pumping costs and also narrowing the gap in terms of extractions. Benefits from groundwater management are the difference in present value of social net benefits under the two regimes. After annualizing, groundwater management benefits are $2.44 acre-1 year-1 for the basin considered here over the 50 year horizon (Table 2). This would be the social gain from managing the resource. Consistent with previous economic studies of groundwater management, this figure is fairly low. Note also that this is only the benefit side of management, it does not consider the costs of management such as reaching initial agreement, monitoring and enforcement, and so on.


SURFACE WATER CUTBACKS We next consider the effects of cutbacks in surface flows to the basin. Such cutbacks have been imposed in the CVPIA Act for several areas in California (dependent on flow conditions) to meet environmental goals [Loomis (1994), Weinberg (1997a)]. We suppose that the specified cutback in surface water deliveries to the basin begins in the first year of the 50 year simulation and continues for every year thereafter. We also assume that with these cutbacks, growers will no longer have to pay for this water. Also to be noted is that the cutback is in total surface water availability (s); since some of this is lost during conveyance, the cutback translates into a smaller cutback in actual surface water available for irrigation, although recharge to the aquifer is also adversely affected. Common Property We begin with common property conditions. Figure 3 illustrates the results for the original supply of 1.97 maf year-1, and for cutbacks of 10% and 20% respectively. With reduced surface water availability, growers will obviously rely more heavily on groundwater, thus further lowering the water table. For example, the 10% cutback lowers the water table by 18 feet after 20 years and by 33 feet after 50 years in comparison to the base run. Comparable effects for the 20% cutback are 35 feet and 67 feet respectively. The effect on withdrawals is more complicated. In particular, reduced surface water availability tends to "tilt" the withdrawal schedule. In the early years withdrawals are greater with larger cutbacks as would be expected. However, at some point later in the horizon, the extra pumping lift makes groundwater sufficiently expensive that withdrawals eventually become less than withdrawals with no cutbacks. Also, recharge is reduced. The bottom panel in figure 3 illustrates what happens to annual net benefits (profits) with cutbacks.


For the 10% cutback, annual net benefits are reduced by $6 per acre per year after 20 years and $9 per acre per year after 50 years. Comparable effects for the 20% cutback are $12 per acre per year and $19 per acre per year. Annualized values are summarized in Table 2. The 10% cutback implies a present

value loss of $111 per acre or (annualized) $5.16 per acre per year; analogous figures for the 20% cutback are $231 per acre and $10.74 per acre per year. An overall conclusion about the effects of surface water cutbacks on groundwater usage in the unregulated case thus depends on the perspective. If one considers the 10% cutback and a short-horizon, then the impacts are modest. If one considers the larger cutback and/or a long horizon, then the effects are more substantial. Economic Efficiency Next we consider the effects of surface water cutbacks on economically efficient allocations and the benefits from groundwater management. The results are illustrated in figure 4 for the original surface allocation and also the 10% and 20% cutbacks. Qualitatively, the effects of cutbacks are comparable to those under common property: cutbacks lower the water table over what it otherwise would be, they tilt the extraction schedule, and they also lower annual net benefits from water use in agriculture. In the case of the 20% cutback, the effect is strong enough that optimal water tables decline over time instead of increasing as in the other two cases. Quantitatively, the 20% cutback after 50 years lowers the water table by 47% under common property (compared to the original allocation), but only 30% under efficiency. Comparable figures for annual net benefits are 13% for common property and 12% for efficiency. Thus optimal management mitigates - albeit only partially - the adverse effects of transfers on the aquifer. Intuitively, water transfers from the basin makes the groundwater resource more valuable and so the gains from management should increase as surface water availability decreases. This is borne out by


