Here we describe the evolution through winter of a layer of in situ supercooled water beneath the sea ice at a site close to the McMurdo Ice Shelf. From early winter (May), the temperature of the upper water column was below its surface freezing point, implying contact with an ice shelf at depth. By late winter the supercooled layer was c. 40 m deep with a maximum supercooling of c. 25 mK located 1–2 m below the sea ice-water interface. Transitory in situ supercooling events were also observed, one lasting c. 17 hours and reaching a depth of 70 m. In spite of these very low temperatures the isotopic composition of the water was relatively heavy, suggesting little glacial melt. Further, the water's temperature-salinity signature indicates contributions to water mass properties from High Salinity Shelf Water produced in areas of high sea ice production to the north of McMurdo Sound. Our measurements imply the existence of a heat sink beneath the supercooled layer that extracts heat from the ocean to thicken and cool this layer and contributes to the thickness of the sea ice cover. This sink is linked to the circulation pattern of the McMurdo Sound.

We investigated the physical conditions of the Spring pack ice environment at Terra Nova Bay to understand their influence on the structure and physiology of sympagic microalgae. Bio-optical methods were used to study the availability and spectral quality of solar radiation, both inside and underneath the ice cover. Pack ice thickness was around 2.5 m, with a temperature between −2 and −7°C. On average, only 1.4% of surface PAR penetrated to the bottom ice and less than 0.6% below platelet ice level. Surface UV-B radiation under the bottom ice was 0.2–0.4%. Biomass concentrations up to 2400 mg Chl a m−3, dominated by two species of diatoms (Entomoneis kjellmannii and Nitschia cf. stellata), showed marked spatial and temporal patterns. Maximum values were in the platelet ice during the first half of November, and in the bottom ice two weeks later. Strong shade adaptation characteristics emerged clearly and explained the relevant abundance of microalgae within the sea ice, with specific absorption coefficients (a*) as low as 0.005 m2 (mg Chl a)−1 and the photo-acclimation index (Ek) in the range of in situ irradiance. The biomass specific production values were low, around 0.12–0.13 mg C mg Chl a−1 h−1. The hypothesis suggesting bottom ice colonization by platelet ice microalgae is supported here.

The internal structure of ice cores from western Ross Sea pack ice floes showed considerable diversity. Snow-ice formation made a small, but significant contribution to ice growth. Frazil ice was common and its growth clearly occurred during both the pancake cycle and deformation events. Congelation ice was also common, in both its crystallographically aligned and non-aligned varieties. Platelet ice was found in only one core next to the Drygalski Ice Tongue, an observation adding to the increasing evidence that this unusual ice type occurs primarily in coastal pack ice near ice tongues and ice shelves. The diverse internal structure of the floes indicates that sea ice development in the Ross Sea is as complex as that in the Weddell Sea and more complex than in the Arctic. The mean ice thickness at the ice core sites varied between 0.71 m and 1.52 m. The thinnest ice generally occurred in the outer pack ice zone. Regardless of latitude, the ice thickness data are further evidence that Antarctic sea ice is thinner than Arctic sea ice.

The seasonal changes of the nutrient regime and the development of algal communities in snow-covered fast ice and the underlying platelet layer was investigated in the eastern Weddell Sea during autumn, winter, and spring 1995. In the upper sea ice, an autumnal diatom community became enclosed during subsequent ice growth in winter, declined, and was replaced by a flagellate dominated community in spring. In this layer, nitrate was completely exhausted at the end of spring, although nutrients had been partly regenerated in early spring. The progressive congelation of platelet ice contributed significantly to sea ice growth thus influencing algal inoculation of the sea ice bottom. Biomass, present in the uppermost section of the platelet layer, could be found in the sea ice bottom after this section congealed to solid ice. After incorporation, species composition changed from larger and chain-forming species to species of smaller cell size. Concurrently, net growth rate slowed down from 0.07 day−1 within the platelet layer to 0.03 day−1 within the sea ice. Despite a thick snow cover of more than 20 cm, maximum biomass yield was 210 mg chl a m−2 in the platelet layer and 40 mg chl a m−2 in the sea ice respectively, while 95% of the latter was located within consolidated platelet ice. Total fast ice biomass observed here is significantly lower than that observed in snow-free fast ice of the Ross Sea, but because snow cover of the southern Weddell Sea is representative of most fast ice areas in the Antarctic, the data presented here are of general value.