March 7, 2007
The weather is what we all worry about when we are planning an IFR cross-country flight. There are many decisions to be made before and after takeoff, and the wrong decision anywhere along the way can lead to disaster.
Some pilots, when they reach the instrument approach phase of their trip, are tempted to bust minimums in an attempt to get the airplane on the ground. It may be that they don't want to disappoint their passengers by having to keep them in the air while finding another airport. Or, perhaps they are in a hurry to get the flight over with if they are tired or not quite up to flying in the conditions they find at the destination airport.
KCRE to KOXC
During April 2002 the pilot of a Piper Seneca called the Anderson Automated Flight Service Station in Anderson, S.C. He asked for and received a standard briefing for a flight from North Myrtle Beach, S.C., (KCRE) to his home base, the Waterbury-Oxford Airport (KOXC) at Oxford, Conn. The accident report does not detail the weather briefing, but it appears that the weather at KOXC that evening was low IFR.
The flight from North Myrtle Beach, which began at 6:07 p.m. Eastern time, was uneventful. The pilot communicated with ATC, and with the exception that his Mode C was not working when he was switched over to the New York Terminal Radar Approach Control (TRACON), nothing seemed unusual or offered the indication that the flight would end in anything other than a normal approach and landing.
There were more than normal transmissions between the controllers and the pilot while the Seneca descended into the New York airspace because of the failure of the aircraft's encoding altimeter. Several times the controllers asked the pilot for his altitude, but at no time did it appear that the aircraft descended below an assigned altitude. Nor did it appear that the pilot did not have complete control of his aircraft while in flight.
At first, the pilot was told to expect radar vectors to the Bridgeport VOR, but shortly afterward he was given a 350 degree heading to fly. The controller then told him to turn 20 degrees to the left because the winds "were blowing pretty strong out of the west and dragging you more east than I expected." He also told the pilot to intercept the localizer at KOXC.
At 9:03 p.m. the aircraft was flying level at 4,000 feet when the pilot was instructed to change frequencies. The new controller told the Seneca pilot to let him know when he had the KOXC weather and he asked what approach the pilot intended to fly. The pilot replied that he wanted the ILS 36 approach and the controller confirmed the pilot was on the localizer.
A couple of minutes later, the controller instructed the pilot to descend and maintain 3,000 feet and he then cleared the aircraft for the ILS 36 Approach. The pilot acknowledged the transmission.
At 9:07 p.m. the pilot reported that he was at 3,000 feet, and the controller responded, "Seneca 284, thank you. You are 15 miles from CLERA [the outer marker] and you're cleared the ILS 36 approach."
Seven minutes later the controller said, "N284, you're about two miles from CLERA. Radar service terminated. Change to advisory frequency. Cancel on 135.1 in the air or on the ground."
The pilot acknowledged the transmission and it was the last one received from the aircraft.
|Waterbury-Oxford (KOXC) ILS Rwy 36 chart. Click here for larger version.
The aircraft crashed into trees and terrain at an elevation of 650 feet MSL about 2,000 feet southeast of the approach end of runway 36. All aircraft components were recovered although the altimeter's needles were missing. The altimeter was set at 30.37.
The nose wheel actuator was found in the extended position. The right main gear and nose gear were separated from the aircraft. Investigators did not indicate that they found any mechanical discrepancies that would have caused the accident.
There were several witnesses to the accident. One was standing in his driveway facing the airport when he heard the aircraft approach. He said that it sounded like it was "off course." When he looked up at the airplane it flew over his house at about 40 to 50 feet above the trees. He said it "continued to fly normally at the same altitude as if to land." He told investigators that the aircraft's landing lights then illuminated the tops of the trees before it crashed. He said that he built engines, and that he thought the Seneca's engines sounded like they were running normally.
Another witness was inside his home when he heard the aircraft approach. He said the engine RPM sounded higher than normal. Then, all of a sudden, it dropped to a real low RPM for a few seconds. The RPM then increased "drastically" for three to four seconds, before it dropped slightly.
A third witness was also inside her home, which was located directly across the street from the accident site. She said the airplane flew very low over her home and sounded like a "motorcycle going full throttle overhead at a steady constant rate." Almost instantly, she heard a loud boom and saw an orange glow in the yard.
The recorded weather at KOXC at 8:55 p.m. was reported as winds from 190 degrees at 8 knots, visibility 2.5 miles, ceiling at 300 feet overcast, and the temperature/dew point at 12 C and 11 C respectively. The altimeter was 30.37.
