An infringement of separation standards occurred 70 NM east of Darwin, NT, between a descending Boeing 737-376 (737) and an Embraer EMB-120 (Brasilia) that was maintaining level flight. The event took place during the hours of darkness and in visual meteorological conditions. The crew of the 737 intentionally flew the aircraft through its assigned level in response to a traffic alert and collision avoidance system (TCAS) warning. The Brisbane sector controller also received a short-term conflict alert (STCA) between the two aircraft from the Australian Advanced Air Traffic System (TAAATS). The STCA alerted controllers when the radar trajectories of two aircraft indicated that separation standards might be infringed. The 737 and the Brasilia passed within 1.6 NM horizontally and 600 ft vertically. The required separation standard was either 3 NM or at least 1,000 ft.

TCAS is an airborne device that functions independently from the ground based air traffic control system and provides collision avoidance protection for a broad range of aircraft. The system fitted to the 737, TCAS II version 6.04, provided recommended escape manoeuvres in the vertical dimension, to either increase or maintain the existing vertical separation between the aircraft. The escape solution was communicated directly to the flight crew via a cockpit display and a synthesised voice attention getter. The TCAS alerts in the 737 used a liquid crystal display (LCD) instantaneous vertical speed indicator (IVSI) with red and green markings to indicate the vertical speeds to be avoided (red), and the desired vertical speed to be flown (green). The display was 70 mm square and had a fixed range of 6.5 NM. Aircraft were depicted using geometric symbols, depending on their threat status. A partial aircraft symbol was displayed at the extremity of the screen for aircraft detected beyond the display range. TCAS information could, in principle, have been incorporated in the electronic flight instrument system display (EFIS) of that aircraft but wiring and space available in the electronics racks of the aircraft would have required a major modification programme. TCAS information that is incorporated in the EFIS display allows the pilot to show traffic at longer ranges.

TCAS equipment interrogates the transponders of other aircraft to determine their range, bearing and altitude. Accordingly, the TCAS does not provide protection against aircraft that do not have an operating transponder. The Brasilia was fitted with an operating mode C transponder (altitude encoding) but was not TCAS equipped and the crew was not immediately aware of the infringement of separation standards. The Brasilia, which had 30 passenger seats and had a maximum take-off weight of 11,990 kg, was not required to be TCAS equipped by the Australian legislation current at the time. That legislation mandated the carriage of TCAS for all turbine-powered aircraft with more than 30 passenger seats or a maximum take-off weight greater than 15,000 kg. When both potentially conflicting aircraft are fitted with TCAS, both TCAS units co-ordinate their intentions to provide appropriate co-coordinated avoidance manoeuvres.

TCAS II can issue two types of alerts:

• Traffic Advisory (TA) to assist the pilot in the visual search for the intruder aircraft and to prepare the pilot for a potential RA; and
• Resolution Advisory (RA) to recommend manoeuvres that will either increase or maintain the existing vertical separation from an intruder aircraft.

Once the risk is over, the TCAS issues a synthesised voice `clear of conflict' message.

The 737 crew were on their fourth consecutive day of duty and completing the last sector of a four-sector day when the incident occurred. It took place at the transfer of control point between Brisbane Centre and Darwin Approach. The 737 was en-route from Cairns to Darwin and was on descent to FL220. The Brasilia was enroute to Groote Island, under the control of Brisbane Centre, with instructions to maintain flight level (FL) 210. Those routes placed the two aircraft on almost reciprocal tracks. As the 737 was passing FL235, the crew were instructed by the Brisbane sector controller to contact Darwin approach control for further descent, but were not advised of the opposite direction traffic.

The Manual of Air Traffic Services (MATS) gave guidance to controllers on how best to manage the situation and advised:

`6.6.5.5 Frequency change management in relation to the transit of an aircraft across airspace boundaries shall be arranged in a manner that normally enables pilot communication with the unit responsible for the airspace within which the aircraft is operating. A lateral tolerance of 10 NM either side of the boundary is permitted, except that when entering controlled airspace, the frequency change shall be within the 10 NM prior to the boundary.

6.6.5.10 Aircraft should normally remain on the frequency appropriate to the airspace in which it is operating. However, aircraft may be transferred to another ATS frequency provided that:

1. significant operational advantage will be gained;
2. workload, communications and equipment capabilities will permit the responsible controller to take such action as is necessary to preserve the separation without delay;
3. the actual separation is in excess of the minimum;
4. there is no possibility of separation being reduced to the minimum with the normal operation of the aircraft.'

