CF188747 Hornet - Epilogue - Flight Safety Investigation Report

Report / November 28, 2016 / Project number: CF188747 - A Category

Location: Cold Lake Air Weapons, Alberta
Date: 2016-11-28
Status: Investigation Complete

FSIR - download PDF version, 2.2MB

Epilogue

The pilot of aircraft CF188747, using the call sign “Swift 32”, was part of a two-ship formation led by “Swift 31” for an air-to-ground training mission. The mission objective was to practice level deliveries of two Mark 83 inert bombs followed by two laser guided training rounds, simulating laser guided bombs, in the Cold Lake Air Weapons Range. The plan was to ingress to the target and drop weapons from 600 feet above ground level. To avoid simulated bomb fragmentation after dropping their bombs each pilot would fly a “breakaway manoeuver” comprising a steep turn through 90 degrees of heading change.

Following his Mark 83 drop, Swift 32 manoeuvred his aircraft in a manner that was suggestive of a pilot attempting to visually spot his weapon impact, losing over 200 ft of altitude in the process. Swift 32 then assumed tactical lead, with Swift 31 flying about 2 miles in trail of Swift 32 and lasing the target for Swift 32, who then dropped his laser guided training round. The ingress to the target was flown at approximately 500 feet above ground level.

Immediately after dropping his laser guided training round Swift 32 initiated a steep left turn, reaching a maximum left bank angle of 118 degrees while pulling approximately 5g. The aircraft nose began to pitch towards and then below the horizon, eventually reaching a nose-down pitch angle of minus 17 degrees and concurrently generating a large descent rate.

About 1.5 seconds before impact the aircraft began rolling right. The bank angle had reduced to approximately 30 degrees left and the pitch angle increased to approximately minus 10 degrees when ground impact occurred. Swift 32 made no radio calls during the turn, did not eject and was fatally injured when the aircraft struck the ground in a descending left turn.

The available evidence did not support a mechanical failure, bird strike or pilot incapacitation scenario. Therefore, it appears that the pilot was capable of controlling the aircraft but did not adequately monitor the aircraft’s flight path while manoeuvring in the low level environment, and allowed the aircraft to enter an overbank situation and the nose to drop well below the horizon. A recovery may have been attempted at the last second but there was not enough altitude available to safely recover the aircraft. While the reason for this lack of flight path monitoring is not knowable with any certainty, circumstantial evidence suggests that the pilot may have been distracted from the critical task of terrain clearance while attempting to spot his weapon impact.

Safety recommendations include the re-enforcement of low level awareness training principles and improved training on Terrain Awareness Warning System reactions.

Return to top of page

CANADIAN ARMED FORCES FLIGHT SAFETY INVESTIGATION REPORT (FSIR)

FINAL REPORT

FILE NUMBER:  1010-CF188-171345 (DFS 2-7)

FSIMS OCCURRENCE NUMBER:  171345

DATE OF REPORT:  12 January 2018

OCCURRENCE CATEGORY:  "A"

AIRCRAFT TYPE:  CF188 Hornet

AIRCRAFT TAIL NUMBER:  CF188747

DATE OF OCCURRENCE:  28 November 2016

TIME OF OCCURRENCE (L):  11:03 (L) MST / 18:03 UTC

LOCATION:   4 Wing Cold Lake Air Weapons Range
N54 55.54 W109 13.32
Elevation: 2132’ ASL

OPERATOR:  401 (Tactical Fighter) Squadron

This report was produced under authority of the Minister of National Defence (MND) pursuant to section 4.2 (1)(n) and 4.2 (2) of the Aeronautics Act, and in accordance with the A-GA-135-001/AA-001, Flight Safety for the Canadian Armed Forces.

The contents of this report shall only be used for the sole purpose of accident prevention. This report was released under the authority of Director of Flight Safety, National Defence Headquarters, pursuant to powers delegated to him by the MND as the Airworthiness Investigative Authority for the Canadian Armed Forces under Part II, section 12 of the Aeronautics Act.

Return to top of page

SYNOPSIS

The pilot of aircraft CF188747, using the call sign “Swift 32”, was part of a two-ship formation led by “Swift 31” for an air-to-ground training mission. The mission objective was to practice level deliveries of two Mark 83 inert bombs followed by two laser guided training rounds, simulating laser guided bombs, in the Cold Lake Air Weapons Range. The plan was to ingress to the target and drop weapons from 600 feet above ground level. To avoid simulated bomb fragmentation after dropping their bombs each pilot would fly a “breakaway manoeuver” comprising a steep turn through 90 degrees of heading change.

Following his Mark 83 drop, Swift 32 manoeuvred his aircraft in a manner that was suggestive of a pilot attempting to visually spot his weapon impact, losing over 200 ft of altitude in the process. Swift 32 then assumed tactical lead, with Swift 31 flying about 2 miles in trail of Swift 32 and lasing the target for Swift 32, who then dropped his laser guided training round. The ingress to the target was flown at approximately 500 feet above ground level.

Immediately after dropping his laser guided training round Swift 32 initiated a steep left turn, reaching a maximum left bank angle of 118 degrees while pulling approximately 5g. The aircraft nose began to pitch towards and then below the horizon, eventually reaching a nose-down pitch angle of minus 17 degrees and concurrently generating a large descent rate.

About 1.5 seconds before impact the aircraft began rolling right. The bank angle had reduced to approximately 30 degrees left and the pitch angle increased to approximately minus 10 degrees when ground impact occurred. Swift 32 made no radio calls during the turn, did not eject and was fatally injured when the aircraft struck the ground in a descending left turn.

The available evidence did not support a mechanical failure, bird strike or pilot incapacitation scenario. Therefore, it appears that the pilot was capable of controlling the aircraft but did not adequately monitor the aircraft’s flight path while manoeuvring in the low level environment, and allowed the aircraft to enter an overbank situation and the nose to drop well below the horizon. A recovery may have been attempted at the last second but there was not enough altitude available to safely recover the aircraft. While the reason for this lack of flight path monitoring is not knowable with any certainty, circumstantial evidence suggests that the pilot may have been distracted from the critical task of terrain clearance while attempting to spot his weapon impact.

Safety recommendations include the re-enforcement of low level awareness training principles and improved training on Terrain Awareness Warning System reactions.

Return to top of page

TABLE OF CONTENTS

1. FACTUAL INFORMATION

2. ANALYSIS

3. CONCLUSIONS

4. PREVENTIVE MEASURES

Annex A - List of Abbreviations.

Annex B - Figures.

1. FACTUAL INFORMATION

1.1. History of the Flight

1.1.1. The occurrence pilot, “Swift 32” and his lead, “Swift 31” (together referred to as “Swift Flight”), both arrived at the 401 Tactical Fighter Squadron (TFS) hangar at 0730 local (L) time [1] on 28 November 2016. Swift flight was scheduled to fly an air-to-ground (A/G) training mission to practise weapon deliveries against pre-planned targets in the Cold Lake Air Weapons Range (CLAWR). The pilots were motivated to fly this type of mission, and in particular to practice level weapon deliveries. The targets were located approximately 90 kilometers northeast of the Cold Lake aerodrome (CYOD). Their training objective was to deliver two Mark 83 inert bombs on simulated oil tanks, followed by two laser guided training rounds (LGTR) simulating laser guided bombs on sea containers. They planned to drop their weapons from 600 feet above ground level (AGL) using straight and level deliveries at a speed of 450 knots indicated airspeed (KIAS). Although the weapons were not live, for training purposes to avoid simulated fragmentation damage, the pilots planned to fly a post-drop breakaway manoeuvre comprising a level steep turn away from the drop point.

