Mapping High-Risk Disease Zones in Strawberry Fields Using Drone Imagery and Random Survival Forests

Early detection of canopy decline in strawberry production is essential for timely management, yet visual scouting often misses subtle or spatially heterogeneous symptoms. We developed a UAV-based monitoring framework that integrates multispectral imagery, plant-level canopy metrics, clustering, and Random Survival Forest (RSF) modeling. This framework was used to predict the onset and spatial progression of soilborne pathogen-associated canopy decline in three commercial strawberry fields in Oxnard, California. Nine UAV surveys collected from December 2022 to June 2023 were processed into 159,220 plant-level monitoring units. NDRE- and Redness Index–based classifications quantified proportional and absolute canopy dieback within standardized hexagonal units and supported a time-to-event modeling approach. RSF models achieved consistently high concordance during periods of active decline, with strongest performance in the field exhibiting the greatest disease pressure. Spatial risk maps revealed early hotspots that expanded into contiguous high-risk zones by June, while fields with minimal visible symptoms showed diffuse but consistent risk patterns. Post-hoc comparison with operational fumigation rates (280, 336, and 392 kg Pic-Clor 60/ha) showed no consistent association with predicted canopy risk, consistent with the possibility that lower application rates may be sufficient in portions of fields with historically low disease pressure. These results demonstrate that UAV multispectral time series combined with survival modeling can track fine-scale spatiotemporal canopy decline and provide an early-warning framework to support spatially targeted disease monitoring and management in commercial strawberry systems.