Engineers had dismissed the danger from the flying insulation while the astronauts were in orbit. But Ron Dittemore, the shuttle program manager, said during a news conference in Houston yesterday that it remains the best lead in the mysterious disaster that claimed the lives of seven crew members.
As he spoke, scientists and engineers were sifting through more than 12,000 pieces of shuttle debris that have been recovered in Texas and Louisiana. Yesterday's discoveries included a piece of the cabin segment and nose cone.
Dittemore said the most important clues are likely to come from any debris found further "upstream" in Arizona or New Mexico. Those would be the first pieces that broke from the craft and a clue to the source of the fatal disintegration.
Meanwhile, scientists will rerun their analysis of what happened when a 2.7-pound sheet of hardened foam peeled off the shuttle's external fuel tank 80 seconds into liftoff, striking the underside of the wing and vaporizing in a puff of tiny pieces of debris.
Caught during a review of film the day after liftoff, it raised concerns about possible damage to the shuttle's critical tiles, which protect the craft against heat that reaches 3,000 degrees on re-entry.
Engineers began an intensive study that concluded on the 12th day of the mission. Dittemore said the team decided that even if the heat shield tiles had been damaged, it would not pose "a flight safety issue."
"At the time, I was not aware of reservations by any individual on our team," he said, noting that the computer simulation used to analyze the incident was well-tested and conservative.
Columbia, the oldest of the National Aeronautics and Space Administration's four shuttles, was destroyed 39 miles above Texas on a landing approach that was to end at Cape Canaveral, Fla. after a successful 16-day scientific mission.
NASA's investigation revealed yesterday that in Columbia's final seconds, its flight computers were in a losing struggle to keep the spacecraft from making an uncontrolled turn to the left.
NASA officials said increasing wind resistance pulling the shuttle to the left caused the shuttle's four right "yaw" jets to fire to assist other flight controls wrestling with the problem.
"It appears we were losing ground," said Dittemore. "It was not long after that point that we lost all data and communications with the crew."
NASA officials planned to take time today for a memorial service for the seven lost Columbia astronauts. President Bush was to attend services at the Johnson Space Center.
There was to be no other interruption in the investigation or the search for the remains of the shuttle and its crew.
In Nacogdoches County, Texas, Sheriff Thomas Kerss said searchers had found what he described as 7-foot piece of the shuttle crew cabin and computer components.
Part of the shuttle's nose was found buried deep in a 20-foot-wide hole on the Texas-Louisiana border.
The debris and crew remains were being collected at Barksdale Air Force Base near Shreveport, La., and at Carswell Air Force base near Fort Worth, Texas.
Dittemore said yesterday that barely 1 percent of the shuttle's bulk had been recovered, none of it of critical importance. Intact, Columbia weighed about 180,000 pounds, but none of the debris struck or injured anyone.
The search zone now encompasses 28,000 square miles from northwestern Louisiana to Fort Worth. Dittemore said investigators are eager to retrieve any shuttle fragments that may have fallen even farther west.
"We are primarily interested in any debris, structure, tiles upstream [west of] Fort Worth," even as far away as New Mexico and Arizona, he said. "If that exists, it is extremely important to us. It's going to be a key to the puzzle."
That's because any debris shed by the shuttle early in its descent might reveal where in its structure the fatal malfunctions began. Heat shield tiles bear codes that would reveal exactly where on the shuttle's skin they were lost.
"Where are they?" Dittemore asked.
NASA's continuing examination of data from the shuttle during its last minutes shows that the first sign of trouble was unusual heating in brake lines in the spacecraft's left landing gear that began at 8:52 a.m., eight minutes before contact with Columbia was lost.
The 50 degrees of extra heat in the wheel well is "interesting," Dittemore said, but does not look like what would occur if the heat shield had failed.
"There's some other missing link that we don't have yet," he said, "and we've got to go find it."
He expressed confidence in the validity of the original finding that there was little risk to the astronauts from damage to the shuttle's heat shield caused by flying insulation. But investigators will review the assumptions, anyway.
Such damage has occurred at least twice before during shuttle launches, without serious incident.
In their analysis this time, Dittemore said, engineers estimated that the piece of hardened foam insulation that broke from the external fuel tank's nose was 20 inches long and 6 to 16 inches wide, weighing about 2.7 pounds.
