Inferring Airborne Total-Column Aerosol Mass Speciation from Optical Measurements: A Step Towards Satellite-Inferred Chemical Composition

Passive satellite-derived total column-integrated aerosol properties are useful to constrain model-predicted AOD. However, this process does not correct the uncertainty associated with simulated vertical distribution of aerosols and aerosol chemical speciation. Uncertainties in aerosol speciation have significant implications in the correct estimation of the IPCC Radiative Forcing due to aerosol-radiation interactions (RFari) estimates. In particular, RFari for individual aerosol species are less certain than the total RFari [Boucher et al., 2013].

Another way to improve estimates of RFari would be to directly compare modeled and satellite-derived total column mass concentration per species. This would help adjust individual aerosol masses when applying data assimilation techniques in the model (and potentially the emission/chemistry/transport processes driving them). However, global coverage of satellite-derived aerosol chemical speciation does not currently exist.

To date, the best product we could infer from passive satellite observations are qualitative aerosol types (e.g., urban industrial, dark and white biomass burning smoke) from A-Train’s POLDER (Polarization and Directionality of Earth’s Reflectances) optical retrievals [Hasekamp et al., 2011]. We use POLDER retrievals of particle size, spectral light absorption and a pre-specified clustering method described in Russell et al. [2014]. These aerosol types are useful to provide spatial context to support other observations of aerosols and clouds or evaluate other aerosol type classifications.

However, they need to be translated into satellite-derived total column mass concentrations per chemical species to be consistent with model output, and hence, effectively correct models. To this end, we have translated aerosol type identifications into an averaged distribution of different chemical components (e.g., organic, sulfate, black carbon, mineral, sea salt) by an extensive dataset of aerosol chemical and physical properties taken onboard the aircraft during the 2013 SEAC4RS (Studies of Emissions and Atmospheric Composition, Clouds and Climate Coupling by Regional Surveys) mission over the United States as a training dataset. These averaged distributions, paired with assumed mass extinction efficiencies and auxiliary measurements were used to estimate aerosol mass speciation at the height of the airplane.

We propose to (1) improve our aerosol classification method and our optical-to-chemical translation, (2) understand the limitations of this type of translation, especially in the presence of a mix of different aerosol types, (3) infer aerosol type and mass speciation from airborne total column polarization measurements and (4) assess the utility of our total column mass speciation estimates in constraining models.