Insufficient in-situ observations from the Antarctic marginal ice zone limit our understanding and description of relevant mechanical and thermodynamic processes that regulate the seasonal sea ice cycle. Here we present high-resolution thermal images of the ocean surface and complementary measurements of atmospheric variables that were acquired underway during one austral winter and one austral spring expedition in the Atlantic and Indian sectors of the Southern Ocean. Skin temperature data and ice cover images were used to estimate the partitioning of the heterogeneous surface and calculate the heat fluxes to compare with ERA5 reanalyses. The winter marginal ice zone was composed of different but relatively regularly distributed sea ice types with sharp thermal gradients. The surface-weighted skin temperature compared well with the reanalyses due to a compensation of errors between the sea ice fraction and the ice floe temperature. These uncertainties determine the dominant source of inaccuracy for heat fluxes as computed from observed variables. In spring, the sea ice type distribution was more irregular, with alternation of sea ice cover and large open water fractions even 400 km from the ice edge. The skin temperature distribution was more homogeneous and did not produce substantial uncertainties in heat fluxes. The discrepancies relative to reanalysis data are however larger than in winter and are attributed to biases in the atmospheric variables, with the downward solar radiation being the most critical.

The land-sea breeze is resonant with the inertial response of the ocean at the critical latitude of 30°N/S. 1D-vertical numerical experiments were undertaken to study the key drivers of enhanced diapycnal mixing in coastal upwelling systems driven by diurnal-inertial resonance near the critical latitude. The effect of the land boundary was implicitly included in the model through the ‘Craig approximation’ for first order cross-shore surface elevation gradient response. The model indicates that for shallow water depths (<~100m), bottom shear stresses must be accounted for in the formulation of the ‘Craig approximation’, as they serve to enhance the cross-shore surface elevation gradient response, while reducing shear and mixing at the thermocline. The model was able to predict the observed temperature and current features during an upwelling/mixing event in 60m water depth in St Helena Bay (~32.5°S, southern Benguela), indicating that the locally forced response to the land-sea breeze is a key driver of diapycnal mixing over the event. Alignment of the sub-inertial Ekman transport with the surface inertial oscillation produces shear spikes at the diurnal-inertial frequency, however their impact on mixing is secondary when compared with the diurnal-inertial resonance phenomenon. The amplitude of the diurnal anticlockwise rotary component of the wind stress represents a good diagnostic for the prediction of diapycnal mixing due to diurnal-inertial resonance. The local enhancement of this quantity over St Helena Bay provides strong evidence for the importance of the land-sea breeze in contributing to primary production in this region through nutrient enrichment of the surface layer.

Artificial sea ice growth is an emerging field that aims to assist in understanding the growth conditions and properties of sea ice. The Meta-Stable Zone Width (MSZW) of a supercooled solution is dependent on volume. Larger volumes have a higher probability of having a seed crystal being introduced to the system which halts supercooling and begins nucleation, hence, small extents of supercooling can be found in the ocean before freezing. Small MSZW’s are not often reflected in artificial sea ice set-up, thus many choose to seed to avoid excessive supercooling. This research focused on determining the MSZW of artificial sea ice grown in four volumes: 5 L, 30 L, 100 L and 370 L. The results showed a decreasing average MSZW as tank volume increased as a decreasing exponential relationship was found between volume and the MSZW.