The pilot, 58, was a hotel owner in the northwest England town of Chester. Although he did not fly professionally, he held commercial licenses for both fixed-wing aircraft and helicopters. In 2002 he had replaced his first airplane, a 160 hp fixed-gear Socata Tampico, with a 1970 Cessna 310, which he flew for 300 hours before grounding it for a thorough refurbishment that ended up taking five years. The 310 returned to service in August 2013, at which time the pilot renewed his multiengine rating.
A few days later, he flew in the 310 to Seo de Urgel in eastern Spain, near the tiny principality of Andorra, where he had a house. After a stay of several weeks during which he made a number of flights between airfields in Spain and France, he flew from Seo de Urgel to Lognes, 10 miles east of Paris. A week later, he and a companion set off on the return flight to Hawarden Airport, just outside Chester.
The flight was uneventful until its final seconds, when, on final approach to Runway 22 in fine weather, the 310 drifted to the left of the runway. Its wings rocked, and there was a sudden increase in engine sound before it pitched up, rolled and plunged to the ground. The pilot died at the scene; his companion, 42, was taken to a hospital, but died shortly after of her injuries.
Two photographs taken by a casual plane spotter confirmed witness accounts of the airplane’s final gyration. In one, it is in a very steep left bank, about 60 feet above the ground. In the other, taken a few hundredths of a second before impact, it is 10 feet above the ground in an inverted, nearly vertical dive.
There was no fire, and so investigators were able to glean a good deal of information from the wreckage. Damage to the blades of the right propeller, and the fact that it had been twisted off the crankshaft, indicated that its engine had been developing power on impact; the blades of the left propeller, bent backward but not in the plane of rotation, indicated that the left engine had not.
Both main fuel tanks — in the 310, the tip tanks, with a capacity of 51 gallons each, are the mains and are used for takeoff and landing — were nearly empty: Half a gallon remained in the left, about two gallons in the right. In addition to its main tanks, the 310 had auxiliary 20.5-gallon wing tanks; each contained eight gallons of fuel. The injector distributor on the left engine was half empty. Since the outlets to the injectors are halfway up the sides of the distributor body, half empty is functionally equivalent to empty.
It seemed that with remarkably unlucky timing the left engine had run out of fuel on short final. The pilot may have attributed the resulting leftward drift to a gust of wind. He added power and raised the nose. It was a reaction that was reasonable for an approach that went unstable at the last moment, but not for a twin flying below its minimum single-engine control speed.
The examiner who renewed the pilot’s multiengine qualification described his flying, including his engine-out work, as competent. He was said by his pilot father and by various acquaintances to have been meticulous, and particularly attentive to fuel planning. He always checked fuel levels with a dipstick before flight and “would have been aware of the exact quantity of fuel required for a flight.” He was also concerned about fuel cost and carefully adjusted his power settings for best economy, and he stayed abreast of fuel prices at different airfields. In fact, just before the accident flight he had phoned ahead to Hawarden to compare the fuel price there with that at his departure airport of Lognes. Avgas was about 53 cents a gallon less at Hawarden.
Investigators examined the fuel system and found no faults. The quantity gauges and fuel flow indicators operated normally and were correctly calibrated. There was no sign of leakage. They then reconstructed the likely fuel purchases and use for the accident flight and the one preceding it, and compared them with the actual groundspeeds and times made good as recorded on the pilot’s tablet computer.
His assumptions about wind for this flight are unknown. The actual headwinds over France were stronger than forecast, and, for unknown reasons — perhaps just sightseeing — the leg from Lognes to the Channel had lasted 15 minutes longer than it should have, even taking the wind into account. The default groundspeed programmed into the computer would have resulted in a time of two hours 34 minutes for the 411 nm trip; the actual elapsed time was three hours five minutes. At the pilot’s customary fuel burn of 21 gph, he would have used about 10 gallons more than the computer predicted for no-wind conditions. (The calculations in the accident report make no allowance for taxi and takeoff.)
But why did the pilot — especially this pilot, who was generally so interested in fuel — allow his main tanks to run dry without using the fuel in his aux tanks? In its analysis, the Air Accidents Investigation Branch concluded that he had carried the aux tank fuel — about 45 minutes’ worth — untouched from Spain and did not intend to use it on this flight; that he took only as much fuel at Lognes as he expected to use, putting it all into the main tanks, but underestimated the duration of the flight; and that, making mental calculations of fuel and distance remaining, he failed to recognize until late in the flight that his main tank reserves were getting extremely low. “Although he would not have intended or expected to land with such a low fuel state in the main tanks,” the report concludes, “the fine weather conditions of the day and his familiarity with Hawarden may have been factors in his apparent acceptance of the situation.”
The other aspect of the accident to reflect upon is the loss of control.
Hawarden has a single runway 6,700 feet long with 1,000-foot concrete underruns and overruns. The wind at the time of the accident was 280 degrees at 5 knots. The 310 was over the centerline and about 65 feet above the ground when it crossed the threshold of the Runway 22 underrun. Its speed was about 70 knots — slightly below the minimum control speed of 75 knots and well below the recommended minimum approach speed of 90 knots. From the gradually decreasing slope of the altitude profile recorded by the onboard GPS, it is apparent that the pilot had already begun his flare. By any normal calculus, the landing was “assured.”
At this moment the 310 began to drift to the left. Since the main tank fuel gauges must have been indicating empty, the pilot should have been expecting an engine to quit at any moment; but a left drift was consistent with a 60-degree right crosswind, and he may not have immediately recognized its real cause.
If he attributed the drift to a gust of wind and added power to arrest his descent, only the right engine would have responded, lifting the right wing and turning the airplane still more to the left. It was now quite close to the surface, over the grass, still descending and 20 or more degrees off heading. The rapid, instinctive addition of power to arrest the deteriorating situation led to the loss of control.
Very likely, in the calm of a hangar this pilot could have given an accurate account of the relationship between power level and minimum control speed. Precise modulation of the right engine’s power might — perhaps — have permitted him to level the wings, maintain height and move back to the runway with room to land. But surprise and precision are seldom found together.
This article is based on the AAIB’s report of the accident and is intended to bring the issues raised to our readers’ attention. It is not intended to judge or to reach any definitive conclusions about the ability or capacity of any person, living or dead, or any aircraft or accessory.
Get online content like this delivered straight to your inbox by signing up for our free enewsletter.
