Rapid Icing Response: Similar Encounters With Different Outcomes

Outboard section right wing
Outboard section right wing, lodged in tree
Credit: NTSB

When winter arrives, as it soon will in the Northern Hemisphere and where it can linger, aviators need to become hyper alert for conditions that invite ice formation aloft. Given the right combination of temperature, moisture content, airfoil design and crew action, the result can be sudden, unexpected, confusing and potentially catastrophic. Two flights with drastically different endings underscore the concern.

On Dec. 20, 2011, at about 1005 EST, a Socata TBM 700, N731CA, collided with terrain following an inflight loss of control near Morristown, New Jersey. Visual conditions prevailed. The flight had departed from Teterboro, New Jersey (KTEB) 15 min. earlier en route to Atlanta. The private pilot and four passengers were killed and the airplane was destroyed.

The 45-year-old pilot and his business colleague, a non-pilot, were in the cockpit. The pilot’s wife, two children and a family pet were in the back. The pilot held a private pilot certificate rated for SEL and also an instrument rating. Approximately five months earlier, he had reported 1,400 total hours of experience when he took his second-class medical. His personal logbooks were not located after the accident.

Records indicate the pilot took a two-day recurrent training course in the single-engine turboprop just over a month before the crash. The training addressed the technical aspects of the aircraft’s installed ice protection and environmental systems including preflight checking or testing, along with normal and emergency checklists. Simulator training consists of system checking, testing and operation including operating in icing conditions, at altitude, and system malfunctions.

Left wing, burned
Left wing, burned, found within main wreckage debris field  Credit: NTSB

The TBM 700’s Pilot’s Operating Handbook — Section 2 Limitations — focuses on accumulation of ice on the upper surface of the wing aft of the protected area. It states that “since the autopilot, when operating, may mask tactile cues that indicate adverse changes in handling characteristics, use of the autopilot is prohibited when any of the visual cues specified above exist, or when unusual lateral trim requirements or autopilot trim warnings are encountered while the aircraft is in icing conditions.” Indications were that the pilot had entered icing conditions with the autopilot engaged.

On the morning of the accident, Teterboro Airport had 10 kt. of wind with relatively high ceilings. Somerset Airport (KSMQ) in Pluckemin, New Jersey, was reporting ceilings of 12,000 ft. just before the crash. KSMQ is 12 mi. from the crash scene. Later estimates on the icing level at the time of the incident put it between 13,000 ft. with some encounters lasting as high as 17,000 ft. There were numerous reports of icing in the hours before and after the accident although the NTSB report states severe icing was not forecasted.

The pilot self-filed an instrument flight plan through the Direct User Access Terminal System (DUATS) in which he listed a cruising speed of 292 kt. and an altitude of FL 260. No record could be found of him receiving a verbal forecast. A clearance was issued at 0930 through Teterboro Airport Clearance Delivery at which point the turboprop began to taxi. The pilot contacted ground and taxied to KTEB’s Runway 6, and at 0948, he reported that he was ready to depart. According to air traffic control (ATC) recorded communications, weather information was not requested by, nor issued to, the pilot.

A captain departing out of nearby LaGuardia Airport (KLGA) in an MD-80 about an hour ahead of the accident aircraft told the NTSB he started picking up ice immediately upon entering the layer and that by the time the aircraft came out the top he estimated there was a 1-1.5-in. cone, 0.75 in. in diameter built up on the wiper blade locknut. He did not recall how thick the cloud layer was but estimated it was only a few thousand feet. His first officer stated that it was the heaviest ice he had ever seen. The captain said he had seen worse but would definitely not have wanted to stay in it for long.

An urgent pilot report was received at 0808 from a flight crew operating an MD-83 at 14,000 ft. over New Jersey’s Morristown Municipal Airport (KMMU). The pilot reported moderate to severe rime icing between 14,000 and 16,500 ft. One of the flight crewmembers reported the icing was the worst he had seen in 38 years of flying and he had never seen ice accumulate so quickly. He described “golf ball-sized” accumulation on the windshield wiper.

An interview with the captain of a Bombardier CRJ that was operating close to the accident aircraft reported that the wing anti-ice system could not “keep up” with the accumulation. He estimated 2.5 in. of ice on the protected areas of the wing, and a 4-in. accumulation on some unprotected areas in about 5 min.

