Aim: The Deepwater Horizon disaster resulted in the largest accidental marine oil spill in history and caused extensive injury to deep-sea habitats, including cold-water coral communities dominated by Paramuricea species. One of the primary difficulties in assessing the full extent of the injury to cold-water coral ecosystems is the extreme paucity of observational data and the subsequent lack of knowledge of their distribution within the affected region. The aim of this study was to use habitat suitability modeling to estimate the number of potentially affected Paramuricea sp. corals across the northern Gulf of Mexico. Location: Northern Gulf of Mexico Taxon: Cold-water corals in the genus Paramuricea Methods: High-resolution (12.5 m) models were built using the maximum entropy (Maxent) approach using remotely sensed data including seafloor topography, seismic reflectivity, temperature, and the amount of productivity exported from the surface. Model outputs were used to estimate the number of potential coral sites in the northern Gulf of Mexico, delineated as areas with both high habitat suitability scores and the presence of hard substrate. The number of coral sites was further adjusted using a ground-truthing procedure using autonomous underwater vehicle transect data. Results: Across the entire study area in the northern Gulf of Mexico, there were 558 predicted coral sites, covering an area of 14.2 km². Within a 2,291 km² region shown to have been directly affected by the spill and subsequent oil plume, there were 66 predicted coral sites covering an area of 1.2 km² with an average of 63 corals per site. Main Conclusions: Our results indicate that the magnitude of injury stemming from the spill was likely far higher than previously known, and will help quantify the full extent of the losses incurred as well as prioritize disturbed areas for future research and restoration efforts.

This is the final habitat suitability model for Paramuricea spp. in the northern Gulf of Mexico. The raster provides the predicted habitat suitability score for each grid cell. The habitat suitability score is a relative score between 0 and 1, with 1 indicating the most suitable habitat and 0 indicating the least suitable habitat. The model was built using Maxent as a presence-only model using the following environmental variables at a scale of 12.5 m: terrain metrics (northness, eastness, topographic position index, plan curvature, and rugosity), temperature, and particulate organic carbon. The model was trained and tested using 245 spatially unique occurrences.  Please see the methods section of the associated publication for full methods behind the model creation, testing, and subsequent analyses. 

The raster is a compressed ASCII file and has a grid size of 12.5 meters and is projected to WGS 1984 UTM Zone 16. No data are indicated by -9999.