AT73, en-route, Roselawn IN USA, 1994

AT73, en-route, Roselawn IN USA, 1994


On 31 October 1994, an ATR 72 exited controlled flight after a flap retraction when descending through 9000 feet was followed by autopilot disconnect and rapid and very large un-commanded roll inputs from which recovery, not within the scope of received crew training, was not achieved. The investigation found this roll upset had been due to a sudden and unexpected aileron hinge moment reversal after ice accretion on the upper wings aft of the leading edge pneumatic de-icing boots during earlier holding in icing conditions which had been - unknown to the crew - outside the icing certification envelope.

Event Details
Event Type
Flight Conditions
Flight Details
Type of Flight
Public Transport (Passenger)
Take-off Commenced
Flight Airborne
Flight Completed
Phase of Flight
Roselawn, Newton County, IN
Inadequate Airworthiness Procedures, Ineffective Regulatory Oversight
Airframe Structural Failure, Uncommanded AP disconnect, Environmental Factors, Extreme Bank, Extreme Pitch
In Flight Airframe Icing
Damage or injury
Aircraft damage
Hull loss
Non-aircraft damage
Non-occupant Casualties
Occupant Fatalities
Most or all occupants
Number of Occupant Fatalities
Off Airport Landing
Causal Factor Group(s)
Aircraft Technical
Safety Recommendation(s)
Aircraft Operation
Aircraft Airworthiness
Air Traffic Management
Investigation Type


On 31 October 1994, an ATR 72-212 (N401-AM) being operated by Simmons Airlines (a wholly owned subsidiary of American Airlines' holding company AMR Corporation) on a scheduled domestic passenger flight from Indianapolis to Chicago O'Hare as American Eagle 4184 was descending through 9100 feet QNH in day Instrument Meteorological Conditions (IMC) and experienced a sudden, severe and sustained roll upset from which the crew were unable to recover. Uncontrolled descent ended in ground impact at an extremely high rate of descent which completely destroyed the aircraft and killed all 68 occupants.


An Investigation was carried out by the NTSB. The FDR and CVR were both recovered from the wreckage and their data were successfully downloaded. Examination of the wreckage indicated that the aircraft had impacted open ground partially inverted and in a nose down left wing low attitude. The disposition of the wreckage in combination with the available recorded data indicated that an EGPWS 'PULL UP' warning had been annunciated and that both pilots had seen the ground shortly before impact. In response to their recognition of close proximity to the ground and the high rate of descent of the aircraft, they had "made a nose-up elevator input that, combined with the high airspeed (about 115 KIAS above Vmo) had resulted in excessive wing loading and structural failure of the outboard sections of the wings".

It was established that during the prior 39 minute hold at 10,000 feet, the crew had selected 15º flap in response to the abnormally high pitch attitude which prevailed and that this had resulted in a reduction in the Angle of Attack (AOA) from 5º to approximately 0º - although this flight parameter was not displayed to the crew. During the hold, which had been in IMC at a marginally sub-zero temperature, airframe de-icing had been selected on and the pilots had made periodic checks on the extent of visible ice accretion on the wings without observing anything notable in that respect. However, it was concluded that although there had been no prior forecast of abnormally significant icing conditions or PIREPS mentioning this from other aircraft in the general area, the accident aircraft had, whilst in the hold, been accreting ice to the upper wing surface aft of the area kept clear by the leading edge pneumatic de-icing boots. The general difficulty in assessing the relevance of the presence or absence of PIREPS from other aircraft types was noted.

Soon after clearance to descend had been received and descent had been commenced, the flaps were retracted to zero. This action was followed immediately by un-commanded disconnect of the AP and the beginning of a series of un-commanded rolls to both right and left. A very high rate of descent developed and impact occurred just over 20 seconds after the AP disconnect. During this period out of control, it was found from the recorded data that the aircraft had completed one full 360º roll to the right and otherwise remained unstable in roll throughout.

It was concluded that slight unevenness in the ice accretion on the wing upper surfaces aft of the leading edge and ahead of ailerons towards each wing tip had, when the flaps were retracted and AoA had consequently increased, caused differential airflow separation and disturbed the normally balanced aileron hinge moments so as to cause sudden self-deflection.

