Human Factors: Crossair Flight 850
Human Factors: an airworthy aircraft involved in an accident caused by decisions taken by people. The pilot in command, in the end, is held responsible, whether the errors in question are latent or active. The pilot in command carries the final decision over the safety of the flight and the landing.
Statistically, we distinguish between pilot error which is weather related and pilot error which are caused by or lead to mechanical issues. In addition, we have other human error which relates to air traffic controllers, weather reporters and operations staff.
In the incident of Crossair Flight 850, in which a Saab 2000 was written off, the answer appears to be “all of the above”.
The Bundesstelle für Flugunfalluntersuchung (German Federal Bureau of Aircraft Accident Investigation, also known as BFU) released their report on Crossair Flight 850 last month. The sequence of events makes for a remarkable read, where just about every person involved could have reacted better. It is a textbook example of the holes lining up, leading to the wheels being ripped off of a Saab 2000 in a small airfield in Germany in visual conditions. Every individual issue was minor … only as they converge do we end up with the house of cards falling down.
On the 10th of July 2002, a Saab 2000 and her crew were assigned at short notice for the Basel to Hamburg flight, which had been scheduled to fly with an Embraer 145.
They had weather reports (METARs, TAFs, wind charts) valid to 18:00 UTC which showed that they could expect some thunderstorms en-route. They took on an additional 570kg of fuel in case they had to wait out a thunderstorm over Hamburg, which could close the airfield for 20-30 minutes. There was no reason to expect any unusual or extreme weather based on the information the crew received.
However, a SIGMET (Significant Meteorological Information) was issued at 15:00 UTC for Bremen UIR and FIR, including their destination area: a line of thunderstorms advancing northeast and extending up to flight level 380. The flight crew were involved in pre-flight preparation and did not see the SIGMET nor the additional warnings of thunderstorms for the area to the east of Bremen. The flight did not depart until 16:09 UTC, over an hour after the Bremen SIGMET was released.
Note: the accident report uses local time for the flight which I have converted to UTC for consistency. All times in this document are given in UTC.
Issue Number One: The Operational Control Centre did not notify the crew of extreme weather warnings in the destination area.
The BFU is of the opinion that if they had been available, these SIGMETs would have alerted the crew that these were not isolated thunderstorm cells, but a thunderstorm front of considerable proportions. This knowledge might have had an influence on the decisions made by the crew during the flight.
The BFU is of the opinion that the fact that the SIGMET issued at 15:00 UTC … did not reach the aircraft even though it remained on the ground until 16:09 UTC (1809 hrs), indicates that the operator’s Operational Control Centre provided insufficient support.
16:09 | Crossair Flight 850 departed Basel |
17:36 | The flight is cleared to descend to 3,000 feet for an ILS approach into Hamburg runway 23 |
17:38 | As a result of severe turbulance, the flightcrew aborted the approach into Hamburg at 3,300 MSL |
17:41 | The flight entered a holding pattern to see if the weather cleared so they could land at Hamburg rather than divert. |
17:49 | The crew requested a diversion to Hannover, their second alternate. |
While in the holding pattern, the crew had time to consider their next decision. The accident report gives the following list of options:
- Another approach to Hamburg
- Approach to one of the standard alternate airports
- To consult the FMS data base for another suitable alternate airport (NAV-Display)
- Make contact with the OCC and request support
- To consult air traffic control with a request for suitable alternatives
Having decided not to attempt a further approach into Hamburg, the flightcrew contacted ATC to ask about their planned diversion airfield, Bremen. It was also beset with thunderstorms and although the storms were passing, it would involve the flight attempting to penetrate the front. They decided instead to fly to their second alternate, Hannover.
Issue Number Two: Whilst in the holding pattern, the crew had ample opportunity to get a full picture of the surrounding area. They did not ask ATC about other airfields in range nor for updated information about the weather in the area. In fact, at that point Lübeck was immediately below and free from storms, while Hannover was already behind the cold front. However, the crew blindly followed the flight plan.
If there could be no landing at the original declared destination, a diversion to a predetermined alternate airport is a standard operating procedure. Because Bremen was behind the weather front and the crew was not prepared to penetrate the front or fly around the frontal area, the decision to divert to the second alternate – Hannover, which by that time was also behind the cold front – is not understandable. This shows the BFU that the list and sequence of alternates was followed diligently instead of trying to get an idea of other options and possibilities.
