Updating satellite-based estimates of aerosol-cloud interactions with refined approximation of cloud condensation nuclei

Some atmospheric particles are capable of initiating droplet formation at a supersaturation. Aerosol-cloud interactions (ACI) mediated by these cloud condensation nuclei (CCN) remain poorly quantified. Satellite observations yield estimates of ACI in ways that complement model-based studies. Updating the estimates with extended satellite records would improve the statistics and illuminate multi-decadal trends. Satellite-based ACI studies simplistically substitute aerosol optical depth (AOD) times its wavelength dependence for CCN. Emerging model and satellite capabilities promise refined CCN approximation, warranting evaluation with direct measurements.

We refine CCN approximation and update satellite-based ACI estimates. To refine CCN approximation, we obtain light extinction and its wavelength dependence from MERRA-2 reanalysis, compute CCN proxies and analyze errors using suborbital measurements. Specifically, the boundary-layer extinction determined by MERRA-2 reanalysis are used to calculate the so-called aerosol index and the recently developed CCN indices, each on a column-integral basis and on an altitude-resolved basis, the latter more relevant to the boundary-layer CCN. These CCN approximations are evaluated with the direct measurements in DOE ARM facilities (e.g., Southern Great Plains and Azores) and NASA research flights (e.g., ARCTAS, DISCOVER-AQ and ORACLES).

To update satellite-based ACI estimates, we match decades-long cloud products to the approximated CCN, run regression and investigate the impact of environmental and technical factors. Specifically, the MODIS derived boundary-layer cloud droplet number concentration since 2003 are co-located with the CCN approximations. Linear regression on log-log scales between each cloud property-CCN pair is studied. Our preliminary analysis suggests that, because Aerosol Index varies more than CCN, the regression slope, which past studies identified with the strength of ACI, may be underestimated.

We plan to extend the analysis to other data source (e.g., Deep Blue, PARASOL GRASP) and cloud properties (e.g., optical thickness, effective radii); estimate the humidification effect from the MERRA-2 aerosol species and relative humidity; deduce the uncertainty for the global CCN approximations; and study the regional, seasonal and multi-year trends of the regression results. The analysis is expected to help make recommendations on future efforts and discuss implications for radiative forcing calculations.