On 13 July 2012, a Gulfstream G-IV (N823GA) being operated by USA-based Universal Jet Aviation on an out of service positioning flight from Nice to Le Castellet with only the three crew members on board left destination runway 13 at speed after making a relatively uneventful touchdown off a visual approach in normal daylight visibility. Passage through the perimeter fence and then collision with trees and rocks was followed instantly by fuel-fed fire from which the three occupants were unable to escape and died. Most of the aircraft was destroyed as a result of the impacts and fire and damage was caused to runway and PAPI lighting as well as the fence.
The direction of landing and location of the wreckage (reproduced from the Official Report)
An Investigation was carried out by the French BEA. Relevant data was successfully downloaded from both the Flight Data Recorder (FDR) and 30 minute Cockpit Voice Recorder (CVR). The 60-year old Captain held an Federal Aviation Administration (FAA) ATPL and was found to have joined the Operator part time in 2010 after a 31 year career as a pilot at American Airlines. At the time of the accident, he had been flying as Captain on the G-IV for just under a year during which he had accumulated 972 hours. He was accompanied by a 24 year old Co-pilot who had 1350 total flying hours including 556 on type and who had been working for the Operator full time for just over a year.
It was found that after a stopover in Athens lasting 9 days, the two pilots and a cabin attendant had positioned earlier on the day of the accident to Istanbul Sabiha Goken where they had boarded three passengers and flown them to Nice, before making the positioning flight to Le Castellet in the early afternoon because there was no parking available for the intended stop of several days. The aircraft was airborne for 22 minutes with the Co-Pilot acting as PF and landed off a manually flown visual approach. It was found that the 'GND SPOILER UNARM' caution had been displayed on the Engine Indicating and Crew Alerting System (EICAS) from the point at which the landing gear had been extended until the end of the flight and noted that it would have been accompanied by a single chime aural alert when first generated. This occurred because the crew had forgotten to arm the ground spoilers during the approach.
An Terrain Avoidance and Warning System (TAWS) 'SINK RATE" caution was found to have been annunciated at 25 feet agl over the runway 13 threshold and after correction of the flight path, main gear touchdown on the 1750 metre-long runway occurred 365 metres beyond the threshold. Braking was commenced, but the initially unrecognised failure of the PM to arm the ground spoilers prior to landing meant that there was only a minimal load on the landing gear which both reduced braking effectiveness and, because of a consequent momentary loss of 'on ground' sensing, led to the thrust reversers retracting almost as soon as they had been (promptly) selected. The rate of deceleration achieved was poor and Master Warnings for 'L-R REV UNLOCKED' were generated because of the failure of the reversers to deploy normally. The nose landing gear initially touched down 785 metres beyond the threshold but pitch increased again and contact was lost. A "strong nose-down input" was applied and the nose landing gear touched down a second time 1050 metres beyond the threshold. At this point, the absence of ground spoiler deployment was noticed and a manual selection was made with "maximum thrust from the thrust reversers was reached one second later". At this point (see the illustration below) there were 655 metres of runway remaining and a slow deviation left began. In response, a sharp input on the right rudder pedal and an application of right brake were made, but the deviation continued and the aircraft, "skidding to the right, ran off the runway to the left 385 metres from the runway end at a ground speed of approximately 95 knots". Thereafter, it struck a runway edge light, and the PAPI installation for runway 31 before passing through the metal perimeter fence, across a road and into an area of trees where both wings were detached and released fuel instantly ignited as the aircraft came to a stop.
Only one of the two Rescue and Fire Fighting Services personnel mandated under the specifically pre-arranged Level 5 Fire Cover at Le Castellet was present and although they responded quickly, they were unable to bring the fire under control. In respect of this manning level failure, it was found that the request for upgrade of cover from the 'baseline' level 1 to level 5 for the accident flight arrival had been correctly made in advance and accepted by the aerodrome operator. The RFFS unit had then been informed but "the second fire-fighter, scheduled to ensure level 5 protection, arrived late" and since the AFIS Officer was not informed of this absence, they were unaware of the lack of level 5 protection. It was also noted that the responding RFFS operator did not have access to keys which would unlock all gates they might (and did) encounter during emergency response.
