Rapid variability in sea surface salinity in northern high latitudes and implications for satellite-derived salinity measurements

Sea surface salinity (SSS) is an important indicator of the state of the polar oceans. SSS is directly related to both surface and deep-ocean processes, including input of freshwater at the surface, formation of sea ice and ocean melt water layers, identification of melt ponds, as well as modulation of deep-ocean convection and halocline formation. Changes in SSS can occur on a variety of time scales, including a pronounced annual cycle associated with large seasonal variation in freshwater input, sea ice formation and poleward transport. Here we are focusing on relatively unknown rapid variations in SSS at daily to sub-monthly periods and their potential contributions to ocean processes in polar regions. Due to unique and challenging atmospheric and oceanographic conditions at high latitudes, in situ observations are extremely sparse and most information comes from either numerical models or satellites. Our analysis uses salinity output produced by the HYCOM project and based on a global, eddy-resolving, high-resolution model run constrained to a variety of ocean data sets. Based on the HYCOM daily time series, a significant part of the total SSS variability is contributed by rapid fluctuations over many coastal regions, including the eastern coast of Greenland, Barents and Kara Seas, and Canadian Arctic coasts. Standard deviation of rapid SSS changes exceeds 0.1-0.2 psu in these regions and can be attributed to high precipitation, particular during winter, sea ice transport and other sources. One of the implications of the existence of the short scale fluctuations is the potential for their aliasing into the climate (typically monthly-averaged) values of satellite-derived salinity measurements, such as SMOS and Aquarius/SAC-D. Such missions have typically temporal sampling rates between 1 and 2 weeks. Our preliminary estimates of the implicit aliasing error suggest that removal of aliased signals, if possible, could reduce errors in monthly mean salinity estimates by more than 0.1 psu in many Arctic coastal regions