The crew who survived last year's Airbus A300 freighter shootdown at Baghdad relive their terrifying experience of coaxing their stricken aircraft back to earth
The Airbus A300 is on final approach to Baghdad's runway 33L, flying controls totally disabled, with the burning outer left wing ripped apart by a missile. As he watches Capt Eric Gennotte manipulate the only remaining control input - the throttles - first officer Steeve Michielsen mentally rates his survival chances at "10-20%" because he knows how much can go wrong in the remaining few seconds.
At about 400ft (120m) on a precarious, but relatively stable long straight-in approach, crosswind and surface heat turbulence suddenly make the stricken twinjet unsteady, lifting the left wing and turning it toward the airport buildings. Gennotte tries to lift the right wing by varying the relative thrust of the two engines while Michielsen and flight engineer Mario Rofail monitor his struggle with the turbulence. Having battled as a co-ordinated team for 25min since the missile attack, they are now leaving throttle control to the captain. The crew know what happened in July 1989 to a United Airlines McDonnell Douglas DC-10, which suffered catastrophic hydraulic failure after an engine disintegrated. Despite a heroic effort from the crew, in the last 100ft of an emergency approach to Sioux Gateway airport, Iowa, the DC-10's nose and right wing dropped, the wingtip hit the ground and the huge airliner flipped before sliding to a halt in a storm of flames. Of the 296 people on board, 111 were killed.
Fast forward to 22 November 2003 and Gennotte will also face attempting to land an aircraft using only throttle control.
That morning European Air Transport/DHL A300B4 arrives in Baghdad from Bahrain earlier. Locally the weather and visibility is good and, on the ground, work to turn the aircraft around begins quickly. The cargo is unloaded and reloaded with about 7t of general avargo for the trip back to Bahrain. At take-off the aircraft's weight is only 105t, the maximum allowable take-off weight is 165.9t.
The crew taxi the A300 out to runway 15L for take-off, with Gennotte the pilot flying. Because of the light weight and the need for a maximum-angle climb from lift-off to gain as much height as possible before reaching the airfield boundary, the selected configuration is "slats only" - no flap - and full power. Climb is to be straight ahead to waypoint LOVEK.
Rocked by explosion
As the aircraft powers up through 8,000ft an explosion rocks the aircraft, a cacophony of aural warnings erupts and lights flash for multiple systems. Rofail tells the pilots almost immediately that all pressure is lost from the Green and Yellow hydraulic systems, and 20s later the Blue system pressure also begins to drop and soon hits zero. The primary flight control surfaces and spoilers go limp as their actuators drain, trailing in the slipstream.The horizontal stabiliser, which controls the aircraft's pitch, is frozen at the trim position for 215kt (400km/h) with climb thrust set - the angle it is at when drained of hydraulic power. Flaps and slats are unavailable.
The crew know something has hit their aircraft. Michielsen believes a missile has hit the rudder because he sees the yaw damper switches trip out, but Rofail does not rule out a collision with an unseen aircraft. While the cause is unknown, the effects are dramatic. Michielsen makes an emergency call to Baghdad approach, then Rofail takes over communication with air traffic control because the pilots are preoccupied with fighting the aircraft's apparently out-of-control state.
From his seat behind the pilots, Rofail, like all flight engineers, can see the big picture: systems panels; primary flight instruments; the real horizon as it rocks and pitches - often disappearing from sight - and he joins the pilots' struggle to understand what - if any - control they have left. He says: "The aircraft was like a piece of paper in the air. We went through a series of steep banks and dives - you could not leave your seat". The aircraft's rollercoaster manoeuvring throws the crew against their harnesses.
All the crew are taking part in everything, doing whatever they see needs to be done. "The rulebook has gone out the window," explains Rofail. "Situations like this are unique every time. You cannot train for them. You cannot write a checklist for them." The crew have since listened to the cockpit voice recorder tape and say they are quite surprised at how calm they all sound. Rofail says: "All you can do is apply common sense and stay calm. We were the right combination of crew."
