The correlation analysis of the Landsat TM data and its derived data with the biomass of the tropical forest vegetation

The correlation analysis of the Landsat TM data and its derived data with the biomass of the tropical forest vegetation is explored here in Xishuangbanna, Yunnan province, P.R. of China. It includes four steps. Firstly, the biomass for each sample is calculated by using the field inventory data of each sample. Secondly, several vegetation indices are derived from Landsat TM. Thirdly, principal component analysis is used to analyze Landsat TM. The remote sensing data including Landsat TM data, its derived data and the biomass of each sample are referenced to the same projection and coordination. Thirdly, the Landsat TM data and its derived data of each sample are obtained. Finally, the correlation between the biomass, the Landsat TM and its derived data is analyzed. It is shown as follows: (1) the correlations between Landsat TM and the forest vegetation biomass are arranged as TM5, TM7, TM4, TM3, TM2, TM1, TM6 in descending order. (2) The correlations between several vegetation indices and the forest vegetation biomass are arranged as VI3, DVI, PVI, SARVI, RVI, NDVI, TSAVI in descending order. (3) The correlations between the principal components of Landsat TM images and the forest vegetation biomass are arranged as the second principal component, the first principal component and the third principal component from higher to lower. (4) The correlations between the spectral indices such as Bright Index (BI), Green Vegetation Index (GVI), and Wetness Index (WI) and the forest vegetation biomass are arranged as BI, WI and GVI in descending order. Among all the correlations, the highest is the correlation between the forest vegetation biomass and the second principal component that is outstanding at the 0.01 level and its coefficient reaches -0.231.