Aspects of southern ocean transport and mixing

Understanding and quantifying the circulation of the oceans and the driving mechanisms thereof is an important step in developing models which can accurately predict future climate change. In particular, model studies have shown that the spatial variability of diapycnal diffusivity, which represents the rate at which deep water returns to shallower depths by means of turbulent diapycnal mixing, is a critical factor controlling the strength and structure of the circulation. Efforts are therefore ongoing to measure diffusivity as extensively as possible, but temporal variability in diffusivity has not been widely addressed. Results from three Southern Ocean studies are presented in this thesis. Firstly, a high resolution hydrographic survey carried out on the northern flank of the Kerguelen Plateau identifies a complex meandering current system carrying a total eastward volume transport of 174 ± 22 Sv, mostly associated with the blended Subtropical Front/Subantarctic Front. Significant water mass transformation across isopycnals is not required to balance the budgets in this region. Secondly, results are presented which cast doubt on the advisability of using density profiles acquired using Conductivity-Temperature-Depth instruments to estimate diapycnal diffusivity (an attractive proposition due to low cost and widespread data availability) in areas of weak stratification such as the Southern Ocean, because the noise characteristics of the data result in inaccurate diffusivity estimates. Finally, a method is developed for estimating diffusivity from profiles of velocity shear acquired by moored acoustic Doppler current profilers. An 18-month time series of diffusivity estimates is derived with a median of 3.3 × 10−4 m2 s−1 and a range of 0.5 × 10−4 m2 s−1 to 57 × 10−4 m2 s−1. There is no significant signal at annual or semiannual periods, but there is evidence of signals at periods of approximately fourteen days (likely due to the spring-neaps tidal cycle), and at periods of 3.8 and 2.6 days most likely due to topographically-trapped waves propagating around the local seamount. More widespread application of this method would allow for an assessment of natural climate variability in diapycnal diffusivity