Joint inversion of physical and chemical state variable data within a multi-model subsurface hydrology framework

Casillas Jesus


Walraevens Kristine,

SCK•CEN Mentor

Rogiers Bart
+32 14 33 31 23

SCK•CEN Co-mentor

Beerten Koen
+32 14 33 32 41

Expert group

Engineered and Geosystems Analysis

PhD started


Short project description

The hydrogeological models used in evaluation of radiological risk and impact require reliable model parameters at the model scale to produce trushworthy results. They need to be calibrated using observations, typically hydraulic heads. However, using a single type of observations often leads to parameter correlation of unsensitivity rendering the estimated parameters less unique and more uncertain. In this PhD topic, the student shall collect other observation types along the hydraulic head observations, such as temperature and flux measurements, and age tracer concentrations and use them in the calibration (or estimating uncertainty of) groundwater flow and solute transport models. The case study is the Neogene aquifer in the Nete catchment. Besides collecting novel types of state variables, the main scientific challenge in this topic consists of integrating the new types of observations in nummerical models at the catchment scale, as the new observations represent processes at different scales. The main innovative aspects are: (1) sampling, analysis and integration of 4He tracer concentration (constraining flow in the deep aquifers) along with the young age tracers; (2) perform spatially distributed material parameter inversion, to deal with heterogeneities which are not part of the conceptual model, across coupled models. The modelling tool which shall be used is this year’s new MODFLOW release (MODFLOW 6), which allows for including multiple models, both loosely or thightly coupled, in a single simulation. It would allow to address properly the integration of catchment and local-scale processes. Although there is no transport code for this version of MODFLOW yet, the heat transport could be performed with MODFLOW 6 itself, and our own random walk particle tracking code (modified for MODFLOW 6) can be used for calculating age tracer concentrations.


- Sample, analyse and interpret novel types of observations, such as groundwater tracers, flow and temperature measurements;

- Develop a multi-scale coupled  modelling framework integrating the novel types of observations;

- Perform spatially distributed material parameter inversion of the integrated model.