Multiple criteria analysis of economic, environmental and water use problems in an irrigation district of Mexico

Abstract

The Alto Rio Lerma Irrigation District (ARLID), located in the state of Guanajuato in Mexico, is an agricultural area whose sustainability depends partially upon groundwater withdrawal for crop irrigation. Because of high pumping demands and current land-management practices, water levels in the regional groundwater aquifers have declined severely, resulting in aquifer overdraft. The potential adverse consequences of this overdraft cannot be overstated, as the long-term economic viability and environmental integrity of the region is threatened. In order to analyze this economic, environmental, and water use problems in this region, simulation of the agricultural system was performed and associated water use impacts quantified under different management scenarios. Linear programming identified the 12 optimal cropping patterns, and then multi-criteria decision-making methodologies were applied to rank and identify the best cropping pattern (satisficing solution). The GLEAMS model was used to simulate the amounts of water, nitrate, and pesticides in both runoff and percolation for each cropping pattern. In order to quantify the economic and environmental impacts of aquifer overdraft, two attributes were used; pumping costs and an aquifer exploitation coefficient. Three multiple criteria methods: Q-analysis, ELECTRE II, and the Range of Value Method, were evaluated and the latter method was selected to analyze the payoff matrix for the ARLID, where 12 alternatives, each with 13 attributes, were considered. The results show the best alternative for effectively balancing environmental with economic considerations was the farming practice, consisting of land leveling, growing vegetables such as red tomato, and controlled groundwater withdrawals to preserve aquifer sustainability. Low water availability does not necessarily reduce farmer's profits, if new water saving alternatives are applied, combined with more profitable crops. Given that ROV method only considers full compensation between objectives, this research contemplate an extension for the nonlinear case in order to assess the whole range of values of the objective function. The final findings suggest that the best methodology was to apply L1 in the lower levels and then L2 for the highest level in the hierarchy. In this manner, the partial compensation between objectives is considered and also the decrease in the uncertainty of alternatives selection.