On 11 January 2017, a Cessna Citation 560 (LN-IDB) being operated by air taxi company Hesnes Air on a non revenue positioning flight from Oslo to the company base at Sandefjord/Torp as HSG03 experienced an uncommanded sudden and violent nose-down manoeuvre in day IMC. A MAYDAY was declared as recovery was achieved with minimal terrain clearance but it was cancelled after full control was regained and the intended flight was completed without further event. Structural damage to the aircraft was subsequently found to be so severe that it was declared a hull loss.
An Investigation was carried out by the Accident Investigation Board Norway (AIBN) after the event was notified the same day by the Gardermoen TWR Supervisor. The CVR and 2 hour FDR were removed from the aircraft and their available data were successfully downloaded. However, it was noted that “unlike most FDRs it did not register the control surface and yoke positions, which would have been highly beneficial in this Investigation”.
It was noted that the Captain, who was PF for the accident flight, was also the Flight Operations Manager and the nominated post holder responsible for both fixed wing and rotary wing operations by the company. He had completed his initial pilot training in 1991 and had joined Hesnes Air in 2014. He had a total of 12,000 hours flying experience which included 300 hours on type. The First Officer had completed pilot training in 2012 and had joined Hesnes Air in 2015 after a period as a light aircraft flying instructor. He had a total of 1,500 hours flying experience which included 240 hours on type.
It was established that the same aircraft and crew had operated a non scheduled revenue flight from Bern to Oslo earlier the same day cruising initially at FL380 and then at FL 430. This flight landed at Oslo at 1603 LT and after passenger disembarkation at the GA Terminal was flight planned to then position out of service from Oslo to Sandefjord/Torp under the same flight number.
It was noted that as per a pre flight conversation in Bern, the crew had agreed that the Oslo ground stop would be “as brief as possible” with no refuelling unless unavoidable, leaving one engine running whilst the passenger disembarked and the First Officer conducted the external pre-flight check of the aircraft. They planned to avoid refuelling and de-icing if weather conditions permitted it. This plan was followed and, as a consequence, the First Officer’s external pre flight check was limited to an inspection of only the left hand side of the aircraft during which he “saw water on the wings and presumed that this was also the case on the other wing and tailplane” but no ice. By the time the First Officer had returned to the flight deck, the Captain had already obtained a departure clearance and had advised GND that they did not want de-icing.
After less than 10 minutes on stand, the left engine was restarted and taxi clearance was requested. Clearance to taxi to the holding point of runway 19L via taxiway C3 was given after which the Captain asked the First Officer to remind GND that “they did not want de-icing”, in the hope that they would then be given a runway 19R departure which would save taxi time and avoid a need for de-icing. However, there was no change to the 19L holding point clearance and the Captain “then realised that they were approaching a need for de-icing and there was a certain urgency to get in the air”. The GND controller then asked for confirmation that they did not want de-icing which was given and once with TWR, there was no other traffic ahead and the flight was given an immediate line up.
With continuous light snow falling and a surface temperature of zero, the wing/engine anti-ice and tailplane de-icing systems were selected to auto as part of the Before Takeoff Checks and once clearance had been given, takeoff was commenced using 88% N1. Cloud was rapidly entered and since the aircraft was very light, it accelerated very quickly during takeoff and in the initial climb and to keep the airspeed below 200 KIAS until the height specified for flap retraction, approximately 1,500 feet aal, the normal 12° climb angle had to be increased to 21.0° and then further to 25.5° and despite this, the speed still reached a recorded 202.5 KCAS before the pitch was reduced which occurred at around the target 1,500 feet aal - approximately 2,100 feet QNH.
Although the normal trim change when retracting flap would be nose up, FDR data showed that in this case, the nose began to drop and a bunt became a sustained dive. The aircraft also simultaneously banked sharply to the left and these two sudden and unexpected changes together with the onset of significant negative g were bound to have caused spatial disorientation for both pilots. The dive, turn and negative g continued and eventually the aircraft reached a recorded 33.6°nose down and the accompanying negative g reached 2.62. By this time, the speed was 230 KIAS and still increasing and a rapid succession of EGPWS Warnings and Alerts was occurring - SINK RATE, PULL UP, TOO LOW TERRAIN PULL, WINDSHEAR, BANK ANGLE.
As the dive had developed, the Captain was “not active on the controls” and the First Officer therefore took control and both initiated and completed a pull-out and recovery to normal flight. Negative g steadily reduced and the maximum nose down attitude of 53° was achieved at +2 g. Positive g continued to increase during the pull out and reached 5.99g with the pitch still 22° nose down 10 seconds after flaps up was initiated. With the aircraft only about 500 feet agl, the aircraft emerged from cloud into visual conditions and the Captain then joined the First Officer in pulling hard back on the control column and this helped in the conversion of the dive to “a relatively quick climb [...] and preventing the aircraft from flying into the ground”. The aircraft nose passed through the horizon at 2.13 g and the speed peaked at 325 KCAS at 7° nose up and +0.45g. The aircraft came to within approximately 170 feet of the terrain before climbing.
As the positive pitch increased, a normal climb was re-established at up to 21.5° pitch. The Investigation noted that the crew reported having “hesitated somewhat before continuing the climb after the pull-out [...] because the Captain did not trust the instruments”. When the aircraft had climbed to a point where the cloud base was approaching with the First Officer still controlling the aircraft, the Captain had “pushed the control column forward [...] to prevent the aircraft from entering the clouds” but the First Officer “was able to convince him that the instruments were correct and that they should climb further into the clouds”.
FDR data showed that engine thrust remained close to that set for takeoff until after the pull out and was only reduced to 39% as the aircraft pitch was approaching 5° and as it continued increasing through 18.6°, thrust was increased to around 83% where it remained until the aircraft levelled off at approximately FL150. The Captain then cancelled the MAYDAY and confirmed that the flight would continue its planned destination. It was noted that as the After Takeoff Checks were not run, the altimeter subscales had remained set to QNH so the level-off was made at a slightly lower altitude than the cleared level but with the altimeters re-set, there was little traffic and the flight was allowed to continue at a non-standard level.
For the remainder of the flight, “further routines and checklists were partially completed in accordance with SOPs” and on nearing their destination, radar vectors were provided until the flight was clear of cloud at about 3,500 feet QNH, after which the approach was completed visually to a landing after a total of 20 minutes airborne.
Due to the high g loads to which the flight was exposed during the upset, deformations had occurred in multiple locations on the upper wing surfaces as well as on the engine and the aircraft was deemed a hull loss due to the prohibitive cost of repair.
Explaining the loss of control
The Investigation noted that an aerodynamic wing stall “would not cause negative g”. It was additionally noted that no evidence that the crew's manoeuvring of the aircraft had contributed to the onset of the loss of control had been found. After very considerable investigative effort, it was concluded that an ice-induced tailplane stall was the only credible explanation for the uncommanded dive. It was, however noted that no previous ice-induced tailplane stall accident investigation was known to have attributed the stall to the process of retracting flaps, rather all known previous instances of ice-induced tailplane stall had been assessed as a result of deploying flaps. However, no evidence was found that the concurrence of flap retraction with the dive represented a causal link. Rather, the initiation of tailplane stall due to the accretion of slush deposits during the takeoff roll which then froze and/or falling snow throughout the time on the ground in Oslo was assessed to be highly credible. It was noted that the Cessna Aircraft Operating Manual for the Citation variant involved contained the explicit warning that “trace or light amounts of icing on horizontal tail can significantly alter airfoil characteristics which will affect stability and control of the aircraft”.
Flight crew performance
On the evidence available, the Captain’s decision not to de-ice was flawed and associated with an apparently overwhelming desire for a quick turnround in Oslo despite the Investigation being unable to find any pressing reason for it. The ‘fixation’ on a quick turnround led to an engine being left running (for which there was no routinely applicable SOP in the company OM) and a consequently abbreviated external check and was considered to be an example of ‘plan continuation bias’ which continued right up to the persistent ‘no de-icing required’ calls made before and after engine start. The First Officer stated that he had considered the ground stop "rushed" and that he "was a bit behind". Noting that this was his first flight where de-icing was an issue, “he wasn't sure if his feeling of not being comfortable with the situation was professionally justified or if it was due to inexperience”.
The different responses of the two pilots to the loss of control were marked. The Captain stated that he had been “overwhelmed (by the) dramatic change in the aircraft's position” and had been convinced that his instruments were wrong and “that they could impact the ground”. He had not believed his instruments until the aircraft came below the cloud layer and visual reference was available and he had still lacked confidence in them as the First Officer had continued to climb back into the cloud layer after taking control and pulling out of the dive. The First Officer explained that he had seen that “all three artificial horizons changed colour at the same time” and thus had not identified any flight instrument malfunction. He had therefore taken over control and started to pull back on the control column. The fact that the recovery was attempted without reducing engine thrust was clearly an oversight attributable to the confusion. However, the Investigation considered that in fact “the rapid increase in speed provided a positive contribution toward clearing the tailplane stall and thus resulted in rapid control response”.
It was noted that it would have required a very hard pull on the control columns to achieve positive g of 5.99 and its achievement was seen as “an indication that the First Officer, and eventually the Captain realised the risk of impacting the ground during the pull-out”. The Investigation considered that the situation which the crew was in was so extreme that it was only this powerful pull-out with a high g load which had saved them from a fatal accident.
Aircraft operator management process and its oversight
Whilst having no bearing on the accident, the Investigation found from its examination of Hesnes Air that “allowing one person in a company to hold several key roles linked to operational safety may have unfortunate consequences” and that there is a discrepancy between the regulatory approval of Norwegian aviation organisations and what appears to be the intention of EASA Regulations.
The Primary Conclusion of the Investigation was as follows:
The probable explanation for the aircraft suddenly diving is that the tailplane stalled. No explanations for this have been found other than slush / spray from the runway and falling snow and sleet settled on the tailplane's leading edge and underside during taxi and takeoff. This contamination is presumed to have frozen to ice.
The aircraft's anti and de-icing systems on the wings and tailplane were switched on, but the tailplane de-icing system had completed a complete ‘cycle’ before takeoff and was in rest mode during takeoff. The Investigation considers that the aircraft's anti- and de-icing systems were not suitable for removing the type of ice and snow that had most likely settled on the aircraft tailplane. The aircraft should have been de-iced before take-off, in line with the operator's de-icing procedure.
This accident shows the significance of good crew resource management (CRM) in the flight deck when an unexpected and extreme flight situation occurs. In this instance, the First Officer's situational awareness and initial pull-out contributed to the aircraft not crashing.
Two Safety Recommendations were made as a result of the Investigation as follows:
- that Textron/Cessna inform all its customers that operate Cessna Citations about this accident and about the risk of contamination on the tailplane in the form of ice or other substances which can result in the tailplane stalling. SL no. 2020/01T
- that the Civil Aviation Authority Norway changes its routines for approving organisations in line with the intent of the EASA regulation. SL no. 2020/02T
The Final Report was published on 30 January 2020.