It was assumed that leaf biomass of shrub and grass did not change much among different years, thus, leaf biomass in 1995 could be directly derived from those which Li et al. (1981) measured in 1980. Branch and root biomass in 1995 could be calculated according to their production.