Hyperspectral Satellite Data for Modelling Spatial Beta Diversity Patterns of Birds along an Environmental Gradient
Leitão, Pedro J.¹; Suess, Stefan¹; Schwieder, Marcel¹; Milton, Edward J.²; van der Linden, Sebastian¹; Hostert, Patrick^1
1Geography Department, Humboldt-Universität zu Berlin, GERMANY; ²Geography and Environment, University of Southampton, UNITED KINGDOM

Human-driven reduction in biodiversity is widely acknowledged, with direct impact on ecosystem functioning and provisioning of services. However, existing patterns of biodiversity and most particularly those of community composition turnover, or beta diversity, are little known. While Earth observation missions provide an excellent tool for describing these patterns, the structural complexity of biotic communities is usually difficult to characterise using data from existing satellite sensors. Forthcoming hyperspectral missions will deliver much more detailed descriptions of the Earth's surface, which will greatly enhance our ability to tackle this issue. In the current study we used simulated EnMAP imagery, derived from geometrically and spectrally highly resolved airborne data from a region in southern Portugal. These data were used to describe the turnover of a bird community along an environmental gradient of shrub encroachment, resulting from land abandonment. For describing the turnover in community composition we adopted generalised dissimilarity modelling, while a sparse canonical correlation analysis enabled making full use of the hyperspectral information. The use of hyperspectral data, when compared to broadband multispectral data, such as Landsat TM, improved the explanatory power of the models by over 25%. Our results thus highlight the potential of hyperspectral satellite data for modelling the spatial patterns of biodiversity and ecosystem functioning. Nevertheless, further studies are still needed to validate the generalised usage of these type of data for tackling complex problems of ecosystem research.