1.1.2. Both pilots reviewed the weather for the Cold Lake and Jimmy Lake Ranges using aviation weather reports (known as a METARs), Terminal Aerodrome Forecasts (TAFs), and the Prairie Region Graphical Area Forecast (GFA). The CYOD TAF indicated that, at least temporarily, the weather would be suitable for visual flight rules (VFR) flight. The GFA valid at 1800 UTC (1100 L) forecasted cloud ceilings of 4,000 to 5,000 ft above sea level (ASL) (which equates to approximately 2,000 ft to 3,000 ft AGL) with extensive lower cloud ceilings from 400 ft to 1,200 ft AGL over the CLAWR. The Fighter Force Training Rules [2] (FFTR) required minimum weather of a 1,500 ft ceiling AGL and a visibility of 3 statute miles for an A/G mission.

1.3. The overall training plan for that day with the assigned aircraft was that Swift Flight would fly the first wave, drop their weapons and then “hot” turn/hot crew swap [3] to the second wave. The second wave would fly an assessed air interdiction mission for a pilot working towards his “Combat Ready” qualification. Following the morning flight, Swift 32 was scheduled for a simulator mission in which he was to brief and lead a combined Low Level Awareness Training (LLAT) and academic range mission. This mission was an upgrade mission as part of his Tactical Instructor Pilot training syllabus.

1.4.  Swift 32 was assigned aircraft CF188747 and Swift 31 aircraft CF188731. Both aircraft were configured for an A/G mission and loaded with a “Sniper® Advanced Targeting Pod” [4] (hereafter referred to as “Sniper Pod”), an LGTR and a Mark 83 inert bomb and three external fuel tanks. The mission was on the approved daily flying program and was authorized by the Squadron Duty Operations Officer (DOO). At the request of the second wave crew, who had been delayed starting their planning cycle, Swift Flight delayed their originally planned step time by 30 minutes, to 1000L. This also gave them a chance to check the latest weather reports which would refresh on the top of the hour. The pilots walked to their aircraft at 1000L, started up normally and took off from CYOD at 1034L.

1.5. The formation transited to the target area (Annex B – Figure 1) with Swift 31 flying between 500 and 900 ft AGL and Swift 32 flying in trail at between 500 ft and 800 ft AGL at a speed of approximately 360 KIAS. Approximately four minutes after Swift flight entered the CLAWR, Swift 32 climbed to check the base of the cloud ceiling and reported to Swift 31 via radio that the cloud base was 800 to 900 feet AGL and that the visibility was good.

1.6. Swift Flight completed a total of five target runs in the CLAWR over target complex E-355L (Annex B – Figure 2). Specifically, the target runs comprised one target verification pass (on the oil tanks), one target run where each aircraft, in trail, released their Mark 83 inert bombs, followed by a second target verification pass (this time on the sea containers), and then two runs for each pilot to release their LGTR. During this sequence, Swift 32 flew breakaway manoeuvres as part of the five target runs. Two of Swift 32’s breakaway manoeuvres were not flown as briefed. The first non-standard breakaway manoeuvre was flown after his Mark 83 release, the second was flown after his LGTR release.

1.7. For Swift 31’s Mark 83 drop, Swift 32 flew in a three mile trail position and used his Sniper Pod imagery to view the target (the oil tanks) and the weapon delivery and impact on one of his cockpit displays. Swift 32 radioed Lead to enthusiastically describe the weapon impact, ricochet, and bounce angle, before flying the briefed breakaway manoeuvre.

1.8. Following his own Mark 83 drop, Swift 32 immediately rolled right to 73 degrees of bank, then reversed the roll to the left to 94 degrees of bank, and then reversed the roll again back to the right to complete the first breakaway manoeuvre. The breakaway manoeuvre was flown at 60 to 74 degrees of bank for the majority of the turn, at 2.6g to 4.3g. Overall Swift 32 lost 230 ft of altitude while manoeuvring, of which 180 ft was lost while at 60 to 74 degrees of bank.

1.9. The first LGTR delivery was flown with Swift 31 in the lead and Swift 32 flying a three mile trail position and lasing the target for Swift 31. For Swift 31’s LGTR delivery, Swift 32 again used Sniper Pod imagery to view the target (the sea containers) on one of his cockpit displays. Swift 32 advised Swift 31 via radio that the LGTR fell short of the target, and had a pretty good ricochet.

1.10. Following this attack run both pilots flew the briefed breakaway manoeuvre. For the last attack, they swapped roles and Swift 32 assumed tactical lead with Swift 31 flying in a two mile trail position and lasing the target for Swift 32. Swift 32 flew the final target run-in at about 450 ft AGL on a heading 044 degrees magnetic (M), and advised Swift 31 via the radio that he had dropped his ordnance.

1.11. After making the radio call, Swift 32, flying at a speed of about 445 KIAS, initiated the breakaway maneuver by rolling left to a bank angle of about 118 degrees (Annex B – Figure 3) with the g increasing to and then remaining at about 5g after about 90 degrees of bank had been reached. The aircraft initially gained 50 feet of altitude before the nose of the aircraft began to pitch towards and then below the horizon, eventually reaching a nose-down pitch angle of approximately minus 17 degrees with a large descent rate.

1.12. After a few seconds the aircraft began to roll back to the right, reaching approximately 30 degrees of left bank just prior to ground impact, while the pitch attitude had increased to about minus 10 degrees. Flying in trail and focussed on laser designating the target on his multi-purpose display indicator (MDI), Swift 31 did not observe Swift 32’s descending turn but did observe a large explosion in his peripheral vision.

1.13. Upon realizing what had occurred, Swift 31 did a couple of passes over the site to look for the pilot of Swift 32 and radioed a Mayday call. On the second pass Swift 31 noted an open parachute at the edge of the main impact area and reported this information via radio. “Odin Flight”, another flight of two CF18s, heard Swift 31’s Mayday call and relayed the information to air traffic control and to the command post at Cold Lake. Swift 31 remained on station as the “on scene commander” until relieved by Odin 21. Hearing the radio call, the base rescue helicopter, which was airborne on a utility mission at the time, returned to Cold Lake to refuel and pick up a Search and Rescue Technician (SAR Tech). It then proceeded directly to the crash site where the SAR Tech was lowered by hoist and was able to confirm that the pilot had been fatally injured in the crash.

Return to top of page

1.2. Injury to Personnel

1.2.1. The pilot was fatally injured on impact.

1.3. Damage to Aircraft

1.3.1. The aircraft was destroyed (Category A damage).

1.4. Collateral Damage

1.4.1. The aircraft crashed on Crown land in the CLAWR. Fuel, oil and hydraulic fluids were released into the surrounding area. Due to the nature of the terrain and the scattering of the debris, this could not be contained.

