In the urgent effort to pinpoint where wreckage from the space shuttle Columbia came to earth, investigators are finding that more "eyes" were watching the pieces fall than they first realized, and they are now trying to decipher what was seen.
Scientists at NASA and the Defense Department are masters at predicting the movements of objects in space. The Federal Aviation Administration has computer software that can make sense of a mass of otherwise confusing radar tracking data. And investigators from the National Transportation Safety Board have vast experience telling search parties where to look for key pieces of wreckage from airliners that crash.
As the search grows more difficult, those skills and tools are being melded with the sciences of meteorology, photographic analysis and astronomy, with secret military technologies about which no one can talk, and with plain, old human instinct. The lead agencies, plus the Energy Department, the National Oceanic and Atmospheric Administration and the U.S. Geological Survey, are learning a lot from each other as they write a new chapter in the book of trajectory analysis.
The cooperative effort has not only helped locate larger pieces of the shuttle in the Texas-Louisiana area, but has also identified a possible wreckage landing site in the rugged area around Caliente, Nev., much farther west along Columbia's flight path than the locations of pieces found so far. Investigators hope to identify other likely sites soon.
"We're sort of cross-training here," said Vern Ellingstad, the NTSB's director of research and engineering. "One of the things I've been impressed with is the level of not just cooperation, but a tightly integrated kind of approach."
The search has been aided by an unusual number of video images and photographs taken by space flight enthusiasts, a television station and others, including the military. The Columbia Accident Investigation Board said this past week that videos taken near the California coast appear to document the first point at which pieces began to shred from the shuttle as it approached the California coastline at an altitude of more than 200,000 feet, traveling at more than 17,000 mph.
The board established an advanced sightings team and directed it to analyze the videos and fix an exact timeline for each anomalous event they can see.
To get useful information from the images, the team must first establish the exact location of each photographer and the angle at which each photo or video was taken. They do this by a number of methods. In some cases, stars and planets are visible in the background. In other cases, power lines, mountains or trees provide a point of reference. All the videos and photos are then blended by computers and by educated eyes in a way that helps analysts determine exactly where each piece separated from the shuttle and what track it took as it fell.
Among other things, a senior official of the investigation board said, the analysis has determined that pieces of the shuttle began coming loose much sooner than he would have expected, at a point when the aerodynamic pressure on the fast-moving shuttle was still light.
From the point at which the videos indicate a piece broke off from the shuttle, especially if the piece remained visible for a long period, scientists can calculate where it likely reached the lower altitudes of about 90,000 feet, where ground-based radar might have picked it up. Light atmospheric drag on any object begins at about 400,000 feet, increasing as the object descends. After years of tracking space junk as it falls toward Earth, NASA and defense scientists have become quite accurate in projecting an object's path in these fringes of space.
For the next step, the FAA, the Air Force, the North American Air Defense Command and the National Weather Service have provided tapes from numerous radar installations along the shuttle's path. The first radar units that could have seen the shuttle were the Air Force's long-range, high-altitude radar systems with height-finding ability. The altitude function was used in previous airliner crash investigations, including that for Trans World Airlines Flight 800 in 1996. An FAA technician said the Air Force has honed that ability to near-perfection since then.
Luckily for investigators, the shuttle passed over a number of major military bases and secret facilities, including the Naval Air Systems Command facility at China Lake in California and Nellis Air Force Base in Nevada.
As the wreckage dropped lower, the FAA's long-range ARSR-4 radar started picking it up. At still lower altitudes, pieces would have shown up on local airport radar and other low-level radar.
Unlike the Air Force and FAA long-range radar systems, however, the lower-level radar installations are not required to record their images, and few did. One radar at the Dallas/Fort Worth International Airport was in record mode and covered part of the area where numerous pieces of the shuttle landed.
The task of analyzing these radar tapes is daunting, however. Falling pieces of wreckage do not have transponders, the devices carried by aircraft that automatically send a signal to an air traffic controller's screen showing the plane's identity, speed, altitude and direction.
The tapes being analyzed by investigators contain radar in its purest form -- called "primary radar" -- a jumble of millions of targets that can be anything, from aircraft, birds and speedboats to ghost images. Among other things, radar does not give a constant image; it only records a particular target every few seconds as its antenna sweeps the sky.
But if analysts with the NTSB and NASA can identify a piece of wreckage and then follow it through the jumble of other targets, radar can tell investigators a lot. A radar signal is reflected with varying degrees of energy, depending on what the object is made of and how big it is. By watching the movements of the target as it arcs deeper into more dense atmosphere, analysts can make an educated guess about its shape and weight. The analysts have already discovered one quirk: The shuttle's protective tile is not a good radar reflector unless it still has its ceramic coating.
The weather service has provided detailed weather records for the flight path, including data on high-altitude winds that could have played havoc with the falling wreckage, but "it was not a hostile kind of weather day," Ellingstad said.
Luckily for the shuttle investigation, several types of computer software have been developed to follow radar tracks, including one devised recently by an FAA technician and a crash investigator in the FAA's New England region.
One FAA technician said he was amazed at how rapidly NASA scientists learned to use the software. "I guess they are rocket scientists," he said.
Radar has other limitations. Most radar systems cannot follow the curvature of the Earth, and they cannot see through mountains. In the mountainous West, the shuttle pieces would have disappeared from radar at high altitudes. "It's not as good as if it had happened in Kansas," said one investigator.
Once the analysis of each shuttle piece sighting is completed, searchers on the ground are given a location as well as the direction of travel. It is only then that the analysts get to know how good they are.
"It's really been a neat thing to watch," Ellingstad said.
The NTSB, which has been tracking low-level pieces of airplanes for many years, has produced increasingly sophisticated techniques to help find them. One of the more famous wreckage searches was the hunt for a large piece of a titanium fan disk that was the key to solving the crash of a United Airlines DC-10 at Sioux City, Iowa, in 1989.
Investigator John Clark, who now heads the NTSB's aviation safety division, calculated the likely location of the large hunk of titanium, but no one could find it. Some time later, a farmer doing spring plowing ran into it. It lay buried in one of his fields -- almost exactly where Clark said it should be.