On February 1, 2003, the Space Shuttle (click for more info) Columbia disintegrated during atmospheric entry, killing all seven crew members. The disaster was the second fatal accident in the Space Shuttle program, after Challenger, which broke apart 73 seconds after liftoff and killed the seven-member crew in 1986.

During the launch of STS-107, Columbia‘s 28th mission, a piece of foam insulation broke off from the Space Shuttle external tank and struck the left wing of the orbiter. A few previous shuttle launches had seen damage ranging from minor to nearly catastrophic from foam shedding, but some engineers suspected that the damage to Columbia was more serious. NASA managers limited the investigation, reasoning that the crew could not have fixed the problem if it had been confirmed. When Columbia re-entered the atmosphere of Earth, the damage allowed hot atmospheric gases to penetrate the heat shield and destroy the internal wing structure, which caused the spacecraft to become unstable and break apart.

                                               

Rear (L-R): David Brown, Laurel Clark, Michael Anderson, Ilan Ramon; Front (L-R): Rick Husband, Kalpana Chawla, William McCool

After the disaster, Space Shuttle flight operations were suspended for more than two years, as they had been after the Challenger disaster. Construction of the International Space Station (ISS) was put on hold; the station relied entirely on the Russian Roscosmos State Space Corporation for resupply for 29 months until Shuttle flights resumed with STS-114 and 41 months for crew rotation until STS-121.

Several technical and organizational changes were made, including adding a thorough on-orbit inspection to determine how well the shuttle’s thermal protection system had endured the ascent, and keeping a designated rescue mission ready in case irreparable damage was found. Except for one final mission to repair the Hubble Space Telescope, subsequent shuttle missions were flown only to the ISS so that the crew could use it as a haven in case damage to the orbiter prevented safe reentry.

The shuttle’s main fuel tank was covered in thermal insulation foam intended to prevent ice from forming when the tank is full of liquid hydrogen and oxygen. Such ice could damage the shuttle if shed during lift-off.

Mission STS-107 was the 113th Space Shuttle launch. Planned to begin on January 11, 2001, the mission was delayed 18 times and eventually launched on January 16, 2003, following STS-113. The Columbia Accident Investigation Board determined that this delay had nothing to do with the catastrophic failure.

At 81.7 seconds after launch from Kennedy Space Center’s LC-39-A, a suitcase-sized piece of foam broke off from the external tank (ET), striking Columbias left wing reinforced carbon-carbon (RCC) panels. As demonstrated by ground experiments conducted by the Columbia Accident Investigation Board, this likely created a six-to-ten-inch-diameter (15 to 25 cm) hole, allowing hot gases to enter the wing when Columbia later re-entered the atmosphere. At the time of the foam strike, the orbiter was at an altitude of about 65,600 feet (20.0 km; 12.42 mi), traveling at Mach 2.46.

The left bipod foam ramp is an approximately three-foot-long (1 m) aerodynamic component made entirely of foam. The foam, not normally considered to be a structural material, is required to bear some aerodynamic loads. Because of these special requirements, the casting-in-place and curing of the ramps may be performed only by a senior technician. The bipod ramp (having left and right sides) was originally designed to reduce aerodynamic stresses around the bipod attachment points at the external tank, but it was proven unnecessary in the wake of the accident and was removed from the external tank design for tanks flown after STS-107 (another foam ramp along the liquid oxygen line was also later removed from the tank design to eliminate it as a foam debris source, after analysis and tests proved this change safe).

 

Close-up of the left bipod foam ramp that broke off and damaged the shuttle wing

 

Space Shuttle external tank foam block

 

Bipod ramp insulation had been observed falling off, in whole or in part, on four previous flights: STS-7 (1983), STS-32 (1990), STS-50 (1992), and most recently STS-112 (just two launches before STS-107). All affected shuttle missions completed successfully. NASA management came to refer to this phenomenon as “foam shedding”. As with the O-ring erosion problems that ultimately doomed the Space Shuttle Challenger, NASA management became accustomed to these phenomena when no serious consequences resulted from these earlier episodes. This phenomenon was termed “normalization of deviance” by sociologist Diane Vaughan in her book on the Challenger launch decision process.