The Captain and First Officer had gained their first airline employment as Simmons' Airlines First Officers in 1987 and 1989 respectively, with the Captain having been promoted to that rank whilst flying the Shorts 360 and transferring with it to the ATR72 eighteen months prior to the accident. The Captain had accumulated a total of 7867 flying hours including 1548 hours on the ATR72 and the First Officer had accumulated a total of 5176 flying hours including 3657 hours on the ATR72. The First Officer had been PF for the accident flight.

Subsequent to the accident, the NTSB recommended that the FAA conduct a Special Certification Review (SCR) of the ATR 42 and ATR 72 aircraft and that "flight tests and/or wind tunnel tests be conducted as part of that review to determine the aileron hinge moment characteristics of the airplanes while operating at different airspeeds and in different configurations during ice accumulation and with varying AoA following ice accretion". The conclusions of this review were that both aircraft types had been correctly certificated in accordance with 14 CFR parts 21 and 25 and JAR 25 but that the accident aircraft exposure to icing appeared to have included Supercooled Water Droplets (SWD) in the size range of about 40 to 400, droplets defined by the SCR team as Supercooled Drizzle Drops (SCDD). The team considered that "while the physics of formation are not the same, freezing drizzle and SCDD can be considered to present the same icing threat in terms of adverse effects" with the difference between them being only that the description "freezing drizzle" refers to precipitation at the surface whereas SCDD "is found aloft with air at temperatures above freezing underneath". It was concluded that "while the effects of ice accreted in SLD may be severe, the clouds that produce them tend to be localized in horizontal and/or vertical extent".

The Investigation noted that Freezing Rain/Drizzle - and therefore their in-cloud SWD equivalents, occur in circumstances not covered by the Part 25 'Appendix C' Icing Envelope and observed that "no airplane should be authorized or certified for flight into icing conditions more severe than those to which the airplane was subjected in certification testing, unless the manufacturer can otherwise demonstrate the safety of flight in such conditions". It noted that "had the FAA acted more positively upon" NTSB Safety Recommendations issued to the Federal Coordinator for Meteorological Services and Supporting Research in their 1981 Report on 'Aircraft Icing Avoidance and Protection', the accident "may not have occurred".

It was concluded that ice-induced aileron hinge moment reversals, autopilot disconnects and rapid, un-commanded rolls can occur if ATR 42 or 72 aircraft are operated in the presence of SWD with the Median Volume Diameter (MVD) typical of freezing drizzle. Whilst the 60 lb force on the control column which would have been required to maintain a wings-level attitude at the initiation of the aileron hinge moment reversal which affected the accident aircraft was within FAA requirements, the rapid, un-commanded rolls and the sudden onset of a requirement for such large corrective forces without any prior warning for the pilot or, in this case, any training for such unusual events, would not be conducive to a timely recovery.

It was noted that "ATR is considering design changes to the lateral control system for current and future ATR airplanes that will reduce the susceptibility to flow separation-induced aileron hinge moment reversals" and that any such changes "could minimize the reliance on the changes to flight operations and pilot training that have already been mandated".

It was also found that both the French DGAC and the FAA had failed to require ATR "to provide documentation of known undesirable post-Stall Protection System (SPS) flight characteristics" and that this had "contributed to their failure to identify and correct, or otherwise properly address" the abnormal aileron behaviour now evidenced "early in the history of ATR icing incidents" of which a number were identified by the Investigation.

Other Findings included the following:

  • Because the DGAC did not require ATR, and ATR did not provide to the operators of its airplanes, information that specifically alerted flight crews to the fact that encounters with freezing rain could result in sudden autopilot disconnects, aileron hinge moment reversals, and rapid roll excursions, or guidance on how to cope with these events, the crew of the accident flight had no reason to expect that the icing conditions they were encountering would cause the sudden onset of an aileron hinge moment reversal, autopilot disconnect and loss of aileron control.
  • Had ice accumulated on the wing leading edges so as to burden the ice protection system, or if the crew had been able to observe the ridge of ice building behind the deice boots or otherwise been provided with a means of determining that an unsafe condition was developing from holding in those icing conditions, it is probable that the crew would have exited the conditions.

