Sea Ice Simulations Based on Fields Generated by the GLAS GCM

View More View Less
  • 1 Laboratory for Atmospheric Sciences, NASA, Goddard Space Flight Censer, Greenbelt, MD 20771
  • | 2 Department of Meteorology, University of Wisconsin, Madison. 53706
© Get Permissions
Full access

Abstract

A four-month simulation of the thermodynamic portion of the Parkinson-Washington sea ice model was conducted using atmospheric boundary conditions that were obtained from a pre-computed seasonal simulation of the Goddard Laboratory for Atmospheric Sciences' Gencral Circulation Model (GLAS GCM). The sea ice thickness and distribution were predicted for the 1 January–30 April period based on the GCM-generated fields of solar and infrared radiation, specific humidity and air temperature at the surface, and snow accumulation. The sensible heat and evaporative fluxes at the surface are mutually consistent with the ground temperatures generated by the ice model and the air temperatures generated by the atmospheric model.

In general, in the Northern Hemisphere the predicted ice distributions and the wintertime accretion and southward advance of the pack ice are well simulated. The computed ice thickness in the Southern Hemisphere appears reasonable, but the Antarctic melt season is extended, causing ice coverage to be less than observed in late March and April. During the Northern Hemisphere winter, the simulated ice accretion is the result of the net deficit of longwave radiation, heat gained from the ocean, am sensible heat lost to the atmosphere. In the early part of the Southern Hemisphere summer, the melting essentially balances the excess of solar over longwave radiation at the surface, while later in the simulation accretion balances the longwave and convective heal losses.

The results show that the Parkinson–Washington sea ice model produces acceptable ice concentrations and thicknesses when used in conjunction with the GLAS GCM for the January to April transition period. These results suggest the feasibility of fully coupled ice-atmosphere simulations with these two models.

Abstract

A four-month simulation of the thermodynamic portion of the Parkinson-Washington sea ice model was conducted using atmospheric boundary conditions that were obtained from a pre-computed seasonal simulation of the Goddard Laboratory for Atmospheric Sciences' Gencral Circulation Model (GLAS GCM). The sea ice thickness and distribution were predicted for the 1 January–30 April period based on the GCM-generated fields of solar and infrared radiation, specific humidity and air temperature at the surface, and snow accumulation. The sensible heat and evaporative fluxes at the surface are mutually consistent with the ground temperatures generated by the ice model and the air temperatures generated by the atmospheric model.

In general, in the Northern Hemisphere the predicted ice distributions and the wintertime accretion and southward advance of the pack ice are well simulated. The computed ice thickness in the Southern Hemisphere appears reasonable, but the Antarctic melt season is extended, causing ice coverage to be less than observed in late March and April. During the Northern Hemisphere winter, the simulated ice accretion is the result of the net deficit of longwave radiation, heat gained from the ocean, am sensible heat lost to the atmosphere. In the early part of the Southern Hemisphere summer, the melting essentially balances the excess of solar over longwave radiation at the surface, while later in the simulation accretion balances the longwave and convective heal losses.

The results show that the Parkinson–Washington sea ice model produces acceptable ice concentrations and thicknesses when used in conjunction with the GLAS GCM for the January to April transition period. These results suggest the feasibility of fully coupled ice-atmosphere simulations with these two models.

Save