集成多源遥感数据与机器学习的三峡库区生态环境质量评估

Integrating Multi‑Source Remote Sensing Data and Machine Learning for Assessment of Eco‑Environmental Quality in the Three Gorges Reservoir Area

摘要: 三峡库区是长江上游重要生态屏障，库区生态环境质量变化对区域可持续发展具有重要影响。利用2000~2020年多源遥感数据、Google Earth Engine平台与机器学习模型，对库区生态环境质量进行综合评估。构建的随机森林模型预测精度良好(R² = 0.89)；库区生态环境质量存在先降后升趋势，空间差异明显；生物丰富度和土壤湿度改善明显，降水量与工业产值是驱动生态环境质量变化的核心驱动因子；尽管当下生态保护区占比高(64.50%)，但部分区域未来存在生态环境质量退化风险。建议关注生态质量持续退化的库区西南部和长江沿岸城市区域，权衡库区生态保护修复与可持续发展关系。

Abstract: The Three Gorges Reservoir Area is an important ecological barrier in the upper reaches of the Yangtze River, and changes in its eco-environmental quality have a significant impact on regional sustainable development. This study used multi-source remote sensing data from 2000 to 2020, the Google Earth Engine platform, and machine learning models to comprehensively assess the eco-environmental quality of the reservoir area. The constructed random forest model achieved good prediction accuracy (R² = 0.89). The eco-environmental quality of the reservoir area showed a trend of first decreasing and then increasing, with marked spatial differences. Biological richness and soil moisture improved significantly. Precipitation and industrial output were the core drivers of changes in eco-environmental quality. Although the proportion of ecological protection areas is currently high (64.50%), some areas face the risk of future degradation in eco-environmental quality. It is recommended to focus on the southwestern part of the reservoir area and urban areas along the Yangtze River, where ecological quality continues to degrade, and to balance the relationship between ecological protection and restoration and sustainable development in the reservoir area.