Assessment of FLD-based algorithms for the retrieval of vegetation solar-induced fluorescence from the in-filling of the telluric O2-A and O2-B lines

Abstract

This study evaluates the performances of different algorithms for the retrieval of solar induced fluorescence of vegetation in both the telluric O2-A and O2-B bands of the atmospheric molecular oxygen, respectively at 760 nm and 687 nm. In particular, we evaluated the performances of three algorithms amongst those already applied by the scientific community: two of them are based on the use of two or three spectral bands (sFLD and 3FLD methods), while the third one exploits the information content of all the spectral channels in certain bands by applying a polynomial model for fluorescence and reflectance (SFM method). These were applied to a synthetic set of fluorescence data corresponding to different types of vegetation. The main technical specifications of the spectroradiometer have been outlined in terms of three different airborne operating scenarios, addressing different flight altitudes and speeds chosen on the basis of typical platforms suitable for operation from low-medium altitudes. The results underline that the high spectral resolution of the instrument plays a fundamental role for the determination of the value of fluorescence with a good precision and accuracy, as expected. Nevertheless, the extraction of the value of fluorescence in the O2-A band is less critical than in the O2-B band and, specifically, it is less sensitive to the spectral resolution of the spectroradiometer. Even at low spectral resolutions, however, the retrieval algorithms based on polynomial fitting provided better results than methods based on the use of spectral bands.