A cloud optical and microphysical property product for the advanced geosynchronous radiation imager onboard China's Fengyun-4 satellites: The first version

Fengyun-4 (FY-4), the latest collection of Chinese geostationary meteorological satellites, monitors the Eastern Hemisphere with high spatiotemporal resolutions. This study developed a cloud optical and microphysical property product for the Advanced Geosynchronous Radiation Imager (AGRI) onboard the FY-4 satellites. The product focuses on cloud optical thickness (COT) and cloud effective radius (CER) using a bi-spectral retrieval algorithm, and also includes cloud mask and phase using machine learning (ML) algorithms as prerequisites for COT and CER retrievals. The ML-based algorithm develops four independent models using Random Forest methods for cloud mask, liquid water, ice, and mixed-phase/multi-layer clouds, respectively. A two-habit ice and sphere water cloud model are employed to give their optical properties. Look-up tables of cloud reflectance in the COT and CER sensitive channels are built for efficient forward simulations, and the retrieval is performed by an optimal estimation algorithm. Compared with collocated active observations, the cloud mask and phase results give true positive rates of ∼95% and ∼85% and are more sensitive to mixed-phase clouds. Meanwhile, the AGRI-based COT and CER agree closely with those given by the collocated MODIS and AHI cloud products, and the correlation coefficients between MODIS and the AGRI results are 0.76 and 0.63 for COT and CER, respectively. The COT and CER retrievals will be persistently maintained and improved as the operational product for FY-4/AGRI.