Susanne M. Scheierling, Robert A. Young, Department of Agricultural and Resource Economics, and Grant E. Cardon, Department of Soil and Crop Sciences, Colorado State University, Fort Collins, CO
Recent interest in the impact of reduced irrigation water use is a reflection of the growing competition between agricultural, urban, and environmental uses for limited water supplies. In many parts of the western United States, agriculture is the main user of the water resource, accounting for 80 to over 90 percent of the water consumed. It is increasingly recognized that more efforts to improve irrigation practices and thus reduce water use in agriculture would have several desirable effects (Howe et al., 1986; Young, 1986a; Colby, 1990; Dinar and Letey, 1991). It would make water available for emerging, higher-valued uses in non-agricultural sectors. By discouraging excess irrigation and reducing drainage volumes it would help control the discharge of agricultural pollutants into surface and ground waters and, as a consequence, diminish the social costs associated with irrigated agriculture. In addition, water not consumed in agriculture can contribute to increased instream flows for improving fish and wildlife habitat, recreation, and other environmental uses.
However, it is often pointed out that a reduction of irrigation water use would also be associated with serious negative economic impacts on agriculture. According to this viewpoint, the use of less irrigation water than "required" by different crops would lead to significant reductions in crop yields and/or irrigated area and, as a result, to high losses in agricultural income. Furthermore, the local and regional economies due to their close linkages to irrigated agriculture would suffer. Based on a more detailed study (Scheierling, in preparation), this paper attempts to show that this viewpoint neglects the strategies available to farmers for adapting to reductions in the quantity of irrigation water available or, equivalently, to increases in the price of irrigation water. Particularly in the long run, farmers can adapt not only by changing the crop mix or reducing the area irrigated but also by substituting other inputs for water, such as capital (e.g., more efficient irrigation systems) or improved management (e.g., variations in the number of irrigation applications) (Young, 1986b; Gibbons, 1986). The specific objective of this paper is to demonstrate via a case study in northeastern Colorado that by allowing for these adaptation strategies the agronomic and economic impacts can be relatively limited over a wide range of irrigation water reductions.
STUDY AREA
The service area of the New Cache La Poudre Irrigating Company located near Greeley in northeastern Colorado was chosen as a study area. This irrigation company holds several rights to flow from the Cache La Poudre River, a major tributary to the South Platte River. The water can be used anywhere within the irrigation company's service area of about 40,000 acres. Water deliveries to individual farmers are based on their ownership of capital stock in the irrigation company, and the stock shares can be transferred freely within the service area (Maass and Anderson, 1978).
According to records of the Office of the State Engineer, the irrigation company distributes an average of 60,000 acre feet of irrigation water per season. Additional water is made available by pumping groundwater from the alluvial aquifer along the Cache La Poudre River and the South Platte River. Records in the Division Engineer's Office in Greeley show that there are 460 wells in the area which are entitled to pump up to 60,000 acre feet annually. More than three quarters of the service area are used to grow five major crops: alfalfa, corn for grain, corn for silage, dry beans, and sugarbeets. The predominant irrigation system is open ditch with siphons (55 percent), followed by flexible pipe (25 percent), gated pipe (15 percent), flexible pipe with surge (3 percent), and grated pipe with surge (2 percent).
Agronomic Model. Due to the lack of on-farm data of yield responses to irrigation water use for the study area, a previously developed agronomic model was used to synthesize the data and provided with parameters representing conditions in the study area. The model employed is known as the van Genuchten-Hanks model (Cardon and Letey, 1992a, 1992b). By combining soil-water, plant-water, and yield-evapotranspiration relationships it estimates yields for each of the major crops as a function of the quantity and quality of water applied at different points in time during a growing season. Water flow in the vertical direction is calculated by the Darcy-Richards equation. Plant water uptake is incorporated by the addition of a sink term, S(z,t):
where K is the soil hydraulic conductivity, C is the soil water capacity, h is soil matric pressure head, t is time, and z is soil depth (positive downwards). Transpiration and water and solute redistribution are calculated for a user-specified length of time until an irrigation, rainfall event, or adjustment in crop-dynamic variables occurs.
