基于Landsat 8和Sentinel-1的广东青云山自然保护区森林生物量反演OA北大核心CSTPCD

Forest is an important part of terrestrial ecosystem and plays an important role in global carbon cycle.Forest biomass research contributes to the deeper understanding of forest ecosystems.The inversion models of forest above-ground biomass for the forests occurring in the Qingyunshan Nature Reserve(QNR)in Wengyuan,Guangdong Province were constructed by employing methods of random forest(RF),support vector machine(SVM)based on such remote sensing data as Landsat 8 and Sentienl-1,and permanent sample plot data.The results showed that 1)the prediction accuracy of RF method was higher than that of SVM in the case of using either Landsat 8 or Sentienl-1,or both.2)From the angle of compre-hensive accuracy of remote sensing estimation,the order of data sources was cooperative remote sensing data＞optical remote sensing data＞synthetic aperture radar data.3)The average biomass of QNR was 132 t/hm2,ranging from 93 t/hm2 to 171 t/hm2,and the total biomass was 892 466 t.The results indicates that collaborative Landsat 8 and Sentienl-1 data sources and RF method plays important roles in the procedure of forest biomass estimation.