Effects of roots on soil slippage were investigated on unstable hillsides covered by chaparral, prescription-burned chaparral, and grassland in the Transverse Ranges of southern California. Direct shear tests were conducted on large (0.25-m diameter x 0.61 m) "undisturbed" soil samples extracted from slip-prone landscapes under the three vegetation types. Roots were decayed from small (61-mm diameter x 25 mm) cores extracted from each large shear surface and were used to establish the strength of soil without roots. Absolute and relative contributions of roots to soil strength were very significant, although average root-strength contributions did not differ among the three vegetation types. Unburned chaparral tended to provide sparsely distributed but occasionally great root reinforcement. Burned chaparral and grassland landscapes provided more uniformly distributed root reinforcement with lower maximum local contributions to soil strength than chaparral. Total shear strength of soils was mobilized at greater horizontal displacements in burned and unburned chaparral than in grassland.
We developed a three-dimensional analysis of slope stability, using representative distributions of root-strength values. This model predicted systematic differences in minimum scar size between vegetation types. These differences arose not from the average root-reinforcement values which were the same for each type, but rather from the differing distributions of root contributions to soil strength found for each vegetation type. This heuristic model suggests that the generally observed higher frequency of landslides in chaparral lands after conversion to grass cover may result from reduced variation in root reinforcement, which permits smaller slips to develop.