The pilot reported that shortly after levelling off and at the commencement of cruise checks there was a loud bang followed by a shudder. The pilots first reaction was to check the caution panel and engine transmission temperatures and pressures. All indications appeared normal. The pilot then reduced power slightly suspecting a possible drive train malfunction. However, his attention was then diverted to a flapping noise at the front of the aircraft, at which time he observed that the right-hand forward float bag was deployed, but uninflated. He continued to reduce power, reduced airspeed to 75 knots, and commenced a turn back towards the Sedco 702 platform.

The pilot reported that he was flying an AT502 Air tractor aircraft in calm weather conditions, applying liquid fertiliser to a cereal crop. Whilst conducting a procedure turn for the clean-up run, he felt one wing drop slightly followed by the other wing. After the third wing drop, he managed to maintain wings level, but the aircraft seemed to be in turbulence and continued to descend. The aircraft struck the ground with the right wing and broke up with the pilot sustaining leg injuries.

The pilot reported that the turn was conducted close to the stall, and he considered that the turbulence experienced was created by the aircraft during the turn.

FACTUAL INFORMATION

Circumstances

The company's primary source of revenue was from scenic flights over the Great Barrier Reef. The flights were usually conducted in single-engine aircraft and at an altitude of 1,500 ft AMSL.

The planned one hour scenic flight was the third in a series of similar flights undertaken by the pilot in the aircraft that day. The pilot reported that, after a normal take-off, he was instructed by air traffic control to make a left turn, maintain 1,000ft, and track to a position abeam the control tower before proceeding to Green Island. When the aircraft was approximately abeam the control tower, the engine began to run roughly. This was followed by a significant power loss. The pilot immediately transmitted a mayday broadcast on the Cairns Approach frequency, advising that the engine had failed and that he would be returning to the runway. However, a short time later, the pilot assessed that the aircraft did not have sufficient altitude to reach the runway and he decided to land in a cleared area a short distance inland from the coast.

The pilot reported that he changed the position of the fuel selector soon after the engine lost power. He did this without looking down at the selector which was positioned on the floor of the aircraft. A short time later, when there was no apparent response from the engine, he moved the selector back to its original position, again without looking at the selector.

In the subsequent forced landing, the aircraft landed heavily in a left wing low, nose down attitude. It slewed left and hit a road sign before crossing a narrow sealed road and coming to rest against the gutter. There was no fire.

Communications

The pilot's distress call was his first transmission after changing to the Cairns Approach frequency. The approach controller acknowledged the call and then asked the pilot to report his altitude and confirm the aircraft's registration. Later, the controller requested that the pilot change to the tower frequency. The pilot subsequently advised that these requests increased his workload and distracted him from the primary task of flying the aircraft.

The Manual of Air Traffic Services, page 17-2-1, stated in part; `Distress or urgency communications should be maintained on the frequency on which it was initiated until it is considered that better assistance can be provided by transferring to another frequency'. It also stated that `Staff shall be conscious of the distracting effect that information requests may have on the aircrew'. The investigation could not determine whether, or to what extent, the pilot's performance was affected by the requests from the approach controller.

Wreckage examination

The impact marks and the nature of the damage indicated that the aircraft struck the ground at low forward speed, but with a high descent rate. The aircraft sustained major structural damage to the forward fuselage area. The main spar of the left wing was broken at about the mid-span position and the nose and left main landing gear assembly had separated from the airframe. The engine/cockpit firewall and cockpit floor on the pilot's side was compressed rearward and upward into the cockpit area. The flaps were up.

The right-wing fuel tank, the fuel filter, and the carburettor bowl were empty. The left-wing tank contained approximately 60 lt of fuel. Distortion of the cockpit floor had locked the cockpit fuel selector control in the right tank position, confirming that it was in this position at impact. All fuel lines were clear of obstructions. The nature of the damage to the propeller indicated that it was either stationary or rotating slowly at impact. The magnetos and carburettor were undamaged. These were retained and the engine was removed from the wreckage and test run. It started and operated normally.

