Near-real-time detection of manure applications on snow-covered croplands in Google Earth Engine using Sentinel-2 and Landsat-8/9

Spreading livestock waste from animal feeding operations (AFOs)—including the larger concentrated animal feeding operations (CAFOs)—on frozen or snow-covered ground poses substantial water-quality risks, yet no freely available tool exists to enable governments, non-profits, or citizens to monitor such spreading events. We present an open, reproducible workflow in Google Earth Engine that maps waste applications on snow using freely available Sentinel-2 and Landsat-8/9 imagery across three Wisconsin regions during the months of January and February 2025. We assemble field-verified and photo-interpreted training data, along with masks for cloud cover, snow cover, and cropland, to detect livestock waste applications and validate performance with a design-based pixel-day assessment. Within the observable domain, the classifier achieved high precision (0.990, 95% CI 0.977–1.000) and recall (0.942, 95% CI 0.934–0.949) for the livestock waste class. Outputs include vectorized spread polygons and optional image chips that support complaint triage and site visits. The workflow emphasizes tolerance to adverse conditions and scalability using free data, and it can be adapted to other snow-prone regions with modest retraining.