基于Sentinel-1/2A的普洱市思茅区城市绿地覆盖分类及碳储量估测  

Construction of Carbon Storage Retrieval Model of Urban Green Space in Simao District of Puer City Based on Sentinel-1/2A

作　　者：胥晓 沈守云 张加龙 XU Xiao;SHEN Shouyun;ZHANG Jialong(College of Landscape Architecture,Central South University of Forestry and Technology,Changsha 410004,Hunan,China;School of Logistics and Management Engineering,Yunnan University of Finance and Economics,Kunming 650221,Yunnan,China;Hunan Big Data Engineering Technology Research Center of Natural Protected Landscape Resources Areas,Changsha 410004,Hunan,China;Faculty of Forestry,Southwest Forestry University,Kunming 650224,Yunnan,China)

机构地区：[1]中南林业科技大学风景园林学院,湖南长沙410004 [2]云南财经大学物流与管理工程学院,云南昆明650221 [3]湖南省自然保护地风景资源大数据工程技术研究中心,湖南长沙410004 [4]西南林业大学林学院,云南昆明650224

出　　处：《西北林学院学报》2025年第2期141-151,共11页Journal of Northwest Forestry University

基　　金：云南省教育厅科学研究基金项目(2023J0652);国家林业局重点学科项目(林人发[2016]21号);湖南省“双一流”培育学科项目(湘教通[2018]469号)。

摘　　要：城市绿地碳储量遥感反演已成为当前城市可持续发展与生态保护研究的热点。以云南省普洱市思茅区城市绿地为研究对象,利用2018年Sentinel的遥感数据源、《城市绿地分类标准》与森林资源二类调查数据对城市绿地进行分类,结合Sentinel和DEM数据提取并筛选出纹理因子(9×9窗口下的均值信息meanR9)、地形因子(坡度、坡向)、地表水分指数(LSWI)、归一化植被指数(NDVI)、B11、红边位置指数(S2REP)、归一化差异水体指数(MDWI)、归一化红边指数(NDRE1)等城市绿地碳储量强相关因子,采用RF、KNN、GBRT模型估测普洱市思茅区城市碳储量。结果表明:1)思茅区城市绿地系统分为公园绿地、防护绿地、附属绿地、生产绿地4类,其面积及占比分别为16.09、19.4、0.425、2.64 km^(2)和41.72%、50.33%、1.10%、6.85%。其遥感分类的总体精度(OA)为89.02%,Kappa系数为0.8471,F_(1)为89.44%;2)RF模型估测绿地碳储量的总体效果(R2、RMSE、P和rRMSE分别为0.762~0.966、5.766~10.684 t/hm^(2)、0.548~0.965、14.765~34.102),优于KNN和GBRT模型;3)选取最优RF模型反演得到的各类绿地碳储量,公园绿地、防护绿地、附属绿地、生产绿地的碳储量密度及占比分别为10.571、9.698、2.395、8.866 t/hm^(2)和45.69%、30.52%、2.53%、21.25%,总体误差在0.53%~1.85%。普洱市思茅区城市绿地有着较高的碳储量,具备较大的碳汇潜力,研究结果可为城市碳吸收提供参考。Remote sensing inversion of urban green space carbon storage has become a hot topic in current research on urban sustainable development and ecological protection.Taking the urban green space in Simao District,Puer City as the research object,the urban green space was classified by using the Sentinels remote sensing data in 2018 the relative national industrial standard“Urban Green Space Classification Standard”and forest resource inventory data for management.Several indices that are significantly relative to the carbon storage of urban green space were extracted and filtered by combining Sentinel and DEM data,including texture index(mean information meanR9 under 9×9 window),topographic index,surface moisture index(LSWI),normalized vegetation index(NDVI),B11,red edge position index(S2REP),normalization differential water body index(MDWI),normalized red edge index(NDRE1).Models of RF,KNN and GBRT were used to estimate the carbon storage of urban green space of Simao District.The results showed that 1)the urban green space system in Simao District was divided into 4 categories:park green space,protective green space,auxiliary green space,and production green space with the areas of 16.09,19.4,0.425,and 2.64 km^(2),accounting for 41.72%,50.33%,1.10%,and 6.85%of the total,respectively.The overall accuracy(OA)of remote sensing classification,Kappa coefficient,and the value of F_(1) were 89.02%,0.8471,and 89.44%,respectively.2)The overall effect of the RF model in estimating green space carbon storage was better than the KNN and GBRT models,the values of R2,RMSE,P,and rRMSE were 0.762-0.966,5.766-10.684 t/hm^(2),0.548-0.965,14.765-34.102,respectively.3)The optimal RF model inversion was selected to estimate carbon storage of various green spaces,the results of the estimation indicated that the carbon storage density of park green space,protective green space,affiliated green space,and production green space were 10.571,9.698,2.395,and 8.866 t/hm^(2),respectively,accounting for 45.69%,30.52%,2.53%and 21.25%of the

关 键 词：城市绿地 碳储量 SENTINEL RF模型 

分 类 号：S718.55[农业科学—林学]