The old man and the sea

image: Conquest Air fleet of Convair 340 | Nathan Coats

The experienced captain made an inexplicable decision to embark on a long overwater flight with a sick engine.

About lunchtime, the young first officer found himself in a horrible predicament.

Through the pain of a head wound from the impact of the glareshield, he saw and felt the flight deck filling with seawater. The captain, a man old enough to be his father, was unconscious and not responding to shouts of his name or shaking of his arm. The first officer could not lift him from his seat.

As the water rose to chest level he realised he had to leave. He kicked open the cockpit door and saw daylight at the far end of the fuselage—the tail had detached. Taking a life raft under his arm, he swam towards the light.

The second aircraft he flew in that day was a US Coast Guard rescue helicopter that arrived after a few minutes. The captain was never seen again. It was 8 February 2019.

Freight run

The working day had begun shortly after 9am with a routine flight from Miami-Opa Locka Airport in Florida to Nassau, in the Bahamas, an independent country about 160 nm east of the US coast. The aircraft was a Convair CV-340, a pressurised twin piston-engine design from the 1940s.

Convair’s 240-340-440 series was one of many aircraft designs that attempted something eventually abandoned as impossible—replace the Douglas DC-3. Nonetheless, the CV-340 found a long-lasting niche as an airliner and later as a freight and charter aircraft. Ansett operated several CV-440s in Australia, for example.

Like all aircraft of its era, the type had a series of accidents, many of them appallingly predictable and avoidable by modern standards.

Two notorious losses were the 1977 fuel exhaustion crash that killed three members of the rock group Lynyrd Skynyrd and the 1989 in-flight breakup of a CV-580 that was traced to counterfeit bolts used in the tail assembly.

The CV-340 being flown as a cargo flight by Conquest Air that morning had started its career as a military variant, the C-131 Samaritan, and had logged a substantial but not outrageous 12,701 hours over its 65-year operating life. The engines, notably, were well within their time between overhaul limits and the aircraft had been inspected the day before.

The captain, aged 68, had had a long and varied aviation career of 23,000 hours. He held an air transport pilot licence with type ratings for the Boeing 727, 737, Learjet and CV‑340, on which he had flown about 725 hours.

Why he should spend the later years of his career flying an unvarying over‑water route in an aircraft almost as old as himself might seem curious, but it offered the undeniable advantage of a daily return to his own bed, rather than multi-day schedules involving early starts, late finishes and dreary airport hotels. It was also a considerably more stable lifestyle than the corporate pilot’s fast-to-sour reality of living at the beck and call of some capricious billionaire, or their offspring. And perhaps there was also the not‑to‑be-discounted attraction of stick and rudder flying in a kind of aircraft now disappearing from commercial aviation.

The 28-year-old first officer was getting this sort of flying experience early in his career. He had 650 hours of total time, 305 of them on the crashed aircraft, with ‘second-in-command privileges only.’

The morning’s flight had been normal except for a brief incident when the first officer, who was pilot monitoring, went to pull back the left engine propeller control for cruise flight. The prop remained at 2400 rpm. The first officer tried to reset the propeller control circuit breaker but was unable to do so. The captain stabilised power on both engines, and normality returned for the remainder of the flight to Nassau.

The engine run-ups and take-off were normal on the return flight to Miami, with the first officer as pilot flying. However, as the aircraft climbed through 4000 feet, the left engine propeller control again stopped working. The captain adjusted it, pulled the circuit breaker twice, and ‘inadvertently increased power to 2700 rpm’, in the words of the US National Transportation Safety Board’s (NTSB) report. He took control from the first officer, levelled the aircraft at 4500 feet, cancelled the IFR flight plan and instead set off on the 160 nm return leg to Miami under VFR.

A statement taken for the NTSB’s accident docket describes how the captain ‘decided to continue with prop turning and power low to baby the engine.’ The first officer noted that brake mean effective pressure (BMEP, a measure of engine power) was at the bottom of the scale. According to the statement, the captain said he had dealt with this sort of problem before and it could be fixed at Miami-Opa Locka upon return.