The aircraft fuel system

Two integral tanks, one in each wing supplied fuel to the engine. Fuel from these tanks flowed through the fuel selector to a reservoir tank under the cockpit floor, then through a fuel shutoff valve and fuel strainer to the engine driven fuel pump. The fuel selector had three positions - LEFT, RIGHT and BOTH. Useable fuel in each wing tank was 91 lt. The aircraft was refuelled to approximately 60 lt per tank (120 lt total) prior to the first flight on the day of the accident. Recent fuel consumption tests conducted by the operator confirmed a usage rate of about 30 lt per hour.

The pilot advised that he normally operated the aircraft with the fuel selector in the BOTH position, as it was his experience that the tanks emptied at about the same rate during normal operations. Another company pilot advised that, because of an earlier indication of fuel imbalance, he had returned from the last flight on the previous day with the right tank selected. He did not move the selector from that position at the conclusion of the flight.

The aircraft operating handbook called for the fuel selector to be in the BOTH position for engine start and for it to be selected to BOTH during the pre-landing checks. A note in the handbook stated that the purpose of this check was to prevent engine failure due to one tank running empty.

Survival aspects

Whilst the outcome of this occurrence was not a ditching event, it was considered relevant to investigate aspects of survivability given that the majority of flights conducted by the company were overwater scenic flights at an altitude of 1,500 ft AMSL. The single-engine land aircraft used by the company to conduct these flights were therefore operated beyond gliding distance of a suitable landing area in the event of an engine failure for a significant proportion of each flight.

The company provided waist pack type life jackets for passengers and pilots. The normal procedure was for passengers to wear the waist packs, but to not undo the pack or don the jacket. Pilots usually stowed their life jackets under the cockpit seat. The expectation was that, in the event of an emergency occurring while the aircraft was over water, there would be sufficient time for the passengers and the pilot to don their life jackets before the aircraft ditched. The pilot of DZJ reported that, when the emergency situation developed, he had insufficient time to consider the passengers, or to instruct them on the use of life jackets, had that been necessary. A similar situation existed concerning his life jacket.

Recorded radar information

The recorded radar data showed that the aircraft's groundspeed for the last 20 seconds of flight was 49 kts. The aircraft handbook stated that the power off, flaps up stalling speed of the aircraft was 54 kts.

ANALYSIS

The evidence indicated that the engine failed because of fuel starvation that occurred when the right tank contents were exhausted. Given the reported contents of each tank at the start of flying on the day of the accident, and the contents of the left tank when the aircraft impacted the ground, it is likely that the fuel selector was not moved from the RIGHT position after the aircraft was refuelled the day before the accident. This conclusion is supported by the evidence that, in the two-hour period the aircraft had operated that day, the engine would have used about 60 lt fuel.

The reserve fuel tank below the cockpit floor would have emptied after the right tank ran dry, causing the engine power loss. When the pilot moved the fuel selector during the emergency (probably to either BOTH or LEFT), fuel would have begun to flow to the reserve tank. However, it is probable that pilot reselected the empty right tank before there was sufficient fuel in the reserve tank to restore engine power.

The recorded radar data indicated that the aircraft speed in the latter stages of the flight was close to, if not below, the stalling speed. The aircraft impact attitude, and the extent and nature of the damage, supports this conclusion. This evidence, along with the fact that the flaps were in the UP position at impact, indicates that the pilot had mismanaged the aircraft during the forced landing.

SIGNIFICANT FACTORS

1. The pilot did not select the fuel system to the BOTH position prior to take-off.

2. The engine ceased operation due to fuel starvation.

3. The pilot did not configure the aircraft appropriately for the forced landing.

4. The pilot did not maintain proper control of the aircraft.

SAFETY ACTION

While it was determined not to be a contributing factor in this occurrence, the Bureau believed that the safety implications of overwater operations in single-engine land aircraft carrying fare-paying passengers, required further investigation.

The Bureau considers that overwater operations in single-engine land aircraft carrying fare-paying passengers, at altitudes that will not allow the aircraft to glide to land (suitable for an emergency landing) in the event of an engine failure, is a high-risk practice. In particular, the outcome of a ditching event in a high-wing aircraft similar to the accident type presents obvious exit problems. The fact that this aircraft was equipped with fixed landing gear further reduces the survivability of a ditching event.