17:52 | Crossair Flight 850 left the holding pattern and proceeded towards Hannover. The flight encountered further thunderstorms and the crew realised they would not be about to fly around the storm cell to reach Hannover. |
18:13 | The crew aborted the flight to Hannover and requested a diversion to Berlin-Tegel, which showed on their weather radar as clear of the storms. |
18:15 | Berlin-Tegel ATIS information cited the weather as CAVOK (Ceiling And Visibility OK) with NOSIG (NO SIGnificant change expected). |
Issue Number Three: The recorded weather information at Berlin-Tegel was incorrect. The approaching weather was already recognisable when the report was assembled. It is incomprehensible why a NOSIG – a forecast that the situation would remain unchanged for the next two hours – was included. In addition, the ATIS was not updated as the storm front approached.
In the 30 minute time period between the two routine weather reports, weather conditions worsened considerably. The worsening weather conditions resulting from the approaching cold front were recognisable and did not occur totally unexpectedly. The BFU is of the opinion that it would have been appropriate to issue a SPECI report in the period between 17:50 UTC (1950 hrs local) and the next routine weather report at 18:20 UTC (2020 hrs local).
18:17 | Crossair Flight 850 established contact with ATC at Berlin |
18:18 | Crew received a low-fuel warning and requested priority approach, stating that they only had fuel remaining for 40 minutes. They are given instructions for runway 08. |
18:20 | Berlin-Tegel weather is updated with TEMPO, including warning of thunder showers |
18:26 | A SPECI (a SPECIal unscheduled report) was released as the line of thunderstorms reached Berlin |
18:28 | Crossair Flight 850 aborts the approach owing to turbulence and requests further airport options in the area. Berlin ATC recommend Eberswalde-Finow which was 27NM away. |
From this point, the flight was treated as an emergency because of the low fuel warnings. However, Eberswalde-Finow was also hit by the thunderstorms.
18:32 | The crew requested information on airfields cited in the onboard Flight Management System and changed heading towards Neubrandenburg, 46 NM away. |
18:34 | Berlin ATC report a thunderstorm over Neubrandenburg. The weather was reported to be clear east of Berlin and ATC recommend a heading towards Werneuchen, 20 NM away. |
Werneuchen was not in the Flight Management System and so the crew did not have a chart for the airfield. However, they were running out of options.
Because of the changing weather and the necessity to alter track several times, the crew found themselves in a situation where, instead of a routine diversion to an airport, the shortage of fuel and the weather situation made a precautionary landing necessary at any airfield with a suitable runway. Although it was still unlikely that an engine would stop very soon because of lack of fuel, there was hardly any other alternative. The crew therefore accepted an airfield that was totally unknown to them, Werneuchen Special Airfield.
ATC contacted the chairman at the flying club. Werneuchen was a former military aerodrome which has been used by civil aircraft since 1990. The original length of 08/26 was 2,400m x 80m. The chairman took the call on his mobile phone and warned ATC that the western end of runway 08 included a displaced threshold. The usable runway at Werneuchen is 1,500m with an earth wall erected to curb the problem of illegal car races.
As a precaution against unauthorised access, the aerodrome operator was given permission by the regional Civil Aviation Office to erect an earthen wall no more than four metres high at the western end of the ex-military airfield. In addition, a 70 cm high wall of heavy clay was placed across the entire width of the runway about 235 m west of the threshold to Runway 08 and about 770 m from the western end of the ex-military airfield.
Issue Number Four: Communication breaks down regarding the displaced threshold. The ATC supervisor was under the impression that there were two runways at Werneuchen. He told the controller, “…he should not take the short runway, there is an earth wall after five hundred metres.” The controller had the aerodromechart but it made no reference to closed runway markings on the disused sections of the old runway. Further, the aerodrome chart bore no markings denoting the displaced threshold to runway 08. The controller passed the runway data to the crew and relayed the message from his supervisor.
The contents of the message were incomprehensible for the flight crew without having the possibility of checking or interpreting the approach and runway plates. The choice of the word “should”, would normally be understood by the crew as a recommendation, and was therefore not perceived as a warning. The message contained none of the key vocabulary with which the crew would have been familiar, such as closed portion, obstacle, blocked, displaced threshold.
