The Arctic sea-ice cover is typically simulated using monthly-mean cloud fractions. To examine the sensitivity of sea-ice thickness and distribution to the more realistic case of cloud fractions that vary over synoptic time scales, we simulate the ice cover with a two-dimensional ice model and simple radiation parameterizations containing linear and non-linear cloudiness terms. Using a common non-linear parameterization with daily-varying, clear-sky/overcast cloud fractions that sum to the monthly mean rather than daily cloud fractions equal to the monthly mean, the simulated ice volume in the Arctic and peripheral seas increases by 38% overall with a net decrease in downwelling shortwave radiation of 26%. Mid-summer ice extent increases by 17%, while ice thickness doubles in the Kara and Barents seas. Since the ice model is shown to respond nearly linearly to day-to-day variations in downwelling radiation, the need to consider synoptic variability of cloud fraction depends on the non-linearity of the relationship between cloud fraction and radiation. ¿ American Geophysical Union 1993