Identification of ecological structure and species relationships along an oceanographic gradient in the gulf of maine using multivariate analysis with bootstrapping

Ecosystem-based fisheries management requires a fundamental understanding of ecosystem boundaries and interspecies interactions. We report a delineation of fish and invertebrate data collected by trawl survey along the Maine, USA, coast. Principal components analysis (PCA) reduced the multidimensionality of the data and created new variables from correlations among species. Bootstrapped PCA was employed to assess PCA structure using eigenvalue variation and species associations using eigenvector variation. A general linear model related structure of fish community identified in PCA to depth, temperature, longitude, and interactions among these variables. Generally, alongshore and onshore-offshore assemblage patterns related to oceanographic gradients, with seasonal variation. PCA-created variables act as indicators of biodiversity and are related to the scale of observation, allowing for multiple scales to be integrated if data are available. Species targeted and gear used along with spatial extent and sampling density must be considered when management zonation is to be undertaken but would allow boundaries to follow biological and physical gradients rather than relying on the typical political and economic variables, avoiding ecologically deleterious spatial patterns in fishing pressure or other instances of de facto zoning. Regardless, this study describes some contributing factors causing division of species assemblages that should be considered.