Comparisons between the distributions of dust and combustion aerosols in MERRA-2, FLEXPART, and CALIPSO and implications for deposition freezing over wintertime Siberia

Abstract. Aerosol distributions have a potentially large influence on climate-relevant cloud properties but can be difficult to observe over the Arctic given pervasive cloudiness, long polar nights, data paucity over remote regions, and periodic diamond dust events that satellites can misclassify as aerosol. We compared Arctic 2008–2015 mineral dust and combustion aerosol distributions from the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite, the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2) reanalysis products, and the FLEXible PARTicle (FLEXPART) dispersion model. Based on coincident, seasonal Atmospheric Infrared Sounder (AIRS) Arctic satellite meteorological data, diamond dust may occur up to 60 % of the time in winter, but it hardly ever occurs in summer. In its absence, MERRA-2 and FLEXPART each predict the vertical and horizontal distribution of large-scale patterns in combustion aerosols with relatively high confidence (Kendall tau rank correlation > 0.6), although a sizable amount of variability is still unaccounted for. They do the same for dust, except in conditions conducive to diamond dust formation where CALIPSO is likely misclassifying diamond dust as mineral dust and near the surface (< ∼ 2 km) where FLEXPART may be overpredicting local dust emissions. Comparisons to ground data suggest that MERRA-2 Arctic dust concentrations can be improved by the addition of local dust sources. All three products predicted that wintertime dust and combustion aerosols occur most frequently over the same Siberian regions where diamond dust is most common in the winter. This suggests that dust aerosol impacts on ice phase processes may be particularly high over Siberia, although further wintertime model validation with non-CALIPSO observations is needed. This assessment paves the way for applying the model-based aerosol simulations to a range of regional-scale Arctic aerosol–cloud interaction studies with greater confidence.