Landmark Accidents: Three Strikes

Making the wrong choices knocks the pilot out

April 1, 2007

When is it reasonable to decide that the weather just isn't cooperating and it's time to rethink your entire approach (pun intended)? Weather changes constantly and requires both controllers and pilots to keep up.

How much experience should a pilot have before attempting approaches to minimums? What role does air traffic control play in leading a pilot down the primrose path to destruction? The first two questions pilots must answer for themselves, and a U.S. District Court answered the last one. The court's decision may surprise you.

On December 12, 2001, at about 3:35 p.m. Eastern Standard Time, the pilot of a Piper Cherokee Six called the Gainesville Flight Service Station in Florida for a weather briefing and to file an IFR flight plan from Fort Lauderdale Executive Airport to St. Augustine, Florida, with Craig Municipal Airport, Jacksonville, as the alternate. A stationary front stretched across the northern Florida panhandle to just south of the Jacksonville area before continuing off eastward into the southwestern Atlantic Ocean. Low ceilings and visibilities were widespread in the frontal area and to the north of the frontal boundary. Visibilities south of the front were mostly unrestricted.

At 4:50 p.m., N7701J departed Fort Lauderdale Executive Airport IFR to St. Augustine. At 6:09, the pilot called the Miami Flight Service Station to request St. Augustine and Craig airport weather. St. Augustine showed two miles in mist and 200 feet overcast (well below the landing minimums for the VOR 13 approach, which required a ceiling above 450 feet). Craig was hovering slightly above minimums with one and a half miles' visibility and a 200-foot broken ceiling. Temperature and dew point were 19 degrees Celsius.

Strike one — St. Augustine

At 6:35 p.m., as the flight checked in with Jacksonville Approach Control, the pilot inquired again about the Craig weather. The controller replied that he'd had one flight miss the approach and advised the weather as one-half-mile visibility and 100 feet overcast. The pilot responded, "All right, I guess I'd like to go ahead and try the VOR in St. Augustine and see what happens. I'm not very optimistic, though."

The controller confirmed that the pilot had the current St. Augustine weather and cleared the Piper for the VOR 13 approach. Anticipating that Craig wasn't such a great option either, the controller suggested Jacksonville International Airport if the approach at Craig didn't work out. At 6:58 p.m. the pilot advised the next Jacksonville sector that he had missed at St. Augustine and was diverting to Craig, and he asked if anyone was getting in. This controller held out more hope, advising that the last two flights had landed using the ILS.

Strike two — Craig

The 6:55 p.m. Craig weather was essentially unchanged from before: wind 010 degrees at 7 knots with one-half-mile visibility, overcast at 100 feet; temperature and dew point at 20 degrees C; and altimeter setting of 30.18 inches. Minimums for the ILS 32 approach are 200 feet agl and one-half mile. At 7:09 p.m. the flight was cleared for the approach and the pilot contacted Craig Tower around 7:11 p.m. At 7:16 the pilot called Jacksonville on the miss and was provided vectors for an ILS approach to Runway 7 at Jacksonville International Airport. Note that good VFR prevailed about 60 miles to the south.

Strike three — Jacksonville

Behind the scenes there was considerable discussion at Jacksonville Tower as the weather changed from bad to worse. It was so marginal that at 7:14 p.m. ground control mentioned to the tracon data controller on the phone: "These guys are getting lost out here on the taxiways, or maybe it's me that's lost, I don't know." The data controller advised that new weather was coming out shortly, and at 7:16 p.m. the special weather observation was wind 050 degrees at 6 knots, visibility one-half mile in fog, clouds 100 feet broken, 500 feet overcast, temperature and dew point 19 degrees C, altimeter setting 30.20. Visibility was now at minimums, the ceiling was below minimums, and the altimeter setting had increased by three-one hundredths. This was passed to the tower at 7:19:02 p.m. just before the Piper was handed off to the next Jacksonville Approach sector. The newly revised automatic terminal information service (ATIS) "November" was broadcast at 7:24:39 p.m. but there is no evidence that the pilot ever received it.

The flight was handed off to another approach sector at 7:19:50, and the new controller verified that the pilot had the current Jacksonville ATIS information "Mike," which was by now more than an hour old. At 7:56 p.m. the wind was from 070 degrees at 7 knots, visibility one and one-half miles in mist, clouds 200 feet broken, 500 feet overcast, temperature 20 degrees C, dew point 19 degrees C, altimeter 30.17 inches. "Mike" was a "teaser report," with the last reported visibility three times better than what was needed to land but with the ceiling right at minimums. It was a roll of the dice that the approach would be successful.

The pilot contacted the final approach controller at 7:21:08 p.m. At 7:29:06, ATC advised the pilot that he was 10 miles from the "Dinns" locator outer marker, provided a vector to intercept the localizer, and confirmed that he was cleared for the ILS approach.

It's axiomatic that weather changes, and at 7:30 p.m. another special report was issued: wind calm; visibility one-quarter mile in fog; ceiling indefinite, 100 feet; temperature and dew point 68 degrees C; altimeter setting 30.20 inches. It does not appear that this was broadcast to the pilot either. This approach was a "Hail Mary" attempt, but the pilot may have been thinking "Mike" was still current.

