Published 2025 | Version 1.0

Sea level rise inundation spatial modelling workflow and dataset for Europe

  • 1. ROR icon International Institute for Applied Systems Analysis

Contributors

Work package leader:

  • 1. International Institute for Applied Systems Analysis

Description

Overview

This dataset contains reproducible R scripts and output raster maps modeling sea-level rise (SLR) inundation across European coastal countries. Developed for Task 2.4 of the ACCREU Project, this custom "bathtub" modeling approach integrates regional DIVACoast relative sea-level-rise (RSLR) projections (Bacca et al., 2026) with a high-resolution Digital Elevation Model (MERITDEM) (Yamazaki et al., 2017) to generate binary and fractional flood multi-band GeoTIFF stacks of coastal flood extent for European countries, plus per-band area statistics and the full processing scripts.

These outputs are designed for downstream impact assessments on biodiversity and ecosystem services. To comply with data size limits and licensing, original input files are not included in this deposit; only the derived flood maps, summary statistics, and processing scripts are provided.

Input Data

To comply with data size limits and licensing, original input files are NOT included in this repository. Users must download the external datasets from the sources below and place them in the input_data/ directory as specified in the R script's CONFIG block.

  1. Digital Elevation Model
    • Source: MERITDEM (https://hydro.iis.u-tokyo.ac.jp/~yamadai/MERIT_DEM/)
    • Format: GeoTIFF
    • Resolution:  1000 m, aggregated from original ~90m
    • License: CC BY-NC 4.0
    • Citation: Yamazaki, D., Ikeshima, D., Tawatari, R., Yamaguchi, T., O’Loughlin, F., Neal, J. C., Sampson, C. C., Kanae, S., & Bates, P. D. (2017). A high accuracy map of global terrain elevations. Geophysical Research Letters, 44, 5844–5853. DOI: 10.1002/2017GL072874
  2. Administrative Boundaries
    • Source: EUROSTAT GISCO (https://ec.europa.eu/eurostat/web/gisco/geodata/administrative-units)
    • Format: Shapefile
    • Resolution: 1 m scale
    • License: CC BY 4.0
    • Citation: © EuroGeographics for the administrative boundaries
  3. Permanent Water Mask
    • Source: JRC Global Surface Water 2021 (https://global-surface-water.appspot.com/download)
    • Format: GeoTIFF
    • Resolution: 1000 m
    • License: CC BY 4.0
    • Citation: Pekel, J.F., Cottam, A., Gorelick, N., Belward, A.S. 2016. High-resolution mapping of global surface water and its long-term changes. Nature 540, 418-422 (2016). DOI: 10.1038/nature20584
  4. Regional Sea Level Rise Projections
    • Source: Developed internally for the ACCREU Project (Task 2.4), utilizing DIVA (Dynamic Interactive Vulnerability Assessment) methodology at the regional/country scale.
    • Format: CSV
    • Content: RSLR in meters per country, scenario (e.g., RCP126, RCP245, RCP370), and year (2015-2100).
    • License: CC BY 4.0
    • Citation: Bacca, S., Haas, J., Lincke, D., & Hinkel, J. (2026). DIVACoast model results: sea-level rise impacts and adaptation response under different adaptation paradigms [Dataset]. Zenodo. DOI: 10.5281/zenodo.20545890

Methods

The workflow executes a spatially explicit bathtub model at the country level, which is then aggregated to a pan-European extent:

