This thesis is based on seven papers describing studies conducted with samples from the Baltic Sea. It discusses some aspects on the use of multivariate methods for evaluation of data on hydrophobic organic compounds (HOCs) found in the aquatic environment. The HOCs studied are primarily polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs), and the matrices mainly samples of recent soft bottom sediments and settling particles.
The multivariate studies, performed in six of the seven papers, were all conducted using the pattern recognition (PARC) methods principal component analysis (PCA) and the classification method SIMCA. The PARC methods are applied on chemical profiles, i.e. the relative levels of compounds in a sample. Methods for pre-processing of data prior to PARC are also discussed.
The results are discussed from a chemical point of view concerning the chemistry of HOCs in terms of environmental behaviour, and also from an ecological perspective concerning their sources, transport and distribution. Most of the multivariate studies concern the spatial distribution of compounds of the same type. These studies can be divided into attempts to differentiate samples taken at or near sources of HOCs from background samples, and to study the distribution of HOCs in remote areas.
PARC was also used to differentiate between the distribution of different groups of HOCs (i.e. pesticides), and to show that the PAH profile changes with depth in a sediment, whereas the PCB profile does not.
Conclusively, the thesis shows that a combination of PARC methods is a versatile tool for different types of studies concerning HOCs in the aquatic environment.
A novel method for estimating the degree of resuspended sedimentary material found in sediment traps is also presented. The method uses stable carbon isotope values of individual PAHs as labels of sedimentary input to the traps.
Some aspects on the variations associated with sampling of sediments for HOC analysis are also discussed. This study clearly shows that the sampling procedure itself introduces most of the variation, followed by spatial variations in the sediments. The analytical procedure has the smallest effect on the variation.
Stockholm: Department of Analytical Chemistry, Stockholm University , 1998. , 39 p.
Härtill sju uppsatser.