Diagnostic Dashboard

IFS-NEMO-ER and IFS-FESOM2-SR effectively reproduce observed climate variability, eliminating the excessive tropical noise found in CMIP6, whereas ICON-ESM-ER systematically suppresses extratropical storm track dynamics while overestimating land-surface volatility.

While IFS-based high-resolution models significantly outperform CMIP6 and ICON-ESM-ER in reproducing global circulation, kilometric-scale resolution fails to rectify persistent structural biases in marine stratocumulus clouds and tropical precipitation.

High-resolution coupled models fail to capture the observed 1980–2014 intensification of the Pacific Walker Circulation and Southern Ocean cooling, exhibiting systematic trend biases indistinguishable from standard-resolution CMIP6 ensembles.

IFS-FESOM2-SR demonstrates exceptional skill in reproducing global thermal and radiative cycles, contrasting with severe cold biases in IFS-NEMO-ER and an overly aggressive hydrological cycle in ICON-ESM-ER and HadGEM3-GC5.

IFS-based models demonstrate superior fidelity in extratropical annular modes and stratospheric QBO dynamics, while tropical variability exhibits a massive spread ranging from the critically damped ICON-ESM-ER to the hyper-active HadGEM3-GC5.

IFS-FESOM2-SR closely matches observational baselines, whereas IFS-NEMO-ER exhibits a self-reinforcing cold/dry bias and ICON-ESM-ER simulates an accelerated hydrological cycle driven by excessive solar input.

HadGEM3-GC5 achieves superior fidelity to the observed mean thermal state compared to the cold-biased IFS and ICON variants, though high-resolution coupling introduces intense, localized ocean-driven trend artifacts in the Antarctic and North Atlantic that diverge from CMIP6 baselines.

IFS-NEMO-ER significantly outperforms CMIP6 and other high-resolution configurations by resolving the Double ITCZ bias, whereas ICON-ESM-ER exhibits a systematic global wet bias and excessive hydrological intensity.

Despite eddy-rich resolutions, models exhibit a 1.5 K spread in global mean SST dominated by ocean formulation choices, with persistent biases in upwelling zones and boundary currents that mirror lower-resolution systematic errors.

HadGEM3-GC5 captures Arctic sea ice decline with high skill, whereas IFS variants demonstrate that ocean model formulation (NEMO vs. FESOM) drives extreme, opposing biases in Antarctic sea ice extent and volume.

Despite eddy-rich resolution, models exhibit significant initialization drifts and persistent systematic errors, including universal Arctic salinification, warm biases in upwelling zones, and divergent global thermal energy balances.