What most concerns the crew is lack of control over airspeed. Initially, Michielsen and Gennotte try to use the control yokes and rudder pedals, but quickly accept they are ineffective. Although the crew know theoretically they could control the aircraft with the throttles alone, it takes them about 10min to learn how to keep the aircraft at an acceptable attitude. During the learning process airspeed lurches wildly between 180kt and 300kt.
Because, like most big jets, the A300's engines are slung below the wings, an increase in thrust causes the nose to pitch up, conversely a thrust decrease causes a pitch-down moment. The problem for Gennotte and his crew is that, on applying power, the increase in nose-up attitude tends to dampen an increase in speed. Then because they cannot apply any direct pitch control or change the pitch trim, when any power-induced increase in speed exceeds the trimmed indicated airspeed (IAS) of 215kt, there is a gradual further pitch-up followed by a loss of speed as the nose-up attitude steepens. Thrust reduction may provide an instant nose-down pitching moment, but then the descending flightpath tends to make the speed increase, and as airspeed rises above the 215kt trim speed, the nose gradually pitches further up. The result is a counter-intuitive secondary effect to every power change. Selection of asymmetric thrust provides roll control, but altering heading or picking up a wing by this method is painfully slow compared with the almost instant primary effect a power change has in pitch.
By the time they have carried out the first gradual, violently pitching and rolling left turn toward the airfield (see diagram opposite page) they are at about 4,000ft and they have the airfield in sight. The captain calls for the gear to be lowered using the emergency gravity system, even though airspeed is slightly higher than the 270kt limit for deployment. While Rofail leaves his seat to do this, Michielsen has to push his own seat as far forward as it can go, making him temporarily almost useless in assisting Gennotte.
To the crew's relief, the gear locks down at the first attempt. Its deployment makes the aircraft noticeably more stable, potentially providing enough control over airspeed and attitude to make a viable attempt at landing. It provides the pilots with an increase in overall drag and a slight pitch-down moment against which to use the pitch-up effect of an increase in power. (Airbus has since told the crew that if they had attempted a gear-up approach, they almost certainly would not have succeeded.)
The crew know they have to land quickly because the wing is still trailing a 50m flame. They cannot see this - Gennotte cannot see the left wingtip from his seat - but ATC and a helicopter circling nearby confirm that their left wing remains on fire; they know that if a part of the wingtip separates they will lose all control of the aircraft.
At present - although the crew does not know it - the leading edge of the wing is complete along almost its entire length, but the fire is gradually destroying the outer wing, creeping forward from the trailing edge. At some stage before they land the rear wing spar separates and the remaining structure is held together only by the forward spar. It is only a matter of time before that also fails. Meanwhile, the asymmetry in thrust needed to compensate for the difference in lift and drag between the damaged and undamaged wings is increasing.
Michielsen suggests to the captain that they position for a long, straight final approach to runway 33R - starting about 37km (20nm) out - and assists the captain with navigation by monitoring the VOR/DME readouts. They lose visual contact with the airport as they turn away to position the aircraft on long final approach, but ATC has no radar to provide the crew with vectors.
Rofail is meanwhile wrestling with another task vital to keeping the aircraft airborne. Although the missile has not damaged the engines, if a fuel flow interruption causes failure of either they will be dead within a minute. When it hit, the missile destroyed the outer left wing tank 1A - so comprehensively that the fuel just fell out of it. There was no explosive ignition of fuel because the tank was full, so there was no fuel-air vapour inside - if there had been, the wing would have been blown off the aircraft.
Wrestling with fuel
But Rofail's problem is more complex. When the missile hit, the engines were both feeding directly from their respective inboard wing tanks (tanks 1 and 2) and he is keeping it that way. But as well as destroying tank 1A, the missile also pierced tank 1, so it is losing fuel. Rofail does not want to open the crossfeed valve to transfer fuel from right wing to left for fear of bleeding away all the fuel; and neither does he want - unless there is no alternative - to break the golden rule that separate engines should be fed from separate sources. So he does not open the crossfeed, but monitors the fuel quantity and feed to both engines and selects ignition on both permanently on.
