Wx Watch: Winter Weather Accidents

Running the numbers on winter icing accidents

November 1, 1997

Winter is just around the corner, and once again it's time for pilots to include the risks of icing conditions in their preflight planning and in-flight decision making. Airframe icing and strategies for dealing with it were addressed in last month's column (" Wx Watch: Wrestling the Iceman," October Pilot), so I'd refer you to that article for more information on the subject.

However, airframe icing, per se, isn't the only weather problem that pilots must consider. It's just that airframe icing, and icing accidents, get the most press. Don't get me wrong — airframe icing is a serious problem. But there are other cold-weather meteorological risks out there. Although they are often overlooked, these other risks either cause or contribute to many winter weather-related accidents.

In preparing the AOPA Air Safety Foundation's Safety Review of General Aviation Weather Accidents, I asked the ASF's computer to come up with all the icing accidents between 1982 and mid-1993. The raw statistics and accident summaries came from the National Transportation Safety Board (NTSB), but it was the ASF's Emil Buehler Center for Aviation Safety that crunched the numbers and did the sorting.

It turned out that the NTSB recorded 637 icing accidents in the 10.5-year study period. Of those, 27 percent (172 accidents) involved fatalities.

The other icing

The ASF's search of icing accidents used ice or icing as keywords. As a result, carburetor icing, induction system icing, and even icy runways turned up in the batch of 637 icing accidents. Some may find fault with this methodology, asserting that airframe icing and all other ice-related phenomena belong in separate categories. Maybe they do, but the decision was made to keep all icing events in the same subset in the interest of consistency and in view of time constraints.

I learned something from the mélange of icing statistics that emerged. Carburetor ice, for example, caused far more accidents than airframe icing encounters. We have no idea of how many carb ice encounters had fatal outcomes, but the accident numbers alone make one sit up and take notice: 326 — or about 51 percent of all "icing" accidents — were attributed to carburetor or induction system icing. Airframe icing was involved in 262 crashes, or 41 percent of icing accidents.

It wasn't so much the carb or induction system ice itself that caused the preponderance of these accidents. Instead, the pilot's failure to use, or improper use of, carburetor heat or alternate induction air was blamed.

Carburetor ice forms when intake air experiences a decrease in temperature as it travels through the carburetor venturi. It can happen in a wide range of temperature conditions — anywhere from 105 to zero degrees Fahrenheit (40 to minus 18 Celsius). No, that 105 degrees was not a typo. This means that if relative humidities are high enough, carburetor ice can form in just about any weather — and you don't have to be in cloud or precipitation for it to build. The envelope for the most severe buildups of carburetor ice is between 65 and 100 percent relative humidity and 25 to 65 degrees Fahrenheit (minus 4 to plus 18 degrees Celsius). That's why carburetor icing happens most often in the colder months of the year.

The initial signs of carburetor ice can include engine roughness, an unexplained drop in revolutions per minute (in airplanes with fixed-pitch propellers) or manifold pressure (in airplanes with constant-speed propellers), and a drop in exhaust gas temperature. Pilots flying airplanes with carburetor temperature gauges can check to see if the instrument's indicator is in the yellow arc, meaning that the carburetor venturi is cold enough for ice to form there.

For advice on proper application of carburetor heat, consult your pilot's operating handbook. Most recommendations advise applying full carburetor heat and leaving the heat on for an extended period of time.

It's essential to apply carburetor heat at the first suspicion of carburetor icing and to avoid any temptation to return the carb heat control to the Off position should any roughness or vibration occur. The roughness is doubtless caused by the water from the melting ice being ingested by the engine. Wait long enough and the roughness should go away. Shut off the carburetor heat prematurely, and the engine will build more ice — and probably quit because of air starvation.

Because carburetor heat is provided by engine exhaust heat ducted to the carburetor, there obviously won't be any carb heat available if the engine quits — a big, big reason to turn on full carb heat and leave it on.

Fuel injection and alternate air

The fact that you fly with a fuel-injected engine doesn't mean you're exempt from induction system icing. True, you have no carburetor and, therefore, no chance of carburetor icing. But ice can still form over the intake air scoop or air filter, depriving the engine of air and thereby causing it to quit.

With icing like this, the first signs are apt to be a loss of power, as evidenced by a drop in manifold pressure or a slight change in engine sound level. This is the signal to activate the alternate induction air door or doors. When these doors open, intake air routes through them, bypassing the ice-blocked normal induction air pathway.

Many alternate induction air systems activate automatically; these designs use spring-loaded doors. Suction in an ice-blocked air intake draws these alternate air doors open. Some older fuel-injected airplanes have alternate air doors that must be manually opened. Knobs or levers — usually located beneath the instrument subpanel — have to be physically moved to the Open position in order for alternate air to reach the engine. With this design, pilots have to take the initiative and take timely action to prevent engine stoppage.

It can be easy to overlook the alternate air in the confusion and stress of an icing encounter. In preparing the ASF's safety review of the Piper Comanche and Twin Comanche, I found numerous instances in which pilots experienced engine stoppages in icing conditions. Most of the time, these pilots forgot to open the alternate air doors. However, there were cases in which alternate air doors in certain Twin Comanches were frozen shut by the very ice they were meant to combat! This unforgivable design flaw was corrected by a service bulletin and an airworthiness directive that required the directing of heated exhaust air to the alternate air door mechanisms.

