Improvement in AOD retrieval from geostationary measurement over the ASIA with obtained AOP from DRAGON-2012 campaign

The long-term aerosol monitoring from ground-based sun photometer such as AERONET has variously used to obtain optimized aerosol optical properties (AOPs) for assumed aerosol type and to validate aerosol optical depth (AOD) retrieved from satellite. Additionally, the meso-scale network campaign such as DRAGON ASIA-2012 is suitable to monitor the aerosol characteristic over localized area. Thus, this study focused on the improvement of AOPs over the East Asia by using the DRAGON ASIA-2012 campaign dataset for two geostationary AOD retrieval algorithms. The algorithms were developed for Geostationary Ocean Color Imager (GOCI) and Meteorological Imager (MI) onboarding the Communication, Ocean, and Meteorological Satellite (COMS), which was launched in June, 2010. GOCI has 8 channels in visible and near IR. Using these multi-spectral bands, we retrieved aerosol optical properties such as AOD, fine-mode fraction (FMF), single scattering albedo (SSA) and aerosol type. In this study, we consider dynamic aerosol models categorized by FMF and SSA from East Asia AERONET data including DRAGON campaign data. Using these aerosol data, AOPs are upgraded for LUT calculation and re-retrieved from GOCI. The validation with AERONET shows the improved results after AOP upgrade. On the other hand, a single channel algorithm developed for MI assumed seasonal aerosol property variation due the algorithm has limitation in select aerosol type. The seasonally analyzed SSAs at 675 nm from the AERONET measurement over the East Asia including the campaign data are 0.92, 0.94, 0.92, and 0.91 for spring (MAM), summer (JJA), autumn (SON), and winter (DJF), respectively, and slightly higher then the originally used values. By using those newly analyzed AOP, the comparison result between retrieved AOD and measured value from AERONET shows the increase of correlation coefficient from 0.551 to 0.702 and the increase of regression slope from 0.538 to 0.615 during months from May to Dec. in 2011.