February 2002 Volume 45 / Number 2
The 90-Percent Solution
Avoiding spatial disorientation equals avoiding IMC
If you have a vacuum failure in IMC and no redundant power source:
- Revert to partial-panel, using the turn coordinator to keep the wings level.
- Use airspeed and vertical speed instruments to ascertain pitch.
- If you have a GPS, even a handheld, heading information can be derived from it to substitute for a heading indicator. If not, refer to the compass.
- Cover the offending instruments—you do have covers readily available, don't you?
- If in a complex aircraft, lower the gear. This helps stabilize the aircraft and minimize speed buildup if it does get away from you momentarily.
- If your aircraft derives autopilot input from the electric turn coordinator, turn on the autopilot. (If you're unsure about the AP's source of information or its operation, now would be an excellent time to ask the manufacturer or your avionics technician.)
- After the aircraft is under control, notify ATC and declare an emergency.
- Fly to VFR weather if feasible. If not, ask for no-gyro vectors to the nearest radar-equipped airport with an ILS.
Simulating spatial disorientation
In the summer of 2000, ASF undertook a study of spatial disorientation, funded by Sporty's Pilot Shop with assistance from FlightSafety, at FlightSafety's Cessna Learning Center in Wichita, Kansas. The objective was to see how pilots handled an un-annunciated vacuum failure. A Cessna 210 nonmotion visual simulator was used and 24 subjects were told that they were part of an instrument flying experiment, but they were not told specifically about what would happen.
The results were sobering. After the pilots had flown about 40 minutes and had become used to the simulator environment, the vacuum pump was discretely failed. Sixty-six percent of the participants lost control and 50 percent "crashed" while attempting a partial-panel approach. It should be noted that at the conclusion of their training, all pilots were able to handle the scenario without loss of control.
ASF decided to test a bit further, working with the FAA's Civil Aerospace Medical Institute. The foundation's Piper Archer was rigged so that a vacuum failure could be induced by the flick of a switch—a trick used on 25 unsuspecting pilots. This allowed a real-world approach to detecting the problem. In addition to a view-limiting device, ASF used special polarizing material that allowed the safety pilot to see, but prevented a peek over the instrument panel by the subject. Only current instrument-rated pilots were selected and the experience level ranged from the newly rated to seasoned ATPs.
The Archer was equipped with a video camera, cockpit voice recorder, and GPS flight data recorder so there would be no subjectivity regarding the results. Comments by the subjects were "colorful" and the surprise factor was universally successful. All pilots correctly identified the vacuum failure, typically about 6 minutes after the system had been failed, but only a few of them declared an emergency.
Eighty-four percent of the group completed the partial-panel approach successfully. Others missed the approach but the aircraft remained under control. When the needles approached full-scale deflection they made the decision to go around. Everyone, when asked, could identify his or her position relative to the airport.
One pilot had suction-cup covers to put over the instruments. One used dollar bills, and three ripped up unused approach charts to paper over the instruments. The rest flew with all instruments uncovered but with noticeable distraction.
While ASF was flying pilots in the Archer, the FAA's Civil Aerospace Medical Institute (CAMI) in Oklahoma City duplicated our FlightSafety experiments in a visual fixed-base Piper Malibu and Cessna 172 simulator. The Malibu pilots had similar results to the FlightSafety Cessna 210 subjects with a substantial loss of control, but the 172 group more nearly resembled the ASF Archer group with no loss of control. This suggests that pilots of complex, retractable-gear airplanes might have more difficulty in managing flight instrument failures. When the Malibu was equipped with an electric-powered horizontal situation indicator (HSI) and a backup attitude indicator, only 8 percent of the pilots lost control.
To explore further, we rigged a Beechcraft Bonanza A36 like the Archer to see if a real high-performance complex aircraft would complicate things. Early results show that it does—there was a higher loss of control in the Bonanza. The research, which was still continuing at press time, has raised many questions. Simulated vacuum failure, both in simulators and in aircraft, may not quite replicate real-world IMC when it comes to spatial disorientation. However, we have seen that it serves as an excellent preventative tool when applied in a training program.—BL
Spatial disorientation has always been a part of aviation and three recent high-profile accidents make it a topic of continuing interest. The National Transportation Safety Board (NTSB) determined that the John F. Kennedy Jr. accident on a dark night over the water was the result of spatial disorientation (SD). The governor of Missouri, Mel Carnahan, was killed in a Cessna 335 at night. Just prior to the crash, the pilot radioed that he had instrument problems. That crash is currently being investigated. A Beechcraft Bonanza crashed in downtown Newark, New Jersey, in instrument conditions after the pilot advised air traffic control that his gyros were inoperative. Judging by these three headline-making accidents in less than a year, one might draw the conclusion that general aviation aircraft are falling out of the sky regularly because of instrument malfunction.
