Crash During Landing Highlights Need for Improved Training

If they aren't doing so already, air carriers need to incorporate bounced landing recovery techniques into their initial and recurrent training.

This is the recommendation of the National Transportation Safety Board (NTSB) following its investigation of the crash of Executive Airlines, doing business as American Eagle, at San Juan, Puerto Rico's Luis Munoz Marin International Airport (SJU) on May 9, 2004.

In addition to flying technique, a number of problems surfaced in the course of the investigation, notably with the flight data recorder (FDR).

The airplane, an ATR-72, bounced twice before control was lost. "The captain added power, he was probably going to go around, but he bounced on the nose gear and got behind the airplane," said Capt. Dave Kirchgessner, an NTSB investigator.

The circumstances that led up to the crash, though, are more nuanced than Kirchgessner's declarative would suggest, and which he annunciated during his briefing on the case.

The first officer was the pilot flying (PF) for this flight. The first officer was new to the aircraft, with barely 20 hours of time flying it. The accident flight was his first after completing company training. He was paired with a captain who had 3,800 hours, 1,120 as an ATR captain.

The first officer had bounced the aircraft before, in a landing at San Juan during his initial operating experience (IOE). A bounce is when the airplane makes contact with the runway, but becomes airborne again.

Accordingly, the captain was actively coaching the first officer, urging him to get the nose down. "Now start [walking] it back," the captain was heard to say on the CVR.

However, the aircraft became airborne after the first touchdown, and the captain took control. The Air Line Pilots Association's (ALPA) submission to the docket on this investigation says, "Bouncing on landing an ATR-72, especially for a pilot new to the aircraft, is common."

Despite having seized the controls, the captain's inputs made the situation worse. One second after the first officer's bounce, power was reduced, indicating an attempt by the captain to land the airplane.

However, the airplane bounced a second and third time before contacting the ground a fourth time in a crash. The left main landing gear broke from the aircraft on the third bounce, after which the aircraft was partially or completely uncontrollable and, in the blunt phrasing of the ALPA analysis, "the crew was essentially 'along for the ride' until the aircraft came to rest."

Although the first officer was trained in an ATR-42 simulator (the ATR-42 is a smaller version of the ATR-72 but they enjoy a common type rating), he flared at the right time and crossed the threshold within limits. His training in an ATR-42 simulator was not considered a factor in the accident, although pilots new to the ATR-72 had a greater propensity to bounce on landing.

The NTSB did not consider the differences between the ATR-42 and ATR-72 sufficient to question the different performance of the two aircraft. What the NTSB did question was the company's lack of guidance on a bounced landing (i.e., bounced landing recovery guidance). Neither formal guidance was contained in manuals nor was bounced landing recovery taught in the simulator.

The lack of training may have contributed to the captain's failure to apply proper pitch and power corrections after he took control, and his failure to execute a go-around, the NTSB concluded. Following the accident, Pinnacle Airlines [PNCL] did include bounced landing recovery guidance in its airplane operating manuals.

The NTSB concluded, "Written company guidance on bounced landing recovery techniques would have increased the possibility that the captain could have recovered from the bounced landings or handled the airplane more appropriately by executing a go-around."

The NTSB went a step further and said training in recovery techniques, during both initial and recurrent training, would be useful.

The NTSB rejected ALPA's call for training in transfer of control when the aircraft is "extremely close to the ground." The NTSB argued that transfer of control was not the problem, but the captain's inappropriate control techniques after taking control.

As a result of the crash, the captain had one of the vertebra in his back fractured when his seat collapsed. His seat collapsed when it was subjected to vertical loads that exceeded those required for certification. However, the NTSB considers the certification standards adequate and did not recommend retrofit of the cockpit seats. The first officer, two flight attendants and 16 of 22 passengers received minor injuries.

Of greater concern to the NTSB than cockpit seats was the status of aileron position sensors that provide input to the FDR. The NTSB found that the left aileron surface position data on the accident aircraft was invalid. Cassandra Johnson, an NTSB recorder specialist, noted a 46 percent failure rate of the aileron position sensors was experienced on this aircraft type in the past 31/2 years, which is a high failure rate (to say the least). The NTSB urged the replacement of these sensors with a more reliable unit, and to review all flight recorder systems that have been modified by a supplemental type certificate (STC) to make sure they are reliable, and are replaced if not.

ALPA went a step further, urging replacement of the L-3 Communications [LLL] Model F-800 flight data recorder with a more reliable unit (this recorder's reliability has come up before; see ASW, Aug. 5, and Jan. 19, 2004).