Southern High-Latitude Ocean Climate Drift in a Coupled Model

Abstract

Abstract Climate drift in coupled models affects the response of the coupled system to an external forcing. In most existing coupled models that employ flux adjustments, the southern high latitudes, in particular, are still affected by some climate drift. In the CSIRO coupled model, within 100 years following coupling, the Antarctic Circumpolar Current (ACC) intensifies by about 30 Sv (Sv ≡ 106 m3 s−1). This happens despite the use of flux adjustments. Many other model fields such as sea ice, surface albedo, and heat fluxes of the coupled system also experience drift from the precoupled spinup states. It is therefore important to study the processes that give rise to these drifts. The primary cause of drift in the CSIRO model is due to changes in the pattern of convection in the Southern Ocean relative to the spinup steady state. Upon coupling, the pattern of convection alters systematically regardless of surface boundary conditions. Consequently, overturning at shallow to intermediate depths (from the surface to about 2000 m) weakens, while that below these depths intensifies. The decline of overturning at shallow to intermediate depths leads to reduced surface temperatures because a lesser amount of warm subsurface water is mixed up into the colder surface mixed layer. The cooler surface temperature leads to an initial increase in sea ice, which is exacerbated by a significant albedo–temperature–sea ice feedback. The resulting increase in sea ice formation at the higher southern latitudes leads to increased brine rejection and a general increase in salinity throughout much of the high-latitude water column. This increase in salinity intensifies deep convection and bottom water formation, driving a stronger ACC. Several additional experiments are performed to trace various oceanic and ocean–atmosphere feedbacks that give the drift its character. It is demonstrated that the feedbacks significant to the drift in the present model are the positive albedo–temperature–sea ice feedback and a negative feedback between sea ice and overturning. The role of these two feedbacks in the interconnection between the drifts in various model fields is discussed.