Determination of structural domain boundaries in jointed rock masses: An example from the Songta dam site, China

This paper presents an application of the Kolmogorov-Smirnov and Wilcoxon rank sum nonparametric statistical tests for identifying structural domain boundaries in jointed rock masses. In the method, the upper hemispherical surface is divided into 100 nearly equal-area windows. The similarity between two samples of joint orientations is measured by comparing the frequencies or the number of joint poles occurring in the windows. Over 2400 joints collected from 8 adjacent exploration tunnels at the Songta dam site in southwest China are used to demonstrate the method. By applying the technique to the study area, structural domain boundaries in the rock mass are determined. Our results suggest that the study area, with an area of approximately 17,850 m2, can be classified into four structural domains. However, the traditional method with the correlation coefficient fails to reveal the structural changes. Since the correlation coefficient is only a measure of strength of the linear relation between two samples, it has limitations in measuring the similarity between joint orientation distributions. A comparison between the proposed method and previous methods indicates that the new technique could provide more reliable results. Besides, the new method can be applied to structural populations with small sample sizes.