18:39 | Werneuchen: “And there is an earthen wall, so he should pay attention, that he does not attend er does not land on this five hundred meter strip – he should land on the one thousand five hundred metre strip” |
18:39 | ATC Supervisor: “Yes good, he’ll do that anyway” |
18:39 | ATC controller to Crew: “We just eh been informed that you should use the easterly part of the runway eh so eh in you eh you are not before landing before the threshold of zeroeight – genau.” |
Issue Number Five: The runway markings were no longer clear. The military runway markings were clearly visible, as you can see from the Google maps image of the airfield. The crosses painted on the disused area had eroded. From a visual point of view, it was not clear that any part of the runway was closed, even on direct approach.
Clear marking of flight operation areas, in particular runways, is an essential safety mechanism for the prevention of accidents. A runway must be clearly visible for pilots in the air, and distinguishable from other flight operation areas. The markings at Werneuchen Special Airfield did not meet the requirements or recommendations of ICAO Annex 14, or the national regulations then in force.
18:40 | Crew: “Ja we’re just abeam the threshold zero-eight now making a left hand eh downwind if you agree for zero-eight.” |
Issue Number Six: The crew had never been to this airfield and did not have a chart or any information other than what had been passed on by ATC. Standard procedure would be to overfly the runway but, no doubt as a result of the deteorating weather and concern about the fuel, the pilots never discused this but instead flew straight in. Although the runway markings were not clear, the earthen walls should have been apparent if the crew had flown along the runway for a visual check. The conditions were VMC at this time.
At 20:40 the crew reported Werneuchen in sight and lined up for an approach to Runway 08. They did not have the frequency for the Airfield and continued to speak to Maastricht Upper Airspace Control. The crew approached runway 08 directly without radio communication with the airfield and – due to the weather and shortage of fuel – without first inspecting the runway from the air.
And now the stars are all in alignment for the wipe-out of the Saab 2000.
20:41 | Final approach of Crossair Flight 850 into Werneuchen Runway 08 |
The co-pilot, who was the pilot flying, asked, “Touch down where?” The PIC answered, “Where-ever you like, my friend.”
After passing through 500 ft AGL the PIC said: “It’s longer than … “It’s longer than… longer than Bern, hä.“
Bern is 1,510 metres – the same length as available in Werneuchen. The crew never realised that there was a displaced threshold. They were reacting to the original military runway (2,400mx80m) rather than the 1,500m usable runway. It was not until immediately before the collision that the crew could see the earth wall across the runway.
20:42 | The Saab 2000 touched down and collided into the earthen wall. All three landing gear legs broke off. The aircraft slid 350 meters and then came to a halt. |
The crew opened the doors and the flight attendants evacuated the passengers. Only one minor injury was sustained. | |
21:00 | The thunderstorms passed over the airfield with wind gusts up to 52 knots. |
The BFU cite the following direct causes of the accident:
Immediate Causes:
- The extent and intensity of the thunderstorm frontal system, plus the speed of change in the weather system.
- Insufficient use of available resources when making decisions in flight (pro-active).
- The loss of alternative landing options, coupled with increasing time pressure (reactive).
- Aircraft touched down outside operational area of an airfield.
- Earth wall was not detected, followed by collision with the same.
Systematic Causes:
- Insufficient information with respect to weather situation and development, both prior to and during the flight
- Insufficient information about Werneuchen Special Airfield due to inadequate chart illustration, plus absence of and misunderstood communications
- Insufficient signs and markings of operational and non-operational airport areas.
No safety recommendations were made.
From the first moment, when the SIGMETs were not passed onto the crew, the options began to narrow. The crew took on an additional 570kg of fuel to deal with delays, enough for approximately 45 minutes additional flight. At Werneuchen, the Saab 2000 had 420 kg of fuel remaining in its tanks. In the end, a series of questionable choices and bad information meant that the flightcrew felt pressured to land in a rush on an airstrip for which they had no details.
The Saab 2000 was a write-off, and although weather (especially thunderstorms) are frequently a problem for pilots, the loss of the plane in this instance was solely an accumulation of human factors.
Until this point, the ATC Supervisor was of the impression that the chairman of the flying club was on site at Werneuchen. It was only after ATC lost contact with the plane that he realised the error. He hung up to call the fire brigade and then phoned the Werneuchen number again to ask if maybe the chairman would drive out to the airfield to see what happened. “No, it would take me an hour and a half to get there,” said the chairman. “There is only just one, he is seventy eight or eighty years old but is very mobile. But he could drive out, should I ask him to do so?”