MATS does not require controllers to provide traffic information to either crew in these circumstances.

In order to comply with MATS 6.6.5.10, the Brisbane sector controller could have either:

• Initially assigned the crew of the 737, FL230 on descent, instead of FL220, then assigned the responsibility for separation to the Darwin approach controller, coordinating the terms of the transfer of the 737 with him, before transferring the aircraft to the Darwin approach frequency; or
• Assumed the separation responsibility, coordinated the terms of the transfer of the 737 with the Darwin approach controller, and transferred the aircraft once the 737 had passed the Brasilia and a horizontal separation standard existed.

Airservices Australia believed that MATS section 6.6.5.10 did not apply to a change of frequency in these circumstances. According to Airservices Australia the action of the Brisbane sector controller, in assigning the crew of the 737 FL220, was consistent with MATS in that the controller applied a 1,000 ft separation standard between the two aircraft and the 737 crew was transferred to the Darwin approach controller within 10 NM of the lateral boundary between the Brisbane sector and Darwin airspace.

Shortly after acknowledging the instruction to change frequency, the 737 crew received an aural `traffic, traffic' warning and a display indication of an aircraft 5 NM ahead. The pilot in command stated that the TA quickly changed to a RA with a `descend, descend, descend' aural alert. As the aircraft was approaching its assigned level he disconnected the autopilot and pitched the aircraft nose down with the intention of following the RA commands. He stated that the required rate of descent shown on the IVSI was 1,200-1,500 ft/min. On passing FL220 the TCAS command abruptly reversed to a climb RA (aural `climb, climb now') which was followed positively. The climb annunciation continued until the aircraft was at FL225. No more commands were issued and there was no TCAS `clear of conflict', which is normally generated once a RA is removed.

Analysis of recorded data indicated that as the 737 descended through FL230, its rate of descent was approximately 2,900 ft/min. At FL227, the automatic flight system commenced a transition manoeuvre to achieve level flight at FL220. At approximately FL225 the autopilot was disengaged and the descent was continued manually at a rate of descent in excess of 3,200 ft/min to FL215. That was followed by a climb to FL225 at 2,900 ft/min as the pilot responded to the RA reversal (`climb now' advisory).

Maintenance files from the TCAS computer were examined and no indication of TCAS failure was found. Technical expertise was requested from the manufacturer of the TCAS equipment. Their evaluation of the event presented two possible scenarios.

`Explanation 1. The reported `descend' advisory was actually a `reduce descent' advisory that was misunderstood by the crew. A `reduce descent' would be consistent with the expected TCAS response per the reported geometry of the aircraft. Because the advisory was misinterpreted, the rate of descent was increased rather than decreased until the aircraft was below the TCAS required 700 feet vertical separation. Thus, the TCAS was required to issue a `climb now' advisory. The lack of a `clear of conflict' annunciation is explained in the following paragraph.

`Explanation 2. There is a possibility that the intruder aircraft's (equipped with mode C transponder) altitude report was not correctly received by the TCAS. There have been instances when a Mode C reply will not contain all the appropriate pulses in the message or it transmits pulses that are too narrow for the TCAS to detect. This could cause differing altitude reports and could result in multiple unstable tracks at different altitudes for the same intruder aircraft. This being the case, the TCAS could have issued a `descend' advisory for the intruder because it appeared (due to erroneous altitude report) that it was actually above its own aircraft. If subsequent replies had the correct altitude, the erroneous track would be dropped by the TCAS and the TCAS would issue a `climb now' advisory on the track with the correct altitude.

`The reason that `clear of conflict' was not annunciated can be attributed to low track firmness of the intruder aircraft. Mode C equipped aircraft are typically only equipped with a single antenna mounted on the lower hull. Since the intruder aircraft was below the 737 aircraft, it is likely that the TCAS was not able to receive regular replies at close proximity. The TCAS computer unit will `coast' the track of a previously established intruder file if it does not receive a valid or reasonable interrogation response. If the track of the intruder that generated the RA is coasted during the time of the associated RA, then the `clear of conflict' is not announced.'

Since the incident, the operator's TCAS software has been updated to Version 7. The objectives of the Version 7 update were to further increase the safety benefits of TCAS, make TCAS more compatible with the procedures used by ATC and to address operational concerns identified by pilots operating the older versions of TCAS. Improvements to the aural annunciations included a change from `reduce descent, reduce descent' to `adjust vertical speed, adjust'.