Gennotte, with Michielsen's guidance, flies the aircraft outbound and crosses the extended centreline for 33R from right to left, setting up for a teardrop turn to the right onto final approach. When the aircraft begins to turn toward the airport, it is about 37km out and at about 3,000ft. Gennotte stabilises the aircraft at that height, heading inbound in approximately level flight. Michielsen monitors the DME to determine when they are approaching a standard 3¡ descent point toward the runway. The visibility is excellent, but Gennotte realises the aircraft is drifting to the right. He calls Michielsen for a wind vector reading and is told it is 290¡ at about 20kt. The crew find they are lined up for 33L and, despite it being shorter than 33R, they decide to use it. Runway 33L is further from obstacles and they cannot guarantee their touchdown point or heading.
Early in the long final approach, Michielsen points out that, counter-intuitively, Gennotte must not retard the throttles before touchdown. It was that move which, after brilliant handling beforehand, led the Sioux City crew to a crash-landing, as the nose and one wing dropped just before the aircraft landed.
The crew's co-operation is impeccable, Gennotte is not merely an able pilot, he is also no autocrat and makes good use of all information fed to him by his crew.
Rofail, the senior crew member by far in terms of age and experience, carries out the final approach checks in his head without disturbing the pilots. He depressurises the aircraft so the doors can be opened on landing. The gear is down and there are no high lift devices to deploy. Gennotte does not call for the landing checklist: he and Michielsen are busy trying to direct the aircraft and manage its stable descent toward 33L. Rofail knows the spoilers will not deploy on touchdown and, since their airspeed will still be about 210-215kt, he knows the aircraft will want to keep flying when it hits the surface, so he readies himself to slam on full reverse thrust.
Gennotte handles the approach, with call-outs from Michielsen to guide the descent profile, and the aircraft is almost stabilised when it passes 1,000ft. At this point, Michielsen still gives himself only a 20% chance of survival, but that percentage is climbing as they close with the runway.
Then, at 400ft, surface-generated turbulence starts to upset the aircraft and the left wing tips upward, turning them toward the buildings between the two runways. Gennotte manipulates the throttles relative to each other to control the aircraft's roll, but its response is agonisingly slow. As landing becomes imminent the aircraft is going to make it to a runway touchdown, but with its heading 8¡ to the left of centreline. In an attempt to line up, the right throttle is retarded slightly, but there is no attempt to close them both, as would be instinctive. The right wing drops and they touch down, right wheel first. Both pilots revert to instinct and are pumping the rudder pedals, stick and control wheel - uselessly. Rofail grabs the throttles and slams them fully into reverse. The aircraft has departed the runway into the sand to its lefthand side, the wheels and thrust kicking up a plume of thick dust, but the sand is helping to slow them down. The aircraft is suffering jolts registering 7.5g vertical acceleration on the uneven ground and several tyres have failed. They career through the razorwire fence to the left of 33L and carry it with them before the aircraft comes to a halt.
Into a minefield
Rofail already has both front doors open and the crew exit down the slide on the right. They run from the aircraft, believing the fuel tanks could explode at any moment, but are stopped by soldiers screaming that they are running through a minefield. They wait until a rescue vehicle is called and they can walk behind it.
Records show the aircraft touched down with a nose-up attitude and a descent rate of about 10ft/s (3m/s) - a normal touchdown rate, with a 10¡ right bank.
That evening the crew were back in their hotel bar in Bahrain when the news item about what they had just been through came onto CNN. Michielsen says: "In the morning we were just freighter pilots. That evening a Lockheed C-5 test pilot from Andrews Air Force Base wanted to shake our hands."