At 7:31:26 the flight was handed off to the tower, which, at the time, was operating from a temporary facility while the main control tower building was undergoing renovations. The temporary tower had no audio recording capability for air-to-ground communications. It was a significant oversight, as will be seen later, and there was no additional transcript of what information was or was not provided to the pilot of N7701J.

Two other single-engine airplanes successfully made the approach to Runway 7 ahead of N7701J, seven minutes and about four minutes before the Piper made its approach. Those flights, according to the trial transcript, were IFR instructional flights. The pilots reported breaking out of the clouds about 50 to 100 feet above decision height and seeing the approach lights, but not the runway lights, at minimums. The tower passed these reports on to N7701J as the flight was passing the Dinns final approach fix. The controller did not see the Piper come out of the clouds.

Aftermath

Recorded radar data show that N7701J approached Runway 7 on the localizer course until about 7:39:04 p.m., when it was about two miles from the runway approach end. At this point the flight was at 500 feet and turned slightly to the right. It continued to descend to 300 feet and turned back to the left. It was not a stellar performance for a minimums approach and it appears that the aircraft did not actually reach decision height. At about 7:39:59, when about 2,000 feet from the end of the runway, the flight turned left to the north and began to climb. The Piper completed one 360-degree left turn, climbing to 1,000 feet, and then another 360-degree left turn, descending to 300 feet, where radar contact was lost at 7:41:23 p.m.

The approach controller advised the tower that radar showed the flight was initiating a missed approach. The tower then heard the pilot report he was making a missed approach; this was followed by some crackling sounds on the radio and then the pilot said "something about his instruments malfunctioning." The tower called Jacksonville Approach to determine if it was in contact with the flight, and multiple attempts to contact N7701J met with no success.

Witnesses at the approach end of Runway 7 reported seeing a red navigation light as the airplane passed overhead and hearing the engine noise increase as the airplane approached the end of Runway 7. The airplane then "appeared to start climbing and turn hard to the north. The airplane then continued to make several circles and descend, followed by the sound of the airplane crashing through trees." The pilot and three passengers were fatally injured. The main wreckage of N7701J was located in a wooded area with 75-foot pine trees about one mile north-northwest of the approach end of Runway 7.

The pilot of another airplane, which was taxiing to Runway 7 for takeoff at the time of the accident, reported that the ground and local control tower ATC positions were combined and he could hear the pilots of airplanes on approach talking to the local controller. He heard N7701J report a missed approach. "The local controller did not respond immediately and then asked N7701J what his position was. About five seconds later the pilot transmitted something like his instruments were 'haywire or goofy,' with stress in his voice."

Two subsequent aircraft on approach were sent around to clear the airspace but landed a few minutes later with weather reported at 300 feet overcast, visibility three-quarters of a mile.

Pilot and aircraft

The pilot had received his private certificate more than 20 years previous to the accident, but had earned his instrument rating only 10 months earlier. His logbooks and other information showed about 965 total flight hours, including 442 flight hours in the Piper Cherokee Six, 17 hours of actual instrument flight time, 24 hours of simulated instrument flight time, and 225 flight hours at night. It could not be determined if the pilot met the FAA recency-of-experience requirements for instrument flight. The lack of instrument experience is noteworthy despite some actual instrument time.

Lessons learned

When temps and dew points are close or the weather is otherwise questionable, ask for the latest report, especially if your previous information is more than half an hour old. In low conditions it may go above or below minimums constantly. Current pireps are extremely valuable — give them and ask for them.

The fact that other flights made it in is a powerful inducement to continue but there are multiple factors to consider:

  • The weather may have changed since their approaches (see above).
  • They may have better equipment — a horizontal situation indicator or flight director, for example.
  • They may have cheated and have a higher risk tolerance.
  • Hard as this one may be to swallow — they may be more skilled.

After multiple attempts at different airports, especially if the airports are in close proximity, the effects of fatigue and stress are not to be ignored, and it may be time to exit the entire situation for a gold-plated alternate — you do have one, don't you?

Most of us don't practice approaches to minimums very often and practice missed approaches in actual conditions even more rarely. Get some training from an experienced instructor and raise your personal minimums to a level appropriate to your level of experience and proficiency. Consider following the professional's rule of not starting an approach if the weather is reported below minimums — just go to the alternate.

Use the autopilot, again like the pros — it will fly a consistently better approach than the human pilot.

Don't fly with a cold or if you are taking cold medication, even if it is a nonprescription drug. The effects on equilibrium can be quite powerful. According to AOPA's Director of Medical Certification, Gary Crump, "Any of the over-the-counter drugs found in common upper-respiratory-tract-infection treatments, particularly pseudoephedrine, could contribute to physical/mental impairment that could affect safety of flight. There are many variables in the equation, including the duration of use of the medication, the dosage, the frequency of use, the person's morphology (body type: obese, slim, fit, unfit), any underlying chronic/acute medical conditions, especially those for which the pilot is self-medicating, and the variety of different drugs being taken together (the cocktail).