The annotated ground track of the aircraft based on FDR data (reproduced from the Official Report)
The Investigation noted that the required Landing Distance for the accident flight was comfortably within the LDA even with up to a 10 knot tail wind component which was not the case. FDR data showed that the approach flown had been stable. It was also noted that the rocks and tress which were struck were outside the Runway Safety Area designated in accordance with International Civil Aviation Organisation (ICAO) guidance.
CVR data showed that although "the atmosphere between the pilots was good and the observable stress level low", several normal Checklists were not requested or not done and, in particular, that the “Before Landing“ Checklist was not fully completed. It also showed that there had been no verbal exchange between the two pilots during the landing roll.
The Investigation noted that the immediate cause of the accident was that the crew had been unable to regain directional control of the aircraft as it deviated left despite initiating intuitive input to the rudder pedals and using differential braking. It was found that a number of nose gear steering system failure modes are not detected and in such cases, no 'STEER BY WIRE FAIL' warning would be generated and, crucially, the steering system would not automatically disengage. Certification assumptions that a failure can be countered by an input on the rudder pedals and differential braking were found to be inadequate in the case of both the investigated accident and in a previous USA accident which had led to the introduction by Gulfstream of a new pilot response drill for "Uncommanded nose wheel steering" in 2004. This required pilots facing this situation to:
- use the rudder and differential braking to control and correct the flight path;
- set the PWR STEER switch to OFF in order to disable the steering system.
The new procedure was found to have been notified in a letter to all G-IV operators advising that "control problems of the steering system may arise during landing and be undetectable or undetected until they have occurred" . The procedure had been subsequently incorporated in both the Aircraft Flight Manual (AFM) and AOM but not also added to the rest of the manufacturer's operational documentation, specifically the manufacturer's Quick Reference Handbook (QRH) (which should always include all abnormal procedures published in the AFM), the 'Cockpit Reference Handbook' and the 'Initial Pilot Training Manual'. It was found that the notification letter had not been sent to all relevant training organisations and specifically not to the contracted pilot training organisation used by the Operator, CAE Simuflite. The Operator stated to the Investigation that they were "not aware of this letter" and the drill was not available to or likely to have been otherwise known to the pilots, who had understandably not immediately set the PWR STEER switch to OFF when the unexpected deviation to the left began.
It was also noted that the fact that the new drill "could not be fully carried out from the right seat had not been identified by the FAA, Gulfstream or training organizations".
The Investigation considered that the accident illustrated the importance of ground spoiler deployment on touchdown and noted that whilst the unintended failure to arm them in this case could be linked to failure to complete the ‘‘Before Landing’’ Checklist, there was no indication to the crew of spoiler position or any normal procedure for the PM to confirm actual activation of the spoilers once the aircraft was on the ground. It was also noted that the G-IV "does not provide for the automatic deployment of ground spoilers when the thrust reversers are selected" which is common on other aircraft types of the same generation. It was additionally considered that the EICAS caution 'GND SPOILER UNARM' which did occur prior to the accident landing was not an effective way to help the crew to detect such an omission. And it was noted that the fact that the 'Before Landing' Checklist order put the EICAS Check ahead of the action to arm the ground spoilers "could habituate crews to the presence of the 'GND SPOILER UNARM' message when checking the EICAS".
The formally-documented Conclusion of the Investigation was that the accident had been caused by (a) combination of the following factors:
- the ground spoilers were not armed during the approach;
- (the) lack of a complete check of the items with the ‘‘before landing’’ checklist, and more generally the UJT crews’ failure to systematically perform the checklists as a challenge and response to ensure the safety of the flight;
- procedures and ergonomics of the aeroplane that were not conducive to monitoring the extension of the ground spoilers during the landing;
- a possible left input on the tiller or a failure of the nose gear steering system having caused its orientation to the left to values greater than those that can be commanded using the rudder pedals without generating any (corresponding) warning;
- a lack of crew training in the ‘‘Uncommanded Nose Wheel Steering’’ procedure, provided to (deal with) uncommanded orientations of the nose gear;
- an introduction of this new procedure (without it being) subject to a clear assessment by Gulfstream or the FAA;
- failures in updating the documentation of the manufacturer and the operator;
- monitoring by the FAA that failed to detect both the absence of any update of this documentation and the operating procedure for carrying out checklists by the operator.
