Rock mass statistical homogeneity investigation along a highway corridor in Vietnam

The mechanical and hydraulic behaviours of discontinuous rock masses and rock slope stability depend very much on the discontinuity geometry pattern of the rock masses. Identification of statistically homogeneous regions is a pre-requisite to model the discontinuity geometry patterns in a rock mass. The review performed on the available geologic information such as the lithology, the existing fault system, and the tectonic history of the site revealed that the studied site had undergone several tectonic activity phases resulting in a quite complex rock mass statistical homogeneity. This finding led to further investigation on rock mass statistical homogeneity. Three different methods, (a) polar equal area pole plot comparison, (b) linear correlation coefficient analysis and (c) contingency table analysis, were then applied to evaluate the statistical homogeneity of rock masses that exist along a 15 km stretch of the 3B Highway corridor in Vietnam. This study clearly illustrates that at present no strongly reliable objective method is available in the engineering geology or rock mechanics literature to firmly demarcate between the statistical homogeneity and non-homogeneity of rock masses. The study shows that the available methods can only be used either to qualitatively assess or to rank rock masses with respect to the strength of statistical homogeneity. The applied different methodologies led to varying ranking orders on the strength of statistical homogeneity of rock masses. That implies the difficulty in making a judgement on rock mass statistical homogeneity using results obtained from only one method. The paper shows how better decisions were made on rock mass statistical homogeneity by first making a qualitative assessment through examination of available geologic information and then comparing the results obtained from the three different methodologies.