Can nitrogen recommendations for corn production be improved through spatially explicit crediting of cover crops and soil organic matter?

•We tested a new model for n recommendations in corn grain production.•The SCAN model explicitly considers n from organic matter and cover crop residues.•Cover crop NDVI, soil EC, and previous yield dictated discrete management zones.•Variable rate n application was used to meet requirements for each zone.•The new model had similar performance to traditional yield-based models. Managing N supply to corn (Zea mays L.) properly is a major challenge, with consequences of mismanagement ranging from N-deficient, low yielding crops to pollution of surface and groundwater. For much of the 20th and 21st centuries, N fertilizer recommendations for corn have been calculated by multiplying the yield goal by a constant equal to the amount of N needed per mass of corn grain. While this approach is often sufficient to support the growing crop without tremendous environmental consequences, it does not explicitly consider N mineralized from soil organic matter (OM) or the dynamics of N mineralization or immobilization caused by previous cover crop residues. While typical N management practices used by farmers have led to relatively low N use efficiency, precision agriculture technologies developed in the late 20th century held the promise of improving N management by considering spatial variability in crop fields and adjusting fertilizer rates accordingly. To date, however, few farms use precision agriculture technologies for managing N. In this experiment, we tested a new N recommendation system based on a biogeochemical model, the soil and cover crop available nitrogen (SCAN) model, which explicitly considers OM and cover crop residue pools of N. We coupled this new N recommendation system with variable rate precision agriculture technology to generate spatially explicit N recommendations in corn fields. We hypothesized that this approach to N management would: (i) reduce the amount of N recommended for a corn crop compared to a yield goal approach while, (ii) maintaining corn yield at a level similar to that in a traditional yield goal approach. We found that the SCAN model reduced N rates at four of the six sites compared to the yield goal approach but recommended greater N application at two of the six sites. At one of the sites where the SCAN model recommended higher N rates, we also observed increased grain yields. However, when the SCAN model recommended lower N rates than the yield goal approach, we typically observed reduced yields. Additional research is needed