Ecological applications of remote sensing data in neotropical rainforests

Abstract

Understanding species' distributions is a central theme of biodiversity studies. A combination of data derived from moderate and high spectral resolution satellite imagery (vegetation indices and hyperspectral narrow bands, respectively) was used to address questions regarding tree species' distributions, vegetation phenology, and influences on bird seasonal movements in tropical rainforests. Vegetation indices were used in ecological niche modeling to predict movement patterns of a tropical canopy frugivorous bird in Central America: the predicted distributions generally recovered observed non-breeding ranges, but estimated lowland areas for the breeding range, which is restricted to middle elevations. Hyperspectral imagery provided sufficient spectral information to discriminate crowns of five different tree taxa that represent food resources for macaws and peccaries in southeastern Peru. Tree spectra showed significant temporal variation, suggesting that it is possible to study tree phenology remotely. Current and future developments of remote sensing techniques permit regional studies of ecosystem functions and structure.