Towards conditioning discrete fracture network models: a Monte Carlo simulation approach including existing site data

Abstract. Crystalline rocks are inherently more or less densely permeated by fractures at very different scales. They often represent the hydraulically dominant flow paths compared to low-permeable matrix. Consequently, finding potential sites for high-level waste repositories in crystalline host rock is challenging. Fracture inter-connectivity significantly affects geomechanical integrity and radionuclide transport (Mönig et al., 2020). Hence, an accurate representation of present-day fracture system a prerequisite describing future evolutions repository. However, information on properties spatially limited. This shows necessity combining existing data stochastic network modelling. In this manner, spaces with lack can be considered modelling (Lei 2017). Mrugalla al. (2020), generic geological models based statistical parameters were used test methodology safety analysis within scope CHRISTA II project. By using representative outcrop data, discrete (DFN) 2D generated stochastically, upscaled, mapped onto finite-element grid. individual realizations seldom accurately reality. Additionally, from must incorporated. work aims understand uncertainties regarding location, extent, other geometrical such as aperture governing patterns. The distributions these exhibit significant variations both laterally depth-wise. Uncertainties addressed superimposing large set realizations. For purpose, Monte Carlo methods applied derive best-guess network. An approximation real will require consideration known (e.g. Dorn 2013) surface subsequent underground explorations. Site-specific demonstration purposes. focus developing holistic model DFN its congruence field observations. methodological treatment underlying data-collection methods, given their resolutions, accuracies, restrictions, important aspect. Moreover, implications effects three-dimensional exemplarily investigated sense what-if scenarios. done special emphasis matrix–fracture diffusion adsorption possible retention mechanisms parameter sensitivities local changes.