1.5. Personnel Information

Pilots - Swift 31 and Swift 32 
PilotsSwift 31Swift 32
Total flying time hours (military) 1022.8 766.2
Flying hrs on type 464.5 451.3
Flying hrs last 30 days 7.9 16.5
Flying hrs last 90 days 26.9 48.1
Duty hrs - day of occurrence 3.5 3.5
Duty hrs - previous 24 hrs 11.5 4.5
Duty hrs - previous 48 hrs 26.5** 4.5
Duty hrs - previous 72 hrs 32.0** 4.5
Annual proficiency check 30 Jun 2016 31 May 2016
Currency Valid Valid
Medical category Valid Valid

**Includes 9.5 hours of rest while holding alert duties

Table 1.  Personnel Information

1.5.1. Swift 31, the formation lead, originally trained in the Royal Navy and had undertaken training on the Harrier aircraft. The Royal Navy flight training on the Hawk and Harrier aircraft involved numerous low-level missions. The Harrier fleet was retired before he flew it operationally. Swift 31 then joined the Royal Canadian Air Force (RCAF), graduated from the CF188 Fighter Pilot Course at 410 Tactical Fighter Operational Training Squadron in November 2013 and was on his first tour as a CF188 pilot. He had accumulated 1022.8 hrs total flying time and 464.5 hours on the CF188. He was a qualified element (two-aircraft) lead and was qualified and current to fly an A/G mission.

1.5.2. The pilot of Swift 32, the wingman, was on his first operational flying tour in the RCAF. He completed his CF188 Fighter Pilot Course in June 2014, finishing as the top student. In July 2014 he was assigned to 409 TFS as a line pilot. From 12 August 2015 to 30 September 2015 he deployed and flew missions in support of Op IMPACT. Upon returning to Cold Lake in October 2015 he was posted to 401 TFS. He had accumulated a total of 766.2 flying hours of which 451.3 hrs were on the CF188. Swift 32 was qualified and current to fly an A/G mission and his last recorded LLAT training flight was flown on 12 October 2016. He was a qualified element lead and due to his demonstrated abilities had begun training to become a Tactical Instructor Pilot and had also commenced the upgrade training to become a section (four-aircraft) lead.

1.5.3. Swift 31 and Swift 32 were of similar CF188 experience levels and socialized together outside of the work environment. Both were qualified as element leads and the authority gradient [5] between them in the roles of lead and wingman was relatively shallow, especially considering the additional upgrade training that Swift 32 had received. Two characteristics of a shallow authority gradient are a more collaborative approach to task completion and a greater willingness on the part of the subordinate person to voice concerns.

Return to top of page

1.6.  Aircraft Information

1.6.1. CF188747 was a single seat CF188 (F/A-18A) fighter aircraft. The aircraft was serviceable at the time of its launch from 4 Wing Cold Lake and had accumulated a total of 5,377.6 airframe hours and had a relatively low total fatigue life expended index of 0.7428 (the maximum possible fatigue life expended index for any CF188 aircraft is 1.0 before end of life).

1.6.2.  A review of the aircraft maintenance record set found no anomalies. The review included the electronic records within the Data Management System and the physical (paper) records for its Periodic (PER) Inspection. PER no. 1 had been completed the month before the accident and the aircraft had flown 47.5 hours since PER no.1. The aircraft had undergone both a Daily Inspection and a Before Flight inspection prior to the flight, with no anomalies noted.

1.6.3. The left engine, serial number 376017, had accumulated 7,750.1 operating hours since new and had completed a periodic inspection in June 2016, and had accumulated 94.7 operating hours since the inspection. The right engine, serial number 376023, had accumulated 6,615.4 operating hours, including 80.5 operating hours since its last periodic inspection in May 2015.

1.6.4. The aircraft was configured for an A/G training mission. Four hundred rounds of ballast were loaded within the M61A1 gun. It was loaded with a Sniper Pod underneath the left wing root at station 4, an LGTR under the left wing at station 2, a Mark 83 1,000 lb inert bomb under the right wing at station 8, a captive air training missile (CATM-9M) on the left wingtip and an airborne instrumentation subsystem (AIS) pod [6] on the right wingtip. It was also equipped with an external fuel tank under each wing and a centreline fuel tank. The aircraft was within limits for weight and longitudinal centre-of-gravity at take-off with an estimated takeoff weight of 46,300 lbs and a mean aerodynamic chord (MAC) of 19 percent. Using expected fuel burn rates, the aircraft was within its weight and balance envelope, with an estimated weight of 38,600 lbs and a MAC of 21%, at the time of the accident.

Flight Controls

General

1.6.5. The following information is extracted from the CF-18AM/BM Hornet Aircraft Operating Instructions (AOIs) (ECP-583R2) [7]. The primary flight controls comprise the ailerons, twin rudders, differential/collective leading edge flaps (LEF), differential/collective trailing edge flaps, and differential/collective stabilators. Hydraulic actuators position the control surfaces. There is no direct aerodynamic feedback to the stick and rudder pedals so stick and rudder feel is provided to the pilot by spring cartridges which in turn receive input from the flight control computers (FCCs). Normally, inputs to the hydraulic actuators are provided by the two FCCs (FCC A and FCC B) through the full-authority control augmentation system (CAS). A direct electrical link (DEL) automatically backs up the CAS. DEL is normally a digital system, but has an analog mode for backup aileron and rudder control.

1.6.6. The lateral control system, used to roll the aircraft around the longitudinal axis, uses ailerons, differential leading and trailing edge flaps, differential stabilator, and rudders to achieve the desired roll. Flight control responsiveness is generally reduced at high speed to prevent the exceedance of structural limits. The flight computers limit the maximum roll rate when carrying any wing pylon mounted stores except missiles (which was the configuration of the accident aircraft).

Redundancy

1.6.7. Multiple redundant paths ensure that single failures in the flight control system have no effect and that multiple failures will have a minimal effect on control. Most flight control malfunctions trigger a "FLIGHT CONTROLS, FLIGHT CONTROLS" voice-alert. For aileron and rudder operation, all four hydraulic circuits would have to be off before they totally shut down.

1.6.8.  If digital DEL fails, a mechanical operation (MECH mode) automatically provides pitch and roll control through a direct mechanical input from the stick to the stabilator actuators. MECH mode bypasses both FCCs, stabilator actuator servo valves, force sensors, all air data, all motion feedbacks, and associated electrical wiring. Stick movement directly controls the stabilator actuators.

Leading Edge Flaps Failures

1.6.9. The left and right LEF systems are controlled independently by the aircraft’s FCCs. Each wing has an inboard LEF extending from the wing fold to the fuselage, and an outboard LEF extending from the wing fold to near the wing tip. The CF188 flight control system has a number of software and hardware monitors, including pre-flight built in tests, to detect LEF failures. If the system detects a failure it will isolate the failed side and apply a brake to hold it within a few degrees of its failed position. The other side will continue to function normally.

1.6.10. To address a series of LEF drive train failures, a CF188 LEF system malfunction Record of Airworthiness Risk Management (RARM) was completed in 2006. The RARM assessed the risk to safety of flight as being within the acceptable level of safety. Nevertheless, a series of maintenance related preventive measures were implemented to reduce the likelihood of recurrence. A review of the Flight Safety Information Management System (FSIMS) database found that since the RARM was completed in October 2006, there have been 11 reported occurrences of LEF system malfunctions (of various types) and no mention in the reports of any notable controllability issues.

Aileron Lock down Failures

1.6.11. In September 2009 a CF188 pilot experienced an uncommanded roll to the right during an A/G training mission. The pilot experienced oscillations in roll and a large yaw to the left but was able to control the aircraft and landed safely. This is the only known instance of this failure in a CF188.

1.6.12. The investigation into this incident found that the left aileron had become locked in the fully down position and was non-functional. The cause of the lock down was found to be a failure of the linear variable directional transducer (LVDT) which is located in the aileron servo-cylinder. This resulted in the aileron remaining stuck in the full deflection position.

