Last month, I was privileged to be able to speak at Spacefest VIII in Tucson, Arizona about the recovery and reconstruction of Columbia.
For those of you who haven’t been able to attend a Spacefest, it’s an incredible experience. You’ll meet Gemini, Apollo, and Space Shuttle astronauts, planetary scientists, futurists, historians, artists, authors, and hundreds of everyday people who are enthused by space exploration. The wonderful people at Novaspace make this an experience you’ll never forget.
As the co-author of the upcoming book on the Columbia accident, I was invited to speak about the events of 2003. Joining me on the dais was astronaut Jerry Ross, who shared his first-hand accounts of working with the crew and in the search for the vehicle’s debris after the accident, as well as the near-miss he had on STS-27—the most heavily-damaged spacecraft ever to return safely.
I recorded the audio from the presentation and from the ensuing question-and-answer period. I’ve since incorporated a few more images and some video editing to help make the photos tell the story even more clearly. Jerry Ross graciously agreed to allow me to include his commentary in this video.
I am NOT a practitioner of “Death by PowerPoint.” You’ll need to have the audio turned on as you watch this presentation, as there are no bullet-point slides—none. It’s all photos and a few maps, which illustrate the story I tell.
The video is one hour long. I hope you’ll enjoy it and learn from what one audience member called “a moving and surprisingly inspirational presentation”!
The exclamation point that provided closure to the Columbia accident investigation was independent of the analysis of Columbia‘s debris and its data recorder.
Two days after the February 1, 2003 accident, the NASA Accident Investigation Team contacted the Southwest Research Institute (SwRI) near San Antonio for assistance in the accident investigation. SwRI had conducted previous studies for NASA on the effects of impacts of much smaller pieces of foam, cork insulation, and ice on shuttle tiles. NASA wanted incontrovertible proof that foam from the external tank was capable of inflicting mortal damage on the shuttle’s thermal protection system.
That foam could damage the wing seemed counterintuitive on many levels. How could a piece of lightweight insulation— about the density of Styrofoam and weighing less than two pounds—fall off the tank and cause that kind of damage? And wasn’t it traveling about the same speed as the shuttle?
In fact, analysis showed there was a significant velocity difference between the shuttle and the foam at the time of impact. NASA estimated that the shuttle was traveling faster than 1,500 mph—and accelerating—when the foam fell off the tank. After falling off, the foam immediately and rapidly decelerated due to air resistance. The block slowed to about 1,000 mph in the 0.2 seconds between when it came off the tank and when the shuttle’s wing impacted the foam. The relative difference in speeds between the shuttle and foam was therefore more than 500 mph.
The piece of foam that struck Columbia was four hundred times larger than the pieces tested previously by SwRI. Using a special compressed air cannon, SwRI planned to simulate the collision by firing foam blocks at more than 500 mph into samples of shuttle tiles and wing leading edge panels. High-speed cameras photographed the test firings and impacts, and over two hundred sensors measured the effects of the collisions.
By the time the equipment and procedures were ready for the first test on the landing gear door, the investigation had already narrowed its focus to the wing’s leading edge as the impact area. SwRI ran its test anyway using a landing gear door—one borrowed from Enterprise and subsequently covered with silica tiles—to check out the test equipment and processes. As expected, a grazing impact of foam, akin to what would have occurred in flight had the foam hit the underside of the wing, caused only minor damage to the tiles on the landing gear door.
Space shuttle wing leading edge panels are large, expensive, and made to order. The reinforced carbon-carbon (RCC) material also wears and becomes more brittle over time, so SwRI could not use newly manufactured panels to get an accurate assessment of potential damage in its impact tests. The test panels would have to come from the wings of Discovery and Atlantis, the two orbiters that had flown about as many times as Columbia.
NASA decided to check out the test process first using Fiberglas leading edge wing panels from Enterprise, which was not designed to fly in space. Several test shots at Enterprise’s Fiberglas panels—which were stronger than the RCC panels on the flightworthy shuttles—produced scuff marks from the foam blocks, but no breakage. After getting its process and equipment calibrated, SwRI was now ready to try the tests with the space-flown RCC panels.
First, a foam block was fired at panel 6 from Discovery. The impact created a crack nearly six inches long in a rib supporting the leading edge, and it moved the panel enough to create a small gap in the T-seal between panels 6 and 7. This test proved that foam could damage the RCC. However, the damage incurred in this test would not have been severe enough to create the burn-through seen on Columbia. NASA estimated a hole of at least ten inches in diameter would have been needed for the wing to ingest a plasma stream large enough to create the damage shown in Columbia’s debris.
