Graff et al. (2025) Sensitivity of winter Arctic amplification in NorESM2

Identification

• Journal: Earth System Dynamics
• Year: 2025
• Date: 2025-10-10
• Authors: Lise Seland Graff, Jerry Tjiputra, Ada Gjermundsen, Andreas Born, Jens Boldingh Debernard, Heiko Goelzer, Yanchun He, Petra M. Langebroek, Aleksi Nummelin, Dirk Olivié, Øyvind Seland, Trude Storelvmo, Mats Bentsen, Chuncheng Guo, Andrea Rosendahl, Dandan Tao, Thomas Toniazzo, Camille Li, Stephen Outten, Michael Schulz
• DOI: 10.5194/esd-16-1671-2025

Research Groups

• Norwegian Meteorological Institute, Oslo, Norway
• NORCE Research AS, Bergen, Norway
• Department of Earth Science, University of Bergen, Bergen, Norway
• Bjerknes Centre for Climate Research, Bergen, Norway
• Nansen Environmental and Remote Sensing Center, Bergen, Norway
• Finnish Meteorological Institute, Helsinki, Finland
• Department of Geosciences, University of Oslo, Oslo, Norway
• National Centre for Climate Research, Danish Meteorological Institute, Copenhagen, Denmark
• Geophysical Institute, University of Bergen, Bergen, Norway

Short Summary

This study investigates the drivers of uncertainty in Arctic climate change projections by performing sensitivity experiments with NorESM2, modifying five key processes. The results show that these modifications consistently enhance future Arctic warming, with the amplitude of additional winter warming varying by approximately 9 K, primarily driven by an enhanced greenhouse effect.

Objective

• To better understand the drivers of uncertainty in Arctic climate change projections by systematically investigating how modifications to key Arctic climate processes (mixed-phase clouds, ocean eddies, Greenland ice sheet coupling, snow on sea ice, and ozone chemistry) in NorESM2 affect future Arctic amplification, particularly winter surface warming, and to identify the underlying mechanisms and associated uncertainties.

Study Configuration

• Spatial Scale: Arctic (poleward of 66° N), Northern Hemisphere (NH) extratropics (north of 20° N), Greenland, Barents-Kara seas, Beaufort-Chukchi seas, Eurasian Basin, Canadian Basin. Model resolution: NorESM2-MM (nominal 1° × 1° horizontal resolution in atmosphere, land, ocean; 32 hybrid-pressure layers in atmosphere; 53 layers in ocean).
• Temporal Scale: Historical experiments (1850–2014), future scenario (ssp585) experiments (2014–2100, some extended to 2300). Analysis periods focus on future response (2071–2100 relative to 1985–2014) and contemporary trends (1979–2017).

Methodology and Data

• Models used: Norwegian Earth System Model version 2 (NorESM2-MM), based on CESM2, with components including BLOM, iHAMOCC, CAM6-Nor, CLM5, and CICE5. Five sets of sensitivity experiments (NEEMS) were conducted with modifications to: (1) mixed-phase clouds, (2) upper ocean eddy processes, (3) Greenland ice sheet coupling (using CISM), (4) snow on sea ice processes (reduced thermal conductivity from 0.30 W m⁻¹ K⁻¹ to 0.15 W m⁻¹ K⁻¹), and (5) interactive ozone chemistry (using TS1 scheme). Additional experiments included CMIP6 ScenarioMIP (ssp126, ssp245, ssp370) and experiments without anthropogenic aerosols (hist-piAerOxid, ssp585-piAerOxid). Analysis involved linear decomposition of surface temperature, Empirical Orthogonal Function (EOF) analysis, and emergent constraints.
• Data sources: CMIP6 baseline and scenario experiments (Earth System Grid Federation - ESGF), NEEMS and piAerOxid experiments (Norwegian Infrastructure for Research Data - NIRD Research Data Archive). Observational data for sea ice area and volume from OSI SAF, PIOMAS, CryoSat-2, and Cavalieri and Parkinson (2012) for sea ice trends. Satellite observations for cloud phase (Cesana et al., 2015).

