The presented dataset forms the basis for investigating present and future coupled effects of rising surface temperatures and temporal trends in groundwater recharge on subsurface pressure and temperature (PT) conditions in the North German Basin beneath the Federal States of Brandenburg and Berlin (NE Germany), for the period 1955-2100.
The study relies on a stochastic weather generator, a distributed hydrologic model, and a 3D thermo-hydraulic groundwater model to evaluate spatio-temporal subsurface feedback to two shared socioeconomic pathways (SSP) for seven general circulation models (GCM).
The results demonstrate a regional variability in both the intensity and maximum depths of projected groundwater warming, driven by hydraulic gradients and the underlying geological structure. The magnitude of groundwater warming primarily depends on the surface temperature scenario. Projected changes in recharge are not sufficient to reverse this trend, although recharge is still a key factor controlling groundwater dynamics within aquifers lying above the Rupelian Clay aquitard.
The dataset can be further utilized for assessing shallow geothermal potential and groundwater storage availability in the Berlin-Brandenburg region under climate change.

The advanced version of the non-stationary Regional Weather Generator (nsRWG; Nguyen et al. (2024)) was used to simulate long-term daily weather series for both historical (1950–2021) and future (2022–2100) periods for different GCM-SSP combinations.
The weather forcing served as input for calculating groundwater recharge fluxes via a distributed conceptual mesoscale Hydrologic Model (mHM; Samaniego et al. (2010)).
Calculated groundwater recharge and projected near-surface air temperatures were applied as a time-variable boundary condition in a regional 3-D groundwater model. A series of transient thermo-hydraulic numerical simulations was conducted using the FEFLOW 8.0 finite-element simulator (Diersch, 2013) to assess long-term changes in hydraulic head, groundwater velocity, subsurface temperatures, and heat-in-place. Simulated results were validated against monitoring data from observation wells.
The modelling approach was applied to the area of Berlin-Brandenburg, NE Germany, where a periglacial geology and geomorphology result in a complex deep-to-shallow groundwater dynamics.

The dataset comprises weather generator output, recharge and baseflow fluxes from the hydrological model, and output files from the groundwater model.
The weather generator and hydrologic model were run at daily resolution for seven GCMs (CNRM-CM6-1, INM-CM5-0, GFDL-ESM4, MPI-ESM1-2-HR, MRI-ESM2-0, CESM2, and IPSL-CM6A-LR) and two SSP’s (SSP245, SSP585). 50 stochastic nsRWG realizations were generated both for the historical case and for each GCM–SSP combination, resulting in a total of 750 realizations.
The groundwater model was calculated for both SSP, using the cross-GCM ensemble mean, 90th and 10th percentiles, as well as for discrete “what-if” recharge scenarios.

Output files are provided in open formats, like NetCDF and CSV. A complete record of groundwater model simulation (DAC format) can be explored using the FEFLOW Viewer that does not need a license.
The nsRWG is available from the GFZ GitLab repository for scientific use under the EUPL1.2 license at https://git.gfz-potsdam.de/hydro/rfm/rwg. The access will be granted by Viet Dung Nguyen upon request. The source code of the mHM (Samaniego, 2023) is publicly available at https://doi.org/10.5281/zenodo.4575390. FEFLOW Viewer can be downloaded from https://www.dhigroup.com/technologies/mikepoweredbydhi/feflow/download-feflow-installation