Developing a set of large-scale 3D groundwater models using iMOD-WQ and global datasets – a pathway towards a new global groundwater model

Groundwater plays a crucial role in drinking water supply, agricultural and industrial production, and ecosystem stability worldwide. Numerical modelling has been applied in past decades to better understand the groundwater flow patterns and future threats to preserving groundwater head levels and quality facing both anthropogenic and natural threats (i.e. aquifer overexploitation and climate change impacts). Most groundwater models focus on local to regional scale groundwater systems, covering areas up to 10,000 km² and relying on local data to set up and calibrate the groundwater model. However, in recent years several attempts have been made to build a global groundwater model based on global datasets and using a combination of high-performance computing and parallel numerical code. One of the main limitations of this approach is the simplified schematization of hydrogeological heterogeneity in these groundwater models.

Therefore, this work aims to increase the realism and complexity of the hydrogeological schematizations of continental to global-scale groundwater models. To this end, we divide the globe into large-scale groundwater regions and apply a novel approach to estimate the regional-scale hydrogeological makeup of large-scale groundwater models. Three main lithological layers are defined, the most recently deposited unconsolidated sediments represent the top model layer while the second layer consists of older unconsolidated sediments. The third lithological layer consists of sedimentary rock formations, whose depth and type are defined from available global datasets (e.g. GLiM and CRUST 1.0). Additionally, the ArchPy Python library is used to further split these three lithological layers into several sub-layers representing the heterogeneous conditions (e.g. clay or sandy sub-layers). The resulting geological model is then used as a base to build a groundwater and variable density flow model, set up with the parallel iMOD-WQ code. This allows us to simulate complex large-scale groundwater processes with extensive amounts of active model cells and thus provide a better understanding of large-scale groundwater flow patterns. In the next steps, these large-scale groundwater models can be further improved by incorporating local data to create more accurate geological models and to calibrate the groundwater model input parameters.

The presented methodology was applied to create a groundwater model spanning the Australian continent, Papua New Guinea island and the continental shelf connecting these two landmasses. By applying this methodology to other large-scale groundwater regions around the world we can eventually create a new global groundwater model with higher and more realistic hydrogeological complexity and thus provide valuable insight into global groundwater flow patterns and input into Earth system models where groundwater processes are often largely simplified or neglected.