Ice on the approach

November 1, 1999

Taking a winter trip when icing is forecast in an airplane without ice protection can be a risky proposition. Pilots are regularly told of icing forecasts that cover vast quantities of airspace at multiple altitudes. Forecasting ice accurately, though, is notoriously difficult, and almost every IFR pilot who's done much flying will tell you that forecasts and reality are two different things. However, every winter we lose a number of aircraft when pilots fail to approach the issue of ice with the caution that it deserves. The accident described in this column has multiple causal factors and shows the importance of having a plan and not abdicating your safety to ATC.

In late December, a Piper Saratoga departed from Albany, Texas, en route to Memphis, Tennessee, with a refueling stop in Wichita Falls, Texas. The pilot and passenger were returning from a hunting trip with a final destination of Cincinnati. The pilot had logged 1,615 total flight hours with 1,450 hours in the Saratoga. His logbook showed 380 flight hours in actual instrument conditions and indicated that he was current on instruments. Most of the flight time was cross-country, indicating that this pilot had some experience with weather.

The night before the trip, the pilot called for a weather briefing and filed an IFR flight plan. During the briefing, the pilot asked for a general synopsis and the best way to get back to Cincinnati. The briefer described a low-pressure system and associated fronts to the north. He suggested a routing that kept the aircraft well south of a direct line back to Cincinnati. A routing over Memphis seemed to be the most reasonable.

At about 8:50 the next morning, the flight made a fuel stop in Wichita Falls, Texas. The lineman noted some ice on the wing tips, which melted before departure. At 9:01 a.m., the pilot called flight service to file IFR from Wichita Falls to Memphis (MEM). The weather briefing in-cluded information on a low-pressure system over the route of flight. An airmet along the route called for occasional moderate rime or mixed icing in clouds or precipitation above the freezing level. Tops of the clouds were reported as flat, running about 5,000 feet from Wichita Falls to Memphis with clear skies above. The Memphis hourly observation was 600 feet broken, 1,100 overcast, visibility three miles in drizzle with a temperature of 42 degrees Fahrenheit. With a standard lapse rate of 3.5 degrees F per 1,000 feet, the freezing level should have been right around 5,000 feet, which is what the forecast predicted. There was no mention of freezing rain or drizzle in any of the forecasts.

Based on that forecast, was it reasonable to make the flight? Many pilots would have decided to go, based on multiple top reports and a freezing level that was high enough to provide some options. The flight might encounter some ice on the way down, but with those surface temperatures, it should melt off before landing. According to some pireps, not received by the accident aircraft, a number of light aircraft without deicing equipment were operating in the Memphis area. There is no evidence that the pilot received any weather up-dates or asked for pireps while en route.

The Memphis International Airport surface weather observation at 12:50 p.m. reported a measured ceiling of 1,400 feet broken, 2,300 feet overcast, visibility seven miles with light drizzle; temperature 39 degrees F, dew point 37 degrees F; winds 300 degrees at 14 knots with gusts to 20 kt, altimeter setting 29.74 inches; the drizzle began at 12:35. The temperature had dropped 3 degrees from the preflight briefing, and the potential icing layer was now about 1,000 feet lower than forecast.

At 12:38, while approaching Memphis, the flight was cleared from 9,000 to 4,000 feet. At 12:41:50, the pilot reported to the approach controller that the aircraft picked up moderate rime icing going through 5,000 feet, but that it "stabilized" at 4,000 feet. At 12:42:39, the pilot requested a lower altitude as soon as possible. At 12:43:21, the flight was cleared to 3,000 feet. About 1:07:13, the flight received vectors to the ILS approach to Runway 27 and was told to descend to 2,000 feet. At 1:07:44, the flight was cleared for the approach. At 1:09:18, the pilot was told to contact Memphis Tower. The pilot acknowledged, and there were no further transmissions from the Saratoga.

Radar data from Memphis Approach Control showed that the Saratoga flew an easterly heading and passed south of the Memphis Airport at 3,000 feet with a groundspeed of about 115 kt. At 1:06:30, the flight initiated a left turn to the north and continued turning until the aircraft was on a northwesterly heading. The groundspeed slowed to 100 kt and continued to slowly decrease while the aircraft descended to 2,000 feet. At 1:09, the aircraft turned to a westerly heading and the groundspeed was 85 kt. At 1:10:07, the groundspeed was 65 kt and the altitude was 1,900 feet. With a 15-kt wind, the Saratoga's estimated airspeed would have been about 80 kt. Only nine seconds later, the altitude was 1,100 feet, which equates to a descent rate of more than 5,000 feet per minute. The last radar return showed the aircraft almost seven miles from the airport.

Witnesses on the ground observed the aircraft descend from a 500-foot overcast layer of clouds in a nose-down attitude. As the aircraft exited the clouds, the nose raised to a steep nose-up angle, and the engine noise increased significantly. The right wing and nose dropped, and the aircraft entered a spin to the right. The engine noise level decreased, and the aircraft crashed in a residential area. Flaps were set at 10 degress. There were no survivors.

The pilots of several air carrier aircraft on the approach to Runway 27, before and after the Saratoga, reported the tops of the clouds at between 5,000 and 6,000 feet. At 12:41, 10 miles south of Memphis, a Douglas DC-9 reported moderate icing at 3,000 feet. Seven miles east of Memphis, only five minutes before the accident, a Boeing 727 also reported moderate mixed ice at 3,500 feet, which continued until the jet broke out of the clouds at between 1,000 to 2,000 feet. A Cessna 206 west of Memphis reported light clear ice from 8,500 to 4,500 feet at about the same time as the accident. The ice melted at 2,000 feet.

