基于GEE和随机森林分类算法研究广西壮族自治区在1995-2022年27 a间的土地覆被变化，采用地理探测器对不同土地覆被类型的时空变化及其主要驱动因素进行分析研究。结果表明:1)土地覆被类型分类结果具有较好的分类精度，整体精度在87%～92%。2)1995-2022年，广西壮族自治区的林地和建设用地面积持续增加，增幅分别达75。94%和132。67%。林地面积从85 382。72 km2 增加到150 224。55 km2，建设用地面积从2 038。73 km2 增加到4 743。46 km2;与此同时，草地及稀疏灌草混交地面积减少了82。46%(66 990。97 km2);其他土地覆被类型变化则相对稳定。3)通过地理探测器模型的分析结果显示，人为因素是影响土地利用/覆被(LULC)变化的主要驱动因素，经济的快速发展、城市人口的增加和工业扩张是建设用地和林地扩张以及植被减少的主要原因。研究结果对于广西壮族自治区的生态环境管理和土地可持续利用规划具有参考价值。
Research on Land Cover Change and Driving Factors in Guangxi Zhuang Autonomous Region Based on Google Earth Engine and Landsat Images
Based on GEE and random forest classification algorithms,the land cover changes in Guangxi Zhuang Autonomous Region for 27 years from 1995 to 2022 were studied.Geodetectors were used to analyze and study the spatiotemporal changes of different land cover types and their main driving factors.The research results showed that 1)the land cover type classification results had good classification accura-cy,with the overall accuracy ranging from 87%to 92%.2)Between 1995 and 2022,the area of forest land and construction land in the region continued to increase,with an increase rates 75.94%and 132.67%,respectively.The forest land area increased from 85 382.72 to 150 224.55 km2,and the construction land area increased from 2 038.73 to 4 743.46 km2;at the same time,the area of grassland and sparse shrub-grass mixed land decreased by 82.46%(66 990.97 km2);other land cover types changed relatively sta-blize.3)The analysis results of the geodetector model showed that human factors were the main driving factors affecting land use/cover(LULC)changes.Rapid economic development,increase in urban popula-tion and industrial expansion were the main reasons for the expansion of construction land,forest land and the reduction of vegetation.The research results have reference value for ecological environment manage-ment and sustainable land use planning in Guangxi Zhuang Autonomous Region.
land useGEELandsatdriving factorgeodetectorrandom forestGuangxi Zhuang Autono-mous Region
万方数据
维普
