Data-driven assessment of soil total nitrogen on the Qinghai-Tibet Plateau
Jiahui Zhao,Peng Jiang,Tongqing Shen,Rongrong Zhang,Dawei Zhang,Nana Zhang,Nie Ting,Kunqi Ding,Bin Yang,Changhai Tan,Zhongbo Yu
DOI: https://doi.org/10.1016/j.scitotenv.2024.169993
IF: 9.8
2024-01-15
The Science of The Total Environment
Abstract:The investigation of soil total nitrogen (STN) holds significant importance in the preservation and sustainability of Earth's ecosystems. The Qinghai-Tibet Plateau (QTP), renowned as the world's most expansive plateau and characterized by its exceptionally delicate ecosystem, demands an in-depth exploration of its STN content. In this study, we use a machine learning approach to extrapolate point-scale measured STN stocks to the entire QTP and calculated STN storage from 0 to 2 m. Our results show that the XGB algorithm performs well in modeling STN despite variations in simulation accuracy for specific depth ranges. The spatial distribution of STN across the QTP exhibits pronounced heterogeneity, especially for the 0-50 cm soil layer, with relatively higher STN stocks in the southeast and lower stocks in the northwest of QTP. The vertical distribution reveals a gradual decrease in STN storage with increasing depth. The 0–50 cm soil layer holds the highest STN stocks, averaging around 0.78 kg/m 2 , which is almost the sum of STN stocks in the 50–100 cm and 100–200 cm soil layers. Meanwhile, the STN stocks are smaller in permafrost zone than that in non-permafrost zone. We also investigate the impact factors that control the spatiotemporal distribution of STN. It indicates that vegetation, precipitation, temperature, and elevation are the major factors for STN distribution, while physical properties of the soil have a relatively smaller impact. These findings are crucial for understanding the distribution and evolution of STN on the QTP.
environmental sciences