Presenter: Dafydd Stephenson (University Corporation for Atmospheric Research)

Description:
Predictability of large-scale ocean metrics (signal) is limited by interference (noise) from different sources such as chaotic fluctuations within the ocean and atmosphere, which themselves have extremely limited predictability and can effectively be treated as stochastic. Here we present new results using adjoint methods to decompose the relative importance of such sources and the nature of the ocean’s response to them. We first determine whether ocean heat content variations are a purely passive (advective-diffusive) response to stochastic surface fluxes, or whether associated circulation changes have a significant role in their manifestation. We show that surface-forced North Atlantic full-depth heat content variations are primarily due to the active redistribution of existing heat by dynamical ocean circulation changes. We then compare the large-scale impact of these surface fluxes with that of the internal mesoscale eddy field for a range of metrics such as heat content and volume transport. We show that, while month-averaged AMOC variations are primarily wind-driven and decade-averaged AMOC variations are the oceanic response to surface buoyancy flux anomalies, year-averaged AMOC fluctuations at 26°N are up to 60% driven by mesoscale eddies. Conversely, in the subpolar region, AMOC variations at all considered timescales are predominantly driven by atmospheric forcing, explained by an overall lower sensitivity of the region to small-scale internal perturbations. These results motivate the further decomposition of the forcing and response into their preferred patterns, for example to evaluate compatibility between atmospheric modes of variability (such as the NAO) and optimal mechanisms of ocean variability excitation.