Construction of hydrostratigraphic grid models for the estimation of subglacial erosion during future glaciations of the North German Basin

During the Elsterian glaciation, subglacial tunnel valleys were deeply incised into the subsurface of the Northwest German Basin. These tunnel valleys typically range in depth from 100 to 400 m but can reach depths of more than 500 m. As the BGE (Federal Company for Radioactive Waste Disposal) plans a high-level radioactive waste disposal at depths between 300 and 1500 m, it is important to consider the subglacial erosion potential during future glaciations, in order to ensure a long-term safety of the potential site. In this project, we first constructed 3D geological subsurface models and currently are carrying out numerical simulations to quantify the meltwater-driven erosive potential during future glaciations in the Northwest German Basin. We are developing and deploying a next-generation dynamical model for subglacial meltwater erosion on soft beds. This hydraulic model, based on principles of dynamical subglacial channel formation and fluvial erosion, is parameterised against tunnel valley formation during past glaciations. Once calibrated, our hydraulic model will estimate meltwater-driven erosion and sediment transport during future glaciations, with particular emphasis on the maximum depth of meltwater erosion. The hydrostratigraphic 3D reservoir grid models of the Northwest German Basin are used as input for the numerical hydraulic modelling. The lithology and hydrology of the subsurface will considerably influence the location and depth of future tunnel valleys. These hydrostratigraphic 3D subsurface reservoir grid models cover Permian to Cenozoic sediments, have an area of about 40,000 square kilometres and reach a depth of 2,000 metres. We constructed these 3D subsurface models by using a layered-structural-model and voxel-grid-models approach, enabling us to generate grid models with varying resolution from the same structural model. To build the layered structural model, we utilised existing stratigraphic 3D models (GTA3D, TUNB3D-NI and small-scale regional models) and additional borehole data. The reservoir grids integrate constant permeability values reflecting the hydrogeological properties of the stratigraphic units. This approach facilitates a rapid construction of grid models of different sizes, despite a highly heterogeneous database. We will compare the outputs of the subglacial erosion modelling with variably resolved grid models to assess the effects of different input data (e.g., lithological data, facies architecture, and related variations in hydrogeological properties). The results are intended to deliver a firm base for future long-term safety considerations of potential repository sites.