A GIS-Based Ground Water Contamination Risk Assessment Tool for Pesticides by Michael G. Sinkevich Jr., M. Todd Walter, Arthur J. Lembo Jr., Brian K. Richards, Natalia Peranginangin, Sunnie A. Aburime, and Tammo S. Steenhuis Abstract A process-based preferential flow transport model was implemented in a geographic information system to locate areas in the landscape with high risk of contamination by agrochemicals, especially pesticides. Protecting ground water resources necessitates a reliable ground water quality monitoring strategy. It is valuable to be able to focus monitoring on areas with the highest risk of contamination because monitoring ground water is an expensive activity, especially at the landscape scale. The objective of this project was to develop a tool that quantifiably estimates distributed ground water contamination risk in order to develop reliable, cost-effective ground water observation networks. The tool is based on a mechanistic model of chemical movement via preferential flow and uses land cover data, information about chemical properties, and modeled recharge to estimate the concentration of chemical reaching the ground water at each point in the landscape. The distributed risk assessment tool was tested by comparing the model-predicted risk with observed concentrations from 40 sampling wells in Cortland County, New York, for atrazine (pesticide) and nitrate, the latter assumed to be an indicator of agricultural pollu- tion. The tool predictions agreed well with observed nitrate concentrations and pesticide detections. An Internet-based ver- sion of this tool is currently being developed for ready application to New York State. Introduction Ground water is an important natural resource that should provide drinking water for future generations. How- ever, in recent decades ground water pesticide contamina- tion from agriculture has become a problem that requires monitoring. Because such monitoring is expensive— especially for ground water over large areas of agricultur- ally dominated landscapes—reliable and flexible tools are needed to identify potential hazard areas in the landscape so that monitoring strategies can focus on the highest risk areas. By focusing monitoring activities on the highest risk areas, fewer observation wells are needed and the potential cost and detection effectiveness of the monitoring is improved. The goal of this project was to develop a distrib- uted landscape-scale, physically based ground water risk assessment tool that can be implemented with readily available open-access data. While landscape-scale water quality risk assessment has been revolutionized by the advent of geographic infor- mation systems (GIS), most, if not all, risk assessment tools regularly used by water quality professionals employ logical factor- or index-based rubrics for assessing risk that are only loosely linked to physical processes and fail to predict actual chemical concentrations. This is espe- cially true for ground water (e.g., Grayman 1977; Hamlett et al. 1992; Nizeyimama et al. 1996; Peterson et al. 1996). Traditional mechanistic predictions of ground water risk are generally only applied at a single point in the landscape and assume that pesticide leaching through soils is accu- rately characterized by the convective-dispersive equation, which neglects preferential flow phenomena. The reliable, mechanistic or physically based models developed to pre- dict the ground water contamination from land-applied chemicals that meaningfully consider preferential flow (e.g., Wagenet and Hutson 1986; Ramos and Carbonell 1991; Steenhuis et al. 1987; Steenhuis and Parlange 1991; Nguyen et al. 1998; Kim et al. 2005) have not been incor- porated into GIS to assess contamination risks at a land- scape scale, probably because they require either a copious amount of input data or data that are not readily available. Perhaps the most important transport process to be included in ground water risk assessment is preferential flow, the rapid nonuniform transport of solutes via these flowpaths, which can result in contaminants reaching the ground water before they degrade or can be adsorbed by the soil (e.g., Stagnitti et al. 1994; Camobreco et al. 1996). Indeed, the surprise discovery of pesticide contamination of the Long Island aquifers in the early 1980s clarified that ground water contamination by toxic chemicals cannot be Copyright ª 2005 The Author(s). Journal Compilation ª 2005 National Ground Water Association. 82 Ground Water Monitoring & Remediation 25, no. 4/ Fall 2005/pages 82–91