Enhancement-classification and spectral mixture analysis of caribou lichen habitats, northern Quebec, Canada

Abstract

We present a new approach for the detection and mapping of lichen in a heterogeneous habitat of northern Quebec using Landsat imagery. Results from the enhancement-classification method (ECM) and from spectral mixture analysis (SMA) were compared and evaluated for their potential to be applied over large territories. ECM is based on guided unsupervised classification of enhanced satellite images, and provided an overall accuracy of 74.5% and a good discrimination between lichen and non-lichen habitat. However, ECM discrimination between lichen habitats was more problematic. SMA is based is based on the fact that each pixel contains different surface features, each of which contributes to the overall pixel-level signal received by a remote sensor. SMA quantifies the spectral contributions from individual scene components to retrieve their sub-pixel scale proportions. Image SMA fractions showed a good agreement with field validated fractions. Tho two methods were efficient for mapping lichen. However, SMA provided new information not available by classification methods and therefore is recommended for further application to larger areas and spatio-temporal studies.