NYT, 17 Feb 2003

After Liftoff, Uncertainty and Guesswork

By DAVID BARSTOW

 HOUSTON, Feb. 15 — LeRoy E. Cain, the crisp, soft-spoken flight director guiding the space shuttle Columbia home, rubbed his lip, then his cheek, then his nose, trying to piece together the meaning of sensor readings that suggested trouble in the left wing.

 The left wing. For days after the Columbia was launched, small teams of engineers and technicians from NASA and the shuttle program's contractors had been consumed by the left wing. Some 81 seconds after liftoff, the wing had been struck by perhaps the largest piece of debris ever to hit a space shuttle. Roughly the size and weight of an empty briefcase, it had slammed the wing's underside at about 500 miles per hour, smashing into fragile, easily damaged tiles that protected the Columbia from searing heat.

 The public knew nothing of the effort to nail down the precise dimensions and location of the damage. NASA's highest officials were only dimly aware of engineering detective work. But Mr. Cain, 39, had been part of the deliberations. He, too, had ultimately dismissed the debris strike as a potential threat.

Now, as the Columbia streaked across California and toward oblivion over Texas, Mr. Cain's mind flashed to its liftoff, and the moment that mysterious chunk of debris exploded against the left wing. "That was the first thing that entered my mind," he said in an interview, recalling that terrible Saturday morning.

In the two weeks since the Columbia disappeared from his monitors, Mr. Cain has not seen anything that has caused him to doubt the conclusion he and other managers had reached. "On the overall, I feel confident in our processes," he said.

Much is still unknown about what doomed the Columbia. Some experts favor the theory that the left wing might have been crippled in space by a micrometeoroid. But investigators inside and outside NASA are aggressively pursuing the possibility that the debris strike played a role in the craft's demise, particularly as evidence mounts that superhot plasma might have penetrated the wing in the area where the debris struck.

While still standing by their initial conclusions that the debris posed no threat to the shuttle, Mr. Cain and other senior NASA managers have been humbled enough to acknowledge that some of NASA's best and brightest might have gotten it wrong.

The teams had worked under crushing time constraints. Using sophisticated computer models, borrowing from old research and consulting with the Columbia's crew, they sought answers to exceedingly complex questions. Yet an examination of their efforts, based on interviews and a review of documents central to their analysis, shows that the process was filled with uncertainty and guesswork. Judgments were made on the basis of incomplete information or with extrapolations based on data compiled from very different circumstances.

 Theirs was a series of decisions and assumptions carried out in the borderland between optimism and fatalism — optimism because no debris strike had ever imperiled a shuttle mission, and fatalism that there was nothing to be done anyway if the damage was grave. But among the many pointed questions now being raised about the analysis of the debris strike is this one: Were the optimism and the fatalism misplaced?

 "My only question to the team is how complacent did the team become at looking at these things," said Carl Meade, a retired astronaut who flew on the Columbia in 1992. "And if complacency ruled the day, that's atrocious. That's a shame."

 He added: "If you hit a bug going Mach 2, you're going to damage a tile. I don't buy the arguments that say that it was not a big deal."

 Officials from Boeing and United Space Alliance, two of the major contractors most involved in the analysis, declined to speak about their work in any detail, other than to defend it as thorough and complete. Some outside experts also praised the studies. "It looked like a reasonable quick study," Seymour C. Himmel, a former member of a NASA advisory board, said.

 Mr. Cain pointed out that NASA's conclusions were also supported by those with the most at stake. "They knew we looked at it, analyzed it, put our best minds on it," he said of the Columbia's crew. "I believe we uplinked some video to them on e-mail so they could see the debris strike."

 The crew was asked to weigh in.

 "They were very comfortable with what we had told them and with what we had done," Mr. Cain said. `'My overall sense is that they were not concerned."

 "We had a beautiful launch," Phil Engelauf, launching director at the Johnson Space Center, told reporters on Friday, Jan. 17, the day after the Columbia lifted off from Cape Canaveral. "We are very, very clean."

 That same day, Armando Oliu and the other members of the debris and ice team began to sift through film of the takeoff. Frame by frame, they searched for anything unusual, including bits of debris that might have flaked off the shuttle's external boosters and struck its tiles. It was a routine precaution. The tiles, 90 percent air, are so fragile — one can carve into them with a fingernail — that they suffer dozens of nicks, dings and small gouges on each flight.

Major damage is rare. Only 11 times have shuttles returned with what NASA considered significant tile damage. Never had a debris strike on a tile been judged a flight risk. Still, years of study had taught NASA not to be complacent: even a few missing or badly damaged tiles in precisely the wrong spot could doom a shuttle.

 Some 81 seconds into the film, Mr. Oliu's team saw it. Where the shuttle's nose joins the external tank, a large light-colored piece of debris seemed to peel away from the tank and fall toward the leading edge of the left wing. According to an internal mission report, the chunk broke into "a spray of white-colored particles" on the wing's underside.

One thing in that report struck Ron D. Dittemore, the shuttle program manager. It was the word "white." He recalls circling it and writing "light" nearby so he would remember to ask about it.

 The distinction was critical.

 Mr. Dittemore and other managers strongly suspected that the debris was foam insulation shed by the external tank. The insulation — which has a long history of breaking free during liftoff — is an orange-coated plastic foam-type material. But the word "white" suggested the debris might have been ice from rain or condensation on the external tank. And ice, Mr. Dittemore knew, could do far more damage than foam.

Yet as he would later point out, he also knew that inspectors had not seen any dangerous ice buildup on the external tank before liftoff. In the days to come, he and NASA's engineers would proceed on the assumption that the debris had indeed been foam.

 Samuel T. Beddingfield, a retired NASA manager and one of the original designers of the shuttle, is among those who are not so sure. "It looks like ice," he said in an interview. "If ice hits tiles, it's going to bust the tile. I don't see anything else that makes more sense than that."

But where exactly had the debris hit? In some spots on the wing, NASA had determined that the shuttle could safely withstand a "burn-through" — jets of superhot gases boring through the aluminum skin like a blowtorch. But in others — over the wheel wells, say, or back by the flaps that controlled flight — burn-through was a terrifying prospect.

 In short, the margin between safe landing and catastrophe could be a matter of inches.

 NASA's engineers spent days scrutinizing film from the nearly 200 cameras trained on the liftoff. They were encouraged by the absence of obvious signs of damage. Still, NASA officials said, they could not pinpoint the hit. They could not make out individual tiles, much less the extent of the damage.

One of those cameras — at Cocoa Beach, 15 miles south of the launching pad — had a prime view of the shuttle's left wing. But it had been mounted on a faulty telescope, and as it had swung up to chase the shuttle it had never come into focus.

There was a way to get a better look.

 NASA managers could have enlisted powerful Air Force telescopes in Hawaii and New Mexico. The option was discussed and rejected as pointless. In 1998, NASA had used the Hawaii telescope to inspect Senator John Glenn's shuttle, which had lost a drag parachute door on liftoff.

"We reviewed the pictures, and they did not reveal a lot of granularity that would help us," Mr. Dittemore said. "We knew that as our background."

 But today both facilities have far more powerful telescopes than they had in 1998. Under good atmospheric conditions, those telescopes could have zeroed in on individual tiles, managers from both facilities said in interviews.

 It is not clear if NASA's mission managers were aware of the stronger telescopes, but they have acknowledged that their decision making was influenced by a second assumption: "Even if I had information, I can't do anything about it," Mr. Dittemore told reporters the day after the accident.

 Either way, Mr. Dittemore's team did not end its inquiry with the blurry film.

 Troubling News

On Day 5 of the mission, Jan. 20, the engineers went to work.

 Their goal was to use sophisticated computer modeling techniques to predict just where the debris hit, and the angle and velocity of the impact. With this information, they would create a detailed portrait of the potential damage and danger.

 They had used similar techniques on past missions, and in less than 24 hours, records show, a team of five Boeing engineers had completed a preliminary assessment. At the daily news briefing on Jan. 21, a NASA spokesman said he had "no significant anomalies" to report; the only "annoyance" facing the Columbia, he said, was overheating in the crew areas. Even so, the engineers' findings contained some troubling news. The predicted impact area included several vulnerable points on the wing, including the leading edge and the left wheel well.

 Two days later, a second Boeing team completed a far more extensive damage assessment.

 Looking back at past missions, the team noted that gouges of less than a quarter of the tile thickness were "within flight experience." But when gouges penetrate more than half the tile, the engineers wrote, "temperatures on structure may exceed limits." When tiles are virtually obliterated, burn-through can occur.

 To predict the depth of the gouges on the Columbia, the team relied on Crater, a NASA software program, as its "official evaluation tool." The program is based on laboratory data gleaned from blasting bits of foam at tiles.

 The Boeing team did not accept Crater's predictions at face value. The program was "designed to be conservative," they noted, and so has sometimes "overpredicted penetration of tile coating significantly."

 Another factor in their judgment was data from a 1992 Columbia mission that they believed had absorbed a "potentially" similar debris impact. In that case, the debris left a gouge only a half-inch deep.

