Accident Recovery: Not Like Riding a Bike

Unlike bicycle riding, certain critical aviator skills require more frequent, consistent practice to retain.

Bicycle riding is the stereotypical example of a durable skill: once learned, it’s just about permanent. Even after a lapse of years, a rider can climb back into the saddle and take off after just a couple of wobbles. The experience of rusty pilots climbing back into the cockpit suggests that some aspects of basic airmanship work in the same way. The view from the pilot’s seat reinforces the feeling that the controls reawaken the memory of physical aircraft control, so simple maneuvers can be accomplished on the first flight back in the cockpit.

Other skills, however, are highly perishable, requiring frequent and consistent practice to retain—particularly those that require interpreting abstract information in potentially counterintuitive ways. Attitude instrument flying is the quintessential example. An April 2023 report by New Zealand’s Transport Accident Investigation Commission (TAIC) identified proficiency with night vision goggles (NVG) as another.

The Mission

On Apr. 22, 2019, a fishing vessel in the Southern Ocean requested an air medical evacuation of a crew member in need of emergency hospitalization. The flight would also provide an opportunity to replenish the ship’s stock of first-aid supplies.

At the time, the boat was 210 nm south of the Auckland Islands. A frontal system was approaching New Zealand from the south, so the air ambulance operator planned to position the aircraft on Enderby Island, where the company maintained a shelter and a fuel supply, before the front arrived. The aircrew would spend the night while the vessel sailed toward a protected anchorage.

The plan was to fly to the rendezvous point with only minimal equipment, return to shore to stabilize the patient, move him to a more comfortable stretcher while the helicopter refueled, then transport him to Invercargill, South Island. The crew expected to meet the ship at 7:50 am the next day.

The Aircraft

The operator dispatched a Kawasaki BK 117 C-1, a twin-engine helicopter whose pair of 692–shaft hp Turbomeca Arriel 1E2 engines drive a four-bladed rigid main rotor and conventional tail rotor. Its maximum gross weight of 3,350 kg (7,385 lb.) provides a useful load of 1,249 kg (2,754 lb.), and a high-mounted tail boom fitted with dual vertical stabilizers allows room for clamshell rear doors that facilitate loading.

The 1996-model aircraft was imported into New Zealand in May 2016, registered as ZK-IMX, and acquired by the air ambulance operator in July 2017. The helicopter came equipped with a radio altimeter primarily visible from the pilot’s (right) seat and NVG–compatible lighting certified by the Civil Aviation Authority (CAA) of New Zealand in simulated blackout conditions in December 2017. The helicopter’s last recorded total flight time before the accident was 6,559 hours on Mar. 23, 2019.

The Crew

The single pilot was accompanied by a paramedic and a winch operator. His commercial pilot certificate was valid for VFR flight only. His 6,673 hours of experience included 43 hours in the BK 117 C-1 and 135 hours in all types during the preceding 90 days. He held a current First-Class Medical Certificate and had received both his initial NVG training and helicopter underwater egress training (HUET) in 2013.

The pilot had logged 73 hours of night flying; his logbook did not specify NVG time, but he told investigators that “all recent night flight” had been on NVG, including 0.5 hours in the previous week and another 0.4 in the preceding 90 days. Those hours combined included three takeoffs and landings to maintain currency.

The paramedic had logged 198 hours of NVG time since undergoing initial training in 2015, including a revalidation check just 14 days before this flight. The paramedic also completed HUET refresher training in May 2018 and received overwater emergency training covering ditching procedures, aircraft evacuation, and life raft operation. The winch operator had taken the same overwater emergency course, completed refresher HUET training in September 2013, and logged a total of 120 hours of NVG time, including a revalidation check in November 2018. Both were contract employees who flew with the operator on an on-demand basis.

The Flight

The crew intended to leave their base at Te Anau, South Island, in time to reach Enderby Island in daylight, but delays in releasing and delivering the required medical supplies pushed their departure back to 3:43 pm local time. The paramedic took the left-front seat, planning to monitor the flight using NVG. The winch operator was still returning to base by road and was picked up en route to a fuel stop at Invercargill, where the crew also donned their immersion suits and life jackets and checked the NVG. The pilot filed a verbal flight plan with local air traffic control (ATC), anticipating three hours of flight time, and departed VFR at 5:03 pm.

The flight extended beyond the ATC zone and lasted past the end of the service day, but the operator maintained contact via a VHF radio repeater on Stewart Island, then by satellite phone beyond radio range. The pilot made regular position and status reports, and a satellite tracking system provided flight-following information.

At 6:50 pm, with the helicopter still in daylight at 3,000 ft. about 90 nm from its destination, the pilot discussed the crew’s progress with the company’s chief pilot and concluded that it was safe to continue. The clouds of the approaching frontal system were visible on the horizon. At 6:58 pm, the pilot reported that he and the paramedic were using NVG; the winch operator, who was not, recalled that it was “pitch black.”

As they approached their intended landing area, it appeared to be covered by a cloud bank. An area with clear weather extended northwest from Port Ross. The helicopter’s GPS database included alternative landing sites where they could camp overnight, but the pilot proposed flying south past Ewing Island to descend in the clear area, then follow the coastline back to Enderby Island. He turned south at 7:34 pm, passing 1.5 nm east of the landing zone, and set the radio altimeter reference to 1,000 ft. as he turned west and began to descend. He subsequently reset the reference to 500 ft. and then 250 ft. while slowing to 75 kt.

About one-quarter mile south of Ewing Island, the paramedic saw cliffs ahead and alerted the pilot. The pilot flared to slow the helicopter and tried to turn right, but at 7:43 pm, the aircraft struck the ocean’s surface in a shallow descent and slight right bank.

