Expanding risk predictions of pesticide resistance evolution in arthropod pests with a proxy for selection pressure

Chemical resistance in pest organisms threatens global food security and human health, yet resistance issues are mostly dealt with reactively. Predictive models of resistance risk provide an avenue for field practitioners to implement proactive pest management but require knowledge of factors that drive resistance evolution. Despite the importance of chemical selection pressure on resistance evolution, poor availability of chemical usage data has limited the use of a general multi-species measure of selection pressure in predictive models. We demonstrate the use of pesticide product registrations as a predictor of resistance status and potential proxy of chemical selection pressure. Pesticide product registrations were obtained for 427 USA and 209 Australian agricultural arthropod pests, for 42 and 39 chemical Mode of Action (MoA) groups, respectively. We constructed Bayesian logistic regressions of resistance status as a function of the number of pesticide product registrations and two ecological traits, phagy, and voltinism. Our models were well-supported with demonstrated power to discriminate between resistant and susceptible observations in both USA and Australian species sets using cross-validation. Importantly, we observed strong support for a positive association between pesticide products and resistance status. Our work expands the horizon for proactive management by quantitatively linking a proxy for selection pressure on pest species to different chemical MoAs, which can be combined with ecological information to build models of resistance evolution risk. Because pesticide product registrations can typically be obtained from publicly available data, we believe they have broad applicability for risk predictions in other agricultural pests, such as weeds and fungi, and to other geographical regions beyond the USA and Australia.

Automated summary

Chemical resistance in pest organisms threatens global food security and human health. Predictive models of resistance risk provide a proactive approach to pest management, but limited chemical usage data hinder the use of a general multi-species measure of selection pressure. This study demonstrates the use of pesticide product registrations as a proxy for chemical selection pressure and resistance status, and provides a quantitative link between selection pressure and different chemical Mode of Action (MoA) groups. The findings have broad applicability for risk predictions in agricultural pests and other geographical regions.