1.6.13. The LVDT had two known failure modes that would cause it to fail in this manner: a pre-loaded nut breaks off or there is a bearing failure. Both failure modes can be caused by improper maintenance techniques. In November 2009 a RARM was completed to address this airworthiness issue. The RARM determined that there was a low probability of failure and that the overall risk index was at an acceptable level of safety. The probability of a catastrophic accident due to an aileron lock down was determined to be extremely improbable. Nevertheless, maintenance activities were updated to minimize the probability of improper maintenance and a Special Inspection was completed to ensure that the LVDT’s that were in service did not have worn bearings. There have been no reported failures of the LVDT in the CF188 since the one in 2009.

Cockpit Displays

1.6.14. The cockpit avionics controls and displays comprise a head-up-display (HUD), left and right MDIs, a horizontal situation display (HSD), an up-front control panel, hands-on throttle and stick switches and standby instruments.

1.6.15. The HUD is the primary flight instrument in the aircraft, providing flight, navigational and weapon delivery information. It projects collimated symbology into the pilot’s forward field of view.

1.6.16. The right and left MDIs can display the situational awareness and electronic attitude and direction indicator pages and project information on the HUD in addition to a number of other tactical and support displays. When tactically appropriate, Sniper Pod imagery is typically displayed on either the right or left MDI, depending on the pilot’s preference. It is not known on which display Swift 31 chose to display his Sniper Pod imagery.

Sniper Pod

1.6.17. The Sniper Pod can be used for target detection and identification, continuous stabilized infrared and optical surveillance, and designating a target using a laser. The aircraft structure limits the field of regard (view) in an upward direction but not downwards. Based on the known flight path and the field of regard of the Sniper Pod, the LGTR target would have been visible by Swift 32’s pod throughout the post-target turn, despite the large bank angles. A large bank angle, however, is not required to keep the target in the pod’s field of regard.

Joint Helmet Mounted Cueing System

1.6.18. Swift 31 and 32 were using the Joint Helmet Mounted Cueing System (JHMCS). The JHMCS displays weapon and aircraft performance information on the helmet mounted display (HMD), directly in front of the pilot`s eyes. To limit the potential for information overload, the pilot can declutter the display by programing the amount of information that they wish to see. The HMD information is automatically removed when looking through the HUD and cockpit blanking region. It is not known what information Swift 32 had selected to display in his HMD.

Terrain Awareness Warning System

1.6.19. The CF188 is equipped with a terrain awareness warning system (TAWS) which functions as a safety back−up system and is intended to provide the pilot with a warning of an impending controlled flight into terrain (CFIT) condition. TAWS installation had been recommended after a series of fighter CFIT accidents in the early 1990’s. It was implemented in the CF188 between 2007 and 2009. The TAWS is an integral part of the aircraft’s operational flight program and uses aircraft performance information and digital terrain elevation data to detect and predict the possibility of CFIT conditions. If the tactical aircraft moving map capability system is not operational, warning of an impending CFIT situation is provided by the ground proximity warning system (GPWS). The TAWS is automatically selected on during the power-up sequence. If it is non-functional or degraded, this will be indicated to the pilot.

1.6.20. During all phases of flight TAWS constantly predicts two recovery trajectories (oblique or turning, and vertical) to determine if or when the trajectories will intercept the ground. Oblique (turning) and vertical trajectories are constantly compared and the last one to intercept the ground determines whether a turning or vertical escape warning cue is initially provided. TAWS is designed to eliminate false warnings and minimize nuisance warnings.

1.6.21. Five voice cues are provided to provide guidance to the pilot on how to manoeuvre the aircraft to avoid an impending CFIT. These cues, which are three to six decibels louder than other aircraft aural cues, include aural commands such as "ROLL RIGHT (LEFT)" (turning escape) or "PULL UP" (vertical escape). TAWS voice alerts may be inhibited by previously initiated voice alerts since once a voice alert has been activated, it cannot be interrupted by a higher priority voice alert. All voice alerts play until completed. Visual recovery cues in the form of large arrows are depicted on the HUD first and are followed by voice alerts.

1.6.22. A timely and properly executed TAWS reaction will provide only minimal terrain clearance. The CF188 Aircraft Operating Instructions state that "All TAWS warnings should be treated as though an imminent flight into terrain condition exists. Pilot response to a TAWS warning should be instinctive and immediate." The operation of TAWS system is covered during the CF188 ground school. There are no specific TAWS recovery procedures included in the Emergency Operating Procedures (time critical emergencies) section of the AOIs. There are no CF188 simulator training scenarios programmed to specifically demonstrate TAWS warnings or to practice TAWS recoveries.

1.6.23. A search of the FSIMS database did not reveal any reported instances of a TAWS failure in the CF188.

Ejection Seat

1.6.24. The CF188 aircraft is equipped with the Navy Aircrew Common Ejection Seat SJU−17B ejection seat. This ejection seat incorporates fully automatic electronic sequencing and is cartridge-operated and rocket-assisted. The ejection sequence is initiated by pulling a seat firing handle situated on the front of the seat bucket between the occupant’s legs. Once the handle is pulled, the canopy is immediately blown off, a bail-out tone and Link 16 [8] message is transmitted and the ejection seat begins to leave the aircraft 0.3 seconds later. At the speed and altitude of the accident aircraft, full parachute deployment would take approximately 2.3 seconds.

Return to top of page

1.7.  Meteorological Information

1.7.1. Cold Lake airport (CYOD), elevation 1,775 ft ASL, is located 48.6 NM (90 KM) southwest of the accident site. The accident occurred at 1803Z. The pertinent weather reports at CYOD were as follows:

SPECI CYOD 281613Z 30004KT 15SM BKN028 M03/M05 A2954 RMK SC5SC1 SLP030 SKY77 T10301054=

METAR CYOD 281700Z 31005KT 15SM BKN018 BKN030 M02/M05 A2955 RMK SC6SC1 SLP036 SKY99 T10201053=

METAR CYOD 281800Z 34005KT 15 SM OVC016 M02/M05 A2957 RMK SC8 SLP040 51020 SKYXX T10191054=

Based on their step time to the aircraft of just after 1700Z, the last METAR available to the crew would have been 1613Z special. The actual weather for departure from Cold Lake was VFR, meaning a ceiling of at least 1,500 ft AGL and a visibility of 3SM or better. The latest CYOD terminal forecast available to the pilots was an amended forecast issued at 1529Z (see below). Note that TAFs are intended to relate to weather conditions only within five [9] nautical miles of the centre of the runway complex at a particular aerodrome and all cloud heights are AGL. The 1529Z CYOD TAF forecast that conditions for their departure time were going to be slightly below VFR requirements with a ceiling of 1,200 ft AGL, with temporary (less than one hour in duration) fluctuations to VFR conditions, in the which the 1,200 ft layer would become scattered and the cloud ceiling would rise to 2,500 ft AGL.

