Cloud radiative forcing at the top of the atmosphere is derived from narrowband visible and infrared radiances from NOAA-12 and NOAA-14 advanced very high resolution radiometer (AVHRR) data taken over the Arctic Ocean during the First ISCCP Regional Experiment Arctic Cloud Experiment (FIRE ACE) during spring and summer 1998. Shortwave and longwave fluxes at the top of the atmosphere (TOA) were computed using narrowband-to-broadband conversion formulas based on coincident Earth Radiation Budget Experiment (ERBE) broadband and AVHRR narrowband radiances. The NOAA-12/NOAA-14 broadband data were validated using model calculations and coincident broadband flux radiometer data from the Surface Heat Budget of the Arctic Ocean experiment and from aircraft data. The AVHRR TOA albedos agreed with the surface- and aircraft-based albedos to within one standard deviation of ¿0.029 on an instantaneous basis. Mean differences ranged from -0.012 to 0.023 depending on the radiometer and platform. AVHRR-derived longwave fluxes differed from the model calculations using aircraft- and surface-based fluxes by -0.2 to -0.3 W m-2, on average, when the atmospheric profiles were adjusted to force agreement between the observed and the calculated downwelling fluxes. The standard deviations of the differences were less than 2%. Mean total TOA albedo for the domain between 72 ¿N and 80 ¿N and between 150 ¿W and 180 ¿W changed from 0.695 in May to 0.510 during July, while the longwave flux increased from 217 to 228 W m-2. Net radiation increased from -89 to -2 W m-2 for the same period. Net cloud forcing varied from -15 W m-2 in May to -31 W m-2 during July, while longwave cloud forcing was nearly constant at ~8 W m-2. Shortwave cloud forcing dominated the cloud effect, ranging from -22 W m-2 during May to -40 W m-2 in July. The mean albedos and fluxes are consistent with previous measurements from the ERBE, except during May when the albedo and longwave flux were greater than the maximum ERBE values. The cloud-forcing results, while similar to some earlier estimates, are the most accurate values hitherto obtained for regions in the Arctic. When no significant melting was present, the clear-sky longwave flux showed a diurnal variation similar to that over land under clear skies. These data should be valuable for understanding the Arctic energy budget and for constraining models of atmosphere and ocean processes in the Arctic. ¿ 2001 American Geophysical Union