Integrating Chemical Modelling and Geostatistics for Improved Radionuclide Retention Models in Crystalline Rock

Ensuring the safety for deep geological repositories for nuclear waste in crystlline host rock necessitates a comprehensive understanding of the far field and it's potential for radionuclide retention. In case of a repository leakage, radionuclides may get mobile and migrate through pathways in rock and aquifers. To asses the uncertainties in forcasting the migration of radionuclides it is essential to incorporate naturally occurring heterogeneities in rock composition and geological structures into the models, e.g. heterogeneities occurring near intrusion margins, tectonically influenced granitic bodies, or metamorphic formations like gneisses. This complexity significantly impacts the modeled radionuclide retention potential compared to simplistic isotropic granite models.
The SANGUR project (Systematic Sensitivity Analysis for Mechanistic Geochemical Models using Field Data from Crystalline Rock) aims to identify crucial parameters and their uncertainties essential for modeling radionuclide retention in crystalline rock. Our study presents a comprehensive workflow modeling how petrological variations in both granitic and metamorphic crystalline host rocks influence radionuclide retention. Utilizing Multinary Random Fields geostatistics, we simulate crystalline rocks based on analyzed spatial rock data to quantify uncertaintiesand to determine the appropriate model scale. The petrological variance is then considered for the chemical modeling through software such as PHREEQC or Geochemist's Workbench©: Surface Complexation Models (SCM) in chemical modeling software calculate partition coefficients (Kd values) for radionuclides, such as uranium, in diverse mineral environments in combination with varying aqueous phases. To enhance and simplify models, global sensitivity anlsysis is applied to determine critical features for radionuclide retention.