Predicting Disease in the Field before It Becomes an Issue

Combining DNA results from airborne samplers placed in active cotton fields with climate, phenological, and agronomic data, we have been able to not only model for but provide early warning for disease emergence as well as other key values of interest across the US.  This system provides unique insights into crop risk at a field level, without waiting for diseases to become visible to sensors or the naked eye, and enables precise timing of pre and post emergent control methods, optimizing cost for the farmer in the field.  A further component, allowing direct feedback from the farmer extends continuous learning to continually improve and personalize recommendations and predictions for individual fields. The system also enables rapid deployment of existing models to a broad audience as well as enabling existing models to be improved or tuned automatically based on field observations. Use of novel airborne spore and bacteria trapping filters that utilize bio-active proteins to ensure accurate pathogen capture has enhanced our ability to further quantify models around biological load as well conditional probability for a disease outbreak. Corynespora, Alternaria, and others pathogens of concern were used for the initial approach validation, and we have since expanded the system to allow for modeling based on other research and field data to allow for predictions around other pathogens, insects, and weeds at both a field and regional level.