Multivariate statistical analysis of hydrogeochemical data towards understanding groundwater flow systems in granites

The present study focuses on hydrogeochemical data from a Hercynian granitic mass where a new underground hydraulic circuit for a hydropower plant is being developed between two reservoirs and near another existing and operating underground hydraulic circuit. The main goal of this work is to assess the use of multivariate statistics to distinguish groundwater from different flow systems, to aid in evaluating the possible infiltration of water from the existing hydraulic circuit into the new one. Sampling points were selected inside the tunnels (eight points at the new circuit and one point at the existing circuit), near the surface (five points along the new circuit major axis) and on Venda Nova dam reservoir (one point). Based on laboratory determinations of major ions, three distinct groups of waters with different physicochemical characteristics and spatial distribution were established by cluster analyses. The relations of these groups with the characteristics of the occurrences are discussed with the support of principal component analysis. A hydrogeochemical conceptual model for groundwater flow systems related to infiltration and water-rock interaction in this granite rock mass is proposed based on multivariate statistical analyses and geological information.