The private pilot held ratings for single and multi-engine land aircraft, and an instrument rating. His logbook indicated that he had a total of 897.5 flying hours. Of that total, 341 hours were flown in multi-engine aircraft. He had 119.7 hours of night flying and 45.2 hours of actual instrument time. In the six months preceding the accident the pilot wrote in the remarks section of his logbook that he had flown eight ILS approaches to Runway 36 at OXC. Separately, the pilot recorded that he had flown two flights 13 days before the accident flight. The note indicated that on the first flight, which was one hour long, he had done "turns, ILS 36 circle to land Runway 18, 3 landings." The second flight consisted of "turns, 3 landings day, 1 ILS 36." It was 1.7 hours in length.
Runway 36 at KOXC is 5,800 feet long of which 5,000 feet is available for landing. It's equipped with a four-light precision approach path indicator (PAPI) system, high-intensity runway lights (HIRL) and runway end identifier lights (REIL). The published inbound course for the ILS is 005 degrees. The decision altitude is 971 feet MSL, which is 250 feet AGL. The airport elevation is 726 feet.
The NTSB requested radar data concerning the flight from the FAA but due to a technical malfunction with the recording equipment no data was saved. However, investigators did recover the pilot's personal Garmin GPSMAP 295 in the wreckage and sent it back to the manufacturer for examination. Technicians were able to collect data from the unit and investigators were able to study the final minutes of the fateful flight.
The data revealed that when the aircraft was approximately 1-3/4 miles south of the clera it was at an altitude of 3,008 feet MSL. It then turned left, passing abeam clera about one mile left of course at an altitude of 2,837 feet, which is 337 feet above the altitude the aircraft should have been at had it been on the glide slope at that point. It should be pointed out that this altitude data comes solely from the handheld GPS, so there's no way to verify the accuracy of the altitude data.
At that point the aircraft turned back to the right and appeared to intercept the localizer, but then it turned left again and flew about three quarters of a mile left of the localizer course before turning back to the right again. The airplane crossed over the localizer course at a reported altitude of 1,271 feet MSL just south of the middle marker, then it began a left turn back toward the airport just before the data ended. The last altitude reported was 779 feet MSL at 9:18 p.m.
Perception v. Reality?
So, how did the aircraft wind up crashing at an elevation that was below that of the airport 2,000 feet southeast of the runway? The NTSB blamed the accident on the pilot's failure to follow the published instrument approach procedure. Contributing factors were the darkness and the weather conditions.
It's obvious that that's what happened, but it does not tell us why the accident occurred. If you refer to the GPS data you can see that the pilot was hardly on the localizer from the time the data begins.
The pilot was told that there was a strong wind from the west at 4,000 feet that was pushing the aircraft to the east during radar vectors to the localizer. Is it possible that the same was true at lower altitudes as well? Did the pilot, once he lined up on the localizer, see that the wind was blowing him east and that he overcorrected by making too large a cut to the west? But, if he did that then why didn't he reduce the amount of correction once the aircraft was back on the localizer and stay closer to it?
Perhaps the pilot was not as proficient at flying the ILS as he thought he was. His logbook indicated that he had flown a number of ILS approaches to Runway 36 over the preceding six months, but how were they flown? The NTSB report does not indicate whether he flew those approaches under the hood with a safety pilot, and there was no sign off by a flight instructor reported.
The same is true for the two flights he made two weeks before the accident occurred. Did the pilot simulate instrument conditions, or did he simply fly the approaches in VFR conditions by himself? There is a big difference in flying an ILS approach when you are looking out the windshield, and when you are confronted with weather conditions that require the ability to fly solely by the gauges.
It seems that accidents that occur on ILS approaches involve the inability of the pilot to fly both the localizer and glide slope at the same time. Do you remember the first time you flew an ILS under the hood? You were probably trying to get the localizer needle pegged and neglected the glide slope for a period. Then, you came back to the glide slope, but it was so difficult to capture that it wasn't long before you lost the localizer again.
While the only altitude information we have on the accident flight as it made its way along the approach is from the GPS, which shows a continuous descent, it's entirely possible that the pilot was not proficient enough to be flying in the weather conditions he encountered that evening and that his performance at maintaining the glide slope was as poor as the localizer track that was recovered from the GPS.
What have we learned from reviewing this accident? First, always make sure you are current and proficient for any flight you are going to make, whether it is in VFR or IFR conditions. If you go out to do instrument approaches for proficiency take a safety pilot and simulate them properly, or get an instrument flight instructor to fly with you and do it in actual conditions.
If you are an inexperienced instrument pilot or one who does not fly IFR regularly, your proficiency is suspect. Make sure you have established personal limitations that recognize your abilities, so you won't be out in conditions that you should not be flying in. And don't violate those personal limitations.
More accident analyses are available in AVweb's Probable Cause Index. And for monthly articles about IFR flying including accident reports like this one, subscribe to AVweb's sister publication, IFR Refresher.