A21E-3092:
A Consistent Aerosol Optical Depth (AOD) Dataset over China
Tuesday, 16 December 2014
Yahui Che1,2, Yong Xue1,3, Hui Xu1,2 and Jie Guang1, (1)RADI Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing, China, (2)University of Chinese Academy of Sciences, Beijing, China, (3)London Metropolitan University, Faculty of Life Sciences and Computing, London, United Kingdom
Abstract:
The Moderate Resolution Imaging Spectroradiometer (MODIS), Multiangle Imaging Spectroradiometer (MISR) as well as Sea-viewing Wide Field-of-view Sensor (SeaWiFS) can provide reliable AOD products over land. For MODIS, version of collection 051 AOD product including AOD data retrieved by Dark Target (DT) algorithm and Deep Blue (DB) algorithm and China collection 2.0 AOD dataset retrieved by synergetic retrieval of aerosol properties algorithm (SRAP) have good accuracy. However, AOD products from these different satellites and different algorithms are not consistent. The objective of this study is to merge multiple satellite AOD products to obtain a new consistent AOD dataset. For this purpose, five satellite AOD products are chosen as the source AOD maps, including three MODIS AOD products (Collection 051 MODIS DT and MODIS DB AOD product, China collection 2.0 AOD dataset), one SeaWiFS AOD product and one MISR AOD product. Because each source AOD product may have different biases for different surface albedo and AOD values, the biases and fusion weights were computed for different ranges of surface albedo and AOD values separately. Using the weight result, fused AOD dataset is produced based on maximum likelihood estimate (MLE) algorithm. The fused AOD dataset has been validated using AOD data from China Aerosol Remote Sensing Network (CARSNET) and Aerosol Robotic NETwork (AERONET). The new AOD dataset shows significant improvement on the correlation with the AERONET and CARSNET derived AOD data with good accuracy. Compared with AERONET AOD data, the correlation coefficient (R) of fused AOD dataset increased to 0.92 from 0.78 (MODIS-DB), 0.85 (SRAP), 0.88 (MISR), 0.88 (SeaWiFS) AOD product. The root mean square error (RMSE) of fused AOD dataset is 0.18, which is nearly equal to MODIS-DT (RMSE=0.18), SRAP (RMSE=0.16) and SeaWiFS (RMSE=0.17) AOD product but much smaller than MODIS-DB (RMSE=0.42) AOD product. Compared with CARSNET AOD data, the R of fused AOD dataset increased to 0.94 from 0.75 (SeaWiFS), 0.87 (MODIS-DB), 0.87 (SRAP), 0.87 (MISR) AOD product. The RMSE of fused AOD dataset is 0.16, which is smaller than MODIS-DT (RMSE=0.18), SRAP (RMSE=0.17) and SeaWiFS (RMSE=0.18) AOD product and much smaller than MODIS-DB (RMSE=0.37) AOD product.
