Alternate title

VI005009

Date (Creation)

2017-12-31

Citation identifier

94672fe5-4213-4722-817b-5fdb33638ba2

Point of contact

Organisation name	Individual name	Electronic mail address	Role
NOAA/NCCOS		bryan.costa@noaa.gov	Owner

GEMET - INSPIRE themes, version 1.0

• Oceanographic geographical features
• Habitats and biotopes

GEMET - Concepts, version 4.1.3

• habitat
• biotope

Access constraints

Other restrictions

Other constraints

no limitations on public access

Use constraints

Other restrictions

Other constraints

Available under the Creative Commons Attribution (CC-BY) License v4.0

Use constraints

Other restrictions

Other constraints

Costa, Bryan M. et al. (2017). Benthic habitat maps for the insular shelf south of St. Thomas and St. John. https://doi.org/10.7289/v5/tm-nos-nccos-241

Spatial representation type

Vector

Language

English

Topic category

• Biota
• Oceans

Begin date

2017-12-31

Reference system identifier

http://www.opengis.net/def/crs/EPSG/0/4326

Distribution format

Name	Version
Esri Shapefile

OnLine resource

Protocol	Linkage	Name
OGC:WMS	https://ows.emodnet-seabedhabitats.eu/geoserver/emodnet_view_maplibrary/wms	vi005009
OGC:WFS	https://ows.emodnet-seabedhabitats.eu/geoserver/emodnet_open_maplibrary/wfs	vi005009
WWW:DOWNLOAD-1.0-http--download	https://files.emodnet-seabedhabitats.eu/data/EMODnetSBHsurvey_VI005009.zip	EMODnet Seabed Habitats VI005009 Download

Hierarchy level

Dataset

Statement

This map has been formatted by the EMODnet Seabed Habitats consortia according to Data Exchange Format v4.1. In total, 652 km2 of seafloor were characterized on the insular shelf south of St. Thomas and St. John. These habitat map products are ‘pixel-based,’ developed from 11×11 meter resolution raster images and mathematical modeling techniques called boosted regression and classification trees. These modeling techniques generate continuous spatial predictions by finding relationships between the geographic distribution of habitats (extracted from 1,005 underwater videos) and environmental conditions that may be influencing or correlated with these distributions (extracted from 20 raster images describing the oceanography, geography, and seafloor topography of the area). The final habitat map products describe the probability of occurrence and associated precision for four individual substrata (i.e., ‘Coral Reef’, ‘Pavement’, ‘Rhodoliths’, and ‘Sand’) and two biological cover types (i.e., ‘Live Hard Coral’ and ‘Live Soft Coral’). These six layers were also combined to create a single, composite benthic map with five habitat classes including: ‘Coral reef colonized with live coral’, ‘Rhodoliths with macroalgae’, ‘Bare sand’, ‘Rhodoliths with macroalgae and bare sand’, and ‘Pavement colonized with live coral.’ The performance and thematic accuracy of these products were evaluated using an independent set of underwater videos (n=348). The performance of the habitat models was considered to be good to outstanding based on four quantitative metrics, including percent deviance explained (x̄ =37.4% ±16.2 SE), area under the receiver operating characteristic curve (x̄ =0.86 ±0.05 SE), mean error (x̄ =-0.04 ±0.03 SE) and root mean square error (x̄ =0.33 ±0.04 SE). The overall accuracy and tau value for the composite habitat map was high at 85.6% and 0.82 ±0.05, respectively.

Metadata

File identifier

94672fe5-4213-4722-817b-5fdb33638ba2 XML

Metadata language

English

Character set

UTF8

Hierarchy level

Dataset

Date stamp

2025-09-18T08:10:40.547882Z

Metadata author

Organisation name	Individual name	Electronic mail address	Role
South Atlantic Environmental Research Institute		datamanager@saeri.ac.fk	Point of contact