The moral: Know your alternate air system and use it as part of a checklist item should you inadvertently enter icing conditions.

Continued flight into known icing

A pilot of a non-ice-protected airplane learns about icing forecasts and reports during a weather briefing, takes off into or continues flying in the ice, and crashes. At least 68 accidents fit that approximate description in the ASF's 1982 to 1993 study period. The mysteries behind this kind of poor decision making are the subject of ongoing speculation, and are best addressed in a separate forum. These kinds of accidents closely fit the mold of the typical VFR-into-IMC accident (there were 580, 226 of which occurred in the winter months), with the added wonderment of contemplating why anyone would knowingly risk flying into icing conditions. It just goes to show that you can preach icing avoidance until you're blue in the face, but a small group of determined pilots still manage to convince themselves that they're exempt from risk.

Winter weather accident sampler

  • A 600-hour, instrument-rated private pilot planning to fly a Beech Bonanza 36 from Springfield to Joplin, Missouri, was told by the Springfield tower that icing conditions existed along his proposed route. He and his passenger took off anyway. Four minutes later he reported picking up rime ice, but he continued flying toward Joplin, accumulating more ice along the way. At Joplin, the weather was 500 feet overcast, with a visibility of 2.5 miles in fog and freezing rain. The pilot made an ILS approach to Runway 13 and planned to circle to land on Runway 31. He crashed while maneuvering, killing himself and his passenger. A pilot who landed a few minutes before reported that he had picked up mixed ice on his windshield and one-quarter-inch of clear ice on his wings in the six to 10 minutes prior to landing.
  • The 4,414-hour, instrument-rated pilot of a Cessna 414 crashed and died after trying to make a night landing at Aurora, Colorado, in 200-and-2 conditions in snow. He, too, was told of icing conditions in the Denver area during a weather briefing. Approaching Denver, the pilot radioed ATC, "Can you get me down, it's a little bit icy up here." He was vectored for an approach, missed the approach, and one minute later said, "Lots of ice…" and "I'm stalling out."
  • A 1,575-hour instrument flight instructor and a passenger escaped injury after a hard landing at Purdue University Airport in Lafayette, Indiana. The Piper Dakota was substantially damaged, with the right main landing gear separating from the airplane at impact. He was advised of reports of light rime ice but took off anyway to do practice approaches. The weather was 800 overcast, visibility six miles in fog. A post-accident investigation showed between one-half and one inch of ice on the wings. The instructor said that he noted ice accumulating on the windshield, but he added that he wasn't concerned because the airplane "could handle light ice very well."
  • During the preflight runup prior to an instructional flight, the pilots noticed that the carburetor heat control would move only one-half of its normal travel. The pilots discussed the problem and ran the engine for another one to two minutes until their Piper PA-28-150 Cherokee's oil temperature rose. Weather conditions were 3,000 feet overcast, visibility 7 miles. After takeoff from Derby, Kansas' Cook Airfield, climbing through 400 feet agl, the engine began to lose power. Emergency procedures were followed, but there was no gain in power. The airplane hit treetops during an attempt to return and land, and one of the pilots sustained minor injuries. A postflight investigation revealed frozen water on the inside of the engine air intake filter. Meteorological conditions were conducive to carburetor icing.
  • In cruise flight between Lawrence, Kansas, and Liberty, Missouri, the 1,196-hour pilot of a Cessna 152 experienced an engine misfire. He applied carburetor heat, and the engine performed normally. He then removed carburetor heat; after two to three minutes, the engine lost substantial power. Carb heat was reapplied, followed by a gain in power — but not to original levels. The pilot decided to make an emergency landing in a field, ran onto a road, and hit a telephone pole with the right wing. According to icing probability charts, conditions were conducive to carburetor icing. The temperature was 26 degrees Fahrenheit, and the dew point was 3 degrees Fahrenheit. Ceiling and visibility were 20,000 feet and 10 miles, respectively.
  • At 7,000 feet, flying his Piper PA-24-180 Comanche from Asheville to Sanford, North Carolina, the 1,444-hour pilot reported that he couldn't maintain altitude because of a power loss. "We're down to about 14 inches, and we just broke out, and I've got the lights," the pilot said, referring to the approach lights of the Asheville Regional Airport. A short time later, the pilot reported a complete power loss, said that he was "going for Henderson," and indicated that he was in the clouds. The airplane crashed into a hillside 4.5 miles southeast of Asheville, on the extended centerline of Runway 16. The pilot and two other occupants were killed. Weather was 200 feet scattered, 1,200 feet overcast, visibility 1.5 miles, with the temperature and dew point both at 49 degrees Fahrenheit. The carburetor heat door was jammed in the cold (Off) position, and the carburetor heat control knob was in the cold (Off) position, too. Two other pilots in the area at the time reported icing at 8,000 and 9,000 feet.

These examples prove that it's not just low-time, VFR-only pilots who make serious errors in judgment when it comes to dealing with winter weather. This winter, let's all minimize our risk by really paying attention to warnings of adverse conditions, not just paying them lip service.


E-mail the author at tom.horne@aopa.org.