As a result, the Bureau of Air Safety Investigation issued the following interim recommendation to the Civil Aviation Safety Authority on 28 Jan 1998:

R970176

The Bureau of Air Safety Investigation recommends that the Civil Aviation Safety Authority reconsider the conditions of the current exemption to CAR 258 as it applies to passenger-carrying charter operations in single-engine land aircraft with a view to:

(a) minimising the likelihood of a ditching event; and

(b) minimising the risks associated with the outcome of a ditching event.

no title

Previous safety action

A previous fatal accident highlighted the safety implications of the practice of not correctly "donning" waist pack type life jackets during operations at or below 2,000 ft AMSL.

As a result of the investigation of that occurrence, the Bureau issued the following interim investigation to the Civil Aviation Safety Authority on 9 December 1997 (only the relevant parts of the interim recommendation have been reported below):

"IR960138

The Bureau of Air Safety Investigation recommends that the Civil Aviation Safety Authority:

(i) review the current orders and regulations to ensure that the intention of Civil Aviation Order 20.11 part 5, governing the wearing of a life jacket is clear and unambiguous, and that jackets worn in accordance with the order afford the wearer the maximum safety benefit; and

(ii) educate the industry on the need to have life jackets worn in such a manner that they afford the wearer the maximum safety benefits."

The following response was received from the Civil Aviation Safety Authority on 21 May 1996 (only the relevant part of the response has been reproduced below):

"Interim recommendation(i)

It is CASA's opinion that the current provisions of CAO 20.11.5 are essentially adequate. However, this issue will be referred to the relevant Technical Committee under the Regulatory Review Program for its review.

Interim recommendation (ii)

CASA supports this proposal and is considering the best means to give effect on this recommendation."

Ongoing safety action

As a result of this and other similar occurrences, the Bureau of Air Safety Investigation intends to conduct a review of Australian aviation occurrences involving fuel starvation and exhaustion. This study is due to be completed prior to July 1999. Any recommendations issued as a result of this safety study will be published in the Bureau's Quarterly Safety Deficiency Report.

During cruise at FL370 in IMC, all aircraft automatic navigational aids, autopilot, autothrottle, flight directors and the attitude director indicator mode annunciators, failed simultaneously. The crew were unable to regain the flight directors or re-engage an autopilot and the inputs to these systems were unavailable. The flight management computers continued to function, but the waypoints had to be sequenced manually, and the navigation log had to be manually upgraded for time and fuel calculations. There was no wind, time or temperature available, but groundspeed was displayed. Attempts to reinstate the automatic equipment by cycling the circuit breakers was unsuccessful. The aircraft was flown manually to Brisbane without further incident.

Ground investigation at Brisbane determined that both sensors in the total air temperature (TAT) probe had failed. The operator is investigating the failure in conjunction with the manufacturer.

The instructor/supervisor reported that the student pilot was engaged in solo crosswind circuit practice. It was the pilot's first landing for this session. It appeared that the pilot misjudged the approach, and the aircraft bounced back into the air following the first touchdown. Instead of entering a go-around, the pilot attempted to recover for a normal landing. During this process the aircraft touched down nosewheel first causing the nosewheel leg to fail in overload. The propeller struck the ground, and the aircraft slid to a stop on the runway.

The pilot started the engine and set about 75 percent RPM. He then experienced an urgent call of nature and left the helicopter unattended with the engine running. While the pilot was absent, the passenger boarded the helicopter (although the pilot had apparently told him not to do so). As the pilot was returning to the helicopter, he saw it become airborne at the same time as the passenger exited the helicopter. The helicopter rolled on to its side and was substantially damaged. The passenger could not recall whether he had moved any of the cockpit controls.

A distress phase was declared after the crew reported they were diverting the aircraft for a landing at Port Macquarie due to a fire warning in the cargo hold. The crew reported that they had discharged the fire extinguisher for the cargo hold in accordance with the emergency procedures. However there was no other indication of fire. The aircraft landed safely a short time later and all passengers disembarked without incident.

Post flight examination confirmed that the fire warning was spurious. Company engineering personnel advised that it is a known problem. The manufacturer is currently modifying the fire detector system for this aircraft type. The operator will submit a Major Defect Report for the incident to CASA Airworthiness.

