Comparison of DDV-algorithm for AOT estimation in Sen2Cor and use of AOT from CAMS data

Kurzfassung

The Copernicus Sentinel-2 mission provides data since the launch of the Sentinel-2A unit in 2015. The launch of the Sentinel-2B in 2017 created a constellation of two polar orbiting satellite units. Both Sentinel-2A and Sentinel-2B are equipped with an optical imaging sensor MSI (Multi-Spectral Instrument) which acquires decametric spatial resolution optical data products. The Sentinel-2 mission serves for observation of land-cover change and deriving biophysical variables related to agriculture and forestry, monitors coastal and inland waters and is useful for risk and disaster mapping. Atmospheric correction processor Sen2Cor was developed to remove the effect of the atmosphere from Sentinel-2 data. Sen2Cor is designed for mono-temporal processing of Sentinel-2 L1C data products providing Level-2A Bottom-of-Atmosphere (BOA) surface reflectance product together with Aerosol Optical Thickness (AOT), Integrated Water Vapour (WV) and Scene Classification (SCL) maps. The processor relies on Dense Dark Vegetation (DDV) pixels for estimation of AOT and uses AOT from Copernicus Atmospheric Monitoring Service (CAMS) as fall-back option in case there are not enough DDV-pixels in the image. The data set for validation was split into two subsets so far investigating the performance of the DDValgorithm and the CAMS-fall-back option. CAMS fall-back option performs better than the DDV-algorithm in some cases and worse in others. However, this is no comparison of DDV-algorithm and CAMS fall-back option because it is based on different images in each subset. The presentation will compare both AOT-options of Sen2Cor on the same images. The analysis will start as before reporting results for the two subsets. Then, Sen2Cor will be forced using CAMS data for reprocessing the subset of images with enough DDV-pixels in the image allowing a direct comparison of DDV-algorithm and CAMS data use. The comparison is done with reference data from AERONET on a dataset of more than 1000 Sentinel-2 images distributed around the globe.