| dc.description.abstract | Coralline algae constitute a cosmopolitan group of calcifying rhodophytes (red algae) known to form important marine ecosystems. Currently, detailed knowledge of their abundance and distribution is lacking in certain regions of their range. A characteristic trait of coralline algae, is their utilization of phycobiliproteins for light harvesting. Phycobiliproteins display prominent patterns of light absorption, and because they only are found in a few organism groups, coralline algae can be considered spectrally conspicuous. The spectrally conspicuous nature of coralline algae makes them potential candidates for optical remote sensing surveys, and recently, a novel optical remote sensing technique has entered the scene of marine research: underwater hyperspectral imaging (UHI). UHI is a form of imaging spectroscopy, where each image pixel is assigned its own light spectrum. The technique is thus capable of recording object-specific optical fingerprints, based on which supervised classification can be carried out. Supervised classification permits qualitative as well as quantitative mapping of biogeochemical objects of interest, which makes UHI suitable for assessing benthic environments.
The aim of the study was to characterize the spectral properties of different coralline algal species, and to assess the potential of UHI as a coralline algal identification and mapping tool. Four species of coralline algae were investigated: Corallina officinalis, Lithothamnion glaciale, Phymatolithon lenormandii, and Phymatolithon tenue. Important coralline algal pigments were identified using spectrophotometry and HPLC. Reflectance spectra of all species were obtained using both a spectrometer and UHI. Multivariate statistical analyses were performed on the reflectance data to identify spectral differences between species and instruments. In the end, supervised classification of coralline algae in UHI transects recorded both in vivo and in situ was carried out. R-phycoerythrin and chlorophyll a were found to be the most dominant coralline algal pigments. The analyzed species of coralline algae displayed highly similar reflectance spectra, and dips in reflectance corresponding to the absorbance peaks of R-phycoerythrin and chlorophyll a were identified in all spectra. Multivariate statistical analyses revealed that wavelengths corresponding to R-phycoerythrin light absorbance were the greatest contributors to interspecific spectral differences, but that the investigated coralline algal species could not be spectrally distinguished with great accuracy. Optical signatures recorded using different instruments were comparable, but inter-instrumental spectral differences were found to be greater than interspecific differences. This was likely a consequence of the distance between the sensor and the target organism being greater for UHI measurements than for spectrometer measurements. As could be expected based on the multivariate statistical analyses, supervised classification was unable to accurately distinguish between different coralline algal species in underwater hyperspectral imagery. As a group, coralline algae could however easily be identified, which suggests that UHI can be used for mapping coralline algal habitats. In the future, further studies on coralline algal spectral characteristics should be carried out. These studies should include more species, and feature larger sample sizes. Furthermore, large-scale UHI surveys of coralline algal habitats should be carried out using platforms such as remotely operated vehicles (ROVs) and autonomous underwater vehicles (AUVs) to enhance our understanding of this widespread and ecologically important organism group. | en |