the results. Referring to Table 2, the 10% cutback implies annualized management benefits of $2.56 acre-1 year-1 (again, this is the difference between common property usage and full efficiency), which is a 5% increase over management benefits with the original surface water allocation. The 20% cutback results in annualized management benefits of $2.71 acre-1 year-1 for an 11% increase over the no cutback case. Thus, in relative terms the surface water reductions do imply significant increases in the incentive to manage the resource. Overall, however, benefits from groundwater management appear to remain rather modest. QUANTITY CONTRACT EXPORTS We now consider a voluntary transfers program of the same magnitude as the pure cutbacks, i.e. sales of 10% or 20% of available surface water in every period over the horizon. Growers continue to pay the base cost of water ($37/a-f) whether it is used or transferred. In this instance, all physical variables (withdrawals, hydraulic head etc.) stay the same as in the analagous cutback case; the only change is the fiscal impact on farmers. In actuality the price received for water by the region will depend on a variety of factors including negotiating and bargaining strengths. Here we will focus on a break-even water price calculated as the cost of water to growers ($37/a-f) plus an amount sufficient to cover the losses identified above. These break-even prices are calculated to leave the district equally well off in present value terms over the 50 year horizon as it would be without the transfers. Common Property One way to calculate such break-even prices is to determine a price in each period that equalizes the loss in annual net benefits in that period from the transfer. Starting with no regulation (common property usage), break-even prices would be $44.90/a-f for water transferred in year 1, $63.80/a-f for water transferred in year 20, and $78.20/a-f for water transferred in year 50. Another approach is to calculate


a single (uniform) break-even price to be paid annually. Using the present-value losses noted previously, we calculate this break-even price as $61/a-f which (again) is an amount to be paid annually. These are minimum annual prices that the region should accept for a uniform level of annual water transfers over the 50-year period in order to achieve at least the same present value of net benefits. (These prices only apply for relatively small transfers - for larger transfers the break-even price could rise given nonlinearities in the net benefit function. Also, these prices still leave the district worse off after year 50, althought this effect will be small in present value terms evaluated at the beginning of the horizon.) The results show that the dynamic groundwater effects and the time horizon are critical in determining the break-even price for water transfers, and that a static analysis ignoring these effects could seriously underestimate break-even prices. In the sequential price approach, there is a $33/a-f difference in the breakeven price from year 1 to year 50, while in the uniform price case there is a $16/a-f difference between the original (first-period price) and the correct annualized price. These differences arise because prices implied by short time horizon calculations (e.g. the first period) do not account for the longer-term impacts of the transfer on the quantity of groundwater available or the costs of using it. Thus a static analysis relying on year 1 results could seriously underestimate appropriate prices to charge for water. Economic Efficiency What happens if we assume an efficient allocation? Break-even prices under a sequenced approach equating annual net benefits in each period are $47/a-f in year 1, $62/a-f in year 20, and $79/a-f in year 50. The uniform (annualized) break-even price equating present values over the 50-year horizon is $60/a-f. Interestingly enough, these various break-even prices are all very comparable to those calculated under the common property regime. Thus in many respects (water table level being a notable


exception), water transfers as analyzed here appear to have similar effects across the two regimes. CANAL-LINING We next consider the effects of canal-lining with a subsequent transfer of [email protected] supplies out of the basin. The idea is that an outside entity invests in canal lining which reduces percolation of surface flows to the water table. The amount of [email protected] water is then transferred to the outside entity. Nominally at least, the outside entity presumably gains from this transaction while the agricultural producing region is at least not worse off. It will be seen that in actuality the situation is somewhat more complicated. It can be shown that this scheme results in a gain in surface water available for irrigation to the district, although this gain is not likely to be large in most instances. Although surface flows to the district are reduced, this reduction is more than made up for by the reduced conveyance losses for what remains. However, the reduction in deep percolation flows affects hydraulic head and hence groundwater extractions. Theoretical analysis (available from the authors) establishes that steady-state hydraulic heads and groundwater extractions are reduced in this scheme under both common property and efficient usage in comparison to the base case. This is true at least for small changes in canal-lining if not all changes. Thus there are somewhat opposing effects on the agricultural region and empirical analysis is needed to determine the net outcome. For the empirical analysis, we considered the canal-lining transfer scheme with a change in the surface water infiltration coefficient equal to .1 and water transfers equal to .971 maf/yr. Thus the new surface water percolation coefficient is ? = .2. and surface flows to the district are s = 1.773 maf/yr. Common Property Consistent with theoretical analysis, this canal-lining project results in a hydraulic head at year 50


