Trends and Variability of Southern Ocean Temperature and Salinity in Models and Observations

Abstract:

Over recent decades, the surface of the Southern Ocean has been cooling while the sea-ice cover around Antarctica has been modestly increasing. These changes are surprising given the rapid warming and sea-ice loss observed in the Arctic over the same period. We argue here that these Southern Ocean changes are best explained in terms of (i) a Southern Ocean ‘thermostat’ (driven by Ekman upwelling south of the Antarctic Circumpolar Current) that has acted to substantially delay surface warming compared to the rest of the world oceans, and (ii) separate processes giving rise to surface cooling and sea-ice expansion. What are these cooling processes? Several compelling hypothesis have been proposed, such as cooling driven by ozone depletion (through surface wind trends changing ocean circulation) and cooling associated with upper ocean freshening (possibly from enhanced melt water from Antarctica). Here we investigate a competing hypothesis that natural variability has played a substantial role. Indeed, unforced model simulations show substantial multi-decadal variability in the Southern Ocean that are as large or larger than the observed trends. To test this hypothesis, we characterize the modes and patterns of Southern Ocean temperature and salinity variability as simulated by CMIP5 models and compare these to the patterns of observed trends. We also consider changes within idealized, ocean-only simulations with the MITgcm forced by a simplified representation of greenhouse gas forcing and hydrologic cycle changes.