The next test target was panel 8, which had flown twenty-six times on Atlantis. Evidence from the reconstructed debris and the OEX recorder indicated that panel 8 was the probable site of the impact on Columbia’s wing.
At the test on Monday, July 7, 2003 the impact from the foam block blew a hole through the panel about sixteen inches by sixteen inches across, created several other cracks, and caused the T-seal to fail between panels 8 and 9. This was entirely consistent with the type of damage that caused Columbia’s demise.
Witnesses were incredulous, but the evidence was incontrovertible. NASA now had the smoking gun matching the fatal wound on Columbia. The test silenced lingering doubts that a foam strike alone was sufficient to damage the wing and doom the ship.
This text in this section is excerpted from the book “Bringing Columbia Home,” (c) 2017 by Mike Leinbach and Jonathan H. Ward. Video material is from NASA.
Throughout February, March, and April of 2003, truckloads of Columbia’s debris retrieved in East Texas arrived at the reconstruction hangar at Kennedy Space Center. In the first month or so, several truckloads arrived every week. By April, those deliveries were beginning to tail off as searchers completed clearing the debris field.
Workers in the field knew that NASA was conducting important analyses on the material they’d found and shipped to KSC, but they had no concept of the scope of the operation until they saw the reconstruction hangar themselves. NASA’s Mike Ciannilli (who now runs NASA’s Columbia Recovery Office and the Apollo 1, Challenger, and Columbia Lessons Learned Program) worked as an aerial spotter in East Texas during the recovery period. He said, “We had no clue what was going on at KSC. The first time I went to the hangar, I was blown out of the water. I was just mesmerized by it.”
By the end of April, the debris had begun to tell a very clear story about Columbia’s demise. Almost all of the debris showed the effects of extreme heat, oxidation, and mechanical stress to varying degrees. But what damage happened before the ship broke apart, and what happened afterward? Painstaking forensic analysis was required to gradually tease apart the sequence of the accident.
Reconstruction workers initially used a computer graphics package (originally developed to track the process of waterproofing the tiles during launch processing) to plot which of the insulating tiles had been recovered from Columbia’s wings. It quickly became obvious that the software was inadequate to provide the information needed for the forensic analysis. Which tiles were damaged? How were they damaged? Were there larger-scale patterns of burns, pitting, or other damage?
The solution was to lay the tiles out on a large, elevated table on which was drawn the outline of the wing and the location and part number of every single one of the 2,800 silica tiles on the left wing. As tiles were recovered and identified (sometimes based only on a few millimeters of thickness), engineers lay them in place on the table.
Patterns quickly emerged.
The carrier panel tiles, which formed the closeout between the leading edge and the rest of the wing, showed where the breach in the wing occurred. The carrier panel tiles behind Panel 9 on the leading edge were slumped and coated with metallic deposits from structures in the interior of the wing.
Burn patterns showed that plasma entering the breach in the wing at thousands of miles per hour pressurized the wing cavity and blew out vents in the upper and lower surfaces of the wing.
As the hot plasma entered and melted its way through the wing’s interior, its heat caused the adhesive that held the tiles onto the wing surface to fail. Those tiles—designed to take heat from the outside surfaces, not from where they were glued onto the ship—peeled off of the wing.
The wing was clearly baking from the inside out.
Finally, the tiles found farthest west in the debris field—those which had come off the earliest in the moments leading to the breakup, as the shuttle was flying from west to east across Texas—were all from the left wing, and all from the area behind leading edge panels 8 and 9.
Leading Edge Frames
Boeing’s Mike Gordon and NASA structures engineer Lyle Davis spent many long hours examining the recovered pieces of the reinforced carbon-carbon (RCC) panels on Columbia’s wing’s leading edges. These pieces ranged in size from larger than one square foot to smaller than a thumbnail.
Since the RCC panels were U-shaped and several feet long, it was impractical to lay them out two-dimensionally. Kennedy Space Center’s shops constructed a set of clear Lexan frames that held the pieces of each RCC panel and its underlying structure, allowing people to see in three dimensions what the front edge of the wing looked like.
The RCC frames also told a compelling story about the accident. Panels 1 though 7, and from Panel 10 outboard, were fractured from forces after the shuttle disintegrated. Their metallic attachment fittings were still partially intact.
However, at Panels 8 and 9, no metallic fittings were recovered. The stainless steel support structure—which melts at 2,500°F—had melted away and was deposited as slag onto the interior of adjacent panels on the wing. The retrieved pieces of RCC from Panels 8 and 9 were knife-edged and heavily eroded, providing evidence that plasma at over 3,000°F was acting as a blowtorch at high pressure on those panels for a prolonged period. This was the only place on the wing where this pattern was observed.
Materials scientists analyzed the cross-section of the slag deposits inside the left wing panels. This analysis showed what the metals were (insulating blankets, aluminum from wing spars, Inconel and steel from fittings, etc.) and how the deposits were laid down. It clearly told the story of the wing’s failure in time sequence.
As with the tiles, the location of the leading edge components in the debris field also gave the sequence of the accident. The leading edge components from the middle to the tip of the left wing (Panels 8 to 22) were found farthest west, between Dallas and Palestine. Pieces of the leading edge of the left wing closest to the shuttle’s fuselage (Panels 1 to 7) were found farther east, between Palestine and Nacogdoches. The right wing leading edge pieces were all found farther east. This provided evidence that the left wing failed before the right wing, with the most likely point of failure being at Panels 8 and 9.
NASA now had the evidence to prove how Columbia broke up.
NASA announced in early May that only one single failure scenario explained all of the evidence contained in the debris, the OEX recorder, and the telemetry received in Houston. Something—most likely the piece of foam from the external tank during ascent to orbit—caused a breach in the leading edge of the left wing, allowing plasma to penetrate the wing during reentry and erode it from the inside out.
Whether the foam actually punctured the RCC at the leading edge, or whether it pushed the spacers apart and created a gap in the panels, will never be conclusively known.
The last truckload of debris arrived at the reconstruction hangar on May 6. Wreckage that had been strewn over thousands of square miles of forest and field—pieces that had been carefully collected by tens of thousands of searchers working in tough conditions for three months—was all here in this one place. The twisted, burned, and shattered debris had told an important story, one that would have been impossible to write without the collective efforts of so many dedicated people in East Texas and Louisiana and of the nation’s wildland firefighting teams.
“Each piece was evidence of how hard Columbia fought to come home to us,” astronaut Pam Melroy said. “We saw every recovered piece as a victory.” Every piece of debris moved the reconstruction team closer to their goal: We will find the problem, fix it, and move forward in their honor.
Shortly after the accident, during the third week in February 2003, a few of us contemplated if a rescue mission of Columbia’s crew could have been conducted. If it could, what were the chances of success?
Under the guidance of Shuttle Program managers we were asked to quietly study it. We were to conduct our studies in part to satisfy our own curiosity and in part knowing the Columbia Accident Investigation Board (CAIB) would no doubt ask us one day. The Flight Directors at Johnson Space Center (JSC) would do the on-orbit assessment, and I would do the Kennedy Space Center (KSC) operations assessment. The two would combine to answer the unfriendly—but necessary—question.
My part concluded that from a pure timing perspective, a rescue was theoretically possible. The result from JSC was the same: theoretically possible. But both required unrealistic assumptions and actions that were not consistent with the mission being flown, or usual program priorities or objectives.
Rescue would have involved having us launch Atlantis—next in line to fly—as soon as possible, rendezvous with Columbia, transfer the astronauts via some sort of tether to Atlantis, and come home. The crew of seven from Columbia would be aboard Atlantis with her rescue crew of four. Four of the crew members would have to ride home strapped to the deck; there were only seven seats on the orbiter. Columbia herself would then be guided to a ditching in the ocean.
At the time of the accident, Atlantis was almost ready to roll out of the Orbiter Processing Facility to the VAB. A full-court press to expedite that and get to the launch pad would be required. The Pad “flow” would be truncated to only those tasks required, the rest omitted to save time. Things like the Terminal Countdown Demonstration Test and cryogenic loading simulation would be eliminated. Other required tasks would be done three shifts per day, seven days per week. Meanwhile, the rescue scenario flight plan would be developed at JSC.
Assuming no significant glitches, launch could have been as early as February 11. This also assumed no significant processing or launch delays occurred, including weather. That also assumed that Atlantis would not have her remote manipulator arm installed, which was almost certainly needed for a rescue mission. Installing the arm would have pushed the earliest launch date to February 13.
If everything went according to plan—and that was a BIG if—the rescue would have happened two days before Columbia‘s consumables ran out. Columbia would have been in orbit for almost a full month by then, two weeks longer than any previous Shuttle mission.