After the Kennedy accident, the AOPA Air Safety Foundation decided to take a closer look at spatial disorientation (SD). This is not a new area of investigation. One of ASF's most famous programs introduced in the 1950s—the 180 Degree Rating Course—was designed to help VFR pilots make the lifesaving turn to exit the clouds before losing control.
How big is the problem? ASF's database of NTSB accident reports, which is restricted to fixed-wing aircraft weighing less than 12,500 pounds, identified 268 accidents investigated by the NTSB as having SD as the cause from the beginning of 1990 through 1998. About 90 percent were fatal. GA accidents typically have a 20- to 25-percent fatality rate, so this type of accident is deadly. Over the past 10 years GA has averaged about 30 SD accidents per year. The trend, however, is definitely down. In the 1980s the average number of SD accidents per year was around 50. That represents a 40-percent reduction in the past two decades.
Pilots, however, not instrument or vacuum system failure, are by far the leading source of these accidents. According to the NTSB, 242 pilots lost control of fully functional aircraft during the period of the review. As you would suspect, noninstrument-rated pilots had the lion's share of problems. In this data set, 164 VFR pilots suffered from SD compared to 78 instrument pilots. Still, many more instrument-rated pilots showed up in the sample than we expected.
Most of the IFR pilots were not on an IFR flight plan. The data does not provide much detail on the background of the instrument pilots, so in many cases we don't know if they were current or how much actual instrument experience they had. Despite holding the rating, the suspicion is that a significant number had not flown instruments for some time. Perhaps that's why they attempted the scud run. They also may have felt more confident that if they encountered clouds they could handle it. Bad assumption. It's unfortunate that our proficiency erodes much more quickly than our confidence level.
VFR flight into instrument meteorological conditions (IMC) is a perennial disaster scenario and the statistics prove it. It's bad during the day, and as you might imagine night compounds the problem significantly. The clouds are not easy to see in the dark, nor is the terrain. Temperature and dew point frequently close to become fog. Pireps and other weather observations are less plentiful.
Why do pilots push into bad weather when they are not prepared to handle it? There are no easy answers, but I'll offer some observations. In some cases pilots use egregiously bad judgment, taking off even though clouds can be seen at the airport and visibility is down. Bluntly, these are Darwin Award candidates.
We found several accidents that would have been avoided had the pilot heeded warning signs just a few minutes earlier. But instead of taking action at the critical moment, the pilot was probably behaving more like a passenger—it was easier not to make a decision to change but to continue on and hope for the best. That is not a long-term survival strategy.
ASF has worked with the FAA and other industry groups over the past several years to determine the root causes of weather-related accidents and what can be done about them. The assumption by the industry/FAA safety analysis team was that most pilots will not take a deliberate risk—they must believe that the flight has a reasonable chance of success and that there is always an option to divert to an alternate.
One interesting revelation was the significant overuse of the weather briefer's term "VFR not recommended." But what does that have to do with SD?
Now you might ask why "VFR not recommended" is used so often when the ceiling and visibility appear to be perfectly safe for VFR flight. The origin of this lies in the briefing manual for flight service station (FSS) weather briefers. Paraphrased, the guidance is that if, in the briefer's judgment, the condition will occur either aloft or on the surface during the current or forecast period and VFR flight is doubtful, the warning should be issued. The National Weather Service (NWS) provides the forecasts, sigmets, and airmets to briefers. An airmet may be issued for a very large area of airspace for IFR conditions even if there is only a very slight chance that the conditions could develop during the forecast period anywhere within the area. Is it any wonder that after a few bogus warnings at least some pilots stop listening? It appears that some VFR pilots develop a disdain for the official weather word and build up experience successfully completing trips until one day the weather really is as crummy as the forecast. There usually aren't survivors to chalk it up to experience.
The reason given for the FAA's conservative approach is that it prefers not to deal with the complexities of weather and the resultant difficulty of attempting to explain the details to pilots whose grasp of weather may be limited. As legal counsel might put it, "It's better to over-warn than be sued for failure to warn." However, this isn't the level of service we need. The NWS would be happy to improve the accuracy of the forecast supplied to briefers and has the technical capability to do it as soon as the FAA agrees. Several suggestions: First, we need truth in forecasting and the agreement between the FAA and the NWS needs to be amended to make that possible. Equally important, the FAA doesn't want to spend time and money in court defending against what many would consider frivolous lawsuits. Pilots also need to understand that they are responsible for the safe operation of their aircraft, not the government. It's not GIC (government in command), although there are times when it may seem that way. Everybody wants to complete the trip on schedule, but the real question is how much risk are you willing to accept for yourself and your passengers?
Let me again push for all pilots to routinely file pireps on cross-country trips. This will improve the forecasts and help FSS to disseminate the real story. If pilots reported on every cross-country trip, in good weather and bad, I believe there would be fewer of these accidents. If there were several reports from other pilots who were either flying IFR, or were VFR types who took a look and decided to bail, it might well make a difference. Would you be as likely to launch if you had four pireps along your route reporting instrument conditions?