some 34 feet lower than that in the base case, and groundwater withdrawals some .21 maf less than base levels in year 50. Present value of net benefits in the region are reduced by $43.65 million or an annualized value of $2.26 per acre of farmland per year. As noted, there are two somewhat counteracting effects of this scheme on the region: First, there is an increase in available surface water for irrigation in each year. This shows up in terms of reduced groundwater extractions and in fact net benefits in the region are higher than without the canal-lining scheme in the early years. However, there is also a negative impact on the region in terms of reduced flows to the water table from percolation below the canal. This latter effect eventually comes to dominate and after some point annual net benefits are below the base level as pumping costs increase due to an increased lift and this in turn reduces withdrawals. Although the program might nominally appear reasonable at first glance since [email protected] water from the canal-lining is being transferred, in actuality the region is somewhat worse off, so either less than the amount of so-called [email protected] water should be transferred or the region would need to be financially compensated in some way. Economic Efficiency Annualized net benefits under the canal-lining scheme are $151.17 acre-1 yr-1 compared to $153.92 acre-1 yr-1 with the original allocation under economic efficiency. As before, one option would be compensation equal to the difference of these two amounts. Another option would be to reduce water transfers below the original savings in conveyance losses by some amount. WATER MARKETS We now consider endogenous water markets as a mechanism of intersectoral transfers and allocation. Here we assume the basin receives its full allotment of surface water, but farmers are allowed


to sell water to the urban sector on an individual basis. We suppose a competitive market with many buyers and sellers in the external market. Each is relatively small compared to the market and takes price as a given. Since competitive markets are economically efficient under the First Welfare Theorem (Varian, 1992), optimizing (1) simulates a competitive market with respect to exports. This market is restricted efficient in that for a given year with a given hydraulic head it would not be possible to achieve a higher level of social net benefits in that year. Whether or not overall efficient water use is achieved depends on whether common property use or economic efficiency is being considered. This is a critical distinction which will have implications for the results. The export demand curve in (1) gives marginal benefits from water exported to areas outside the region under calculation. It is net of transport costs and thus reflects prices at the region=s border. The export demand curve was estimated from Vaux and Howitt (1984). They divide the California urban sector into two regions with estimated supply and demand curves for each. Each of these were aggregated, then horizontally subtracted to determine an aggregate net import demand curve for this sector. This curve was then adjusted for inflation, transport costs, and Kern county's surface supply as a fraction of California agriculture's surface supply, to determine the export demand curve for Kern county. This curve is fairly inelastic reflecting residential and industrial users greater degree of ability to pay for water (Table 1). Other definitions, equations and constraints are as before. Common Property Usage We first consider water markets under common property usage, that is, there is no explicit groundwater management. In this situation, over the 50 year horizon, water exports from Kern county begin at .405 maf/yr and decline to .396 maf/yr. This decline is due to the fact that as the water table falls,


the shadow value of surface flows rises and hence less is offered for sale on the market. Previous studies of water transfers are generally static. Qualitatively, these results demonstrate that the structure and flow of transfers could evolve over time as the aquifer responds. Quantitatively, this effect is quite small and is due to the inelastic export demand curve noted earlier. This level of water exports under common property represents just about 20% of the available surface inflows to Kern county. Thus the other hydrologic variables under competitive water markets would be approximately as depicted in figure 3 for the 20% cutback. In year 1, social net benefits are $138 million without markets and $236 with markets, for an increase of 71%. In year 50, the comparable percentage increase due to water markets is 61%. Thus competitive water markets can result in substantial societal gains; they are driven by the greater WTP for water in the urban sector and the consequent inelastic demand. As noted earlier, regional net benefits are given by (1) minus CS accruing to the urban sector. In the first year, we find regional net benefits essentially identical with and without water markets. In the last year (year 50), regional net benefits with water markets are actually 7.5% less than without markets. The first result is explained by the fact that marginal pumping costs are essentially flat at the equilibrium level; this means that exports can be supplied with a very minimal increase in water prices in the region. In other words, the export supply curve is very flat so there is minimal producer surplus from opening up markets; almost all the gain in social net benefits is accruing to external water consumers. For the second result, the surface water exports imply larger groundwater extractions which drives the water table level lower than it would otherwise be. In this example, at least, the (static) gains from trade to the region are outweighed by the increased pumping cost and the region is worse off.


