Spatio-temporal Dynamics and Drivers of Soil Loss in the Poyang Lake Basin Based on GEE and Multi-Source Remote Sensing Data
Keywords: Soil loss, RUSLE Model, Spatio-temporal Analysis, Google Earth Engine platform, Multi-Source Remote Sensing Data
Abstract. Soil loss presents significant threats to environmental safety and agricultural security in the Poyang Lake Basin. Therefore, accurately assessing the spatio-temporal distribution of soil loss and identifying its key driving factors is essential. Utilizing the Google Earth Engine (GEE) platform and multi-source remote sensing data, this study estimates annual average soil loss and examines its spatio-temporal dynamics through the Revised Universal Soil loss Equation (RUSLE). The analysis incorporates rainfall erosivity (R factor) from CHIRPS precipitation data, soil erodibility (K factor) from OpenLandMap, topographic factors (LS factor) from DEM data, vegetation cover (C factor) from Landsat7/8 NDVI, and conservation practices (P factor) from MODIS land cover data. The results indicate significant changes in soil loss patterns from 2001 to 2020, with average annual soil loss decreasing from 13.4 t/hm2 in 2001–2010 to 7.2 t/hm2 in 2011–2020, reflecting a trend line coefficient of −0.5. Areas of severe erosion were identified at the confluence of the Yangtze River and Poyang Lake, particularly in Zhaisang, Lianxi, and Lushan Districts, as well as in flood-prone regions southeast of Poyang Lake and urban areas in Nanchang City. Within the same year, soil loss distribution correlates with precipitation and slope. While over 20 years, strong relationships were found between soil loss and cropland (with a Pearson correlation coefficient of 0.58) and impervious surfaces (−0.72), indicating that human activities primarily drive soil loss in the Poyang Lake Basin. The research findings align with previous studies that utilized the RUSLE model to calculate soil erosion in Poyang Lake, based on historical and geospatial data from various domestic sources. This demonstrates the effectiveness of GEE and remote sensing in assessing soil loss, providing reliable data to support the sustainable use and protection of land in the watershed.