The key to the entire study was that consumables on board Columbia needed to preserved as much as possible, extending Columbia’s time on orbit awaiting Atlantis’ arrival. Food, water, etc. all needed to be stretched to the max. The limiting commodity however were the lithium hydroxide (LiOH) canisters needed to scrub carbon dioxide from the cabin air. Not food, not water, not power, but the ability to provide breathable air for Columbia’s crew.
The assumption made for the study was that we needed to put the crew on alert for extending LiOH no later than Day 4 of the mission. The crew would have had to go into a very low activity mode to keep their respiration as low as possible. This would have had the effect of terminating the mission’s objectives, effectively ending the reason for the mission. To do this would have been one of the unrealistic moves required. AND, to even get to this posture would have required either proof that the Orbiter was fatally damaged by that day, or assuming so. That was another unrealistic assumption, since the request for additional imagery didn’t occur until Day 6 of the mission, by which time it would already have been too late to conserve the consumables.
But when the two studies were combined, we saw that it would have been technically possible to rescue the crew. That’s the cold, data-driven answer. The truth is that the assumptions I mentioned above, and a few others, would have required extraordinary efforts in both ground and mission operations AND management decision making while we were lacking definitive damage information. All this would have been far outside the normal Shuttle practices at the time.
It should also be noted that the decision to actually launch the rescue mission would have been an extraordinary thing in and of itself. Would we commit a crew of four on Atlantis to rescue Columbia’s, crew possibly facing the same damaging foam loss during its launch? A tough decision to say the least, bigger than NASA alone could make. I believe the President would have had a role in that decision.
But it never got to the point that we’d find out.
No rescue mission was ever contemplated during Columbia’s time on orbit, let alone one early enough to give it a fighting chance of success. We just didn’t have the evidence to support making such a decision, and there was no realistic way in which we could have had that evidence by the time that decision needed to be made.
The CAIB asked us about the scenario in early May 2003. Admiral Gehman, a superior leader, intentionally waited to ask the question until some of the raw emotions had time to subside a little.
When we saw the analyses, there was no grumbling, but there was grief. We couldn’t save the ship. Columbia was doomed, no matter what. Maybe we could have saved the crew. But there were so many what-ifs and assumptions, so many things that had to go completely differently from the very first hours of the mission. Would it have been successful? I don’t know. But we never even had the chance to try.
As much as it hurt people to think about the remote possibility of saving Columbia’s crew, the study helped prompt discussions on how to save a future crew of a damaged shuttle. The studies led to the safe-haven scenario, in which damaged Orbiters could dock at the International Space Station to enable the crews to wait there for a later rescue mission.
KSC and JSC used the Columbia rescue scenario to design a one-time rescue mission that could back up the final Hubble servicing mission. After the successful completion of STS-121 in July 2006, proving that we’d finally solved the foam-shedding problem, NASA Administrator Mike Griffin formally approved the Hubble servicing mission.
On May 11, 2009 Atlantis was poised for launch to the Hubble from Pad 39A at Kennedy. Standing on Pad 39B two miles to the north was Endeavour, ready to go into orbit if there were any problems with Atlantis. For the first and only time, NASA had two shuttles in launch countdown simultaneously. We were ready to launch Endeavour one day after Atlantis if necessary. Tremendous dedication and work went into getting us to this dual launch posture. Fortunately—like many other things in the space business—this contingency capability was assured but never needed.
Atlantis’s flight went flawlessly, so the rescue mission never flew. Atlantis’ crew successfully prolonged Hubble’s life and upgraded its instrument package.
In a roundabout way, Columbia had once again contributed to the advancement of scientific discovery.
Fourteen years ago, in early April 2003, we were about 2/3 of the way through recovering Columbia’s debris from Texas, although we didn’t know it at the time. But the number of debris trucks arriving at the reconstruction hangar at Kennedy Space Center had begun to tail off in the preceding weeks, so we knew at some point they’d stop altogether.
Two initiatives were being worked at that time. First, what to do with the debris, and secondly, how would debris found after operations ended in Texas find its way to us? What were the people finding items after the recovery operations ended to do with the material they found?
I’ll briefly address both now, with the intent to more fully discuss them in subsequent postings.
As stated in a previous entry in this blog, Administrator O’Keefe was instrumental in the decision to learn from Columbia’s accident and in particular, from the debris. Having gotten his unofficial “go” to develop the concept to study the debris, the task to actually put the concept into practice fell on a few of us in the hangar. I asked Scott Thurston, Columbia’s NASA Vehicle Manager, to develop the necessary plans. He did an outstanding job. He and a very small group debated where to store the debris, how to “advertise” that it even existed for study, the requirements for organizations to obtain select pieces, the logistics of lending it to them (it’s not easy lending government property to private organizations), and the proper approval authorities and documents. And, by the way, how to do this for many years to come – also not easy.
