A New Method to Determine Central Wavelength and Optimal Bandwidth for Predicting Plant Nitrogen Uptake in Winter Wheat

Abstract

Plant nitrogen (N) uptake is a good indicator of crop N status. In this study, a new method was designed to determine the central wavelength, optimal bandwidth and vegetation indices for predicting plant N uptake (g N m−2) in winter wheat (Triticum aestivum L.). The data were collected from the ground-based hyperspectral reflectance measurements in eight field experiments on winter wheat of different years, eco-sites, varieties, N rates, sowing dates, and densities. The plant N uptake index (PNUI) based on NDVI of 807 nm combined with 736 nm was selected as the optimal vegetation index, and a linear model was developed with R2 of 0.870 and RMSE of 1.546 g N m−2 for calibration, and R2 of 0.834, RMSE of 1.316 g N m−2, slope of 0.934, and intercept of 0.001 for validation. Then, the effect of the bandwidth of central wavelengths on model performance was determined based on the interaction between central wavelength and bandwidth expansion. The results indicated that the optimal bandwidth varies with the changes of the central wavelength and with the interaction between the two bands in one vegetation index. These findings are important for prediction and diagnosis of plant N uptake more precise and accurate in crop management.