Abstract: Changes in lake area and water volume exert significant impacts on the ecological environment of arid regions. Targeting East Dabuxun Lake，Golmud River Basin，Qinghai Province，this study developed a multi-index random forest algorithm based on the Google Earth Engine （GEE） cloud platform and Landsat imagery to extract the lake area from 1987 to 2021. Then，an area-water level relationship was established using laser altimetry data from ICESat and CryoSat satellites to estimate changes in water volume. Finally，the impacts of natural factors and human activities on the lake were evaluated，using ERA5-Land climate data and records of potash mining，along with correlation analysis and random forest-based contribution assessment. The results indicate that the temporal changes in lake area over time can be divided into five stages：expansion，shrinkage，recovery，re-shrinkage，and rapid recovery. Spatially，the lake exhibited a pattern of shrinkage in the south and expansion towards the northwest. From 2003 to 2021，the water volume of East Dabuxun Lake showed an upward trend. Temperature，glacier and permafrost melting，and solar radiation were identified as the main natural factors influencing lake area，with contribution rates of 31.0%，29.4%，and 15.5%，respectively. In terms of human activities，potash mining emerged as a major driver of lake area changes after 2010. Based on predictions by the auto-regressive moving average model （ARIMA），the lake area is projected to decline to 302.78 km² by 2030.