To take into account basic plant growth dynamics, growth-stage-specific stress tolerance and stress-induced growth reductions, the equation for the sink term used in the model is as follows:
where S(z,t) is the crop water uptake at depth z and time t, S'max(t) is the stress-adjusted value of the time-dependent Smax which is the maximum possible crop water uptake, h is the soil matric pressure head, is the osmotic head, 50 is the osmotic potential at which Smax is reduced by 50 percent, a(t) is a time-variable equal to 50/h50 where h50 is the soil matric pressure head at which Smax is reduced by 50 percent, p is an empirical constant (which is approximately 3 for many crops), and (z,t) is a depth- and time- dependent root distribution coefficient.
The model does not calculate crop yield directly, rather values of the sink term S(z,t) are summed for the season and must then be converted to yield. The following equation for yield-evapotranspiration relationships is used:
where Y/Ymax is relative yield, k is a crop-specific yield response coefficient, ET is evapotranspiration, and PET is potential evapotranspiration. Model-predicted values for relative yield were calculated by substituting cumulative S divided by cumulative S'max for ET/PET and using k values suggested by Doorenbos and Kassam (1979).
All data input was chosen to reflect the situation for crop production in the Greeley area. It was assumed that up to nine irrigation events may occur during the irrigation season, each with a net irrigation application of 3 inches. The dates chosen 6/1, 6/20, 7/10, 7/20, 7/30, 8/10, 8/25, and 9/8. On each of these dates crops can be irrigated or not irrigated, depending on the availability of water. This results in 29, or 512 combinations of irrigation events. All possible 512 combinations were used as model inputs to calculate the respective values for evapotranspiration and yield for each of the five crops. In Figure 1 model estimations of evapotranspiration and yield for alfalfa are presented. The estimations for the other crops are similar. These results show that enormous yield variations may occur depending on the timing of irrigation events. However, with careful management of available water the adverse yield impacts can be minimized. It is possible to reduce the number of irrigations and thus the quantity of water applied over a certain range without hardly impacting crop yields if the remaining irrigations are well-timed.
Economic Model. The yield estimations for the five crops as a function of the quantity of irrigation water applied provide the physical component of the economic model developed to estimate the impacts of reduced irrigation water use on the irrigation company's service area under alternative assumptions regarding adaptation strategies. A liner programming model was chosen, since it can be easily adapted to represent a range of irrigation water use availabilities. The model maximized the service area's net return to fixed factors of production and has the following general form:
where TNR is total net revenue, cj is unit net revenue of output j (revenue less variable and cost), xj is level of output j produced using the j th technological process, aij is amount of input i necessary for production of one unit of output j, and bi is total available quantity of input i.
Farmers are assumed to be economically rational, i.e. to apply irrigation water well-timed so that the highest possible yield for a given number of irrigations can be achieved. Thus in the case of alfalfa, only nine combinations of the 512 shown in Figure 1 need to be considered in the economic model. However, each combination can be used with any of the five surface irrigation systems typical in the study area. Depending on the respective irrigation efficiency, the amount of water actually applied varies with the irrigation system. Net revenue is calculated for each of the relevant combinations. The objective function is maximized subject to constraints on total irrigation water availability and total and individual crop acreage.
To analyze the effect of strategies for adapting to reductions in irrigation water, two scenarios are examined. Scenario A models economic impacts when reducing the area irrigated is the only adaptation allowed. Scenario B models economic impacts assuming farmers can adapt by varying the number of irrigation applications and/or by switching to a more efficient irrigation system. Both scenarios start out with a total irrigation water (surface and well) availability of 120,000 acre feet and a crop mix typical for the area. In scenario A additional constraints are formulated so that the current use of different irrigation systems remains unchanged. Initially with the full amount of irrigation water available, farmers are assumed to choose the number of irrigations for each crop under each irrigation system so that all water is used and net return maximized. As irrigation water is reduced (or as costs go up), farmers can adjust only by having crops gradually drop out of production one by one as they become uneconomic to produce. In scenario B, choices can also be made in the number of irrigations per crop or the irrigation system used.
