Work

Multivariate analysis of ecotoxicological data using ordination: demonstrations of utility on the basis of various examples

Year: 2003
Type: other
Abstract: This paper describes the usefulness of the group of multivariate techniques belonging to ordination for the analysis of ecotoxi cological data sets. It is argued that although ecotoxicologists often gather multivariate data sets, they usually do not evaluate them w ith techniques that can handle multivariate data. Ordination techniques enable the researcher to extract an underlying structure out of a data set (eg. differences in composition of macro-invertebrate community between sites) and, if measured, relate this structure to explanator y variables (eg. concentrations of toxicants at the same sites). Five example data sets are presented to illustrate the underlying theory a nd the possibilities of ordination techniques. Two methods are presented, one based on weighted summation (eg. Principal Component Analysis, PCA) an d one on weighted averaging (eg. Correspondence Analysis, CA). These techniques differ in the shape of the modelled response (lin ear versus unimodal) as the type of data they model (absolute versus relative). Results of these two methods are illustrated using a data set comprising levels of different PCB congeners measured in the blood of Adelie penguins in three periods. After this the constrai ned forms of PCA and CA are discussed, ie. constrained means that they are able to optimally display the differences in species compositi on (here levels of PCB congeners) due to explanatory variables (here period). Further examples illustrate the use of covariables, contin uous and nominal explanatory variables, supplementary explanatory variables and forward selection of explanatory variables. Finally, two examples of Principal Response Curves (PRC) analyses are given. PRC is a technique that is especially developed to analyse time-series i n which a control or reference is present. The PRC results are discussed for a designed experiment and a monitoring data set. The paper ends with a discussion focussing on the comparison between ordination techniques and other multivariate techniques used in the field of e cotoxicology.
Language: English
Open Access status: green
OpenAlex ID: W2184244351
Cites: 43
Cited by: 138
Related to: 10