Validation of Sentinel-2 Auxiliary Data: Focus on Aerosol Optical Depth at 550 nm and Total Column of Water

Kurzfassung

The Copernicus Sentinel-2 mission comprises a constellation of two polar-orbiting satellites placed in the same sun-synchronous orbit, phased at 180° to each other. Since 2015 it has been monitoring the variability in land surface conditions thanks to its wide swath width (290 km) and high revisit time (10 days at the equator with one satellite, and 5 days with 2 satellites under cloud-free conditions which results in 2-3 days at mid-latitudes). Since the beginning of the mission, in complement of sentinel-2 imagery files, meteorological auxiliary data from European Centre for Medium-Range Weather Forecasts (ECMWF) are provided to the users within L1C and L2A products. The format (GRIB V1) and contents of this file are identical in L1C and L2A products. This auxiliary data is provided in a gridded format which results from a temporal (linear) and spatial interpolation (bilinear with a Ground Sample Distance of 12.5km) of the raw ECMWF global forecast dataset into a geographical area covering the Level-1C tiles footprint. Grid points are provided in latitude/longitude using WGS84 reference system. This meteorological auxiliary data file initially provided three parameters: - Total column ozone (TCO3) [Kg/m2]; - Total column water vapour (TCWV) [Kg/m2]; - Mean sea level pressure (MSL) [hPa]. Since the 25th of January 2022, the number of ECMWF meteorological parameters has been extended with three other parameters: - 10-meter V wind component (10v) - 10-meter U wind component (10u) - Relative Humidity @ Isobaric Surface , together with a new set of 10 atmospheric parameters from the Copernicus Atmosphere Monitoring Service (CAMS) that generates every day, five-day forecasts of aerosols, atmospheric pollutants, greenhouse gases, stratospheric ozone, and the UV-Index. This contribution reports on the monitoring of the two following parameters: - Aerosol Optical Depth at 550 nm (extracted from AUX_CAMS_FO file), - Total Column of Water (extracted from AUX_ECMWFT file). The Aerosol Optical Depth at 550 nm can be used as atmospheric information fallback in the L2A processor when performing the atmospheric correction for certain Sentinel-2 tiles when not enough dark dense vegetation pixels are present to perform an independent aerosol retrieval . The Total Column of Water is not used in the L2A processor when performing the atmospheric correction. However, it is interesting to check how its performance compares with the L2A processor outputs. The monitoring is performed on 24 different locations distributed over all continents and all climate zones using 24 AERONET stations. The data period covers more than 18 months of data, between 25/01/2022 to 25/08/2023, starting with Sentinel-2 PB 04.00, since the AUX_CAMS_FO files are embedded in the L1C and L2A products. The values of AUX_CAMS_FO and AUX_ECMWFT are extracted at the AERONET site location using its geographic coordinates. The Aerosol Optical Depth at 550 nm AERONET values are spectrally and temporally interpolated to the Sentinel-2 acquisition time. The Precipitable Water AERONET values are temporally interpolated to the Sentinel-2 acquisition time. We present the results of the monitoring in the form of scatter plots of all the concomitant Sentinel-2 auxiliary CAMS data with respect to the AERONET data for all the 24 AERONET test sites. These consolidated statistics on all test sites show that 53% of Aerosol Optical Depth at 550 nm CAMS data is within the uncertainty goal with an overall uncertainty of 0.17 without significant bias. It should be noted however that depending on the test site this value can range from 23% for the lowest agreement (Ilorin) up to 84% for the best agreement (Kangerlussuaq). In general, the agreement is better for test sites with lower aerosol load. For water vapour, these consolidated statistics on all test sites show that 88% of Total Column of Water ECMWF data is within the uncertainty goal with an overall uncertainty of 0.27 g.cm-2 with a slight positive bias of 0.15 g.cm-2 . It should be noted however that depending on the test site, this value can range from 75% for the lowest agreement (OHP_OBSERVATOIRE) up to 98% for the best agreement (Kangerlussuaq). In general, the agreement is better for test sites with dryer atmosphere.