Benthic Habitat Mapping
As directed by the National Coral Reef Action Strategy, the mission of the PIBHMC is to produce comprehensive digital maps of all coral reef habitats in the U.S. and its Trust Territories. These habitats cover over 400,000 sq. km. and are dispersed across the entire central Pacific, extending over 40 degrees in latitude and 60 degrees in longitude. Comprehensive mapping of these areas is an extensive and complex undertaking.
In 2000-2002 NOAA conducted studies to determine what type of techniques and equipment were most suitable for habitat mapping in and around coral reefs. For mapping in Pacific waters up to 30 meters, this has been accomplished primarily through the development of maps based on IKONOS satellite imagery. NOAA’s Biogeography Program leads the effort to comprehensively map the distribution of shallow coral reefs and other benthic habitats. In waters with depths greater than 20 meters, hull-mounted multibeam sonars that produce both bathymetric and acoustic backscatter data were chosen as the most appropriate technology for benthic habitat mapping. [See Pacific Implementation Plan (36.7 MB PDF).] In addition, it was determined that optical validation data are required to aid in interpretation of both the satellite imagery and the multibeam data.
During this same period NOAA was able to substantially enhance the capability of its Hawai‘i-based fleet when the NOAA Ship Oscar E. Sette replaced the Townsend Cromwell in 2002 and the Hi‘ialakai joined the fleet in 2004. PIBHMC and HMRG scientists working in collaboration with NOAA Marine and Aviation Operations (NMAO) and Office of Coast Survey (OCS), developed a plan for multibeam and optical mapping from these ships. In 2001, a program was initiated to collect optical validation data, primarily during night operations on Reef Assessment and Monitoring (RAMP) cruises, first on the NOAA Ship Townsend Cromwell and later on the Oscar E. Sette. The next year a procurement was initiated to build and outfit a 25-ft survey launch, R/V AHI with a RESON 8101ER multibeam sonar, which at 240-kHz is designed for mapping in depths between 5 and 250 m. The R/V AHI was designed to be deployed independently or from either the Oscar E. Sette or the NOAA Ship Hi‘ialakai.
Because of the extensive areas to be mapped in the Pacific, the Hi‘ialakai was also equipped with two multibeam sonars. The ship has a Kongsberg 300 kHz EM3002 that can map in depths up to 150 m, and a Kongsberg 30 kHz EM300 that maps in depths between 50 and 3000 m. These sonars were chosen to enable the ship to collect comprehensive ecosystem habitat information on the bank tops as well as on the adjoining slopes, where important bottom fish habitats are located. In addition, vertical reference, CTD, data collection, and data processing hardware and software were chosen that are entirely compatible with the AHI’s multibeam system; this allows scientists and survey technicians to easily collect and process on both the Hi‘ialakai and the AHI.
Because the mapping operations are largely dependent on the availability of the NOAA ships, a variety of operational scenarios has been developed to maximize the amount of mapping that can be accomplished. During a dedicated mapping cruise the Hi‘ialakai surveys round-the-clock and the AHI is deployed during daylight hours to map near-shore or shoal areas that might endanger the ship. Ten mapping personnel typically support these dedicated operations. During combined mission cruises the ship is usually dedicated to another mission during the day but can perform mapping during the night. The AHI may or may not be used depending on the other missions and the resources available for mapping. Two to four mapping personnel usually support a combined mission. The Hi‘ialakai’s survey technician is also fully trained in multibeam surveying and can, if called upon, collect a limited amount of multibeam data during night operations on other cruises without additional support from the mapping team.
Except for very small areas, it is not logistically feasible to optically image more than a small fraction of the seafloor deeper than about 30 m within a coral reef ecosystem using any technology currently available. Accordingly, optical imagery of the seafloor is collected primarily to ground truth or validate the interpretation of benthic habitat characteristics from bathymetry, backscatter imagery, or other data. To date, optical data have been collected using two different types of systems. In shallow waters, at depths between a few meters and 20–30 m, towboard divers with the NOAA Pacific Islands Fisheries Science Center’s Coral Reef Ecosystem Division (CRED) collect video or still imagery while being towed behind a small boat over track lines pre-programmed into a handheld GPS unit. For a complete discussion of this method and the data it generates see Kenyon et al, 2004 (PDF).