The Probable Cause of the Accident was finally determined as:

"The loss of control, attributed to a sudden and unexpected aileron hinge moment reversal, that occurred after a ridge of ice accreted beyond the deice boots while the airplane was in a holding pattern during which it intermittently encountered supercooled cloud and drizzle/rain drops, the size and water content of which exceeded those described in the icing certification envelope. The airplane was susceptible to this loss of control and the crew was unable to recover."

Five Contributory Factors were also identified as follows:

  1. the inadequate oversight of the ATR 42 and 72 by the French Directorate General for Civil Aviation (DGAC) and its failure to take the necessary corrective action to ensure continued airworthiness in icing conditions
  2. the failure of the DGAC to provide the FAA with timely airworthiness information developed from previous ATR incidents and accidents in icing conditions
  3. the failure of the Federal Aviation Administration (FAA) to ensure that aircraft icing certification requirements, operational requirements for flight into icing conditions, and FAA-published aircraft icing information adequately accounted for the hazards that can result from flight in freezing rain
  4. the inadequate oversight of the ATR 42 and 72 by the FAA (to the extent necessary) to ensure continued airworthiness in icing conditions
  5. the inadequate response of ATR to the continued occurrence of ATR 42 icing/roll upsets which, in conjunction with information learned about aileron control difficulties during the certification and development of the ATR 42 and 72, should have prompted additional research and the creation of updated airplane flight manuals, flight crew operating manuals and training programs related to operation of the ATR 42 and 72 in such icing conditions.

During the course of the Investigation, on 7 November 1994, the following 5 Safety Recommendations were made:

  • that the FAA should conduct a special certification review of the ATR 42 and ATR 72 airplanes, including flight tests and/or wind tunnel tests, to determine the aileron hinge moment characteristics of the airplanes operating with different airspeeds and configurations during ice accumulation and with varying angles of attack following ice accretion. As a result of the review, require modifications as necessary to assure satisfactory flying qualities and control system stability in icing conditions. (A-94-181)
  • that the FAA should prohibit the intentional operation of ATR 42 and ATR 72 airplane in known or reported icing conditions until the effect of upper wing surface ice on the flying qualities and aileron hinge moment characteristics are examined further as recommended in A-94-181 and it is determined that the airplane exhibits satisfactory flight characteristics. (A-94-182)
  • that the FAA should issue a general notice to ATC personnel to provide expedited service to ATR 42 and ATR 72 pilots who request route, altitude or airspeed deviations to avoid icing conditions. Waive the 175 knot holding speed restriction for ATR 42 and ATR 72 airplanes pending acceptable outcome of the special certification effort. (A-94-183)
  • that the FAA should provide guidance and direction to pilots of ATR 42 and ATR 72 airplanes in the event of inadvertent encounter with icing conditions by the following actions: (1) define optimum airplane configuration and speed information; (2) prohibit the use of autopilot; (3) require the monitoring of lateral control forces; (4) and define a positive procedure for reducing angle of attack. (A-94-184)
  • that the FAA should caution pilots of ATR 42 and ATR 72 airplanes that rapid descents at low altitude or during landing approaches or other deviations from prescribed operating procedures are not an acceptable means of minimizing exposure to icing conditions. (A-94-185)

Later during Investigation, on 6 November 1995, the following 4 Safety Recommendations were made:

  • that the FAA should require the Air Traffic Control System Command Center to retain all flow control-related facility documents for 15 days, regardless of title, name or form number, for reconstruction purposes. (A-95-103)
  • that the FAA should develop a list of documents to be completed by the Air Traffic Control System Command Center personnel in the event of an incident or accident. (A-95-104)
  • that the FAA should revise Order 8020.11, “Aircraft Accident and Incident Notification, Investigation and Reporting,” to include the Air Traffic Control System Command Center (DCC) facility. Ensure that the DCC facility is assigned specific requirements to be included in an accident/incident package. (A-95-105)
  • that the FAA should revise FAA Order 7210.3, “Facility Operation and Administration,” Chapter 3, “Facility Equipment,” Section 4, “Recorders,” paragraph 3-41, “Assignment of Recorder Channels,” to include the Air Traffic Control System Command Center facility, listing the recorded positions and their priority. (A-95-106)

A total of 26 further Safety Recommendation were made upon the conclusion of the Investigation as follows:

