Bidirectional and Spectral Reflectance Properties of the Cladina stellaris and Flavocetraria nivalis lichens and the Racomitrium lanuginosum moss

Abstract

In this thesis we present a scientific investigation of the spectral and bidirectional reflectance properties of two lichen species, <i>Flavocetraria nivalis</i> and <i>Cladina stellaris</i>, and one moss, <i>Racomitrium lanuginosum</i>. The spectral and bidirectional characterizations were based on high-quality laboratory measurements made in the European Goniometer facility, at the Joint Research Center, Ispra, Italy. We found that the estimated reflectance factors were within 2% of their mean value, after corrections for an intensity drift in the illumination source, and non-Lambertian properties of the reference panel. We found that the spectral profile of the <i>Racomitrium lanuginosum</i> was similar to that of dry or stressed green vegetation. Moistened lichen canopies exhibited a higher reflectance factor than dry lichens in the visible and near-infrared spectral bands. Moistened moss however, had a lower reflectance than dry moss. We compared our laboratory measurements with <i>in situ</i> spectral field measurements, and found an excellent agreement. Our goniometric measurements showed that the lichens are strongly backscattering canopies in both the visible and near-infrared parts of the spectrum. The moss however, scattered isotropically in the visible part of the spectrum, whereas it was strongly backscattering in the near-infrared part of the spectrum. We characterized our experimental bidirectional reflectance factors of the targets by means of a parametric and a physical model. By means of a statistical analysis, we showed that the parametric model was superior in terms of numerical convergence and flexibility. We applied cross-validation and bootstrap techniques to assess the accuracy of reflectance factor estimates. Lastly we demonstrated the possibility of simulating the top-of-the-atmosphere lichen reflectance factors based on laboratory measurements, and we explained how these results can be used to identify areas with a lichen cover of av Landsat-5 TM scene.