A Decision Support System for Watershed-Scale Management of Ecosystem Services using Evolutionary Algorithms

Abstract

Increased production of ecosystem services (e.g., sediment and nutrient load reduction, flood peak reduction, wildlife habitat, and biodiversity) is likely to be achieved through strategic management of agricultural landscapes. This paper explores the role of landscapes in improving ecosystem service generation on a watershed scale. Four ecosystem service related objectives are considered; they involve reduction of flood peaks and sediment, nitrogen and phosphorus loads. The authors have developed a model to identify land uses and management practices that most cost-effectively generate these ecosystem services. Cost effectiveness is integrated by considering resulting impacts to landowner income, interpreted as a fifth objective. The spatial decision support model used to solve the underlying watershed management problem couples the U.S. Department of Agriculture’s Soil and Water Assessment Tool (SWAT) and a fairly new, multi-objective evolutionary algorithm known as SPEA2. Application of the model to the Big Creek watershed, a complex basin located in southern Illinois, demonstrates that it allows the assessment of tradeoffs between competing objectives, namely agricultural commo dity production and generation of ecosystem services. Moreover, it confirms that this integrative modeling approach could enable stakeholders and policy makers in identifying a suitable set of land use and management practices at a watershed scale.