  • that the FAA should direct Principal Operations Inspectors (POIs) to ensure that all 14 Code of Federal Regulations (CFR) Part 121 air carriers require their dispatchers to provide all pertinent information, including airman’s meteorological information (AIRMETs) and Center Weather Advisories (CWAs), to flight crews for pre-flight and in-flight planning purposes. (A-96-48)
  • that the FAA should require that Hazardous In-flight Weather Advisory Service (HIWAS) broadcasts consistently include all pertinent information contained in weather reports and forecasts, including in-flight weather advisories, airman’ s meteorological information (AIRMETs), significant meteorological information (SIGMETs), and Center Weather Advisories (CWA’s). (A-96-49)
  • that the FAA should encourage Principal Operations Inspectors (POIs) and operators to re-emphasize to pilots that Hazardous In-flight Weather Advisory Service (HIWAS) is a source of timely weather information and should be used whenever they are operating in or near areas of potentially hazardous weather conditions. (A-96-50)
  • that the FAA should revise the existing aircraft icing intensity reporting criteria (as defined in the Aeronautical Information Manual (AIM) and other Federal Aviation Administration (FAA) literature) by including nomenclature that is related to specific types of aircraft, and that is in logical agreement with existing Federal Aviation Regulations (FARs). (A-96-51)
  • that the FAA should publish the definition of the phrase “icing in precipitation” in the appropriate aeronautical publications, emphasizing that the condition may exist both near the ground and at altitude. (A-96-52)
  • that the FAA should continue to sponsor the development of methods to produce weather forecasts that both define specific locations of atmospheric icing conditions (including freezing drizzle and freezing rain) and produce short-range forecasts (“nowcasts”) that identify icing conditions for a specific geographic area with a valid time of 2 hours or less. (A-96-53)
  • that the FAA should revise the icing criteria published in 14 Code of Federal Regulations (CFR), Parts 23 and 25, in light of both recent research into aircraft ice accretion under varying conditions of liquid water content, drop size distribution, and temperature, and recent developments in both the design and use of aircraft. Also, expand the Appendix C icing certification envelope to include freezing drizzle/freezing rain and mixed water/ice crystal conditions, as necessary. (A-96-54) (this supersedes A-81-116 and A-81-118)
  • that the FAA should revise the Federal Aviation Regulations (FAILs) icing certification requirements and advisory material to specify the numerical methods to be used in determining median volumetric diameter (MVD) and liquid water content (LWC) during certification tests. (A-96-55)
  • that the FAA should revise the icing certification testing regulation to ensure that airplanes are properly tested for all conditions in which they are authorized to operate, or are otherwise shown to be capable of safe flight into such conditions. If safe operations cannot be demonstrated by the manufacturer, operational limitations should be imposed to prohibit flight in such conditions and flight crews should be provided with the means to positively determine when they are in icing conditions that exceed the limits for aircraft certification. (A-96-56)
  • that the FAA should require all aircraft manufacturers to provide, as part of the certification criteria, information to the FAA and operators about any known undesirable characteristics of flight beyond the protected (stall system and related shaker/pusher) flight regime. (A-96-57)
  • that the FAA should develop an icing certification test procedure similar to the tail plane icing pushover test to determine the susceptibility of airplanes to aileron hinge moment reversals in the clean and iced-wing conditions. Revise 14 CFR Parts 23 and 25 icing certification requirements to include such a test. (A-96-58)
  • that the FAA should encourage ATR to test the newly developed lateral control system design changes and upon verification of the improved or corrected hinge moment reversal / un-commanded aileron deflection problem, require these design changes on all new and existing ATR airplanes. (A-96-59)
  • that the FAA should revise 14 CFR Parts 91.527 and 135.227 to ensure that the regulations are compatible with the published definition of severe icing, and to eliminate the implied authorization of flight into severe icing conditions for aircraft certified for flight in such conditions. (A-96-60)
  • that the FAA should require all principal operations inspectors (POIs) of 14 CFR Part 121 and 135 operators to ensure that training programs include information about all icing conditions, including flight into freezing drizzle/freezing rain conditions. (A-96-61)
  • that the FAA should develop an organizational structure and a communications system that will enable the Aircraft Evaluation Group (AEG) to obtain and record all domestic and foreign aircraft and parts/systems manufacturers’ reports and analyses concerning incidents and accidents involving aircraft types operated in the United States, and ensure that the information is collected in a timely manner for effective AEG monitoring of the continued airworthiness of aircraft. (A-96-62)
  • that the FAA should review and revise, as necessary, the manner in which the FAA monitors a foreign airworthiness authority’s compliance with U.S. type certification requirements under the Bilateral Airworthiness agreement (BAA). (A-96-63)
  • that the FAA should establish policies and procedures to ensure that all pertinent information is received, including the manufacturer’s analysis of incidents, accidents or other airworthiness issues, from the exporting country’s airworthiness authority so that the FAA can monitor and ensure the continued airworthiness of airplanes certified under the Bilateral Airworthiness Agreement (BAA). (A-96-64)
  • that the FAA should evaluate the need to require a sterile cockpit environment for airplanes holding in such weather conditions as icing and convective activity, regardless of altitude. (A-96-65)
  • that the FAA should amend the Federal Aviation Regulations to require operators to provide standardized training that adequately addresses the recovery from unusual events, including extreme flight attitudes in large, transport category airplanes. (A-96-66)
  • that the FAA should revise FAA Order 8400.10, Chapter 7, Section 2, paragraph 1423 (Operational Requirements - Flight crews) to specify that Center Weather Advisories (CWAs) be included and considered in the flight crew’s pre-flight planning process. (A-96-67)
  • that the FAA should revise FAA Order 7110.65, “Air Traffic Control,” Chapter 2, “General Control,” Section 6, “Weather Information,” paragraph 2-6-3, “PIREP” Information, to include freezing drizzle and freezing rain. Additionally, these conditions should be clearly defined in the Pilot/Controller Glossary. (A-96-68)
  • that the FAA should conduct or sponsor research and development of on-board aircraft ice protection and detection systems that will detect and alert flight crews when the airplane is encountering freezing drizzle and freezing rain and accreting resultant ice. (A-96-69)
  • that the National Oceanic and Atmospheric Administration should develop methods to produce weather forecasts that both define specific locations of atmospheric icing conditions (including freezing drizzle and freezing rain), and that produce short range forecasts (“nowcasts”) that identify icing conditions for a specific geographic area with a valid time of 2 hours or less. Ensure the timely dissemination of all significant findings to the aviation community in an appropriate manner. (A-96-70)
  • that American Eagle should require Dispatchers to include in the flight release airman’s meteorological information (AIRMETs) and center weather advisories (CWAs) that are pertinent to the route of flight so that flight crews can consider this information in their pre-flight and in-flight decisions. (A-96-71)
  • that American Eagle should encourage Captains to observe a “sterile cockpit” environment when an airplane is holding, regardless of altitude, in meteorological conditions such as convective areas or icing conditions, that have the potential to demand significant attention of a flight crew. (A-96-72)
  • that American Eagle should conduct a procedural audit to eliminate existing conflicts in guidance and procedures between the aircraft Flight Operations Manuals and other published material. (A-96-73)

