Application of hyperspectral technology for leaf function monitoring and nitrogen nutrient diagnosis in soybean (Glycine max L.) production systems on the Loess Plateau of China

Farmland mulching and nitrogen application are the main agricultural management measures to improve soybean (Glycine max L.) production in dryland agriculture on the Loess Plateau. However, the interaction effects of various farmland mulching materials and nitrogen application rates on soybean leaf functionality and nitrogen nutrition status remains contradictory and inconclusive, and effective and rapid diagnostic of the leaf functionality and nitrogen status are still not well-understood. Therefore, the aim of this study is twofold: 1) to investigate the interactive effects of different mulching methods and nitrogen application levels on soybean leaf growth, physiological traits, and final grain yield throughout the entire growth period; 2) to establish a chlorophyll content monitoring model based on spectral parameters and a diagnostic model for soybean nitrogen status, and to assess the effectiveness of the latter in accurately diagnosing soybean nitrogen requirements. A two-year (2021 and 2022) field experiment, which consisted of four nitrogen levels (0, 60, 120 and 180 kg N ha−1) and four mulching strategies (NM: no mulching, SM: straw mulching, FM: film mulching, SFM: straw and film mulching), was conducted in the soybean growing seasons. Subsequently, nitrogen dilution curves were established based on chlorophyll content and the optimal spectral index to compute corresponding nitrogen nutrition index (NNI) values. The results indicated that: 1) FM, in comparison to SM and SFM, had the ability to delay leaf senescence, increase chlorophyll content, and improve physiological growth parameters, thereby increasing seed yield; 2) the application of 120 kg N ha−1 stood out, boosting soybean photosynthetic productivity and providing a solid foundation for high soybean yields; 3) the Random forest (RF) model had the highest accuracy in chlorophyll content estimation (R2 =0.854, RMSE=2.765 and MRE=5.758 % in validation set); and 4) both the nitrogen dilution curves constructed based on chlorophyll content and the optimal spectral index proved to be reasonably accurate in quantitatively diagnosing the nitrogen status of soybeans. Specifically, N0 and N180 indicated sustained nitrogen deficiency and excess, respectively, while N60 and N120 effectively met soybean's nitrogen requirements. In summary, the FMN120 effectively stabilized seed yield by regulating key functional traits in soybean leaves. This study provides a theoretical basis for the application of precision agriculture and remote sensing technology in monitoring and diagnosing plant nitrogen nutrition in soybean production.