USING MULTIVARIATE ANALYSES FOR SEPARATING SPATIAL AND TEMPORAL EFFECTS WITHIN SPECIES-ENVIRONMENT RELATIONSHIPS

We present a multivariate approach for the analysis of contingency tables involved in the study of species-environment relationships. The first table (species x sample) contains the abundance of p species collected in n samples. The second table (environmental variables x sample) contains values for q environmental variables measured in the n samples. The third table contains the indication of 'where' and 'when' samples were taken. In this paper we demonstrate: (1) how to match an environmental table and a faunistic table using co-inertia analysis; (2) how to take into account a spatial effect using between-class analyses; and (3) how to combine point (1) and (2) to determine the spatial structure shared by fauna and habitat. We illustrate such an approach hv using a set of hydrobiological data concerning 13 Ephemeroptera species and ten physical and chemical variables which were collected in the same site at the same dates in a small river of the Prealps.