Reasons Behind Fatal Accidents

Some pilots are, by nature, worriers. They worry about fuel, about engine failures, about hazardous weather, about midair collisions. Bluntly, pilots worry about things that can kill them. But do they worry about the right things? In other words, does the risk framework that most of us construct in our personal aviation universe reflect the reality of the serious killers in aircraft accidents?

Our guess is that it does not, unless pilots are out there really sweating about stalls, spins and controlled flight into terrain. And even if the pilot population is wide awake about these hazards, it could do a better job of avoiding them. Stalls and CFIT (controlled flight into terrain) pop up as the two biggies in fatal accidents in general aviation to a degree that, frankly, startles us.

To put numbers on this, we recently carefully reviewed NTSB accident reports for two years-2000 and 2008. Theres no magic to these picks; we simply wanted to compare data from two recent years.

We considered only accidents that occurred in the U.S. involving certified aircraft. We didnt include experimental aircraft, nor did we include N-number aircraft that crashed outside of the U.S. To gain some sense of where the risks lie, we sorted the accident causes into 15 categories, ranging from fire in flight, to spatial disorientation, to stalls to mid-airs.

The problem with interpreting NTSB accident reports is that some of them arent especially complete. And even when they are, the listed probable cause doesnt always jibe with the factual report. For example, if the pilot runs the airplane out of gas and spins into a cornfield, whats the accident cause? Or if the airplane clips the tops of trees on approach, is that loss of control or CFIT?

Sometimes, its impossible for the accident investigator to determine this, either due to lack of data or time to carry through the investigation. Our abiding caveat here is that in sorting these accident causes into our own categories, we have done our best to interpret the ambiguous findings fairly. Another analyzer might reach different conclusions on some of these reports. But in our view, the overall conclusions would be similar. For the record, in our interpretation, the pilot who runs out of gas and spins in died because of fuel exhaustion, not a stall/spin.

Having said that, we were somewhat astonished to learn how many pilots really do die in stall-related accidents. According to our analysis, its 18 percent or nearly one in five. Thats a big number. Combine it with CFIT fatals-another 16 percent-and fully a third of all fatal accidents are caused by stall-related issues or just running a perfectly good airplane into the dirt, the rocks or trees.

The actual number, we think, is even higher, because loss of control accounted for another 17 percent of accidents during the two years we examined. Although we cant be certain, we think that some of the loss-of-control accidents are actually stalls or stall/spins that the accident report simply doesnt or cant illuminate. Some are probably CFIT, too.

How and why do so many pilots get into stall trouble in airplanes? We wish we could say theres a typical scenario, but we cant say we found a smoking-gun repetitive pattern. But there are some commonalities. The so-called “moose stall” is one.

In almost every part of the country, light aircraft are used to observe or hunt animals from the air. These flights involve low-altitude maneuvering-as low at 100 feet above trees or terrain-and steep turns. This is a deadly combination that has killed dozens of pilots and passengers. In Florida, for example, four people died when a Cessna 172 collecting bird activity data spun in from 300 feet. The airplane was five pounds over gross, which would have bumped up the stall speed a bit, but not nearly as much as high bank angles do.

Speaking of which, its no particular surprise that many stall/spins occur in the traffic pattern for the same reasons they have always occurred: over steep turns at low speeds and altitudes. Although CFIs are supposed to hammer into their students the need for maintaining airspeed in the pattern-really, it should be maintaining the right angle of attack-and avoiding more than 30 degrees of bank, pilots ignore these admonitions nearly a dozen times a year, and probably more often.

If we were to guess here, based on our reading of the accident reports, we would say that as with the moose stalls, the problem isnt airspeed itself, but bank angle and a weak grasp of wing loading versus stall angle of attack. This is nothing new, by the way. Barrels of ink and entire forests have been denuded describing the dangers of hurrying the base-to-final turn with steeper bank or inside rudder. Neither the aerodynamics nor the consequences of ignoring them have changed since Wolfgang Langewiesche described both in Stick and Rudder 65 years ago.

Again, another qualification. Controlled flight into terrain is generally associated with instrument operations, often when the airplane is wings level, climbing or descending, and collides with terrain or obstacles. For our purposes, we included accidents that met this definition, but also scud running, canyon crawling-theres a lot of that, for some reason-and botched takeoffs or approaches that wound up in the trees. Many of these occurred in visual conditions-including the scud runs, for that matter-but resulted in a perfectly flyable airplane being smashed into an unyielding object.

First, the scud running. True scud running is not a common type of accident. Of the 338 fatal accidents we looked at, six would qualify as classic scud runs; barely two percent of the total. Whats a classic scud run? The pilot takes off under a low overcast or in poor visibility with the intent of flying a cross country. One of several things happens. He runs into a tower, tree or terrain he couldnt see even though he maintains altitude; the visibility decreases and he simply loses control; the ceiling lowers, forcing him to hit something at an altitude lower than he ever intended. (Editors note: the pronoun “he” is intentional here; women dont seem to pull this stunt.)

