The NDVI and spectral decomposition for semi-arid vegetation abundance estimation

Abstract

This paper describes the use of the Normalized Difference Vegetation Index (NDVI) and spectral decomposition techniques to estimate vegetation abundance using ground-based spectroradiometric data of semi-arid vegetation. It is found that the NDVI provides a better measure of the proportion of directly irradiated leaves within the field-of-view of the spectroradiometer than it does the Leaf Area Index or biomass. Moreover, it is found that spectral decomposition isolates a factor which is strongly influenced by spectral variation at the region of the red edge. This factor is highly correlated with the NDVI -(R2=0.91) and as such also provides a good estimate of the proportion of directly irradiated leaves. It is suggested that spectral decomposition techniques provide a unique framework in which to analyse the factors affecting the spectral response of vegetation.