Comparison of the Copernicus Sentinel-2 L2A Core Product distributed by ESA and the Sen2Cor Toolbox ‘user-generated’ product

Kurzfassung

Sen2Cor is a Level-2A (L2A) processor whose main purpose is to correct mono-temporal Copernicus Sentinel-2 (S2) mission Level-1C (L1C) products from the effects of the atmosphere in order to deliver radiometrically corrected Bottom-of-Atmosphere (BOA) data. Byproducts are Aerosol Optical Thickness (AOT), Water Vapour (WV) and Scene Classification (SCL) maps. The Sen2Cor Toolbox can be downloaded from the ESA website for autonomous processing of S2 L1C data by the users, thus generating BOA products here referred as ‘user’ products. In parallel, Sen2Cor is used for systematic processing of Sentinel-2 L1C data thus generating the S2 L2A products systematically distributed to users. Operational global L2A processing with the integration of Sen2Cor in the Sentinel-2 Ground Segment started in December 2018. These S2 L2A core products can be downloaded from the Copernicus SciHub. The S2 L2A core products agree with ‘user’ products as long as both are generated with the same Sen2Cor version and configuration. However, sometimes new versions of Sen2Cor Toolbox are released to public at a different time than their implementation in the S2 Ground Segment processing chain. This may result in a disagreement between ‘user’ and core products for some time. Additionally, differences between the outcome of the two processing lines can arise due to the DEM used and different minor patches included. Patches are usually included faster in the ESA-L2A core products. Whereas ESA-L2A core products have been using Planet-DEM for a while, the current DEM available to the users is SRTM-DEM. However, the new Copernicus DEM is now available also to users and is going to be used also to generate the S2 L2A core products, this limiting those differences. Both production lines are compared for what concerns AOT and WV maps, and BOA outputs per band. Both AOT and WV retrievals are validated by comparison with reference data provided by AERONET sunphotometers. Comparison of BOA outputs is performed by statistical metrics for pixelby-pixel differences between both processing lines over 9 km x 9 km areas. The influence of the different DEMs on resulting BOA will be discussed. ESA-L2A core products are generated with default configuration values, which are suitable for most situations of this operational mission. ‘User’ production gives the opportunity to apply individual configuration settings, which may prove favourable in some cases. This will be demonstrated on a few examples.