DISCUSSION OF RESULTS AND IMPLICATIONS
Figure 2 shows the results of the economic model for scenarios A and B. For both scenarios irrigation water availability was varied and a linear programming solution found for each water quantity available, all other constraints remaining unchanged. When water availability is high, farmers would be expected to allocate water even to less valued crops and to irrigation systems with lower irrigation efficiency. As the water availability is reduced, the less valuable crops and the more water-intensive irrigation systems are excluded from the solution, and the marginal value of water rises. The set of shadow prices derived at various levels of water availability is a water demand schedule (the marginal value of water in Figure 2) for the irrigation company's service area. The model's objective function at various levels of water availability makes up the net return schedule shown in Figure 2.
There are important differences in the water demand schedules and net return schedules of scenario A and B. For purposes of water reallocation policy, the relevant parts of the schedules are those which show the impacts of the initial changes in irrigation water availability on the right hand side of the graphs. Under the assumptions of scenario A even relatively small reductions in water availability would result in significant losses in net return to farmers. Under the more realistic assumptions in scenario B water availability could be reduced by a quarter and hardly impact net returns. This is because farmers can adjust, for example, by applying not five but only four irrigations for alfalfa which has only a limited effect on yield. When irrigation water is further reduced, farmers can adapt again, this time for example, by changing the irrigation system from flexible pipe to flexible pipe with surge which allows to keeps the numbers of irrigation constant while reducing irrigation water use and only slightly reducing net revenue (due to higher irrigation costs).
The main implication of this study is that in the long run, when farmers have the opportunity to adjust their irrigation practices and the timing of irrigations is flexible, reduced irrigation water use does not necessarily lead to serious negative economic impacts on agriculture and rural communities. The models used in this study do not consider risk and uncertainty involved in irrigated agriculture, such as variations in rainfall, water availability, or crop prices, but at the same time they allow for only a few of many possible technical adaptations. therefore, it seems quite likely that in actual situations of reductions of irrigation water use in the south Platte-Cache La Poudre Area the agronomic and economic impacts would be less dramatic than often assumed. The model results indicate that efforts towards improving irrigation practices and reducing water use in agriculture are advisable.
References
Cardon, G.E., and J. Letey. "Soil-Based Irrigation and Salinity Management Model: I. Plant Water Uptake Calculations." Soil Sci. Soc. Am. J., 56, 1881-1887, 1992a.
Cardon, G.E., and J. Letey. "Soil-Based Irrigation and Salinity Management Model: II. Water and Solute Movement Calculations." Soil Sc. Soc. Am. J., 56, 1887-1892, 1992b.
Colby, B.G. "Enhancing Instream Flow Benefits in an Era of Water Marketing." Water Resources Res., 26, 1113-1120, 1990.
Dinar, A., and J. Letey. "Agricultural Water Marketing, Allocative Efficiency, and Drainage Reduction." J. Environmental Economics and Management, 20, 210-223, 1991.
Doorenbos, J., and A.H. Kassam. Yield Response to Water. Irrigation and Drainage Paper No. 33. Food and Agriculture Organization, Rome, 1979.
Gibbons, D.C. The Economic Value of Water. Resources for the Future, Washington, D.C., 1986.
Howe, C.W., D.R. Schurmeier, and W.D. Shaw, Jr. "Innovative Approaches to Water Allocation: The Potential for Water Markets." Water Resources Res., 22, 439-445, 1986.
Maass, A., and R.L. Anderson. ...and the Desert Shall Rejoice. Conflict, Growth, and Justice in Arid Environments. R.E. Krieger Publishing Company, Malabar, FL, 1978.
Scheierling, S.M. Ph.D. Dissertation. Colorado State University, Fort Collins, CO, in preparation.
Young, R.A. "Why Are there so few Transactions among Water Users?" Am. J. Agric. Econ., 68, 1143-1151, 1986a.
Young, R.A. "Local and Regional Economic Impacts." Water Scarcity: Impacts on Western Agriculture, eds. E.A. Engelbert with A.F. Scheuring, University of California Press, Berkeley, CA, 244-265, 1986b.
Return to 1995 Platte River Basin Ecosystem Symposium 
Last updated by Darren A. Jack on 9/8/97