Effect of surface heterogeneity on energy fluxes during the SHEBA winter

Using the extensive Surface Heat Budget of the Arctic Ocean (SHEBA) data set, a method has been devised for obtaining the spatial distribution of ice surface temperature, surface sensible (and latent) heat fluxes, snow depth and ice thickness on the Arctic pack ice. This method relies on a simple 1-D snow and ice model and on the ability to assign a formation time to any pixel identified as being first-year ice (FYI) using synthetic aperture radar (SAR) data. Validation shows that the 1-D model and this SAR/1D method both produce reasonable and consistent results for the SHEBA mid-winter conditions. However, extending the technique forward in time may be difficult, since the complex deformations of the pack ice later in the winter make it difficult to use the semi-subjective technique for identifying various FYI categories and assigning formation dates.

The method developed allows the estimation of aggregate surface temperature (Ts) and sensible heat flux (Hs) (as well as other parameters) over a GCM-scale grid surrounding the SHEBA site during a two-month-long period on an hourly basis for both clear and cloudy conditions. The best estimate for the average aggregate Hs is 6.2 W m-2 greater than the simultaneous average Hs value on the multi-year ice (MYI) at the SHEBA site. The aggregate Hs are typically greater than the MYI values for both cloudy and clear conditions. This difference is less than the 10-12 W m-2 difference suggested by most of the AVHRR assessments by Overland et al. (2002), which was done for only clear-sky conditions. The new method also obtains these larger differences if only the clear-sky times of the AVHRR measurements are used.

A consistent application of three different assumptions regarding the heat transfer coefficient (CH) in the 1-D model and the aggregation methods suggests that the computation of CH based on local stability is inappropriate because the resulting Ts are too cold. The differences in the aggregate Hs between the other two assumptions are much smaller than suggested by static aggregation applications to a given spatial Ts distribution.

The SAR/1D technique is also capable of providing spatial fields of snow depth and ice thickness. The heterogeneity in these two parameters produces the large variability in Ts, which in turn may produce mesoscale circulations in the atmospheric boundary layer. Such fields are needed as a lower boundary condition to 3-D mesoscale models.