Identification of phytoplankton taxonomic groups using spectral absorption characteristics
The aim of the work described here is to use the optical absorption characteristics of phytoplankton to identify the taxonomic groups present in seawater. The spectral absorption of phytoplankton is determined by the pigment assemblage within each cell and taxonomic groups can have specific pigment combinations. Each individual pigment has distinct absorption characteristics, therfore phytoplankton communities containing different pigment assemblages should have absorption spectra that have distinctive characteristics. An initial investigation into pigment distribution within the taxonomic classes reveals that the pigments fucoxanthin and 19'hexanoloxyfucoxanthin which are used as speficic biomarkers for diatoms and coccolithophores respectively are present in other taxonomic groups and so can not be considered to be indicative of diatoms of coccolithophores alone The chlorophyll a to accessory pigment ratios within taxonomic groups were also investigated and found to be inconsistent, varying due to light and nutrigent conditions. This is important as the pigment assemblage and relative concentration will have an effect on the ability to distinguish between the absorption spectro of phytoplankton. The absorption spectra of pure culturers of six different phytoplankton taxonomic groups were measured using the filter pad technique. Investigation of the extent to which the pigment assemblage needs to differ for the absorption sepctra to be statistically different, revealed that phytoplankton containing pigment assemblages that differ only in the chlorophyll a to major accessory pigment ratio have absorption spectra that can be distinguished. Fourth derivative analysis was used to locate the in vivo absorption peaks on the absorption spectra, multiple regression analysis was then used to identify the pigments responsible. Only one absorption peak located at 650nm provides a phytoplankton taxonomic identification, of CHLb containing chlorophytes. A spectral library was constructed using the absorption spectra of each of the six different taxonomic groups. Multiple regression analysis performed between the spectral library and the absorption spectra of mixed phytoplankton samples, was successful in identifying the pigments present in samples providing fucoxanthin was not present. These analysis techniques were applied to natural samples collected from the Atlantic Meridional Transect (AMT) cruise 10 in May 2000. This transect encompasses several different oceanographic provinces, Tropical gyres, Upwelling and Shelf waters which were experiencing a spring bloom. The absorption of the phytoplankton communities in each of these zones showed varied characteristics. The fourth derivative method identified high chlorophyll b containing areas. Comparison of the spectral library with the absorption spectra of eachprovince correctly predicted the pigment assemblage. From this the taxonomy was estimated, this concurred the microscopy analysis, with exception to the pigment HEX which was misleading predicting a dominance of coccolithophores, when other taxonomic groups were responsible for the presence of this pigment. Conclusions drawn from this investigation are that pigments FUCO and HEX can not any longer be considered to be biomarker pogments for diatoms and coccolithophores. It is possible to identify the presence of chlorophyll a, b, c and total carotenoids from fourth derivative analysis. Of these pigments, only thepresence of chlorophyll b is useful in identifying the presence of a phytoplankton taxonomic group. It is possible with limitations to identify the taxonomic groups in a sample by regression of the absorption shape with a library of pure phytoplankton absorption spectra, although this is only an indication of the dominant pigments not phytoplankton biomass.