Delineation of Coral and Sponge Significant Benthic Areas in Eastern Canada Using Kernel Density Analyses and Species Distribution Models

Published: 16-10-2018| Version 1 | DOI: 10.17632/hnp4xr2sy3.1
Ellen Kenchington,
Lindsay Beazley,
Camille Lirette,
Francisco Murillo,
Javier Guijarro-Sabaniel,
vonda wareham,
Kent Gilkinson,
Mariano Koen-Alonso,
Hugues Benoit,
Hugo Bourdages,
Bernard Sainte-Marie,
Margaret Treble,
Tim Siferd


Significant Benthic Areas are defined in DFO’s Ecological Risk Assessment Framework (ERAF) as “significant areas of cold-water corals and sponge dominated communities”, where significance is determined “through guidance provided by DFO-lead processes based on current knowledge of such species, communities and ecosystems”. Here we provide maps of the location of significant concentrations of corals and sponges on the east coast of Canada produced through quantitative analyses of research vessel trawl survey data, supplemented with other data sources where available. We have conducted those analyses following a bio-regionalization approach in order to facilitate modelling of similar species, given that many of the multispecies surveys do not record coral and sponge catch at species level resolution. The taxa analyzed are sponges (Porifera), large and small gorgonian corals (Alcyonacea), and sea pens (Pennatulacea). We applied kernel density estimation (KDE) to create a modelled biomass surface for each of those taxa, and applied an aerial expansion method to identify significant concentrations, following an approach first applied in 2010 to this region. We compared our results to those obtained previously. KDE uses only geo-referenced biomass data to identify “hot spots”. The borders of the areas so identified can be refined using knowledge of null catches and species distribution models that predict species presence-absence and/or biomass, both incorporating environmental data. We present such predictive models produced using a random forest machine-learning technique, and in one region compare the biomass random forest models for sea pens to those produced by generalized additive models (GAMs). Together, these distribution maps can be used to identify significant concentrations of corals and sponges in eastern Canada; an essential first step in the identification of Sensitive Benthic Areas.