Location: Crop Production and Protection
Project Number: 0500-00102-001-096-A
Project Type: Cooperative Agreement
Start Date: Aug 10, 2026
End Date: Aug 9, 2027
Objective:
This project will utilize a data modeling framework previously developed to create predictive disease risk tools for soybeans and pulse crops. For soybeans, extensive disease and spore trap datasets have already been collected across multiple regions of the United States for frogeye leaf spot. The spore trap data includes the detection of multiple airborne foliar soybean pathogens that have not yet been modeled and would be highly desired by farmers across the U.S. These datasets, including ongoing contributions from the 2025 growing season, will be aggregated and aligned with high-resolution weather data to develop and validate regional prediction models.
Weather inputs will initially utilize the IBM Environmental Intelligence Suite, which has been successfully used in prior disease modeling efforts, while simultaneously evaluating alternative weather data sources to improve flexibility and long-term deployment potential.
Approach:
Model development will closely follow methodologies previously established in the Corn RAC, particularly those used in the development of tar spot prediction models. These approaches incorporate robust feature selection, machine-learning, and logistic regression modeling techniques, and independent validation to ensure biological relevance and practical usability by producers. To address existing knowledge gaps, this project will also emphasize the need for expanded, high-quality datasets. New field trials will be initiated to collect standardized disease ratings, and then environmental data can be incorporated into future model iterations. These efforts directly address current limitations in spatial and temporal coverage that restrict our predictive accuracy.