Hyperspectral remote sensing of closed forest canopies: Estimation of chlorophyll fluorescence and pigment content

Quantitative assessment of vegetation physiological condition from remote sensing requires the development of methodologies to estimate biochemical constituents and physiological measures through the use of radiative transfer models. This dissertation provides a description of the investigations carried out to assess whether two indicators of stress in vegetation at leaf level, chlorophyll content and chlorophyll fluorescence, can be predicted using hyperspectral reflectance data. This study was focussed on twelve sites of Acer saccharum M. in the Algoma Region (Canada), where field measurements, laboratory-simulation experiments, and hyperspectral CASI imagery have been carried out in 1997, 1998, 1999 and 2000 campaigns. Radiative transfer theory and modelling assumptions are applied at laboratory and field scales in order to define the link between leaf reflectance and transmittance and canopy hyperspectral data.; Research work related to chlorophyll fluorescence estimation consisted of a series of laboratory and field measurements of spectral reflectance under artificial and natural light conditions which demonstrate that effects of natural chlorophyll fluorescence are observable in the reflectance red edge spectral region. Leaf samples from the study sites were used for reflectance and transmittance measurements with the Li-Cor Model 1800 integrating sphere apparatus coupled to an Ocean Optics Model ST1000 fibre spectrometer with and without fluorescence-exciting radiation. A study of the diurnal change in leaf reflectance spectra, combined with fluorescence measurements with the PAM-2000 Fluorometer show that the difference spectra are consistent with observed diurnal changes in fluorescence. Small canopies of Acer saccharum M. were used for measurements in the laboratory and under natural light conditions in diurnal trials, in which the variation of measured reflectance was shown experimentally to be consistent with a fluorescence signature imposed on the inherent leaf reflectance signature. The Fluorescence-Reflectance-Transmittance (FRT) leaf radiative transfer model was developed to theoretically simulate the effects of fluorescence emission on the apparent reflectance. Optical indices were identified to track fluorescence emission effects on the observations of apparent reflectance at the leaf and canopy level.; Infinite reflectance and canopy reflectance models were used to link leaf to canopy levels through radiative transfer simulation. Optically thick simulation formulae and SAILH and MCRM canopy reflectance models were used for chlorophyll content estimation by scaling-up and by numerical model inversion approaches through coupling to PROSPECT leaf radiative transfer model. Study of the merit function showed that red edge optical indices perform better than single spectral reflectance channels from hyperspectral airborne CASI data, and effects of shadows and LAI variation were minimized. Estimates of leaf pigment by hyperspectral remote sensing of closed forest canopies were shown to be feasible with root mean square errors ranging from 3 to 5.5 μg/cm 2. MERIS sensor onboard ENVISAT (to be launched in 2001) is shown to be a satellite sensor with the potential to provide global mapping of foliar pigment content.