Close Window

Assessment of multi-scale optical and microwave imagery for mapping biodiversity surrogates at an Amazon floodplain site

Laura L Hess, University of California, Santa Barbara, lola@icess.ucsb.edu (Presenting)
Evlyn MLM Novo, INPE, evlyn@ltid.inpe.br
Adriana Gomes Affonso, INPE, affonso@ltid.inpe.br (Presenting)
Robert L Pressey, University of Queensland, rpressey@uq.edu.au
Ana LKM Albernaz, Museu Paraense Emilio Goeldi, anakma@museu-goeldi.br

The Amazon River floodplain, or varzea, is remarkable in its extent and in the spatial complexity of its habitats. Although species diversity for most taxonomic groups is lower in floodplain than in upland communities, rates of endemism are high. The ecological integrity of varzea habitats, and the traditional livelihood of its human inhabitants, are threatened by a variety of extractive activities. Since the current extent of protected areas is inadequate, additional reserve areas must be delineated. However, conservation planners face a severe lack of data on species distributions. Appropriate surrogates for biodiversity are needed as a basis for location and management of protected and sustainable use reserves. This study assessed the feasibility of using remote sensing datasets to delineate geographical units to serve as biodiversity surrogates. The surrogates, to be input to the C-Plan and MARXAN decision support systems for conservation planning, were proposed by a multi-disciplinary working group of scientists with expertise in the várzea environment. They include vegetation types, inundation extent and duration, extent of interface between várzea and upland, and catchment effects. Time series of high-resolution active microwave (JERS-1, RADARSAT, Envisat) and optical data (Landsat) were assembled and coregistered for a floodplain reach west of Manaus. A three-year time series of MODIS data was analyzed to determine the utility of these globally available products for mapping surrogate units.The potential for continental-scale mapping with ScanSAR products from the recently launched ALOS PALSAR was tested using JERS-1 scenes. The results show that integration of high-resolution optical and SAR data, and analysis of floodplain environments in conjunction with their catchments, allow optimal mapping of biodiversity surrogates for varzea environments. The lower spatial resolution of simulated ALOS ScanSAR resulted in lower accuracies in mapping floodplain units, but the scale and temporal frequency are appropriate for regional-scale surrogate mapping.

Science Theme:  LC (Land Use and Land Cover Change)

Presentation Type:  Poster

Abstract ID: 102

Close Window