Validation and evaluation of weather-based predictive decision support systems for frogeye leaf spot of soybean in the United States.

Frogeye leaf spot (FLS), caused by Cercospora sojina Hara, is a major foliar disease of soybeans (Glycine max) worldwide. While a fungicide application is often recommended at the beginning of the pod fill (R3) growth stage, farmers lack a decision-support system (DSS) to guide whether and when to apply. A model consisting of three risk-based action thresholds was developed in 2024, and field trials were performed in 2024 and 2025 across 37 site-years and 15 states to validate it. Our objectives were to evaluate the effects of DSS-based fungicide applications on FLS suppression and yield protection, and to assess the predictive model's performance under field conditions. FLS severity levels were low in most site-years. In a combined-environment analysis of FLS severity, spraying at the R3 growth stage, and at a 40% or 50% action threshold outperformed the non-treated control (NTC). Yield was improved in the R3 application and low-threshold treatments compared to the NTC. At individual site-years, the model performed as well as the standard R3 spray in most sites but overestimated FLS risk in others. Overall model accuracy at the highest-risk threshold reached 86% and performed better at the lower FLS severity sites than at higher severity sites. Sensitivity was best with the lowest-risk action threshold, while specificity reached 100% at the highest-risk action threshold. This work is the first attempt to develop and implement a DSS based on predictive models for timing fungicide applications for FLS in the U.S.