Coupling field‐scale and watershed models for regulatory modeling of pesticide aquatic exposures in streams

Abstract

Estimating exposure in receiving waterbodies is a key step in the regulatory process to evaluate potential ecological risks posed by the use of agricultural pesticides. The United States Environmental Protection Agency (USEPA) currently uses the Variable Volume Water Model (VVWM) to predict environmental concentrations of pesticides in static waterbodies (ponds) that receive edge-of-field runoff inputs from the Pesticide Root Zone Model (PRZM). This regulatory model, however, does not adequately characterize potential pesticide concentrations in flowing water systems (streams and rivers) drained from watershed areas. This study aims at addressing this gap by coupling the regulatory PRZM model with a watershed-level hydrological model, the Soil and Water Assessment Tool (SWAT), to predict pesticide concentrations in flowing water habitats for aquatic organisms. This coupled PRZM-SWAT model was applied in a test watershed (~HUC12), a headwater watershed of Goodwater Creek in Missouri, and simulation results at the outlet of this watershed were compared to daily and near-daily measured streamflow and atrazine concentration data from a decade-long sampling campaign. Overall, the PRZM-SWAT model captured (1) the general magnitude and temporal trend of daily atrazine concentrations, (2) the observed high-end of exposure levels (>3 ppb) of atrazine concentrations, and (3) the 90 th centile annual maximum for various exposure durations (1-, 4-, 7-, 21-, and 60-day rolling average), which are important exposure metrics used in assessing the potential ecological risks posed by the application of pesticides. The PRZM-SWAT model is expected to expand the utility of the field-scale regulatory model to include pesticide exposure prediction capability in flowing waterbodies from agricultural watersheds.

Proposed schemes on more integrative ecological risk assessment of pesticides

“…These researchers also highlighted the need for and the potential of integrating retrospective with prospective ERAs for pesticides. A different example of model integration was presented recently by Ghebremichael et al (2022) for the assessment of pesticide exposure in flowing waterbodies. Finally, a critical review of how population effect models integrate data from different sources distinguished three population model types: unstructured, structured, and agent‐based (Accolla et al, 2020).…”