A spectral parameter for the estimation of soil total nitrogen and nitrate nitrogen of winter wheat growth period

Abstract

AbstractReal‐time monitoring of crop nitrogen and soil nutrient status can provide an important basis for the rational application of nitrogen during fertilization. Two field experiments were conducted in Henan Province, China, using three main wheat cultivars and four nitrogen levels, across two consecutive growing seasons. Canopy spectral reflectance, plant nitrogen and soil nitrogen were synchronously measured during the wheat's main growth stages. Furthermore, the quantitative relationship between the soil nitrogen and plant nitrogen concentrations was analysed, and the relationship with existing spectral parameters and several kinds of hyperspectral indices, including normalized difference spectral indices (NDVI), ratio spectral indices (RVI) and difference spectral indices (DVI), was extracted. All the combinations of two wavebands between 350 and 1 050 nm were calculated, and their quantitative relationships with soil nitrogen content were analysed. The results were used to build a remote monitoring model of soil total nitrogen and nitrate nitrogen (NO3‐‐N) content by using the canopy spectral reflectance. The results showed that the correlation between soil NO3‐‐N and plant nitrogen concentration was superior to the correlation between total nitrogen and plant nitrogen concentrations. After systematically analysing the quantitative relationship between wheat canopy spectra, soil total nitrogen content and NO3‐‐N content, we found that the correlation with the original reflectance in the visible region was superior and that the first derivative spectrum significantly improved the correlation in the near‐infrared region. Furthermore, the soil total nitrogen content, NO3‐‐N content and canopy spectral data were analysed in different growth stages (jointing stage, flowering stage and filling stage). The results showed that NDVI (FD747, FD699) was the best indicator for monitoring soil total nitrogen and NO3‐‐N content modelling, as the modelling decision coefficient and prediction decision coefficient were both above 0.8 for all test cases. Thus, during wheat growth, crop canopy spectroscopy can be used to estimate soil nitrogen nutrition status, and NDVI (FD747, FD699) can be used as an effective spectral parameter for estimating soil total nitrogen and NO3‐‐N content.