Can normalized difference vegetation index and climate data be used to estimate soil carbon, nitrogen, and phosphorus and their ratios in the Xizang grasslands?

作者：Shaohua Wang, Huxiao Qi, Tianyu Li, Yong Qin, Gang Fu, Pan Xu, Xinjie Zha · 发表于：Frontiers in Earth Science · 年份：2024 · DOI：10.3389/feart.2023.1340020 · 被引用次数：7 · 研究领域：Soil Carbon and Nitrogen Dynamics、Soil Geostatistics and Mapping、Remote Sensing in Agriculture

Accurately quantifying the relative effects of climate change and human activities on soil carbon, nitrogen, and phosphorus in alpine grasslands and their feedback is an important aspect of global change, and high-precision models are the key to solving this scientific problem with high quality. Therefore, nine models, the random forest model (RFM), generalized boosted regression model (GBRM), multiple linear regression model (MLRM), support vector machine model (SVMM), recursive regression tree model (RRTM), artificial neural network model (ANNM), generalized linear regression model (GLMR), conditional inference tree model (CITM), and eXtreme gradient boosting model (eXGBM), were used for modeling soil organic carbon (SOC), total nitrogen (TN), total phosphorus (TP), the ratio of SOC to TN (C:N), the ratio of SOC to TP (C:P), and the ratio of TN to TP (N:P) at depths of 0–10, 10–20, and 20–30 cm under non-grazing and free-grazing scenarios in the Xizang grasslands. Annual radiation (ARad), annual precipitation (AP), and annual temperature (AT) were used as independent variables under non-grazing scenarios, whereas ARad, AP, AT, and growing season maximum normalized difference vegetation index (NDVI max ) were used as independent variables under free-grazing scenarios. Overall, the RFM and GBRM were more accurate than the other seven models. However, the tree numbers of the GBRM were much larger than those of the RFM, indicating that the GBRM may have a greater model complexi...