Prioritization of Farmlands in a Multicriteria Irrigation Water Allocation: PROMETHEE and GAIA Applications

Abstract. In this study, we propose irrigation water allocation as a multicriteria problem and demonstrate the potential for application of PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluations), a multicriteria decision making (MCDM) method, for solving a water allocation problem in the lower Indus basin of Pakistan. Five criteria were used for allocating irrigation water to 16 farmlands: (1) increasing dependency on agricultural income, (2) generating revenue for improving operational and maintenance capability of the provincial irrigation department (PID), (3) enhancing water allocation efficiency, (4) increasing annual farm income, and (5) supplementing canal water deficiencies with fresh groundwater. The relative importance of the five water allocation criteria was determined based on a survey of 62 farmers located in the lower Indus basin of Pakistan. Increasing annual farm income was found to be the most important criterion in water allocations, as the highest weights (37.8%) were attached to this criterion. The survey participants assigned the lowest weight (only 4%) to the “generating revenue for improving operational and maintenance capability of PID” criterion, which makes this the least important criterion to be considered in water allocations. Depending on the performance of these 16 farmlands for the five water allocation criteria, a rank order of the farmlands competing for water allocations was developed using the PROMETHEE method through D-Sight software. Farms 10, 4, and 6 were found to be the first, second, and third most preferred farmlands to which irrigation water should be allocated on priority basis. Farm 15 was identified as the least preferred farmland, as the lowest net flow score (I• = -0.544) was calculated for this farmland. We stress that the application of the proposed MCDM water allocation method in larger irrigation systems such as the warabandi system may be difficult, as estimating values for the decision criteria may not be feasible in terms of money, human resources, and time. However, this method could be applied in smaller-scale irrigation schemes and closed watersheds where the estimation of water allocation criteria is economical and consumes relatively less time.