At first glance these results may seem at odds with economic theory since water sales are voluntary; however, this is not the case. In each period considered individually, water sales leave the agricultural region at least as well off as no water markets in that period given the existing hydraulic head and other hydrologic conditions at the beginning of the period. Furthermore, water sales are by individuals in a competitive market, thus there is no market power being exerted. The fact that they lose over the horizon compared to no markets cannot by overcome by individual agents acting separately since, under common property usage with many agents and no regional collusion, it wouldn=t pay for individual agents to try and hold back on water sales and their own groundwater usage in order to stem the falling groundwater table. It=s rational for them to operate solely on a period-by-period basis without regard for the future. Economically efficient management We now consider a competitive water market in conjunction with economically efficient groundwater management. This is simulated by optimizing (7) subject to the equation of motion (4), and the other definitions and constraints, but with no upper bound on surface water exports we as in the t previous section. This optimizes groundwater usage over time as well as the allocation of surface water between the agricultural region and urban/residential use. In this instance we find that water exports from the region are a constant (over time) .39 maf/yr. This is somewhat less than exports under common property usage, the reason being that with optimal management the user cost of groundwater extractions is now accounted for. This means that growers are somewhat less willing to substitute current extractions for reduced surface flows, hence somewhat less surface flows are exported. Nevertheless the difference is relatively small. This and the constancy over time are again due to the relatively inelastic export demand curve.


This level of exports roughly corresponds to a 20% cutback in surface water availability to the region. Accordingly, the physical variables are roughly similar to those for the 20% cutback under efficient management as illustrated in figure (4): hydraulic head declines slightly over time as do withdrawals. Social net benefits as defined by (1) are almost constant over the horizon: they begin at $230 million/year in year 1 and end at $229 million/year in year 50. The gain in social net benefits for the combined agricultural/urban system from opening up markets is again quite substantial. Annualized social net benefits over the horizon are $230 million/yr and $138 million/yr with and without markets respectively. This represents an annualized gain in overall social welfare of $92 million/yr or some 67%. As before, regional net benefits are defined as (1) with the exception that the area under the export demand curve is replaced by export revenue defined as the export price at the border of the agricultural region times export quantities. Annualized regional net benefits over the 50 year horizon are $141.41 million/yr and $138.529 million/yr with and without water markets respectively. Thus opening up water markets in conjunction with optimal groundwater management implies a gain of $3.20/(acre-year) for agricultural producers who are the water sellers in this problem, or a gain of approximately 2%. Note that this assumes that the gains from groundwater management are internalized in the region. This could occur through quantity restrictions on groundwater withdrawals, in which case users capture the benefits of management, or through a pumping charge which is rebated back to users in a lump-sum fashion so as not to distort the incentive effects of the charge. Implications Water markets can result in both large efficiency gains and a reduction in seller welfare when


groundwater dynamics are taken into account. In the policy arena, many regions express concern over water transfers out of the region due to perceived losses from water sales. This is typically countered by economists on the grounds of static supply and demand analysis in which everyone gains from trade. However, the results here provide some support for the concerns expressed by exporting regions in actual trading situations. One way around the possible losses is that with the large efficiency gains, clearly it would be possible for individual water sellers in the region to band together and collectively negotiate a mutually beneficial price, e.g. through the water district. In reality this may be how water trades occur anyhow since the property rights to water by individuals in the district may often be somewhat ambiguous. This would essentially be the fixed price case analyzed previously. Another possibility is for forward-looking groundwater management that either explicitly considers the inefficient externalities being generated under common property usage, or places physical restrictions on water sales and/or extractions in order to maintain current water table levels. In this instance we did find that economically efficient usage does allow water sellers (agriculture here) to gain from trade. At the same time, the gains are relatively small. If the gains from water sales are this small in actuality, then it might not be worth it to agriculture to participate in the opening up markets in view of the various transition and transaction costs that would have to be borne. Aside from this particular numerical example, whether or not economically efficient management will always guarantee gains to the sellers when groundwater dynamics are accounted for appears to be a somewhat open question.


