A 30-year-old commercial pilot and a passenger left Plant City, Florida, on a Saturday morning in March for a time-building trip to Las Vegas. The plan was to return around the middle of the coming week, depending on weather. The pilot had rented a Diamond DA40 for the trip, a four-seater with a 180 hp engine, a gliderlike wing of nearly 40-foot span and a cross-country speed of around 150 knots. The three-year-old airplane was equipped with a Garmin G1000 glass panel, a system with which the pilot had several hundred hours of experience.
Leaving Plant City at 1100 EST and stopping at Mobile, Alabama, they reached Huntsville, Texas, at 1920 CST. At this point they had been airborne for 7½ hours, and the pilot had put in an additional hour being checked out in the DA40 before their departure. They had originally intended to continue after refueling, but after discussing how tired they both felt, they decided to stay the night at Huntsville.
They were back at the airport early and took off at 0715. At some point in the middle of that afternoon, having apparently stopped to refuel somewhere in southeastern New Mexico, they ate, refueled again and filed two flight plans at Sedona, Arizona. Las Vegas had apparently lost its charm; they were headed instead for Arcata, California, with a stop at Palmdale, near Los Angeles. They took off at 1730 MST, overflew Palmdale and landed at Bakersfield at 1841 PST. At this point they had most likely been airborne for 8½ of the past 15 hours, and had covered around 2,000 nm since leaving Plant City the previous day.
At 1936 PST, they left Bakersfield for Arcata, picking up an IFR clearance once airborne.
About half an hour after takeoff, the pilot called Oakland Flight Watch to inquire about turbulence reports. The specialist responded with a general update on unsettled weather affecting central and northern California. There was icing in clouds, but the freezing level was high enough that it should not pose a threat to the flight. An area of moderate precipitation was just ahead, but rain would become more scattered as the flight advanced northwestward. There were some reports of moderate turbulence. Arcata was reporting nine miles’ visibility, with scattered clouds at 5,000 feet and an overcast layer at 9,500.
As the flight neared Arcata, the pilot switched from Oakland to Seattle Center. After a few minutes the controller, apparently noticing that the Diamond had strayed from its assigned altitude, first gave the pilot the current altimeter setting and then asked his altitude. He replied that he was at 9,300 feet. The controller reminded the pilot that his assigned altitude was 10,000 feet, and asked whether he was “having difficulty.” The pilot replied that he was experiencing lots of updrafts and downdrafts and moderate to severe turbulence, but he was correcting to his assigned altitude.
Shortly afterwards, the controller cleared the flight to descend to 9,000 feet. There followed some discussion about which of Arcata’s several instrument approaches the pilot wished to use; he selected the Runway 14 RNAV/GPS, an overwater approach from the north, just off the coastline. The controller cleared the pilot direct to CULDU, the initial approach fix, which is 12.4 miles northwest of the airport. The controller spelled out the name of the fix phonetically twice before the pilot, saying that it was “really turbulent right now,” asked if he could “get back” to the controller for the information. He was back after 15 seconds, and again the controller spelled out the name and repeated that it was the initial approach fix — in other words, the pilot should not have to hunt for it; it would be displayed in any depiction of the approach.
There was a long silence — 80 seconds — before the pilot returned to ask the controller to verify the ID, which he, the pilot, now misspelled as CUDLU. The controller corrected the error, and the pilot at last said, “OK, I got it this time.”
A few minutes later, the controller cleared the flight down to 8,000, warning the pilot that this was the minimum IFR altitude and that it “would be bad” if he got below it. The controller also advised the pilot that he would be encountering moderate to heavy rain along the approach.
After a series of descent clearances, when the flight was at 6,000 feet, the pilot reported that he was painting an area of heavy precip ahead and that he was deviating to the west. The controller cleared him to descend to 5,000 on a heading of 250. A minute and a half later, he cleared him to descend to 4,000 and proceed direct to CULDU. Shortly after, the controller cleared the flight for the approach, instructing the pilot to cross CULDU at or above 4,000. The pilot acknowledged the clearance, but read it back as “at or below 4,000.” The controller corrected his error.
The airplane proceeded to CULDU, executed the course reversal and was inbound to CULDU at 3,600 — the charted minimum crossing altitude is 3,000 — when the controller told the pilot that radar service was terminated and that he could change to the Arcata traffic advisory frequency.
The controller continued to monitor the flight. He observed that, although it was between CULDU and UYFOR, the final approach fix, crossing altitude 2,100, it had already descended to 1,400 feet. When the airplane continued to descend in spite of not yet having crossed UYFOR, the controller repeatedly attempted to contact the pilot on the Center frequency, without success. The last radar return from the airplane came near UYFOR, 5.8 miles from the runway; it was then 300 feet above the water. The weather observation at Arcata at the time was wind 190 at eight, visibility six, ceiling 1,900 scattered, 2,800 broken, 3,400 overcast.
Beginning about four days later, a dozen fragments of the DA40 washed ashore on beaches north of the airport. Several of them came from the left wing or the left side of the fuselage, suggesting that the left wing might have been the first part of the airplane to hit the water. The rest of the wreckage was not found, nor were the remains of the occupants.
The NTSB determined that the probable cause of the accident was “the pilot’s failure to maintain proper altitude and glidepath while executing a night instrument approach,” and that a contributing factor was “the pilot’s fatigue.”
As is often the case with NTSB probable causes, this one is merely a restatement of what happened: The cause of the accident was that the pilot flew into the water. The fatigue theory, however, has some relation to causality.
As the NTSB notes, between the pilot’s arrival at the Plant City airport on Saturday morning and the accident 42 hours later, he accumulated nearly 23 hours of flying time and was “on duty” — meaning awake and involved in activities somehow or other related to flying — for almost 31 hours. Sunday had been a long day, and if the pilot’s trouble holding altitude and copying the name of a fix was any indication, a draining one.
Fatigue is frequently cited as a factor in accidents. Discussions of its importance, however, usually take the form of recitals of accidents that took place at the end of long, or overlong, duty periods. Since we are not given comparable data for long duty periods that did not end in accidents, or for comparable accidents that did not come at the end of long duty periods, we cannot logically judge how powerful or pervasive the influence of fatigue in accidents really is.
In this case, fortunately, we do not have to resort to the hard-to-prove hypothesis of fatigue to understand what may have happened. Take away the fatigue, and the accident scenario remains a familiar one: A pilot making a visual approach at night over unlighted terrain or water hits the surface a mile or two short of the airport. What is involved is a well-documented illusion that convinces pilots that they are higher up than they really are.
It is significant that the ceiling at Arcata was such that by the time the flight was approaching the final approach fix, it was below the clouds. The runway lights were probably in sight. As his ever-decreasing altitude suggests, the pilot was almost certainly no longer flying the instrument approach. Very possibly, he was instead relying on his eyes, and — perhaps with the help of physical and mental fatigue — they were deceiving him.
This article is based on the NTSB’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.
_Read more information and analysis on other aircraft accidents here.
_
