Abstract:

Four high-resolution global ocean–sea ice reanalysis products were evaluated against observations in the Ross Sea and Amundsen Sea, Antarctica. The assessed quantities included climatological sea-ice concentration, sea-ice thickness, temperature, salinity, and mixed layer depth. The assessment revealed that the performance of the reanalysis products varied among variables and areas. Overall, the results from UR025.4 (University of Reading) were more consistent with observed sea-ice concentration and thickness, whereas the results from C-GLORS (Global Ocean Reanalysis System) had the largest discrepancies from observations. In the surface layer, during ice-freezing seasons, temperature and salinity data from C-GLORS agreed better with the WOA13 (World Ocean Atlas 2013) observations, whereas those from ORAP5 (Ocean ReAnalysis Pilot 5) deviated most from observations. Temperature and salinity data from the four reanalyses were generally consistent with the WOA13 values in deep layers below 700 m. In the subsurface layers, none of the reanalyses stood out from among the four products in hydrography simulations; the multi-model (reanalyses) mean results can reduce the error if the four reanalyses depart from observations in different directions.

Key words:

"> ocean-sea ice, reanalyses, sea ice, hydrography, mixed layer depth, Ross Sea, Amundsen Sea