A new generation aerosol optical depth dataset based on AVHRR data over China from 1981 to 2000

The Advanced Very High Resolution Radiometer (AVHRR) series onboard the National Oceanic and Atmospheric Administration (NOAA) and the EUMETSAT Meteorological Operational Satellite (Metop) polar-orbiting satellites have provided continuous Earth observation data since 1979, which facilitates the development of long-term global climate data records. In this paper, a new version of the algorithm for the retrieval of the Aerosol Optical Depth (AOD) over Land (ADL v2.0) using AVHRR data is proposed with improved accuracy, in particular for high AOD values. The surface reflectance estimation scheme is based on a regression model established using simulated AVHRR reflectances spectrally transferred from the Moderate Resolution Imaging Spectroradiometer (MODIS) MOD09/MYD09 product. To address limitations in retrieving high AOD, the surface reflectance is determined using the maximum Normalized Difference Vegetation Index (NDVI) during a certain period of time. To this end, a dynamic NDVI search window is proposed to identify the NDVI that is least affected by aerosols. ADL v2.0 has been applied to provide an AOD dataset covering Mainland China (70o -140◦E, 15o -60o N) for the years from 1981 to 2000. This dataset has been evaluated by comparing with AOD data available from the application of the broadband extinction method (BEM) to ground-based solar radiation measurements and from the AVHRR Deep Blue (DB) AOD dataset. The AOD variations retrieved using the BEM data at seven stations (two in North China, two in Northeast China, one in East China, one in Central China, and one in the southwest mountainous region) are well reproduced by the ADL v2.0 algorithm. The comparison with the AVHRR DB AOD dataset shows good agreement with ADL v2.0 retrieval results even though with less valid retrievals for high AOD in Eastern China, Sichuan, and the Guanzhong Basin, as well as over North India.