Abstract
The Sentinel-3 satellites, S3A and S3B, have been respectively launched in 2016 and 2018, by the European Space Agency (ESA) as part of the European Commission’s (EC) Copernicus program, and have continuously acquired images of the Earth since then. Both satellites carry the Ocean and Land Colour Instrument (OLCI) dedicated to scanning the ocean and land surface to harvest information of Earth visible and NIR radiances returned from the Earth’s surface to the sensor. The quality of the OLCI data is ensured through rigorous calibration and validation activity, but also thanks to the provisioning of prognostic uncertainties on the measurements. Recently, a comprehensive model of the uncertainties associated with the level-1 radiometry of OLCI has been developed. This improvement will be transferred to operation later this year by the ESA. Following the GUM methodology, the OLCI level-1 per-pixel radiometric uncertainties have been estimated from the level-0 measurements by propagating all radiometric sources of uncertainty at each step of the global level-1 processing chain. At this point a major difficulty is to provide a validation of these uncertainties. For this purpose, we took advantage of the 6 months tandem phase at the launch of S3B. The validation of the new OLCI uncertainties has been performed using the relative intercomparison of OLCI-A and B uncertainties-normalised difference, a metrological methodology already applied to the SLSTR instrument (Hunt et al. 2020) and adapted here to the OLCI specificities. This paper presents an overview of the OLCI level-1 per-pixel radiometric uncertainties and then provides details of the validation results of the uncertainty product using the OLCI-A&B uncertainty-normalised difference.
OLCI L1 Radiometric Uncertainties Validation
The Sentinel-3 satellites, S3A and S3B, have been respectively launched in 2016 and 2018, by the European Space Agency (ESA) as part of the European Commission’s (EC) Copernicus program, and have continuously acquired images of the Earth since then. Both satellites carry the Ocean and Land Colour Instrument (OLCI) dedicated to scanning the ocean and land surface to harvest information of Earth visible and NIR radiances returned from the Earth’s surface to the sensor. The quality of the OLCI data is ensured through rigorous calibration and validation activity, but also thanks to the provisioning of prognostic uncertainties on the measurements. Recently, a comprehensive model of the uncertainties associated with the level-1 radiometry of OLCI has been developed. This improvement will be transferred to operation later this year by the ESA. Following the GUM methodology, the OLCI level-1 per-pixel radiometric uncertainties have been estimated from the level-0 measurements by propagating all radiometric sources of uncertainty at each step of the global level-1 processing chain. At this point a major difficulty is to provide a validation of these uncertainties. For this purpose, we took advantage of the 6 months tandem phase at the launch of S3B. The validation of the new OLCI uncertainties has been performed using the relative intercomparison of OLCI-A and B uncertainties-normalised difference, a metrological methodology already applied to the SLSTR instrument (Hunt et al. 2020) and adapted here to the OLCI specificities. This paper presents an overview of the OLCI level-1 per-pixel radiometric uncertainties and then provides details of the validation results of the uncertainty product using the OLCI-A&B uncertainty-normalised difference.