  1. Data Preparation: Inputs are loaded. The JRC water mask is inverted so that 1 = Land and 0 = Water.
  2. Country-Level Bathtub Modeling: For each unique coastal country, scenario, and year:
    • The DEM and land mask are cropped to the country geometry plus a defined coastal distance (default: 10 km) to optimize memory and limit inland artifacts.
    • Potential flooding is identified where DEM elevation ≤ the projected RSLR value.
    • A distance-to-coast constraint is applied.
    • Scientific Note: The permanent water mask is applied during this step to prevent the model from erroneously calculating "new" flooding over existing oceans/lakes.
  3. Spatial Aggregation (Fractions): The 1 km binary maps are aggregated to 10 km cells. The pixel value represents the fraction of valid land pixels within that 10 km cell that are flooded (sea/NoData are strictly excluded from the denominator to prevent dilution).
  4. Years are stored as bands within multi-layer GeoTIFF stacks (one stack per country/scenario).
  5. Country stacks are then merged into pan-European ("global") stacks. Landlocked countries are explicitly coded as 0 (never flooded).
The workflow executes a modified, spatially explicit bathtub model (Fitchett et al., 2016; Williams and Lück-Vogel, 2020; Yunus et al., 2016) at the country level and then aggregates to a pan-European extent. Unlike standard bathtub models that apply elevation thresholds across the entire landscape, our approach includes key modifications to enhance accuracy while maintaining computational efficiency:
  1. Land-sea boundaries (i.e., coastlines) are identified directly from the DEM's NoData values, rather than applying simple elevation thresholds blindly.
  2. Flooding is restricted to a 10 km inland proximity to the coast. This accounts for the DEM’s vertical uncertainty and prevents the misclassification of inland depressions as flooded.
  3. Existing water bodies (lakes, rivers) that are covered with water >90% of the time (using the JRC Global Surface Water dataset) are masked out. This prevents double-counting areas that are already underwater.
  4. Regional (country-level) estimates of relative sea level rise from the DIVA model are incorporated instead of uniform global values.
  5. Checks are implemented for landlocked countries (which are explicitly coded as 0 to maintain spatial continuity) and proper geographic alignment.
The 1 km binary maps are aggregated to 10 km cells. The pixel value represents the fraction of valid land pixels within that 10 km cell that are flooded (sea/NoData are strictly excluded from the denominator to prevent dilution).
 
Years are stored as bands within multi-layer GeoTIFF stacks (one stack per country/scenario). Country stacks are then merged into pan-European ("global") stacks.

Dataset Description

Note: "Global" in the filenames refers to the pan-European study area, not worldwide. All spatial rasters share the same Coordinate Reference System: ETRS89 / LAEA Europe (EPSG:3035), measured in metres. Within all multi-layer GeoTIFFs, bands are named chronologically by year, e.g., `Y2015`, `Y2030`, `Y2050` for years 2015, 2020, 2030, 2040, 2050, 2060, 2070, 2080, 2090, 2100. Scenario refers to the DIVACoast RCP scenarios for RSLR (e.g., RCP126, RCP245, RCP370).

Folder 1: 01_country_binary_res1000m_buffer10000m
Description: Multi-band GeoTIFFs of binary flood maps at the country level.
Naming Convention: <Country>_<Scenario>_binary_stack.tif (e.g., Netherlands_126_binary_stack.tif), where res1000m is the native 1 km^2 spatial resolution of the binary flood model, and buffer10000m (10 km^2) is the raster buffer extent for each country.
Resolution: 1000 m x 1000 m
Pixel Values: 1 = Flooded land; 0 = Dry land; NA = NoData / Permanent water / Outside borders.
 
Folder 2: 01_country_fraction_res10000m_from1000m
Description: Multi-band GeoTIFFs of fractional flood maps at the country level aggregated to 10000 x 10000 m.
Naming Convention: <Country>_<Scenario>_fraction_stack.tif
Resolution: 10000 m x 10000 m
Pixel Values: 0.0 to 1.0 (fraction of valid land flooded); NA = NoData / Permanent water.
 
Folder 3:. 02_global_binary_res1000m_buffer10000m
Description: Pan-European multi-band binary flood maps (merged country outputs).
Naming Convention: global_<Scenario>_binary_stack.tif (e.g., global_126_binary_stack.tif)
Resolution: 1000 m x 1000 m
Pixel Values: 1 = Flooded land; 0 = Dry land (including explicitly coded landlocked countries); NA = NoData / Permanent water.
 
