The role of vegetation in preventing shallow soil mass movement is now fairly well understood, particularly at the individual plant level. However, how soil is reinforced on a larger scale and the influence of changes in vegetation over time has rarely been investigated. Therefore we carried out a study on the temporal and spatial changes within stands of Cryptomeria japonica D. Don, growing in the Sichuan province of China, an area where shallow landslides are frequent. Soil cores were taken from three neighbouring stands of C. japonica aged 9, 20 and 30 years old and growing on steep slopes. Cores were taken from around trees and the root (<10 mmin diameter) biomass density (root density (RD)) present in each core wasmeasured at different depths. The spatial position of trees at each site was noted and soil shear strength was measured. The tensile strength of a sample of roots from each stand was measured. Using the RD data, the root area ratio (RAR) could be estimated. RAR and root tensile strength were used as input to a model of root reinforcement which determines the additional cohesion, cr, or contribution of vegetation to soil. Data were then incorporated into a twodimensional model of slope stability developed in the finite element (FE) code, Plaxis, which calculates the safety factor (FOS), or likelihood of a slope to fail under certain circumstances. We calculated the FOS of slopes with and without C. japonica, taking into account the spatial position of trees at each stand. Results showed that RD was highest in the 9-year-old stand, but that root tensile strength was lowest. In the 30-year-old stand, RD was low but a higher root tensile strength compensated for the decrease in RAR. The FOS increased by only 15-27% when vegetation was present, with the greatest augmentation in the 9-year-old stand. The older stands had been thinned over the years, resulting in large gaps between trees,whichwould be prone to local soil slippage. This spatial effect was