Diop et al. (2025) Climate change impacts on floods in West Africa: new insight from two large-scale hydrological models

Identification

• Journal: Natural hazards and earth system sciences
• Year: 2025
• Date: 2025-09-09
• Authors: Serigne Bassirou Diop, Job Ekolu, Yves Tramblay, Bastien Dieppois, Stefania Grimaldi, Ansoumana Bodian, Juliette Blanchet, Ponnambalam Rameshwaran, Peter Salamon, Benjamin Sultan
• DOI: 10.5194/nhess-25-3161-2025

Research Groups

• Laboratoire Leïdi “Dynamique des Territoires et Développement”, Université Gaston Berger, Saint-Louis, Senegal
• Centre for Agroecology, Water and Resilience, Coventry University, Coventry, UK
• Water for Production Department, Ministry of Water and Environment, Uganda
• Espace-Dev, Université Montpellier, IRD, Montpellier, France
• European Commission, Joint Research Centre (JRC), Ispra, Italy
• Université Grenoble Alpes, CNRS, IRD, Grenoble INP, IGE, Grenoble, France
• UK Centre for Ecology & Hydrology, Wallingford, UK

Short Summary

This study provides a large-scale analysis of flood frequency and magnitudes across West Africa using two hydrological models driven by CMIP6 climate models, projecting consistent increases in flood frequency and magnitude under future climate change scenarios.

Objective

• To assess the impacts of climate change on flood frequency and magnitudes across West Africa using two large-scale hydrological models driven by five bias-corrected CMIP6 climate models under two Shared Socioeconomic Pathways (SSP2-4.5 and SSP5-8.5).

Study Configuration

• Spatial Scale: West Africa (approximately 18° W to 25° E and 4° N to 25° N), covering 58 selected catchments. Hydrological models operated at 0.1° × 0.1° (HMF-WA) and 0.05° × 0.05° (LISFLOOD) spatial resolutions.
• Temporal Scale: Historical period (1950–2014 for model evaluation, 1985–2014 as reference), near-term future (2031–2060), and long-term future (2071–2100). Climate model simulations span 1950–2100.

Methodology and Data

• Models used:
  • Hydrological Models: HMF-WA (Hydrological Modeling Framework for West Africa), LISFLOOD (open-source LISFLOOD v4.1.3).
  • Climate Models: Five bias-corrected CMIP6 GCMs (MPI-ESM1-2-HR, MRI-ESM2-0, IPSL-CM6A-LR, UKESM1-0-LL, GFDL-ESM4) under SSP2-4.5 and SSP5-8.5 scenarios.
  • Statistical Models: Generalized Extreme Value (GEV) distribution (stationary and non-stationary forms), Generalized (penalized) Maximum Likelihood Estimation (GMLE) for parameter estimation.
  • Trend Detection: Modified Mann–Kendall test for GEV parameters.
  • Model Selection: Akaike Information Criterion (AIC), Deviance test (Likelihood Ratio test).
  • Significance Testing: Parametric bootstrapping, False-Discovery-Rate (FDR) procedure.
• Data sources:
  • Observed Streamflow: African Database of Hydrometric Indices (ADHI) daily streamflow data (1950–2018) for 58 catchments.
  • Reanalysis Data: ERA5 reanalysis (for LISFLOOD calibration and bias correction reference), EWEMBI dataset (for HMF-WA bias correction reference).
  • Climate Forcings: Daily rainfall and temperature outputs from five CMIP6 GCMs.
  • Geospatial Data: Terrain morphology, soil characteristics, land use, and water demand (for LISFLOOD).

Main Results

• The LISFLOOD hydrological model significantly outperformed HMF-WA in simulating extreme flows in West Africa, performing satisfactorily at 64% of stations compared to 24% for HMF-WA. HMF-WA showed consistent negative relative bias (median -46% to -52%), while LISFLOOD exhibited lower bias (median -14% to 7%).
• Both hydrological models consistently project an increase in flood frequency and magnitude across West Africa under both SSP2-4.5 and SSP5-8.5 scenarios.
• Flood magnitudes are projected to increase at 94% (96%) of stations for the 2-year (20-year) event in the near-term future (2031–2060) and at 88% (93%) of stations for the 2-year (20-year) event in the long-term future (2071–2100). Some locations are expected to experience increases exceeding 45%.
• Both models project upward trends in the GEV location and scale parameters across the region, indicating more frequent/severe floods and greater variability, with strong agreement between the models.
• The double-linear-trend GEV model (GEV3), which accounts for linear trends in location and scale parameters before and after a breakpoint, was identified as the best-fitting model for most stations (e.g., 66% to 79% for LISFLOOD under SSP2-4.5).
• Most significant flood trends were detected from the 1980s onward, with a predominant pattern of decreasing flood trends before a breakpoint transitioning to increasing trends after.

Contributions

• Provides the first large-scale analysis of flood frequency and magnitudes across West Africa using two large-scale hydrological models driven by bias-corrected CMIP6 GCMs under two Shared Socioeconomic Pathways.
• Utilizes an unprecedented set of 58 catchments and robust statistical methods to assess both the magnitude and field significance of future flood changes.
• Offers regional-scale insights into the evolving flood risks in West Africa, highlighting the urgent need for climate-resilient strategies.
• Demonstrates that climate forcing has a more significant influence than hydrological model representation itself in projecting future flood changes in this region, as both models show consistent projections despite differences in calibration and process representation.
• Identifies the onset of significant flood trends, revealing that shifts in extreme flood patterns began as early as the 1970s in several basins.

Funding

• PhD grant of Serigne Bassirou Diop: AFD/IRD project CECC.
• PhD grant of Job Ekolu: Centre for Agroecology Water and Resilience (CAWR) of Coventry University, UK.
• Yves Tramblay and Bastien Dieppois: PHC ALLIANCE grant.
• Juliette Blanchet: Agence Nationale de la Recherche – France 2030, PEPR TRACCS programme (grant number ANR-22-EXTR-0005).
• Ponnambalam Rameshwaran: Natural Environment Research Council, NC-International program (NE/X006247/1).

Citation

@article{Diop2025Climate,
  author = {Diop, Serigne Bassirou and Ekolu, Job and Tramblay, Yves and Dieppois, Bastien and Grimaldi, Stefania and Bodian, Ansoumana and Blanchet, Juliette and Rameshwaran, Ponnambalam and Salamon, Peter and Sultan, Benjamin},
  title = {Climate change impacts on floods in West Africa: new insight from two large-scale hydrological models},
  journal = {Natural hazards and earth system sciences},
  year = {2025},
  doi = {10.5194/nhess-25-3161-2025},
  url = {https://doi.org/10.5194/nhess-25-3161-2025}
}