For the second half of January in 2003, the seven men and women of shuttle Columbia’s STS-107 crew—Commander Rick Husband, Pilot Willie McCool, Mission Specialists Dave Brown, Mike Anderson, Laurel Clark and Kalpana “K.C.” Chawla, plus Payload Specialist Ilan Ramon of Israel—labored around-the-clock on 80 scientific experiments spanning a variety of disciplines from life sciences to fluid physics and from materials processing to Earth observations. Eighty-seven missions after the calamitous loss of Challenger, shuttles had serviced the Hubble Space Telescope (HST), docked nine times with Russia’s Mir orbital complex and begun building the International Space Station (ISS).
The shuttle, of course, remained an inherently dangerous machine. But the robustness of the four surviving orbiters—Discovery, Endeavour, Columbia herself and Atlantis—had been amply demonstrated, time and again, and their shortfalls were well understood. Or so it seemed. For those shortfalls came home to roost with horrifying suddenness two decades ago, on the morning of Saturday, 1 February 2003.
Across more than two decades of operational service, the shuttle fleet weathered last-second shutdowns on the launch pad, main engine problems during liftoff and later in the hazardous ascent regime, as well as severe Thermal Protection System (TPS) damage during the high dynamism of atmospheric re-entry. And when a briefcase-sized chunk of insulating foam was spotted on the STS-107 launch video falling from the External Tank (ET) at T+82 seconds and hitting Columbia’s left wing—at precisely the spot where Reinforced Carbon Carbon (RCC) would later guard the ship against the most severe re-entry temperatures—concern was elevated, but ultimately dismissed.
It was a dismissal as ill-judged as declaring the Titanic to be unsinkable.
The video footage from STS-107’s launch on 16 January offered little indication of what, if any, damage the foam strike had caused, save for a huge shower of particles. It was unclear if these particles originated from the impact of the foam itself or from shattered pieces of the RCC panels. If the latter, it did not bode well for Columbia’s re-entry, for the panels guarded the vehicle against the brunt of 3,000-degree-Celsius (5,400-degree-Fahrenheit) extremes during the hypersonic return to Earth.
Senior managers doubted that a foam strike—an event which had occurred on earlier missions—could conceivably be a “safety of flight” issue. This did not, however, prevent an analysis of a possible scenario in which the RCC had been breached.
On 31 January, engineer Kevin McCluney offered a hypothetical description to his colleagues at the Johnson Space Center’s (JSC) flight control team of the kind of data “signature” they could expect to receive if the worst should happen.
Let us suppose, said McCluney—outlined in depth by Michael Cabbage and William Harwood in their harrowing account of the tragedy, Comm Check—that a large hole had been punched through one of the shuttle’s RCC panels, enabling super-heated plasma to enter the airframe.
“Let’s surmise,” he told them, “what sort of signature we’d see if a limited stream of plasma did get into the wheel well [of Columbia’s main landing gear], roughly from entry interface until about 200,000 feet (60 kilometers); in other words, a 10–15-minute window.” Little could McCluney possibly have guessed that his “signature” would almost exactly mirror the events which befell Columbia on Saturday, 1 February 2003.
“First would be a temperature rise for the tires, brakes, strut actuator and the uplock actuator return…”
At 8:52:17 a.m. EST, nine minutes after entry interface—the point at which the shuttle began to encounter the tenuous upper traces of the “sensible” atmosphere—Entry Flight Director LeRoy Cain and his team saw the first unusual data on their monitors. Cain had begun his shift on console earlier that morning, with an up-tempo “Let’s go get ’em, guys,” before giving Husband the go-ahead to perform the irreversible de-orbit burn to drop Columbia out of orbit and onto an hour-long path to land at the Kennedy Space Center’s (KSC) Shuttle Landing Facility (SLF) at 9:16 a.m. EST.
Much of the entry profile was controlled by the shuttle’s General Purpose Computers (GPCs), but with 23 minutes remaining before touchdown, Maintenance, Mechanical Arm and Crew Systems (MMACS) Officer Jeff Kling spotted something peculiar in his data. It was what flight controllers termed an “off-nominal event”.
As later described by Cabbage and Harwood, Kling noticed that two downward-pointing arrows appeared next to readings from a pair of sensors deep within Columbia’s left wing. They were designed to measure hydraulic fluid temperatures in lines leading to the elevons. A few seconds later, two more sensors failed. The attention of Kling was captivated; it looked for all the world that the wiring to all four sensors had been cut.
Kling and his team tried to fathom a common “thread” to explain the fault. But none was forthcoming.
Kling spoke directly to Cain. “FYI, I’ve just lost four separate temperature transducers on the left side of the vehicle,” he began, cautiously. “Hydraulic return temperatures. Two of them on System One and one in each of Systems Two and Three.”
“Four hyd return temps?” queried Cain.
“To the left outboard and left inboard elevon.”
Cain’s thoughts mirrored those of Kling: was there a common root cause for all four sensors to have failed in such close physical and temporal proximity to one another? When Kling asserted that there was “no commonality” between the failures, Cain was perplexed, but instantly thought back to the foam strike. In subsequent interviews, he would admit that his immediate fear was that hot gas had worked its way through a breach in Columbia’s left wing and was somehow affecting the interior systems.
However, Guidance, Navigation and Control Officer (GNC) Mike Sarafin assured him that overall vehicle performance as it crossed the California-Nevada state line at 22.5 times the speed of sound remained nominal. Was Kling happy with all other hydraulic systems? Kling replied that, yes, everything else was functioning normally.
“Tire pressures would rise, given enough time, and assuming the tires don’t get holed,” continued Kevin McCluney’s chilling prediction of what might happen. “The data would start dropping out as the electrical wiring is severed.”
Suddenly, at 8:58 a.m. EST, Husband made his first radio transmission since entry interface a quarter-hour earlier. He started to call Houston, but his words were abruptly cut off. A few seconds later came a loss of temperature and pressure data from both the inboard and outboard tires of Columbia’s landing gear in the left well.
If the tires were holed or losing pressure, it was very bad news, for STS-107 was a “heavyweight” mission with the fully loaded Spacehab science module and experiment pallet. A “wheels-up” belly landing was likely not survivable. The astronauts would need to perform a never-before-tried bailout, utilizing an escape pole system implemented after Challenger, but this could not be attempted until Columbia was at much lower altitude and at much lower relative airspeed.
“Data loss would include that for tire pressures and temperatures, brake pressures and temperatures,” concluded McCluney.
After hearing Kling’s report, astronaut Charlie Hobaugh—the lead Capcom on duty that morning—called Husband to inform him of the anomalous tire pressure messages. Hobaugh also asked Husband to repeat his last comment.
But there was no reply from Columbia. By now, Cain was pressing Kling for answers on whether the messages were due to faulty instrumentation but was advised that all associated sensors were reading “off-scale-low”—they had simply stopped working.
Seconds later, at 8:59:32 a.m., Husband tried again to contact Mission Control. These were to be the last words ever received from Columbia.
“Roger,” he said, presumably acknowledging Hobaugh’s earlier pressure call, “uh, buh…” At that point, abruptly, his words were cut off in mid-sentence, together with the flow of data from the orbiter.
Communications were never restored. Thirty-two seconds later, a ground-based observer with a camcorder shot video footage of multiple debris contrails streaking like tears across the Texas sky.
With the telemetry broken, the atmosphere in Mission Control was becoming increasingly uncomfortable. Kling told Cain there was no common thread between the tire pressure messages and the hydraulic sensor failures; moreover, other instrumentation for monitoring the positions of the orbiter’s nose and main landing gear had also been lost.
As the seconds of radio silence stretched longer, Cain asked Instrumentation and Communications Officer (INCO) Laura Hoppe how long she expected the intermittent “comm” to last. She admitted that she expected some ratty comm but was surprised and puzzled by how protracted and “solid” it was.
“Columbia, Houston, comm check,” radioed Hobaugh at 9:03 a.m. His words were greeted only by static and by the echo of his own voice in the deathly-silent Mission Control. A minute later, he repeated the call. Again, there was no reply.
Half a continent away, at KSC, astronauts Jerry Ross and Bob Cabana were chatting outside the convoy commander’s van at the SLF, when they heard that communications with Columbia had been lost. At first, they were unconcerned—that is, until they were informed that powerful long-range radars at the Cape, meant to lock onto the incoming orbiter at 9:04 a.m. and track its final approach, saw nothing coming over the horizon.
In Cain’s words, that offered the final punch-in-the-stomach confirmation that all hope was lost. “That was the absolute black-and-white end,” he said later. “If the radar is looking and there’s nothing coming over the horizon, the vehicle is not there.”
Unlike an aircraft, which can adjust its flight profile to make secondary approaches, the shuttle had only one shot to make a pinpoint landing. Its trajectory through the atmosphere could be timed to the second and touchdown was expected at 9:16 a.m. Weather data also made it possible to predict how far down the runway—about 1,500 feet (460 meters)—the shuttle would land.
At the Cape, the assembled crowds saw the countdown clock tick to zero…and then begin ticking upwards again as 9:16 came and went. No trademark sonic booms had been heard. No sign of the tiny black-and-white dot of the orbiter had been seen. Veteran shuttle commander Steve Lindsey—later to become Chief of the Astronaut Office—was one of the escorts for the STS-107 families and his blood ran cold. Something was terribly wrong.
Standing next to NASA Administrator Sean O’Keefe that morning was former astronaut Bill Readdy, now serving as the agency’s Associate Administrator for Space Flight. O’Keefe would describe the former fighter pilot and veteran shuttle commander as ashen-faced and visibly trembling. Jerry Ross was a few months into his new job as head of the Cape’s Vehicle Integration Test Team (VITT) and his first act was to say a brief prayer.
In Texas, police were being inundated with 911 calls, reporting strange lights in the sky, loud explosions, and instances of falling debris. CNN quickly picked up on the stories and began reporting them. In Mission Control, however, televisions were not tuned to outside broadcasts. It was an off-duty NASA engineer, Michael Garske, who watched the shuttle pass overhead from the roadside, south of Houston, and called colleague Don McCormack in Mission Control with the news.
“Don, Don, I saw it,” Garske cried, paraphrased by Cabbage and Harwood. “It broke up!”
“Slow down,” McCormack replied. “What are you telling me?”
“I saw the orbiter. It broke up!”
Sitting behind LeRoy Cain at the same time, veteran flight director Phil Engelauf received a call from off-duty flight director Bryan Austin, who provided first-hand testimony. By this time, although no one in Mission Control had physically seen the evidence of the disaster, they had resigned themselves to it.
At 9:05 a.m., Cain asked Flight Dynamics Officer (FDO) Richard Jones when he could expect tracking from the long-range radars in Florida. One minute ago, came the reply.
Now, as Engelauf relayed Austin’s emotional report to Cain, the flight director slowly shook his head, composed himself and turned to the silent control room to declare an emergency. At 9:12 a.m., he instructed Ground Control (GC) Officer Bill Foster to “lock the doors”—a de facto admission that all hope was gone—and ordered flight controllers not to leave the building, but to begin preserving their data and writing up their logbook notes for use in the subsequent investigation.
“OK,” Cain began, “all flight controllers on the Flight loop, we need to kick off the FCOH [Flight Control Operations Handbook] contingency plan procedure, FCOH checklist, page 2.8-5.” He then proceeded to talk them through the required actions: preserving logbook entries and display printouts, communicating only on the Flight loop and restricting outside telephone calls and transmissions. “No phone calls, no data, in or out,” he told them.
Nine hundred miles (1,500 kilometers) to the east, in Florida, the STS-107 families were shepherded from the landing site to the crew quarters by 9:30 a.m. It was left to Bob Cabana to break the terrible news—one of his worst jobs in his astronaut career.
Mission Control, he explained, had not picked up any radio beacon signals which would have been activated if the crew had managed to bail out of Columbia. Regardless, the orbiter was at an altitude of around 40 miles (64 kilometers) and traveling at nearly 15,000 miles per hour (24,000 kilometers per hour) when it disintegrated. That alone offered not even the faintest hope of being survivable.
Later that morning, near Hemphill, Texas, Roger Coday found some human remains. He said a brief prayer and built a tiny wooden cross by the roadside.