How Eddy Heat Fluxes Shape the Resolution-Dependent Atmospheric Response to North Atlantic SST Anomalies

Atmospheric responses to extratropical sea-surface temperature (SST) anomalies are known to be sensitive to model resolution, yet the exact mechanisms controlling this sensitivity remain an open question. In the North Atlantic (NA), where strong air-sea coupling and stormtrack dynamics interact, resolving mesoscale frontal processes may be essential for correctly representing SST-driven atmospheric variability and its feedback onto large-scale circulation patterns such as the North Atlantic Oscillation (NAO).

We analyze a new ensemble of variable-resolution CAM6 simulations with prescribed SST anomalies. The atmospheric grid is globally 110 km and refined over the North Atlantic to 28 km and 14 km, allowing explicit representation of weather fronts and associated mesoscale circulations at the highest resolution. Imposed SST anomalies are derived by regressing the observed NAO index onto SSTs over 1958–2018, producing a cold–warm–cold tripole for a controlled comparison of NA SST feedbacks onto the NAO across resolutions.

We find that the NA SST tripole anomaly induces a positive feedback onto the NAO in the 14-km simulations, whereas this feedback is absent in the 28-km and 110-km configurations, which exhibit weaker and structurally different circulation responses. The atmospheric adjustment pathways also differ markedly across resolutions, with strongly contrasting responses in both the vertical and meridional eddy heat fluxes. In comparison to the lowest resolution, the intermediate resolution exhibits enhanced horizontal eddy heat flux responses, whereas the highest-resolution simulations respond to the positive Gulf Stream SST anomaly primarily through vertical eddy heat fluxes. While the mean states of the two higher-resolution simulations are in closer agreement, the SST-forced responses are more similar between the two lower-resolution simulations, suggesting that resolving the mesoscale might be particularly crucial for correctly representing ocean–atmosphere coupling.

As climate models move toward increasingly high resolution and computationally demanding coupled configurations, these results offer important guidance on the atmospheric resolution required to realistically represent ocean-atmosphere coupling and the climate response to SST perturbations, including those arising from changes in ocean circulation.