Correlations of the biomass of the main tropical forest vegetation types and LANDSAT TM data in Xishuangbanna of P. R. of China

Abstract

The tropical forest biomass estimation is very important for global research. The correlation analysis of the tropical forest biomass and remote sensing data is the base for the tropical forest biomass estimated by remote sensing data. Therefore, the correlations of the biomass of the main tropical forest vegetation types in Xishuangbanna of P. R. of China and their Landsat TM data were analyzed here. It includes several steps. Firstly, the biomass for each forest sample plot with the size of 28.28 times28.28 meters is calculated by using its field data. Secondly, the forest sample biomass spatial database is created according to its coordinate. Thirdly, the Landsat TM image was geometrically corrected and registered with the forest sample plot biomass spatial database. Fourthly, the Landsat TM data of each forest sample plot were achieved by overlying the forest sample plot biomass layer with the Landsat TM layers. Fifthly, the forest sample plots are stratified as three main forest types such as the monsoon tropical forest, the mountainous tropical forest and the seasonal evergreen broadleaf forest. Finally, the correlations of the biomass and the Landsat TM data of the forest sample were analyzed for each forest type. The study shows: The strongest correlation relationships which are obvious at 95% level of confidence exist between the forest biomass and Landsat TM6 and TM7 for monsoon tropical forest, and are 0.549 and -0.606 respectively, and the correlation relationships which are not obvious at the level 0.05 exist between the forest biomass and the Landsat TM for both the mountainous tropical forest and the seasonal ever-green broadleaf forest..