Simulation of the reflectance of ground vegetation in sub-boreal forests

Abstract

To study the reflectance spectra of ground vegetation in forests a series of field measurements was performed by a GER-2600 spectrometer in Estonian and Swedish sub-boreal forests. The measured reflectance spectra of grasses, regenerating trees, mosses and dwarf shrubs were analysed with a two-layer model of vegetation canopy reflectance. By the inversion of the two-layer model a set of model parameters was estimated for several types of understorey vegetation. The estimated sets of understorey parameters can be used to reproduce the spectral signatures of the respective types of ground vegetation which are needed in simulating the reflectance spectra of sub-boreal and boreal forests. Principal component analysis showed that at least six biochemical components are needed to reproduce the spectral signatures of understorey vegetation in the spectral range of 400-2400 nm. Some vegetation indices (VI) which are in use in remote sensing are rather well correlated with canopy parameters estimated in the inversion. At the same time several VIs have very low sensitivity and/or almost random behavior.