Ice - ocean - atmosphere interactions in the Southern Ocean and implications for phytoplankton phenology

The annual advance and retreat of sea ice in the Southern Ocean is recognised as one of the largest seasonal events on Earth. Such considerable physical changes have profound effects on the vertical structure of the water column, and hence controls the availability of both light and nutrients to phytoplankton. This means that in the region seasonally covered by sea ice (the SSIZ), the timing of the growth and decline (phenology) of phytoplankton is determined to a large degree by the dynamic interactions between ice, ocean and atmosphere. However, this region is simultaneously one of the most poorly observed in the global ocean, and one of the most complex. This has led to significant gaps in our understanding of how sea ice modulates the exchanges of heat and momentum between atmosphere and ocean, as well as the implications this has for phytoplankton phenology in the SSIZ. This study seeks to address these gaps by combining both model and observationallybased methods. The lack of observational data are directly tackled through an analysis of BGC-Argo float data sampling under ice. Such data reveal high growth rates in the presence of near full ice cover and deep mixed layers, conditions previously thought to prevent growth. These results suggest a revision of our current understanding of the drivers of under ice phytoplankton phenology, which should take into account the unique character of Antarctic sea ice and its effect on the under ice light environment. In addition, results obtained from several numerical process studies indicates that phytoplankton may have a higher affinity for low light conditions than previously thought. From a modelling perspective, an analysis and intercomparison of 11 Earth System Models (ESMs) and their representation of vertical mixing and phenology is presented. This revealed that misrepresentations in phenology where driven by model biases in sea ice cover and vertical mixing. That is, only models with either too much or too little ice cover were able to simulate phenology close to observations. Furthermore, a strong correlation between the location of the ice edge and the extent of vertical mixing suggested that ESMs overly dampen ocean-atmosphere fluxes as mediated by sea ice. This led to the development of a regional ocean-sea ice model of the Atlantic sector of the Southern Ocean, from which experiments enhancing both heat and momentum fluxes could be conducted. It was found that the model responded more uniformly to enhanced heat flux, generally deepening the mixed layer closer to observations in winter. On the other hand, the effects of enhanced momentum flux (implemented by increased air-ice drag) where more complex and spatially heterogeneous, with contrasting responses depending on the initial vertical density structure of the water column. Overall, the argument is made that the unique features of Antarctic sea ice should be included in models if we are to improve the representation of the SSIZ mixed layer, and hence phenolog