3D modelling of rock mass heterogeneities in unsaturated karst using geophysics, clustering and geostatistics

Conserving an underground heritage site in a karst environment poses significant challenges because of the complexity of the physical phenomena involved. The deterioration of cave walls is mainly caused by condensation and infiltration, which originates from the thermal behaviour of the cave and the surrounding rock mass. This phenomenon is difficult to describe because it depends on the heterogeneity and water content of the unsaturated rock mass. However, it is clear that the conservation of cave art is a matter of concern for all caves worldwide. The purpose of this study is to analyse the heterogeneities surrounding Lascaux Cave and identify those which are water-saturated (during the wet season) that could significantly influence the cave thermally. The rock mass was characterised using the non-destructive 2D electrical resistivity tomography (ERT) method. Before kriging the ERT data, Hierarchical Agglomerative Clustering (HAC) was performed on the dataset. This classification provides class boundaries for categorising the kriging-estimated resistivities. The results of this study significantly improve our understanding of rock mass from geological and hydrogeological perspectives. The creation of a 3D model helped to identify four areas of lower resistivity (water storage reservoirs), indicating the possible presence of water as the rock mass was partially saturated. Those water storage reservoirs coincide with the water flows recorded during wet periods inside the cave and are associated with known geological fractures. Moreover, identifying the different materials surrounding the cave and the saturated areas in the 3D model allows us to understand the deviations in the temperature measurements in different parts of the cave and will improve the accuracy of the heat transfer simulations.