Rapid mapping of winter wheat yield, protein, and nitrogen uptake using remote and proximal sensing

Abstract

Farmers’ interest in precision nitrogen (N) fertilization has grown in the inland Pacific Northwest due to significant within-field spatial variability in yield and protein content. Mapping the spatial variability of yield, protein content, and total grain N uptake (Ng) can be a useful tool for post-harvest evaluation of N sufficiency that, in turn, can guide precision N fertilization. We evaluated the applicability of combining RapidEye satellite imagery-derived vegetation indices with topographic variables derived from high-resolution data obtained from proximal sensors to estimate winter wheat yield, protein content, and Ng. Results indicate that the normalized difference vegetation index (NDVI), the normalized difference red edge index (NDRE), terrain curvature, slope, aspect, topographic wetness index, and solar radiation are the most relevant co-variables that contribute to spatial variability of yield and Ng. Yield and Ng exhibited strong relationships with NDVI and NDRE from late June through early July during grain filling stage, with nearly all R2 > 0.6. We found greater yield and Ng values in concave-shaped terrain; 0-5° and 10-20° slopes; and flat, eastern, or northern aspects. In contrast, protein content exhibited weak relationships with vegetation indices and nearly all terrain factors. Prediction models demonstrated that these variables can provide good estimations of yield and Ng, with R2 of 0.848 and 0.864, respectively, and RMSE of 48.47 and 0.0136, respectively. Combining RapidEye-based NDRE with proximal sensor-based topographical factors shows great potential for accurately and efficiently estimating winter wheat yield and Ng, which can help guide N fertilizer management.