CONCLUSIONS Water transfers from agriculture to environmental and urban/residential/industrial uses are likely to become increasingly common over the next few decades in California and other parts of the western U.S and world. This is a result of currently fully allocated supplies with few prospects for expanding those supplies, growing urban and industrial uses, increased attention to maintaining and preserving natural habitats, and the fact that agriculture is currently by far the predominant water user. It is difficult to estimate the magnitude of these transfers, but a review of existing studies suggests that, on average, transfers of 5% to 15% from agriculture to the environment and other sectors of the economy might not be unreasonable over the next 20-30 years. Of course this figure will vary significantly from one area to the next. For example, water districts relying heavily on federal water will be more severely impacted by the CVPIA Act than those with state and local supplies. Agricultural areas close to urban areas or with poorer growing conditions would be prime targets for transfers to other sectors of the economy. To understand the qualitative impacts of transfers as well as get order of magnitude estimates, we analyzed the effects of several water transfer programs using Kern county, California as a representative example of a major agricultural area overlying a groundwater basin. Surface water transfers from the region (or reduced surface water availability) imply increased groundwater usage over time. Considering first common property usage, the effects of an annual 10% transfer level on the water table are fairly modest, ranging up to a 33 foot drop over 50 years in comparison to no transfers. The effects of a 20% transfer level are more significant, here ranging up to a 67 foot drop after 50 years in comparison to the no-transfer case. A 20% level of transfers reduces annual net benefits from agricultural production by 8.3% and 12.8% after 20 and 50 years respectively (in comparison to the


original no-transfer case), while the 10% level of transfers implies reductions about half these levels. Analogous results are obtained under economic efficiency. For voluntary quantity contract transfers, we estimated a break-even price for water sales from the region. Since growers will rely more heavily on groundwater if surface supplies are transferred, this price needs to account for increased future groundwater pumping costs. For a sequenced scheme where the break-even price is set in each year to equalize annual net benefits with and without transfers, we found break-even prices ranging from $45/a-f in year 1 to $78/a-f in year 50. Alternately, a uniform (annualized) break-even price equating present values with and without transfers over the 50-year horizon was found to be approximately $60/a-f. These are the minimum prices a district should accept in order to sell water while accounting for groundwater dynamics and also assuming that the district continues to bear the original base charge of the water. These prices roughly apply to both the unregulated (common property) and economically efficient cases, but differ substantially from what might be calculated in a static analysis under year 1 conditions. We also evaluated a canal-lining scheme in which the outside entity finances the construction in return for receiving the reductions in conveyance losses. This has the two-fold effect of increasing net surface water availability to the region after accounting for canal losses but also reduces recharge to the aquifer. Although the region gains from this scheme in the short-run, it can actually lose over the longrun. Thus either less water would have to be transferred out of the district or some sort of compensation would need to be made. The fourth mechanism examined was water markets. The results suggest that some intuition about water markets from static analyses may need to be modified when one considers groundwater


