Evaluation and improvement of MODIS aerosol optical depth products over China

Abstract

In this study, we use the level-1.5 and level-2.0 aerosol optical depth (AOD) from the Aerosol Robotic Network (AERONET) Version 3 dataset at 12 sites in China to evaluate the MODIS Collection 6.1 (C6.1) AOD products retrieved using three distinct algorithms: Dark Target (DT), Deep Blue (DB), and DTB merged from DT and DB from Terra and Aqua. The performance of the three algorithms is evaluated. Based on these evaluations, a simple and efficient AOD retrieval algorithm at a high spatial resolution of 1.0 km for China is proposed. AODs at this spatial resolution over China are then retrieved using the methods of DT and DB, respectively. The evaluation results showed that: (1) there is little difference in the AOD products derived from Terra and Aqua. The differences in the determination coefficient (R2) between the two platforms at most sites are less than 0.05. (2) There are relatively large differences in AODs between the AERONET level-1.5 dataset and MODIS dataset while these differences are mostly filtered out by the quality assurance process of the AERONET level-2.0 dataset. The statistical tests indicate that the MODIS DTB AOD product is generally better than those from the other two algorithms. Using the new algorithm developed in this study, the AODs at a high spatial resolution of 1 km in the whole of China are determined and the results are compared with the MODIS DTB product. The results show that the AODs determined using the new method agree reasonably well with those of the MODIS DTB dataset though the new results have slightly negative biases in the wintertime. However, these negative biases may not be a negative sign due to the fact that the MODIS AODs are subject to a positive bias relative to the AERONET AODs.