Karstification of aquifers interspersed with non-soluble rocks: From basic principles towards case studies

Abstract

We have developed a numerical model able to describe the karstification of aquifers in fractured rocks containing soluble (limestone or gypsum) and insoluble layers. When water is flowing along fractures crossing the soluble layers, it is able to dissolve the material there, to increase the aperture width of the conduit, and consequently to increase the local hydraulic conductivity. Depending on the thickness and the distribution of these layers, the dissolution can be active only for limited periods, or during the whole evolution time. Fractures located in insoluble layers do not change at all. We are interested in the integral effect of these local processes and study four simplified scenarios of karstification along a prominent wide conduit crossing a fractured limestone block. We keep the initial and the boundary conditions the same for all scenarios and vary only in the amount and the distribution of the soluble material. We demonstrate that aquifers in 100% limestone, without any insoluble layers, develop along areas with high hydraulic conductivities and high hydraulic gradients, creating channel like pathways. On the other hand aquifers containing soluble layers with limited thickness develop faster and exhibit diffuse patterns determined by the chemical properties of the rock. The second part of the paper is a step towards modeling of real karst systems. We present the evolution of an aquifer located in the vicinity of a large hydraulic structure. All initial and boundary conditions, except the amount and the distribution of the soluble rock, remain the same for all scenarios. As a material example for the bedrock, we chose Gipskeuper from an aquifer along the Birs river in Switzerland. This rock consists of soluble gypsum layers and insoluble clays and marls, with typical layer thickness in the range of millimeters to centimeters. The basic processes discussed in the first part of the paper remain valid. We demonstrate that large insoluble zones can impair the karstification process and even completely block it, while areas with thin soluble layers can provide a preferential pathway and decrease the evolution times considerably. Finally we show that the evolution of the leakage rates and the head distribution within the aquifer can sometimes reveal misleading information about the stage of karstification and the safeness of the dam. Our model can be used not only to study simplified geological settings and basic processes, but also to address some of the complications arising when modeling real aquifers.