Estimation of concentration percentiles for pesticide surface water monitoring data

Abstract

Presentation final title:
Upper bound concentration estimates based on pesticide surface water monitoring data

Abstract:
Extensive surface water monitoring data exist for many pesticides from various water quality monitoring programs, each often having differing objectives. These data may be collected at different frequencies and/or times of sampling or targeted at watersheds vulnerable to storm-induced runoff from agricultural fields. However, data from many such sites and many years should contain information from both the high end as well the low end of the likely concentration distributions embodied in the historical multi-year and multi-region monitoring.
We will describe a weighting method that uses data of both low and high frequency sampling to obtain conservative estimates of “typical peak concentrations” at site-years within spatial regions. By “typical” we mean the median of the distribution of the target annual percentile (e.g. 90th, 95th, 99th) across site-years. The method is illustrated using data obtained from the Atrazine Ecological Monitoring Database (AEMD) which includes all atrazine measurements collected at extensive monitoring programs in the United States, both targeted and non-targeted. Percentile estimates obtained are shown to be conservative by empirical comparison to direct estimates obtained using sites where higher frequency sampling data is available.