Multivariate dissemination of species relationships for use in marine spatial planning

Employing ecological approaches to fisheries management or comprehensive marine spatial planning requires that species assemblage structure be accounted for. Fish and invertebrate spatial distributions from the National Marine Fisheries Service demersal trawl survey conducted in the Georges Bank - Gulf of Maine region were analyzed by bootstrapped principal component analysis (PCA_Btsp) and normal PCA (PCA_Nrml). PCA_Btsp produced confidence limits for eigenvalue stopping rules and for eigenvectors to identify significantly correlated species. Stopping rules identified the first six principal components (PCs) as relevant. Initially, summer and fall survey data from 1963 to 2006 were analyzed, but high eigenvector variation led to reductions in the species and time series used. Confidence interval variation was achieved through removal of highly migratory species and restriction of the time series. PC scores were mapped using inverse distance weighted interpolation to reveal multispecies spatial arrangements. Core areas of species groupings and overlapping zones of higher diversity can be delineated and, even under high fishing pressure with large compositional changes, the assemblages maintained robust spatial organization. This spatial organization could be employed to protect appropriate species groups and minimize bycatch. Careful analysis of survey data can help ensure area-based management schemes are consistent with ecological scales.