基于哨兵光学的兴安盟土攘有机碳密度遥感反演研究

打开文本图片集

关键词：土壤；有机碳密度；哨兵光学；遥感；反演；兴安盟

中图分类号：TP237 文献标识码：A

文章编号：0439-8114（2025）11-0175-07

DOI：10.14088/j.cnki.issn0439-8114.2025.11.024 开放科学（资源服务）标识码（OSID）：

A remote sensing retrieval study of soil organic carbon density in Xing'an League based on Sentinel optical data

WANG Xin-xin ！1 ，YUJing2，34，ZHU Hua-chen4， ZHAO Zhen-ni4， CHEN Xiao-long4 （1.hoolaCioJUsig;aen University，Chogqing4o44，hina;3.ChongqingSchol，UniversityofCineseAcadeyofienes，Chongqing0oa; 4.Chongqing Geomatics and Remote Sensing Application Center，Chongqing 401120， China）

Abstract：AspatialestimationofsoilorganiccarbondensityinXing’anLeagueoftheInnerMongoliaAutonomousRegionwasconductedfortheyear2O23usingremotesensingandgeographicinformationtechnologies.Multipleenvironmentalvariableswereextractedbasedonremotesensingdataandevironmentaldata（e.g.，terrain，climate，soil），withsoilpropertydataobtainedfromfieldsoil samplingserving as theresponse variable.Regressionmodeling andaccuracycomparisonwereperformedusing RandomForest（RF）， Extreme Gradient Bosting（XGBoost），and Multilayer Perceptron（MLP）.Basedontheacuracy evaluationresults，thebest-performingodelwasselectedtoultimatelycompletethespatialmappingofsoilorganiccarbondensityinthestudyarea.Theresults showed that the RF model performed the best （ Rz=0.86 ， RMSE=1.51kg/m2 ），the XGBoost model performed slightly worse than the RFmodelbutstillwell，whiletheMLPmodelperformedsignificantlyworseinthistask，withitsaccuracyeing muchlowerthanthe othertwo models.Thespatialdistributionofsoilorganiccarbondensityin Xing'anLeagueshowedobvious north-southdiferences， generallydecreasingfromnorthwestosoutheast.Intemsofverticaldistribution，theorganiccarbondensityfirstdeereasedanden increased with increasing sampling depth.

Key Words： soil;organic carbon density；Sentinel optical； remote sensing; inversion；Xing'an League

土壤有机碳密度作为全球碳循环和碳储量的重要组成部分，在应对气候变化和实现碳中和目标中具有关键作用[1]。（剩余9451字）