Assessing the declining trend in soil erodibility across China: A comparison of conventional and digital K-factor maps

• Published in: International Soil and Water Conservation Research Vol. 13; no. 1; pp. 15 - 26
• Main Authors: Tian, Zhiyuan, Zhao, Yan, Cao, Longxi, Zhao, Yuan, Liang, Yin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2025; KeAi Communications Co., Ltd
• Subjects: algorithms; China; Environmental variables; erodibility; prediction; Random forest; regression analysis; Soil erosion; soil organic matter; Soil survey database; soil surveys; Spatiotemporal variation; species; water conservation; water erosion; China; Spatiotemporal variation; Random forest; Soil erosion; Soil survey database; Environmental variables
• ISSN: 2095-6339
• DOI: 10.1016/j.iswcr.2024.05.005

Soil erodibility is a measure of soil susceptibility to water erosion and serves as an essential element, also known as the K-factor, in empirical soil erosion prediction models, such as USLE, RUSLE, and CSLE. The currently available map of the K-factor for China was generated based on the conventional soil polygon linkage method and soil species survey conducted in the 1980s. For update, an investigation of 4262 samples from the soil series survey in the 2010s and a random forest regression model were used to generate a new K-factor map for China. A digital K-factor map at the 250 m spatial resolution was generated by calculating the K values from soil survey points as training data and using environmental information as predictive variables. The comparison results between the digital and conventional K-factor maps show that there has been a decreasing trend in the K-factor over recent decades. The K value decrease was mainly attributed to the update of soil survey data (the mean K value changed from 0.03193 t ha·h/(MJ·mm·ha) in the soil species database to 0.02988 t ha·h/(MJ·mm·ha) in the soil series) and was less influenced by the replacement of the mapping methods (the mean K value changed from 0.03197 t ha·h/(MJ·mm·ha) in the soil polygon linkage method to 0.03193 t ha·h/(MJ·mm·ha) in the random forest). This study quantified the sources of change between previous and updated national K-factor maps and demonstrated that there is a decreasing trend in K values, which is consistent with the increasing soil organic matter and improved ecological environment in China. [Display omitted]