Multiangular observation of canopy sun-induced chlorophyll fluorescence by combining imaging spectroscopy and stereoscopy

Abstract

The effect that the canopy structure and the viewing geometry have on the intensity and the spatial distribution of passively measured sun-induced chlorophyll fluorescence at canopy scale is still not well understood. These uncertainties constrain the potential use of fluorescence to quantify photosynthesis at this level. Using a novel technique, we evaluated the diurnal changes in the spatial distribution of sun-induced fluorescence at 760 nm (F760) within the canopy as a consequence of the spatial disposition of the leaves and the viewing angle of the sensor. High resolution spectral and stereo images of a full sugar beet canopy were recorded simultaneously in the field to estimate maps of F760 and the surface angle distribution, respectively. A dedicated algorithm was used to align both maps in the post-processing and its accuracy was evaluated using a sensitivity test. The relative angle between sun and the leaf surfaces primarily determined the amount of incident Photosynthetic Active Radiation (PAR), which in turn was reflected in different values of F760, with the highest values occurring in leaf surfaces that are perpendicularly oriented to the sun. The viewing angle of the sensor also had an impact in the intensity of the recorded F760. Higher viewing angles generally resulted in higher values of F760. We attribute these changes to a direct effect of the vegetation directional reflectance response on fluorescence retrieval. Consequently, at leaf surface level, the spatio-temporal variations of F760 were mainly explained by the sun-leaf-sensor geometry rather than directionality of the fluorescence emission. At canopy scale, the diurnal patterns of F760 observed on the top-of-canopy were attributed to the complex interplay between the light penetration into the canopy as a function of the display of the various leaves and the fluorescence emission of each leaf which is modulated by the exposure of the individual leaf patch to the incoming light and the functional status of photosynthesis. We expect that forward modeling can help derive analytical simplified skeleton assumptions to scale canopy measurements to the leaf functional properties.