Assessing Pesticide Risk Using Mesocosm Studies

OUR CHALLENGE

Our client was seeking a more realistic way to model potential pesticide risks. Doing that involves an incredible number of variables. We proposed using data from existing mesocosm (outdoor experimental systems under controlled conditions used to study ecosystems) studies and modeling software to create a model generic farm pond to then model against known toxicity data for the pesticide.

OUR APPROACH

We established the important project data quality objectives (DQOs) from the intended use of the model and metrics for characterizing populations and the ecosystem. Decisions were made about the fit of the model to scenarios that might occur in the ecosystem including checking these assumptions against observations of patterns occurring within the mesocosms. The model was parametrized to reflect a generic farm pond food web at a single geographic region in central Europe and included representative phytoplankton, zooplankton, periphyton, macrophytes, benthic and phytophilous macroinvertebrates, and fish. We calibrated the generic farm pond model, which entailed running successive iterations of the model with slight changes in selected parameters until the project team reached consensus that the model was sufficiently consistent with the project DQOs.  Integral applied its 3-Factor Approach for Model Testing during examination of model calibration. We then completed model testing and analysis which involved verification, sensitivity analysis, and validation.

 

OUR IMPACT

Integral developed a mesocosm mimic model that was calibrated to consider the original studies’ seasons and durations. The calibrated mesocosm mimic model was then applied to evaluate alternate seasons, durations, food web compositions, toxicity thresholds, and exposure profiles beyond those included in the original mesocosm studies. We were able to establish model persistence under control conditions in different individual seasons, across multiple seasons, and across successive years. Using both experimental data and mathematical modeling, we assisted our client to evaluate potential effects of the pesticide beyond the temporal, spatial, and biological scenarios represented by the mesocosm studies.