TAF CYOD 281730Z 2818/2918 32008KT P6SM BKN015 BKN030 TEMPO 2818/2906 SCT015 BKN030

TAF AMD CYOD 281529Z 2815/2912 30008KT P6SM BKN012 OVC025 TEMPO 2815/2824 SCT012 BKN025 FM290000 32005KT P6SM OVC012 TEMPO 2900/2912 SCT012 RMK NXT FCST BY 281800Z=

TAF CYOD 281430Z 2815/2912 30008KT P6SM FEW008 OVC020 TEMPO 2815/2816 2SM -FZDZ BR OVC008 FM281600 32008KT P6SM OVC012 TEMPO 2816/2824 SCT012 OVC025 FM290000 32005KT P6SM OVC012 TEMPO 2900/2912 SCT012 RMK NXT FCST BY 281800Z

1.7.2. An automatic (unmanned) weather station is located at the Jimmy Lake Range (CWHN), elevation 2,090 ft ASL. This station is located 25 NM (46 KM) west of the accident site. The latest CWHN weather report available to the pilots before their flight was the 1637Z special (SPECI) report. The weather recorded at CWHN prior to and around the time of the occurrence was:

SPECI CWHN 281606Z AUTO 32005KT 9SM OVC024 M03/M05 A2950

SPECI CWHN 281637Z AUTO 30004KT 250V340 9SM BKN010 OVC024 M03/M05 A2950

METAR CWHN 281700Z AUTO 30005KT 250V350 9SM FEW011 BKN024 M03/M05 A2951 RMK SLP026 51020 T10311053=

SPECI CWHN 281710Z AUTO 29006KT 260V320 9SM BKN012 OVC024 M03/M05 A2952 RMK T10291055=

METAR CWHN 281800Z AUTO 30006KT 260V340 9SM OVC012 M03/M06 A2953 RMK SLP034 52019 T10321059=

1.7.3. The most recent Prairie Region GFA (GFACN32) available to the pilots prior to their flight was issued at 0533Z on 28 Nov 2016 and valid at 1800Z (1100L) on that same day. The GFA for the planned area of operations forecasted an overcast layer based at 4,000 to 5,000 ft ASL with a visibility of six or more statute miles below the clouds. It also forecasted extensive (meaning greater than 50% area coverage) ceilings of 400 to 1,200 ft AGL and visibility locally reduced to two statute miles in freezing drizzle and mist.

Note: The 4,000 to 5,000 ft ASL overcast cloud layer would equate to a ceiling of about 2,000 to 3,000 AGL given an average ground elevation of about 2,000 ft ASL.

1.7.4. Swift 32 twice radioed to Swift 31 that the base of the overcast cloud layer was between 800 ft and 900 ft AGL. A review of Swift 31’s HUD recording showed good visibility below the overcast layer. Swift 31 observed that the weather to the east of the target area was notably worse.

1.8. Aids to Navigation

1.8.1. Both aircraft were navigating using their Global Positioning System aided Inertial Navigation System. No anomalies were reported.

1.9. Communications

1.9.1. Swift Flight maintained communications on two VHF/UHF radios and a Link 16 voice channel. Because of the low altitude of the aircraft and the distance of the target area from the aerodrome, direct 2-way communications with the Cold lake aerodrome was not possible.

1.9.2. Crash scene co-ordination and information passage was conducted via a combination of aircraft satellite communications, VHF/FM, VHF/AM and UHF radios relayed through the top cover aircraft to both CYOD tower and the 4 Wing command post, as applicable.

1.10. Aerodrome Information

Not applicable.

Return to top of page

1.11. Flight Recorders

1.11.1. The CF188 is fitted with neither a crash survivable cockpit voice recorder nor a crash survivable flight data recorder. The impact forces were such that most of the non-crashworthy recording or memory devices were destroyed or never located.

1.11.2. The CF188 also has non-crashworthy mission computers that contain non-volatile memory modules that can retain up to 45 seconds of flight information. Mission computer 2 was recovered from the crash site but had sustained severe damage. It was forwarded to the National Research Council (NRC) flight recorder and playback center for analysis, but it was severely damaged and no flight data information could be extracted. Mission computer 1 was not recovered.

Removable Memory Module

1.11.3. The CF188 has a digital cockpit video recording system (D-CVRS) that records up to three pilot displays at the same time, including: HUD, JHMCS, and/or three multi-function displays, plus audio. The D-CVRS was in operation at the time of accident. Data from the DVRS is stored on a digital removable memory module (RMM). CF188747’s RMM was recovered from the crash site but had sustained severe impact damage (Annex B – Figure 8). The RMM was transported to the NRC flight recorder and playback center for analysis; however, the degree of damage to the individual electronic chips precluded any data recovery and no flight information or audio could be extracted.

1.11.4. Swift 31’s RMM was reviewed. It did not capture the accident sequence and did not provide any useful investigative data other than some of the inter-aircraft communications and a confirmation that the visibility below the cloud layer was good.

ACMI

1.11.5. The accident aircraft flight path was captured on the ACMI system. The airborne terminal of the CF188 ACMI System, the AIS pod, was installed on the aircraft’s right wingtip. The pod autonomously and continuously determines its own time-space-position information, including position, velocity, altitude and attitude by means of Global Positioning System and inertial data. Airspeed and angle of attack data is provided by a pitot-static system located on the front of the AIS pod. Other ACMI data is derived or calculated and is not based on direct inputs from the aircraft systems. All data is recorded to a data cartridge located in a slot in the aft end of the AIS pod. The on-board data collection rate of the AIS pod is 10 Hertz but the real-time external transmission of this data is limited to a 1 Hertz rate (i.e. once per second) thus there could be relevant data that is not captured due to the relatively slow collection rate. ACMI transmissions are received by a communications tower in the CLAWR and then transmitted to and saved by the Cold Lake CUBIC facility (home station). The on-board AIS pod and its data was not recovered from the crash site; therefore, only the 1 Hertz data was available for review.

1.11.6. For ACMI data playback, the data is automatically smoothed (interpolated) to fill in the gaps between the data points and therefore the fidelity of the information between the data points is an approximation. As such, while it is useful in generating a general picture of the aircraft’s performance and flight, there can be some limited errors in the accuracy of the data. Despite these limitations, overall the data was judged by engineering experts to be of sufficient fidelity to be useful. To refine the data and improve the accuracy of the flight parameters, the ACMI data was then compared to the outputs generated by a high fidelity CF188 engineering simulator and, where required, the flight path data and parameters were adjusted.

1.11.7. The 1 Hertz AIS data was used to review Swift 32’s accident flight with respect to bank angle, flight path angle, angle of attack, heading, altitude, indicated speed, and g loading. The ACMI data for the Mark 83 pass (i.e. the weapon drop prior to the accident pass) showed that upon releasing the weapon from approximately 500 ft AGL, Swift 32 pitched up slightly then initially banked steeply right (73 degrees) for approximately three seconds, then reversed the roll and banked steeply left (94 degrees) for approximately four seconds, after which Swift 32 reversed the turn back to the right and completed the breakaway manoeuvre using 60 to 74 degrees of bank with g ranging between 2.6 and 4.3g. The overall altitude loss during this manoeuvring was approximately 225 ft and the lowest altitude recorded was 2,360 ft MSL (270 ft AGL).

