USE OF SPECTRAL ANALOGY TO EVALUATE CANOPY REFLECTANCE SENSITIVITY TO LEAF OPTICAL-PROPERTIES

The spectral variation Of canopy reflectance is mostly governed by the optical properties of the elements such as the leaves. Since leaf intrinsic scattering properties show very little spectral variation, leaf optical properties are related to their absorption properties. Spectral analogies are thus observed between two wavelengths for which the optical properties (absorption, reflectance, or transmittance) of the elements are similar. The red edge for green plants shows the full range of variation of leaf optical properties. The relationship between canopy reflectance and leaf reflectance measured concurrently at the red edge over sugar beet canopies was thus used to simulate canopy reflectance over the whole spectral domain from leaf reflectance spectra measured over the whole spectral domain. The results show that the spectral analogies found allows accurate reconstruction of canopy reflectance spectra. Explicit assumptions about the very low spectral variation of leaf intrinsic scattering properties are thus indirectly justified. The sensitivity of canopy reflectance (rho(c)) to leaf optical properties is then investigated from concurrent spectral variations Of canopy (partial derivative rho(c)/partial derivative lambda) and leaf reflectance (partial derivative rho1/partial derivative lambda): partial derivative rho(c)/partial derivative rho1 = (partial derivative rho(c)/partial derivative lambda) (partial derivative rho(l)/partial derivative lambda)-1. This expression is strictly valid only when the optical properties of the soil background or of the other vegetation elements such as bark are either spectrally flat or do not contribute significantly to canopy reflectance. Simulations using the SAIL and PROSPECT models demonstrate that the sensitivity of canopy reflectance to leaf reflectance is significant for large vegetation cover fractions in spectral domains where absorption is low. In these conditions, multiple scattering enhances the leaf absorption features by a factor that can be greater than 2.0. To override the limitations of the SAIL model for the description of the canopy architecture, we tested the previous simulation results on experimental data. Concurrent canopy and leaf reflectance spectra were measured for a range of sugar beet canopies. The results show good agreement with the theoretical findings. Conclusions are drawn about the applicability of these findings, with particular attention to the potential detectability of leaf biochemical composition from canopy reflectance sensed from space.