Estimating forest diversity with hyperspectral data

Biodiversity conservation requires knowledge about species’ distribution patterns. Species richness and endemicity indices, calculated using species’ distribution data, catalyze conservation initiatives. Remotely sensed imagery is crucial for investigating biodiversity patterns. Discriminating plant species with imaging spectroscopy relies on lower spectral variation within individuals of a species than among species, whereas quantifying species richness capitalizes on spectral heterogeneity of a landscape. Here I estimate plant species richness and diversity using spectral clusters from Hyperion hyperspectral imagery of tropical forests sites, across an environmental gradient. I show that, in general, spectral clusters can function as proxies of plant species richness and diversity.