Loading...
Loading...
Contributors:
Dataset URI:
10.5281/zenodo.23733524This repository provides the dataset accompanying the GEMS-GER benchmark for machine-learning-based groundwater modeling. It includes long-term groundwater-level time series, meteorological and hydrological forcing data, site-specific environmental properties, and benchmark results. The data originate from official public sources and have been harmonized across Germany’s 16 federal states (Bundesländer). The benchmark is described in Ohmer et al. (2026), “GEMS-GER: a machine learning benchmark dataset of long-term groundwater levels in Germany with meteorological forcings and site-specific environmental features”. Contents of this repository: Groundwater-level time series (GEMS-GER_data/dynamic/MW_*.csv): Weekly data from 3,207 monitoring wells for 1991–2022, including: Daily temperature (mean, minimum and maximum) Precipitation and humidity (HYRAS/DWD) Actual, potential and reference evapotranspiration Soil moisture and soil temperature at 5 cm depth Snow water equivalent, snowmelt and runoff (ERA5-Land) GWL_flag, indicating observed and imputed values Static site descriptors (GEMS-GER_data/static/static_features.csv): Hydrogeology and soil properties Land use and climate classifications Elevation and derived topographic parameters, such as slope and TWI Benchmark model performance (GEMS-GER_data/model_performance/): Per-well NSE, RMSE, R² and Bias for the single-well, global dynamic-only and global dynamic-plus-static models. Scores are calculated from the pointwise median predictions of 20 ensemble members per model. model_performance_single.csv — per-well scores for the single-well model model_performance_global_dynonly.csv — per-well scores for the global dynamic-only model model_performance_global_dynstat.csv — per-well scores for the global dynamic-plus-static model median_predictions.csv.gz — pointwise median predictions for all three models GEMS_GER_scores.pdf — summary table of the revised results README.md — methods, comparison with the published results and instructions for reproducing the scores Time-series figures (GEMS-GER_data_figures/DYN_Feat_MW_*.pdf): Plots of groundwater levels and selected forcing variables for each well. The figures are provided separately to reduce the size of the main dataset download. Directory structure: GEMS-GER_data/ ├── dynamic/ # Weekly time series for 3,207 wells │ ├── MW_1.csv │ ├── MW_2.csv │ └── ... │ ├── static/ # Site-specific static descriptors │ └── static_features_MW_1toMW_3207.csv │ └── model_performance/ # Updated 20-member benchmark results ├── median_predictions.csv.gz # Median predictions for all three models ├── model_performance_single.csv # Single-well CNN scores ├── model_performance_global_dynonly.csv # Global LSTM scores: dynamic inputs ├── model_performance_global_dynstat.csv # Global LSTM scores: dynamic and static inputs ├── GEMS_GER_scores.pdf # Summary table of the results └── README.md # Methods and score reproduction GEMS-GER_figures/ ├── DYN_Feat_MW_1.pdf # Input time series visualizations ├── DYN_Feat_MW_2.pdf └── ... The dataset is intended for research and benchmarking in hydrogeology, data-driven groundwater modeling, and environmental machine learning. It forms the basis of the GEMS-GER benchmark, as described in the associated preprint. All data originate from public sources and have been harmonized across administrative and institutional boundaries to enable consistent large-scale analysis. All associated code, documentation, and update announcements are maintained in the project's GitHub repository: https://github.com/KITHydrogeology/GEMS-GER ensuring transparency, traceability, and reproducibility.
Keyword(s):
groundwater levels, hydrogeology, environmental data, Germany, machine learning benchmark, meteorological forcing, groundwater monitoring, geospatial dataset, hydrology, time series.
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)