The Final Report of the Investigation was adopted by the NTSB on 9 July 1996. The dissenting opinions of the French BEA on a wide range of issues in the Report including factual matters, the analysis made thereon and the Conclusions documented were issued as a separate document.

However, after publication of the Report, the NTSB received 'Petitions for Reconsideration' of it from the French DGAC (on 29 November 1996) and from the aircraft manufacturer ATR (on 22 September 1999) which made essentially similar points. Comments in support of the DGAC position from the FAA were communicated on 28 May 1998. After review of these petitions and the comments made in respect of them by the FAA and other qualified parties, the Board, "partially granted" them and on 13 September 2002 published their Response which concluded with the adoption by the Board of revisions to the text of 7 of the 43 original 'Findings', revisions to the text of the Probable Cause Statement and the removal of two paragraphs from Section 1 of the Report.

The Probable Cause and Contributory Factors reproduced above and the summary provided prior to this are respectively taken from /based upon the revised version of the Final Report. The changes adopted are detailed at the end of the 'Response' document referenced above and also on pages a1, a2 and a3 which have been added to the beginning of the as-originally-issued Final Report, which is otherwise unchanged i.e. the changes adopted have not been incorporated in it.

Related Articles

Further Reading

SKYbrary Partners:

Safety knowledge contributed by: