Jiangxi Province, as one of the first national ecological civilization pilot zones in China, holds a significant responsibility in ecological protection and sustainable development. Ecological vulnerability assessment is of great guiding value for ecological protection and restoration in Jiangxi Province. Based on the sub-watershed and raster evaluation units, combined with remote sensing image data, land use data, soil data, meteorological data, socio-economic data, etc., an ecological vulnerability assessment framework was established using the ecological sensitivity-resilience-pressure (SRP) model to evaluate the spatiotemporal dynamics of ecological vulnerability in Jiangxi Province from 2000 to 2020, and the driving factors of ecological vulnerability changes were revealed using an interpretable machine learning model (XGBoost-SHAP). The results indicate that: ① The areas with relatively low ecological vulnerability of Jiangxi Province were primarily distributed in the northeastern, northwestern, and southern mountainous regions, while areas with higher vulnerability were concentrated in the plains and riverbanks where human activities are intensive, such as the Poyang Lake plain area. The overall distribution was primarily characterized by mild and light vulnerability. ② In the years 2000, 2010, and 2020, the average ecological vulnerability index values were 0.224, 0.219, and 0.206, respectively, indicating a downward trend in the ecological vulnerability index. Among these, the areas where the ecological vulnerability index decreased accounted for 75.75% of the total area. ③ The changes in soil erosion intensity, FVC, percentage of soil erosion above moderate, and land use were key factors driving ecological vulnerability changes, with relative importance weights of 34.66%, 25.99%, 10.83%, and 10.63%, respectively. Moreover, the contributions of these factors exhibited significant spatial variation. The research findings can provide theoretical support for ecological environment protection in Jiangxi Province, while also offering important references and insights for the application of machine learning methods in the study of ecological vulnerability.
扫码关注我们
求助内容:
应助结果提醒方式:
