Groundwater models are central to demonstrating long-term containment in deep geological repositories within the multi-barrier concept, where geological, engineered, and geotechnical barriers jointly limit radionuclide release to the biosphere. In the German context, safety assessments must address a one-million-year evaluation period with stringent cumulative (≤ 10⁻⁴) and annual (≤ 10⁻⁹) fraction release constraints, sharpening requirements on model scope, coupling, and reliability.

This contribution summarizes current challenges and opportunities for groundwater flow and transport code development across crystalline, clay-rich, and salt host rocks. Key modeling challenges include 1) fracture-network representation, 2) variable boundary conditions over geological times, and 3) integration of multi-scale, coupled thermo-hydro-mechanical-chemical processes that govern low-concentration radionuclide migration over kilometer scales and up to a million years. Host-rock-specific flow regimes including saturated flow in fractured crystalline rocks, matrix-diffusion and sorption in clays, and extremely low hydraulic conductivity in low-permeability rocks (e.g., salts, granites) dictate the modeling task and code requirements.

We critically review the groundwater flow and transport codes used internationally for repository evaluations. The primary software platforms for the developed codes include MODFLOW, FEFLOW, SUTRA, HYDRUS, OpenGeoSys, COMSOL, CRUNCHFLOW, and SPRING, among others, mapped by their process coverage. Practical repository-scale groundwater models need to be capable of 1) representing regional 3D geology (e.g., faults, fractures), 2) performing saturated and unsaturated variable-density flow, and 3) coupling surface-water features (e.g., rivers) and transient recharge, to perform long-term prognoses using expanding time steps. Multi-component transport (e.g., tracers and radionuclides with decay chains) is handled via integrated transport modules (e.g. random-walk solute transport model). Global/local sensitivity analysis and automated calibration are generally feasible through built-in or external optimization frameworks, yet end-to-end workflows remain labor-intensive for multi-scenario uncertainty exploration.

We propose a selection framework linking host rock and assessment objectives to code capabilities: 1) fracture representation accuracy (continuum vs. discrete modeling), 2) scale-bridging and boundary evolution over geologic time, 3) variable-density flow and reactive transport, and 4) verification/validation against observations over million-year periods. Applying this framework clarifies trade-offs among variable software platforms and highlights development priorities, particularly THMC integration, and probabilistic workflows for radionuclide transport, for credible, regulation-ready safety assessment of deep geological repositories.

References

COMSOL AB. (2023). COMSOL Multiphysics Reference Manual.

FEFLOW, Finite Element Subsurface Flow & Transport Simulation System, WASY GmbH, Berlin, Germany, 2007.

König, C., Becker, M., Diehl, A., Seidel, T., Rosen, B., Rüber, O., and Zimmermann, C., (2018). SPRING - Simulation of Processes in Groundwater. Witten, 5. Edition.

Kolditz, O., et al. (2012). OpenGeoSys: An open-source initiative for numerical simulation of thermo-hydro-mechanical/chemical (THMC) processes in porous media. Environmental Earth Sciences, 67(2), 589–599.