You are here:
Integration of Climate Model Projections and Pesticide Application Scenarios for Probabilistic Risk Assessment with a Bayesian Network Approach
Citation:
Moe, J., S. Mentzel, R. Benestad, John F. Carriger, J. Hader, T. Kunimitsu, R. Nathan, R. Oldenkamp, A. Pitman, AND W. Landis. Integration of Climate Model Projections and Pesticide Application Scenarios for Probabilistic Risk Assessment with a Bayesian Network Approach. Presented at SETAC North America 2022 meeting, Pittsburgh, PA, November 13 - 17, 2022.
Impact/Purpose:
To present case study approach and development from a SETAC Pellston Workshop on incorporating climate change into risk assessment at an international meeting.
Description:
We present a Northern European case study from the SETAC Pellston workshop in June 2022 on integration of global climate change (GCC) modeling into ecological risk assessment. In Northern Europe, GCC is expected to result in increased temperatures and precipitation. The changes in weather patterns are expected to increase the occurrence of crop pests such as weeds, fungal disease, and insect pests. Increased pest pressures can in turn be expected to alter agricultural practices such as the frequency and combination of pesticide applications. Additionally, GCC may potentially have more direct effects on the environmental exposure of pesticides through changes in the transport, fate, and degradation of pesticides. A Bayesian network (BN) has previously been developed as a meta-model for incorporating future climate projections and pesticide application scenarios with data on toxic effects to support environmental risk assessment for streams in agricultural areas in Northern Europe. This BN model was initially parameterized for a Norwegian case study with predicted environmental concentrations from a process-based pesticide exposure model and species sensitivity distributions derived from toxicity tests data. Within the Pellston workshop, we aim to improve the existing BN model by incorporating more recent and realistic climate change scenarios, a larger number of climate models, and better methods for regional downscaling. The exposure prediction model WISPE has already been calibrated for the case study area with information local conditions and chemical properties of the pesticides used. This model can now be run using alternative climate model projections, as well as more realistic pesticide application scenarios. Our experiences from this case study will aid efforts to better account for uncertainty related to climate change in exposure modeling, effect assessment, and risk characterization. The graphical display of the BN model approach can also aid communication of risk under climate change scenarios to stakeholders such as policy makers and regulators. The views expressed in this presentation are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency.