Operational Overview
Mapping in Depths >20m

In 2001 the Coral Reef Ecosystem Division (CRED) began a program to systematically map coral reefs of the U.S.-affiliated Pacific Islands that occur in depths of greater than 20 m. In these water depths satellite-based imagery cannot routinely be used for mapping, vessel-based acoustic systems must be employed instead. Acoustic mapping systems provide high resolution depth information, and the acoustic response can be analyzed to extract information about the seabed’s character, e.g. roughness or hardness. The regional variation in acoustic characteristics can be used to delineate changes in the seabed and identify regions of similar characteristics. Acoustics alone cannot identify specific habitats; therefore direct optical observations are made to identify specific habitats and correlate them with the acoustic regions. The tools required to perform deepwater benthic habitat mapping include mapping sonars and optical observing systems, such as underwater towed camera sleds, remotely operated vehicles, manned submersibles, and in 20-30m depths, direct diver observations, still photography and videos. The optical observing systems already used by CRED include underwater towed camera sleds and diver observations; in the near future an ROV will also become available. Because of the time required to plan, procure and field such systems, we implemented this capability in three phases:

Phase 1: Mapping from vessels of opportunity

A single-beam based bottom classification system and a towed optical assessment device were procured in 2001; these instruments were used on numerous cruises aboard the NOAA Ship Townsend Cromwell. When the Cromwell was decommissioned in 2002 this equipment was transferred to the Oscar Elton Sette. In 2002, our scientists also participated in a mapping cruise in the NWHI aboard the University of Hawaii’s R/V Kilo Moana that collected over 38,000km2 of multibeam data in 20-5000m water depths.

img/07a-TutuilaCoral.jpg
Corals on banks near Tutuila, American Samoa

On the Townsend Cromwell, acoustic mapping was conducted using a Simrad EQ-50 (50-kHz) single-beam echosounder that was integrated with a Quester Tangent Corporation View* data acquisition and bottom classification system. When transferred to the Sette, the QTC View was interfaced to one quadrant of the 38 kHz split-beam transducer used in conjunction with a Simrad EK-60 echosounder. Optical validation data were collected using an underwater camera platform in 20-100 m water depths. The acoustic and underwater camera data collection was typically done at night when other operations such as rapid ecological assessments, towed diver surveys and oceanographic instrument deployment cannot be conducted. In 2001 and 2002 CRED collected over 7500 linear kilometers of single-beam and bottom classification data in the NWHI and 1600 linear kilometers in the central Pacific. Thousands of photographs were taken with the tow/drop camera, but only approximately 850 in the NWHI and 300 in the Central Pacific were of sufficient quality optically or well enough geo-registered to use for benthic habitat identification.

img/tut3d.gif
Acoustic variability values displayed over single-beam bathymetry around Tutuila, American Samoa. USGS 10m DEM values shown in green

The Quester Tangent Corp. bottom classification system uses principal component analysis to analyze signals from an echosounder. QTC data from the 2001/2 cruises were studied to determine if similar bottom characteristics could be detected throughout the NWHI, the Pacific Remote Island Areas (PRIA), and American Samoa. This analysis showed promise, but one issue in particular limited its usefulness. The QTC can be used with a variety of sonars and different frequencies; however, data collected at different sonar frequencies or transmission characteristics cannot be combined for analysis. An example of this limitation occurred when the echosounder settings were changed on the bridge during the course of American Samoa cruise in 2002. These problems led us to look at the spatial variability of the data classes, rather than the classes themselves. This avenue of research allowed us to use class variability as a predictor of bottom type when we returned to American Samoa in early 2004, thus enabling us to target particular areas for optical sampling. We are planning to re-assess the 2002 Tutuila QTC data and compare it with multibeam data collected in early 2004 in order to better determine the long-term value of collecting single-beam bottom classification data. The inability to combine data sets is of particular concern, because we are working on several different ships with different echosounders and frequencies.

img/07c-toad.jpg
Towed Optical Assessment Device

Our initial tow-drop camera system (named TOAD for Towed Optical Assessment Device) integrated a digital still camera, an underwater video camera, lights, and laser pointers on a Guildline Minibat tow body. The Minibat design allowed limited control of the vehicle’s position while towing by changing the angle of the Minibat wings. This device was navigated using the Minibat software in which the operator entered an estimated layback which was integrated with the ship’s position to provide an estimated position of the tow body. Estimated positional accuracies using this method were on the order of 50 m. Various operational scenarios were tested including towing of the fish at speeds of 1-3 knots as well as using it in a drifting mode. After several cruises, it was determined that the drift technique was more effective and presented less risk to the tow fish and the operator’s nerves in rugged coral-rich areas. One of the drawbacks of this ship-based camera system is that when operating adjacent to islands with steeply sloping sides the areas of interest are often too close to the island for a ship to maneuver safely.

