Probabilistic risk assessment of pesticides under future agricultural and climate scenarios using a Bayesian network

Abstract

AbstractThe use of Bayesian networks (BN) for environmental risk assessment has increased in recent years. One reason is that they offer a more transparent way to characterize risk and evaluate uncertainty than the traditional risk assessment paradigms. In this study, we explore a new approach to probabilistic risk assessment by developing and applying a BN as a meta-model for a Norwegian agricultural site. The model uses predictions from a process-based pesticide exposure model (World Integrated System for Pesticide Exposure - WISPE) in the exposure characterization and species sensitivity data from toxicity tests in the effect characterization. The probability distributions for exposure and effect are then combined into a risk characterization (i.e. the probability distribution of a risk quotient), which is a common measure of the exceedance of an environmentally safe exposure threshold. In this way, we aim to use the BN model to better account for variabilities of both pesticide exposure and effects to the aquatic environment than traditional risk assessment. Furthermore, the BN is able to link different types of future scenarios to the exposure assessment, taking into account both effects of climate change on pesticides fate and transport, and changes in pesticide application. We used climate projections from IPCC scenario A1B and two global circulation models (ECHAM5-r3 and HADCM3-Q0), which projected daily values of temperature and precipitation for Northern Europe until 2100. In Northern Europe, increased temperature and precipitation is expected to cause an increase in weed infestations, plant disease and insect pests, which in turn can result in altered agricultural practices, such as the use of new crop types and changes in pesticide application patterns. We used the WISPE model to link climate and pesticide application scenarios, environmental factors such as soil properties and field slope together with chemical properties (e.g. half-life in soil, water solubility, soil adsorption), to predict the pesticide exposure in streams adjacent to the agricultural fields. The model was parameterized and evaluated for five selected pesticides: the herbicides clopyralid, fluroxypyr-meptyl, and 2-(4-chloro-2-methylphenoxy) acetic acid (MCPA), and the fungicides prothiocanzole and trifloxystrobin. This approach enabled the estimation and visualization of probability distribution of the risk quotients representing the alternative climate models and application scenarios for the future time horizons 2050 and 2075. The currently used climate projections resulted in only minor changes in future risk directly through the meteorological variables. A stronger increase in risk was predicted for the scenarios with increased pesticide application, which in turn can represent an adaptation to a future climate with higher pest pressures. Further advancement of BN modelling as demonstrated herein is anticipated to aid targeted management of ecological risks in support of future research, industry and regulatory needs.