Abstract
PARASOL is a satellite launched in December 2004 and providing multidirectional and polarized observation of the Earth reflectances for the visible and near infrared spectral range. Acquisitions are performed by a POLDER-like instrument which has no on-board calibration device. Despite the fact that there is no on-board reference, it was possible to assess a refined in-flight radiometric (and geometric) characterization of the instrument through various calibration techniques based on acquisitions over natural targets. This result was possible through the use of various types of calibrations sites, both in term of spectral behaviour or reflectance magnitude, such as Rayleigh scattering, sunglint, desert sites (in Africa and Arabia), clouds (DCC), and snowy sites (Domes over Antarctica).
As for most of the on-orbit sensors (if not all), temporal changes appear on the radiometric calibration. This evolution is not the same for all the spectral bands. In addition, these evolutions were identified as non-uniform inside the field-of-view.
We report on this paper how a physical model was proposed to describe the observed phenomenon. This model initially simply considers as parameters a time constant and amplitude for each spectral band. Based on an 8 years archive of acquisitions over oceanic sites, desert sites, DCC, and Antarctica, parameters of the model were adjusted by an iterative process in order to best fit these various sets of data from multiple calibration methods. This analysis reveals that a unique time constant is sufficient to describe the behaviour observed for all spectral bands, and consequently, after fixing this “instrumental” time constant, only the amplitude has to be estimated for each band.
After a brief reminding of each calibration approaches, the adjustment of this simple model will be illustrated for various spectral bands. This comparison also provides a good overview of the accuracy that can be expected from each method in term of temporal monitoring.
For the in-field of view calibration, also called multi-angular calibration, a combination of results from the various methods will be detailed, considering their respective advantages and drawbacks. The same kind of physical model was used to allow a better estimation of this in-field of view calibration. The model will be presented as well as the resulting parameters.
This in-flight characterization of the radiometric trending based on the full 8-years archive will be considered as inputs for the coming PARASOL end-of-life recalibration.
Physical Model to Describe the PARASOL Radiometric Trending: Definition, Adjustment, and Validation
PARASOL is a satellite launched in December 2004 and providing multidirectional and polarized observation of the Earth reflectances for the visible and near infrared spectral range. Acquisitions are performed by a POLDER-like instrument which has no on-board calibration device. Despite the fact that there is no on-board reference, it was possible to assess a refined in-flight radiometric (and geometric) characterization of the instrument through various calibration techniques based on acquisitions over natural targets. This result was possible through the use of various types of calibrations sites, both in term of spectral behaviour or reflectance magnitude, such as Rayleigh scattering, sunglint, desert sites (in Africa and Arabia), clouds (DCC), and snowy sites (Domes over Antarctica).
As for most of the on-orbit sensors (if not all), temporal changes appear on the radiometric calibration. This evolution is not the same for all the spectral bands. In addition, these evolutions were identified as non-uniform inside the field-of-view.
We report on this paper how a physical model was proposed to describe the observed phenomenon. This model initially simply considers as parameters a time constant and amplitude for each spectral band. Based on an 8 years archive of acquisitions over oceanic sites, desert sites, DCC, and Antarctica, parameters of the model were adjusted by an iterative process in order to best fit these various sets of data from multiple calibration methods. This analysis reveals that a unique time constant is sufficient to describe the behaviour observed for all spectral bands, and consequently, after fixing this “instrumental” time constant, only the amplitude has to be estimated for each band.
After a brief reminding of each calibration approaches, the adjustment of this simple model will be illustrated for various spectral bands. This comparison also provides a good overview of the accuracy that can be expected from each method in term of temporal monitoring.
For the in-field of view calibration, also called multi-angular calibration, a combination of results from the various methods will be detailed, considering their respective advantages and drawbacks. The same kind of physical model was used to allow a better estimation of this in-field of view calibration. The model will be presented as well as the resulting parameters.
This in-flight characterization of the radiometric trending based on the full 8-years archive will be considered as inputs for the coming PARASOL end-of-life recalibration.