CRED also participated in an extensive multibeam mapping cruise in the NWHI in October/November 2002 aboard the R/V Kilo Moana. A discussion of the mapping techniques used during this cruise is available, as are grids of the resulting bathymetric data.

Phase 2: Launch-based mapping systems

In 2001, planning began to deploy a high-resolution, shallow-water mapping system that would be used from 20 meters into waters as deep as 250 meters. The result was the R/V AHI (Acoustic Habitat Investigator), a survey launch that was commissioned in 2003. In 2004 we also upgraded our underwater camera systems by replacing the TOAD with two more capable underwater towed camera sleds. One of the primary design goals of these camera systems was the ability to deploy them from a small boat as well as from ship.

img/07d-ahi.jpg
R/V AHI

The R/V AHI is an 8 m (25’) aluminum-hulled vessel with enclosed cabin. The vessel was specially designed by SAFE Boats International to transport personnel and survey electronics for seabed mapping surveys in the U.S. Pacific islands. The vessel is outfitted for daytime operations, either working in conjunction with a support ship, or for independent operations based out of small boat harbors in inhabited island groups. The AHI was specially built for high-resolution surveying of the seabed in depths from 10 to 250 meters. The hull and cabin house a 240 kHz RESON 8101 multibeam echosounder, a POS/MV position and orientation sensor (with two GPS antennas and an inertial measurement unit), and a SAIC ISS-2000 data acquisition and survey control system. The boat is typically operated by a coxswain and a surveyor.

During its first year of operation the AHI was transported to the work area by the NOAA Ship Oscar Elton Sette. Because the Sette was not designed to deploy a boat of this size the AHI was launched only in a calm harbor; the AHI then worked independently from the mother ship in populated areas where small boat facilities were available. During AHI’s first year of operation, over 750 km2 of seafloor in water depths between 10-350m were mapped in the NWHI , CNMI, Guam, and American Samoa .

In 2004 CRED’s optical data collection capabilities were significantly upgraded by the acquisition of two identical Deep Ocean Engineering TARAS camera sleds. The TARAS sleds are constructed using both underwater and surface control components available from the company’s Phantom ROV product line and are designed to be deployable from either ships or small boats. These camera systems are integrated with a Garmin GPSMAP 188 Sounder and a laptop computer running Hypack Survey; this configuration will provide more accurate positioning of optical validation data than was previously possible with the TOAD.

img/07e.jpg
TARAS ROV and data acquisition system

Phase 3: Ship-based mapping systems

In August 2004 the NOAA Ship Hi’ialakai arrived in Hawaii. This ship is equipped with davits that allow the AHI to be deployed from the ship in moderate sea conditions, thus extending AHI’s mapping range to isolated islands and banks. The Hi’ialakai will also have two multibeam echsounders installed in late 2004 – both a Kongsberg Simrad EM300 30-kHz sonar for mapping in 100-4000m and an EM3000 300-kHz sonar for mapping in 20-160m depths. When working together the Hi’ialakai and the AHI will be able to cover up to 40 km2/day of shallow banktop regions that are typical in the NWHI. They will also be able to effectively map in areas such as the northern Mariana Islands or the U.S. Line Islands where the Hi’ialakai can quickly survey offshore areas while the AHI surveys inshore areas that would be dangerous for a large ship.

The Hi’ialakai will also be equipped with a remotely operative vehicle (ROV) and underwater positioning system that will also allow much more efficient and precisely located collection of optical validation data than has previously been possible from either the Townsend Cromwell or the Sette.

Data Processing

Single-beam based bottom characterization data are processed using QTC IMPACT software. Spikes and outliers are first removed from the data manually and a principal components analysis is used to analyze the shape of the electronic signal and separate the data into QTC “classes”. The number of classes, the optimal groupings, and the spatial and depth limits are determined by the operator. Bottom classes are displayed in conjunction with linked optical samples in ESRI’s ArcView 3.3 software and examined for groupings and relationships between the two data types. A depth model is also generated using the single-beam depths and multibeam data where available, and the QTC class data are displayed in 3D draped over the topographic model. As described previously, due to operational problems, an analysis was also done of the variability among the QTC classes. This analysis avenue shows potential, and further analysis of the QTC data in relationship to recently acquired multibeam bathymetry and backscatter data in selected test areas is planned.

Multibeam data are acquired with SAIC’s ISS-2000 which records the data and provides survey control and underway quality control displays. SAIC’s SABER processing software is used to process the raw soundings, analyze the results, manually edit the sounding data to remove outliers, and derive average gridded data values. The data are corrected for predicted tides during field operations; if necessary the data are later updated with observed tides after the fact. GMT is used to reformat the SABER grid into final form. Interactive Visualization System’s Fledermaus and ESRI ArcGIS are used for viewing results and creating browse objects.

*Identification of particular vendor’s equipment does not imply endorsement by NMFS.