Folder 4: 02_global_fraction_res10000m_from1000m
Description: Pan-European multi-band fractional flood maps (merged country outputs) aggregated to 10000 x 10000 m.
Naming Convention: global_<Scenario>_fraction_stack.tif
Resolution: 10000 m x 10000 m
Pixel Values: 0.0 to 1.0 (fraction of land flooded); NA = NoData / Permanent water.
 
Folder 5: slr_workflow_v1
Description: The R code required to reproduce the dataset and generate visualizations.
Content: slr_workflow_v1.R (Main processing script)

Known Limitations

The methodology has several recognized limitations that should be considered when interpreting results for policy applications:

  1. The bathtub approach assumes static inundation without accounting for coastal defense structures, hydrodynamics, wave action, or storm surges.
  2. The 1 km resolution DEM generalizes small-scale topographic features, which may affect local inundation accuracy.
  3. The 10 km coastal buffer may exclude some potential flood areas in regions with very low relief that extend further inland.
  4. MERITDEM heights are referenced to the EGM96 geoid, while DIVA SLR projections are relative to local mean sea level. Residual offsets (≤ 1 m) are not explicitly corrected in this version (vertical datum mismatch).

Related Works

For the broader methodology and context, please refer to:

Funding Acknowledgement

This work was supported by the Assessing Climate Change Risk in EUrope (ACCREU) project, funded by the European Commission under the Horizon Europe programme.
Project website: ACCREU Website

Files

01_country_binary_res1000m_buffer10000m.zip

Files (3.8 MB)

Name Size
md5:e114b10179374b912544ffd5b84d8972
1.8 MB Preview Download
md5:b72d2a3a31e6496f741500905c993483
439.1 kB Preview Download
md5:c3267482a69cd9f237abe0d6332b9e84
1.3 MB Preview Download
md5:d1ba08901a857f4d6d7ae9a085f982d9
238.1 kB Preview Download
md5:ae0fd8be40ee047e0e7c8a7d4507b648
4.7 kB Preview Download

Additional details

Funding

European Commission
ACCREU - Assessing Climate Change Risk in EUrope 101081358

References

  • Bacca, S., Haas, J., Lincke, D., & Hinkel, J. (2026). DIVACoast model results: sea-level rise impacts and adaptation response under different adaptation paradigms [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.20545890
  • Palazzo, A. , Nakhavali, A. , Gerber, G. , Arbelaez Gaviria, J., Jung, M. , Drews, M., Meuriot, O.G.M., Martin, J.S., Bacca, S., Hinkel, J., & Lincke, D. (2025). ACCREU Report: Impacts on ecosystems & biodiversity. ACCREU (Assessing Climate Change Risk in EUrope). [https://www.accreu.eu/wp-content/uploads/2026/04/ACCREU_D2.4-with-annexes_revised-April-2026.pdf]
  • Fitchett, J.M., Grant, B., Hoogendoorn, G., 2016. Climate change threats to two low-lying South African coastal towns: Risks and perceptions. S. Afr. J. Sci. 112, 9. DOI: 10.17159/sajs.2016/20150262
  • Williams, L.L., Lück-Vogel, M., 2020. Comparative assessment of the GIS based bathtub model and an enhanced bathtub model for coastal inundation. J Coast Conserv 24, 23. DOI: 10.1007/s11852-020-00735-x
  • Yunus, A., Avtar, R., Kraines, S., Yamamuro, M., Lindberg, F., Grimmond, C., 2016. Uncertainties in Tidally Adjusted Estimates of Sea Level Rise Flooding (Bathtub Model) for the Greater London. Remote Sensing 8, 366. DOI: 10.3390/rs8050366
  • Yamazaki, D., Ikeshima, D., Tawatari, R., Yamaguchi, T., O'Loughlin, F., Neal, J. C., Sampson, C. C., Kanae, S., & Bates, P. D. (2017). A high accuracy map of global terrain elevations. Geophysical Research Letters, 44, 5844–5853. DOI: 10.1002/2017GL072874
  • Pekel, J.F., Cottam, A., Gorelick, N., Belward, A.S. 2016. High-resolution mapping of global surface water and its long-term changes. Nature 540, 418-422 (2016). DOI: 10.1038/nature20584