The pilot and passenger of a Robinson R22 helicopter were conducting an aerial inspection and cattle mustering flight. During the flight, cattle were observed outside the fenced area and the pilot descended the helicopter to direct the cattle back towards the fence. The passenger then directed the pilot to fly the helicopter along the eastern fence to check its security. The passenger reported that the pilot had just commenced to climb higher, at his request, when the helicopter collided with a single-wire powerline. Recorded Global Positioning System data indicated that the helicopter struck the powerline at a speed of 55 kts. The wire did not break and the helicopter pitched nose down. The main rotor severed the tail boom and the helicopter collided with the ground 69 m beyond the powerline. It impacted in an inverted attitude, facing back along its direction of travel and rolled forward onto its left side. There was no evidence of fire in-flight or after the impact.

The main rotor, mast and upper right side of the helicopter cabin took the main impact and the mast was partially driven into the cabin. The pilot, who occupied the right seat, received fatal injuries. The passenger, although seriously injured, walked 200 m to a track and waited almost 2 hours until found by a passing motorist. The passenger reported that injury prevented him from picking up and activating the portable Emergency Locator Transmitter that was ejected from the helicopter cabin.

The powerline supplied electrical power to a nearby property and was aligned approximately east-west at right angles to the helicopter's flight path. It spanned a distance in excess of 500m from a pole in a saddle on a ridge east of the fence, to another pole set among trees in a timbered paddock. Strike marks on the wire indicated that the helicopter had struck the powerline at approximately mid-span. There were no markers on the powerline. Australian Standards (AS 3891.1-1991) specified markers on powerlines where the height of a cable exceeded 90m. The powerline did not exceed that height and at the point of contact was about 20m above ground level (AGL).

The pilot held a current Commercial Pilot Licence (Helicopter) and was appropriately qualified for cattle mustering operations. He held a valid Class 1 medical certificate and did not require any vision correction. An article on the visual aspects of wire detection by Dr Gordon Cable in the Directorate of Flying Safety-Australian Defence Force's special wirestrike edition (1997) of their safety journal "Spotlight" stated that under ideal conditions, the human eye can resolve detail down to an angle of 30 seconds of arc. That equated to being able to see a 5 mm thick wire from a distance of 150 m. However, contrast between the wire and the background against which it is viewed and the medium through which it is viewed, affect detection. The pilot was wearing a helmet with tinted visor. The fragments of perspex from the helicopter's windshield did not show any sign of being scratched or crazed.

The operator's Operations Manual required pilots to familiarise themselves with the area to be mustered, including any obstructions, before commencing mustering operations. Although the passenger had engaged the operator on previous occasions for the same work, no map of the area to be mustered was held on the operator's file. The passenger was aware of the existence of the powerline and stated that on previous occasions some pilots had him draw a "mud map" on the ground to indicate the area to be mustered, including powerline hazards. The pilot had reportedly not asked the passenger about any powerline hazards prior to the flight. There was no evidence that the pilot had previously flown over or inspected the area to determine the presence of hazards.

Forty minutes after departing Cairns en-route to Japan, the flight crew of the Boeing 747-300 aircraft noticed the number 5 left main entry door warning light illuminate. The flight engineer investigated and found that the door handle had moved from its fully locked 4 o'clock position to an unlocked 3-o'clock position. The flight engineer, with the assistance of one of the cabin crew, attempted to move the handle back to the fully locked position but was unable to do so. Returning to the flight deck, a check of the aircraft's technical logbook revealed that this problem had occurred on previous occasions. The pilot in command contacted the operator's maintenance headquarters, where he was advised that the door would not be able to open due to the cabin being pressurised and the flight could continue on to destination.

The cabin crew were advised to monitor the door for the rest of the flight. Shortly before landing, the flight attendant seated adjacent to the door observed the handle moving slowly upwards. Just prior to touch down, the door handle jumped to the 2-o'clock position at which time a loud wind noise could be heard. Leaving his seat, the flight attendant grabbed the handle and forced it down. Paper was observed being sucked under the door as the passenger seated directly in front of the door (adjacent to the window) turned and grabbed the door handle giving assistance in pushing the handle down towards the locked position. The handle reached the horizontal 3-o'clock position with the flight attendant keeping weight on it until the aircraft had landed and taxied to the terminal.

