Measuring leaf nitrogen concentration in winter wheat using double-peak spectral reflection remote sensing data

Abstract

Remote sensing is proving to be a rapid non-destructive method for crop nitrogen (N) status assessment. In this study, quantitative relationships between leaf N concentration (LNC) and ground-based canopy hyperspectral reflectance in winter wheat (Triticum aestivum L.) were investigated. Winter wheat field experiments were conducted over three years at different sites (Zhengzhou, Jiaozuo and Kaifeng) in Henan, China. Different N rates and wheat cultivars were tested, and a novel double-peak area index was developed to improve the prediction accuracy and stability of LNC measurement. The common optimal red-edge spectral indices were used to monitor the LNC models. Analysis of the relationship between existing vegetable indices and LNC indicated that red-edge spectral parameters were the most sensitive in this case. Integrated linear regression of LNC with mND705 and REPle was performed to describe the dynamic nature of the LNC patterns, giving coefficients (R2) of 0.83 and 0.82, and the standard errors (SE) of 0.414 and 0.424, respectively. These novel double-peak area parameters were constructed based on analysis of the red-edge characteristics, and the optimal normalized difference of the double-peak areas based on REPig division (NDDAig), in the form of (R755+R680−2×RREPig)/(R755−R680), were calculated and found to be highly correlated with LNC (highest R2=0.85; lowest SE=0.385). When independent data were fit into the derived equations, the average relative error (RE) values were 14.1%, 13.7% and 11.5% between measured and estimated LNC using mND705, REPle and NDDAig, respectively, indicating a superior fit and better performance for NDDAig. These results suggest that the models can accurately estimate LNC in wheat, and the novel double-peak area index is more effective for modeling LNC than previously reported red-edge indices.