【摘要】 生态空间是科学制定国土空间规划的重要内容，对区域高质量发展至关重要。以三峡库区重庆段为例，基于随机森林模型和POI数据识别生态空间，并利用生态空间转移矩阵等方法研究其时空演变特征。研究表明：随机森林模型等机器学习方法与多源数据的结合能提升生态空间识别的准确性与科学性。2012—2022年，三峡库区重庆段生态空间整体沿西部向中部的方向变迁，生态空间的分布范围和集聚区域扩大，并从区县逐渐向中心城区延伸，分布更加均衡。研究区域内生态空间总体呈扩张趋势，各地区间生态空间面积变化差异明显，主要表现为林地、水域两类生态用地面积增加，而草地生态用地和非生态用地面积缩减。更多还原

【Abstract】 Ecological space is a crucial component in the scientific formulation of territorial spatial planning and plays a vital role in regional high-quality development. Taking the Three Gorges Reservoir Area(Chongqing Section) as a case study, this research employs a random forest model combined with POI data to identify ecological space within the region. Furthermore, it investigates the spatiotemporal evolution characteristics of this ecological space through ecological space transfer matrix methods. The study reveals that the integration of machine learning approaches with multi-source data significantly enhances the accuracy and scientific rigor of ecological space identification. Over the past decades, the ecological space has generally migrated along the west-central axis. The distribution range and agglomeration areas of ecological space have expanded noticeably, and extended from counties towards the central urban areas, resulting in a more balanced distribution. The research indicates an overall expansion trend of ecological space within the region, with notable disparities in the magnitude of change among different areas. This transformation primarily manifests through two contrasting land-use dynamics: significant increase in forest and water ecological areas, coupled with substantial decrease in grassland ecological areas and non-ecological land uses.更多还原