Observed and reanalysis cloud fraction

Because the Arctic climate is particularly sensitive to cloud-radiative interactions, climate models must represent Arctic clouds realistically in order to capture the variations and feedbacks in high-latitude climate. Observations of clouds and radiative fluxes for the North Slope of Alaska are available from the Department of Energy's Atmospheric Radiation Measurement (ARM) program. Reanalysis models also calculate cloud and radiative variables. In this study, ARM measurements and North American Regional Reanalysis output for four midseason months are used to show that boundary layer clouds are not only the most common type of cloud observed on the North Slope but they are also a major cause for error by the reanalysis. Near-surface clouds are associated with large overestimates of the cloud fraction during the cold season and large underestimates during the warm season. These results were synthesized with other data to produce a comprehensive picture of synoptic conditions that are commonly present when the reanalysis fails to simulate the cloud fraction. When errors in the simulated cloud fraction are largest during the cold season, anomalously high pressure is observed north of the Bering Strait, with the departure being largest in magnitude and most widespread spatially in January. Large undersimulations in the summer are associated with a + 9 hPa deviation from climatology over the Arctic Ocean, a configuration that favors onshore flow from the northeast and east. More generally, large undersimulations in the summer clouds at Barrow are almost exclusively associated with onshore flow.