We provide a single file (exodus II format) that contains all results of the modeling efforts of the associated paper. This encompasses all structural information as well as the pore pressure, temperature, and fluid velocity distribution through time. We also supply all files necessary to rerun the simulation, resulting in the aforementioned output file.

The model area covers a rectangular area around the Central European Basin System (Maystrenko et al., 2020). The data publication is compeiment to Frick et al., (2021). The file published here is based on the structural model after Maystrenko et al., (2020) which resolves 16 geological units. More details about the structure and how it was derived can be found in Maystrenko et al., (2020). The file presented contains information on the regional variation of the pore pressure, temperature and fluid velocity of the model area in 3D. This information is presented for 364 time steps starting from 43,000 years before present and ending at 310000 years after present.

This model was created as part of the ESM project (Advanced Earth System Modelling Capacity; https://www.esm-project.net). This project looks at the development of a flexible framework for the effective coupling of Earth system model components. In this, we focused on the coupling between atmosphere and the subsurface by simulating the response of glacial loading, in terms of thermal and hydraulic forcing, on the hydrodynamics and thermics of the geological subsurface of Central Europe. For this endeavor, we populated the 3D structural model by Maystrenko and Coauthors (2020) with rock physical properties, applied a set of boundary conditions and simulated the transient 3D thermohydraulics of the subsurface. More details about this can be found in the accompanying paper (Frick et al., 2021)

For creating this 3D structural model numerous datasets have been integrated. For this we first visualized all data, that is geological cross-sections, drilled well tops, water depths, seismic lines and larger scale models using the commercial software Petrel (©Schlumberger). We then split those datasets into the desired output horizons, removing inconsistencies between them, and using the scattered information of each of the units top elevations to interpolate to regular grids. This was done by the convergent interpolation algorithm of Petrel and a regular grid resolution of 100 m. Especially for the deeper units where only sparse information can be obtained from drilled well tops, we relied on existing models of the Central European Basin System and of the Northeast German Basin which integrated all available GDR seismic lines and are gravity constrained. These have been used along with the 3D Brandenburg model to provide the carcass for the model where no local information was available. Therefore, the crust, mantle and Pre-Permian sediment configuration was derived from larger scale models. For the overlying model units available deep seismic lines along with all deep wells were integrated. For the shallower model units (i.e. Cenozoic) highly resolved geological cross-sections and a dense population of wells were integrated along with the seismic lines. In a final step, high resolution data of the topography (i.e. lake surface and earth surface) were combined with lake bathymetry data to derive the geological surface of the model.

The grids provided are space separated ascii files for a) the elevation of the top and b) the thickness of each unit, with their structure being identical. The columns for a) are 1: x-coordinate, 2: y-coordinate, and 3: elevation (meter above sea level). For b) the columns are 1: x-coordinate, 2: y-coordinate, and 3: thickness (meter). The horizontal dimensions are 43.5 x 53 km. The resolution of the files is identical, each having a spacing of 100 m. The associated coordinate system is Gauß-Krüger DHDN Zone 4. The naming of the files includes the layer name (geological unit) as well as a number representing the structural position in the model in ascending order. Hence, recomposing the model one would have to order the grids by ascending number to build the model from top to bottom. The vertical resolution of the model is heterogeneous since model units have heterogeneous distributions. A thickness of "0" is denoted where the unit is absent.