Mangrove ecosystems are in serious decline around the world and various initiatives are underway to assess their current coverage and loss in cover. These ecosystems occur as thin strips along coastlines or rivers and, due to the strong environmental gradients present, mangroves show high spatial variability along short transects. Remote sensing tools that offer high spatial resolution mapping and high information content are needed to provide good differentiation of the various mangrove zones and types. The added complexities of tropical atmospheric conditions provide further challenges in terms of the selection of sensors and image analysis methodologies. This paper explores the possibility of combining a high spatial/spectral resolution scanner, 'Compact Airborne Spectrographic Imager' (CASI), with the airborne National Aeronautics @ Space Administration's polarimetric radar, 'AIRSAR', for mapping and monitoring of mangrove estuaries. The Daintree River estuary in far North Queensland, Australia was chosen for this study due to its diversity of mangrove species. Imagery acquired by both the CASI airborne scanner (14 bands, 2.5 m pixel) and the AIRSAR (L- and P-band polarimetry, C-band interferometry, 10 m pixel) has been used to produce detailed maps of the mangrove zones in the estuary. The advantages and difficulties associated with multi-source data integration are investigated in this paper. While radar provides general structural information in relation to mangrove zonation, high-resolution hyperspectral scanners allow for finer-detailed analysis and green-biomass information. Classifications (maximum likelihood) of both the individual and integrated datasets are performed, with the latter producing more accurate results. Application of a hierarchical neural network classification is also explored, where the more general mangrove zones are separated first based on structural information, then species or specie-complexes are extracted in subsequent levels using spectral differences.
22 May 2007 in Aquatic Conservation: Marine and Freshwater Ecosystems
A full understanding of radar backscatter from urban areas is necessary in order to develop a robust methodology for monitoring and classifying urban characteristics using remotely sensed Synthetic Aperture Radar images.This paper examines the dominant backscattering mechanisms such as single bounce from roofs, double bounce from wall-ground structures and possibly triple bounce from wall-wall-ground structures, and their relative contributions to the backscatter. With the use of quad-polarized image data such as those acquired by the NASA/JPL AirSAR system, the backscatter can be decomposed into components caused by different backscattering mechanisms, offering a promise for urban monitoring and classification.