dynamics. Static intuition is that since water sales are voluntary, everyone by necessity gains (aside from third-party effects). In the dynamic case, this is true in the short run; however, it may not be in true in the long-run absent groundwater management, and even with management the gains from participating in water markets may not necessarily be that great depending on the institutional structure, and even when the social net benefits from trade are great. Thus some of the fears of water-exporting regions about water markets may have a legitimate grounding. In the policy arena, it may therefore be necessary to take extra steps for everyone to share in the net benefits of establishing water markets beyond just setting up a textbook competitive water market and accounting for the third-party effects already wellnoted in the literature. We now summarize some general observations which hold across the particular mechanisms being considered: C The transfer levels considered here appear to represent plausible levels - if not upper bounds of water transferred out of California agriculture for the forseeable future on average. These amounts are consistent with both existing projects in Kern county, existing CVPIA cutbacks for environmental uses, and previous studies of water transfers. They also represent a roughly 50% to 100% increase in urban and residential water use if all the transferred water were to go to those uses. This observation, combined with the above numerical results, suggests that water transfers - on average - will have hydrologic and economic effects that are relatively moderate on California agriculture, and even less so financially when compensation is paid. Certainly the effects are something less than catastrophic. C The social net benefits from allowing water transfers can be extremely large. For example,


allowing water trades to the urban sector with its relatively large willingness-to-pay and inelastic demand increases social net benefits by some 60% to 70% depending on the year and institutional structure. C A major concern with water transfers is the equity effects on the regions supplying the water. Certainly regions with involuntary cutbacks will lose, but, as noted, the effects do not appear large when one considers the range of possible adjustments. Regions negotiating collectively say through the district - are in a position to gain substantially from transfers depending on relative bargaining strengths. Under competitive water markets, regions without groundwater management can lose from transfers, although this may be mitigated by instituting management. C The major agricultural basins in California are generally considered to be unregulated with respect to groundwater withdrawals at the current time. Consistent with previous economic studies, under baseline surface flows we find a modest level of benefits from groundwater management ($2.44 acre-1 year-1) over the 50-year horizon. C Reductions in surface water availability due to transfers increase stress on the aquifer and benefits from management increase as would be expected. In percentage terms the increase can be substantial. For example, management benefits increase by 5% and 11% for the 10% and 20% cutback transfer program respectively. Nevertheless, the overall level of management benefits remains rather small. C It is also true, that, at least in some locales, water cannot be transferred out of basins without a groundwater management plan. Thus, these results do not support the idea that groundwater management should necessarily be required as a precondition for transfers, at least on


quantity/water table level grounds and where the magnitude of transfers is at the level analyzed here. Taken together, the results provide a synopsis of how the evolution over time of the agricultural production/groundwater aquifer system is altered by transfers away from that system. REFERENCES Brooke, A., D. Kendrick, and A. Meeraus. GAMS: A User=s Guide, Release 2.25. The Scientific Press, San Francisco, CA, 1992. Brown, G., and R. Deacon. AEconomic optimization of a single-cell [email protected] Water Resources Research 8:1972, pgs 557-564. Burt, O.R. AOptimal resource use over time with an application to ground [email protected] Science 11(1):September 1964, pgs 80-93. Dinar, A. AImpact of energy cost and water resource availability on agriculture and groundwater quality in [email protected] Resource and Energy Economics 16:1994, pgs 47-66. Dixon, L.S. Models of groundwater extraction with an examination of agricultural water use in Kern County, California. PhD Dissertation, University of California, Berkeley, 1988. Feinerman, E., and K.C. Knapp. "Benefits from groundwater management: magnitude, sensitivity, and distribution." American Journal of Agricultural Economics 65(4):1983, pgs 703 - 710. Gisser, M., and D.A. Sanchez. ACompetition versus optimal control in groundwater [email protected] Water Resources Research 16(4):August 1980, pgs 638-642. Howe, C. W., D. R. Schurmeier, and W. D. Shaw, Jr. "Innovative Approaches to Water Allocation: The Potential for Water Markets." Water Resources Research 22:1986, pgs 439-445. Management


