Prediction of Benthic Biotopes on the Norwegian Shelf and Slope Using Multivariate Analysis and Modelling

This study is part of the multidisciplinary seabed mapping programme MAREANO (Marine AREAdatabase for NOrwegian coast and sea areas). The mapping programme includes acquisition of multibeam bathymetry and acoustic backscatter data together with a comprehensive, integrated biological and geological sampling programme. The equipment used includes underwater video, boxcorer, grab, hyperbenthic sled, and beam trawl.

A procedure for producing maps of predicted biotopes is described that combined information on the distribution of biological communities with environmental factors and indicators. Detrended correspondence analysis (DCA) was used to relate bottom environment [including multiscale physical descriptors of the seabed derived from multibeam echosounder (MBES) data] and faunal distribution to find the best physical biotope descriptors. Analyses of large data sets tend to reveal general patterns with limited usefulness for management and poor improvement of the ecological understanding or species distribution. However, sequential analyses of subsets (identified with DCA) of the data provide a useful method for identifying biotopes at a local scale.  The biotopes were characterized by different compositions of species, substrata, depths, and values for terrain parameters. Prediction of biotope distribution was performed using a supervised GIS classification and maximum entropy distribution modeling (Maxent) with the MBES-derived physical seabed descriptors with the strongest explanatory ability (depth, backscatter, and broad-scale bathymetric position index) identified by the DCA. The species diversity of the identified biotopes was described from the content of the bottom samples.