Deman et al. (2025) Future changes in runoff over western and central Europe: disentangling the hydrological behavior of CMIP6 models

Identification

• Journal: Earth System Dynamics
• Year: 2025
• Date: 2025-09-04
• Authors: Juliette Deman, Julien Boé
• DOI: 10.5194/esd-16-1409-2025

Research Groups

• CECI, Université de Toulouse, CERFACS/CNRS/IRD, Toulouse, France

Short Summary

This study characterizes future runoff changes over western and central Europe using a large ensemble of CMIP6 models under a high-end emissions scenario, identifying diverse hydrological responses grouped into clusters. It disentangles inter-model uncertainties, highlighting the roles of large-scale circulation and the physiological effect of CO2, while finding the soil moisture-precipitation feedback important for the ensemble mean but not the inter-model spread.

Objective

• To characterize future changes in runoff over western and central Europe (WCE) in the late 21st century under a high-end emissions scenario (SSP5-8.5).
• To gain a better understanding of the mechanisms responsible for inter-model uncertainties in these hydrological changes.
• To assess the robustness of projected hydrological changes and evaluate how well models represent the hydrological cycle in the current climate.
• To investigate the causes of inter-model differences in hydrological changes, leveraging sensitivity experiments from various CMIP6 Model Intercomparison Projects (MIPs).

Study Configuration

• Spatial Scale: Western and Central Europe (WCE) region, as defined by IPCC climate reference regions.
• Temporal Scale:
  • Current climate: 1985–2014.
  • Future changes: 2081–2100, relative to 1995–2014.
  • LS3MIP simulations: 1980–2100.

Methodology and Data

• Models used:
  • An ensemble of 36 global climate models from the Coupled Model Intercomparison Project Phase 6 (CMIP6).
  • Sensitivity experiments from the Coupled Climate-Carbon Cycle Model Intercomparison Project (C4MIP) (hist-bgc and ssp585-bgc, with deactivated radiative effect of CO2).
  • Sensitivity experiments from the Land surface, Snow and Soil Moisture Model Intercomparison Project (LS3MIP) (amip-lfmip-rmLC and amip-lfmip-pdLC, with prescribed time-evolving or constant present-day soil moisture).
  • Hierarchical clustering algorithm (Ward's linkage with Euclidean distance) to group models based on seasonal relative changes in precipitation, evapotranspiration, and runoff.
  • Potential evapotranspiration calculated using the Penman–Monteith equation.
• Data sources:
  • ERA5-Land reanalysis for land surface variables.
  • Global Land Evaporation Amsterdam Model (GLEAM) version 4.1a for remote-sensing-based evapotranspiration and soil moisture.
  • Climatic Research Unit (CRU) Time Series (TS) version 4 for gridded monthly precipitation and potential evapotranspiration from weather station observations.

Main Results

• Runoff changes over WCE are highly uncertain, with models projecting significant increases, decreases, or no changes in annual runoff.
• A robust signal is a significant decrease in summer runoff projected by the vast majority of models (31 out of 36 models show a decrease, with 24 exceeding the variability threshold).
• Hierarchical clustering identified eight distinct hydrological response groups. Half of these clusters (C1, C7, C2, C4) project a substantial decrease in annual runoff (ranging from -10% to -25%), while the other half (C3, C6, C5, C8) show no change or a slight increase (ranging from -5% to +10%).
• Similar annual runoff changes can be driven by very different precipitation and evapotranspiration changes, even in sign. For example, some clusters show runoff decrease despite precipitation increase, due to a comparatively larger increase in evapotranspiration.
• Present-day evaluation against observational datasets does not allow for the rejection of any specific cluster's future projections, due to large observational uncertainties and variable model performance across different metrics. A generalized underestimation of climatological transpiration by CMIP6 models was noted.
• Large-scale circulation in winter, specifically sea level pressure changes over central and northern Europe, partly explains the inter-model spread in winter precipitation changes. Models like TaiESM1 and those in C8 (CCCma models) project strong winter precipitation increases linked to unusual negative sea level pressure anomalies.
• The soil moisture–precipitation feedback is important for shaping the multi-model ensemble mean changes in summer precipitation and evapotranspiration over WCE, leading to substantial drying. However, it has little influence on the inter-model spread in hydrological changes.
• The physiological effect of CO2 (possibly combined with anthropogenic aerosols) accounts for a substantial part of the large inter-model spread in summer hydrological changes over WCE.

Contributions

• Provides a comprehensive characterization of future runoff changes and their uncertainties over western and central Europe using a large CMIP6 ensemble.
• Introduces a novel hierarchical clustering approach to classify CMIP6 models based on their hydrological responses, revealing distinct "storylines" of future changes and highlighting that the multi-model mean can be unrepresentative.
• Systematically investigates the roles of key mechanisms (large-scale circulation, soil moisture–precipitation feedback, physiological CO2 effect) in driving inter-model spread, leveraging specialized CMIP6 MIP experiments (C4MIP, LS3MIP).
• Offers a detailed evaluation of CMIP6 models against multiple observational and reanalysis datasets for present-day climatology and trends, identifying model biases (e.g., transpiration underestimation) and highlighting observational uncertainties.
• Emphasizes the critical importance of two-way coupling between precipitation and evapotranspiration changes for accurate hydrological projections, a factor often simplified in impact studies.

Funding

• The World Climate Research Programme’s Working Group on Coupled Modelling for CMIP6.
• Climate modeling groups for producing and making their model output available.

Citation

@article{Deman2025Future,
  author = {Deman, Juliette and Boé, Julien},
  title = {Future changes in runoff over western and central Europe: disentangling the hydrological behavior of CMIP6 models},
  journal = {Earth System Dynamics},
  year = {2025},
  doi = {10.5194/esd-16-1409-2025},
  url = {https://doi.org/10.5194/esd-16-1409-2025}
}