3D City-scale groundwater flow and heat transport modelling using an archetype approach

Groundwater beneath urban areas is affected by a number of natural and anthropogenic factors, such as geological and hydrogeological characteristics, as well as dense anthropogenic infrastructure, such as surface land use, heated basements, underground car parks, and train tunnels. Understanding groundwater flow and heat transport processes in such complex urban areas is therefore essential not only for planning thermal and hydraulic subsurface uses, but also for ecological and sustainable management of urban aquifers. This also involves examining the key hydrogeological or anthropogenic characteristics that influence subsurface thermal conditions, thereby supporting decision-making for management purposes. Physics-based numerical models excel at simulating flow and heat transport at different scales, adopting various assumptions. However, a realistic 3D city-scale model in a complex geological setting requires an accurate and computationally efficient approach, as management purposes need iterative simulation of different usage scenarios.

To this end, a 3D groundwater flow and heat transport model is developed for Berlin, for an area covering 118 km². It is based on a detailed 3D geological model created with Leapfrog Geo software using available geological data. To reduce computational time, the area is divided into smaller blocks, each representing the hydro-geological and anthropogenic characteristics of the heterogeneous urban area. Initially, all blocks are simulated at low resolution for different combinations of geological and anthropogenic characteristics to generate input-output relationships between these characteristics of each volume and the groundwater temperatures. Resulting input-output pairs are then used to cluster the modeled volumes into archetypes with similar hydrogeological behavior and thermal states by utilizing a decision tree approach. Finally, simulated archetypes are spatially re-combined to create a city-scale temperature map. Furthermore, to assess the key characteristics governing the thermal status of groundwater, contribution of each parameter to the decision tree is calculated.

The results reveal that, among all, the area of heated basements contributes most to groundwater temperature distribution, especially for the blocks where there is no substantial groundwater flow and conductive heat transport processes dominate. However, for the blocks with sufficient flow, upstream temperature entering the blocks is the main characteristic. Moreover, our findings highlight a close link to surface land use. As an example, the average temperature in Tiergarten location, which is a green area with a less anthropogenic influence is found to be around 11.8 °C to 12.0°C, on shallowest depth of blocks i.e. 0-50 m below ground, while in densely built-up areas, such as Alexanderplatz, the temperature could raise to 14.8°C. Furthermore, using the resulting temperature map, enables us to identify local hotspots or low-spots, which is not possible to observe in alternatives such as interpolated maps created from a dense number of measurements.

 This study presents a computationally efficient city-scale modeling approach, which in future could be utilized as a tool for performing iterative simulations for different subsurface use scenarios such as geothermal use, drinking water use, or protecting ecological habitats. Also, it provides a tool for investigating long-term climate change effects on groundwater quality and for assessing groundwater quantity and quality at city-scale.