Nonpoint-Source Pollution Reduction for an Iowa Watershed: An Application of Evolutionary Algorithms

"We apply an integrated simulation-optimization framework to search for cost-efficient mix and location of agricultural conservation practices in a typical agricultural watershed for two types of nitrogen reduction targets: control of mean annual nitrogen loadings, and a "safety-first" type constraint, insisting that nitrogen targets be met in every weather realization (weather-resilient solutions). Evolutionary algorithms are developed for each of the appropriate water quality targets. Our approach allows for the derivation of a watershed-level total and marginal nitrogen abatement cost curve. Controlling for the probability of meeting water quality targets (looking for weather-resilient solutions) is found to be significantly more costly than controlling the average nitrogen loadings. Both types of solutions are assessed for robustness with respect to weather uncertainty: solutions selected to reduce average loadings do well under weather uncertainty, while the robustness of solutions selected to be resilient decreases with the stringency of the water quality goal". Copyright (c) 2010 Canadian Agricultural Economics Society.