During climb-1wout, while passing 8,000 ft. for 10,000 ft., the TBM pilot was directed to climb and maintain 14,000 ft. The controller then advised the pilot of moderate rime icing from 15,000 ft. through 17,000 ft., with light rime ice at 14,000 ft. The controller asked that the pilot advise him if the icing got worse, and the pilot responded, “We’ll let you know what happens when we get in there and if we could go straight through, it’s no problem for us.” At 0958:24, the controller directed the pilot to climb and maintain 17,000 ft. and to contact New York Center (ZNY). While climbing between 12,800 and 12,900 ft., at 116 kt. ground speed, the pilot acknowledged and advised that they were entering instrument meteorological conditions (IMC).

At 1002:17, the ZNY controller advised the pilot that he would be cleared to a higher altitude when ATC could provide it, and that light icing would be encountered at 17,000 ft. The pilot responded, “I can confirm that light icing . . .” and added that, “. . . light icing has been present for a little while and a higher altitude would be great.” The altitude of the airplane at that time was at 16,800 ft. and its ground speed was 101 kt.

At 1002:34, the TBM pilot reported, “We’re getting a little rattle here. Can we ah get ah higher as soon as possible please?” The ZNY controller responded with “stand by” and coordinated for a higher altitude with an adjacent sector controller.

At 1002:59, the ZNY controller directed the pilot to climb and maintain FL 200 and the pilot acknowledged. At 1004:08, the airplane reached an altitude of 17,800 ft. before it turned about 70 deg. to the left and entered a descent. At 1004:29, while descending through 17,400 ft., and at 90 kt. ground speed, the pilot transmitted, “and N731CA’s declaring . . .” No subsequent radio transmissions were heard from the pilot.

The final radar return at 1005:17 was observed at an altitude of 2,000 ft., about 600 yd. west of the main wreckage impact site. The previous return, recorded 9 sec. earlier, indicated 6,200 ft.

Accident investigators interviewed numerous witnesses who observed the airplane during the accident sequence. A consistent observation was that the airplane descended at a rapid rate, and was trailing smoke. According to some, it was “corkscrewing like an airshow airplane.” At least five witnesses saw pieces of the airplane separate during flight or they observed the airplane descending without a wing attached.

The NTSB determined the probable causes of the accident to be: the airplane’s encounter with un-forecasted severe icing conditions that were characterized by high ice accretion rates, and the pilot’s failure to use his command authority to depart the icing conditions in an expeditious manner, which resulted in a loss of airplane control.

Five years, two seasons later and an ocean away, conditions repeated, but the outcome, thankfully, did not.

While departing Edinburgh Airport in Scotland on June 5, 2017, a Saab 340 regional turboprop (G-LGNB) with 33 passengers and three crew aboard encountered severe icing and turbulence. The flight had launched at 1402 for Sumburgh Airport, 250 nm north in the Shetland Islands.

Fully equipped for all-weather operations, the aircraft had wing and stabilizer deicing, engine and propeller deicing, and heating for the windshield, outside air temperature probe and angle of attack (AOA) sensor. Bleed air goes to the wing and stabilizer boots and the engine intake while electrical power activates the remainder of the ice protection system.

In the climb, the aircraft started pitching up and down as the autopilot, which was engaged, attempted to maintain the selected IAS. Suddenly, the turbulence intensified and ice began to form quickly. The stick shaker activated and the autopilot disconnected.

The copilot, who was the pilot flying (PF), attempted to accelerate the aircraft by reducing the pitch attitude. He then reengaged the autopilot but after 13 sec., the stick shaker activated again and the autopilot again disconnected. Shortly afterward, the stick shaker activated for a third time and the copilot began a descent to accelerate the aircraft. The Saab lost around 500 ft. in the maneuver, during which it accelerated and recovered to normal flight. The crew did not select maximum continuous power during the recovery.

From a review of the flight data recorder, it was seen that as the aircraft climbed through FL 100, perturbations of normal acceleration started to increase, consistent with the aircraft encountering light turbulence. The OAT was -5°C. The turbulence then continued to increase in intensity, with variations in AOA that closely correlated with changes in load factor, pitch attitude and airspeed; the average airspeed was 162 KIAS at that time. The aircraft then briefly leveled off at FL 103 before climbing again.