"For a severe head cold, the big concern in addition to the side-effect potential of the drugs taken is the symptoms themselves that could, alone, cause spatial disorientation and/or subtle, if not sudden, incapacitation. Barotrauma, or ear block, could affect the sensory receptors in the inner ear that help us stay straight and level, so if those sensors are upset by inflammation and/or fluid buildup, VFR flight, let alone flight on instruments, could be risky at best." — BL

The autopsy on the pilot showed he had ingested significant amounts of ephedrine/pseudoephedrine and acetaminophen. There were several over-the-counter cold remedies and some caffeine pills in his briefcase, implying that the pilot was suffering from an upper respiratory infection. Could this have been a factor? (See " Lessons Learned," right.)

N7701J, a Piper PA-32-260, was manufactured in October 1968 and had accumulated about 7,850 total flight hours. The airplane and engine were last inspected on November 21, 2001, about 35 flight hours before the accident. The engine was last overhauled on August 1, 1997, with 1,267 flight hours before the accident. The engine, flight instruments, and the S-Tec autopilot all appeared to be functioning normally at the time of the accident.

Examination of the altimeter showed that it had received impact damage and all three hands had come loose on the face. The barometric setting was 30.20 inches, the "approximate altimeter setting," according to the NTSB, at the time of the accident. This was to become a major source of contention.

NTSB probable cause and the court

The NTSB determined the probable cause as "the pilot becoming spatially disoriented and losing control of the airplane during a missed approach...." That seems pretty cut-and-dried until we collide with the convoluted logic of the U.S. District Court of middle Florida. The pilot's family members sued the FAA for failing to exercise due care in the conduct of its duties by not providing the pilot of N7701J with current weather and altimeter information. This, according to the family's theory, caused the pilot to become disoriented and lose control of the aircraft on the missed approach.

The trial judge held, "All parties agree that the crash occurred because the pilot became spatially disoriented, causing him to lose the ability to control his airplane. I conclude that the plaintiffs have proven by a preponderance of the evidence that FAA air traffic controllers failed to give...the pilot...the current weather information on that night which would have alerted him that weather conditions were rapidly deteriorating and that this failure contributed to the pilot's spatial disorientation.

"However, I further find that the pilot himself also contributed to creating his spatial disorientation by forgoing the other options available to him and attempting instead to make his third instrument approach landing of the flight [after two missed approaches] when he was fatigued, ill, and on medication.

"Applying Florida comparative negligence principles, I hold that the FAA's negligence was the legal cause of 65 percent of the accident and that pilot...negligence was the legal cause of 35 percent of the accident."

The judge went on to cite plaintiff expert witness testimony that, given the outdated weather information on ATIS "Mike," "the pilot...would have reasonably begun to look for external reference points at an altitude of approximately 500 feet. At that elevation, he could expect one-eighth to three-eighths of the cloud cover to be breaking, giving him obvious reference to the ground.

As the pilot descended farther to 300 feet or less, he likely continued searching for visual references outside, which he could not see because of the deteriorating weather conditions that were known to air traffic controllers at the time. The medical and aeronautical experts agreed that by repeatedly turning his focus from his instruments to the outside environment and back, [he] likely began to experience spatial disorientation."

The lack of a current altimeter setting also was cited as a problem. ATIS "Mike" gave a setting of 30.17, and at the time of the accident the setting was 30.20, which equates to 30 feet. Since the pressure was rising, although not by very much, the aircraft would have been approximately 30 feet higher than indicated — or on the safe side.

Granted, ATC may have been remiss in not providing current ATIS weather but there was no way to conclusively prove that, given the absence of recording equipment. The tower controller did provide two timely pireps that should have prepared the pilot for exactly what he was getting into. Somehow that was overlooked in the judge's reasoning, and the minor change in altimeter setting was completely irrelevant. The judge's comment that "he could expect one-eighth to three-eighths of the cloud cover to be breaking, giving him obvious [my emphasis] reference to the ground" is breathtakingly ignorant of the realities of low instrument approaches. The fact that flights immediately before and after the accident were able to land shows that although the weather was at minimums, with the right skills and equipment, it was flyable.

What this judge failed to understand, and in my opinion added to some extremely bad case law, is that pilots, not air traffic controllers, are the final authority as to the safe operation of the aircraft. Landing expectancy out of an instrument approach is a dangerous mindset and is never an excuse for losing control. Following the judge's "reasoning," anytime pilots miss an approach because the weather is somewhat worse than what was reported, it's somebody else's fault if the pilot loses it.

ATC is there to assist you, not the other way around, but as pilot in command you make the decision on how much risk to take. The weather is what you experience from the left front seat, not what anyone has told you it was, is, or might be. Hope for the best, ask lots of questions, and be prepared for the worst. By the way, the government settled the case for something more than $9 million and is unlikely to appeal.

Bruce Landsberg is executive director of the AOPA Air Safety Foundation.