Abstract: Weestimated Hippophae rhamnoide plantation biomass using satellite images of Landsat TM(2011)based on RS and GIS. In one-dimensional linear regression analysis models,the NDVI-based and RVI-based simple linear model were the optimal model for estimating biomasswith the coefficient correlation(R²) of 0.908 6 and 0.868 5, respectively. The multiple regression model established by four kinds of VI could be used for estimating forest biomass with the coefficient of determination(R²) of 0.909. Themultivariable linear regression models were better than one-dimensional linear regression analysis models, and the predicted biomass based on regression analysis models was higher than the theoretical biomass.

Key words: Hippophae rhamnoide, plantation biomass, vegetation index(VI), remote sensing, vegetation index