 "So you can understand some of our database and some of our thinking that when this happened before, yes, it can impact the tile," Mr. Dittemore told reporters. "Yes, it could take some of the coating off. Yes, it can even gouge out some of the tile. But it has never represented anything more than that."

 There was a difference between the flights, though. In 1992, the debris was thought to have struck at an angle of 3.2 degrees. This time, the predicted angles were far sharper, which the Boeing team was careful to point out would alone significantly increase the damage. Still, for reasons not spelled out, the engineers concluded that "even for worst case," the foam debris would leave behind at least the last quarter-inch of tile, giving the wing at least some thermal protection.

 As the Boeing team noted, previous NASA research had documented the dangers of ice debris: if it struck the wing's leading edge at a sharp enough angle — greater than 15 degrees — it could penetrate the tough carbon coating that protects the edge from re-entry heat, potentially leading to a catastrophic burn-through. In the Columbia's case, NASA calculated that the debris could have hit the edge at an angle of 22 degrees.

 Outside experts, including many with deep institutional ties to NASA, are expressing polite concern that the engineers erred by not considering more possibilities.

"It was very good as far as it went," Dr. William Schneider, former assistant director of engineering for NASA, said of two reports prepared by Boeing engineers. But, he added, "They still needed to do more tests."

Others are troubled that the Boeing team appeared to discount Crater's predictions. "There's nothing here that makes me feel better," said Paul S. Fischbeck, an engineering professor at Carnegie-Mellon University who has co-written two NASA studies on tile damage. "If I saw this report, what I would say is that we are at least out on the edge of danger on re-entry."

Having calculated the dimensions of the damage, the engineers struggled to understand its effect during re-entry. For this, they turned to NASA studies overseen by Dr. Schneider, now a professor of mechanical engineering at Texas A&M University.

In the early 1990's, Dr. Schneider's team mapped the entire surface of the shuttle and assessed the implications of damage by micrometeoroids or small bits of orbital debris. They established detailed guidelines for how much damage the shuttle could absorb and still make it home.

  In four cases, they concluded that the shuttle would return safely, though perhaps requiring significant repairs.

 But for the two direst possibilities — those involving the loss of multiple tiles — the engineers reached no conclusion in their Jan. 23 report. Indeed, they said their overall conclusion, that the Columbia would return safely, was "contingent" on the multiple-tile-loss analysis "showing no violation" of Dr. Schneider's guidelines.

 It is unclear what that final bit of analysis concluded. Kari Allen, a spokeswoman for Boeing, said that "all open work was completed," and that the Boeing team had given mission managers an oral briefing on Jan. 24. She declined to say what the completed analysis showed and said Boeing had no written records of the analysis.

 Three days later, the Columbia's mission management team, with the support of NASA's senior managers, signed off on the findings of the technical teams. According to Mr. Dittemore, everyone involved was in agreement. But even then, he said, they asked themselves: What if they were wrong? What if the wing was crippled? Could they minimize the heat on it during re-entry?

 "We asked ourselves, `Is there any other option?,' " Mr. Dittemore recalled. "There's no other option."

The next day, NASA and the Columbia's crew paused to mark the anniversary of the Challenger disaster.

`Bad Time to Get Surprised'

While Mr. Dittemore considered the matter closed, others at NASA continued to worry about the debris strike — and the conclusion that it was of little consequence.

 Five days before the Columbia was to return, Carlisle C. Campbell, a NASA engineer, telephoned an acquaintance at NASA's Langley Research Center in Hampton, Va. The center has vast expertise in runway disasters, and Mr. Campbell, part of the landing gear team, asked how a deflated tire would affect the spacecraft's return to earth.

  "Admittedly this is over the top in many ways, but this is a pretty bad time to get surprised and have to make decisions in the last 20 minutes," Mr. Daugherty wrote.

"Your input is beneficial," another member of the team replied on Jan. 31, one day before the Columbia's re-entry. "Like everyone, we hope that the debris impact analysis is correct and all this discussion is mute [sic]."

 NASA officials play down the e-mail exchange, characterizing it what one called as routine "what-iffing" on a hypothetical scenario. They say flight controllers engage in this process as training and preparation for crises. Moreover, they say, the team had agreed with the conclusion that the foam had not done serious damage, and thus he had not bothered to pass the engineer's warning up the chain of command.

 "They weren't part of our playbook at the time, because they didn't surface, they didn't come forward," Mr. Dittemore said of the exchange.

 The matter also surfaced, very briefly, at the daily news conference.

 "Given that you may have lost a bit of tile during liftoff, I'm wondering if there's going to be anything different about the entry profile, taking that into regard?" a reporter from The Associated Press asked.