“Seismic Stability Chart For Vegetated Homogeneous Dry Soil Slopes”

Abstract

Nepal is mountainous country perched in the subduction zone of Indian & Tibetan plate. Most of the landscape being covered with vegetation and being placed at the seismically prone area, a seismic stability chart that incorporates the effect of vegetation can aid in the preparedness against seismically induced landslide hazards. Development of stability chart is a tedious, time consuming and computationally intensive task. So this thesis work presents a set of codes written in Visual Basic 6.0 that exploits the potential of command driven software SPECFEM3D_GEOTECH in batch processing. Using batch processing, factor of safety of 8640 models have been computed. The models include 4 slope geometry 2H:1V, 1.5H:1V, 1H:1V & 0.5H:1V, having soil cohesion 1, 5, 10, 15, 20, 25 KN/m2, and friction angle, 5, 10, 15, 20, 25, 30, 35, 40 degrees. Effect of vegetation has been modelled with a different layer of top soil; which has thickness equal to that of root depth and soil cohesion increased by a value equal to that of root cohesion. Root cohesion value 1, 5, 10, 15, 20 KN/m2 and root depth 1, 2, 3 m have been adopted. And, the models are subjected to horizontal pseudo static seismic excitation of 0.1, 0.2, 0.3g along the direction of slope. As the final outcome of this thesis work a set of nine charts have been developed in the format as prescribed by Michalowski (1998). These charts can be used to assess factor of safety of vegetated homogenous dry soil slopes. However, as these charts cannot distinguish whether the stability of slope is due to application of vegetation or the soil itself, three additional charts have been proposed to quickly assess the significance of vegetation in slope stability. Further using SPSS, multi variable linear regression analysis has been performed to formulate an equation for FOS of vegetated dry homogenous soil slopes as given below; = 0.533 + 5.397

+ 1.344 tan ∅ − 0.301 tan + 0.035 + 0.006 − 1.521 The equation has regression coefficient R2 0.916 and standard error 0.137. Verification of the work has been carried out using commercial software Phase 2; which shows a strong correlation with the results from SPECFEM3D_GEOTECH.