The results of their labor and Scott’s leadership are clear. The debris loan program is very much still alive, with several hundred pieces either actively out for study or with studies already concluded. The material is stored in a climate-controlled room in the Vehicle Assembly Building, also allowing easy access for employees to view it. It has a full-time NASA curator—Mike Ciannilli—who also developed and runs NASA’s Apollo, Challenger, Columbia Lessons Learned Program. Mike was very active in the debris recovery as an aerial searcher in Texas, and his passion for sharing the lessons makes him the perfect person for the job.
As with debris from Challenger, some pieces of Columbia continue to be found. To deal with this in Texas, a program involving local authorities is charged with taking calls from anyone finding pieces that may be from Columbia. They in turn call Ciannilli, who is responsible for determining the authenticity of the find and returning the material to KSC to join the other 84,000+ pieces of the ship’s debris.
The most “famous” piece found in this manner was a cryogenic tank from Columbia‘s fuel cell system that had been submerged in Lake Nacogdoches since February, 2003. A severe drought in the summer of 2011 lowered the lake level to the point that the tank was high and dry.
Numerous other pieces have been found by farmers, ranchers, hikers, etc. I suspect debris will continue to be found from time to time. We know for certain that three of the six main engine turbine pumps are still out there somewhere. But like the three that we recovered, they are no doubt buried deep in the East Texas or Louisiana dirt or at the bottom of a body of water. They will probably never be found.
We officially wrapped up recovery operations in early May, 2003. The vast majority of Columbia that we will ever find is already home. And some of it is being used to advance our understanding in materials and structures subjected to extreme conditions. The goal is to design future spacecraft that can better withstand such conditions. One such example is a seat design capable of withstanding very high torsional forces.
Columbia continues her scientific and research missions, well after her last space flight. That legacy would have made her final crew proud.
There are so many aspects of the recovery and reconstruction of Columbia to discuss no one book could possibly tell them all. And this observation is just from my perspective—one person out of the many, many people who contributed one way or the other to the overall effort. Other accounts would add many more personal stories and technical information. There are at least a dozen books on the accident from varying perspectives. One thing we all agree on is that the response of NASA to the loss of Columbia and her crew was vastly different from the loss of Challenger and her crew.
What are the factors that made this true? What evidence is there supporting it? And, most significantly, why is important? Briefly recalling the two accidents begins to tell the story.
Challenger was lost January 28, 1986. It was the 25th Shuttle flight, not yet five years into the program designed to be America’s single launch system. (Recall that at the time, NASA needed to fly as many commercial and military payloads as possible to cost-justify the shuttle.) Although the public was already starting to tune out the Shuttle program, this was a high-profile mission, with “Teacher in Space” Christa McAuliffe on board. The explosion was seen in person and on TV by millions of people. It was horrific. Most of us will never forget the images of the expanding plume of vapor, the solid rocket boosters careening out of control, and then the innumerable vapor trails of objects plummeting from the vapor cloud into the ocean.
Columbia was lost February 1, 2003. It was the 113th shuttle mission, and NASA had been flying space shuttles for more than twenty years. Most Americans probably didn’t even know the mission was being flown. People were blasé about the program and this purely research mission. If you hadn’t gotten up early on that Saturday morning in Dallas or along the debris path in sparsely-populated East Texas, all you saw later that day was a few videos of some trails in the sky. That accident was equally horrific as Challenger‘s. Just as many astronauts died. But precious few people outside of NASA or Texas remember the accident at all. (As crazy as it sounds, there aren’t even any public domain, NASA-taken photographs of the accident itself, since it happened far from the nearest NASA facility.)
The responses to the two by NASA leadership were as different as the missions themselves. The reason is most often described as “the mood of the agency was different.” In 1986 NASA wanted to move on after Challenger, put the accident behind us— writing it off as a one time horrible event—and get back to flying the Shuttle. In 2003 we knew very early after the loss it would probably end the program, but we still wanted to “find the cause, fix it, and fly again.” (We were bound by international treaty to finish the International Space Station.) But we also wanted to learn from the loss.
That’s the most significant difference between Columbia and Challenger. Learn from it. Don’t pretend it didn’t happen.
