Back analysis of an ancient rockslide at Lake Fundudzi, Limpopo Province, South Africa.
Chiliza, Sibonakaliso Goodman.
MetadataShow full item record
A massive rockslide event occurred in the Soutpansberg Mountains of the Limpopo province approximately twenty thousand years ago that blocked the course of the eastern flowing Mutale River forming Lake Fundudzi, which is the only known true inland lake in South Africa. The landslide occurred in the northward dipping and highly jointed Fundudzi Formation of the Soutpansberg Group. The failed slope was observed to be a homogenous jointed sandstone rock slope which is quartzitic in places. The techniques of back analysis have been employed in establishing the conditions that may have caused the landslide. Kinematic analysis using Rocscience Inc. Dips revealed three major joint sets (JS1, JS2 and JS3) in the rock mass which are oriented northeast-southwest, north west-south east, and north-south respectively. The main failure plane best correlated with JS2, which dips towards the lake. The kinematic analysis also revealed that the mechanism of failure was planar which is in agreement with site observations. The strong rock material classification of sandstone from laboratory index and mechanical tests was one of the indicators that failure could have been mostly structurally controlled than rock material degradation. Rock mass classification using Rock Mass Rating (RMR) and Slope Mass Rating (SMR) showed that prior to the inclusion of discontinuity adjustment in the ratings, which reflects the relationship between the slope face and joint dip, the rock slope classified as Class II for „good rock mass‟. However, upon incorporating rating adjustment in RMR and SMR for discontinuity orientation, the rock mass quality of the slope dropped to „poor‟ and „bad‟ respectively. The techniques of back analysis were employed using Slide and RocPlane in Rocscience Inc. suite in order to investigate the geotechnical conditions that may have resulted in failure. Prior to conducting the back analysis, sensitivity analysis was carried out for a realistic range in values of the shear strength parameters, in both Slide and RocPlane in order to investigate the parameter which has the most impact on the factor of safety (F). The results obtained from the sensitivity analysis in both Slide and RocPlane showed that the joint roughness coefficient (JRC) followed by the residual basic friction angle (ϕr) were the main influential shear strength parameters on F as opposed to the joint wall compressive strength (JCS) which has very little impact on F for the range of values considered. The shear strength parameters derived from Slide by back analysis were found to be notably higher than the parameters obtained in RocPlane under various groundwater and seismic ground acceleration scenarios considered in the back analysis. This showed that different methods of stability analysis can give different results. Using RocPlane seems more plausible than Slide as the rockslide could have failed as a single unit. Results also showed that seismic ground acceleration has more impact on the back analysed ϕr and JRC values from a pseudostatic analysis in Slide and RocPlane than groundwater. This suggests that ground acceleration due to seismicity was more likely the trigger for the Lake Fundudzi rockslope failure than groundwater. This study shows the usefulness of back analysis techniques in assessing various conditions of failure that may have resulted in ancient landslides and avoids many of the problems associated with laboratory tests. Although it may not result in a unique set of geotechnical parameters that may have resulted in failure, it however gives an insight into probable causes of failure and thus can be useful in mitigation of future failures. Results obtained from back analysis must be treated with caution as many uncertainties exist which can be reduced by making sensible, careful and realistic assumptions.