Kern County Water Agency (KCWA). Water Supply Report 1995. Bakersfield, CA, January 1998. Knapp, K.C., and L. J. Olson. "The economics of conjunctive groundwater management with stochastic surface supplies." Journal of Environmental Economics and Management 28: 1995, pgs 340 - 356. Loomis, J.B. AWater transfer and major environmental provisions of the Central Valley Project Improvement Act: A preliminary economic [email protected] Water Resources Research 30(6):June 1994, pgs 1865-1871. Saliba, B.C. ADo water markets >work=? Market transfers and tradeoffs in the Southwestern [email protected] Water Resources Research 23(7):July 1987, pgs 1113-1122. U.S. Bureau of Reclamation, Central Valley Project Improvement Act: Draft Programmatic Environmental Impact Statement, (Sacramento, CA), January 15, 1998. Varian, H.R. Microeconomic Analysis. 3rd edition, W.W. Norton & Co., New York, 1992. Vaux, H. J. and R. E. Howitt. "Managing Water Scarcity: An Evaluation of Interregional Transfers." Water Resources Research 20:1984, pgs 785-792. Weinberg, M. AFederal water policy reform: implications for irrigated farms in [email protected] Contemporary Economic Policy 15:April 1997a, pgs 63-73. Weinberg, Marca. Water Use Conflicts in the West: Implications of Reforming the Bureau of Reclamation's Water Supply Policies. Congressional Budget Office Study, Congress


of the United States (GPO, Washington DC), August 1997b. Weinberg, M., C. L. Kling, and J. E. Wilen. "Water Markets and Water Quality." Journal of Agricultural Economics 75:1993, pgs 278-91. American



1. The water table level is the height of the water table relative to some datum (sea level in our case) with no pumping. As groundwater is extracted, the water table in the immediate vicinity of the wells is lowered, creating what is known as a cone of depression or drawdown. This downward gradient is necessary for water to flow into the well, and is ultimately due to the finite transmissivity of the aquifer. The drawdown implies additional energy costs needed to lift water to the land surface beyond those implied by the regional water table level, hence its inclusion in the pumping cost formula.


Table 1. Parameter values for the Kern County agricultural production/groundwater aquifer


Parameter a0 a1 a2 k e s ps s ? ß



water demand intercept$146.49 (a-f) -1 water demand slope -$45.21 (maf)-1 (a-f) -1 water demand quadratic $3.31 (maf) -2 (a-f) -1 well/pump O&M costs $6.16 (a-f) -1 energy cost $.148 (a-f) -1 foot-1 drawdown 60 feet surface water price $37 (a-f) -1 annual surface inflows 1.97 maf/yr surface flows infiltration .3 coefficient deep percolation coefficient .2

A _ h h sy ? h1 b0 b1 b2 b3

aquifer area land height aquifer bottom specific yield natural recharge initial hydraulic head

1.29 million acres 385 feet above MSL -233 feet above MSL .13 .052 maf/yr 168 feet above MSL $509.41 (a-f) -1 -$1342.54 (maf) -1 (a-f) -1 $2318.18 (maf) -2 (a-f) -1 -$4486.4 (maf) -3 (a-f) -1

export demand intercept export demand slope export demand quadratic export demand cubic


Table 2. Annualized net benefits and management benefits for the Kern county agricultural production/groundwater aquifer system.

Surface Annualized net benefits over 50 year horizon Annualized Groundwater water cutbacks Common Property Efficiency Management Benefits (%) ($ per acre per year) ($ per acre per year)

0 10 20

151.48 146.32 140.74

153.92 148.88 143.45

2.44 2.56 2.71

Net benefits are returns to land and management in agricultural production and do not include any financial compensation for water transfers. Per acre figures are per-acre of farmland in Kern County (approximately .9 million acres).


Figure Legends

1. Schematic of the agricultural production/groundwater aquifer system. (ET is crop evapotranspiration; other symbols are defined in the text.) 2. Common property and economic efficiency for the Kern County agricultural production/groundwater aquifer system. 3. Effects of surface water transfers (reduced surface water availability with no compensation) under common property usage for the Kern County agricultural production/groundwater aquifer system. 4. Effects of surface water transfers (reduced surface water availability with no compensation) under economically efficient management for the Kern County agricultural production/groundwater aquifer system.




41 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate


You might also be interested in

Chapter 7 Water Transfers
t: An argument for the importance of equity as a criterion in evaluating water policy, with examples in wide-ranging case studies from North and South America and Europe. For more information, or to purchase this title, visit:
A Water Resources Planning Tool for the Jordan River Basin