A STEP Method based multiple objective methodology for irrigation water management to model preferences and tradeoffs

Management decisions relating to river basins are difficult due to the complex interaction of biophysical, economic and social variables. Moreover, conflicting objectives of different stakeholders have to be considered. Lack of adequate information to determine water allocations, environmental flow requirements and the increasing intensity of cropping systems require a better seasonal distribution of water to satisfy consumptive and in-stream environmental demands. Water demand management that considers system constraints on water conveyance and losses in addition to environment requirements will result in optimum productivity of irrigation areas and better management of river flows. Many decision support systems in agricultural enterprises use conventional linear programming approach to optimize a single objective function such as total gross margin. However, as agricultural systems become more complex, multiple objectives that are in conflict with each other need to be addressed. Mathematical programming techniques are required to formulate the problem and find a compromise solution. A methodology based on the techniques of STEM (Step Method) is provided that allows for the progressive articulation of preferences. This is an iterative procedure that narrows the region on the “efficiency frontier” in which the final compromise solution is found. The procedure involves iterations where preferences are based on the objective space of previous iterations. The decision making process is entirely in the objective space and results are presented in the form of graphs and tables of objective values and utility values. A utility function is used to select the best objective function at each iteration and can take various forms (linear, convex or concave curves). The choice of the utility function is subjective and there is scope for investigating various utility functions. We have adopted a linear utility function in this study. The technique is applied to a hypothetical nodal network shown in Figure 1. For the purpose of this paper, the irrigated area is divided into eight regions with a total irrigable land of 121,808 ha and a potential for growing fourteen crops (rice, wheat, oats, barley, maize, canola, soybean, winter pasture, summer pasture, lucerne, vines, summer vegetables, winter vegetables, citrus and stone fruit). In the current analysis groundwater pumping under the irrigable area is permitted to satisfy crop water demand if surface water supplies are not sufficient.