Familiarity can breed confusion when a procedure that you have flown before has been revised and you don't brief for the approach. Even in a single-pilot situation, an approach briefing will ensure you have everything set up properly. Recently in IFR Refresher, Brian Jacobson followed the chain of events that led to an accident during an approach, when the pilot wasn't where he thought he was, and didn't notice the signals warning him.
This article appears in the June 2002 issue of IFR Refresher magazine, and is reprinted here by permission.
The approach briefing is one of the most important elements of an instrument flight. As you will see in the following accident report, a faulty briefing can lead to errors, especially if the pilot is familiar with a particular approach and does not notice changes that have been made to it.
Before we look at the accident though, let's consider the approach briefing for a moment. The term "briefing" tends to give the impression that one person is going to brief another. But a single pilot must brief himself on the approach to be flown, though many pilots don't think of their perusal of the approach plate as such.
Failure to pay close attention to the plate during the self-briefing can lead to a serious deviation from the published approach and may result in an accident.
(Click chart for hi-res version)
The Beech 200 Super King Air departed Las Vegas, Nevada, on a March evening in 1997, bound for Salt Lake City, Utah. The weather was IFR, and the pilot, when handed off to Salt Lake Approach Control, was told to expect the ILS Runway 34R approach at KSLC.
The pilot was vectored to the approach at 15,000 feet, then told to "turn right heading zero one zero. Cross the PLAGE intersection at or above eleven thousand. Cleared ILS runway three four right approach." Immediately after the pilot acknowledged the approach clearance, the controller informed him that the RVR for Runway 34R was out of service and that "the visibility is one-half mile." It appears that the pilot acknowledged this transmission, though the first part of his response was unintelligible.
As the aircraft intercepted the localizer, the approach controller asked the pilot to say his airspeed. The pilot responded, "Indicating one eight zero." The controller then instructed him to "maintain best forward speed." About a minute later, the controller said, "Maintain best forward speed until SCOER, cross KERNN at one seven zero knots, contact tower now, one one niner point five." Again the pilot's response was somewhat unintelligible, but it appears that he acknowledged the transmission.
Recorded radar data indicates that the aircraft crossed PLAGE intersection at 11,800 feet, 800 feet above the minimum crossing altitude specified in the clearance that was issued to the pilot and 1,300 feet above the minimum crossing altitude on the approach plate.
About a minute and a half after passing PLAGE, the aircraft crossed SCOER at 10,500 feet, which is 1,500 feet above the minimum crossing altitude listed on the approach plate.
Two minutes later, the tower controller told the pilot, "King Air One One Seven Whiskey Mike, caution wake turbulence, Boeing seven five seven six miles ahead. Wind three six zero at one five, runway three four right, you are cleared to land."
After SCOER, the approach plate allows for a descent to 7,100 feet to intercept the glideslope. Investigators determined that the intercept should take place about 8.9 miles from the runway and about 6.5 miles beyond SCOER.
Radar data indicated that the aircraft crossed the KERNN outer marker 5.5 miles from the runway at 7,000 feet at a radar-indicated speed of 164 knots. If the aircraft had been on the glideslope when it crossed KERNN, it would have been at 6,095 feet.
Recorded radar indicates that the aircraft remained above the glideslope from KERNN until it was about 1.8 miles from the runway threshold. At that point the aircraft was at 4,900 feet, 478 feet above decision height, and it had a radar-indicated groundspeed of 103 knots.
The aircraft remained on the glideslope for approximately 28 seconds, during which its radar-indicated groundspeed continued to deteriorate. The King Air began a descent well below the glideslope as its radar-indicated groundspeed deteriorated steadily to 70 knots. The loss of 200 feet of altitude (from 4,700 to 4,500 feet) over four seconds, combined with the decreasing airspeed, was computed to correspond to an average downward vertical flight path angle of 20.3 degrees below the horizontal. The last radar hit on the airplane indicated that it was at 4,400 feet, approximately 200 feet above the touchdown zone elevation, and at 71 knots.
The aircraft crashed approximately 1.3 nautical miles short of the Runway 34R threshold and about 1/4 mile left of the localizer centerline. The pilot and two passengers were seriously injured in the accident. A third passenger received fatal injuries.
The pilot was seemingly well-qualified to fly this flight. He was the chief pilot for a hotel chain in Las Vegas, and he flew part-time for the owner of the accident aircraft.
He held an ATP with type ratings in the Cessna 500 series, Mitsubishi 300, and Beech 400 aircraft. He was a CFII with airplane single-engine ratings.
He had 8,172 total hours, 1,841 of which were in the Beech King Air 200. He had received simulator training in the King Air 200, but the last session was 10 years before the accident occurred.
He had 39.8 hours of King Air 200 time in the six months prior to the accident, about one third of which was instrument time. He had flown 61 hours in the 90 days preceding the accident, including six hours in the King Air 200.
The accident occurred at 1913 MST. The METAR weather observation taken at 1851 indicated that the wind was from 340 degrees at 18 knots, the ceiling was indefinite with a vertical visibility of 1,100 feet, and the visibility was 1/2 mile in moderate snow showers.
The RVR for Runway 34R was not in service. At 1927 a special observation was taken, and it was essentially the same as the 1851 observation.
The pilot told investigators that when he flew single-pilot, he used the autopilot for approaches if it was working properly. He did the same on the evening of the accident.
He said that he did not believe that the 170-knot speed restriction to KERNN was "unmanageable." He would prefer to have been at 140 knots at that point. He said that he did not recall any abnormal engine or instrument indications, nor did the aircraft accumulate any ice during the approach. He told investigators that he was "tracking the localizer and glideslope well." He captured the glideslope from above, using a power reduction and a two- to three-degree pitchover to increase the descent rate in order to capture the glideslope.
The flaps were extended to 20 degrees, and the landing gear was extended. He thought that the cause of the accident was wake turbulence from the preceding Boeing 757.
The pilot told investigators that he did not recall hearing any audible warning horns or seeing any warning or caution lights during the approach. Both surviving passengers reported hearing what they thought was a warning horn of some type during the accident sequence.
There were no reports of turbulence or wind shear from other aircraft on the approach. The pilot said that the ride was smooth on final approach and that the weather was "exactly like the ATIS" with heavy snow.
The first indication of trouble, according to the passengers, was when the aircraft suddenly rolled left and struck the ground two to three seconds later. The pilot had no recollection of the accident itself.
When investigators studied the instrument panel, they found that the DME selector switch was in the number-two position. That means the DME information that the pilot saw came from the number-two nav radio. The number-one nav radio was tuned to 109.5, the ILS frequency, and would have displayed localizer and glideslope navigation information on the pilot's HSI. The number-two nav radio was tuned to 116.8, the Salt Lake City VORTAC frequency.
If you study the approach plate for the Salt Lake City Runway 34R ILS approach, you will note that the PLAGE, SCOER, and KERNN intersections are defined using the DME associated with the ILS approach facility, not the SLC VORTAC.
Six months before the accident, a new approach facility was installed for the ILS Runway 34R approach. It included a co-located DME from which the approach fixes were referenced.
Before that, the DME fixes that identified the PLAGE, SCOER, and KERNN intersections were referenced from the SLC VORTAC. The VORTAC is located 4.7 miles northwest of the ILS facility. With the DME selector on the VORTAC, the pilot would have thought that he was farther from the airport than he was, and that would explain why the King Air was well above the glideslope for as long as it was.
The pilot told investigators that he increased his rate of descent, using the autopilot controls, to capture the glideslope from above. At the same time, he reduced power. He said that he did not recall his airspeed when he increased the rate of descent, but he remembered the glideslope "capture" light illuminating when the aircraft did make the interception. He said that after the interception was completed, pitch control was assumed by the autopilot to maintain the glideslope, but he did not recall adjusting the power.
