Using remote sensing to explore the spectral and spatial characteristics of wetland vegetation.

Wetlands play an important role as ecotones between terrestrial and aquatic habitats and, as a result, represent an environment of high biodiversity and important hydrological function. Ecological understanding in these environments is hampered by difficult terrain and the dynamic and heterogeneous nature of the vegetation. Remote sensing can provide large amounts of contemporaneous data quickly, objectively and over large areas. This study utilises remote sensing data in conjunction with field data and habitat maps derived from traditional ecological surveys to investigate the use of remote sensing as a tool to aid the ecological understanding and monitoring of wetland environments. This study investigated three main objectives; the first two involved the use of field spectrometry from six habitat types in a freshwater wetland in the north of Scotland. Multivariate analyses demonstrated the possibility of distinguishing between these habitat types using field spectra alone. Detailed vegetation datasets were also collected and the relationship between these and variation in the associated spectra was investigated. Significant relationships were established between ordination axes and spectral bands in the green and NIR regions of the spectrum. Results also demonstrated the potential for remote sensing data to characterise the nature of habitat boundaries. The third objective involved the use of airborne imagery to classify remote sensing data into ecologically meaningful classes. Classification accuracies of over 70% were obtained. Work over the last decade has seen a bridging of the relationship between remote sensing and ecology although it is widely acknowledged that our ecological understanding of the remote sensing-vegetation relationship is still limited at many scales and in many ecosystems, not least the wetland environment. This study provides a much needed basis to research in this cross-disciplinary field and identifies further areas that would benefit from future work