As it happened, STS-112 had been the first flight with the “ET cam”, a video feed mounted on the ET for the purpose of giving greater insight to the foam shedding problem. During that launch a chunk of foam broke away from the ET bipod ramp and hit the SRB-ET attach ring near the bottom of the left solid rocket booster (SRB) causing a dent four inches wide and three inches deep in it. After STS-112, NASA leaders analyzed the situation and decided to press ahead under the justification that “the ET is safe to fly with no new concerns (and no added risk)” of further foam strikes.

Video taken during lift-off of STS-107 was routinely reviewed two hours later and revealed nothing unusual. The following day, higher-resolution film that had been processed overnight revealed the foam debris striking the left wing, potentially damaging the thermal protection on the Space Shuttle. At the time, the exact location where the foam struck the wing could not be determined due to the low resolution of the tracking camera footage.

Columbia lifting off on its final mission. The light-colored triangle visible at the base of the strut near the nose of the orbiter is the left bipod foam ramp.

Meanwhile, NASA’s judgment about the risks was revisited. Linda Ham, chair of the Mission Management Team (MMT), said, “Rationale was lousy then and still is.” Ham and Shuttle Program manager Ron Dittemore had both been present at the October 31, 2002, meeting where the decision to continue with launches was made.

Post-disaster analysis revealed that two previous shuttle launches (STS-52 and -62) also had bipod ramp foam loss that went undetected. In addition, protuberance air load (PAL) ramp foam had also shed pieces, and there were also spot losses from large-area foams.

Columbia was scheduled to land at 09:16 EST.

  • 02:30 EST, February 1, 2003: The Entry Flight Control Team began duty in the Mission Control Center.
The Flight Control Team had not been working on any issues or problems related to the planned de-orbit and re-entry of Columbia. In particular, the team had indicated no concerns about the debris that hit the left wing during ascent, and treated the re-entry like any other. The team worked through the de-orbit preparation checklist and re-entry checklist procedures. Weather forecasters, with the help of pilots in the Shuttle Training Aircraft, evaluated landing-site weather conditions at the Kennedy Space Center.
  • 08:00: Mission Control Center Entry Flight Director LeRoy Cain polled the Mission Control room for a GO/NO-GO decision for the de-orbit burn.
All weather observations and forecasts were within guidelines set by the flight rules, and all systems were normal.
  • 08:10: The Capsule Communicator (CAPCOM) (and astronaut) Charles O. Hobaugh told the crew that they were GO for de-orbit burn.
  • 08:15:30 (EI-1719): Husband and McCool executed the de-orbit burn using Columbias two Orbital Maneuvering System engines.
The Orbiter was upside down and tail-first over the Indian Ocean at an altitude of 175 miles (282 km) and speed of 17,500 miles per hour (28,200 km/h) when the burn was executed. A 2-minute, 38-second de-orbit burn during the 255th orbit slowed the Orbiter to begin its re-entry into the atmosphere. The burn proceeded normally, putting the crew under about one-tenth gravity. Husband then turned Columbia right side up, facing forward with the nose pitched up.
  • 08:44:09 (EI+000): Entry Interface (EI), arbitrarily defined as the point at which the Orbiter entered the discernible atmosphere at 400,000 feet (120 km; 76 mi), occurred over the Pacific Ocean.
As Columbia descended, the heat of reentry caused wing leading-edge temperatures to rise steadily, reaching an estimated 2,500 °F (1,370 °C) during the next six minutes. (As former Space Shuttle Program Manager Wayne Hale said in a press briefing, about 90% of this heating is the result of compression of the atmospheric gas caused by the orbiter’s supersonic flight, rather than the result of friction.)

Columbia at about 08:57. Debris is visible coming from the left wing (bottom). The image was taken at Starfire Optical Range at Kirtland Air Force Base.

  • 08:48:39 (EI+270): A sensor on the left wing leading edge spar showed strains higher than those seen on previous Columbia re-entries.
This was recorded only on the Modular Auxiliary Data System, which is similar in concept to a flight data recorder, and was not sent to ground controllers or shown to the crew.
  • 08:49:32 (EI+323)Columbia executed a planned roll to the right. Speed: Mach 24.5.
Columbia began a banking turn to manage lift and therefore limit the Orbiter’s rate of descent and heating.
  • 08:50:53 (EI+404)Columbia entered a 10-minute period of peak heating, during which the thermal stresses were at their maximum. Speed: Mach 24.1; altitude: 243,000 feet (74 km; 46.0 mi).
  • 08:52:00 (EI+471)Columbia was about 300 miles (480 km) west of the California coastline.
The wing leading-edge temperatures usually reached 2,650 °F (1,450 °C) at this point.
  • 08:53:26 (EI+557)Columbia crossed the California coast west of Sacramento. Speed: Mach 23; altitude: 231,600 feet (70.6 km; 43.86 mi).
 

Columbia debris (in red, orange, and yellow) detected by National Weather Service radar over Texas and Louisiana
The Orbiter’s wing leading edge typically reached more than 2,800 °F (1,540 °C) at this point.
  • 08:53:46 (EI+577): Various people on the ground saw signs of debris being shed. Speed: Mach 22.8; altitude: 230,200 feet (70.2 km; 43.60 mi).
The hot air surrounding the Orbiter suddenly brightened, causing a streak in the Orbiter’s luminescent trail that was quite noticeable in the pre-dawn skies over the West Coast. Observers witnessed four similar events during the following 23 seconds. Dialogue on some of the amateur footage indicates the observers were aware of the abnormality of what they were filming.
  • 08:54:24 (EI+615): The MMACS officer, Jeff Kling, informed the Flight Director that “four hydraulic fluid temperature sensors in the left wing had stopped reporting.” In Mission Control, re-entry had been proceeding normally up to this point.
  • 08:54:25 (EI+616)Columbia crossed from California into Nevada airspace. Speed: Mach 22.5; altitude: 227,400 feet (69.3 km; 43.07 mi).
Witnesses observed a bright flash at this point and 18 similar events in the next four minutes.
  • 08:55:00 (EI+651): Nearly 11 minutes after Columbia re-entered the atmosphere, wing leading-edge temperatures normally reached nearly 3,000 °F (1,650 °C).
  • 08:55:32 (EI+683)Columbia crossed from Nevada into Utah. Speed: Mach 21.8; altitude: 223,400 feet (68.1 km; 42.31 mi).
  • 08:55:52 (EI+703)Columbia crossed from Utah into Arizona.
  • 08:56:30 (EI+741)Columbia began a roll reversal, turning from right to left over Arizona.
  • 08:56:45 (EI+756)Columbia crossed from Arizona to New Mexico. Speed: Mach 20.9; altitude: 219,000 feet (67 km; 41.5 mi).
  • 08:57:24 (EI+795)Columbia passed just north of Albuquerque.
  • 08:58:00 (EI+831): At this point, wing leading-edge temperatures typically decreased to 2,880 °F (1,580 °C).
  • 08:58:20 (EI+851)Columbia crossed from New Mexico into Texas. Speed: Mach 19.5; altitude: 209,800 feet (63.9 km; 39.73 mi).
At about this time, the Orbiter shed a Thermal Protection System tile, the most westerly piece of debris that has been recovered. Searchers found the tile in a field in Littlefield, Texas, just northwest of Lubbock.
  • 08:59:15 (EI+906): MMACS informed the Flight Director that “pressure readings on both left main landing-gear tires were indicating “off-scale low”.”
“Off-scale low” is a reading that falls below the minimum capability of the sensor, and it usually indicates that the sensor has stopped functioning, due to internal or external factors, not that the quantity it measures is actually below the sensor’s minimum response value.
  • 08:59:32 (EI+923): A broken response from the mission commander was recorded: “Roger, uh, bu – [cut off in mid-word] …” It was the last communication from the crew and the last telemetry signal received in Mission Control. The Flight Director then instructed the Capsule Communicator (CAPCOM) to let the crew know that Mission Control saw the messages and was evaluating the indications, and added that the Flight Control Team did not understand the crew’s last transmission.
  • 08:59:37 (EI+928): Hydraulic pressure, which is required to move the flight control surfaces, was lost at about 08:59:37. At that time, the Master Alarm would have sounded for the loss of hydraulics, used to move flight control surfaces. The shuttle would have started to roll and yaw uncontrollably, and the crew would have become aware of the serious problem.
  • 09:00:18 (EI+969): Videos and eyewitness reports by observers on the ground in and near Dallas indicated that the Orbiter had disintegrated overhead, continued to break up into smaller pieces, and left multiple ion trails, as it continued eastward. In Mission Control, while the loss of signal was a cause for concern, there was no sign of any serious problem. Before the orbiter broke up at 09:00:18, the Columbia cabin pressure was nominal and the crew was capable of conscious actions. Although the crew module remained mostly intact through the breakup, it was damaged enough that it lost pressure at a rate fast “enough to incapacitate the crew within seconds”, and was completely depressurized no later than 09:00:53.
  • 09:00:57 (EI+1008): The crew module, intact to this point, was seen breaking into small subcomponents. It disappeared from view at 09:01:10. The crew members, if not already dead, were killed no later than this point.
  • 09:05: Residents of north central Texas, particularly near Tyler, reported a loud boom, a small concussion wave, smoke trails and debris in the clear skies above the counties east of Dallas.
  • 09:12:39 (EI+1710): After hearing of reports of the orbiter being seen to break apart, Entry Flight Director LeRoy Cain declared a contingency (events leading to loss of the vehicle) and alerted search-and-rescue teams in the debris area. He called on the Ground Controller to “lock the doors”, meaning no one would be permitted to enter or leave until everything needed for investigation of the accident had been secured. Two minutes later, Mission Control put contingency procedures into effect.