He did not remember seeing any indication that the autopilot lost the glideslope any time after capture, nor did he recall seeing any full deflection indications.
To study the possibility of a wake-turbulence encounter, investigators checked the radar track of the Boeing 757 that flew the ILS course ahead of the King Air. They found that the aircraft did not deviate any more than 93 feet above the glideslope at any point during the approach. The aircraft was computed to be 52 feet above the glideslope at 1.8 nautical miles from the threshold, the point where the King Air intercepted it approximately two minutes and 33 seconds later.
Investigators asked the pilot whether he recalled any buffeting or shaking during the accident event. He said that he did not. Nor could he recall adjusting the power after the aircraft intercepted the glideslope. When queried about the approximate stall speed in the approach configuration, the pilot told investigators that it would have been in the "low 70s." He was shown the radar data for his approach, and he could not offer an explanation or any additional recollections regarding the apparent deceleration.
The NTSB concluded that the accident was the result of the pilot's failure to maintain adequate airspeed during the approach, resulting in a stall. Additional factors were the low visibility, the pilot's selection of the improper DME for the approach, his failure to attain the proper descent profile for the approach, and insufficient altitude available for the stall recovery.
The pilot told investigators that the accident approach was his first at Salt Lake City in the preceding 90 days. He did not say how long it had been since he flew an approach into that airport.
Had the pilot thoroughly briefed himself on the approach and set up his instruments and radios properly, his DME would have been selected to the ILS frequency instead of the VORTAC frequency.
By using the improper DME indication, the aircraft was much closer to the airport than the pilot thought that it was throughout the approach, and well above the glideslope until the pilot instructed the autopilot to descend and intercept the glideslope from above. That should have been a signal to the pilot that all was not as it should have been.
The approach plate clearly indicates that glideslope interception should be made from below after passing the SCOER intersection.
The pilot may not have suspected a problem, however, because he knew that his airspeed was higher than normal as he complied with the controller's instruction to maintain 170 knots to KERNN, and because it appears that he had not made an aggressive descent to ensure that he crossed what he thought was SCOER at the proper altitude.
We know from the radar indications that because of the DME switching error, the King Air was 4.7 miles inside of SCOER when the pilot thought that he was at the intersection. He would have had a full-down deflection of the glideslope at that point. It should have been a warning that something was amiss, yet he obviously did not see it.
At some point he must have pulled the throttles to idle, or at least significantly reduced the engine power, to slow the airplane to approach speed and to initiate the descent that ultimately allowed the airplane to intercept the glideslope 1.8 miles from the runway threshold. It appears that he did not increase power upon intercepting the glideslope, and as the autopilot pitched the nose of the airplane up to maintain the glideslope, it ultimately stalled.
Is it possible that the pilot was looking out the windshield for the runway and never noticed the airspeed bleed off to the point where the aircraft stalled? It doesn't seem likely that any pilot would forego looking at his instruments for that long of a period, but it is hard to come up with another explanation.
At the point where the aircraft intercepted the glideslope it was 1.8 miles from the runway and 478 feet above decision height. From there things deteriorated rapidly, because the airplane crashed 1.3 miles from the threshold.
This accident probably would not have occurred had the pilot set the DME switch to the number-one nav radio as specified on the approach plate. It is likely that he would have been much closer to the specified altitudes and the autopilot would have intercepted the glideslope from below, as was intended.
Had that happened, there should have been no reason to significantly decrease the power to the point where the airplane would have stalled. It appears that this aircraft never was properly stabilized on the approach. Pilots must ensure a complete approach plate briefing before beginning any approach. Don't rely on memory, because changes can be made to approaches that will impact how we fly them.
About the author...
Brian M. Jacobson is a corporate pilot and aircraft owner who lives in White Lake, Michigan.