1.11.8. On the accident pass, the adjusted ACMI data showed that after the LGTR drop Swift 32 entered a left turn at approximately 420 ft AGL at 445 KIAS, rolling left at an average rate of about 35 degrees per second until about 80 degrees of bank where the roll rate slowed to about nine degrees per second until the maximum bank angle of 118 degrees left was reached. After reaching 118 degrees of bank (see Annex B – Figure 3) and about 1.5 seconds before impact, the aircraft rolled right at a rate of at least 45 degrees per second, reaching about 30 degrees left bank just before impact. The aircraft climbed slightly at the beginning of the turn and peaked at 470 ft AGL three seconds into the turn but then quickly began to descend. In a few seconds the pitch attitude changed from an initial nose-up attitude of 5 to 10 degrees to about 17 degrees nose-down, then back to 10 nose-down just before impact. The g remained at about 1.5g until the bank reached 80 degrees at which point it rapidly increased to about 5g and then varied between 4.5g and 5g for the remainder of the flight. The engineering simulator indicated that this g rate required a near full aft longitudinal stick input of about 33 pounds force. The required lateral stick input was comparatively light at between 5 and 10 lbs. Total time from initiation of the turn until ground impact was approximately seven seconds.

Terminal Radar

1.11.9. The departure of Swift 31 flight for the range area was captured on the Cold Lake terminal radar. Swift 31’s transponder Mode C (altitude) transmission showed his aircraft at 1,800 ft ASL immediately after take-off (which equates to the runway elevation) followed by an initial climb to 2,700 ft ASL (800 to 900 ft AGL) and a turn to the northeast around Cold Lake (see Annex B - Figure 1). Once north of the lake, Swift 31’s Mode C transmission indicated a descent to 2,600 ft (600 ft AGL). The transponder was still indicating an altitude of 2,600 ft ASL when radar tracking was lost approximately 28 NM northeast of the airfield due to the radar’s inherent line of site limitations of altitude versus distance. Swift 32, as the wingman, did not have his transponder transmitting so no radar information was received from his aircraft.

1.12. Wreckage and Impact Information

1.12.1. The aircraft struck the ground at 1103L at position N54 55.54 W109 13.32, where the local ground elevation was 2,132 ft ASL and the muskeg covered terrain was lightly treed and generally level.

1.12.2. Based on corrected ACMI data, tree strike marks and ground scars, the aircraft struck the ground on a track of approximately 330 degrees M at over 440 knots indicated airspeed, left wing low at a flight path angle between minus 10 and minus 15 degrees. The impact crater was initially two distinct depressions, but these quickly filled with water and became one oblong depression 50 ft across and 100 ft long, which then froze over. The depth is unknown.

1.12.3. Most of the aircraft fragmented into small pieces with a few larger pieces comprising the right wing, both vertical stabilizers and sections of the engines (see Annex B – Figures 5, 6, 7). The main aircraft wreckage trail was oriented approximately north-northwest. The wreckage was distributed throughout a V pattern from the initial impact point, with the highest concentration of debris along the center of the wreckage trail and heavier items further down path. The majority of the wreckage was dispersed over an area approximately 2,500 ft long and 1,000 ft wide (see Annex B – Figure 4). No evidence of a windscreen bird strike was found. There was no sustained post-crash fire and most of the fuel was burned off in the initial impact explosion.

Return to top of page

1.13. Medical

1.13.1. The impact forces were immediately fatal to the pilot.

1.13.2. An autopsy was completed at Saskatoon City Hospital with the Directorate of Flight Safety Flight Surgeon and the 4 Wing Flight Surgeon in attendance. The autopsy, a review of the deceased pilot’s medical records, and interviews with persons close to the deceased revealed no relevant medical conditions that would be contributory to the accident. Potential for fatigue was also evaluated and deemed not to be contributory. Tissue samples were taken and sent to the United States Federal Aviation Administration Civil Aerospace Medical Institute laboratory for toxicological analysis. Toxicology sample testing of the accident pilot and the Lead pilot did not reveal the presence of any substances hazardous to aviation.

1.13.3. The g profile (time versus g) experienced by Swift 32 as recorded by the ACMI data was evaluated by a physiologist from the Defence Research and Development Canada Toronto Research Centre with extensive experience in g physiology. A review of the published research, prior knowledge and historical RCAF centrifuge training data suggests that generally, an imbalance of high g and anti-g protection would need to be maintained for at least four seconds for a possible g-induced loss of vision or consciousness to occur. Swift 32 experienced only a moderate amount of g over a four second period. Based on this data the physiologist concluded it was unlikely that the pilot suffered a g induced loss of consciousness and that there was a low probability, depending on the amount of anti-g straining manoeuvre used, of any severe g induced vision loss.

1.14. Fire, Explosives Devices, and Munitions

1.14.1. Although a fireball was reported by Swift 31, there was no ground evidence of any significant post-crash fire.

1.14.2. The aircraft carried no explosive ammunition. It was, however, equipped with various explosives associated with the ejection seat escape system as well as explosive cartridges for the separation of racks and training munitions, in addition to the cartridges for the actuation of the engine fire extinguishers. Not all of these were recovered.

1.15. Survival Aspects

1.15.1. The impact was not survivable.

1.15.2. The physical evidence indicates that the ejection seat was properly configured to function as designed. The ejection seat handle was found in the ARMED position, the seat safety pins were found stowed in the pilot’s helmet bag and the canopy was on the aircraft at impact. No bail-out tone was heard on the radio and no ejection symbol was generated in the Link 16 network. Collectively, this evidence indicates an ejection was not initiated.

1.15.3. Based on the known capabilities of the escape system and the aircraft’s speed, bank angle and flight path, the probability of a survivable ejection was negligible once the nose of the aircraft dropped below the horizon and the aircraft began to descend while in the overbanked condition.

1.15.4. The parachute came free of its ejection seat container during the breakup sequence and subsequently became caught up in some of the trees at the crash site, making it appear from the air as if it had deployed (see Annex B – Figure 9).

Emergency Transmitters

1.15.5. The CF188 is equipped with a bailout tone that activates once an ejection sequence is initiated. No bailout tone was received from the accident aircraft and no Link 16 symbol was noted by other aircraft on the network.

Aviation Life Support Equipment

1.15.6. Swift 32 was wearing a JHMCS helmet, the HALP/CF oxygen mask, MSV980 life preserver survival vest, a PCU-56 torso harness, an MSF830 anti-g suit, winter flight jacket, issue flight suit, non-issue flight boots and was wearing dual layer clothing including long thermal underwear and flight gloves with the inner liner.

1.15.7. Swift 32 was wearing his G2 anti-g suit, which is normally fitted for wear with his immersion suit. Records show it had been issued to him in October 2014. Damage to the anti-g suit precluded any conclusions about the fitting of the g-suit.

Emergency Response

1.15.8. The initial search was conducted by Swift 31 with top cover provided by a second formation of CF188s operating in the same area (call signs Odin 21 and Odin 22). Odin 21 relayed initial notification of the accident from Swift 31 to Cold Lake Tower at 1106L with positive confirmation of the crash passed at 1107L. The Cold Lake base rescue helicopter, already airborne on a utility mission, heard the transmissions regarding the accident and was tasked by the Tower at 1110L to proceed to the crash site. Before it could proceed to the accident site it had to return to CYOD for fuel and to pick up the remainder of the search and rescue (SAR) crew. The base rescue helicopter was airborne and enroute to the crash site at 1143L. It arrived on scene at 1207L and a visual lead-in to the site was provided by Odin 21, who was orbiting the crash site. No electronic distress signals were received by the rescue helicopter crew. The SAR Tech was hoisted down to the crash site to search for the pilot.

1.15.9. An attempt to reach the crash site was also made by ground SAR, but they were forced to turn back due to unpassable terrain. SAR response was in accordance with mandated times, however; it had no effect on the survival outcome.

