Landslide Analysis with Incomplete Data: Examples from Colorado and Wyoming

摘要

Although there is no substitute for careful investigation for slope stability analysis, in reality we are often missing important parameters, yet we still must proceed with the study. Several tools allow us to reduce error and estimate both current and future stability. First, analysis of several cross-sections permits iteration as parameters are modified to meet an expected factor of safety of unity at initial failure for all sections. Concurrently, unknown parameters can be analyzed within reasonable bounds, and these bounds tightened during the iterations. For large slides, there are often smaller failures off the toe that can be used as additional calibration points to constrain input parameters. Sensitivity analysis can help focus testing on vital parameters, reduce error, and improve confidence in the results. Sensitivity analysis unique to the site is preferred, but general studies have shown that cohesion is equally important as friction angle for deeper failures, but is about twice as important for shallow failures. Groundwater levels are about half as important as strength parameters. A third important item is attention to material strength values, including inhomogeneity, anisotropy, peak versus residual, slaking, porewater pressure effects in the capillary zone, and critical state effects such as dilation or contraction. Example applications of these tools are given for sites in Colorado and Wyoming.