Preparatory Studies for FLEX: Contributions to the radiometric calibration and the atmospheric characterization

Abstract

European Space Agency's (ESA's) 8th earth explorer mission, the Fluorescence Explorer (FLEX), aims to deliver the fluorescence signal emitted by terrestrial vegetation. It is planned to launch the satellite FLEX in 2025. FLEX will carry the high resolution Fluorescence imaging spectrometer (FLORIS) which has a spatial resolution of 300x300 m² and a swath width of 150 km. FLORIS will measure the complete fluorescence spectrum between 500 and 780 nm with a high spectral resolution between 0.3 and 3 nm. To disentangle the small fluorescence signal from the top-of-atmosphere (TOA) measurement, the sensor must be very sensitive and precise. Therefore, a good calibration and validation is necessary. Within this thesis, a validation strategy for FLORIS is developed by analysing pre-studies for the FLEX mission. Furthermore, an aerosol layer height retrieval is developed which can be extended to improve the atmospheric correction within the fluorescence retrieval. The Sentinel-3A and -3B tandem mission of 2018 was exploited to mimic the future tandem constellation of Sentinel-3 and FLEX. The spectral bands of the Ocean and Land Color imager (OLCI) on the Sentinel-3B satellite were shifted to a comparable band setting of FLORIS. This thesis presents a transfer function that enables the comparison of the radiance data of OLCI in FLEX configuration (OLCI-FLEX) and of OLCI in the nominal configuration. The transfer function overcomes the difference in spectral resolution by transferring information about the atmosphere and the surface from the spectral high resolution instrument to the lower resolution instrument. This information is used in a radiative transfer model to simulate OLCI-A measurement. The resulting simulated measurement at the nominal OLCI bands can be compared to the original OLCI-A measurement. The satellite-satellite comparison showed sensitivity to the known systematic measurement bias and also revealed processing errors. It was also studied how the satellite-satellite validation can be complemented by comparison of the satellite products with ground-based measurements. Ground-based instruments can be operated at low costs and with high accuracy. However, differences in spatial resolution result only in comparable signals if the surface is homogeneous. Thus, a measure for homogeneity of the surface was defined to identify the best possible ground site within a study scene by exploiting airborne measurements. The uncertainty due to differences in spatial resolution and spatial mismatch was also quantified. For the studied case, the validation uncertainty was small within the spectral range between 720 and 800 nm. The ground-based instrument should cover at least 13.5x13.5 m². The fluorescence retrieval is based on a spectral fitting method that exploits the well known structure of the oxygen absorption band at 760 nm. The depth of the absorption lines is determined by the surface pressure. Additionally, a present aerosol layer changes the depth of the lines depending on the height of the layer. Hence, the aerosol layer height is an important information for the fluorescence retrieval. A retrieval algorithm of the aerosol layer height based on OLCI measurements in the oxygen absorption band was developed. This algorithm was applied to two test cases over the ocean and for known aerosol types. The found height was comparable to results from a similar instrument, the Tropospheric Monitoring Instrument (TROPOMI). Both algorithms estimated the aerosol layer height too low compared to simultaneous measurements from an active instrument, the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP). To extent the OLCI algorithm also for land cases, the surface reflectance must be well characterized and the aerosol type must be estimated. All in all, this study contributes valuable information for the development of a successful validation strategy of FLEX. This strategy should include the validation of the top of atmosphere radiance data by a comparison of FLORIS and OLCI. Especially within the oxygen absorption band at 760 nm, the application of a transfer function is necessary. Large uncertainty sources of the validation with ground-based measurements were quantified and recommendations for the choice of sites are provided. Finally, a first retrieval of the aerosol layer height from OLCI could be presented. This information can be used to improve the fluorescence retrieval.