Answers are grounded in this paper and cite their sources.
Two sets of simulations for 1993–2005 are carried out with a medium-resolution ocean and sea-ice model covering the North Pacific, Arctic and North Atlantic Oceans. The first set, using the same model parameters and three different atmospheric forcing datasets (DFS5.2, JRA55-do and ERA5), all show too fast melting of Arctic in spring and summer compared with the ice concentration based on satellite remote sensing. The simulation using ERA5 obtains the smallest ice concentration (largest deviation from satellite data) in summer, and the smallest ice thickness in both summer and winter, corresponding to the largest warm bias of surface air temperature over the Arctic sea-ice. In the second set of simulations using ERA5, changing either the snow conductivity (in W m−1 K−1, from the constant value of 0.31 to 0.15 during April –September and 0.5 during October–March) or the albedo of bare puddled ice (from 0.53 to 0.63) leads to an increase in ice concentration in summer, and ice thickness in both summer and winter. The simulation using ERA5 with both parameters altered is from October 1993 to March 2023, and obtains seasonal, interannual and long-term variations of ice area generally consistent with satellite data.
No takes yet. Share an insight, caveat, or question.
MacDermid et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: