NTSB Factual Report

HISTORY OF FLIGHT

On January 2, 2006, at 1439 Pacific standard time, a Saab-Scania AB SF340B+, N390AE, operated by American Eagle Airlines, Inc., as flight 3008, encountered icing conditions during the en route climb over Santa Maria, California. The airplane was at 11,700 feet mean sea level (msl) when it departed controlled flight, and descended to an altitude of about 6,500 feet msl. The pilots recovered control of the airplane, and continued to their scheduled destination of Los Angeles International Airport (LAX), Los Angeles California, where they landed at 1540 without further incident. The 2 flight crew members, 1 flight attendant, and 25 passengers were uninjured, and the airplane did not sustain substantial damage. Instrument meteorological conditions prevailed for the 14 Code of Federal Regulations (CFR) Part 121 scheduled domestic passenger flight that was operating on an instrument flight rules (IFR) flight plan. The flight originated from San Luis County Regional Airport (SBP), San Luis Obispo, California, at 1414.

A review of the American Eagle Airlines, Inc., flight log disclosed that the accident flight was scheduled to be the flight crew's fifth trip of the day, and their second trip in the accident airplane. The 1 hour 26 minute accident flight was scheduled to depart from Santa Barbara at 1408, and terminate in Los Angeles at 1513.

The flight crew members stated in post incident interviews that before the incident flight, they had encountered light rime icing and moderate turbulence on the inbound leg to San Luis Obispo as they were descending from 9,000 to 5,000 feet. The pilots stated that, while preparing for the return flight to Los Angeles, they reviewed the weather conditions for the intended route of flight. The operator's dispatch package noted two AIRMET (airmen's meteorological information) reports for icing in clouds and two PIREPs (pilot weather reports) for turbulence. The pilots discussed the conditions that they had encountered on the way in, as well as the conditions for the intended route of flight outbound. Because of the gusty wind conditions and the short runway at San Luis Obispo, the captain decided to perform the departure. He was going to turn control of the airplane over to the first officer after completing the climb checklist at the acceleration altitude.

In accordance with American Eagle's minimum equipment list (MEL), the incident airplane was dispatched with the continuous mode of the boot deice system inoperable for the inbound flight and the return incident flight. The flight crew reported that they performed the manual test of the deicer boots as called for in the MEL, and observed the operation of the inboard and outboard wing boot segments. However, they could not see the stabilizer segments, and did not have qualified ground personnel available to observe the test as required by the MEL. The pilots reported that they did confirm illumination of the green boot inflation lights on the overhead panel when they pressed the manual buttons.

In accordance with company procedures that require flight crews to activate the deice system at the first sign of ice accretion and operate the deice boots continually, the crew stated that they were prepared to operate the deice boots in manual mode as needed during the flight. They departed in level 2 weather conditions (defined as 10 degree Celsius or colder with visible moisture) and with the engine anti-ice on.

The pilots stated that the weather radar was on, and they did not observe any activity on it. The captain had the autopilot engaged in the medium (M) climb mode. Shortly after taking the controls about 2,500 feet, the first officer changed the autopilot to vertical speed (VS) mode, which gave pitch attitude commands to maintain the vertical speed existing at the time of mode engagement.

As the airplane climbed through 11,000 feet, the captain noted light rime ice accumulating on the windshield wiper blades and about a 1/2-inch-wide area of ice on the left wing.

The captain reported that, as he began to reach up to activate the manual deice boot system, he felt a heavy vibration in the airframe. He said that the windscreen immediately turned white. Immediately thereafter, the airplane's nose dropped, the left wing dropped, and the autopilot disconnected. He grabbed the yoke to take control of the airplane. He said that the clacker sounded (indicating an imminent stall), the stick shaker activated, and the ground proximity warning system emitted a "bank angle" aural warning.

The flight crew reported that the airplane vibrated again, but less violently than the first episode. The captain leveled the wings, and began pulling up on the control yoke. At this point, he instructed the first officer to manually operate the deice boots. The captain stated that he pushed the condition levers to the maximum position, and brought the power levers to idle. The airplane stabilized in roll, and he could hear chunks of ice shedding off and hitting the fuselage. He kept the airplane in a nose-down attitude, maintaining a 500 feet per minute rate of descent until the airplane was below the freezing level.

PERSONNEL INFORMATION

Captain

The operator reported that the 34-year-old captain held an airline transport pilot (ATP) certificate with a rating for airplane multi-engine land. He held a commercial pilot certificate with ratings for airplane single-engine land and instrument airplane. He had a type rating in the SF340.

The captain held a first-class medical certificate issued on September 21, 2005. It had no limitations or waivers.

The captain had a total flight time of 6,764.08 hours, with 3,981.87 hours accumulated in Saab 340 airplanes, of which 2,519.46 hours was as the pilot-in-command (PIC). He had a total of 970 hours of instrument experience and between 1,700 and 1,900 hours of night flight. During the preceding 90 days, 30 days, and 24 hours, he reported that he had flown in both the capacity of PIC and second-in-command (SIC) approximately 172, 47, and 7 hours, respectively. He added that he had acquired numerous hours of aerobatic flight time in a Cessna 150 Aerobat airplane.

First Officer (FO)

The operator reported that the 32-year-old FO held a commercial pilot certificate with airplane instrument and multi-engine land ratings. He additionally held an SF340 Type Rating, with the limitations of SF340 SIC privileges only and circling approaches to be completed only in visual meteorological conditions (VMC). The FO was also a certified flight instructor (CFI) for instrument and multi-engine land airplane.

The FO's second-class medical certificate was issued on May 25, 2005, with the limitation that he must wear corrective lenses.

The FO had a total flight time of 1,367.48 hours, with 132.48 hours accumulated in Saab 340 airplanes. He had a total of 94 hours of instrument experience and 185 hours of night flight. During the preceding 90 days, 30 days, and 24 hours, he reported that he had flown approximately 120, 71, and 5 hours, respectively.

AIRCRAFT INFORMATION

The airplane was a Saab SF340B+, serial number 340B-390. The airplane had a total airframe time of 17,291 hours at the examination following the incident.

Systems

National Transportation Safety Board investigators reviewed the airplane's maintenance records and logbooks. The day prior to the incident, a flight crew reported that during an en route deice boot check, the timer light illuminated. The deicer timer failure light was later deferred in accordance with the operator's MEL. The deferral procedures required a placard to be placed adjacent to the deicer timer switch and the auto cycling switch to remain in the "off" position. Investigators did observe an MEL placard (sticker) next to the deice system controls in the cockpit.

Initial examinations revealed that the airplane's deice systems were operational; the deicer timer failure light illuminated.

Aileron Interconnect

The airplane was equipped with an Aileron Spring Unit. This would allow the flight crew to initially maintain authority in the roll axis if one aileron seized, until the aileron disconnect handle in the cockpit was pulled by a pilot. In the event an aileron seized, the pilot would have to overpower a preloaded spring unit to manipulate an aileron. When the pilot reduced control input pressure, the unit would close allowing the ailerons to be normally coupled. The FO stated that he believed that he did not have his hands on the controls after the captain assumed authority. The roll disconnect handle was not pulled during this incident.

WEATHER

A Safety Board meteorologist prepared a factual report, which is part of the public docket. AIRMET Zulu Update 4 for icing (SFOZ WA 022045) was issued on January 2, 2006, at 1345, and valid until 2000. It noted occasional moderate rime/mixed icing in clouds and in precipitation between the freezing level and FL220. The freezing level in central California was 6,000 to 8,000 feet; the freezing level in southern California was 7,000 to 11,000 feet.

The specialist reviewed San Joaquin Valley, California (HNX) Level II Doppler weather radar Base Reflectivity Images. At 1437:38 at the location of the icing encounter, the HNX beam center was about 16,500 feet with a beam width of about 8,000 feet. The top of the beam was about 20,500 feet, and the bottom of the beam was about 12,500 feet.

At 1442:36 at the location of the icing encounter, the HNX beam center was about 7,500 feet with a beam width of about 8,000 feet. The top of the beam was about 11,500 feet, and the bottom of the beam was about 3,500 feet.

A GOES-10 infrared image at 1441 PST at the location of the icing incident recorded a radiative temperature of 244 degrees K (-29 degrees C). Using NAM12 upper air data, this temperature corresponded to a cloud top of about 21,000 feet.

The report contained experimental Current Icing Potential (CIP) plots that a scientist at the National Center for Atmospheric Research in Boulder, Colorado, provided. It noted that the CIP product (Supercooled Liquid Droplets (SLD) and Ice) combines sensor and numerical model data to provide a three-dimensional diagnosis of the icing environment. The current CIP output consists of a likelihood field ranging from 0 (no icing) to 100 (certain icing). While this is not yet calibrated as a true probability value, CIP has value in pointing out real differences in the likelihood of encountering icing at a given location.

The plots were: icing severity category composite, maximum SLD potential in the column, maximum potential in column for experiencing icing field, icing severity at 12,000 feet, potential for SLD ice at 12,000 feet, potential for experiencing ice at 12,000 feet, icing severity at 9,000 feet, potential for experiencing ice at 9,000 feet, potential for SLD ice at 9,000 feet, and current icing potential.

DIGITAL FLIGHT DATA RECORDER (DFDR)

A Safety Board specialist examined the DFDR data, and the factual report is part of the public docket.

