Retrieval of aerosol single scattering albedo over land using geostationary satellite data

Absorbing aerosols from dust, industrial emissions and biomass combustion have a strong impact on solar radiation in the atmosphere, and they are considered to be an important source of regional air pollution[1], especially in East Asia, where the degree of variation of absorbable aerosols is very large, so it will have a significant impact on regional climate change. Aerosol single scattering albedo (SSA) is a key variable of aerosol absorption and a key metric of climate impact. Accurate estimation of SSA is crucial to reduce uncertainties in the study of atmospheric pollution and climate effects.

The aerosol model of the current aerosol inversion algorithm is several typical regional aerosol candidate models obtained by cluster analysis of selected ground observation data[2]. These independent candidates' aerosol models result in the estimation of SSA being just a few simple constants, leading to a large bias in the results of SSA.

In the current study, an atmospheric radiative transfer model parameterized by a two-stream approximation is used to construct a genetic algorithm for application to geostationary satellite Himawari-8 / AHI data to retrieve the aerosol SSA. The inversion process is constrained by AOD and the surface bidirectional reflectance distribution function (BRDF). Using this algorithm, hourly SSA data were retrieved was retrieved during day time.

The algorithm was tested using the hourly L1 grid data of AHI from 00:00 to 07:00(UTC) from January to March 2020. Examples of the 10km satellite-retrieved SSA on March 15, 2020, are shown in Fig. 1. It can be seen that the SSA value in East China is significantly lower than in other regions. This may be due to the developed industries in southern cities that emit a large number of black carbon aerosols, while the temperature in winter is low, and the aerosol particles are not easily diffused, resulting in the SSA value of some areas maintaining a stable low value.

To evaluate the retrieval SSA results, AERONET V3 datasets were used for validation. The AERONET datasets were selected in East Asia. Fig. 2 shows the scatter plots of AHI SSA retrievals versus AERONET at 470nm, (a) all AOD, and (b) only high aerosol loading (AOD>0.4 at 470nm), respectively. This indicates that the algorithm has great advantages for SSA inversion of heavy pollution conditions.

The retrieval results from three months of AHI data were evaluated against the ground-based AERONET measurements. The AHI SSA shows good agreement with AERONET measurements, especially in heavy pollution conditions. This algorithm has been proven to be to characterize the temporal and spatial distribution of aerosol SSA.