PIOSL-5 With Nitrogen as an Independent Input Parameter and Accurate Estimation of Leaf Nitrogen

The vegetation radiation transfer modeling (RTM) describes the interaction relationship between remote sensing observation signals and ground properties. The PROSPECT considering the internal optical structure of leaves (PIOSL) model assumes the leaf as two layers with different optical properties, achieving accurate simulation of leaf reflectivity. However, the PIOSL model does not consider nitrogen, an important physicochemical parameter. This article introduces a new input parameter, nitrogen concentration (Cn), and calibrates the specific absorption coefficient of nitrogen ( ${K}_{\mathrm { Cn}}$ ) and dry matter ( ${K}_{\text {Cm}}$ ) for nitrogen removal. The PIOSL model considering leaf nitrogen distribution (the PIOSL-5 model) is constructed. The PIOSL-5 model was used to simulate the reflectance spectra of the leaf optical properties experiment (LOPEX) and RICE dataset, compared with the PIOSL and the PROSPECT-PRO models. The results showed that the root mean square error (RMSE) mean of the PIOSL-5 model in 400–2500 nm for LOPEX and RICE dataset were 1.35 and 3.42, respectively, which were 0.72 and 3.49 lower than the RMSE of the PROSPECT-PRO model. The coefficients of determination ( ${R}^{2}$ ) of the PIOSL-5 model inverts nitrogen on the RICE, LOPEX, and Fresh Leaf Spectra of the California ecosystem (CALIFORNIA) datasets are 0.789, 0.733, and 0.713, respectively, which are 0.193, 0.019, and 0.016 higher than data-driven methods. The proposal of the PIOSL-5 model provides a new approach for the construction of vegetation leaf RTMs, which is of great significance for crop nitrogen nutrition management and improving nitrogen utilization efficiency.