The full accident report is available in English as a PDF on the Bundesstelle für Flugunfalluntersuchung website.
If you found this post interesting you might enjoy the following:
- 15 Years since ValuJet Flight 592
- Biggin Hill Accident Report
- Alaska Airlines Flight 1866
- PULL UP PULL UP: Garuda Indonesia flight GA200 in Yogyakarta
If you enjoy this blog, consider buying my ebook: You Fly Like a Woman for under a dollar at Amazon.
So often, the cause of an accident isn’t one catastrophic failure, but a whole series of small failures that add up to something significant. This was a really fascinating read demonstrating not only a series of breakdowns in communication, but an apparent lack of CRM leading to loss of situational awareness. It’s incredible that after all that, only one person was injured.
I can only imagine the pilots’ momentary relief at finally landing safely turning to sheer horror when they spotted the earth wall coming up fast.
Thanks for sharing this.
Given the massive amount of misinformation and confused communication, it’s amazing that they all walked away from this one. Thanks for sharing – good reminder to know the weather, know the alternates, and to remember that it’s always on the pilot to make the safest decision!
I can only imagine the pilots’ momentary relief at finally landing safely turning to sheer horror when they spotted the earth wall coming up fast.
Yeah, I wince just thinking about it.
A further thing that comes out of the accident report (which I just didn’t have the space to include) is that the pilots had not had any training on decision-making skills. The report doesn’t cite this as a cause but clearly they felt it was a contributing factor to the initial failure to gather more information.
Gwen, I couldn’t help but think how easy it seemed for it all to start snowballing. It makes me feel a bit better about my obsessive overplanning, though! ;)
A further thing that comes out of the accident report (which I just didn’t have the space to include) is that the pilots had not had any training on decision-making skills. The report doesn’t cite this as a cause but clearly they felt it was a contributing factor to the initial failure to gather more information.
Surely under JAR-Ops, CrossAir crew would have been given CRM training, which involves Decision Making as part of its core curriculum. I wonder how the airline failed to give such basic and legally required training to their flight crew.
The last report I read with this much of a SNAFU was the British Midland 737 crash at Kegworth, which was the final nail that got the CAA (then JAA) onto CRM in the first place.
It makes me want to weep that twenty two years on, some lessons still haven’t been learnt.
Nice article.Being a pilot is a difficult profession. It really needs a good training in a good training center to avoid some errors especially in landing. Landing is the most dangerous moment for any pilot.
In general very good comments. But we must not forget that it is easier to reason and work out what should have been done with those beautiful tools that are available to us but were not at the time to the crew:
Tool one: a comfortable armchair, perhaps a glass of wine at the open fire when reading all this (laptop on the knees, the main discomfort we suffer).
How different from the perspective of a cockpit in an aircraft caught by surprise by extreme weather conditions, perhaps severe turbulence and lightning all around them.
Tool two: the beautiful science of wisdom with hindsight. We can at our own pace read the reports and recommendations and draw the right conclusions.
But I am also missing one item: Lost skills.
The SAAB 2000 is a relatively advanced aircraft, equipped with modern instrumentation.
I remember once coming into Muenchen Riem (the old now disused airport EDDM) and encountered very similar weather conditions.
I was just a brand-new captain on a Ce550. My previous airplanes had either no weather radar at all, or old-fashioned black-and-green scopes.
Learning to “read” the picture in those days was an art. I had been trained by pilots who had extensive experience in the tropics and knew all about extreme weather, flying through the “ITF” and knew all tricks to get the maximum information from their radar.
The Ce550 (N121C) had an at the time advanced weather radar in full colour. Using all the tricks I had learned from my training captains I managed to stay in or, with ATC permission near, the holding pattern. There was lightning all around us but we even were in reasonably smooth air.
When the squall had passed (air temperature on the ground had dropped from 32 down to 23 degrees C) we had to stay airborne for another 10 minutes because the runway was flooded.
We landed uneventfully, but there still was standing water on some of the taxiways, here and there about a foot (!!!) deep.
Was I that good ? Not really. I just had the good luck of having been trained by very experienced old-style pilots who had learned their trade the hard way. Not by relying on electronics but by having learned to FLY first and foremost.
Our dependence on computerized systems is eroding basic flying skills.