Data Management Plan

This benthic habitat map was created from a semi-automated habitat mapping process, using a combination of bathymetry, satellite imagery, aerial imagery and underwater videos to classify the 0-35m deep benthic habitats of the region around Northeast Puerto Rico (NEPR) and Culebra Island, including the shallow part of the Northeast Ecological Reserves, Puerto Rico. The area of interest includes the nearshore waters of Fajardo and Luqillo to the Former Roosevelt Roads Navy Base, the Vieques Sound, La Cordillera Reserve, the Luis Pena Reserve, and the waters around Culebra Island. The benthic habitat map is classified based: on 1) geomorphological structure, 2) hard bottom cover, 3) topographic complexity, 4) dominant biological cover, 5) live coral cover and 6) dominant type of coral cover. Underwater photo and video data was collected to provide ground validation for remote sensing imagery, and to assess the accuracy t of the final map. Habitat features are described by varying levels of detail (i.e., major and minor categories nested within them), so users can depict the habitat information that best suits their research or management needs.

Process Steps:

• 2013-05-01 00:00:00 - A multiresolution bathymetry model was created for northeast Puerto Rico and Culebra Island. The model was derived using geostatistical modeling of depth sounding data from 1900-2013, including led soundings, single-beam sonar, multibeam sonar surveys and LiDAR surveys. The source depth layers were kriged into multiple resolutions (4m, 20m and 100m depending on source data density) and merged using Global Mapper, Surfer and ArcGIS to produce a continuous 4m surface. | Source Produced: NEPR Bathymetry Model (Citation: NEPR Bathymetry Model)
• 2015-02-01 00:00:00 - The geostatistical bathymetry model was processed into derived morphometric layers using Raster Calculator and Jenness Surface Tools in ArcGIS. The slope, slope of slope, rugosity, curvature profile, curvature plan, standard deviation, and mean were derived from the bathymetry model and analyzed for its three first principle components in ENVI. The PCA image was then saved as a 3 band raster (Geotiff). The PCA 3 band image was included as a layer to the semi-automated feature extraction and classification process that produced the benthic habitat map. | Source Produced: NEPR Principle Component Analysis (Citation: NEPR Bathymetry Model)
• 2013-06-01 00:00:00 - An enhanced satellite imagery mosaic was geNEPRated from the mosaic of high resolution WorldView-2 satellite data. Cloud and land cover were masked and the near infrared1 band were used to deglint the RGB bands in the imagery. The Lyzenga method was then used to water-column correct the enhanced RGB bands to better represents underwater benthic habitats. The final mosaic color-balanced using 12 individual scenes divided into 27 sub-scenes using a neighborhood analysis in PCI. The final mosaic was included as a layer to the semi-automated feature extraction and classification process that produced the benthic habitat map. | Source Produced: WV2 Satellite Imagery Mosaic (Citation: WV2 Satellite Imagery Mosaic)
• 2013-01-01 00:00:00 - The Ground validation sites were planned by manually placing points on a draft habitat map. The sites were selected to include all types of remote sensing signatures that could be identified, covering habitat features across the whole project area (approximately two GV sites/km2). Each ground validation site was sampled either by an underwater drop-camera (SeaViewer Sea-Drop 950 and a GoProHero 3 Black Edition) or by snorkeling to the site with handheld underwater video camera (GoProHero 3 Black Edition, 1440*1920 pixels video, 12 MP photo/5s). Drop Camera Videos were positioned and recorded using a Trimble Geo XH GPS system, while snorkel videos were positioned using a handheld Garmin GPS on a dive buoy. The video data was then used to extract habitat data used to classify the habitat map, as well as presence of selected key species such as ESA listed Acropora palmata and Acropora cervicornis corals. | Source Produced: NEPR Ground Validation Videos (Citation: NEPR Ground Validation Videos)
• 2014-01-01 00:00:00 - Accuracy Assessment points for the NEPR were selected using a stratified random sampling method based on the benthic habitat map schema. Each accuracy assessment site was entered into a Garmin GPS navigation, located by the Biogeography Team in the field, and sampled either by an underwater drop-camera or by snorkeling to the site with handheld underwater video camera. Videos were analyzed by the Biogeography Team and recorded into a database using a Trimble Geo XH system. The video data was then used to accuracy assess the Benthic Habitat map. Presence of selected key species such as ESA listed Acropora palmata and Acropora cervicornis corals were also recorded. | Source Produced: NEPR Accuracy Assessment Videos (Citation: NEPR Accuracy Assessment Videos)
• 2013-02-01 00:00:00 - The 2015 benthic habitat map for Northeast Puerto Rico and Culebra Island, covers 711 km2 of Benthic Habitats with a minimum mapping unit of 100m2. The map was digitized using semi-automated feature extraction techniques and remote sensing imagery (a bathymetry model [1900-2013], WorldView-2 satellite imagery [2011-2013], and Aerial Imagery [2007, 2010]). Ground Validation Videos were collected and classified based on a habitat classification scheme to verify the habitat features segmented using remote sensing imagery. The map was classified using Boosted Regression Tree modelling, using the remote sensing data, the GV data and a suite of spatial predictors including data extracted from the CariCOOS nearshore oceanographic model. The model map was manually edited and adjusted based on visual interpretation. The final map product was adjusted based on the result from a separate accuracy assessment field data collection. | Source Produced: NEPR Benthic Habitat Map (Citation: NEPR Benthic Habitat Map)