Investigators believe mismatched aeronautical data catalysed a serious approach incident, in which a BMed Airbus A321 came close to striking terrain while attempting to land at Khartoum in a sandstorm.
But the inquiry criticises the crew for failing to follow procedures as they put the aircraft into a steep descent, mistakenly thinking that the jet was not following the correct approach profile, and for not executing a timely go-around or efficient terrain-escape manoeuvre.
While arriving from Amman on 11 March 2005 the aircraft passed below the minimum descent altitude of 1,650ft (503m) - just 390ft above ground level, and already 400ft below the correct approach path - while still descending at 1,200ft/min.
Miscommunication between the two pilots led each to believe the other had visual contact with the runway, when neither had. When this became apparent the captain executed a go-around but failed to arrest the descent in time to prevent a ground-proximity warning from sounding.
At this point, says the UK Air Accidents Investigation Branch, the aircraft was just 121ft above the ground and "hazardously close" to striking terrain.
Given the descent profile, it states, a delay of just 6s in making the go-around decision could have resulted in a ground collision. It adds: "It is likely that the [ground-proximity] alert would not have provided sufficient warning time to prevent a controlled-flight-into-terrain accident."
While the AAIB questions the crew's actions, it also states that BMed's operational and written procedures were "incomplete" and the airline's training standards were "inconsistent" for the type of approach being attempted. It adds that the UK Civil Aviation Authority's guidance and regulatory requirements for such approaches were "fragmented and ill-defined".
It stresses, however, that both the airline and the CAA have since taken "significant safety action" to prevent a recurrence.
The A321 had been programmed to perform a VOR/DME approach to Khartoum's runway 36, with the autopilot coupled to the flight-management system.
This approach demanded visual meteorological conditions. The crew had been given incomplete and inaccurate weather information, leading them to believe visibility was appropriate, when in fact the sandstorm had reduced visibility to instrument conditions.
While the approach chart indicated the final descent point was 5nm (9.25km) from the runway threshold, the flight-management system had been updated with a figure of 4.4nm. The information had been sourced from different suppliers but the pilots failed to check whether the data agreed.
This discrepancy meant the aircraft started descending 0.6nm later than the crew expected, leading them to think the aircraft was too far above the approach path and prompting the captain to select increasingly high descent rates - at one point the A321 was losing height at 1,728ft/min.
The AAIB says the sequence of event suggests the pilots either did not monitor the flight-management system or "did not understand what the aircraft was doing", and it adds: "Had they monitored the descent profile they would have realised that the aircraft was descending on the approach profile in its database and this would have acted as a second cue that the approach chart and navigation database were not in agreement."
In instrument conditions the aircraft should have been on the correct descent path at 1,000ft above ground level, says the AAIB: "The flight data clearly shows that the approach was far from stable and a go-around should have been initiated at 1,000ft."
Statements by the pilots, it adds, suggest they did not make the appropriate cross-check calls as the jet neared, then crossed, the minimum descent altitude. The sandstorm proved disorientating and a misunderstanding between the pilots led each to think the other had visual contact with the runway lights.
"Had appropriate calls been made at the critical moments, they would have almost certainly prevented the confusion that allowed the aircraft to continue below [minimum descent altitude] without the required visual references," says the AAIB.
When the captain executed the go-around he pulled the Airbus' side-stick back to about 40% rear deflection. The ground-proximity warning sounded seconds later but the deflection increased to just 54% instead of the fully-back position required for an escape manoeuvre.
The aircraft made a second attempt to land at Khartoum - this time without reference to the flight-management system - but aborted this approach at the minimum descent altitude because neither pilot could see the runway lights. After receiving confirmation from air traffic controllers that visibility had dropped to 200m the crew decided against a third approach and diverted to Port Sudan where it landed safely. None of the 19 passengers and eight crew members was injured.
In its report into the incident the AAIB is recommending greater clarity from Airbus regarding the need for fully-back side-stick during terrain-escape manoeuvring, and guidance from the manufacturer about operations in sand-blown conditions. It also says the CAA should clarify guidance to pilots faced with an aeronautical data conflict.
But it also says that safety regulators should examine the criteria for certification of terrain-awareness and warning systems, particularly regarding timeliness of warnings when landing aircraft are close to the runway. The AAIB highlights improvements in terrain-avoidance technology, such as the availability of direct Global Positioning System data feed into ground-proximity warning systems, but says required minimum standards have not advanced.
Two years before the Khartoum incident another BMed aircraft, an A320, nearly collided with terrain after its crew was misled by an erroneous signal from a ground-based navigation beacon while the jet was performing a non-precision approach to Addis Ababa.
BMed had been operating as a British Airways franchise partner at the time of both occurrences but the airline has since been acquired by UK carrier BMI.
Source: Flight International
