Analyzing Heavily Censored Surface Water Pesticide Concentration Data Using Innovative Statistical Techniques

Surface water bodies are not only important for water supply but also vital for aquatic ecosystems and other important environmental and economic benefits. Surface waters are particularly vulnerable to pesticide contamination. Pesticides enter surface water bodies through runoff, wastewater discharges, atmospheric deposition, spills, and groundwater inflow. The uses and ecological significance of surface water, combined with its vulnerability to contamination, make it particularly important to understand the extent, long-term trends, and significance of pesticide exposure patterns in surface water systems. The presence of pesticides not only has adverse impacts on human health and the ecosystem but also incurs a relatively high operational cost and may cause secondary pollution such as sludge formation. Given the concerns for environmental safety and the likelihood of increased public health risks, monitoring pesticide concentrations in surface waters is important. Since 1990, California has required detailed reporting for all types of agricultural and non-agricultural pesticide uses. The California Department of Pesticide Regulation’s Surface Water Database has created a database to collect and make available information concerning the presence of pesticides in California surface waters. Pesticide monitoring data often contain a substantial number of samples where concentrations are below levels of quantification for the analytical methods employed. Due to this reason, conventional statistical techniques are not applicable. In this study, established statistical techniques for handling left-censored data, including the Kaplan-Meier product-limit estimator, robust regression on order statistics, and maximum-likelihood estimation, etc., are used to analyze pesticide concentrations in heavily censored datasets.