What about the cases of mechanical failure leading to a spatial-disorientation accident? In August 2001, ABC's PrimeTime news magazine broadcast a one-sided view hyping the idea that light-aircraft pilots were in great danger from vacuum equipment failure. The facts show things differently. During a 10-year period, there were 20 accident flights when vacuum pumps failed and another nine when an actual instrument, such as an attitude indicator, failed. In only about three accidents per year, or about 11 percent of all SD accidents, was the NTSB able to find a mechanical fault. Put another way, only one out of every 1,000 general aviation accidents is the result of vacuum failure. There may have been some other cases that went undetected but even if the NTSB missed half the failures, the numbers are still extremely low compared to the self-inflicted damage caused by pilots. Pilots flew fully functioning aircraft into the ground almost 90 percent of the time. We have no way of knowing how many pilots successfully coped with a mechanical failure.
A 1998 accident cited by ABC illustrates the anecdotal nature of the problem, and we can empathize with the 450-hour instrument-rated pilot in that incident. A Piper Dakota crashed after the pilot experienced spatial disorientation at night. The pilot was on an instrument approach when he advised air traffic control that he had a "full failure." One minute later he reported a "static failure," and shortly afterward he stated that the "DG" was out of order. Radio and radar contact was lost, and the controller who was tracking the airplane stated that it appeared as if the airplane was "going around in circles" just before it disappeared from the radarscope. A survivor reported that the pilot tapped on the instrument "with the blue horizon" and stated, "This instrument is not working."
The 1969 Piper was a rental, had been through annual inspection the month before the accident, and had accumulated 3,503 hours. A post-accident examination of the vacuum pump revealed an in-flight failure. The single vacuum pump was replaced in 1982 and had more than 1,200 hours in service. Airborne, the pump's manufacturer, recommends that the pumps be overhauled or replaced regularly, typically after 600 to 800 hours. However, unless the aircraft is operated under Part 135 for charter, there is no requirement to adhere to the manufacturer's recommendation. There was inconclusive evidence that the electrically driven turn coordinator was functioning at impact. So it is possible that not only did the vacuum system fail, but that the electric turn coordinator was inoperative as well. The NTSB determined the probable cause to be "the pilot's inability to maintain control of the airplane after experiencing spatial disorientation. Factors were the total failure of the vacuum pump, fog, drizzle, and night conditions."
In slightly more than 5,000 hours of flight, I've had two vacuum pumps go south. One was a nonevent in visual conditions where a backup system took over automatically and advised me of the situation. That was far preferable to the first episode in solid IMC, which necessitated a partial-panel ILS approach with a 400-foot ceiling and one mile's visibility in fog. The pump that failed had accumulated only 25 hours. The mechanic neglected to clean the vacuum lines thoroughly and thus the second pump became contaminated from residual debris from the original pump failure. After that, I no longer took vacuum pumps for granted. I had a backup system installed immediately on the Piper Arrow I was flying. One thing that really helped the transition to partial-panel was the Arrow's low vacuum annunciator, which identified the failure immediately.
The statistics show that the aircraft systems are relatively reliable, at least as far as reported accidents go. Ask a group of pilots how many have actually suffered loss of instrumentation, especially in instrument weather, and typically only a few will raise their hands. However, if you do much serious instrument flying, a backup system and some type of flagging or annunciation to call attention to the problem is an excellent idea. So is adhering to the manufacturers' recommended time limits for overhaul or replacement. Most of the airframe manufacturers started installing redundant systems as standard equipment on IFR-certificated aircraft in the mid-1980s. The litigation was significant at that time and it continues to this day. Airborne, a division of Parker Hannifin and one of the largest pump manufacturers, is withdrawing from the business.
Periodic refresher training on partial-panel skills is required for instrument proficiency checks. It doesn't take long to practice and I heartily recommend it. If you should be unfortunate enough to have a failure in actual IMC, declare an emergency, and if you can get to VFR weather easily, go there. If not, divert to the nearest airport that has radar facilities and an ILS. Despite the training, this isn't the time to be testing procedure turn and nonprecision approach skills. Slow down and if it's a retractable, put the landing gear down to provide further stability and more time to recover if the aircraft should get away from you.
Redundant systems on IFR aircraft solve about 10 percent of the problems and they're worth installing. It's also a good idea to take a conservative approach to maintenance and replace equipment before it fails.
Better forecasts and more pireps will help pilots to make better decisions. The 90-percent solution, the one thing that we could do to significantly curtail fatal spatial-disorientation accidents, would be to not fly into IMC when not rated or current. Some pilots are lucky — but we know what happens to at least some of the transgressors.
See also the index of "Safety Pilot" articles, organized by subject. Bruce Landsberg is executive director of the AOPA Air Safety Foundation.