Initial investigation

 

Mock-up of a Space Shuttle leading edge made with an RCC-panel taken from Atlantis. Simulation of known and possible conditions of the foam impact on Columbia‘s final launch showed brittle fracture of RCC. NASA Space Shuttle Program Manager Ron Dittemore reported that “The first indication was loss of temperature sensors and hydraulic systems on the left wing. They were followed seconds and minutes later by several other problems, including loss of tire pressure indications on the left main gear and then indications of excessive structural heating”. Analysis of 31 seconds of telemetry data which had initially been filtered out because of data corruption within it showed the shuttle fighting to maintain its orientation, eventually using maximum thrust from its Reaction Control System jets. The investigation focused on the foam strike from the very beginning. Incidents of debris strikes from ice and foam causing damage during take-off were already well known, and had damaged orbiters, most noticeably during STS-45, STS-27, and STS-87. After the loss of Columbia, NASA concluded that mistakes during installation were the likely cause of foam loss, and retrained employees at Michoud Assembly Facility in Louisiana to apply foam without defects. Tile damage had also been traced to ablating insulating material from the cryogenic fuel tank in the past.

Columbia Accident Investigation Board

Following protocols established after the loss of Challenger, an independent investigating board was created immediately after the accident. The Columbia Accident Investigation Board, or CAIB, was chaired by retired U.S. Navy Admiral Harold W. Gehman, Jr., and consisted of expert military and civilian analysts who investigated the accident in detail. Columbia‘s flight data recorder was found near Hemphill, Texas, on March 19, 2003. Unlike commercial jet aircraft, the space shuttles did not have flight data recorders intended for after-crash analysis. Instead, the vehicle data were transmitted in real time to the ground via telemetry. Since Columbia was the first shuttle, it had a special flight data OEX (Orbiter EXperiments) recorder, designed to help engineers better understand vehicle performance during the first test flights. After the initial Shuttle test-flights were completed, the recorder was never removed from Columbia, and it was still functioning on the crashed flight. It recorded many hundreds of parameters, and contained very extensive logs of structural and other data, which allowed the CAIB to reconstruct many of the events during the process leading to breakup. Investigators could often use the loss of signals from sensors on the wing to track how the damage progressed. This was correlated with forensic debris analysis conducted at Lehigh University and other tests to obtain a final conclusion about the probable course of events.

Beginning on May 30, 2003, foam impact tests were performed by Southwest Research Institute. They used a compressed air gun to fire a foam block of similar size and mass to that which struck Columbia, at the same estimated speed. To represent the leading edge of Columbia‘s left wing, RCC panels from NASA stock, along with the actual leading-edge panels from Enterprise , which were fiberglass, were mounted to a simulating structural metal frame. At the beginning of testing, the likely impact site was estimated to be between RCC panel 6 and 9, inclusive. Over many days, dozens of the foam blocks were shot at the wing leading edge model at various angles. These produced only cracks or surface damage to the RCC panels.

