Environmental Modeling and Assessment 6: 195–208, 2001.  2001 Kluwer Academic Publishers. Printed in the Netherlands. Characterizing agrochemical patterns and effective BMPs for surface waters using mechanistic modeling and GIS Steven A. Cryer a,∗ , Michael A. Fouch b , Alan L. Peacock a and Patrick L. Havens a a Dow AgroSciences, 9330 Zionsville Road, Indianapolis, IN 46268, USA b Sphere Software, Inc., 820 Hendricks Drive, Lebanon, IN 46052, USA Monitoring programs in the agriculturally intense San Joaquin River Valley of California have periodically found organophosphate (OP) insecticide concentrations, predominantly chlorpyrifos, diazinon and methidathion, at levels high enough to cause mortality for the aquatic invertebrate Ceriodaphnia dubia. These detections are likely the result of off-site movement from treated fields. However, the relative significance and magnitude of off-site transport pathways cannot be readily deduced from monitoring data alone. Therefore, a comprehensive modeling system has been constructed to estimate temporal and spatial pesticide source magnitudes and to follow the pesticide dissipation pathways once in surface water. The USEPA models HSPF and PRZM3 were used for the hydrology and non-point source predictions, respectively. Spray drift was accounted for using the mechanistic model AgDrift. The Orestimba Creek Watershed in the San Joaquin Valley was characterized and used as a typical watershed for this region. Representative transport pathways were ranked and quantified, and numerical implementation of best management practices (BMPs) determined which practice may have the highest likelihood for reducing pesticide loadings. Approximately 85% of the predicted chlorpyrifos mass detected between May 1, 1996, and April 30, 1997 resulted from drift, with the largest contributions coming from walnut orchards immediately adjacent to Orestimba Creek. Various simulated drift mitigation measures suggest chlorpyrifos mass loadings can be decreased by over 90% depending upon the type of mitigation chosen. Imposed drift BMPs should be effective in reducing chlorpyrifos levels found in surface waters of the San Joaquin valley if the Orestimba creek watershed is considered representative of watersheds found in this area of California. Keywords: exposure, chlorpyrifos, drift, pesticide, runoff, water quality 1. Introduction The U.S. Geological Survey’s (USGS) National Water- Quality Assessment (NAWQA) Program is the largest scale environmental monitoring program in the United States. This program was designed to populate a historical data- base such that trends in the quality of the United States’ surface and groundwater resources can be addressed. Fifty- nine study units are located throughout the country as part of this program (http://water.wr.usgs.gov/pnsp/). Both NAWQA [1] and the California Environmental Protection Agency (EPA) [2,3] have been collecting surface samples in the San Joaquin Valley of California and analyzing the wa- ter for organic contaminates and/or toxicity using the sensi- tive aquatic invertebrate Ceriodaphnia dubia. Unfortunately, only qualitative information about many significant issues (non-point source locations, types of sources, effectiveness of management practices, and so forth) can be approximated using this monitoring data. Monitoring has indicated several of the surface water samples collected during different times of the year have the potential to acutely affect non-target organisms such as Ceriodaphnia dubia (an organism extremely sensitive to Organophosphate (OP) insecticides such as chlorpyrifos). In fact, samples found toxic to Ceriodaphnia dubia contain OP insecticide concentrations of sufficient magnitude to account for the observed toxicity [2]. Detections of acute insecticide concentrations found in surface water tend to be episodic ∗ Corresponding author: Tel.: (317) 337-3428; Fax: (317) 337-3628; E-mail: sacryer@dowagro.com. in nature and of modest magnitude such that effective best management practices may be able to reduce these peak con- centrations below toxic threshold values. USGS or California EPA sampling has generally been conducted at a single location on a tributary near its con- fluence with the San Joaquin River and at sampling intervals of one to two weeks. This sampling approach was accept- able for addressing long-term trends but has the potential to miss peaks of contaminants that may be transported through the tributary and into the river between the sampling inter- vals. Additionally, these monitoring programs have not char- acterized the specific crops, soil types, agricultural manage- ment practices, and so forth, for the agricultural fields within the watershed which are presumed to be the source for the insecticides being monitored. This makes correlating mea- sured exposure concentrations with non-point sources diffi- cult, even with the use of existing numerical modeling tools. The relative contributions of different transport processes for pesticide movement such as surface runoff, spray drift, or ir- rigation tail-water were not defined or accounted for. A more comprehensive monitoring program has been per- formed on Orestimba Creek, a tributary of the San Joaquin River [4]. This study was designed to provide specific mon- itoring information for chlorpyrifos, one of the insecti- cides detected in previous surface water monitoring stud- ies. Daily time weighted water samples from three locations in Orestimba Creek were taken and analyzed for chlorpyri- fos residues over a period of one year (05/01/96 through 04/30/97). Thus, resolution of daily peaks and pulses of