摘要
内蒙古新一轮退耕还林还草工程于2015年启动,是内蒙古生态环境建设恢复的重大举措。本文利用2022年生长季Sentinel-2A影像及地形数据,结合实地测量数据,基于GEE平台采用梯度提升树回归模型(GBRT),对内蒙古新一轮退耕还林工程区林冠覆盖率进行反演分析。结果表明,模型验证决定系数(R^(2))为0.87,均方根误差(RMSE)为0.079,平均绝对误差(MAE)为0.062。内蒙古新一轮退耕还林工程区林冠覆盖率均值为0.147,其分布存在明显的空间差异,受植被配置类型、种植年份和地理环境等因素影响,总体从西向东逐渐递增。利用Sentinel-2A影像和地形数据可以有效地估测退耕还林工程区林冠覆盖率,对低林冠覆盖率区域的反演提供参考。
The new period of returning farmland to forestland and grassland project is launched in 2015,which is amajor measure to restore the ecological environment of the Inner Mongolia.Using the Sentinel-2A images and terrain data in the 2022 growing season,combined with field measurement data,and based on the GEE platform,the gradient boostingreg ression tree(GBRT)is used to invert the canopy cover of the new period of farmland to forestland project area in the Inner Mongolia.The results show that the model verification coefficient of determination(R^(2))is 0.87,the root mean square error(RMSE)is 0.079 and the mean absolute error(MAE)is 0.062.The average canopy cover in the new period of returning farmland to forestland project areas in Inner Mongolia is 0.147.There are obvious spatial differences in its distribution,which is affected by factors such as vegetation configuration type,planting year,and geographical environment.It generally increases gradually from west to east.Sentinel-2A images and terrain data can be used to effectively estimate canopy cover in the conversion of farmland to forestland project area,providing a reference for the inversion of low canopy cover areas.
作者
格根塔娜
王天璨
王建和
沈通
月亮高可
Gegentana;WANG Tiancan;WANG Jianhe;SHEN Tong;YUE Lianggaoke(Forestry and Grassland Work Station of Inner Mongolia,Hohhot 010010,China;Aerospace Information Research Institute,Beijing 100094,China;University of Chinese Academy of Sciences,Beijing 100101,China;Chongqing Jiaotong University,Chongqing 404100,China)
出处
《测绘通报》
CSCD
北大核心
2024年第6期139-145,共7页
Bulletin of Surveying and Mapping
基金
能力培育项目-MFST-蒙古高原土地退化零增长决策支持系统与示范应用。
关键词
退耕还林
林冠覆盖率
梯度提升树
内蒙古
conversion of farmland to forestland
canopy cover
gradient boosting regression tree
Inner Mongolia
