Parameterization of predictive fate models: A case study

Abstract

Accurate prediction of chemical fate and persistence using general aquatic fate models requires model parameterization, i.e., the determination of site‐specific environmental and chemical parameters for model input. The capability of one model, the Exposure Analysis Modeling System (EXAMS), to predict the fate of endothall, an aquatic herbicide, in a reservoir was compared using two different parameterization methods. The first method, limited parameterization, used only literature and limited field data. The second method, intensive parameterization, employed laboratory, experimental pool and field data. Differences of less than one order of magnitude were observed among the endothall fate predictions from EXAMS in this reservoir using either method. Predicted endothall aqueous half‐lives were greater than the observed half‐life by a factor of 5 to 9. Predicted endothall concentrations in sediment were consistently below the minimum detectable level (0.01 mg kg−1) for endothall, whereas endothall concentrations were measured in reservoir sediments in the field. In this case, the results indicate that limited parameterization of EXAMS provides predictions of endothall persistence that are as accurate as those provided by intensive parameterization, thus saving time and reducing costs. Limited parameterization produced relatively accurate predictions in this study, possibly because only one fate process, biotransformation, was important. For chemicals affected by numerous fate processes, the errors associated with each fate process input could significantly affect the accuracy of predictions.

Photochemical oxidation of terbufos

“…As a result, primary fate processes such as transformations must be well understood and rates accurately measured to adequately describe the risk associated with these pesticides. Several studies using predictive fate models have found that results generated by the models were significantly different than observed values because process rates were inaccurate or unavailable [1,2].…”