Observational Evidence for Aerosol Induced Variability in Ice Cloud Properties

Aerosols significantly affect the micro- and macrophysics of ice cloud by scattering and absorbing solar radiation (radiative effect) and by serving as ice nuclei (microphysical effect), but the magnitude of these effects bears the largest uncertainty among external influences on climate change. Natural aerosols, especially mineral dust, are typically known to nucleate ice efficiently and to dominate the global ice nuclei population. However, some observational studies have implied that anthropogenic aerosols may play a similar or even more important role in nucleating ice than mineral dust in China, which is characterized by high anthropogenic aerosol loadings. In this study, we systematically investigate the impact of aerosols on ice cloud properties and the consequent radiative forcing using collocated observations from multiple satellites in the A-Train constellation during 2007-2015, with special focus on the impacts of different aerosol types. We combine the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) and CloudSat to determine the exact locations of ice cloud relative to aerosols, and select coexisting and mixed aerosol and ice cloud layers. We find that, for mixed aerosol-cloud conditions, the ice cloud fraction increases significantly with increasing AOD for moderate aerosol loadings (AOD<0.4), and then becomes saturated for higher aerosol loadings. The ice crystal effective radius (r_e) increases with increasing AOD in pristine conditions (AOD<0.1), indicative of the transformation from homogeneous ice nucleation to heterogeneous nucleation. For pixels with AOD>0.1, r_e decreases steadily with AOD, indicative of the formation of more and smaller ice crystals. Ice water path (IWP) is not found to change significantly with AOD. Using the aerosol classification provided by CALIPSO, we find that the occurrence frequency of ice cloud layers mixed with air pollution aerosols or smoke is relatively lower compared with those mixed with dust or polluted dust. However, the relationships between ice cloud properties (e.g., ice cloud fraction, r_e, IWP, and cloud top pressure) and AOD are quite similar for air pollution aerosols, dust, polluted dust, and smoke, implying that all these types of aerosols can significantly impact the formation and evolution of ice clouds. We demonstrate that meteorological conditions, including relative humidity, lower tropospheric stability, updraft velocity, precipitation, wind speed, and wind direction have very limited influence on the abovementioned AOD-ice cloud relationships. Therefore, we argue that the changes of ice cloud properties are very likely to be primarily driven by aerosols. Based on these analyses, we estimate the aerosol-cloud radiative forcing by ice cloud in East Asia for the first time. We first estimate the aerosol–ice cloud radiative forcing for each aerosol type identified by CALIPSO using the sensitivity of cloud albedo and cloud fraction to AOD and the changes of AOD relative to preindustrial era. Then we estimate the overall aerosol-ice cloud radiative forcing by averaging the radiative forcing for each aerosol type weighted by the occurrence frequency of each aerosol type.