We also found a handful of what we would call modified scud runs. These consist of pilots who actually had been on an IFR flight plan or an approach and cancelled it to complete the trip visually, often in marginal conditions. Why pilots who appear to be well trained do this is one of many great puzzlements in aviation.

Also in the CFIT category are duck unders, blown mins approaches and missed approaches that just never got in gear. One example of a duck under occurred at Marthas Vineyard in October of 2000. A Mitsubishi MU-2 pilot departed for the Vineyard at night, having obtained no briefing nor bothering with a flight plan. When he arrived, the weather was a 100-foot overcast in two miles of vis, so he requested the ILS 24, which has a marker crossing altitude of 1500 feet. Post-accident review of approach radar data revealed that the airplane crossed the marker at 700 feet and was at 300 feet at 1.5 miles from the airport. Although the aircraft appeared to be on the centerline, the pilot evidently never intercepted or flew the glideslope. Four people died.

Although stalls and spins qualify as loss of control, the NTSBs probable causes frequently list “pilot loss of control” as an accident cause with little additional detail. Some 17 percent of fatals fall into this category. As noted above, some of these probably are stall/spins, but many are simply inexplicable; a perfectly good airplane arrives at the bottom of a smoking hole.

In its monthly review of accident records for a particular type of aircraft, our sister publication, Aviation Consumer, finds that for most types of airplanes, loss of control on the runway is the leading cause of accidents. These are groundloops, overruns and excursions into the runway ditches that are rarely fatal. But they can be. We found several examples of overruns and excursions that were so fast and violent that some or all of the aircraft occupants were killed. One particular danger is landing too fast and exciting porpoising which, if not arrested, can cause the airplane to flip, making survival less likely.

Some loss-of-control accidents occur out of what appears to be normal cruise flight. The airplane simply…crashes. Some of these may be undetected mechanical issues or pilot incapacitation or, an overlapping factor, spatial disorientation caused by VFR flight into IMC. These are rarely survivable, nor are they infrequent. The words “spatial disorientation” were cited in 10 percent of the accidents we analyzed.

A like number of fatal accidents qualified as having unknown causes or were simply too weird to categorize. We would put structural breakups-very few of them these days-in this category, along with inflight fires, turbulence and pilot incapacitation. These risk categories have too few incidences to make much of.

There were some. The aviation press-including us-makes much of convective weather and icing as accident hazards. And although both do represent significant risk, they are all but no factor in recent fatal accident trends. Icing was listed as causal in under two percent of fatals during the two years we analyzed. Convective weather was also cited in under two percent. The actual numbers may be higher due to misinterpretation of the data or lack of detail, but even doubling these numbers makes icing and thunderstorms a minor fatal risk factor. The larger weather risk is low visibility, which leads to loss of control or CFITs.

How about midairs? Some owners worry so much about this that spending many thousands of dollars on traffic equipment seems justified. Our data sweep revealed that six percent of fatal accidents are caused by midairs. While thats not insignificant, the data suggests there are larger fish to fry, mostly in the area of basic stick and rudder skills and judgment.

And judgment-really, the lack of it-leads inevitably to an examination of fuel exhaustion, which happens in GA about once a week. Thankfully, most of these arent fatal, but we found that five percent are when an otherwise perfectly flyable airplane is pushed into a corner when the engine quits and the pilot cant pull off the emergency landing. Some of these lead to stalls or stall/spins. Nearly identical stats apply to genuine mechanical engine failures not related to fuel exhaustion: Just under five percent were fatal.

“Risk management” is a buzz phrase thats frequently used, but less often explained. In our view, to manage risk, you need to have some means of ranking hazards contextually. In other words, even though we publish articles about the terrors of convective weather and icing, the actual risk of a fatal accident as a result of either one is relatively low. Its higher if you fly in such conditions regularly, but zero if you never do.

But everyone who flies is always subject to the stall/spin and the CFIT risk. Focusing on the stall risk, we think that based on our findings, a strong case can be made that many pilots simply dont understand the dynamics of stalls and are caught by surprise when they get into an unrecoverable one at low altitude. Obviously, pilots who fly low-altitude game research, pipeline patrols and aerial application work are at high risk, but these operations represent only a portion of stall/spin accident history.

Pilots flying routine patterns get into stalls, so do pilots taking off or, for that matter, cruising at high altitude. And as the Colgan crash in Buffalo, N.Y., last winter evidently shows, airline pilots arent immune.

Were not sure if this represents an indictment of stall avoidance training so much as a failure of pilots to understand the more subtle aspects of what is, in the end, a simple concept. Either way, many of us would be well advised to explore more than just the obligatory stall series when the next flight review comes due.