However, the majority of seafloor in most Pacific coral reef ecosystems is deeper than the approximately 30 m maximum depth attainable by standard CRED towboarding methods. For these deeper areas an evolving series of underwater camera sleds, usually referred to as the Towed Optical Assessment Device (TOAD), have been used to acquire optical validation data. See the Towed Camera Systems page for a detailed description of the camera sled systems.
Camera sled deployments were conducted at night, usually between 1800 and midnight, to avoid interfering with daytime small boat and diver operations. The TOAD was originally deployed from a pot hauler mounted on the starboard side of the fantail on NOAA Ship Townsend Cromwell. On the NOAA Ships O.E. Sette and Hi‘ialakai various iterations of TOADs were deployed off the portside J-frame amidships or the starboard side J-frame respectively. In all cases the sleds were lowered slowly to the bottom by the deck crew with the use of a capstan. The TOAD operator monitored a live video feed from the camera and began recording data on two video tape recorders. When the camera reached bottom the deck crew was notified by radio to stop lowering, and the ship moved off along a predetermined course, towing the TOAD astern for the first TOAD, or by drifting with subsequent camera sleds.
TOAD Tracks (ArcGIS shapefile): These files allow an ArcGIS user to plot the positions of the camera device during the survey period. The attributes of the shapfile (stored in the DBF file) include the results of video classification, and allow the Arc user to sort points by, for example, substrate type or living cover category. Zipped file includes DBF, PRJ, SBN, SBX, SHP, SHP.XML, and SHX files; the metadata and JPG files are also included.
Metadata: A text file documenting all pertinent information specific to the cruise, year, research vessel, and camera sled model used during the collection period.
JPG: A high-resolution image of the track map. Separate track maps exist for each year of data collection, as well as a Master track map which displays all tow tracks together.
Benthic Habitat Classification Codes: A complete listing of the categories used in photo/video classification, including the category codes. These codes are the values which appear in the shapefiles’ attributes table.
ArcGIS-compatible symbology files: Two separate ESRI LYR files which enable the ArcGIS user to match the symbology of the downloaded shapefiles in their projects to that displayed in the maps on the PIBHMC website.
The geomorphological data layers of substrate, slope, rugosity, and bathymetric position index (BPI) produced at the Pacific Islands Benthic Habitat Mapping Center (PIBHMC) are derived from multibeam bathymetry. The substrate classifications (hard bottom vs. soft bottom) is derived from a combination of backscatter imagery and bathymetric variance, and they are constrained by optical validation. These products all support NOAA Coral Reef Conservation Program goals. Goal 1 is to map all U.S. Coral Reef Ecosystems. These data sets specifically address Objectives 1 and 4: to develop high-resolution benthic maps and to characterize priority deep water reefs and associated habitats. The derivatives provide Geographic Information Systems (GIS) layers that may be used for benthic and essential fish habitat characterization, and for the study of geologic features. By combining these geomorphological layers with bathymetry, backscatter, other derivatives and in situ data, they collectively compose benthic habitat maps which are designed to be used to understand and predict moderate depth (~20m – 150m) benthic habitats for different organisms that inhabit coral reef ecosystems.
These data are experimental layers that are part of a standard set of benthic habitat products. PIBHMC has developed these data layers for two pilot study sites: Tutuila, American Samoa and French Frigate Shoals (FFS), Northwestern Hawaiian Islands. Managers can use these base layers for addressing a wide range of management and research questions related to individual species, trophic levels, or entire coral reef ecosystems. Although the individual maps have many uses, they are often most effective and powerful when synergistically combined using a GIS. By selecting features ofdifferent layers in unique combinations using queriesand other tools built into the GIS, this suite of map layers can be used to generate new maps, the values of which may be indicative of a particular habitat, benthic structure, or process occurring within the ecosystem.
The production of each derivative is accomplished in ArcGIS 9.1 using Spatial Analyst. For rugosity and BPI structures and zones, the Benthic Terrain Modeler (BTM), downloadable from NOAA’s Coastal Services Center, was the primary tool. The substrate classifications were derived in ENVI and ArcGIS 9.2 with Spatial Analyst neighborhood statistics.