A minor adjustment of the door upper gate was carried out by maintenance engineers before the aircraft's next departure. The next two flights were uneventful, however on the third flight the left main entry door-5 warning light again illuminated and the handle moved from the fully locked to the unlocked, 3-o'clock position. On arrival at the destination, an inspection by engineers found the force required to move the door handle from the locked to unlocked position was lower than required, necessitating further adjustments to the door. An entry was also made in the aircraft's technical log for a full rigging check and an inspection of the door bearings to be carried out on the aircraft's return to its main base. The aircraft then flew for two further flights, without any reported defects, before returning to its main base.

Upon returning to the operator's main base, the door was removed and disassembled for an inspection of the door bearings. No defects were found and the door was returned to the aircraft where it was refitted and a rigging check carried out. The aircraft was then returned to service.

The Boeing 747 aircraft was being operated on a scheduled passenger flight from Melbourne to Perth with an estimated time of arrival (ETA) at Perth of 0945 WST. The flight crew had been provided with an aerodrome forecast (TAF) for Perth, valid until 0600 the following day, which indicated that the visibility and cloud base would be above the alternate criteria throughout the period of the forecast. As there were no operational requirements, the aircraft departed Melbourne without alternate or holding fuel being carried for Perth.

Three minutes after the aircraft's departure from Melbourne, an amended TAF for Perth was issued with fog being forecast until 0800. After that time, conditions at Perth were forecast to improve above the alternate criteria. A trend type forecast (TTF) for Perth, also issued soon after the aircraft's departure, indicated that the visibility would be reduced in fog until 0800. Subsequent Perth TTFs issued from 0635 until 0759 indicated that the visibility would be reduced in fog or mist until 0800 when the conditions were forecast to improve above the alternate criteria.

The FLAKE waypoint, 465 NM east of Perth, was the flight planned point of safe diversion (PSD) based on a return to Adelaide. The crew obtained the 0730 Perth TTF that indicated an improvement in conditions above the alternate criteria from 0800 and, as the aircraft proceeded to the FLAKE waypoint, the Perth TTFs continued to forecast that improvement. The aircraft passed the FLAKE position at 0824 and soon after, a Perth TTF was issued that indicated the meteorological conditions would be below the alternate criteria until 15 minutes after the ETA of the aircraft at Perth.

As the aircraft was not fitted with an aircraft communications addressing and reporting system (ACARS), the operator's flight dispatch section attempted to contact the crew via high frequency (HF) radio but the attempts were unsuccessful. At 0845, a message from the operator about the 0825 TTF was provided to the crew by air traffic services. As the aircraft had flown past the PSD and fuel was not carried to divert to an alternate airport, the crew decided to continue the flight to Perth. The aircraft made an uneventful landing at 0938.

A report from the Bureau of Meteorology (BoM) indicated that a weak cold front with a fragmented cloud band was expected to move through the Perth area between 0800 and 0830. No showers were expected ahead of the front and the BoM aviation forecasters assessed the formation of fog in the expected weather situation as being unlikely. However, a weak ridge formed overnight ahead of the front and in the light wind regime caused by the weak ridge, advection fog formed at Perth Airport.

As the front approached Perth it became apparent to the aviation forecasters that the frontal passage was being enhanced by north-easterly winds in the Perth area. This meteorological situation resulted in reduced visibility and low cloud being observed until 0930. The forecasters had not previously encountered this type of situation, which produced a mix of advection fog and frontal fog that was difficult to forecast.

On 26 March 2002, VH-KEQ, a Saab 340A aircraft, was being operated on a regular public transport flight from Sydney to Canberra. During the pre-flight planning, the crew identified a line of thunderstorms and associated weather moving through the weather forecast area (ARFOR) 21, in which the flight was to be conducted. Consequently, they anticipated that to avoid the adverse weather, they would probably have to divert to the left of the flight-planned track.

At the time, the weather in ARFOR 21 was under the influence of a low-pressure system that was situated over eastern New South Wales (NSW), with troughs aligned to the north and south. An upper-level disturbance was also present over central NSW. Because of this complex weather system, showers and thunderstorms were forecast over the eastern part of NSW during the afternoon and evening. A significant meteorological forecast (SIGMET) for thunderstorms was also current for ARFOR 21 at the time of the flight.

The forecast wind at 10,000 ft and 14,000 ft was 330 degrees true at 40 kts. The freezing level was forecast to be 11,500 ft to 12,000 ft.

The flight departed normally and, at 1750 ESuT, on first contact with Melbourne Centre, at approximately 50 nm to the southwest of Sydney, the air traffic controller advised the crew that there was significant weather on track between Marulan and Canberra. The controller told the crew that preceding aircraft diverted to the left of track to avoid this weather. The crew indicated that they also would divert left of track. The controller cleared the crew to divert to the left of track and once clear of the weather, to track direct to Canberra.