1.16. Test and Research Activities

CF188 Simulator Trials

1.16.1. Multiple attack-runs and breakaway maneuvers were flown in the CF188 Advanced Distributed Combat Training Simulator by an experienced CF188 pilot using the same aircraft configuration and weight and balance as CF188747. The simulator is used for pilot training and the fidelity is such that its performance does not exactly replicate that of the actual CF188 aircraft. While recognizing this limitation, the objective was to duplicate as closely as possible Swift 32’s ACMI inflight data parameters during the last target run. These trials demonstrated that any delay in initiating a maximum effort recovery following the establishment of 118 degrees bank angle and an 18 degree nose down attitude would preclude recovery before ground impact. A TAWS "ROLL RIGHT" aural warning was heard between 1 and 2 seconds before impact.

TAWS Trials

1.16.2. CF188747 ACMI engineering data was sent to the high fidelity CF188 System Level Test Station to determine when TAWS cueing would have been provided to Swift 32 during the accident manoeuvre. The System Level Test Station provides the most accurate representation of CF188 flight performance available along with the most accurate representation of TAWS cueing, other than the actual aircraft.

1.16.3. The flight profile, as provided by CF188747 ACMI data, was flown by a CF188 pilot and a CAE Inc Operational Specialist (a former CF188 pilot). Despite having advance knowledge of the situation, these pilots were unable to consistently recover the aircraft with the aid of the provided TAWS warnings. If there was any delay in the pilot’s reaction or if a less than perfect recovery technique was used (in which the g loading was not reduced during the roll to wings-level) the recovery attempt would usually be unsuccessful.

1.16.4. By adding the measured TAWS warnings cues over the ACMI reference data, the aircraft flight parameters at key moments were computed. From this it was determined that:

  •  Swift 32 should have been provided TAWS visual and audio warning cues shortly before the predicted ground intercept. It was also established that the backup GPWS audio and visual warning cues would have provided a similar warning period to the pilot.
  •  At the time of the HUD TAWS warning the aircraft parameters were: 365 ft AGL, 5.3g, 113 degrees left bank and a minus14.5 flight path angle (FPA).
  • The recovery manoeuvre was started 1.30 seconds prior to impact. For test purposes, the assumed recovery point is assessed as the point at which bank angle is noted to decrease and at which the pitch is noted to increase. The aircraft parameters were: 302 ft AGL, 4.4g, 118 degrees left bank, and a minus 19.5 FPA.

  •  At impact the aircraft parameters were: 4.5g, 57 degrees left bank and a minus 13.2 FPA.

Flight Control Failure Engineering Simulator Analysis

1.16.5. The Directorate of Technical Airworthiness and Engineering Support contracted L3 Technologies to examine, with the aid of a high fidelity CF188 engineering simulator, the nominal aircraft performance and evaluate the aircraft’s performance under various flight control malfunction scenarios. These failures included: an aileron that had departed the aircraft, an aileron that had jammed at the commanded left roll position, an aileron hard-over and a LEF failure. In all cases the performance and resultant flight path of the aircraft did not match the corrected ACMI data, even if there were no pilot inputs made to counter the adverse aerodynamic affects.

Return to top of page

1.17. Organizational and Management Information

1.17.1. The 401 TFS daily flying schedule is the result of coordination meetings between the Squadron Weapons and Tactics Officer, the Squadron Operations Officer, and the Maintenance Officer. Aircraft availability, internal/external taskings, the short and long range training plan, and the short and long range training or currency needs of the pilots are considered. Scheduling does not dictate mission specifics to be flown, rather the schedule defines whether the mission will be air-to-air, or A/G, and will specify what the aircraft configuration and weapon’s load will be. The schedule which included the accident flight was produced at the end of the work day on Friday 25 November 2016 for missions to be flown on Monday 28 November 2016. Scheduling took into account the forecasted weather which was produced by the Joint Meteorological Centre.

1.17.2. Swift 31 and 32 read the schedule and planned the specifics of their mission taking into account currencies, qualifications, squadron qualification upgrades, and upcoming deployments. Flight leads can plan a back-up mission to avoid the loss of training opportunities due to changes in weather, aircraft configuration, and/or aircraft serviceability. Applicable training rules must be briefed for the back-up mission. Flight leads are entrusted to plan and fly their missions in accordance with all applicable orders and regulations.

1.17.3. All CF188 squadrons use a software program called “FlightPro” to electronically manage all aspects of aircrew currencies and signing authority/acceptance of aircraft prior to launch. The DOO will electronically sign in FlightPro to signify that mission authority is granted and CF188 pilots will sign electronically to signify their mission acceptance.

1.17.4. The accident flight was on the daily schedule and duly authorized by the DOO. There was no discussion between the DOO and the Swift Flight pilots about the particular details of the occurrence mission, nor was there required to be. As per the Squadron Flying Orders, the DOO confirmed the pilots had reviewed the Cold lake Airport and CLAWR weather (TAF, METAR/GFAs), Notices to Airmen, aircraft tail numbers and loads (one inert Mk83 and one LGTR per aircraft), and that Swift 31 and 32’s FlightPro currencies and qualifications were valid for this mission.

Pressure to Expend Weapons

1.17.5. The pilots of Swift flight were aware that immediately following their mission the two CF188 aircraft they were flying were to be “hot turned/hot crewed” for another mission. The follow-on mission was viewed as important in that it involved a pilot flying an upgrade mission for which there had been extensive pre-coordination carried out with external agencies such as intelligence, contracted air support, 42 Radar Squadron and the ACMI facility. It was clear to both pilots of Swift Flight that their weapons had to either be dropped during the mission, or if not dropped, downloaded from the aircraft after landing because the aircraft could not be hot refueled with weapons on board and orders precluded the aircraft from flying an air-to-air mission with air-to-ground stores. A weapons download would take some time and the delay would jeopardize the timing of, or even result in the cancellation of, the follow-on mission. For these reasons, prior to their departure, Swift 31 and 32 were told by a member of the Squadron operations staff to not come back with any weapons.

1.18.  Additional Information

1.18.1. In accordance with the FFTR, mission commanders/flight leads are to select and brief the applicable training rules items for their mission. In addition to the minimum weather requirements, the briefing items for air-to-ground missions include the minimum recovery altitudes for air-to-ground deliveries to ensure both safe ground clearance and fragment avoidance for either live or simulated weapons delivery. For air-to-ground (A/G) missions below 5,000 ft AGL (the “transition” altitude), additional briefing items include the primacy of flight path and terrain clearance and GPWS/radar altimeter settings and reactions. These items were briefed.

Low Altitude Flying

1.18.2. The altitude chosen by the pilots for their weapons drop was based on minimizing aircraft damage due to weapon fragmentation. This altitude would not be enough time for the weapon to fuse before impact. For various reasons, these weapons would typically be dropped from higher altitudes.

1.18.3. The FFTR in effect at the time of the accident stated that the “minimum enroute altitude or floor (area permitting)” was 1,000 ft AGL, or 250 ft AGL if conducting a “show of force.” No specific weapon delivery minimum altitude was listed but the FFTR stated that for A/G missions that the briefing was to include information that “minimum recovery altitudes for A/G deliveries are to ensure frag avoidance for live/simulated weapon delivery and safe ground clearance”. The FFTRs did not specifically state that weapons drops were restricted to 1,000 ft AGL or higher.

1.18.4. The National Defence Flying Orders (NDFO) [10] state that fixed wing aircraft shall not be flown below 1,000 ft AGL except when on an approach, taking off or landing; low flying has been authorized for prescribed low flying areas (for example, to practice LLAT), or over specifically prescribed low level cross‑country routes.

