Regional sensitivity analysis using a fractional factorial method for the USDA model GLEAMS

A sensitivity analysis using the fractional factorial design technique of Plackett and Burman (PB) for the Ground water Loading Effects of Agricultural Management Systems (GLEAMS) model is presented. A software system has been developed for automated implementation of the PB technology. The application can Vary up to 54 GLEAMS inputs and deduce input parameters of importance. Characteristic pesticide runoff and leaching results for the GLEAMS model are presented using seven geographically diverse soil/climate scenarios and two pesticides characteristic of a hydrophobic [chlorpyrifos: (O,O-diethyl-O-(3,5,6-trichloro-2-pyridinyl)phosphorothioate)] and hydrophillic [atrazine: (2-chloro-4-(ethylamino)-6-(isopropylamino)-s-triazine)] chemical. Scenarios representative of the Northern and Southern High Plains wheat, Southeast peanut, Midwest corn, Red River Valley sugar beet, Delta cotton, and Mid-Atlantic tobacco markets are presented. Geographic conditions (weather, soil, crop, and management practices) and physicochemical pesticide properties define both the number and sensitivity ranking of important GLEAMS inputs, illustrating the importance of specific field scale sensitivity analysis. In general, several parameters governing water hydrology were consistently found as sensitive. These include runoff curve number, soil porosity and soil evaporation parameter, with the runoff curve number being the most sensitive GLEAMS input parameter. Pesticide properties of importance were the soil/water equilibrium partition coefficient and soil degradation half-life, with the sensitivity of these inputs largely dependent upon the nominal value initially chosen and geographic region simulated. (C) 1999 Elsevier Science Ltd. All rights reserved.