According to the NTSB report: ‘The first officer suggested that they return to the departure airport, but the captain elected to continue. The first officer indicated that he did not want to disagree with the captain’s decision given his extensive experience.’ The first officer then busied himself checking fuel and alternates on his iPad, which told him there was sufficient fuel, no need to transfer between tanks, and an alternate at Bimini, the westernmost Bahamian island.

After passing Bimini, the captain expressed concern over fuel imbalance, that the right wing would be light. He requested the first officer look up fuel transfer in the checklist. The first officer could not remember if the captain touched any fuel valves while he was looking at the checklist.

About the time the first officer determined the quick reference handbook did not have anything, the right engine—not the one that had earlier given propeller trouble—started losing power, backfiring and shaking violently, ‘with variations in the brake mean effective pressure, fuel pressure, fuel flow indications, rpm and manifold pressure.’

The captain called to feather the right engine, which the first officer did. The captain did not want the engine secured in case they might need it later, so he set it to idle cut-off and closed the cowl flaps. The captain slowly advanced the throttle on the left engine and, shortly after, it started surging. The captain tried to keep the left engine going and asked the first officer to try to restart the right engine. Neither engine responded.

The first officer declared an emergency soon before ditching. He recalled checking his four-point harness and putting his seat fully back. He held the yoke with his left hand and braced his right hand on the glareshield. He could not recall if the captain had put on his four-point harness.

The aircraft went into the sea about 32 nm east of Miami-Opa Locka.

The NTSB determined the probable cause of this accident to be, ‘the captain’s decision to continue with the flight with a malfunctioning left engine propeller control and the subsequent loss of engine power on both engines for undetermined reasons, which resulted in ditching into the ocean. Contributing to the accident was the first officer’s failure to challenge the captain’s decision to continue with the flight.’

Deeper questions will go unanswered. There was no attempt to recover the engines and the aircraft did not carry flight recorders.

Cockpit gradient

Why the captain made the decision to fly 160 nm over water with a dubious engine is unfathomable. Even in the era of the Airbus A350’s ETOPS 370 certification (extended twin operations allowing a theoretical time of 370 minutes or seven hours from the nearest airport), the inflexible rule is to land as soon as possible at any sign of engine trouble. It hardly needs saying that the baroque and complicated technology of late 1940s piston engines is a lot less reliable than the turbofans that made ETOPS possible. The operator, Conquest Air, told the NTSB the flight crew should have landed as soon as practical after the first sign of a mechanical issue.

Why did the first officer not challenge the captain’s decision? The answer has to do with a subtle but powerful concept: cockpit gradient. This is the difference in authority and experience between flight crew members—or between flight crew and cabin crew in passenger aircraft. Accident analyses using cockpit gradient are often harrowing tales of overbearing or even bullying captains leading meek subordinates to their deaths, as in the case of Korean Air Cargo flight 8509, where a captain disregarded warnings about a frozen attitude indicator, or Airblue flight 202, where the CVR recorded the captain ignoring the first officer’s call to ‘turn left, pull up sir, sir pull up!’

But cockpit gradient is an insidious hazard because it applies just as much to the affable personality as the objectionable. There are notes of confusion and respect in the first officer’s statement about not questioning the captain, who had 35 hours in the air for every hour the first officer had. ‘I questioned his decision, and suggested we turn back to land at MYNN (Nassau) and await maintenance. The captain again convinced me that what we were doing was not only acceptable, but had been done before,’ he told the NTSB. And the slim evidence we have about the captain’s command style suggests he was affable—his fiancée told a Florida TV news reporter he had spoken highly of the first officer.

Cockpit gradient is an insidious hazard because it applies just as much to the affable personality as the objectionable.

Seventy-five years before this crash, in a different ocean, US admiral Chester Nimitz wrote the Pacific Fleet Confidential Letter after three destroyers were sunk in a typhoon, drowning 790 sailors. It is an involved document, worth reading for sailors and flyers, and one sentence in its conclusion stands out: ‘Safety and fatal hazard are not separated by any sharp boundary line, but shade gradually from one into the other.’