The crew of another aircraft that had diverted to the left of track reported to the controller that they had experienced moderate turbulence at flight level (FL) 140. The controller passed this information to the crew of KEQ, which at the time was climbing to FL120. At 1800, the controller advised the crew of KEQ that they would soon be leaving controlled airspace. Soon after, the crew reported that they would continue on their present heading as the weather radar indicated that conditions were unsuitable for tracking direct to Canberra. This heading resulted in the aircraft continuing to diverge from the flight-planned track. At 1805, they advised that they were descending from FL120 to 9,000 ft. The pilot in command later reported that this was to remain clear of the freezing level.

The pilot in command reported during a post occurrence interview that when the aircraft became clear of cloud they were over the coastline. Recorded radar information indicated that the aircraft was over the Batemans Bay area on the NSW south coast. After conducting a right orbit to confirm their position and to assess the weather, the crew decided to track direct to Canberra.

The pilot in command also reported that the crew was using the airborne weather radar in maximum gain mode, with the antenna tilt set at approximately 3 to 4 degrees up, and with the range at either the 25 or 50 NM setting. He reported that the radar displayed mostly green returns with an occasional yellow return. He also stated that there were no red returns on the radar. The crew interpreted information provided by the weather radar as being suitable to track direct to Canberra.

After setting course for Canberra, the conditions suddenly became dark, associated with an increase in the turbulence level and rain intensity. Shortly after, the ground proximity warning system (GPWS) sounded a `TERRAIN' warning. The crew applied maximum power and began to climb the aircraft, and at 1817, broadcast on the area frequency that they were climbing due to a GPWS warning. The pilot in command reported that during this event the outer pane of the front left windscreen shattered.

The crew continued the climb to FL110 and shortly after levelling the aircraft it became clear of cloud. They then climbed to FL120 and tracked direct to Canberra for an uneventful landing. Post flight inspection of the aircraft revealed that in addition to the shattered windscreen, the strobe lights had been damaged and several vortex generators were missing.

The pilot in command subsequently reported that he thought hail had been falling at the time the GPWS warning activated. Analysis of the recorded transmissions from the aircraft revealed background noise consistent with hail impacting the airframe.

The Bureau of Meteorology (BoM) `Letterbox weather radar' was situated about 27 NM south-south-west of Sydney, and about 90 NM to the north-east of Batemans Bay. The 1803 image from that weather radar was derived at a scan angle of 0.5 degrees. The returns from that area would, therefore, have been from an altitude of about 5,000 ft. Those returns revealed that a thunderstorm cell with reflected rainfall rates of over 100 mm per hour was located about 14 NM to the north-west of Batemans Bay. The 1823 image showed that the thunderstorm was almost overhead Batemans Bay, and on the direct track to Canberra.

BoM subsequently reported that the thunderstorm caused damaging winds and a tornado about 5 NM east of Braidwood, and that the storm was likely to have been a `supercell' thunderstorm.

Weather Radar

The aircraft was equipped with a WXR 200 colour weather radar. The radar antenna transmitted microwave energy in the form of pulses, which, if reflected off precipitation ahead of the aircraft, would be returned to the antenna. The radar beam was a narrow cone with a beam width of 8 degrees. The amount of energy reflected back to the antenna depended on the intensity of the precipitation, and was converted into a colour code for presentation to the crew on their flight instruments. There were four colour codes that were directly related to precipitation intensity, ranging from black (no precipitation), green (minimum detectable moisture), yellow (medium moisture level), to red (strong to extreme moisture level).

The Saab 340A operations manual described the operation of the airborne weather radar system. This information included the expected radar displays equating to reflected precipitation returns, and descriptions of how to avoid areas of severe weather based on weather radar displays. Information was provided on the reflectivity of various types of precipitation, with wet hail being the most reflective, and dry hail and dry snow being the least reflective.

The manual contained information in the use of the weather radar variable gain control. By reducing the gain, red areas of precipitation targets would eventually be displayed as yellow, and yellow areas as green. The red area that was the last to change to the next lowest level (yellow) would be the strongest part of the precipitation target.

The manual also contained information on weather radar antenna tilt setting, and the importance of precision tilt management for detection, analysing, and avoiding hazardous convective weather. The manual recommended that the antenna be set to the one degree down position whenever tilt and range were not being used for weather analysis. That setting provided pilots with a fast and certain means to ensure that the radar was functioning, and ensured the detection of weather returns ahead of the aircraft.

