Abstract: In Xing'an League of Inner Mongolia,we made use of Landsat8 OLI imagery and a survey of 189 plots in Armeniacasibirica,Caraganakorshinskii shrubberies,and multiple regression to establish remote sensing ground carbon storage model and analyze the prediction accuracy of the model.The results show that 27 independent variables were selected out,including single band,band combination,tasseled cap transformation,vegetation index and principal components.The Pearson correlation analysis between the 27 independent variables and shrubbery carbon storage has been calculated to select the better characteristic variables.The coefficients of determination of natural Armeniacasibirica,artificial Armeniacasibirica and artificial Caraganakorshinskii remote sensing above ground carbon storage model are 0.61,0.86 and 0.74 respectively;The prediction accuracy values of the model are 71%,77% and 73% respectively,the optimization of three kinds of remote sensing models can be applied to a range within the Inner Mongolia natural Armeniacasibirica,artificial Armeniacasibirica and artificial Caraganakorshinskii above ground carbon stocks assessment.

Key words: Landsat8 OLI, shrubbery, carbon storage model, characteristic variable selection, stepwise regression