A member of the Columbia Accident Investigation Board said today, for the first time, that a foam strike during the shuttle's launching is the "most probable cause" of the disaster. He also said analysis of recovered debris indicates a large portion of the ship's left wing broke off in the shuttle's final seconds at the point where the catastrophic breach occurred.

"This is probably the first time you've ever heard me say it's highly probable that the foam is the cause of the accident," said Roger Tetrault. "It's probably the first time you've heard members of the board say it with that kind of strength. We think, when you look at the analysis of all the things that are pointing to the same area in terms of the hole versus where the foam hit, that that's a fairly compelling story."

Columbia's left wing leading edge was struck by a suitcase-size chunk of foam insulation that broke off the shuttle's external fuel tank 81 seconds after liftoff Jan. 16. The foam hit the underside of the leading edge at a velocity of 500 mph or so. Investigators have long believed the foam strike played a role in the disaster, but Tetrault's comments today were the first by a board member to elevate the impact to the level of "most probable cause."

In other developments today, NASA released nearly 10 hours of videotape and still photography shot by Columbia's crew that was recovered in the ship's wreckage over the past several weeks and months. The photography provides no fresh insights into the cause of the disaster, but it no doubt provides sad comfort to family members, giving them one more glimpse of their loved ones as they cheerfully worked through their final days in orbit.

If nothing else, the video put a fresh human face on the disaster and the high stakes involved in the CAIB's efforts to make sure its final report properly addresses the "root cause" of the mishap as well as contributing factors.

CAIB chairman Harold Gehman provided an impromptu preview of what that final report will include and said none of the recommendations would, by themselves, prevent NASA from resuming shuttle flights within nine months.

"I don't speculate on the date of return to flight," he told reporters at a news conference. "I would say that having read every word of the draft report and having gone over what might be possible recommendations, I don't see any recommendations which are so difficult to accomplish that they shouldn't be able to return to flight in six to nine months. Other than that, I wouldn't put any numbers on it."

Many NASA insiders believe launch likely will be delayed well into next year, but senior agency officials continue to hold out hope for a flight by the end of the year. One major wild card in all such speculation is what actions Congress might take when it begins considering the CAIB report later this fall.

In any case, Gehman said a "goodly portion of the report, perhaps half, is going to deal with the issue of management and management techniques at NASA."

"We will not tell NASA how to organize, we will not draw a wiring diagram for them," he said. "But we will tell them what we believe are the characteristics of what we believe to be essentially a flight development program that would help ensure safe operations. We won't tell them how to do it, but we will tell them what needs to be done."

By including the phrase "flight development program," Gehman was serving notice that the board views shuttle operations as more of a series of on-going test flights than as any sort of routine, operational program. He said the CAIB report will force NASA to address the issue of foam shedding, debris impacts and on-orbit inspection and repair of possible post-launch damage.

"I believe the report is going to suggest that you have to take action in each of four areas," he said. "First, you have to take action to either minimize or prevent as best you can foam loss and certainly you have to prevent the egregious foam loss, the big pieces.

"But we're also going to suggest you have to toughen the orbiter's ability to take debris hits because the orbiter is going to continue to take debris hits. It was designed not to, but that's now proven to be not the case. So you have to increase the orbiter's ability to take hits.

"You also - number 3 - you also have to improve your ability to recover from a hit," Gehman said. "That means you have to be able to inspect the orbiter after it's launched and if you find something wrong, you have to be able to make emergency, temporary, one-mission repairs in case the first two steps don't accomplish it."

The fourth item on his list was crew escape. While the board will not make any specific recommendations in that area, Gehman said "we're going to comment on how we got to the present status. But whether or not improvements need to be made, we're leaving that to NASA."

An interim recommendation that will require NASA to develop an in-flight ability to repair damage to heat shield tiles and reinforced carbon carbon leading edge panels is nearing completion. Its release has been held up while the board struggles to fine-tune the wording. "We're trying to state what we want to happen, not how to do it," Gehman said.

A detailed overview of NASA's efforts to develop just such a tile repair capability is available here.

Tetrault provided a fresh look at Columbia's final moments today, outlining how the shuttle's left wing broke apart in the ship's final seconds. He reviewed earlier evidence pointing toward a breach at or near reinforced carbon carbon panel No. 8, the largest of 22 such panels that make up the left wing's leading edge. The panel is located at the point where the wing's sweep angle changes and as such features a complex, curved shape.

Tetrault briefly reviewed recorded sensor data and the timing of sensor failures as a plume of super-heated air burned its way through wire bundles inside the left wing, both of which indicate a breach at or near RCC 8.

He reminded reporters that very little debris has been recovered from the area of RCC panels 8 through 10. No attachment fittings, known as spanner beams and spar fittings, have been found. No portions of the lower halves of the RCC panels themselves have been recovered. So-called knife-edge erosion, indicative of extreme heating, is present in the few fragments of RCC 8 and 9 that have been found, suggesting the breach occurred at or near the lower half of panel 8.

"The fact that there's such a significant amount of missing material from this particular area is very telling and it certainly in my mind points to a problem between panels 8 to 10," Tetrault said. "And it also points to the fact that we are likely to have burned up much of this material, particularly in the spanner beams and spar fittings from these areas."

He showed a map pinpointing where every piece of recovered wing debris had been found. Interestingly, the tiles found farthest to the west, indicating they fell away early in the shuttle's breakup, were located directly behind RCC panels 8 and 9.

"What that indicates is there was probably a breach somewhere in the 8 to 9 area, the hot gas flowed into that breach, it heated up the inside of that wing, the RTV (adhesive) which holds the tiles to the outside skin of the wing heated up and basically lost its adhesion capability at approximately 400 degrees Fahrenheit and fell off to the west," Tetrault said.

"One of the other important parts are in this area, we have a number of tiles which have a light brown deposit on them. And we've taken some preliminary chemical analysis of that light brown deposit. And what we find out from that preliminary analysis is that deposit is high in iron and it's also high in nickel.

"You may recall the spanner beams which hold the RCC, which appear to be missing, are, in fact, high in nickel," he said. "And the spar fittings, which hold the panels to the spar itself, are stainless steel, so they would be high in iron. What you can surmise from that, in fact, is the hole was somewhere in the 8 and 9 area and as the airflow was flowing in, some of this molten material from the spar fittings and from the spanner beams was in fact being deposited on the lower side of the wing. And that's what we're seeing when we do the chemical analysis."

He reminded reporters that so-called "knife-edge" erosion patterns were found on what few pieces of RCC 8 and 9 material that were recovered, indicating the flow of extreme heat from 8 toward 9. He said unusual slag deposits found on the interior surface of fragments from the upper half of RCC 8 showed high concentrations of nickel, "again indicating one of the first things to melt inside the wing leading edge after the breach occurred were the spanner beams."

The spherical slag deposits indicate molten material was splattered back on the inner surface of the panel from a breach on the lower side of RCC 8.

Showing a plot of recovered wing debris, Tetrault said debris from RCC panels 8 through 22 were found farther west than any other fragments from the left wing leading edge. Next came fragments of RCC 1 through 7, then debris from the shuttle's vertical stabilizer and finally, wreckage from the right wing.

"What this suggests is that at breakup, what we first saw was the left wing or a portion of the left wing left the aircraft, fell to the ground, followed by the tail, followed by the right wing," Tetrault said. "What's really important in all of this, when you look at the left wing it's a very long (debris) stream.

"What we believe happened is somewhere in this area a portion of the left wing came off," he said. "A portion of the left wing continued to ride with the aircraft downstream and that would have contained panels 1 through 7. And at some point later on, panels 1 through 7 began to come off the aircraft, we believe it probably ablated (melted) as much as fell off, that the pieces were just open to the airstream.

Interestingly, fragments of panel 8 were found in widely separated areas.

"That begins to indicate to us that, in fact, the breach, not only the breach but also the wing, broke apart at panel number 8 region," Tetrault said. "Otherwise, you wouldn't have those pieces strewn over that lengthy piece of territory. So those are the indicators that we have from the debris.

"I think when you look at it you could probably conclude from the debris alone that the most likely breach that we had in the wing occurred at panel No. 8 or in the vicinity of panel number 8. ... I would also include the T-seals on either side of it. So we feel fairly certain on where the breach was."

As for exactly how the left wing came apart, Tetrault said investigators don't yet known whether "half the wing that came off at area 8 or just the leading edge of half the wing."

"I think the prevailing theory right now is it was more than likely the leading edge of wing (that came off) and heavy spars held for some period later. But if you lost the leading edge, you would lose the RCC panels from 8 through 22."

Looking at all the data and how they fit together, Tetrault said, "they certainly are pointing us to the area of RCC panel 8 as being an indicator of where the breach occurred in the wing. Not only that, but as you well know we have the photographic analysis and evidence that indicates the foam struck on panels 6 through 9. And when you put all of those pieces of Swiss cheese together, it's a pretty compelling story that, in fact, the foam is the most probable cause of the shuttle accident."

Gehman said the full board has yet to decide on what terminology will be used in the panel's final report. But a final round of tests at the Southwest Research Institute in San Antonio, Texas, could go a long way toward determining how strongly the board links the foam strike to Columbia's destruction.

Using a powerful nitrogen gas cannon, researchers have been firing foam debris at a wing leading edge mockup to determine whether a strike like the one observed during Columbia's launching could actually break one of the carbon composite panels or an associated T-seal. A test shot at RCC panel No. 6, taken from the shuttle Discovery, resulted in multiple cracks and showed the impact forces were transmitted to adjacent panels and T-seals. Just as important as showing cracks were possible, the impacts showed the leading edge components responded to the impact as an integrated system.

But panel 6 is smaller than 8 and does not have the same complex curvature. As a result, CAIB member Scott Hubbard has decided to shoot at RCC panel 8 from the shuttle Atlantis. Adjacent panels and mounting hardware also will be flight articles.

"The thing that I'm trying to do with these tests, and I think the board is looking at this information as a piece of that most probable cause, is to connect that dot, from foam to breach," Hubbard said today. "So that's where I stand. I think the panel 6 tests that we did showed that we've got a plausible failure scenario, we created a substantial crack, five inches long, but we haven't created yet a breach that is like what has been described by the debris.

"Maybe I'm one step behind Roger in coming to a conclusion, but as Admiral Gehman said, that's part of our discussion here."

Gehman said the board will have to decide "what words we want to use when we describe the degree of certainty we use to say the foam caused damage to the leading edge of the left wing. Do we want to say we 'think' it did, we're 'sure' it did, it 'might have,' we think 'most likely' it did, the board is 'confident that?' I have 13 different opinions on that and at some time I'm going to have to lock everybody in a room and come out with one set of words."