During June, further analysis of information from Columbia‘s flight data recorder narrowed the probable impact site to one single panel: RCC wing panel 8. On July 7, in a final round of testing, a block fired at the side of an RCC panel 8 created a hole 16 by 16.7 inches (41 by 42 cm) in that protective RCC panel. The tests demonstrated that a foam impact of the type Columbia sustained could seriously breach the thermal protection system on the wing leading edge.

Conclusions

On August 26, 2003, the CAIB issued its report on the accident. The report confirmed the immediate cause of the accident was a breach in the leading edge of the left wing, caused by insulating foam shed during launch. The report also delved deeply into the underlying organizational and cultural issues that led to the accident. The report was highly critical of NASA’s decision-making and risk-assessment processes. It concluded the organizational structure and processes were sufficiently flawed and that a compromise of safety was expected no matter who was in the key decision-making positions. An example was the position of Shuttle Program Manager, where one individual was responsible for achieving safe, timely launches and acceptable costs, which are often conflicting goals. The CAIB report found that NASA had accepted deviations from design criteria as normal when they happened on several flights and did not lead to mission-compromising consequences. One of those was the conflict between a design specification stating that the thermal protection system was not designed to withstand significant impacts and the common occurrence of impact damage to it during flight. The board made recommendations for significant changes in processes and organizational culture.

On December 30, 2008, NASA released a further report, entitled Columbia Crew Survival Investigation Report, produced by a second commission, the Spacecraft Crew Survival Integrated Investigation Team (SCSIIT). NASA had commissioned this group, “to perform a comprehensive analysis of the accident, focusing on factors and events affecting crew survival, and to develop recommendations for improving crew survival for all future human space flight vehicles.” The report concluded that: “The Columbia depressurization event occurred so rapidly that the crew members were incapacitated within seconds, before they could configure the suit for full protection from loss of cabin pressure. Although circulatory systems functioned for a brief time, the effects of the depressurization were severe enough that the crew could not have regained consciousness. This event was lethal to the crew.”

The report also concluded:

  • The crew did not have time to prepare themselves. Some crew members were not wearing their safety gloves, and one crew member was not wearing a helmet. New policies gave the crew more time to prepare for descent.
  • The crew’s safety harnesses malfunctioned during the violent descent. The harnesses on the three remaining shuttles were upgraded after the accident.

The key recommendations of the report included that future spacecraft crew survival systems should not rely on manual activation to protect the crew.

Other contributing factors

Unintended consequences of decisions contributed to the failure: the original tank white paint was removed to save 600 lb (270 kg), exposing the rust-orange-colored foam; the tank foam chemical composition was altered to meet Environmental Protection Agency requirements, weakening it; upgrades to the leading edge proposed in the early 1990s were not funded because NASA was working on the later-cancelled VentureStar single-stage-to-orbit shuttle replacement.

Recovered power-head of one of Columbia’s main engines

Possible emergency procedures

One question of special importance was whether NASA could have saved the astronauts had they known of the danger. This would have to involve either rescue or repair – docking at the International Space Station for use as a haven while awaiting rescue (or to use the Soyuz to systematically ferry the crew to safety) would have been impossible due to the different orbital inclination of the vehicles.

The CAIB determined that a rescue mission, though risky, might have been possible provided NASA management had taken action soon enough. Normally, a rescue mission is not possible, due to the time required to prepare a shuttle for launch, and the limited consumables (power, water, air) of an orbiting shuttle. Atlantis was well along in processing for a planned March 1 launch on STS-114, and Columbia carried an unusually large quantity of consumables due to an Extended Duration Orbiter package. The CAIB determined that this would have allowed Columbia to stay in orbit until flight day 30 (February 15). NASA investigators determined that Atlantis processing could have been expedited with no skipped safety checks for a February 10 launch. Hence, if nothing went wrong, there was a five-day overlap for a possible rescue. As mission control could deorbit an empty shuttle, but could not control the orbiter’s reentry and landing, it would likely have sent Columbia into the Pacific Ocean; NASA later developed the Remote Control Orbiter system to permit mission control to land a shuttle. NASA investigators determined that on-orbit repair by the shuttle astronauts was possible but overall considered “high risk”, primarily due to the uncertain resiliency of the repair using available materials and the anticipated high risk of doing additional damage to the Orbiter. Columbia did not carry the Canadarm, or Remote Manipulator System, which would normally be used for camera inspection or transporting a spacewalking astronaut to the wing. Therefore, an unusual emergency extra-vehicular activity (EVA) would have been required. While there was no astronaut EVA training for maneuvering to the wing, astronauts are always prepared for a similarly difficult emergency EVA to close the external tank umbilical doors located on the orbiter underside, which is necessary for reentry in the event of failure. Similar methods could have reached the shuttle left wing for inspection or repair. For the repair, the CAIB determined that the astronauts would have to use tools and small pieces of titanium, or other metal, scavenged from the crew cabin. These metals would help protect the wing structure and would be held in place during re-entry by a water-filled bag that had turned into ice in the cold of space. The ice and metal would help restore wing leading edge geometry, preventing a turbulent airflow over the wing and therefore keeping heating and burn-through levels low enough for the crew to survive re-entry and bail out before landing. The CAIB could not determine whether a patched-up left wing would have survived even a modified re-entry, and concluded that the rescue option would have had a considerably higher chance of bringing Columbia‘s crew back alive.

In 2008, NASA released a detailed report on survivability aspects of the Columbia reentry. In 2014, NASA released a further report detailing the aeromedical aspects of the disaster. The crew were exposed to five lethal events in the following order:

Depressurization

After the initial loss of control, Columbia‘s “cabin pressure was nominal and the crew was capable of conscious actions”. During this period the crew attempted to regain control of the shuttle. As Columbia spun out of control, aerodynamic forces caused the orbiter to yaw to the right, exposing its underside to extreme aerodynamic forces and causing the orbiter to break up. Depressurization began when the shuttle forebody separated from the midbody 41 seconds after loss of control. The crew module pressure vessel was penetrated when it collided with the fuselage, and the “depressurization rate was high enough to incapacitate the crewmembers within seconds so that they were unable to perform actions such as lowering their visors.” The crew lost consciousness, suffering massive pulmonary barotrauma, ebullism and cessation of respiration.

Off-nominal dynamic G environment

The shuttle’s separated nose section rotated unsteadily about all three axes. The crew (now unconscious or dead) were unable to brace against this motion, and were also harmed by aspects of their protective equipment:

  • Lack of upper-body and arm/leg restraints: the crew’s torsos were free to move because “the strap velocity was lower than the locking threshold velocity of the inertia reel system” and because the seat restraints did not prevent lateral movement. Fractures consistent with flailing arms and legs were also observed.
  • Non-conformal helmets: unlike a racing helmet, the ACES suit helmets allowed the crew’s heads to move inside the helmet, causing blunt force trauma during collisions. The helmet neck ring acted as a fulcrum for cervical vertebrae fractures as the skull whipped backwards, as well as inflicting jaw injuries when wind blasted the helmet off.

Separation of the crew members from the crew module and the seats

As the crew module disintegrated, the crew received lethal trauma from their seat restraints and were exposed to the hostile aerodynamic and thermal environment of re-entry, as well as molten Columbia debris.

Exposure to high-speed / high-altitude environment

After separation from the crew module, the deceased crewmembers entered an environment with “lack of oxygen, low atmospheric pressure, high thermal loads as a result of deceleration from high Mach numbers, shock wave interactions, aerodynamic accelerations, and exposure to cold temperatures.” NASA stated that despite not being certified for those conditions, the ACES suit “may potentially be capable of protecting the crew” above 100,000 feet,  although in Columbia’s case they had already been destroyed by the cabin’s thermal environment during breakup.

Ground impact

The crew members had lethal-level injuries sustained from ground impact. The official NASA report omitted some of the more graphic details on the recovery of the remains; witnesses reported finds such as a human heart, a portion of an upper torso, and parts of femur bones.

All evidence indicated that crew error was in no way responsible for the disintegration of the orbiter, and they had acted correctly and according to procedure at the first indication of trouble. Although some of the crew were not wearing gloves or helmets during reentry and some were not properly restrained in their seats, doing these things would have added nothing to their survival chances other than perhaps keeping them alive and conscious another 30 or so seconds.

