Investigators probing the Columbia disaster are developing an increasingly detailed scenario that explains the sequence of events that led to Columbia's destruction, a scenario that matches up with telemetry and recorded data as well as the damage seen in recovered debris. Only two out of 10 initial scenarios are still being actively developed by NASA investigators, officials say, but one of them, which assumes a breach in the left wing at or near leading edge panel No. 8, has emerged as the leading contender.

This scenario, No. 2 on the original list of 10 being assessed by agency managers and engineers, matches the telemetry downlinked from the shuttle before its breakup as well as data recorded on board by Columbia's payload experiment support recorder, or OEX, which was recovered near Hemphill, Texas, March 19. It also explains unusual communications dropouts and unexplained flashes seen in the wake of the shuttle as it descended across the southwest United States.

The scenario matches up well with the known point of impact where foam debris from Columbia's external fuel tank slammed into the left wing at 450 mph just 82 seconds after liftoff. In fact, OEX data from thermocouple V07T9895, located on the left wing spar just aft of reinforced carbon carbon panel No. 9, shows a slight temperature increase after the strike that may be indicative - this is not yet confirmed - of leading edge damage.

OEX data recorded during re-entry also is consistent with the severe heat damage seen in recovered debris, which strongly suggests a breach at or very near the underside of RCC panel 8, one of 22 such carbon composite panels making up the leading edge of the left wing.

The U-shaped panels, held in place by so-called T-seals that are bolted to the front face of the wing spar with inconel fittings, are designed to protect the wing from 3,000-degree re-entry temperatures and to handle aerodynamic loads as the shuttle falls into the thickening atmosphere.

"Independent teams examining the recovered debris are finding that the most likely location of (the) initial breach into the vehicle was into the left hand wing RCC panel 8/9 area," according to a summary of the scenario. Scott Hubbard, a member of the Columbia Accident Investigation Board, said much the same thing during a news conference last Tuesday.

The only other scenario still under active consideration by NASA engineers involves a breach in RCC panels closer to the shuttle's fuselage. But it does not fit the facts as closely as scenario No. 2.

NASA's development of failure scenarios is in support of the Columbia Accident Investigation Board, which is responsible for determining the root cause of the disaster. All such scenarios must be confirmed by the accident board and readers are cautioned that the scenario presented below is preliminary and subject to change. That said, here is how scenario No. 2 currently plays out.

Columbia fell into the discernible atmosphere 400,000 feet above the Pacific Ocean northwest of Hawaii - entry interface, or EI - at 8:44:09 a.m. on Feb. 1. Scenario No. 2 assumes the shuttle began its descent with significant damage to the underside of RCC panel No. 8. The scenario assumes spar insulation behind the RCC panels was directly exposed to re-entry heating.

During a news conference Tuesday, members of the Columbia Accident Investigation Board said investigators were focusing on the possibility a lost "T-seal" between RCC panels 8 and 9 might have left a gap an inch wide or greater between the adjacent panels, providing a direct path into the leading edge.

What caused the damage has not yet been pinned down, although most investigators believe it likely was caused by a piece of foam insulation that broke off Columbia's external fuel tank 81 seconds after liftoff and slammed into the left wing's leading edge one second later at some 450 mph. That impact may have cracked or penetrated an RCC panel or damaged a T-seal enough to result in failure.

The day after launch, military radars detected debris floating away from Columbia after a routine maneuver. The debris has not yet been identified, but radar tests are underway at Wright Patterson Air Force Base to determine whether a T-seal or a piece of an RCC panel could explain the sightings.

Regardless of the exact location of the breach - RCC panel 8 or an adjacent T-seal - it did not take long for hot gas to penetrate Columbia's left wing during re-entry. At 8:48:39 a.m., 270 seconds after entry interface, a strain gauge (V12G9921) in the leading edge near RCC panel 9 first showed an unusual increase.

"Thermal stresses build due to the breach in the wing and hot gas impingement on the spar," the scenario reads. "Pressure also starts to build in the RCC wing cavity adding to the load on the spar. The damaged wing causes a change in the load path that increases the wing spar strain. All these loads combine and put an off nominal strain on the spar."