So much could be said about how this was manifested, but none of it could have happened without the strong leadership of Sean O’Keefe, NASA’s Administrator in 2003. His treatment of the loss, the crew families, the Columbia Accident Investigation Board, and the ‘NASA family’ was nothing short of inspirational. It was exactly what we all needed.
The evidence of his leadership is easy to see. Columbia is preserved and used in research into the effects of hypersonic re-entry on materials and structures. It is available for study by any organization with appropriate research goals. It is stored in the VAB. It is also used to educate NASA and contractor employees about the risks of spaceflight and the need for everyone to be vigilant in doing their job the best they can at all times.
By comparison, Challenger is buried in an abandoned Minuteman missile silo at launch complex 31/32 on the Cape Canaveral Air Force Station. The debris fills the deep silo and the side rooms and is sealed with a concrete cap. It is locked away, only accessible in the most extreme cases (once for a Shuttle study on an issue with fuel system flowliner cracks). Not even a sign or marker exists to commemorate it.
This one example of the comparison of the response to the two is easy. To fully describe how it was so completely different would take another book on the organizational, political, and social influences existing at the two times.
And I’m certainly not the right person to do that or one to point fingers at the Challenger leaders and their decisions. But I am one to celebrate the response NASA had to Columbia. It was the right thing to do, lead by the right person to do it.
In later blog entries, we’ll dive deeper into the decision to preserve Columbia and the benefits that have already accrued from that decision.
One of NASA Administrator Sean O’Keefe’s first actions after Columbia disintegrated on February 1, 2003 was to activate an independent investigative board. NASA had at least fourteen formal internal task forces and teams designated to respond to the accident and investigate technical issues. However, policy mandated the formation of an independent board after a significant incident and loss of the crew. This board named itself the Columbia Accident Investigation Board, or CAIB (pronounced kabe) soon after its first meeting.
O’Keefe asked Admiral Harold “Hal” Gehman to chair the board. Gehman had recently completed an investigation into the terrorist bombing of the USS Cole. Other members of the Board included leaders from the military, NASA, FAA, research institutions, and other experts.
One of the truisms of a catastrophic accident such as Columbia‘s destruction is that it is almost never due to a single piece of hardware failing. Complex systems like the space shuttle and commercial airplanes are designed and built with all sorts of fail-safe mechanisms and components. For example, if a guidance computer fails, there is a backup (and sometimes several) that can take over. History has shown repeatedly that it usually takes a cascading series of failures to cause the destruction of something like the space shuttle.
Those failures are not always mechanical. They are often compounded by breakdown of processes and how the organizations and people work within the system.
Therefore, the CAIB was empowered to examine not only the physical causes of the accident, but also to look into any organizational, cultural, procedural, policy, design, or other factors that may have played a role.
So while NASA’s internal teams and task forces collected and analyzed the debris from the accident, looked into the telemetry that was received from Columbia in its final minutes, and ran checks on the history of the ship’s hardware, the CAIB interviewed personnel who were involved in various discussions and decisions prior to the accident.
The CAIB held the first of its public hearings on March 6, 2003. Space Shuttle Program Manager Ron Dittemore was one of the first to testify, and the CAIB questioned him at length about subject such as recent changes to program staffing that might have affected operational safety. Johnson Space Center director Jefferson Howell, like Dittemore, disagreed with any suggestion that NASA underestimated shuttle risks or that safety did not receive the highest possible emphasis. However, former Ames Research Center director Harry McDonald suggested that he believed NASA had moved too far toward considering the shuttles as “operational” vehicles rather than complex research and development spacecraft.
Testimony later in the week focused on issues such as whether the insulating foam on the shuttle’s external tank could absorb water and potentially freeze, making the foam heavier and thus more of a threat to the orbiter. Other witnesses spoke about the decision to fly two shuttle missions after Atlantis was struck by insulation on STS-112 in October 2002.
Although many of the points that were raised were difficult for NASA to hear, Administrator O’Keefe had made it abundantly clear that he wanted full and open cooperation with the investigation. The only chance NASA had of finding and fixing the issues that doomed Columbia was to bring everything out into the spotlight for intense scrutiny.
Members of the Columbia Accident Investigation Board. From left to right, seated: Scott Hubbard, Dr. James N. Hallock, Dr. Sally Ride, Board Chairman Admiral (retired) Hal Gehman, Steven Wallace, Dr. John Logsdon, Dr. Sheila Widnall. Standing from left to right: Dr. Douglas Osheroff, Maj. General John Barry, Rear Admiral Stephen Turcotte, Brig. General Duane Deal, Maj. General Kenneth W. Hess, and Roger Tetrault.