Aerosol Fine-Mode-Fraction Retrieval From GEO-KOMPSAT-2A/AMI Using a Deep Neural Network and Spectral Deconvolution Algorithm

Aerosol size information is important to the understanding of aerosol dynamics, which change rapidly during wildfire, dust transport, and volcanic eruption events over Asia. In this study, a deep neural network (DNN) model was trained using Advanced Meteorological Imager (AMI) Level 1B observations, AMI Yonsei aerosol retrieval (YAER) aerosol products, and observation geometries to retrieve the aerosol optical depth (AOD),Ångström exponent (AE), and spectral derivatives of AE (AE ${}^{\prime }$ ). The fine-mode fraction (FMF) was calculated with a spectral deconvolution algorithm (SDA) using retrieved AE and AE ${}^{\prime }$ when AOD >0.2. The retrieved aerosol products were validated using Aerosol RObotic NETwork (AERONET) (AOD at 550 nm: $R =0.837$ , root-mean-square error (RMSE) =0.219, and mean bias error (MBE) $= -0.066$ ; AE: $R =0.726$ ; RMSE =0.231; MBE $= -0.007$ ; FMF: R =0.875; RMSE =0.072; and MBE =0.007). Case studies of dust transport, wildfire, and haze events in Asia revealed that the retrieved aerosol size products may be used for analysis of sudden pollution events. Results of this study indicate the potential for a comprehensive analysis of aerosol properties in Asia using continuous aerosol size data from geostationary Earth orbit (GEO) satellite observations.