Comparisons between hourly forced data simulations and daily average forced simulations using a one-dimensional thermodynamic sea ice model show that diurnal changes in the surface energy balance that directly affect the snow depth, albedo, and surface temperature ultimately affect modeled seasonal ice evolution. Hourly forcing produces earlier onset of snowmelt, and open water duration is increased by 21 days. 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↓) and longwave (L↓) fluxes and albedo over diurnal timescales is assessed by comparison to field observations made during the Seasonal Ice Monitoring and Modeling Site (SIMMS) 1992--1993 field experiments near Resolute Bay, Northwest Territories. Mean K↓ errors (modeled - observed) were -23 W m-2, with a standard deviation of 89 W m-2, and mean L↓ errors were -1 W m-2, with a standard deviation of 18.8 W m-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 1992 on-ice field data as replacement for model K↓, L↓, and albedo parameterizations shows very different spring period simulations compared with using hourly forcing from Resolute land-based observations. Simulated net surface flux, surface temperature, and snow-ice ablation were significantly improved. ¿ 1999 American Geophysical Union