Crew cabin video

Among the recovered items was a videotape recording made by the astronauts during the start of re-entry. The 13-minute recording shows the flight crew astronauts conducting routine re-entry procedures and joking with each other. None gives any indication of a problem. In the video, the flight-deck crew puts on their gloves and passes the video camera around to record plasma and flames visible outside the windows of the orbiter (a normal occurrence). At one point on the tape, Dr. Clark participated, as Mission Control asked her to perform some small task. She replied that she was currently occupied but would get to it in a minute. “Don’t worry about it,” she was told. “You have all the time in the world.” The recording, which on normal flights would have continued through landing, ends about four minutes before the shuttle began to disintegrate and 11 minutes before Mission Control lost the signal from the orbiter.

Following the loss of Columbia, the space shuttle program was suspended. The further construction of the International Space Station (ISS) was also delayed, as the space shuttles were the only available delivery vehicle for station modules. The station was supplied using Russian unmanned Progress ships, and crews were exchanged using Russian-manned Soyuz spacecraft, and forced to operate on a skeleton crew of two.

Less than a year after the accident, President Bush announced the Vision for Space Exploration, calling for the space shuttle fleet to complete the ISS, with retirement by 2010 following the completion of the ISS, to be replaced by a newly developed Crew Exploration Vehicle for travel to the Moon and Mars. NASA planned to return the space shuttle to service around September 2004; that date was pushed back to July 2005.

On July 26, 2005, at 10:39 EST, Space Shuttle Discovery cleared the tower on the “Return to Flight” mission STS-114, marking the shuttle’s return to space. Overall the STS-114 flight was highly successful, but a similar piece of foam from a different portion of the tank was shed, although the debris did not strike the Orbiter. Due to this, NASA once again grounded the shuttles until the remaining problem was understood and a solution implemented. After delaying their re-entry by two days due to adverse weather conditions, Commander Eileen Collins and Pilot James M. Kelly returned Discovery safely to Earth on August 9, 2005. Later that same month, the external tank construction site at Michoud was damaged by Hurricane Katrina. At the time, there was concern that this would set back further shuttle flights by at least two months and possibly more.

The actual cause of the foam loss on both Columbia and Discovery was not determined until December 2005, when x-ray photographs of another tank showed that thermal expansion and contraction during filling, not human error, caused cracks that led to foam loss. NASA’s Hale formally apologized to the Michoud workers who had been blamed for the loss of Columbia for almost three years.

The second “Return to Flight” mission, STS-121, launched on July 4, 2006, at 14:37:55 (EDT), after two previous launch attempts were scrubbed because of lingering thunderstorms and high winds around the launch pad. The launch took place despite objections from its chief engineer and safety head. This mission increased the ISS crew to three. A 5-inch (130 mm) crack in the foam insulation of the external tank gave cause for concern, but the Mission Management Team gave the go for launch. Space Shuttle Discovery touched down successfully on July 17, 2006, at 09:14:43 (EDT) on Runway 15 at the Kennedy Space Center.

On August 13, 2006, NASA announced that STS-121 had shed more foam than they had expected. While this did not delay the launch for the next mission—STS-115, originally set to lift off on August 27 —the weather and other technical glitches did, with a lightning strike, Hurricane Ernesto and a faulty fuel tank sensor combining to delay the launch until September 9. On September 19, landing was delayed an extra day to examine Atlantis after objects were found floating near the shuttle in the same orbit. When no damage was detected, Atlantis landed successfully on September 21.

The Columbia Crew Survival Investigation Report released by NASA on December 30, 2008, made further recommendations to improve a crew’s survival chances on future space vehicles, such as the then planned Orion spacecraft. These included improvements in crew restraints, finding ways to deal more effectively with catastrophic cabin depressurization, more “graceful degradation” of vehicles during a disaster so that crews will have a better chance at survival, and automated parachute systems.

A makeshift memorial at the main entrance to the Lyndon B. Johnson Space Center in Houston, Texas

Columbia memorial on Mars Exploration Rover Spirit

Columbias window frames on display as part of the “Forever Remembered” installation at Kennedy Space Center Visitor Complex in 2018

Close Menu