资源环境科技发展态势分析平台

This study presents a comprehensive evaluation of eight aerosol optical depth (AOD) products from the latest MODIS C6.1 DT, DB and DTB, MISR V23, and VIIRS V1 DB dataset against the Aerosol Robotic Network (AERONET) measurements on global scale during 2012 to 2019. The latest MODIS DB products are found to be superior over DT in most cases. MODIS DTB products generally have a better performance than either DT or DB. VIIRS shows the best overall performance on global and regional scale while MISR has a systematic underestimation in moderate and high AOD conditions. Median AOD bias shows distinct regional dependence with the biggest divergence between satellite AOD products appearing in Asia. All these AOD products tend to have better performance in high latitude region and poorer performance in the tropics likely due to the complex aerosol condition and common high cloud cover in the tropics. Time series stability evaluation shows a distinct temporal pattern of statistics, in which the observation number and median AOD bias tend to be high in summer while low in winter and correlation R has a reversing trend. Among these products, VIIRS V1 DB has the highest time series stability of bias. Influence of aerosol loading, particle size, land use, elevation, and precipitable water and temperature on AOD retrievals is broadly consistent between products. High aerosol loading, small particle size, bright surface, high altitude, and high precipitable water and temperature tend to decrease the retrieval accuracy, e.g. lower expected error (EE) within ratio and larger median bias. The latest AOD products need to be furthermore improved under extreme conditions like very fine aerosols, elevated terrain and high precipitable water and temperature.