Error Analysis for the Himawari-8 Aerosol Optical Depth Basing on Parts of Aerosol Model and Sun Position over Wuhan, Central China

Aerosols attract great attention as having critical influence on the Earth’s energy budget and human health. Geostationary satellites like Himawari-8 process advantages on temporal resolution that allows rapidly changing weather phenomena tracking and aerosol monitoring. This work aims at providing a novel error analysis for the Advanced Himawari Imager (AHI) aerosol optical depth (AOD) retrieval from the aspect of aerosol model and sun position combing with the high quality ground-based observation in Wuhan, central China. Three-year co-located AOD dataset from AHI and sun-photometer are used. AHI underestimates AOD in all the seasons. Aerosol size distributions and phase functions are discussed as parts of aerosol model to explain the underestimation of AOD. AHI sets a low fine-mode particle median radius comparing with the in-site measurement in Wuhan that increases backscattering, and finally leads to the underestimation of AOD. Sun position also affects AHI AOD retrieval, and we use solar zenith angle (SZA) and scattering angle to represent sun position. Geostationary satellites get fixed satellite position for one site that provides convenience to the discussion. SZA influences AOD retrieval mainly through the length of transfer path and higher percent of samples within expected error often appears at low SZAs. Scattering angle also has obvious influence on AOD retrieval through the simulation of phase function and causes the difference of correlation performance between AHI and sun-photometer in aspect of SZA in morning and afternoon. Finally, we applied the dark target method to retrieve AHI AOD. The comparison of AODs reveals that the retrieval method of AHI performs better in Wuhan. The better performance of AHI AOD may be due to high aerosol loading and lack of enough prior information of aerosol properties in Wuhan. Our work could also be performed on other areas or other geostationary satellites, and help us to further understand the controlling factors that affect AOD retrieval accuracy, then contribute to better AOD retrieval.