Remote estimation of above ground nitrogen uptake during vegetative growth in winter wheat using hyperspectral red-edge ratio data

Abstract

Hyperspectral sensing can provide an effective means for fast and non-destructive estimation of crop above ground nitrogen (N) status. In this study, quantitative correlations between above ground N uptake and ground-based canopy hyperspectral reflectance in winter wheat (Triticum aestivum L.) were investigated. Field experiments were conducted over four years at different sites (Xinyang, Zhengzhou and Shangshui) in Henan, China. Different N rates, planting densities, basal/topdressing N ratios and wheat cultivars were tested, and a novel spectral index was developed with improved predictive capacity for above ground N uptake estimation. Linear regression was integrated with previously reported CIred-edge3 and MTCI indices to investigate the dynamic nature of above ground N uptake, which resulted in coefficients (R2) of 0.761 and 0.760, and square error values (SE) of 4.527 and 4.534, respectively. The optimum combination of SR (759,742) and ND (759,742) were derived from two waveband-based algorithms that corresponded to red-edge ratio spectral indices. R2 for the novel SR and ND models was 0.794 and 0.788, respectively, confirming the superiority of the SR index. The modified red-edge ratio (mRER) was constructed based on the ratio vegetation index (RVI) that was derived from the third waveband (λ3) using the formula (R759−1.8×R419)/(R742−1.8×R419). This novel index was highly correlated with above ground N uptake and had the highest R2 (0.813) and lowest root mean square error (4.005) of all models tested. Fitting independent data to the equations resulted in RMSE values of 21.9, 20.4, 18.6 and 16.4% between measured and estimated above ground N uptake values for CIred-edge3, MTCI, SR(759,742) and mRER, respectively, further indicating a superior fit and better performance for mRER. These models can therefore accurately estimate above ground N uptake in winter wheat, and the novel mRER index is superior for estimating above ground N uptake on a regional scale in heterogeneous fields under variable climatic conditions.