The forecast missed some local conditions that were detected by the Memphis Terminal Doppler Weather Radar system (TDWR) and later analyzed by Massachusetts Institute of Technology's Lincoln Laboratory to assist the National Transportation Safety Board. Their report, in condensed form, stated:

"The TDWR data showed the aircraft penetrated two medium-reflectivity echoes that were located to the southwest and west of the airport. The strongest echo only extended to an altitude of 1,300 meters (about 4,000 feet). Each of these penetrations lasted about one to two minutes. The winds were northwesterly, both at the surface and aloft, which does not indicate a significant icing episode [ unless the aircraft is not equipped to cope with ice — BL]. The freezing layer was located at an altitude of 500 meters (about 1,500 feet). The surface temperature at MEM was above freezing (e.g., 39 degrees) at the time of the accident. The airport only reported light drizzle/fog throughout this event, so if there was frozen precipitation aloft, it melted before reaching the surface. It is possible that the precipitation above 500 meters was in primarily a frozen or supercooled liquid water state. Since the aircraft penetrated the two echoes above the freezing level, it could have received some ice accumulation. According to the altitude plot, the aircraft was above the freezing layer for most of the analysis period, so any ice that did accumulate would not have melted. The significance of supercooled liquid water in this event should not be overlooked."

Pilots who choose to operate in a potential icing environment should be aware of some systemic issues. The ATC transcript does not show that the pireps from the DC-9 or the Boeing 727 were given to the Saratoga pilot. The FAA's response was that the pilot knew of ice in the area because of his own experience and had reported it as "stabilized." If there is hazardous weather, do not assume that ATC will volunteer other pireps. Ask if the controllers have received any. Although the controller's handbook requires pireps to be passed along, it is on a time-available basis, so if the controller is busy you may not get some potentially life-saving information.

Routine pireps may not get into the system to result in an immediately updated forecast or reissuance of an airmet. Icing is an area where much more data from the pilot community needs to be communicated back to the forecasters. The present system is cumbersome. If the pirep is given to flight service, it will be disseminated, but if given to a controller, it might or might not get into the system. The AOPA Air Safety Foundation has been having discussions with the FAA on this for some time.

The National Weather Service (NWS) did not update the icing airmet to reflect the lower icing altitudes based on the pireps that were transmitted at 12:45 and 1:05. Their response was that "the forecaster would have been busy composing the area forecasts for the north central and south central areas [covering one-third of the United States]. It is very possible that the forecaster may not have immediately been aware of the information contained in these reports." From a practical standpoint, this would not have made a difference for the Saratoga because it was already in the terminal area.

While the Saratoga was not approved for flight into icing conditions, many pilots might think that undertaking this trip was reasonable. However, the weather was not quite as benign as it appeared, and the pilot made some tactical errors. We are learning through ongoing research about small pockets of moderate icing, and it sounds as though several aircraft found them in the Memphis area. When a large aircraft reports moderate ice, that is a mandate for light aircraft to avoid the area. One strategy is to divert to an alternate or start negotiating with the controller.

Air traffic procedures play a significant part in this. ATC frequently will bring aircraft down early to allow departures to get up to higher altitudes. This is a fuel factor for turbine aircraft, but it can put a light aircraft that is trying to stay above the icing cloud tops into a compromising position. The pilot can insist on staying high, but that may result in some aggravation for ATC and, ultimately, a discussion about special handling and the need to declare an emergency. There is no emergency as long as the aircraft can descend through a narrow band of icing to above-freezing temperatures below. Being forced to operate in an icing layer is a problem. On the other hand, the system cannot tolerate light aircraft totally rearranging arrival and departure procedures because they have pushed the limits on ice. There are no easy answers to this. If you happen to get into a bad situation — declare the emergency. Solve the life-threatening problem, and then work out the legal one. The sanctions are far less permanent.

Based on this well-investigated accident report, the pilot might have been able to land normally had he maintained a higher-than-normal speed while maneuvering. The radar plot indicated a normal slowing upon intercepting the localizer. It appears that the aircraft stalled around 80 kt, which is at least 20 kt higher than normal. Ice can dramatically increase the stalling speed. That means the pilot must fly as fast and as straight as possible. Go for an airport with a long runway because the stopping distance will be much longer than usual. MEM was a good place to go.

The other absolute is not to use flaps when carrying airframe ice. This could result in a tail stall and cause the aircraft to nose over abruptly. At a low altitude, there is no room for recovery.

In this accident, a moderately experienced pilot took what appeared to be a slight risk, based on the forecast and pireps. The icing turned out to be more significant than expected. Because of traffic and ATC procedures, the flight flew at less than optimum altitudes. The pilot should have been more insistent about either staying higher longer or requesting lower altitudes sooner. Additionally, he should have requested weather updates, including pireps, while en route to Memphis. Taking an active interest in the weather is one of the best survival strategies.

Finally, it appears that the aircraft was under control until the final approach — when accumulated airframe ice, a reduced speed, and a configuration change caused the aircraft to stall. What would you have done?


See also the index of "Safety Pilot" articles, organized by subject. For more information about icing, visit the ASF Web site. Bruce Landsberg is executive director of the AOPA Air Safety Foundation.