Spectral reflectance of carbonate sediments and application to remote sensing classification of benthic habitats
Date of Award
Doctor of Philosophy (Ph.D.)
Marine Biology and Fisheries
First Committee Member
Pamela Reid - Committee Chair
Remote sensing is a valuable tool in marine research that has advanced to the point that images from shallow waters can be used to identify different seafloor types and create maps of benthic habitats. A major goal of this dissertation is to examine differences in spectral reflectance and create new methods of analyzing shallow water remote sensing data to identify different seafloor types quickly and accurately. Carbonate sediments were used as a model system as they presented a relatively uniform, smooth surface for measurement and are a major bottom type in tropical coral reef systems. Experimental results found that sediment reflectance varied in shape and magnitude depending on pigment content, but only varied in magnitude with variations in grain size and shape. Derivative analysis of the reflectance spectra identified wavelength regions that correlate to chlorophyll a and chlorophyllide a as well as accessory pigments, indicating differences in microbial community structure. Derivative peak height also correlated to pigment content in the sediments. In remote sensing data, chlorophyll a, chlorophyllide a, and some xanthophylls were identified in derivative spectra and could be quantified from second derivative peak height. Most accessory pigments were attenuated by the water column, however, and could not be used to quantify pigments in sediments from remote sensing images. Radiative transfer modeling of remote sensing reflectance showed that there was sufficient spectral variation to separate major sediment types, such as ooid shoals and sediment with microbial layers, from different densities of seagrass and pavement bottom communities. Both supervised classification with a spectral library and unsupervised classification with principal component analysis were used to create maps of seafloor type. The results of the experiments were promising; classified seafloor types correlated with ground truth observations taken from underwater video and were comparable to existing maps of seafloor type. Creation of accurate seafloor type maps is an important step in constructing maps of benthic habitats.
Biology, Oceanography; Remote Sensing
Louchard, Eric Michael, "Spectral reflectance of carbonate sediments and application to remote sensing classification of benthic habitats" (2003). Dissertations from ProQuest. 1993.