Bias and Sensitivity of Boundary Layer Clouds and Surface Radiative Fluxes in MERRA-2 and Airborne Observations Over the Beaufort Sea During the ARISE Campaign

The representation of Arctic surface radiative fluxes in atmospheric models and reanalyses is integral to understanding relevant physical processes, yet testing of these models is confounded by a scarcity of in situ observations of near-surface atmospheric state profiles, cloud vertical structure, cloud phase, and surface properties. Here, airborne measurements obtained from the Arctic Radiation IceBridge Sea&Ice Experiment (ARISE) during fall 2014 are compared with concurrent products from the Modern-Era Retrospective analysis for Research and Applications, version 2 (MERRA-2). Both data sets are then used as input to a radiative transfer model to produce surface radiative fluxes over multiple locations in the Beaufort Sea under the various conditions observed during ARISE. The sensitivity of the simulated fluxes is assessed and compared between these two data sets. Then, the relative contribution of atmospheric state, boundary layer clouds, and their properties to the sensitivity of the simulated surface fluxes is assessed. In our comparisons with ARISE observations we found that MERRA-2 has a warm temperature bias near the surface and it underestimates near-surface clouds and cloud liquid and ice water content. These prevail over both open water and sea ice surfaces. Our sensitivity analysis showed that boundary layer cloud vertical structure and water content account for more than 70% of the difference between MERRA-2 and the radiative fluxes calculated from airborne observations and that differences in boundary layer atmospheric state parameters contribute about 10–20% to the positive bias in the longwave surface flux.