Sensitivity of Arctic Warming to the Sea Surface Temperature Distribution over Sea-ice Melted Region in the AGCM Experiments

As Arctic Ocean becomes warmer in recent decades, the regions fractionally covered by sea-ice have been expanding and the proper estimation of atmospheric response to the reduced sea-ice cover is now an important issue in both seasonal forecast and future climate prediction. Here, we show that the atmospheric response related to sea-ice variability especially during a cold season (October-March) not only depends on the sea-ice fraction but also is sensitive to the in-situ sea surface temperature (SST). To examine the dependences of atmospheric response to these two distinctive surface conditions, we have performed several atmospheric general circulation model (AGCM) simulations; two different sea-ice conditions are forced, one with the averaged sea-ice condition for 1982-2000 and the other with averaged for 2006-2010, and diverse SST conditions are specified over major sea-ice changing regions. From an analysis on the model simulations, we find that the amplitude of atmospheric warming over the sea-ice melted region depends on in-situ SST conditions. Large discrepancies in the surface heat fluxes among the model simulations are keys for the sensitivity: longwave radiation during early winter (October-December), sensible and latent heat fluxes during whole cold season. Longwave radiation and sensible heat fluxes warm the air at near surface, and latent heat flux moistens the air. Changes in these near surface variables also contribute to vertical expansion of warming (and increasing moisture contents) and consequent heating by condensation that decreases the static stability in the lower troposphere. Since SST observations over SMR are very rare and usually the estimation relies on the statistical relation, the present model result suggests that the uncertainty in the estimation of atmospheric response associated with the sea-ice variability is relatively high and needs to be carefully interpreted.