An approaching release of stand alone GIS projects for Tutuila and FFS is expected by the end of 2007. The stand alone projects (available in ArcMap and ArcReader formats) include all of the benthic habitat data available on the website as well as other basemaps for terrestrial and nautical reference. Optical validation data have been collected on both bank tops using three separate methods: Tethered Optical Assessment Device (TOAD) underwater video and photo surveys, towed diver optical surveys, and Rapid Ecological Assessments (REA). Each method is designed for a certain depth range and spatial scale, and some of the optical surveys overlap the multibeam data while others do not. The Tutuila project will include TOAD and towed-diver transects analyzed for bottom type and bottom cover. The FFS project will, for now, only include TOAD classifications.
Substrate, Hard bottom vs. Soft bottom: This is a preliminary product. Cell values reflect whether the seafloor is hard bottom or soft bottom based on an unsupervised classification run in ENVI software. The classifications are based on backscatter, bathymetry, acoustic derivatives and optical data. Click here to see a conference abstract discussing this classification strategy.
Slope: Cell values reflect the maximum rate of change (in degrees) in elevation between neighboring cells.
Rugosity: Cell values reflect the surface area and (surface area) / (planimetric area) ratio for the area contained within that cell’s boundaries. They provide indices of topographic roughness and convolutedness (Jenness 2003). Distributions of fish and other mobile organisms are often found to positively correlate with increased complexity of the seafloor. Investigations of which of the many methods available for quantifying benthic complexity best correlate with fish distributions in Pacific coral reef ecosystems are underway. Results of the Jenness (2003) method are provided as a standardized and well-documented interim product.
Bathymetric Position Index (BPI): BPI is a second order derivative of bathymetry. The derivation evaluates elevation differences between a focal point and the mean elevation of the surrounding cells within a user defined annulus or circle. A negative value represents a cell that is lower than its neighboring cells (depressions) and a positive value represents a cell that is higher than its neighboring cells (crests). Larger numbers represent more prominent features on the seafloor, which differ greatly from surrounding areas. Flat areas or areas with a constant slope produce near-zero values. (Lundblad et al. 2006)
- Variables of BPI
- Scalefactor = input (bathymetry) grid resolution * the outer radius (orad) of the annulus (or the radius of the circle).
- For example: 5m grid * orad of 50 results in a scalefactor of 250; 5m grid * orad of 4 results in a scalefactor of 20.
- Classified optical data collected with the TOAD around Tutuila in 2002 and 2004 were used to validate transitions between BPI classes. By using the sand and scleractinian coral classifications, transitions were measured to confirm that 20 m and 250 m scalefactors would adequately represent the environment.
- For Tutuila BPI, 5 experimental broad scalefactors and 10 experimental combinations were tested, and the two scales of BPI (20 & 250) were chosen based on the diameter of reef structures seen in the bathymetry (also after Lundblad et al. 2006).
- For other islands similar combinations of scalefactors were tested, and the most appropriate scales of BPI were chosen based on the scale of the reef features seen in the bathymetry.
Jenness, J. 2003. Grid surface areas: Surface area and ratios from elevation grids [Electronic manual]. Jenness
Enterprises: ArcView Extensions. http://www.jennessent.com/arcview/arcview_extensions.htm (scroll to “Surface Areas and Ratios from Elevation Grid”)
Lundblad et al., 2006, A benthic terrain classification scheme for American Samoa, Marine Geodesy, 29(2):89 – 111. http://www.csc.noaa.gov/digitalcoast/tools/btm/index.html
Predicting Seafloor Facies from Multibeam Bathymetry and Backscatter Data, Dartnell, P. and Gardner, J.V., Photogrammetric Engineering and Remote Sensing. Vol. 70, No.9, pp 1081-1091, 2004.
In order to support these acoustic and optical mapping operations, PIBHMC also must perform a variety of office-based tasks. All currently available data must be obtained and synthesized to prepare the best available maps of the area and thus provide a focus for the planned operations. Formal instructions must be provided to the ship and any necessary permits must be obtained. Survey plans must be developed and predicted tide information must be obtained. After the data have been collected it must be processed, analyzed and transformed into data products that can be used by the resource managers that are the mapping center’s primary customers. All of these activities are described in a series of documents that provide an overview of PIBHMC activities.