Spatial assessment of geomechanical rock mass properties by geostatistical analyses

The forecast of geomechanical rock mass properties can be an important goal in civil and mining engineering planning, especially when a wide area is involved. This work is a contribute in assessing how the geometrical and mechanical properties of the rock masses can be treated as regionalized variables. Indeed the present study deals with the reconstruction of main rock mass characteristics in an extended area, from punctual measurements of the rock mass features, through the use of different geostatistical techniques. The research site is located in the Italian Central Alps, along the Chiavenna Valley (SO), which is composed by San Giacomo and Bregaglia valleys; it covers an area of about 200 km2. The regional geological setting is related to the Pennidic Nappe arrangement, characterized by the emplacement of sub-horizontal gneissic bodies resulting from the Mesoalpine isoclinalic folding of crystalline basements (“Tambò” and “Suretta” Units) emplaced throw East and separated by metasedimentary cover units. More than one hundred detailed structural and geomechanical field surveys, mainly located in San Giacomo Valley, unevenly distributed, were carried out in order to characterize the rock masses, in accordance with the I.S.R.M. suggested methods for the quantitative description of discontinuities in rock masses. The investigated lithologies belong to the “Tambò” Unit, to its meta-sedimentary cover and to the “Suretta” Unit. Geomechanical surveys allowed to identify the number of joint sets, their average orientations, spacing, persistence, roughness, aperture, presence and nature of filling, wall strength, weathering and moisture. From the collected data the rock mass quality indexes, such as the Rock Mass Rating and the Geological Strength Index, have been evaluated in each surveyed site. The geomecanichal parameters derived from the measurement stations are local data, therefore in order to know how they are distributed outside the measurement points, in the entire study area, different geostatistical techniques were applied. First of all, the field data of the main rock mass features, such as the intercept of discontinuities, fracture intensity and aperture, were analyzed, and the main statistical parameters were calculated. Then a procedure of Gaussian anamorphosis was performed when necessary. Afterward the correlation structures of the defined rock mass features were investigated, at different scales, by means of a semivariogram analysis, taking into account the occurrence of anisotropic correlations. Both kriging method and conditional simulation techniques were applied to estimate the geomechanical variables. The best results, supported by the validation process, are obtained by conditional simulation method, which respect local extreme values, compared with kriging technique which tends to produce smoothened distributions.