Diagnosing Energy Transfer in an Idealized Ocean-Atmosphere Model: A Frequency-Domain Approach

Abstract

Climate variability is an area of great interest. The ocean and atmosphere, two major components of the climate system, are inherently coupled, and there has been recent interest in deciphering whether oceanic and atmospheric variability is primarily due to intrinsic processes driven by nonlinear advection, due to forcing from the other fluid, or due to the inherently coupled nature of the ocean-atmosphere system. In this work, we investigate the oceanic, atmospheric, and coupled sources of variability by calculating a frequency-domain spectral transfer diagnostic, applied to the energy budget of a coupled ocean-atmosphere model.

Spectral transfers are particularly well-suited to study energy transfer, as they are able to pick out energy sources and sinks from specific terms in the energy budget, as a function of (in this study) frequency. While the majority of previous literature on spectral transfers is in the wavenumber domain, our work focuses instead on the frequency domain to investigate how the transfers of energy vary across timescales. We use the Quasi-Geostrophic Coupled Model (Q-GCM; citealt{QGCM_userguide}) - an idealized, turbulent, eddy-resolving, double-gyre ocean coupled to a channel atmosphere. The simplified nature of Q-GCM allows for the quantification of all terms in the energy budget, and additionally allows for the adjustment of coupling parameter strength. Spectral transfers are applied to the Q-GCM energy budget over timescales ranging from 2 days to 100 years, and reveal the relative magnitudes of energy sources and sinks in each fluid. We then attribute the observed behavior as driven by either intrinsic or forced dynamics.

In both the ocean and the atmosphere in the fully coupled model configuration, nonlinear advection of kinetic energy is found to be the dominant source of low-frequency variability, while potential energy advection is the largest source at high frequencies. In the ocean, we identify dynamically distinct regions that display strikingly different behavior: the western boundary current separation is found to be a large source of energy in the ocean at all frequencies, while the western boundary itself is a substantial sink of energy at nearly all frequencies. We argue that there is an important connection between these two regions, whereby energy generated in the current separation region is dissipated along the western boundary, with both regions characterized primarily by intrinsically-driven ocean variability.

Motivated by the finding that the ocean (at non-eddy-resolving scales) only affects the atmosphere at long timescales citep{Bjerknes1964,Gulev2013}, we are particularly interested in determining if the inclusion of eddies in our turbulent model will reveal a high-frequency ocean imprint in the atmosphere. We run a partially coupled and an atmosphere-only configuration of Q-GCM in order to start identifying the mechanisms responsible for certain behavior in each fluid. Overall, the energy transfer terms appear quite robust to changes in model coupling; however, there are some interesting differences. Ocean-driven variability is found at a timescale of 17 years that is seen throughout the atmosphere only when it is fully coupled to an ocean model. We also find that eddy interactions in the ocean appear to damp high-frequency atmospheric energetics by up to 10%. The exact mechanism responsible for this high-frequency behavior is still under investigation, but our results, in contrast to traditional results with a non-eddying ocean, indicate that ocean eddies can impact the atmosphere at daily timescales.