Résumé
In karst aquifers, groundwater flow heavily depends on underground conduits called karst networks. To model this flow spatially accurately, explicit representation of these conduits is necesary through distributed approaches. However, the thickness of conduits varies greatly, making their direct exploration and mapping with methods like seismic reflection or electrical resistivity quite challenging. Stochastic simulation of discrete karst networks is a helpful approach to overcome the large uncertainties on conduit position and geometry. Yet, among existing methods in the literature, many are limited in capturing the full range of possible cave patterns determined by the specific conditions of karstification, including branchwork, anastomotic and angular patterns.To overcome this issue, we propose to use the newly developed karst simulation method KarstNSim, which solves a shortest path problem between sinks and springs – respectively the inlets and outlets of the network – with the use of an anisotropic cost function defined on an unstructured mesh conformal to geological and structural heterogeneities. This cost function represents the physico-chemical processes that govern speleogenesis – such as erosion and chemical weathering – providing simplified control over the morphometry of the generated networks. The method encompasses geological parameters such as inception surfaces, fractures, permeability, ghost-rocks and solubility of layers, along with considering the hydrological context of recharge by assigning relative weights to the inlets.We will explore the variability of morphologies that the method can reproduce through different tests based on simple (hydro)geological models with varying geology and boundary conditions, represented by different parameters for the cost function.