A spatial approach to matching marine fish diversity and abundance with habitat features

Current issues in marine resource management have in common a geospatial component and a need to integrate both biotic and abiotic data from various sources. We propose a practical approach to address these issues looking at the American plaice (Hippoglossoides platessoides) and the demersal fish fauna in the Gulf of St. Lawrence (Canada). Central to our approach was the use of a common spatial grid and three different methods to match biotic and abiotic features at a broad regional scale, (1) matching plaice distribution with habitat categories determined a priori on the basis of abiotic features (cluster analysis), (2) habitat categories determined taking into consideration both plaice density and abiotic features (simple regression tree), and (3) habitat categories determined taking into consideration demersal fish species density (70 fish species) and abiotic features (multivariate regression tree, MRT). Hot spots and cold spots of plaice abundance in summer were described and matched with specific habitats. The spatial distribution of habitats was similar whether biotic variables were used in the classification or not. The MRT, however, identified 56 different fish species in the plaice habitat (median species richness by 100 km² cell = 12), pointing to potential interactions with other fish species.