A total of 24 Safety Recommendations were made as a result of the Investigation as follows:
- that the FAA assess the appropriateness of making inhibition of the nose gear steering system at high speed on G-IVs mandatory, to prevent the nose gear from being oriented at large angles at high speed. [2015-031]
- that the EASA assess the appropriateness of making inhibition of the nose gear steering system at high speed on G-IVs mandatory, to prevent the nose gear from being oriented at large angles at high speed. [2015-032]
- that the FAA carry out a study to identify the aircraft that may be affected by the previous recommendation. [2015-033]
- that the EASA carry out a study to identify the aircraft that may be affected by the previous recommendation. [2015-034]
- that the FAA in conjunction with Gulfstream evaluate the effectiveness of the 'Uncommanded Nose Wheel Steering' procedure in order to ensure that the actions proposed in this procedure effectively enable control of the aeroplane to be regained in the case of a lateral deviation due to a malfunction in the speed range in which this procedure is required. [2015-035]
- that the FAA ensure that G-IV operators and organizations training G-IV pilots complete their training by adding the training on the 'Uncommanded Nose Wheel Steering' abnormal procedure and ensure that this training is actually followed by the G-IV pilots and is adapted for the seat occupied in the cockpit. [2015-036]
- that the EASA ensure that G-IV operators and organizations training G-IV pilots complete their training by adding the training on the 'Uncommanded Nose Wheel Steering' abnormal procedure and ensure that this training is actually followed by the G-IV pilots and is adapted for the seat occupied in the cockpit. [2015-037]
- that the FAA ensure that the updating process of the documentation for operators and training organizations ensures that the procedures and training programmes provided for crews contain the latest updates of the manufacturer’s procedures. [2015-038]
- that the EASA ensure that the updating process of the documentation for operators and training organizations ensures that the procedures and training programmes provided for crews contain the latest updates of the manufacturer’s procedures. [2015-039]
- that the FAA ensure that training organizations are systematically sent the information and safety warnings issued by manufacturers. [2015-040]
- that the EASA ensure that training organizations are systematically sent the information and safety warnings issued by manufacturers. [2015-041]
- that the FAA ensure that G-IV operators and Gulfstream set up procedures conducive to verifying the activation of the ground spoilers during landing, similar to that used for thrust reversers. [2015-042]
- that the EASA ensure that G-IV operators and Gulfstream set up procedures conducive to verifying the activation of the ground spoilers during landing, similar to that used for thrust reversers. [2015-043]
- that the EASA in coordination with FAA assess the compliance of the G-IV with the certification requirements relating to the indication of the position of the ground spoilers. [2015-044]
- that the EASA ensure that the Certification Specifications (article 25-699 of the CS 25 regulations) require that information on the position of the ground spoilers be available on landing. 
- that the FAA ensure that the Certification Specifications (article 25-699 of the FAR 25 regulations) require that information on the position of the ground spoilers be available on landing. [2015-046]
- that the FAA ensure that Gulfstream review the warning logic when the ground spoilers are not armed, and the spoiler arming procedure, in order to cover the situation of a failure to arm the spoilers brought to light by this investigation. [2015-047]
- that the FAA ensure that carrying out checklists in ‘‘challenge and response’’ mode becomes systematic practice at Universal Jet Aviation. [2015-048]
- that UJT remind crews of the significance and importance of carrying out checklists in ‘‘challenge and response’’ mode. [2015-049]
- that CAE Simuflite remind crews of the significance and importance of carrying out checklists in ‘‘challenge and response’’ mode. [2015-050]
- that the FAA, in conjunction with Gulfstream, review the relevance of changing the order of the items on the ‘‘before landing’’ checklist in order to place the EICAS check after arming the ground spoilers. [2015-051]
- that the DGAC ensure that aerodrome operators have defined procedures that guarantee that the level of protection provided corresponds to that indicated, including:
- informing the ATC service (or the AFIS agent) in case of an unexpected decrease in the level of RFFS protection initially guaranteed;
- the definition of the conditions for carrying out each of the ancillary tasks requested of aerodrome fire-fighters so that they do not jeopardize the outcome of the operational objective. [2015-052]
- that the DGAC ensures that (the procedures described in [2015-052] are in fact applied. [2015-053]
- that the Operator of Le Castellet Aerodrome ensures that the RFFS service has access at all times and to all places in the aerodrome area. [2015-054]
The Final Report was published in English translation on 23 October 2015. It was also published on the same date in French. The BEA advise that "as accurate as the translation may be, the original text in French is the work of reference".