TWA Flight 800 -- Picking Up the Pieces

Sophisticated systems comb ocean floor for clues to TWA flight 800 disaster.

by / January 31, 1997
In a guarded hangar on Long Island, N.Y., experts are piecing together the Boeing 747 that was once TWA flight 800. Investigators from the National Transportation Safety Board (NTSB) and the FBI are reassembling the aircraft in an effort to learn the cause of the explosion that blew the Paris-bound jet out of the sky shortly after taking off from JFK International on the night of July 17, 1996, killing all 230 aboard and scattering wreckage over five miles of the Atlantic.

From wreckage recovered to date, investigators have determined that the jetliner's center fuel tank exploded, but though most of the victims and parts of the plane -- including all four engines -- have been found, neither forensic examination nor analysis of the assembled sections has yet revealed the cause of the explosion. As of this writing, the search for the tank and remaining parts of the aircraft continues. If these prove to be too small for electronic detection, scallop trawlers may be used to comb the sea floor. Otherwise, the remaining pieces will probably be found with advanced technologies similar to those used by the National Oceanic and Atmospheric Administration (NOAA) ship, the Rude (pronounced "Rudy"), and its support group in the first few weeks following the crash.

One of 15 ships in the NOAA fleet, the Rude routinely conducts hydrographic survey operations off the Atlantic and Gulf Coasts, using sophisticated side-scan sonar and multibeam echo sounder systems to locate and chart submerged wrecks and obstructions to navigation. Data provided by the vessel are used to update U.S. nautical charts.

At the time of the crash, the Rude was en route to Rhode Island after completing a hydrographic survey of the approaches to New York. She was one of the first ships to reach the crash site, following the Coast Guard Cutter Juniper. After assisting in recovering floating wreckage, the Rude's commanding officer, NOAA Corps Cmdr. Sam De Bow, was requested to begin side-scan sonar sweeps. Within two hours, the Rude located the wreckage, 110 feet down on the ocean floor.

Originally designed for military applications, side-scan sonar has become the primary search tool for locating sunken ships, downed aircraft and other objects on the sea bottom. In side-scan operations, the sensor, or "towfish," is towed behind the ship at a predetermined depth. As it moves through the water, the sensor radiates sound waves in a fan-shaped pattern, directly down and out to each side of the towfish. Echoes, or contacts, reflecting from objects protruding from the ocean floor are detected by the sensor and sent to the side-scan recorder aboard the ship, where they are processed, recorded and translated into imagery on a thermal printer. As each sonar contact is made, it is time stamped and geocoded with differentially corrected GPS coordinates. Since the Rude uses corrections from the Coast Guard Differential GPS (DGPS) Navigation Service, the positional accuracy of each contact was in the range of 3 meters to 5 meters.

The effectiveness of side-scan imagery is largely determined by the depth at which the sensor is towed. The closer to the bottom, the more defined the shadow created by the sound waves as they pass over the object (comparable to shadows cast by lamplight). Shadows not only highlight the presence of objects, they are used to estimate the height of the object off the bottom. "When we wanted to get higher-resolution imagery," explained NOAA Hydrographic Survey chief, Capt. Andy Armstrong, "we dropped the towfish closer to the bottom and used a smaller range scale."

Once contact was made -- using an AS600 side-scan sonar from EdgeTech -- the Rude began a series of parallel runs (track lines) that produced overlapping sonar images. As the survey progressed, the sonar data, track and position of the ship were processed, combined, recorded and displayed by a data acquisition system (DAS). The DAS combines two processing systems: the HYPACK, from Coastal Oceanographics, and the ISIS, from Triton Technology. The DAS, interfaced with a DGPS receiver, is the core of the vessel's navigation system. It can direct the ship along a preset course -- in this case, track lines.

Since side-scan data requires complex processing and interpretation before it can be useful to people other than hydrographic surveyors, NOAA established a shore support group to process data and produce maps that would be needed by those involved in the search-and-recovery effort. Lt. Cmdr. Emily Christman was a member of the group. "We were there to translate the sonar data into a GIS format that could be used by salvage ships, dive boats, Navy personnel, the press -- by everyone involved."

The NOAA group worked out of a trailer hastily set up at the Coast Guard Station on East Moriches Beach, Long Island. Christman recalled that there were more than 30 trailers of every size, plus recreational vehicles and tents. "Our trailer was very small; we had three computers, a plotter and an HP 650 printer, plus a post-processing system with its own plotter and computer. At the peak, we had five people working out of that trailer."

In addition to the side-scan sonar, the Rude simultaneously ran a SeaBat multibeam echo sounder from RESON Inc. to obtain depth, elevation measurement and bottom contour imagery. The multibeam recorder translates sound waves reflecting off the bottom into elevations, or into imagery of the bottom including the objects on it. The system can calculate the elevation of objects and bottom features with 5-centimeter accuracy.

Armstrong said the multibeam system was used not so much in searching for the wreckage, as in doing detailed surveys of certain patches in the wreckage area, and for locating anchoring sites for the Navy salvage ships, the USS Grapple and the USS Grasp. "They needed to put multiple-point moorings over wreckage areas without dropping anchors on the wreckage. The Rude made detailed micro-surveys of the mooring locations using both multibeam and side-scan sonars."