The pilot reported that the helicopter landed heavily in a carpark, after the engine lost power during takeoff. On lift-off, engine indications were normal with 28-29 inches boost while in the hover. As the helicopter accelerated, the pilot felt a deterioration in performance and observed that the boost had fallen to 15-18 inches. During the subsequent landing, the helicopter's main rotor blades struck a tree and both landing skids were bent. The helicopter remained upright. Post flight examination found that the turbo charger had failed.

FACTUAL INFORMATION

The B737 had departed from Melbourne on a flight to Cairns, Qld, two minutes ahead of the B747 on the same route. Both aircraft were climbing to an initial altitude of flight level (FL) 200 and the crews had been approved to vary their climbing speed. The crew of the B737 increased their speed to 300 knots and the crew of the B747 were approved to conduct a "high speed climb". Due to an active restricted area, both aircraft were given a radar vector to take them off track and jurisdiction for the flights was transferred to the Inner North controller. Both crews contacted the Inner North controller whilst on a heading of 010 degrees, with the B747 paralleling the track of the B737 to the right with approximately 8 - 10 NM separation.

The closing speed at the time of the transfer was approximately 50 knots. The B737 was then cleared to climb to flight level (FL) 310 and the B747 to FL390. When the aircraft were clear of the restricted area, the controller issued instructions for both crews to track from their respective positions direct to the reporting point TOBOB, at which they were to intercept their planned route. The planned track to TOBOB was approximately 355 degrees and the instruction would result in a turn of approximately 20 - 25 degrees to the left for both aircraft. The controller did not realise that the second aircraft was a B747 and thought it was an aircraft of a lower performance. He was not familiar with this type of aircraft on this route and, although he knew that the registration belonged to a B747, did not reconcile that particular type with the traffic situation.

The wind was from the north-west, with forecast speeds of 70 knots at flight level (FL) 185. The crew of the B737 were the first to turn their aircraft in accordance with the controllers instructions and chose an initial heading that was to the north-west and directly into wind. This turn had the effect of reducing the ground speed by almost 60 knots. The turn was approximately 55 degrees, more than that required to track direct to TOBOB, and an adjustment was made about 1 minute later. Because of the position of the B747 relative to the B737, its turn of approximately 20 degrees resulted in a more northerly heading and, as such, the reduction in its ground speed was less than 10 knots.

These changes resulted in a significant increase in closing speed at a time when the Inner North controller became occupied with other work-related tasks. Consequently, he did not initially notice the reduction in separation. Shortly after, as he conducted a scan of his area of responsibility, he saw that the B747 was closing rapidly on the B737 and that there was less than the required vertical separation with the B747 below. He immediately attempted to achieve vertical separation by instructing the crew of the B737 to maintain FL210. The intent was to allow the B747, which had a greater rate of climb, to achieve the necessary vertical separation before the horizontal separation standard was infringed.

However, it became obvious that this plan would not achieve the objective and he issued a radar vector to the crew of the B747 in order to maintain radar separation. Before the distance between the aircraft started to increase, the horizontal separation had reduced to 4.2 NM while vertical separation was 200 ft. The required standard was either 5 NM horizontally or 1,000 ft vertically. A breakdown of separation had occurred. Examination of the radar data indicated that the closing speed had increased to 160 knots for a short period prior to the radar vector taking affect.

SIGNIFICANT FACTORS

1. The Inner North controller assumed that separation would be achieved through differences in aircraft performance without employing separation assurance techniques.

2. The crew of the B737 executed a turn which was greater than that expected by the controller.

3. The Inner North controller did not immediately recognise the combination of events that led to a significant increase in closing speed.

The pilot was attempting the first take-off from a grass paddock on an agricultural operation. Take-off direction was towards the west with 700 m of ground roll available. The wind was variable, but mainly a gusting crosswind from the right at around 10 kt. The aircraft was below maximum weight with a full hopper load and about half fuel.

When it became evident that the aircraft would not lift off by the end of the available take-off area, the pilot rejected the take-off and attempted to stop in the paddock remaining.

The aircraft veered to the left, collided with a fence and ground-looped before coming to a stop.

The pilot reported that engine operation was normal, and he was unable to explain why the aircraft did not lift off in the distance available, which should have been adequate. He believed the changing wind strength and direction may have been a factor.