Publications

Wang, L; Cai, K; Si, YD; Yu, C; Zheng, H; Li, SS (2020). Evaluation of Himawari-8 version 2.0 aerosol products against AERONET ground-based measurements over central and northern China. ATMOSPHERIC ENVIRONMENT, 224, 117357.

Abstract
Himawari-8, a new generation geostationary satellite, carries the Advanced Himawari Imager (AHI), which can provide high temporal resolution observations (10 min) in monitoring the distribution and variation of aerosol over large areas. Recently, the AHI aerosol algorithm was refined, and the updated AHI version 2.0 Level 2 aerosol products were released. However, these products have not been comprehensively evaluated. This study aims to assess the accuracy of AHI version 2.0 Level 2 aerosol products over central and northern China. We compared nine hourly (Beijing time: 08:00 to 16:00) AHI Aerosol Optical Thickness (AOT) retrievals from 2016 to 2017 with AErosol RObotic NETwork (AERONET) measurements over four merged sites: Beijing-urban, Xianghe, Xuzhou and Baotou. The results show that the AHI AOT product provides accurate retrievals at 09:00, 10:00 and 11:00, but these values overestimated AOT at high aerosol loadings. In contrast, AHI AOT retrievals at 13:00, 14:00 and 15:00 exhibit significant overestimation. More than 67% of these retrievals fell above the Expected Error (EE: +/- 0.15 x AOT(AERONET) +/- 0.05). The AHI Quality Assurance (QA_flag) performs poorly in constraining overestimation of the AHI AOT product. The "no confidence" angstrom ngstrom exponent (AE), additional cloud and surface reflectance factors in the QA_flag have negative impacts on AOT retrieval accuracy. For different aerosol models, the AHI AE and Optical Depth Ratio fine (RF) products exhibit large differences between the forenoon (09:00, 10:00 and 11:00) and afternoon (13:00, 14:00 and 15:00). However, the AHI Single-scattering albedo (SSA) values mainly distribute in the range from 0.93 to 0.94, exhibiting minor differences in different times, sites and aerosol models. These results indicated that these products have errors induced by the AHI algorithm.

DOI:
10.1016/j.atmosenv.2020.117357

ISSN:
1352-2310