Shortly afterward, the AOA increased rapidly over 1 sec. from just over 0 deg. to a recorded value of 5.3 deg., which coincided with the autopilot disconnecting; the airspeed was 160 KIAS and the pitch attitude was 6.3 deg. nose-up. The disconnect of the autopilot meant that the peak AOA value was in excess of the 5.3 deg. recorded as the system, in icing, disconnects the autopilot at an indicated AOA of 5.9 deg. The pitch attitude then reduced quickly to 2.8 deg. nose-up, before increasing to 13 deg. nose-up in 4 sec. During this period, the AOA varied rapidly, reducing to a minimum of -7 deg.

The autopilot was then re-engaged and the pitch attitude reduced to about 2 deg. nose-up. The airspeed reduced to 149 KIAS, after which it started to increase toward 160 KIAS. Thirteen seconds later, the aircraft pitched up quickly to 5.6deg.; the recorded AOA also increased rapidly, reaching 6 deg., and the autopilot disconnected once again; the airspeed was 159 KIAS.

The aircraft then briefly leveled off, during which the pitch and AOA both increased rapidly again, with the recorded AOA peaking at 6.3 deg. The aircraft then descended about 500 ft. to FL 105, during which time the airspeed progressively increased to about 190 KIAS. At no time during this period of the flight was engine power increased to the “maximum continuous” setting. The autopilot was then reengaged, which coincided with a reduction in turbulence, and the aircraft climbed to its cruise altitude of FL 170.

The conditions at FL 130 were capable of supporting the formation of severe icing. Analysis by the manufacturer concluded that the aircraft was affected by a large increase in aerodynamic drag. This could have been due to ice or downdrafts or a combination of the two. Even if the downdrafts had been twice as great as the forecast calculated, the increase in aerodynamic drag indicated that the aircraft had probably encountered severe icing conditions.

Saab also noted that the ice ridges, which the pilots saw behind the inflatable area of the deicing boots, corroborated the presence of severe icing. The copilot observed this about the time the airframe vibration began, but he did not discuss it with the commander because they were in the process of descending, to vacate the icing conditions. Also, the operating manual did not state that this was an indication of severe icing.

The evidence suggested severe icing conditions were encountered by the time the aircraft climbed past FL 125 and that it remained in severe icing after leveling. The operating manual states maximum continuous power should be used if ice accumulation due to “extreme icing conditions” causes a “large impact on performance” and the IAS decreases toward the minimum safe speed.

The aircraft has a stick-shaker channel for each control column and their vibration is reinforced with an aural warning in the form of a continuous clacker. At the same time as the stick shaker activates, the autopilot disengages. If sufficient action is not taken after the shaker and aural warnings are triggered, the stick push system provides a forward movement of the control columns to pitch the aircraft to a slightly nose-down attitude. If the stick push activates, visual warnings on the central warning panel and on the instrument panels also illuminate. The stall warning is generated by a combination of AOA, flap position and information from the wing anti-ice system.

A modification to the stall-warning computer adjusted the logic of the stick shaker and introduced the ice-speed system. This increased the stall-warning speed trigger levels to compensate for possible ice accretion on the wings. Notably, the triggering AOA for the stick shaker activation was lowered to 5.9 deg. from 12.1., but the stick-push logic remained unchanged. This ice-speed function is activated by switching on the engine anti-ice system. It remains activated even when the engine anti-ice system is selected off because a separate ice-speed switch also must be selected off. The engine anti-ice system must remain on for 5 min. after exiting icing conditions.

The Saab was climbing in IAS mode in which the flight control computers adjust the pitch attitude of the aircraft in order to maintain the selected IAS. The mode was engaged with an IAS of 163 kt. The stick shaker was triggered three times by the aircraft AOA reaching 5.9 deg. The aircraft was in turbulence with its pitch and IAS varying. Meteorological reports showed the weather in both Edinburgh and Sumburgh was affected by the presence of an occluded front lying just to the north of Edinburgh, moving eastward. There also was a warm front lying parallel to the occlusion. Visibility outside the cloud was good but there were isolated moderate or heavy showers. There also were isolated, embedded cumulonimbus clouds with bases from 1,500 to 3,000 ft. ASL and tops above 10,000 ft. Freezing levels were between 4,000 and 5,000 ft.

In reviewing the encounter and actions taken, the UK’s Air Accidents Investigation Branch determined that the aircraft’s stall warning system “functioned as it was designed,” and that the crew “did not initially address the problem sufficiently.” However, after the third activation of the stick shaker, the pilots “descended the aircraft to regain a safe airspeed” and were able to “clear the icing and turbulent conditions” and continue on to Sumburgh without further incident.