Coupling Crop Simulation Modeling and Crop Sensing to Improve within-Field Nitrogen Management.

Increasing nitrogen (N) use efficiency could be possible if in-season N fertilization management is optimized. The goal of this study was to test and develop a methodology for combining Normalized Difference Vegetation Index (NDVI) data and simulation modeling to assess spatial variability of corn N stress and in-season N rate. The yield, soil, and NDVI data used for this was collected from a precision agriculture project conducted on five corn fields located in Italy. The spatial model calibration and simulation were conducted in CERES-Maize model in DSSAT using the interface with the Geospatial Simulation (GeoSim) tool in Quantum GIS. The interface of GeoSim with DSSAT allowed the spatial calibration of soil parameters (initial soil water, drain upper limit, lower limit) for better simulation of spatial yield variability. The spatial optimization of these soil properties allowed reducing the yield RMSE to 92% with respect to a not calibrated model. The integration of the NDVI values for improving leaf area index simulation around 63 days after planting was possible through the use of the modified Beer’s law. The optimization of the PHINT coefficient was necessary to improve LAI spatial simulation. The correlations between NDVI and simulated LAI increased from 0.07 to 0.90 after the calibration. After calibration, the model was able to assess the spatial variability on crop N stress around silking and to assess the variable N rate necessary to minimize the N stress. The potential for conducting spatial simulation modeling provides enormous opportunities for site specific nitrogen management.