DAVID LEARMOUNT / LONDON
ADDITIONAL REPORTING BY ARI EGOZI IN TEL AVIV
A year before Baghdad, a similar terrorist strike on an Israeli Arkia Boeing 757 leaving Mombasa, Kenya - two SA-7s missed by metres - sparked a flurry by the US and other governments to fast-track development of civil aircraft-based counter-Manpads (man-portable air-defence systems) technology. Baghdad - although in a war zone - only served to highlight, for many politicians, the urgent need to get systems to market.
Israeli airline El Al's first aircraft with flare-based countermeasures flies soon, with further aircraft to be equipped after tests. In the USA, BAE Systems and Northrop Grumman are working with airlines towards flight tests next year of their laser-based systems. El Al aircraft flying to high-risk destinations will be protected by the Flight Guard system developed by Elta, an Israel Aircraft Industries subsidiary. The system dispenses pyrotechnic "dark flares" developed by Israel Military Industries, said to be safe and nearly invisible.
The US Department of Homeland Security (DHS) rejected a flare-based solution when, in August, it selected laser-based directed infrared countermeasures (DIRCM) systems for Phase 2 of its Counter-Manpads programme. Northrop Grumman and BAE have 18-month, $45 million contracts that will lead to supplemental type certification of their DIRCM systems on widebodies in January 2006.
BAE's DIRCM will fly later next year on a Delta Air Lines Boeing 767-300, to collect data such as false-alarm incidences, says Steve DuMont, business development manager countermeasures programmes. Delta Tech Ops is leading integration and installation work. Northrop Grumman will test its Guardian DIRCM on a FedEx Express MD-11 and Northwest Airlines 747 in the third quarter of 2006, says Jack Pledger, director IRCM business development. Derived from Northrop Grumman's AAQ-24 military DIRCM, Guardian uses AAR-54 ultraviolet missile-warning sensors and a smaller, lighter laser-only turret, or "jam head".
BAE's system is derived from its ALQ-212(V) military DIRCM and uses its AAR-57 UV sensors combined with a smaller Agile Eye 2 jam head with high-power multi-band laser. Both companies mount the sensors and jam head in an underbelly pod that can be fitted and removed quickly once structural and wiring provisions have been installed during heavy maintenance. "We can modify the aircraft in less than four days and bolt on the pod in less than an hour," Pledger says. In each case the pod is common across aircraft ranging in size from the 737 to the 747. "Unit and operating costs are the two biggest challenges," says DuMont. Both companies say they are significantly below the DHS's price target of $1 million installed by the 1,000th unit, while the operating costs are "in the 1/10th of a cent per available seat mile range for widebody aircraft", DuMont says.
Pledger says: "The pod weighs 350lb [160kg]. With modifications and cables the total is 500lb - the equivalent of two passengers and their bags." Increasing reliability is critical to reducing support costs, and is a key part of Phase 2. Military DIRCMs typically have a mean time between failures (MTBF) of 300h, about what the aircraft fly each year. Commercial flying is "an order of magnitude" higher, says DuMont, adding: "We need to get tens of thousands of hours MTBF at the box level to get the system-level reliability required."
Phase 2 of the Counter-Manpads programme will provide the DHS with data for presentation to Congress, which must decide whether to mandate countermeasures systems on US airliners - and who will pay for them. The DHS is looking at a possible Phase 3, akin to a military operational evaluation, which would gather data from systems in revenue service on a larger number of aircraft.
While the flare-based Flight Guard is Israel's immediate answer to the Manpads threat, the ministry of defence has selected a DIRCM from Elbit Systems and Rafael as a permanent solution. The system will use a fibre laser developed by Elbit arm El-Op, and an infrared missile-warning sensor from Elisra. The DIRCM will be ready for installation by the end of 2005, says Haim Rousso, El-Op general manager. The laser generator will be mounted in the cargo bay, allowing use of a small turret to house the detection sensors and the tip of the fibre laser. Developers say the more powerful laser will be able to jam missiles with imaging infrared seekers.
GRAHAM WARWICK / WASHINGTON DC
ADDITIONAL REPORTING BY ARI EGOZI IN TEL AVIV