A Deep Learning-based Daily Baseflow Dataset across the United States

Figure1
41597_2025_4389_Fig2_HTML
Figure3
Figure5
Figure6
previous arrow
next arrow
Figure1
41597_2025_4389_Fig2_HTML
Figure3
Figure5
Figure6
previous arrow
next arrow

Dataset Description: Daily baseflow data, along with input datasets for 1661 basins for the hydrologica years from 1981 to 2022, can be downloaded in CSV format from the DeepBase repository on FigShare. The baseflow datafiles for the basins are zipped into archives named ‘Daily_Baseflow_Cluster[cluster_number].zip’, corresponding to their respective clusters. All the static inputs for 1661 basins are provided in a csv file named ‘Static_Inputs.csv’. The statistic attributes for the static inputs, calculated for each cluster, are provided in the file ‘14Clusters_statistics.csv’. All the dynamic forcings for 1661 basins are provided in csv files with the format of ‘Daymet_[basin_id].csv’ and are zipped into an archive named ‘Daily_DayMet_Forcings.zip’. The USGS gauge IDs of training basins (mentioned as gauged basins) are provided at ‘530basins_ids.txt’. The associated shapefiles for each cluster, including the polygons of the basins titled ‘DeepBase_Clusters.zip’ along with the PDF version of the cluster map titled ‘DeepBase_Clusters_map.pdf’ are accessible via the DeepBase repository.

Related Publications: For further details on the methodology and applications of this dataset, refer to the paper "Ghaneei, P., and H. Moradkhani (2025), DeepBase: A Deep Learning-based Daily Baseflow Dataset across the United States, Scientific Data, doi:10.1038/s41597-025-04389-y."

The dataset can be accessed at: https://doi.org/10.6084/m9.figshare.27312927.

GitHub Repositoryhttps://github.com/hmoradkhani/DeepBase