Most landslides in the Saratoga foothills, California, occur along previously sheared surfaces formed in overconsolidated claystone units of the Plio-Pleistocene Santa Clara Formation. Landsliding occurs as the result of progressive shear failure, and the primary geologic factors that control landsliding are the presence of a smectite clay mineral and the overconsolidated nature of the claystone.
Preexisting landslide surfaces in the claystone units are usually at, or near, residual shear strength, which is a measure of the frictional strength of the clay minerals present. To evaluate the effects of clay mineralogy on residual shear strength, laboratory testing and x-ray diffraction were performed on claystone samples collected from landslide-prone areas in the Saratoga foothills, California.
Claystone samples tested consisted of thirteen disturbed bulk samples and one intact block sample, carved out of the sidewall of a 30-in. (76-cm) diameter drill hole at a depth of 30 ft (9 m), which contained a well-developed landslide shear surface. The bulk samples were analyzed for Atterberg limits and gradation, and their clay minerals were identified and quantified using x-ray diffraction techniques. Thirteen claystone specimens were fabricated from the bulk samples and tested for residual shear strength. The block sample was analyzed for Atterberg limits, and the residual shear strength of the intact landslide surface was measured.
It was concluded from x-ray diffraction analysis that the predominant clay mineral is a smectite and is present in quantities ranging from roughly 13 to 37 percent of the total sample and roughly 34 to 100 percent of the clay fraction (less than .002 mm). The 13 fabricated claystone specimens possess tan ϕr′ values ranging from 0.105 to 0.456, with an average value of 0.210. The intact landslide shear surface possesses a tan ϕr′ value of 0.080, which is the lowest frictional strength measured during this study. Based on published experimental data, pure sodium smectite possesses tan ϕr′ values generally less than 0.200.
Five correlations were observed for the claystone samples. Three correlations useful for estimating residual shear strength for landslide repair design are tan ϕr′ versus plasticity index, liquid limit, and percent clay fraction (less than .002 mm). Least squares regressions were performed for the three tan ϕr′ correlations and the correlation between liquid limit and plasticity index. Good correlations involving tan ϕr′ were found as indicated by r2 regression values equal to, or greater than, 0.795. A good correlation between liquid limit and plasticity index was observed, as indicated by an r2 regression value of 0.895. A direct 1 to 1 ratio was calculated for percent clay fraction versus plasticity index, and an activity ratio of 1.07 was calculated. An activity ratio of about 4, judged to be more realistic, was calculated from the plasticity chart which is not susceptible to the masking effects of other clay minerals.