Evaluation Of Peatland Cover In Hulunbuir Region,Inner Mongolia,China Based On Sentinel-2 Remote Sensing Data

Peat swamp is an important carbon reservoir of terrestrial ecosystem and one of the important wetland types,and its unique ecological functions play an important role in maintaining ecological balance and sustainable development.Currently,with the increasing scope of human activities,the degradation of peat bog wetlands is becoming more and more serious.To better protect the existing peat bog resources,reduce the destruction of peat bogs and develop scientific management measures,it is a prerequisite and basis to determine the spatial distribution of peat bogs.Peat bogs are widely developed in Inner Mongolia due to its suitable climatic and hydrological conditions.Therefore,this study takes the Hulunbuir region of Inner Mongolia as the study area,and applies traditional classification methods to extract peat bog information in the study area based on Google Earth Engine(GEE)platform,combined with Landsat 8,Sentinel-2 and ground survey data,in order to quickly and accurately extract the spatial location distribution of peat bogs.The main findings of this paper are as follows.(1)The study used Landsat 8 data from April 15,2020 to October 15,2020,a single-window algorithm to obtain surface temperature data of the study area,and combined with vegetation index data,two classification methods,random forest(Random Forest,RF)and decision tree classification(Classification and Regression Trees,CART),were used to produce two peatland maps of the Hulunbuir region.Among them,the RF-based classification estimated the peatland area of Hulunbuir to be 44737 km2,while the CART result was 49565 km2 of peatland,and both classification results showed highly accurate results with an overall accuracy rate of over 80%.(2)By comparing the two classification methods of RF and CART for peat bog information extraction in the GEE platform,the study found that the accuracy of CART was slightly lower than that of RF,and the Sentinel-2 data used were easily available in the GEE platform,which also provided important guidance for subsequent studies on peatland information extraction at large regional scales.