The development and application of a charge-coupled device based instrument for at-site monitoring of algae and cyanobacteria in freshwaters
The research presented in this thesis describes the development and application of a portable, high-resolution instrument, specifically designed for the at-site monitoring of algae and cyanobacteria in freshwaters. The instrument incorporates a miniature charge-coupled device (CCD) based spectrometer and a low power combined deutelium and tungsten light source, enabling the absorbance to be measured between 200 - 850 nm at a resolution of 1.3 nm. A transmission dip probe with removable tips of 5, 10 and 40 mm pathlengths forms the sampling device. A specifically developed control program allows easy operation of the instrument. A linear response from 0.0 - 1.2 AU and a combined signal to noise ratio of 576: 1 for the instrument components resulting in a high baseline stability of 1.0 mAU drift over five hundred measurements being observed. The instrument provides in-vivo absorbance characteristics with high resolution across the visible spectrum. Up to twelve specific spectral features were commonly identified in the absorbance spectra of algae and cyanobacteria between 400 - 750 nm. Individual spectral features were linked to specific pigments, some of which were found to be taxonomically distinct. Fourth derivative analysis was proven to provide further enhancement of subtle spectral features. The instrument has a linear range for chlorophyll a up to 1000 !lg rl and a detection limit of 8 )lg rl using the 40 mm pathlength probe. Physiological adaptation to light and nutrient conditions were shown to have a significant effect on the in-vivo absorbance spectrum, therefore providing potential information on physiological status and health of a natural sample. Spectral analysis using principal component analysis (PCA) with classification based on the soft independent modelling of class analogy (SIMCA) method was used to classify nine species from three taxonomic classes, including four cyanobacteria (Microcystis aeruginosa, Anabaena variabilis, Aphanizornenon flos-aquae, Synechnococcus sp.), four chlorophyceae (Chlorella vulgaris, Scenedesmus acuminatus, Spirogyra mirabilis, Staurastrurn chaetoceros) and a single bacillariophyceae (Asterionella Formosa). Classification using the SIMCA method proved to be highly reliable and robust. Moreover, the addition of noise was found to have very little effect on the classification. Under laboratory conditions all nine species were correctly classified using 'unknown' spectra. At-site classification of natural samples and laboratory simulations have shown the robustness and reliability of the developed portable instrument. In combination with the data analysis techniques, the instrument is well suited to the proactive at-site assessment of algal and cyanobacterial blooms in eutrophic freshwater environments.