A fuzzy knowledge-based model for assessing risk of pesticides into the air in cropping systems

Pesticide use in current cropping systems has become a key input to improve productivity. However, their potential risk to nature demands tools for designing a sustainable use. In this work, a fuzzy knowledge-based model was devel-oped for assessing risk of pesticides into the air. The model was based on fuzzy logic theory which provides a means for representing uncertainty by including knowledge about different processes related to pesticide dynamics using func-tions, control rules and logical inference systems. All these elements were built through a literature review. Results from the sensitivity analysis on the final model structure showed that the Henry's law constant was the most influential input variable related to the active ingredient identity, while the most influential management and environmental input variables on the pesticide air risk values were the droplet size together with the application method and the cur-rent wet bulb temperature depression value, respectively. Results from an independent model validation showed a sig-nificant goodness-of-fit between the simulated risk of drift and volatilization and the observed values under experimental conditions. Long-term simulations in a real soybean production system in Argentina showed results of drift reduction in post-emergence conditions of the crop under aerial application condition, and a significant effect of the identity of the active ingredient in the risk values. Simulated risk values from the developed model allow to iden-tify ex ante the combination of agronomic decisions, together with environmental conditions that can reduce the risk of pesticides in the air in real production systems. Further combination with ecotoxicological classification tools should improve pesticide use assessment in agricultural systems.