The manual also contained information that heavy rainfall could reduce the ability of the weather radar to provide a complete picture of the weather ahead. That phenomenon is termed `radar attenuation'. If a radar beam is fully attenuated, the display will indicate a radar shadow that appears to be the end of the precipitation area, but which actually extends further than is apparent from the display. Attenuation may reduce reflected precipitation readings by as much as 20dBz, which the weather radar interprets as an area of decreased rainfall. The effect of this is a downward colour shift that the colour returns displayed to the crew indicate a lower level of precipitation intensity than is actually occurring. What would normally be displayed as a red return (indicating strong to extreme rainfall rates, with the possibility of associated hail) is displayed as a yellow return.

Ground Proximity Warning System (GPWS)

The aircraft was equipped with a GPWS that provided 5 modes of protection. One of those modes provided protection against excessive closure rate to terrain. The GPWS processed radio height, flap and gear position logic, vertical speed and indicated airspeed to determine if a dangerous situation was developing with respect to the aircraft's height above the ground. The mode had two warning envelopes. Penetration of the first envelope resulted in a `TERRAIN' aural warning with the corresponding warning lights. If the second envelope was penetrated, the aural warning changed to a `WHOOP WHOOP - PULL UP' that was repeated until the aircraft gained 300 ft radio altitude.

Analysis of recorded flight data indicated that at the time the crew received the GPWS warning, the aircraft was at an altitude of 9,248 ft above mean sea level and flying at an indicated airspeed of 231 knots. The radar altimeter was indicating 2,154 ft and the derived rate of descent from the radar altimeter was 17,760 ft per minute. These parameters were within the warning envelope for the GPWS computer and resulted in a `TERRAIN' warning being provided to the crew.

Technical Analysis Report: Examination of a Failed Rolls-Royce RB211-524 Turbofan Engine Boeing Commercial Aircraft Group, 747-436, G-BNLD

FACTUAL INFORMATION

History of the flight

On the evening of March 1, 2002, Boeing 747-436 aircraft G-BNLD sustained the failure of the number 3 (right inboard) engine during a scheduled regular passenger transport flight from Sydney to Bangkok. The flight crew experienced vibrations and received an ENG 3 REVERSER engine indicating and crew alerting system (EICAS) message. The crew shut down the number 3 engine and completed checklist items before returning the aircraft to Sydney.

An initial engineering examination found that a fan blade from the number 3 engine had failed, and that debris had punctured the engine cowl, the right wing leading and trailing edge flaps and the fuselage, damaging a structural member above the wing root area. The inspection found fractured fasteners and other components beneath the fan cowls and damage to the structure associated with the thrust reverser assembly. Debris from the number 3 engine was also found embedded within the intake cowl of the adjacent number 4 engine.

The AS350B2 Squirrel helicopter was being operated on a private flight with the pilot and two passengers on board. The pilot reported that shortly after lifting the helicopter into an approximately 1.2 metre hover, he noted that the main rotor system had a pronounced vertical, once per revolution, vibration. The pilot then elected to terminate the hover and land the helicopter. He further reported that when the skid landing gear touched the ground, the helicopter began to oscillate violently. The pilot then activated the emergency fuel cut-off. Subsequently, the engine and main rotor revolutions per minute (RPM) began decreasing. The pilot and passengers reported that the oscillations of the helicopter became more violent and pronounced as the main rotor RPM decreased. Once the main rotor ceased rotation, the occupants exited the helicopter. One passenger received minor injuries.

The helicopter sustained substantial damage to the main rotor assembly, the right landing gear skid, the forward cargo mirror mount bracket, the left and right structure keel beams, and the right rear passenger seat support. The principle damage to the main rotor assembly consisted of the fracture and separation of the yellow and blue starflex rotor arm outboard segments.

An examination of the helicopter main rotor head and blades did not reveal any anomalies, other than the separated starflex rotor arm outboard segments, that could have resulted in the vertical vibration reported by the pilot. A witness near the helicopter during its hover flight did not report any foreign objects or birds in the area of the main rotor disc during the flight.

The damage to the main rotor starflex rotor assembly was consistent with the damage documented in a technical report compiled by the Australian Defence Science and Technology Organisation relating to a previous military AS350 helicopter occurrence. That report indicated that the starflex rotor arms failed due to severe upward bending due to excessive loading. That investigation determined that the circumstances of the accident were consistent with a ground resonance event.