基于随机森林算法的吉林省延边地区土地生态敏感性评价

Evaluation of ecological sensitivity in Yanbian area of Jilin Province based on random forset algorithm

以吉林省延边地区为研究对象,运用随机森林分类方法训练数据并建立模型,进行生态敏感性评价。在选取的植被覆盖度、坡度、水环境、人口密度和土地利用类型5个因子中,水环境因子对延边地区生态敏感性影响程度最高,植被覆盖度次之,土地利用类型中等,人口密度和坡度影响较小。通过评价,将延边地区生态敏感性划分极度敏感、高度敏感、中等敏感、轻度敏感和不敏感5个等级,占研究区总面积的比例分别为40.13%、44.96%、8.98%、4.99%和0.94%,并在此基础上针对不同土地生态敏感性区域提出差别性的用地管控措施。

Taking the Yanbian area of Jilin Province as the research object,random forest classification was introduced to train data and establish a model for ecological sensitivity evaluation.Among the five factors selected,namely vegetation coverage,slope,water environment,population density,and land use type,water environment factor exhibited the most significant influence on ecological sensitivity in the Yanbian area,followed by vegetation coverage and land use type,with population density and slope demonstrating relatively minimal impact.Subsequently,the ecological sensitivity in the Yanbian area was divided into five levels:extremely sensitive,highly sensitive,moderately sensitive,slightly sensitive,and insensitive,accounting for 40.13%,44.96%,8.98%,4.99%,and 0.94%of the total study area,respectively.Based on the evaluation,differential land use management measures were then proposed for specific ecologically sensitive areas.