They calculated that it would have struck a glancing blow to the wing near the landing gear door. That could have resulted in a complete loss of one tile or shallower damage to an area 32 inches long and 2 to 7 inches wide, extending to the wing's trailing edge.
In the end, Dittemore said, his engineers concluded that "you would not have had damage sufficient to cause a catastrophic event, and no impact on the flying qualities of the vehicle."
Michael Kostelnik, deputy associate administrator for NASA's office of space flight, said the engineers' conclusions were reported to NASA headquarters. "We were in complete concurrence with their assessment, and their plan ahead," he said.
Even if engineers had concluded that the damage was a risk to the crew, Dittemore said, there would have been no good options for saving the astronauts.
Steering the shuttle to favor the left wing and protect it from fatal heat damage would have risked overheating the right wing, he said.
Even if such a maneuver had been possible, he said, the shuttle, because it returns to earth as an unpowered glider, could not have reached the spaceport in Florida.
The crew would have had to ride the spacecraft to a safe speed and altitude and bail out, allowing the shuttle to crash.
"I just don't believe we ever would have had the type of information to make that type of risk trade," he said.
The cascading failures aboard the shuttle Columbia may have begun as far west as California, where freelance newspaper photographer Gene Blevins was waiting to photograph it as it flew over the Owens Valley Observatory south of Bishop at 5:52 a.m. local time.
"It was right on the money," Blevins said. "Here comes this big, red glow, coming over the mountain. ... It looked like a huge meteor blasting through the sky."
"About midway, we saw these little red lava-rock [objects] coming from underneath it, breaking away from it in small chunks ... like a meteor that enters the atmosphere and starts breaking up."
Just before the shuttle disappeared over mountains to the east, he said, "we saw this big red flare, like it [the shuttle] had a bomb bay door, and it dropped out and burned for about a second and a half, then disappeared."
Mark Drella, a professor of aerodynamics at Massachusetts Institute of Technology, said that if a few tiles fell off during the launch, the resulting friction could have forced more tiles to fall off during re-entry, a chain reaction known as a "zipper effect."
"The heat diffuses to a neighboring tile, and they fall off like a zipper unzipping," he said.
Charles Oman, an aeronautical researcher at MIT who was prinicipal investigator for a 1998 Columbia mission, said that if enough of the shuttle's 24,000 tiles fail, its aluminum exterior will break apart in the intense heat. Aluminum melts at 1,000 degrees.
"The tiles have always been a key safety feature," he said.
To check on the status of the tiles after each flight, NASA crews probe the shuttles with suction cups to check on their stability.
Donald Emero, a retired vice president of shuttle engineering at Boeing who managed the shuttles' thermal protection systems, said the tiles are made of pure silica and are so delicate that they could be damaged flying through a heavy rain.
Tiles have been dislodged in past launches by ice that forms on the fuel tanks. But Emero said the key is which tiles come off.
"You can withstand the loss of 20 tiles in one area, but not the loss of one tile in some areas," he said.
He said one critical area is the bottom of the vehicle, where a loss of tiles could expose the shuttle's wires, landing gear, wheel wells and other key components.
"In some areas, without the tiles, it would be like hitting the shuttle with a blow torch," Drella said.
NASA is under scrutiny for budget overruns and for cost-cutting measures that some critics say compromise safety.
The agency has increasingly relied on private contractors as it has trimmed costs, shifting work previously handled by federal employees, including much of the shuttle and unmanned space programs. "I don't think it's been a great improvement," said Robert Park, a University of Maryland expert on the space program.
In 1996, NASA handed over day-to-day operation of the shuttle fleet to United Space Alliance, which then consisted of Rockwell International and Lockheed Martin Corp. Boeing took over Rockwell's role that year.
Nine of every 10 dollars the space agency spends go to its roughly 120 subcontractors. United Space Alliance's six-year contract, from 1996 to 2002, was valued at $9 billion. Last year, it was extended for two more years for $2.9 billion. The alliance was the space agency's largest contractor in 2001, with procurements totaling $1.7 billion.
Columbia was shipped to the Boeing's Orbiter Assembly facility at Palmdale, Calif., for upgrades in 1999. It was returned to service in 2001, and resumed flying in 2002.
Dittemore pledged that investigators would go to Palmdale to re-examine every change that was made during the overhaul.
Sun staff writer Dennis O'Brien and wire services contributed to this article.