About 6 minutes after takeoff, the airplane was passing through 9,200 feet. The airspeed began to decline from 180 knots, and the pitch angle began to increase. Around 2 minutes later, at 1439:36, the pitch of the airplane was 14 degrees up and the roll was neutral. One second later, the altitude reached a maximum recorded value of 11,712 feet, and the airplane was in a 16-degree left roll. During the next second, the autopilot disconnected, and the airspeed registered 118 knots.

The rate of airspeed decay accelerated in the final 10 seconds before the autopilot disconnected. The airplane departed controlled flight at an airspeed of 130 knots indicated airspeed (KIAS), and before the stall warning activated. The DFDR data also revealed that about 26 seconds before the stall roll departure, while the airplane was at a speed of 144 KIAS, the airplane began to experience a slight rolling anomaly that was counter to the direction of the aileron input. Aileron input from the autopilot arrested this slight rolling motion.

The airplane rolled to 86 degrees left wing down, and then went through a series of roll and pitch movements. It reached 140 degrees of right roll, and a maximum pitch down angle of 48 degrees. It rolled to 75 degrees left wing down, and a pitch of 31 degrees nose down. It then rolled to 94 degrees right wing down, followed by a pitch angle to 40 degrees nose down. Starting at 1440, the altitude and outside air temperature parameters stopped recording valid data for a period of 15 seconds. At 1440:06, the airplane's pitch angle began to increase. It passed through 0 degrees about 6 seconds later at an airspeed of 219 knots, and a recorded maximum vertical acceleration of 2.5 g's. The pitch reached 23 degrees nose up at 1440:24; the minimum recorded airspeed value of 105 knots occurred 11 seconds later while the airplane was at an altitude of 7,840 feet. The parameters began to stabilize after this time.

The DFDR data disclosed that 14 seconds after the initial stall, both ailerons simultaneously traveled to the full up position for approximately 14 seconds.

DFDR Study

The DFDR specialist participated in a ground test to validate the control wheel and aileron position data recorded on the event flight and gather additional data related to a breakout scenario. The test did demonstrate that, under a normal scenario, the control wheel being manipulated would lead the control wheel not being manipulated. Also, it showed that moving the control wheel would drive the ailerons to their maximum range values, but moving the ailerons would not drive the control wheels to full range. Additionally, in manually manipulating the ailerons, the left one did not reach its full upwards range.

TESTS AND RESEARCH

Performance Study

A Safety Board specialist conducted a vehicle performance study, which is a part of the public docket. It showed that the airplane's aerodynamics degraded with time until the airplane stalled. This stall occurred at a lower angle of attack than would be expected for an uncontaminated airframe. The study concluded that the aerodynamic degradation and early stall was consistent with airframe icing.

ADDITIONAL INFORMATION

Training

American Eagle had a recurrent and requalification simulator training syllabus for captains and first officers. It included approach to stalls in the takeoff, clean, and landing configurations. It also included unusual attitudes recoveries from nose low and nose high positions. One section dealt with normal and abnormal emergency situations operations including anti-icing and deicing systems, stall warning, and stick pusher.

American Eagle's advanced aircraft maneuvering program (AAMP) includes a review of phenomena that cause upset events and unusual attitude recovery procedures.

The Operations Group chairperson interviewed several American Eagle pilots. All the pilots interviewed, including the incident captain and first officer, stated that they never practiced encountering a stall in icing conditions as part of their simulator training. Additionally, the pilots could not recall ever having the opportunity to practice a complete stall in the simulator, as they were always instructed to recover at the first indication of an impending stall.

Minimum Airspeeds for Flight in Icing Conditions

American Eagle's 340B+ Airplane Operating manual (AOM) calls for flight crews to compute a final clean airplane climb speed, or Vcln, and to add 15 knots to that value to determine the minimum speed (Vcln+15) in icing conditions. For the incident flight, Vcln was computed to be 126 knots indicated airspeed (KIAS), and the minimum speed in icing conditions was 141 KIAS.

Use of Autopilot in Icing Conditions

The limitations section of the American Eagle 340B+ AOM stipulates that the indicated airspeed (IAS) mode is the only authorized flight director/autopilot mode if an airplane is climbing when ice accretion is occurring, or with residual ice on the airframe. In IAS mode, the flight control computer gives pitch attitude commands to maintain the indicated airspeed existing at the time of mode engagement. In the vertical speed (VS) mo

NTSB Probable Cause

An in-flight loss of control due to ice accreted in supercooled liquid droplets (SLD) conditions, and the flight crew's failure to maintain the specified minimum airspeed in icing conditions. Contributing to the accident was the flight crew's decision to climb the airplane in known and forecast icing conditions using an autopilot mode contrary to that specified in the operator's Airplane Operations Manual for climbs during periods of ice accretion or when ice was present on the airframe.