Aftermath: Imprecision Approach

Although the term “precision approach” merely means that vertical guidance is included, it seems to imply that the approach ought to be executed with precision by both pilots and controllers.

On a December evening in 2013, a Cessna 310 carrying a pilot, 60, and his two daughters, 17 and 20, crashed while executing a missed approach at Jacksonville Executive (CRG) in Florida. The pilot filed an instrument flight plan, but declined the specialist’s offer of a weather briefing because, as he said, the weather “looked good.” At the time, CRG was reporting a 400-foot overcast and 2 miles visibility in mist.

Darkness was almost complete when the 310 approached Jacksonville. The approach controller gave the pilot the freshly updated ATIS calling for calm winds, 400-foot ceiling and 1.5-mile visibility, and cleared him to descend to 2,000 feet. Crossing altitude at the final approach fix (FAF) is 1,900 feet. When the 310 was 7 miles from the FAF, the controller instructed the pilot to turn to a heading of 350 degrees and to maintain 2,000 feet until established on the localizer, and cleared him for the ILS approach. A minute later, he handed the flight off to CRG Tower.

The 310 intercepted the localizer about 12 nm from the airport, turning inbound a little to the left of the approach course and slowly correcting toward it. By the time he reached the FAF 5.8 nm from the runway, however, he had crossed the localizer and drifted to the right of center by more than three-quarters of a mile, or three times the half-width of the localizer course. He then corrected to the left, regained the centerline about 2 miles from the runway, overshot widely, and then corrected again.

His altitude control was no better. Although the clearance called for remaining at 2,000 feet until established, he was at 1,700 when he first crossed the localizer about 3 miles outside the FAF. He passed abeam the FAF 900 feet low, or nearly three times the glideslope half-height, and continued to descend until he triggered a low-altitude alert in the tower. The local controller told the pilot to check his altitude. He calmly acknowledged that he was at 600 feet, corrected briefly upward, and continued inbound as his altitude fluctuated randomly between 600 and 900 feet.

Despite the unpromising appearance of the approach, the 310 — somewhat like the broken clock that is right twice a day — was on the localizer and on the glideslope when, about a mile short of the runway, the pilot declared a missed approach.

The published procedure for the miss is a straight-ahead climb to 700 feet (the airport elevation is 42 feet) followed by a right turn to 180 degrees while continuing to climb to 1,900 feet. In fact, however, controllers often ordered a left turn to 280, and that is what the tower controller, on instructions from Approach, did. The tower controller did not specify an altitude.

From its position well short of the missed approach point, the 310 began a climbing left turn. It turned through 280 to a heading of 180 or so, and then, from an altitude of 900 feet, dropped almost vertically into a pond in a housing development about a mile south of the airport. All three aboard perished.

The National Transportation Safety Board attributed the accident to loss of control due to spatial disorientation and lack of instrument proficiency. There were no indications of mechanical or instrumentation problems, or of pilot incapacitation.

The pilot had over 1,600 hours. He had had an instrument rating since 2002 and had gained his multiengine rating in 2007.

The airplane, a well-maintained R model with an EFIS panel, certainly had an autopilot with, at the very least, the ability to track a localizer. Either it was not available for some reason or the pilot chose not to use it. That the approach was unstabilized cannot have escaped him; perhaps that is why he decided to abort even before reaching the missed approach point.

The pilot may have been confused by the instruction to turn left to 280 degrees. For one thing, it was not what he expected, assuming that he studied the approach plate before beginning the approach. But how, exactly, was he to interpret the instruction? Was he to turn immediately, as he did? Turn after crossing the MAP? Climb to 700 feet straight ahead and then turn? Climb to 1,900 feet and then turn?

The last seems the least likely option, and yet the NTSB report, after stating that there was no requirement for the controller to provide an altitude with the new clearance, says that “the pilot would be expected to climb to the altitude of the published missed approach and then turn to the alternate heading that was provided by the controller.” Contrarily, an interview record in the accident docket refers to “the requirement for an alternate missed approach to include an altitude.”

In interviews 10 days after the accident, the Jacksonville Tracon supervisor told accident investigators that he found it “shocking” that no altitude had been included with the missed approach clearance; but since the pilot did turn and did climb, the lack of a specified altitude seems not to have been a factor in the loss of control unless it distracted the pilot by obliging him to refer anew to the approach plate. The 280 heading was customarily accompanied by an altitude of 3,000 feet; the approach controller said that he had assumed the tower controller would supply it, but the tower controller said that he did not supply it because it was his practice to pass on just what the approach controller said to him.

Neither the tower controller nor the approach controller had observed the 310’s difficulty in capturing the localizer and glideslope. The approach controller said if he had, and if he had been in communication with the aircraft, he would have asked the pilot whether he was still receiving the localizer — a polite hint — and, if he still failed to intercept, he would have canceled the approach and sent the pilot around for another try. The tower controller, who was able to follow the approach on a radar repeater, said that, apart from the single low altitude alert, he did not recall anything “remarkable” about it. He was not concerned about the out-of-tolerance altitudes, he said, because the approach course is free of obstructions.

Fog was general that night, and precision, it seems, had taken the evening off.