Impact of climate change on changes in streamflow intermittence over European reaches

1. Overview and background 2. Repository content 3. Data format and projection 4. License and citations 4.1 License agreement 4.2 Citations and acknowledgements 1. Overview and background This documentation describes the data used/produced for the research article “Abbasi, M., Messager, M. L., & Döll, P. (2025). Increased streamflow intermittence in Europe due to climate change projected by combining global hydrological modeling and machine learning. Earth's Future, 13(12), e2024EF005868. https://doi.org/10.1029/2024EF005868" In this study, we quantified the impact of climate change on the streamflow intermittence status for about 1.5 of European reaches for two emission scenarios in three different periods such as a reference period (1985-2014), the 2050s (2041-2070), and the 2080s (2071-2100). This research follows up on Doll et al. (2024), which developed a statistical framework for predicting streamflow intermittence at high resolution on a continental scale. Reference: Döll, P., Abbasi, M., Messager, M. L., Trautmann, T., Lehner, B., & Lamouroux, N. (2024). Streamflow intermittence in Europe: Estimating high‐resolution monthly time series by downscaling of simulated runoff and random forest modeling. Water Resources Research, 60(8), https://doi.org/10.1029/2023WR036900 2. Repository content The data repository has the following structure: data: This directory is designated for storing GCM data, predictors, shapefiles, and Random Forest (RF) model objects. data/ ├── others: This directory contains the downstream points of river segments in the DRYvER network for European reaches, along with the classifications of the six climate zones. ├── predictors: This directory contains the predictors required to run the Random Forest (RF) model. │ ├── HR: This directory contains only the monthly downscaled streamflow at high resolution for the five GCMs, two emission scenarios, and three periods, used to generate Figure 2. The HR predictors are generated using the compute_hr_predictors function. │ ├── LR: This directory contains the total runoff data at low resolution for the five GCMs under RCP8.5, used to generate Figure 2. The select_modify_lr_predictors function is executed to produce two dynamic low-resolution predictors. │ ├── statics: This directory contains the static predictors for European reaches. │ ├── min_five_GCMs: This directory contains the minimum monthly values in the rime series of dynamic predictors along with other predictors for three distinct sets 1) reanalysis-based 1985-2014, 2) GCM-based reference period, and 3) GCM-based 2080s period. ├── raw: This directory contains the raw monthly NetCDF files of downscaled streamflow from the GHM WaterGAP, organized by the five GCMs, two emission scenarios, and three time periods. ├── rf_models: This directory contains the trained sequential Random Forest (RF) model objects. └── shp: This directory contains the necessary shapefiles, including European country boundaries, the European river network, and European rivers with their corresponding climate zone codes. results: This directory serves as the storage location for the figures and streamflow intermittence predictions across the five GCMs, two emission scenarios, and three time periods. These represent the final outputs of the workflow. All the figures included in the paper are generated from these products using the workflow. results/ ├── figures: This directory contains the shapefiles required to create the figures presented in the paper. └── predictions: This directory contains the streamflow intermittence prediction status across the five GCMs, two emission scenarios, and three time periods. └── aoa: This directory contains the area of applicability (AOA) analysis for three distinct sets 1) reanalysis-based 1985-2014, 2) GCM-based reference period, and 3) GCM-based 2080s period. 3. Data format and projection The geometric network (lines) and gauging stations (points) datasets are distributed both in ESRI® file shapefile formats. Each shapefile consists of three main files (.dbf, .shp, .shx), and projection information is provided in an ASCII text file (.prj). The attribute table can be accessed as a stand-alone file in dBASE format (.dbf) which is included in the Shapefile format. These datasets are available electronically in compressed zip file format. To use the data files, the zip files must first be decompressed. All data layers are provided in geographic (latitude/longitude) projection, referenced to datum WGS84. In ESRI® software this projection is defined by the geographic coordinate system GCS_WGS_1984 and datum D_WGS_1984 (EPSG: 4326). 4. License and citations 4.1 License agreement This documentation and datasets are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License (CC-BY-4.0 License). For all regulations regarding license grants, copyright, redistribution restrictions, required attributions, disclaimer of warranty, indemnification, liability, waiver of damages, and a precise definition of licensed materials, please refer to the License Agreement (https://creativecommons.org/licenses/by/4.0/legalcode). For a human-readable summary of the license, please see https://creativecommons.org/licenses/by/4.0/. 4.2 Citations and acknowledgements. Citations and acknowledgements of this dataset should be made as follows: Abbasi, M., Messager, M. L., & Döll, P. (2024). Impact of climate change on changes in streamflow intermittence over European reaches. Zenodo. https://doi.org/10.5281/zenodo.14535981