Cloud phase and microphysical properties control the radiative effects of clouds in the climate system and are therefore crucial to characterize in a variety of conditions and locations. An Arctic-specific, ground-based, multisensor cloud retrieval system is described here and applied to 2 yr of observations from Barrow, Alaska. Over these 2 yr, clouds occurred 75% of the time, with cloud ice and liquid each occurring nearly 60% of the time. Liquid water occurred at least 25% of the time, even in winter, and existed up to heights of 8 km. The vertically integrated mass of liquid was typically larger than that of ice. While it is generally difficult to evaluate the overall uncertainty of a comprehensive cloud retrieval system of this type, radiative flux closure analyses were performed in which flux calculations using the derived microphysical properties were compared with measurements at the surface and the top of the atmosphere. Radiative closure biases were generally smaller for cloudy scenes relative to clear skies, while the variability of flux closure results was only moderately larger than under clear skies. The best closure at the surface was obtained for liquid-containing clouds. Radiative closure results were compared with those based on a similar, yet simpler, cloud retrieval system. These comparisons demonstrated the importance of accurate cloud-phase and cloud-type classification, and specifically the identification of liquid water, for determining radiative fluxes. Enhanced retrievals of liquid water path for thin clouds were also shown to improve radiative flux calculations.