The big chill

Looking at the numbers, one might be tempted to think that general aviation has finally learned its lesson about ice. In-flight icing is blamed for about a dozen accidents in a typical year, around 40 percent of which are fatal. That’s roughly 1 percent of all fixed-wing accidents and 2 percent of fatal accidents, which could be interpreted as evidence that the flying community has generally gotten the message: Light aircraft without known-ice certification have no business playing in the freezing stuff. Even those approved for flight into known icing require constant vigilance for accumulations more severe than their de-icing systems can manage, an ironclad plan of escape, and no hesitation in executing it when circumstances require.

A closer look may temper that optimism. For all the attention we pay to thunderstorm avoidance, about twice as many accidents are attributed to icing—which have caused 20 percent more deaths even though thunderstorm encounters are almost twice as likely to be fatal. Even those figures may be overly optimistic, since evidence of ice may disappear by the time the wreckage is discovered.

On Nov. 18, 2011, a turbocharged Cessna T337G Skymaster departed Rawlins, Wyo., about 1:15 p.m. on an instrument flight plan to Casper, Wyo. Sixteen minutes later, Casper Approach Control cleared the flight to descend from 13,000 feet to 10,000 feet. The controller told the pilot to expect the ILS to Runway 3 and provided vectors to the initial approach fix; four minutes later, he received his approach clearance with instructions to maintain 7,600 feet msl until established.

Problems arose almost immediately. It took him three tries to confirm the altitude restriction and read the clearance back correctly, but less than two minutes after that the approach controller issued a low-altitude alert. After acknowledging that, the pilot was cleared to tower frequency, but by the time he checked in he was a quarter-mile right of the localizer. When queried by the tower controller, he replied, “Yeah, we’re having a little rodeo up here.” The controller immediately instructed him to fly runway heading, climb, and maintain 8,000 feet, and expect vectors for a second approach.

The weather was not good. A special METAR recorded when the approach clearance was issued listed overcast ceilings at 2,500 feet, two scattered layers beneath, and visibility of a mile and three-quarters in light snow and mist. By the time the pilot began the missed approach, however, the ceiling had dropped to 1,300 feet overcast and visibility had fallen to half a mile in moderate snow and freezing fog. Airmets for IFR conditions, moderate turbulence below 16,000 feet, and moderate icing below 20,000 feet were all in effect, and the terminal forecast for Casper warned of a 35-knot wind shear at 1,000 feet. The National Weather Service’s Current Icing Potential estimated a 60- to 80-percent chance of moderate to severe icing during the approach—and at 1:43 p.m., the approach controller advised that runway visual range (RVR) had dropped to 2,000 feet on Runway 3, below minimums for the ILS. The pilot was offered the choice of continuing on the approach with a possible hold on the localizer or holding at the Muddy Mountain VOR. He elected to continue toward the final approach course.

It seemed like the right decision. Within four minutes the Runway 3 RVR was back up to 2,800 feet, 400 feet above minimums, and the controller advised that it appeared to be increasing steadily. Twenty seconds later, though, the controller transmitted that the Skymaster was 90 degrees off its assigned heading and told the pilot to turn to 250 degrees, which he read back. Eighteen seconds later, the controller transmitted, “Low altitude alert. Check your altitude immediately. Minimum vectoring altitude in your area is seven thousand five hundred. Climb and maintain eight thousand six hundred. Are you having trouble maintaining altitude?” There was no reply, and subsequent calls went unanswered.

The wreckage was found about two hours later in an open field eight miles northeast of the airport. It had been almost entirely consumed by fire, making it impossible to tell whether the airframe had accumulated ice. The Skymaster was equipped with the factory known-ice package, including a windshield hot plate; boots on the wings, tail, and propellers; and heated pitot and static ports. Still, the NTSB concluded that “Based on the weather conditions that the pilot encountered while maneuvering … it is likely that the pilot failed to maintain adequate airspeed due to structural icing, which resulted in a loss of airplane control and subsequent impact with terrain.”

The pilot was experienced; his 4,500 hours as pilot in command included 700 in type and at least 645 in actual IMC. He was a Wyoming resident, was familiar with local conditions, and had gotten two weather briefings before the flight. The airplane was about as well equipped for icing conditions as any light piston twin ever is—but ice can build up fast enough to overwhelm any aircraft’s systems. Even the overall degradation of performance that results is not as dangerous as the unpredictable stall behavior—particularly in turbulence, at low altitude, in IMC.