Validation trial of predictive fate models using an aquatic herbicide (endothall)

Abstract

Environmental risk assessment of pesticides in aquatic environments requires predictions of their persistence and compartmentalization. A strategy for developing confidence in predictive fate models, such as the Exposure Analysis Modeling System (EXAMS) and the Simplified Lake and Stream Analysis model (SLSA), is to test the models using carefully chosen chemicals in semi-controlled and field situations. For example, endothall, a relatively water-soluble aquatic herbicide, essentially has a sole fate process, biotransformation. This dicarboxylic acid was used to test the predictive capabilities of EXAMS and SLSA and to identify sources of variance in those predictions. The models were parameterized using laboratory, experimental pool and field measurements. Persistence and compartmentalization of endothall in the water, sediments and aquatic plants (Eurasian watermilfoil-Myriophyllum spicatum L.) contained in experimental pools were measured using gas chromatography. Both EXAMS and SLSA predicted aqueous compartment half-lives ranging from 7.3 to 7.8 d, whereas an aqueous half-life of 4 d was observed in the experimental pools. Endothall was introduced to areas within Pat Mayse Lake, a 2,400-ha reservoir in north central Texas, as part of an aquatic plant management program for M. spicatum. Concentrations of endothall were below detection limits (0.002 mg L water, 0.01 mg kg sediment) in 2 to 3 d and in 4 d in the water and sediment compartments, respectively. EXAMS and SLSA predicted half-lives ranging from 3 to 6 d in the water column. Observed aqueous half-lives ranged from 0.1 to 0.23 d. Horizontal dispersion and dilution contributed significantly to the dissipation of endothall at these sites. Predicted concentrations of endothall in sediment were similar to the measured concentrations. This study is one validation trial of EXAMS and SLSA in which the models were not considered validated.

Parameterization of Predictive Fate Models: A Case Study

“…1). However, only one site, Lamar Point, was chosen for intensive parameterization and fate study [4,6]. The Lamar Point study site received a measured initial dose of 0.41 mg L-' (acid equivalents) endothall.…”

“…More applicable data were not available from other studies because endothall concentrations in target plant species are usually not reported because of the rapid death and decay of the plants [ 101. Chemical-and site-specific model inputs for the intensive parameterization method were obtained using laboratory studies of sediment sorption [ 111, sediment penetration [4] and biotransformation [12] of endothall (Table 3). Pat Mayse Lake water and sediments were utilized in these studies.…”

“…Intensive parameterization, the second method, employed laboratory-, experimental pool-and field-derived site-and chemical-specific input data. The intensively parameterized model was studied first as a validation trial of EXAMS [4], which was presented at the Fifth Annual Meeting of the Society of Environmental Toxicology and Chemistry in November 1984. This article compares the predictive capability of EXAMS in a north Texas reservoir treated with endothall, a relatively water-soluble aquatic herbicide, using two methods of parameterization.…”

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].…”

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].…”