Statistical Approaches to Assessing Pesticideconcentrationsinthedpr Surfacewaterdatabase

Various studies on measured levels of pesticides in water have been completed and data from these studies are currently available on a surface water database maintained by the Department of Pesticide Regulation (DPR). W e examine portions of two databases in this report and suggest statistical methodology for approaching questions of importance in monitoring pesticide levels in water. The question of primary interest is the merging of concentrations, measured over time, space and sampling technique, into a coherent pesticide signal from a given region. This involves handling time series that are irregularly observed, due to missing observations or different sampling rates We consider models over time and space that lead to prediction limits for the pesticide signal during episodes where the concentrations may be exceeding regulatory standards. A recommended exploratory data analysis and modeling procedure is developed and applied to concentrations of various pesticides, focusing on chlorpyrifos and diazinon, measured in a Dow Chemical Study on Orestimba Creek and diazinon and simazine in a U.S. Geological Survey study along the San Joaquin River.