Main Results

• All five model modifications consistently lead to enhanced future Arctic warming compared to the baseline NorESM2-MM CMIP6 experiments.
• The amplitude of additional winter Arctic warming varies significantly among experiments, reaching approximately 8.99 K between the strongest (cloud) and weakest (ozone) warming experiments by the end of the 21st century, a range comparable to the ScenarioMIP uncertainty (10.6 K).
• Winter warming is primarily driven by an enhanced greenhouse effect due to increased cloud cover, near-surface humidity, and resulting increased downwelling longwave radiation (LWCS contribution of 9.43 K). Contributions also come from changes in cloud radiative effect (CRE, 1.88 K) and ocean heat storage/transport (HSTORE, 2.87 K), partially offset by turbulent heat fluxes (HFLUX, -2.82 K).
• The most pronounced warming and ensemble spread occur in sea ice retreat regions, particularly on the Atlantic side (Greenland–Barents seas).
• The cloud experiment shows the strongest additional surface warming (3.78 K Arctic average) and fastest sea ice loss. The snow experiment, with reduced snow thermal conductivity, yields the most realistic historical sea ice, leading to weaker future sea ice loss than the baseline but earlier ice-free summer conditions.
• An emergent constraint reveals a strong correlation (correlation coefficient of -0.78 for sensitivity experiments, -0.79 for CMIP6 models) between future Arctic winter surface temperature changes and simulated historical trends in Northern Hemisphere winter sea ice area.
• Applying this emergent constraint, the observationally constrained estimate for future Arctic winter warming is 14.4 ± 4.0 K, which has slightly smaller uncertainty than the unconstrained CMIP6 multi-model mean of 14.2 ± 4.2 K.
• Dynamically coupling the Greenland ice sheet and interactive ozone chemistry have relatively weak impacts on Arctic amplification compared to other modifications.

Contributions

• Presents the first coordinated ensemble of sensitivity simulations using a single CMIP6 Earth System Model (NorESM2) to systematically quantify the impact of five key Arctic climate processes on future Arctic amplification.
• Quantifies the substantial uncertainty (up to ~9 K in winter warming) introduced by different representations of these processes within a single model framework, highlighting their importance relative to inter-scenario uncertainty.
• Identifies the dominant physical mechanisms (enhanced greenhouse effect from clouds and humidity, ocean heat release) driving the additional warming and its spatial variability, particularly in sea ice retreat regions.
• Demonstrates a robust emergent constraint linking future Arctic winter warming to contemporary Northern Hemisphere winter sea ice area trends, reinforcing the importance of accurate sea ice representation for future projections and providing a constrained estimate of 14.4 ± 4.0 K.
• Provides detailed documentation and data for the NorESM2 Ensemble Exploring Model Sensitivity (NEEMS) experiments, making them available for further research.

Funding

• Research Council of Norway (RCN) projects: KeyCLIM (project no. 295046), INES2 (project no. 270061), BASIC (project no. 325440), GREASE (324639), TopArctic (314826), NAVIGATE (352142).
• European Union’s Horizon 2020 research and innovation program: PolarRES (grant no. 101003590).
• European Research Council grants: no. 101081661, no. 101045273.
• Sigma2 – the National Infrastructure for High Performance Computing and Data Storage in Norway (projects NN9252K, NS9034K, NN8006K, NS2345K, NS9560K, NS9252K, and NS8006K).
• ESA project SMOS and CryoSat-2 Sea Ice Data Product Processing and Dissemination Service.

Citation

@article{Graff2025Sensitivity,
  author = {Graff, Lise Seland and Tjiputra, Jerry and Gjermundsen, Ada and Born, Andreas and Debernard, Jens Boldingh and Goelzer, Heiko and He, Yanchun and Langebroek, Petra M. and Nummelin, Aleksi and Olivié, Dirk and Seland, Øyvind and Storelvmo, Trude and Bentsen, Mats and Guo, Chuncheng and Rosendahl, Andrea and Tao, Dandan and Toniazzo, Thomas and Li, Camille and Outten, Stephen and Schulz, Michael},
  title = {Sensitivity of winter Arctic amplification in NorESM2},
  journal = {Earth System Dynamics},
  year = {2025},
  doi = {10.5194/esd-16-1671-2025},
  url = {https://doi.org/10.5194/esd-16-1671-2025}
}