The helicopter flipped over and began to fill with water. The pilot and paramedic escaped underwater, and the paramedic rescued the winch operator, who was knocked unconscious. The sea was “dead calm.” The paramedic tried to retrieve the life raft and the survival supplies packed in the aircraft’s emergency bag but couldn’t find them in the dark before the helicopter sank. The winch operator regained consciousness.

The crew’s immersion suits provided enough buoyancy to enable them to paddle 100 m to shore, where they climbed through a kelp bed onto the rocks. They took shelter in the bush but, without the personal locator beacons and other gear lost in the emergency bag, were unable to attract the attention of a P-3 Orion airplane that passed overhead or vessels they could see offshore.

The Rescue

The operator’s chief pilot saw that the flight-tracking signals had stopped updating and tried to reach the crew by satellite phone. After confirming that the satellite tracking had not malfunctioned, the chief pilot contacted the Rescue Coordination Centre New Zealand (RCCNZ) at 8:08 pm, 25 minutes after the accident. The RCCNZ in turn arranged for a Royal New Zealand Air Force P-3 Orion to initiate a visual search and asked five fishing vessels in the vicinity to assist.

The P-3 took off at 10:51 pm and arrived on the scene the next morning about 1:20 am, but low cloud cover prevented its crew from conducting a low-altitude search, so they dropped flares and attempted an infrared search for warm bodies. The first fishing vessel arrived in the area at 11:23 pm. One of the fishing boats diverted to Bluff with the patient who’d suffered the initial emergency. The other vessels established a search grid that located the helicopter’s left sliding door. Three rescue helicopters launched at 10:15 am the next day and reached the scene in 27 minutes, where they spotted the aircrew’s brightly colored immersion suits and evacuated them to the hospital in Invercargill.

The Investigation

Eighteen days after the accident, a private contractor retrieved the main wreckage from the sea floor. The tail boom had separated and drifted away and was not found. Examination confirmed that the helicopter was functioning normally at the moment of impact. The pattern of damage to the fuselage and main-rotor hub informed the TAIC’s reconstruction of the angle and velocity of impact.

The TAIC report goes into some detail about the technology underlying NVG, which multiplies available light, particularly infrared light not otherwise visible to the human eye, but do not create ambient lighting that doesn’t exist otherwise. On the night of the accident, the moon was below the horizon and celestial illumination was minimal. Limitations on contrast resolution are particularly acute over featureless surfaces such as large bodies of calm water. The available light the night of the accident, as estimated by the UK Meteorological Office on request from the TAIC, was one one-­thousandth that of a full moon.

Interviews with the crew along with GPS tracking data from both the satellite link and an onboard unit provided more detail into the flight’s final moments. As the helicopter descended, the paramedic began providing altitude callouts that were neither expected nor requested by the pilot—instead they were based on the barometric altimeter, which the pilot had not reset for local atmospheric pressure since leaving Invercargill. The difference resulted in readings about 50 ft. higher than given by the radio altimeter, which was not easily seen from the left seat. The pilot set the radio altimeter for progressive descents but did not monitor its alert light as he scanned for surface references; the altimeter did not give audible alarms.

Though their employer’s operating specifications called for a maximum descent of 300 ft. per minute (fpm) on NVG, an initial descent rate of 500 fpm increased to 1,200 fpm as the pilot tried to drop down into what he thought was a clear area. The paramedic, who had no training in aviation instrumentation or phraseology, warned the pilot to check his “speed” rather than “descent rate,” which the pilot interpreted as “airspeed.” The cliffs they saw at the last minute were only 20 m (66 ft.) high, and the pilot recalled being “surprised” that the helicopter was so much lower than expected.

TAIC investigators took pains to note the latitude provided by regulations regarding NVG operations. Pilots and nonpilot crew alike were required to take a ground training course and complete three takeoffs and landings in the preceding 90 days. Pilots had to log five hours of supervised NVG flight for certification, and nonpilots only two hours. The pilot’s last three takeoffs and landings had occurred over the course of two flights that included less than one hour of NVG time, and his last NVG competence check four months earlier had not involved low-altitude flight over water.

The TAIC report concluded that, “With more currency on NVG, the pilot might have questioned [the] impression of a dark area beyond the cloud,” particularly since the crew had seen a layer of low fog during the daylight portion of the flight.

The Takeaway

There’s little question that the crew’s HUET training saved their lives—but the immersion suits didn’t contain the flares, personal locator beacons, or cutaway knives specified by company procedures. (One unidentified crew member did have a flashlight and a pocketknife.)

The stowage of the life raft and emergency gear “go-bag” in the cabin didn’t anticipate the impossibility of retrieving them from an inverted fuselage in the dark before the helicopter sank. Thoughtful placement and perhaps water-activated lights could help assure emergency equipment can reliably be accessed and deployed in the shock of an actual emergency.

The TAIC report details how the flight’s risk profile progressively escalated. The distance from Invercargill to Enderby Island exceeded the helicopter’s standard range, requiring supplemental fuel in an external pod. Past a certain point, returning to the mainland was impossible, necessitating a landing somewhere in the Auckland Islands.

The original plan would have gotten the crew to the destination before dark, but the delayed departure left them relying on NVG to land on uninhabited terrain during the darkest part of the night. The pilot was also nearing his duty-day limitations even without accounting for the additional fatigue of NVG flight, estimated by an advisory circular as 2.3 times that of VFR daytime flight. His lack of an instrument rating increased reliance on visual cues, including depth perception—known to be prone to error using NVG—and probably contributed to lack of attention to the radio altimeter and the consequent failure to stabilize the aircraft at an altitude safely above the terrain.

The outcome demonstrates that the difference between nominal currency and genuine sharpness becomes increasingly crucial as the margin for error diminishes.