Quantifying the subfield heterogeneity in winter cereal cover crop biomass across farms in the Mid-Atlantic Coastal Plain

Winter cover crops can reduce erosion and improve water quality. Ecosystem services generally increase with higher cover crop biomass, but they are known to be variable within the field. We generated a large remotely-sensed biomass dataset from more than 8000 fields enrolled in a cover crop cost share program to study biomass heterogeneity from fields in the Choptank and Chester-Sassafras watersheds in the Delmarva Peninsula, across four years. We also used soil and topography data to determine drivers of subfield heterogeneity in cover crop biomass. We found that winter cover crop biomass heterogeneity is higher in fields with large cover crops, and that low-lying or poorly-drained positions on the landscape often had lower biomass in the most variable fields. Random forest model performance at predicting subfield cover crop biomass was slightly improved at 9 m compared to 3 m spatial resolution. We demonstrate that remotely-sensed normalized difference vegetation index, soil, and topography data can be used to quantify variation in cover crop biomass at a watershed scale.