Twenty seconds later, at 8:48:59 a.m., heat entering the RCC cavity through the presumed breach in panel 8 is first registered by a temperature sensor on a fitting between RCC panels 9 and 10 (V09T9910). Almost simultaneously, insulation on the forward face of the spar was compromised, allowing super-heated air to begin eroding the spar structure itself.

As the leading edge heating continued, temperature sensors on Columbia's left-side orbital maneuvering system rocket pod began sensing cooler temperatures than usual. This trend began developing at 8:49:49 a.m., or 340 seconds after entry interface. Computational fluid dynamics calculations indicate this was due to leading edge RCC panel damage affecting the flow of air over the wing.

At 8:50:00 a.m., the first in a series of brief communications drop outs were observed. Engineers now believe these interruptions were caused by the release of molten metals from the eroding wing spar and spar insulators into the hot air surrounding the space shuttle. The effect was similar to that of "chaff" released from military aircraft to foil weapons sensors.

At 8:50:09 a.m., the disrupted airflow over the leading edge of the left wing resulted in a cooling trend on the left side of Columbia's fuselage. Ten seconds later, at 8:50:19 a.m., increasing damage to the leading edge caused a thermocouple on the lower side of the wing near RCC panel 9 to begin sensing an off-nominal increase.

The inside surface of the left wing spar then began showing an unusual temperature rise (sensor V09T9895) at 8:51:14 a.m. This reading, following the RCC interface temperature increase noted at 8:48:59 a.m., confirms hot gas penetrated the cavity behind the RCC panels before entering the interior of the wing itself.

At 8:52:05 a.m., telemetry from the shuttle showed the start of an unusual yawing motion pulling the nose of the orbiter to the left. This was due to the growing damage to the leading edge affecting the aerodynamic behavior of the spacecraft. Computational fluid dynamics confirms this unbalanced force would continue to increase as the damage worsened.

Just four seconds later, at 8:52:09 a.m., the plume of super heated air in the RCC cavity finally burned through the wing spar itself, allowing hot gas to flow into the wing's interior. The plume of hot air impinged on the main landing gear wheel well box, about halfway down its outboard side, and began burning through cable bundles, cutting off data from scores of left wing sensors.

Based on the order of the cable bundle burn throughs, it is possible the hot gas plume impacted the upper skin of the wing and then was deflected along the skin to the wire bundles running along the upper part of the wheel well box. The top two cable bundles were damaged first, followed 30 seconds later by a third bundle lower down the side of the wheel well.

Analysis of the timing of the cable burn throughs appears to confirm the general location of where the burn throughs are believed to have occurred and they are consistent with a breach at RCC panels 8/9.

The hot gas quickly spread throughout the interior of the unpressurized wing, triggering a sharp rise in the temperature being recorded by sensor V09T9895 - the same interior sensor that first detected higher spar temperatures a minute earlier - at 8:52:10 a.m. Seven seconds later, hot gas in the wing was first registered by a brake line sensor in the left main landing gear wheel well.

"As heat enters wing, thermal stresses are created and structural failures start to occur," reads a scenario entry for 8:52:19 a.m. "Hot gas continues to flow into wing, with pressure and temperatures building in the wheel well cavity." At 8:52:41 a.m., a second wheel well temperature sensor begins indicating an unusual increase.

In the meantime, the leading edge continued to deteriorate, affecting the flow of air over the wing. At 8:53:29 a.m., OEX data indicated the left fuselage began heating up as a vortex of disturbed air began moving forward along the side of the orbiter. The movement of this shock wave occurred at the same time as a sudden increase in yaw motion, which appears to confirm increasing leading edge damage.

It was at roughly this point, beginning at 8:53:44 a.m., that observers along Columbia's ground track began noticing debris falling away from the shuttle. At 8:54:20 a.m., the shuttle's "roll moment" changed sign, going from negative to positive. One explanation is an unexplained increase in lift on the left wing, presumably due to increasing structural damage. Another explanation, however, is damage to the upper surface of the wing.

"Hot gas continues to progress down the RCC and may burn through top of wing shedding skin and creating a hole," the scenario reads. "Damage pushes shock wave/vortex onto vertical tail leading to large increase in rolling moment."

Ground controllers at the Johnson space Center in Houston did not notice anything unusual until 8:54:24 a.m. when mechanical systems officer Jeff Kling informed flight director Leroy Cain "I've just lost four separate temperature transducers on the left side of the vehicle, hydraulic return temperatures. Two of them on system one and one in each of systems two and three."

Nine seconds later, at 8:54:33 a.m., a bright flash was noted by ground observers. This is now believed to be the result of maneuvering jet firings and interactions with debris falling away from the shuttle.

Then, at 8:56:16 a.m., the super-heated air entering the wing from the breach in the leading edge finally burned through the outboard wall of the left landing gear wheel well, triggering dramatic temperature increases in sensors located inside the wheel well. The plume is believed to have impinged directly on the left main landing gear strut. One such strut has been recovered and while engineers have not yet determined whether it was from the right or left landing gear, it shows severe melting consistent with a direct plume impingement.

Forty-four seconds after the wheel well was breached - at 8:57:00 a.m. - the plume burned through the forward inboard corner of the main landing gear door, providing an exit path for the hot gas in the wing. As a point of reference, the famous Kirtland Air Force Base telephoto view of Columbia, a photograph showing obvious signs of distress at the leading edge, was taken at 8:57:14 a.m.

At 8:58:03 a.m., a sharp aileron trim change was noted, an indication large pieces of the wing's overheated skin were blowing off and falling away. At the same time, the wheel well continued to heat up. All tire pressure and temperature data were lost during a 20-second period beginning at 8:58:38 a.m.

Ten seconds later, commander Rick Husband radioed, "And, uh, Hou(ston)..." The transmission was garbled. Thirty seconds later, Kling told Cain "We just lost tire pressure on the left outboard and left inboard, both tires."

"And Columbia, Houston, we see your tire pressure messages and we did not copy your last," astronaut Charles Hobaugh called form mission control.

"Roger, uh, buh..." Husband replied at 8:59:32 a.m., interrupted again by a comm drop out. It was the final transmission from the crew.

"The aerodynamic forces and the aero heating become more intense leading to further structural degradation of the vehicle," the scenario concludes. Columbia ultimately became aerodynamically unstable and broke up around 9:00:21 a.m.

Evidence supporting scenario No. 2 includes:

• Heavy slag-like deposits on the interior of RCC panel 8 fragments. The slag presumably is the result of molten aluminum spraying away from the point where hot gas impinged the wing spar.
  
  
• Severe heat damage on a tile carrier panel located between the lower edge of RCC panel 9 and the heat shield tiles permanently bonded to the lower side of the wing; erosion and slumping of the heat-resistant tiles indicate prolonged exposure to extreme temperatures.
  
  
• Severe heat damage to a tile carrier panel located just behind the upper edge of RCC panel 8.
  
  
• "Knife-edge" erosion of pieces from RCC panels 8 and 9; RCC erosion requires temperatures in excess of 3,000 degrees Fahrenheit; the knife-edge pattern, only seen on debris from panels 8 and 9, is indicative of such extreme heating.
  
  
• Recovery of an upper wing tile farther west than any other confirmed shuttle debris; the tile, originally mounted on the upper surface behind RCC panel 9, presumably fell off earlier than others because of the disturbed airflow and heating caused by the breach.

"Data is emerging that shows very significant and unusual damage at the interface between (RCC) panels 8 and 9," Hubbard said Tuesday. "For example, the splattered metal on panel 8 is much heavier than elsewhere and there is erosion of the reinforced carbon carbon elements, the ribs, what's called the lug where the attachment occurs.

"What we see is something that's not seen thus far anywhere else on the wing leading edge or indeed in the orbiter debris, which is pieces of a very tough material, this reinforced carbon carbon, eroded to knife edges. Where a normal piece is a half an inch (thick), it's been eroded to about the thickness of a dime. This kind of heating event indicates long duration, very extreme heating.

"We don't know quite what to make of this yet, other than what I said, a very severe heating event in the intersection between panels 8 and 9," Hubbard said. "The carrier panel, that is, the piece of material that goes between the reinforced carbon carbon and all the tiles on the bottom of the orbiter, that carrier panel also shows severe heating. It's slumped, like you overheated Styrofoam. That's another indication of some very severe heating events."