Since time was a critical factor in the recovery effort, the Rude used the widest scan possible while maintaining the resolution needed. "In the depths at the crash site," Armstrong said, "the towed side-scan sensor could cover a swath 200 meters to 300 meters (660 feet to 990 feet), while the multibeam could cover only 30 meters to 40 meters (99 feet to 132 feet). "As it was, it took the Rude two weeks to cover the whole area. The multibeam system would have taken much longer."

How small an object can the side-scan detect? Christman said that depends to some degree on the type of bottom. "Where the wreckage field is located, about 10 miles to the south, off East Moriches Beach, Long Island, the bottom is flat, without a lot of clutter. The reports we got back from the salvage teams said one of the contacts was a coffee pot. Another turned out to be a pile of Coke and Pepsi cans."

Data from both sonar recorders and from the ship's navigational system were recorded and brought ashore, where they were loaded into an HP 9000 series 300-based post-processing system, essentially a special hydrographic data acquisition and processing system (HDAPS) used by NOAA. Following a series of complex corrections, data from the HDAPS were converted to a deliverable GIS format.

Using MapInfo Professional GIS, the NOAA group produced maps displaying separate databases of the aircraft's flight path. Since each pixel of the mosaic was geocoded, nautical charts could be used as base maps, providing a familiar, user-friendly backdrop for all types of data.

"Our first category of products," Christman said, "was the dive sheets. The second was for Navy Rear Adm. Edward Kristensen, who held daily press briefings, and for a while, daily briefings for the families of the victims. For these we were able to produce poster-size prints showing where the salvage ships were, what activities were going on and where wreckage had been found. The admiral said he found the maps extremely helpful."

The Navy, NTSB, FAA, FBI, TWA and Boeing all required maps with various scales and databases. Salvage ships needed enlarged sections of the wreckage field to avoid anchoring on or near objects. Operations coordinators required maps indicating what part of the wreckage field had been cleared. Investigators needed access to the various databases. Divers whose working range was restricted by air hoses and poor visibility needed map enlargements of very small areas in order to be placed as close as possible to the objects they were going to recover. "With MapInfo," Christman said, "we were able to produce maps with various databases, at the scale and size needed."

Cooperation between the NOAA support group, the Navy and the dive teams contributed to the rapid recovery of most of the victims. The group was able to provide divers with descriptions of the contacts so they would know what to look for in the darkness. Divers, in turn, confirmed the identity of the contacts. "We were getting enough feedback that we were able to say, well, this contact turned out to be three to four seats, so we would go back and examine the data for more contacts that looked like that. The information was extremely sensitive," Christman said, "and carefully guarded."

Since the Rude was not designed for salvage work, it departed for Rhode Island shortly after completing the survey. Recovery operations were carried out by the USS Grasp and the USS Grapple. According to Christman, the NOAA shore group stayed several days longer. "We continued interpreting and reviewing side-scan records to make sure nothing was missed and to continue providing the paper products needed by the Navy and NTSB."

Using teamwork and leading-edge technologies, NOAA performed a timely and valuable service in locating the wreckage of TWA flight 800, as well as in providing critical mapping and database information needed for the search-and-recovery effort. The work by the Rude and its support group enabled Navy salvage teams to bring up remains of the victims and most of the wreckage in the shortest possible time.

Sections and pieces of the plane recovered to date have enabled NTSB to determine that the jetliner's center fuel tank exploded. Clues to what caused the explosion may be among the missing pieces on the floor of the Atlantic. If these are found, it will most likely be with efforts and technologies used so effectively in the days following the crash.

Bill McGarigle is a freelance writer specializing in communication and information technology. E-mail:.



PROBLEM/SITUATION: Locating sunken wreckage and quickly providing digital mapping and hydrographic imagery for salvage teams.

SOLUTION: Side-scan sonar and other advanced technologies.


VENDORS: EdgeTech, RESON Inc., Coastal Oceanographics, Hewlett-Packard, MapInfo Corp., Triton Technology Inc.

CONTACT: Capt. Andy Armstrong, Hydrographic Survey Division, Office of Coast Surveys, NOAA, 301/713-2698 x 124, e-mail: . Lt. Cmdr. Emily Christman, Operations Branch, NOAA, 301/713-2702, e-mail: .

ISIS, sonar and seismic data acquisition and processing system:
Triton Technology Inc.
125 Westridge Dr.
Watsonville, CA 95076

AS600 side-scan sonar:
455 Fortune Blvd.
Milford, MA 01757

SeaBat 9001S portable multibeam echo sounder:
300 Lopez Rd.
Goleta, CA 93117

Coastal Oceanographics
11-G Old Indian Trail
Middlefield, CT 06455

HP 9000 series 300-based post processing system:
3000 Hanover
Palo Alto, CA 93034

MapInfo Professional GIS:
MapInfo Corp.
One Global View
Troy, NY 12180-8399