Late winter snow and ice characteristics of first-year floes in the Bellingshausen and Amundsen Seas, Antarctica : results of investigations during R.V. Nathaniel B. Palmer cruise : NBP 94-5 in August and September 1993

Abstract

In August and September 1993, the R. V. Nathaniel B. Palmer operated for 37 days in the pack ice of the Bellingshausen and Amundsen seas supporting a sea ice research program, which included snow cover characterization, snow and ice thickness measurements and ice core analysis. The objective of the study was to improve our knowledge of the conditions and processes that contribute to first-year sea ice development and ice thickness variability in this region, and their impact on active microwave backscatter from the ice. The mean snow density was 247kg/m3. The principal snow type was composed of facetted crystals and depth hoar (48%), indicating strong temperature gradient metamorphism had been common. Ice layers (9%) indicated that snow metamorphism also included melting. Wet snow and slush layers (6%) at the base of the snow cover, with a mean salinity of 21.6‰, were due to seawater flooding of the snow/ice interface and brine wicking. Negative freeboard values, i.e., the water level was above the ice surface which was flooded with seawater, occurred at 18% of the drill holes on the snow and ice thickness transects. The majority of snow depth values (76%) ranged from 0.05-0.35m. The mean snow depth value was 0.23 ±0.16m. The majority of ice thickness values (62%) ranged from 0.3-0.8m. The mean ice thickness was 0.90±0.64m. Both the ice thickness and snow depth probability density functions have long tails of high values: the ice thickness distribution due to thickening by rafting and ridging, and the snow depth distribution due to snow drift formation on the flanks of ridge sails and other protuberances at the surface. In addition to the widespread flooding of the snow/ice interface, there was strong flooding potential as indicated by the large number (59%) of freeboard values in the range 0-0.05m. Three types of floe, X, Y and Z, are classified according to their coefficients of ice thickness variation. Their snow depth, ice thickness, draft and freeboard characteristics, and structural composition are described. Ice <0.5m thick was colder (mean temperature: -3.9°C) than ≥0.5m thick (mean temperature: -2.8°C), mainly because it had a thinner snow cover and was, therefore, less well insulated from the atmosphere than the thicker ice. Ice <0.5m thick was also more saline (mean salinity: -7.3‰) than ice ≥0.5m thick (mean salinity: 5.6‰), primarily because the thicker ice was older and had undergone greater brine loss. The temperature-ice thickness and the salinity-ice thickness relationships each have an abrupt change of slope at 0.5m. In ice ≥0.5m thick, mean core salinity was largely independent of ice thickness and generally lower than Arctic ice of similar age and thickness. The composite salinity profile for ice <0.5m thick was close to the C shape that is typical of Arctic first-year ice. Ice ≥0.5m thick, however, had a characteristic S shape profile with the highest values at the top due to snow ice formation, and a trend to lower values towards the bottom due to brine loss. Much of the brine loss may have been due to gravity drainage, since the majority of brine volumes were >5%, a value at which brine pockets coalesce and gravity drainage occurs. Ice crystal texture and δ¹⁸O values show that the major structural components of the floes were frazil ice (44%), congelation ice (25%) and snow ice (25%). The frazil ice originated primarily in the pancake cycle. Frazil and congelation ice layers both had an average thickness of 0.12m, indicating that neither reached a substantial thickness by thermodynamic growth alone; each thickened primarily by dynamic processes, i.e., deformation. The snow ice layers had a mean thickness of 0.2m and indicated that, by the end of winter, the thermodynamic development of the ice cover was dominated by seawater flooding of the snow/ice interface and snow ice formation. Snow ice layers were composed of 7-13% snow, which contributed to 2-4% of the total ice mass.