Kalacska, MargaretSanchez-Azofeifa, G ArturoRivard, BenoitCaelli, TerryWhite, H PCalvo-Alvarado, JC2015-12-070034-4257http://hdl.handle.net/1885/28416We evaluated the use of EO-1 Hyperion hyperspectral satellite imagery for mapping structure and floristic diversity in a Neotropical tropical dry forest as a way of assessing a region's ecological fingerprint. Analysis of satellite imagery provides a means to spatially appraise the dynamics of the structure and diversity of the forest. We derived optimal models for mapping canopy height, live aboveground biomass, Shannon diversity, basal area and the Holdridge Complexity Index from a dry season image. None of the evaluated models adequately estimated stem or species density. Due to the dynamic nature of the leaf phenology we found that for the application of remote sensing in Neotropical dry forests, the spectro-temporal domain (changes in the spectral signatures over time-season) must be taken into account when choosing imagery. The analyses and results presented here provide a means for rapid spatial assessment of structure and diversity characteristics from the microscale site level to an entire region.Keywords: Biomass; Forestry; Imaging techniques; Neural networks; Satellite communication systems; Wavelet decomposition; Holdridge Complexity Index; Hyperion structure; Hyperspectral remote sensing; Tropical dry forest; Ecosystems; aboveground biomass; biodiversit Biomass; Costa Rica; Holdridge Complexity Index; Hyperion; Hyperspectral remote sensing; Neural network; Structure; Tropical dry forest; Wavelet decomposition200710.1016/j.rse.2006.11.0072015-12-07