Zonally Asymmetric Response of Southern Ocean Heat Content to Wind, Heat and Freshwater Forcing at Multi-decadal Time Scales

Abstract

Since it was recognised that the Southern Ocean plays a crucial role in global climate, the region has generally been understood in a zonally symmetric framework. While this simplification has provided valuable insight, recent work suggests that significant zonal asymmetries exist in several key parameters such as SST, mixed layer depth and air-sea CO2 flux. This is true both for the mean state and for changes at multiple time scales. Of particular interest here are changes in ocean heat content (ΔOHC) over the past three decades. Using an eddy-permitting ocean model forced by ERA5 reanalysis, we find significant asymmetries in ΔOHC both within and north of the ACC, which is robustly reflected in a suite of hindcast and reanalysis models, as well as observation based temperature reconstructions. In our model, asymmetry stems largely from a southward displacement of ΔOHC in the Indian basin, where warming occurs primarily within ACC, as opposed to north of it in the Atlantic and Pacific. However, significant asymmetries are also found within the sea ice zone south of 60o S, where the Ross Sea warms to a much greater degree than other basins. In order to better understand the sources of this asymmetric warming, we run several model experiments which decompose the total OHC into components originating from wind, heat and freshwater flux changes. We find roughly equal contributions from wind and surface heat flux north of the ACC, with asymmetric changes in the westerlies driving anomalous convergence of heat. Within and south of the ACC all three forcings play an important role, although this depends strongly on the basin. Overall, we conclude that much of the asymmetries in ΔOHC originate from asymmetries in the surface flux changes, with an important secondary role played by variability in the mean state. These findings have two important implications. First, studies which only consider zonally averaged quantities will likely mask significant variability, and therefore miss important regional and local processes. Second, the impact of multi-decadal climate variability on the Southern Ocean is not manifested in a zonally symmetric fashion, which may have important implications for future changes.