The role of diurnal processes in the seasonal evolution of sea ice and its snow cover

Diagnostic comparisons between hourly and daily average forcing simulations using a one-dimensional thermodynamic sea-ice model show significant differences due to diurnal changes in the surface energy balance that directly affect the snow layer and indirectly affect modeled ice conditions. Hourly forcing produced earlier onset of snow melt compared to daily average forcing and a 21 day longer open water duration. These differences are due to nonlinearities associated with absorbed shortwave radiation as well as latent and sensible heat fluxes. The accuracy of model parameterizations of downwelling shortwave (K/spl darr/) and longwave (L/spl darr/) fluxes and albedo over diurnal time scales is assessed by comparison to field observations made during the Seasonal Ice Monitoring and Modeling Site (SIMMS) 1992/93 field experiments near Resolute Bay, NWT. Mean K/spl darr/ errors (modeled-observed) were -23 W m/sup -2/ with a standard deviation of 89 W m/sup -2/ and mean L/spl darr/ errors were -1 W m/sup -2/ with a standard deviation of 18.8 W m/sup -2/. Modeled dry/new snow and wet snow albedos were about 0.1 and 0.05, respectively, lower than those observed, allowing greater amounts of shortwave energy absorption into the snowpack. Using SIMMS'92 on-ice field data as replacement for model K/spl darr/, L/spl darr/ and albedo parameterizations shows very different spring period simulations compared to using hourly forcing from Resolute land-based observations. Simulated net surface flux, surface temperature and snow/ice ablation were significantly improved.