Interannual and decadal variability of sea ice drift, concentration and thickness in the Weddell Sea

Abstract

Sea ice concentrations in the Weddell Sea are subject to regional climate variability. The magnitude and origin of local trends in the sea ice coverage were studied using the bootstrap algorithm sea ice concentration data from the NSIDC for 1979-2006. The impact of atmospheric forcing such as air temperature, wind speed, sea level pressure and cloud coverage, gained from NCEP/NCAR reanalysis, was assessed by analyzing correlation coefficients between the respective atmospheric component and the sea ice concentrations. In addition, the variability of sea ice drift was analyzed using the Polar Pathfinder sea ice motion vectors, and the correlation with sea ice concentration was tested after an assessment of the product s uncertainties. The connection to the variability of sea ice thicknesses was derived by model simulations from the Finite Element Sea ice-Ocean Model (FESOM). It was found that sea ice concentrations increased in the eastern and decreased in the western Weddell Sea, predominantly in the marginal sea ice zone. There, and in coastal regions, temperatures are strongly negatively correlated to sea ice concentrations, whereas in the central Weddell Sea, mostly a positive correlation was assessed, especially during winter. From analyses of the wind field it was found that the prevailing westerlies at the Antarctic Peninsula frequently show a shift towards the south. The enhanced southward winds are expected to bring warmer air into the western and central Weddell Sea and are further expected to redistribute the sea ice from the west into the central and eastern regions. This would increase the sea ice concentrations in the central Weddell Sea due to enhanced compactness, although temperatures are increasing. The correlation between sea ice concentrations and sea ice drift is only robust for the central Weddell Sea, where both parameters are mainly anti-correlated. Hence, strong sea ice drift is connected to lower sea ice concentrations and vice versa. This finding is consistent with the connection to the wind fields, since stronger northerly winds would reduce the north-eastward drift of sea ice in this region and enhance the sea ice concentrations. From model simulations with FESOM it was found that sea ice thicknesses predominantly show the same tendencies for changes as the simulated sea ice concentrations, which are basically decreasing in the central Weddell Sea and increasing in the eastern Weddell Sea. The overall changes in sea ice thickness and concentration result in an increase of the total sea ice volume by 1 % per decade in the simulations. A sensitivity study with a free drift model, forced by 10-m wind speeds and ocean currents from FESOM showed that the trends in the modelled sea ice drift are driven by the atmospheric fields, since ocean currents show barely any trends. Further it was revealed that sea ice drift velocities in the model are overestimated, especially in the zonal direction. Nevertheless, despite the overestimation, the mean sea ice export rate of 22 x 10^3 m^3/s is only about half of the export rates found in previous studies, which is certainly an effect of underestimated sea ice thicknesses in the western Weddell Sea.