1.18.5. Low level flying and training effectively ceased (except for LLAT) in the CF188 following its 1997 certification to drop precision guided munitions. The cessation of most low level flying was not due to safety concerns per se but rather because the precision guided weapons were designed to be dropped from higher altitudes and low level flying was a significant driver of aircraft structural fatigue. Minimizing low level flying would extend the fatigue life of the CF188. Direction to CF188 operators on the cessation of low level operations was issued by 1 Cdn Air Div circa 2000; however, a copy of the actual message(s) could not be located by the investigation.

1.18.6. Interviews with various CF188 pilots from 4 Wing indicated that at least some pilots were not aware (prior to the accident) that dropping weapons from altitudes below 1,000 ft was not authorized, and stated they had flown mission profiles similar to the accident profile within the past couple of years. Generally, this was done because they believed that in the future there may be a tactical requirement for them to operate in the low-level environment. However, for other pilots, it was clear in their mind that this was prohibited. Due to the small sample size, no conclusions could be drawn about just how widespread the practice of dropping weapons from below 1,000 ft was, but it did occur on training missions other than the accident flight.

Weather Limits

1.18.7. For fixed wing VFR flights, the NDFO specify that the cloud ceiling must be at or above 1,500 ft AGL, and the visibility be three miles or greater. The minimum vertical distance from cloud allowed is 500 ft. Forecast ceilings down to a minimum of 1,000 ft AGL are permitted for flight planning purposes if aircraft are authorized to fly below 1,000 ft AGL and a vertical separation (aircraft to cloud base) of 500 ft is forecast. The FFTRs stated that for A/G missions with no adversaries the minimum weather limits were a cloud ceiling of 1,500 ft AGL, 3 NM flight visibility, with 500 ft vertical and 1 NM horizontally separation from cloud.

Low Level Awareness Training

Principles

1.18.8. CF188 pilots receive specialised training called LLAT which teaches pilots how to operate safely and effectively in the low level environment (defined as altitudes below 1,000’ AGL). The training takes place in the form of briefings, simulator sessions and dual/solo flights. Specific currency requirements must be met to maintain proficiency in low level flying. Experience and the previous losses of aircrew have demonstrated that one of the more hazardous aspects of low level flying is inadvertently descending while turning.

1.18.9. Turning in the low level environment is not necessarily dangerous but the proximity to the ground and the high velocities involved demand that if errors are made they must be rectified immediately. For example, if an aircraft flying at 450 knots at 500 ft AGL acquires a 10 degree down FPA for whatever reason, it will generate a descent rate of 134 feet per second, resulting in ground impact in just under four seconds if nothing is done to correct the downwards vector.

1.18.10. The recommended technique for turning safely at low level is to roll into the turn, keep the aircraft nose attitude or velocity vector on or slightly above the horizon and set the desired bank angle while pulling enough g to maintain level flight. The pilot must continuously and effectively monitor the aircraft’s nose position relative to the horizon. At low level, once the turn is established and the nose is confirmed to be tracking along or slightly above the horizon, then and only then is it safe for the pilot to look away from the HUD in the direction of the turn to look, for example, for rising terrain or other aircraft. Only a maximum of a one second look away from flight path monitoring is recommended before rechecking the flight path. If any descent or “nose slice” is detected, it must be corrected immediately by decreasing bank angle and then applying g.

1.18.11. With CFIT representing the largest threat, no mission task can take priority over the time critical tasks involved with terrain clearance. Common errors include not clearing the turn/checking terrain, improper g and bank combinations, not noticing and correcting nose slice immediately, and looking while turning or spending more than one second on non-critical tasks.

Training

1.18.12. 419 Tactical Fighter (Training) Squadron provides fighter lead-in training and flies the CT155 Hawk aircraft. It provides one Electronic LLAT (ELLAT) simulator (SIM) mission and one LLAT flight. Students at 419 Squadron also receive low level flying experience with conventional weapons delivery missions (three SIM sessions and three flights), plus low level air interdiction missions (one SIM session and two flights). 410 Tactical Fighter (Operational Training) Squadron trains pilots to fly the CF188 and provides one ELLAT SIM mission and one LLAT dual training flight. LLAT is typically flown as an “academic” mission, with prescribed manoeuvres that would not include weapons delivery profiles. However, evidence was also provided that some LLAT missions involved weapons drops and tactical manoeuvring. CF188 tactical fighter squadrons provide a LLAT brief annually, and to stay current CF188 pilots must fly at least one LLAT mission every 180 days. If a LLAT is not flown within 180 days then the pilot is required to receive one ELLAT supervised SIM prior to conducting a LLAT mission in the aircraft. Additional training requirements are specified if LLAT is not flown within 365 days. Swift 32 had received all the requisite training.

TAWS Training

1.18.13. Limited TAWS training is provided to CF188 pilots and focuses primarily on how the system functions. TAWS is installed and operational in the CF188 simulator, however, dedicated TAWS simulator training is not provided, therefore the majority of CF188 pilots are not specifically trained to react and recover to a TAWS warning. The only time a CF188 pilot may receive TAWS simulator training is if the student or pilot errs during a LLAT simulator maneuver resulting in a TAWS warning, thus necessitating a recovery. A TAWS warning requires “instinctive and immediate pilot action, with no hesitation until situational awareness is re-established.” [11]

1.18.14. For civilian aircraft that are required to be equipped with GPWS/EGPWS or TAWS, Transport Canada, as one example, requires air operators to provide CFIT avoidance training in a synthetic flight training device to include GPWS or TAWS escape manoeuvres [12]. The purpose of this training is to ensure flight crew can respond effectively to GPWS cautions and warnings and are aware of factors that can reduce effectiveness of GPWS.

Bird Activity

1.18.15. On the day of the accident bird activity in the CLAWR was reported as low. The majority of the bird migration season was completed by late November, and with the arrival of colder weather, the farmers’ fields were mostly snow covered and frozen, limiting forage available to wildlife. Most open water (lakes and ponds) had frozen over, further discouraging waterfowl from remaining in the area. No bird remains were found in the crash site area.

1.19. Useful or Effective Investigation Techniques

Not applicable.

Return to top of page

2. ANALYSIS

2.1. General

2.1.1. The severe fragmentation of the aircraft and the nature of the terrain at the crash site limited the physical evidence available for detailed forensic examination. The lack of any type of crashworthy recording device and the impact damage to those non-crashworthy recording devices that were recovered meant that the investigation had to rely on the ACMI data alone to reconstruct the accident flight path. No direct evidence was available that would enable an unequivocal determination of what exactly was going on in Swift 32’s cockpit. Lead did not observe the accident aircraft’s final turn or its impact with the ground. Lead’s RMM did not capture the actual accident sequence but did record some inter-aircraft radio communications and provided evidence regarding the ambient weather conditions and horizon definition in the area at the time of the occurrence.

2.1.2. The aircraft transmitted ACMI data at 1 Hz intervals to the Cubic ground station in Cold Lake. For unknown reasons, within the last 10 seconds of data, there were two non-consecutive one second intervals of data that were not transmitted. As a consequence, while the overall flight path is known with certainty, the exact roll rates, pitch rates, etc. between these missing data points cannot be known with precision and are subject to some interpolation. Nevertheless, the remaining points place boundaries on the actual aircraft performance and the data proved useful to the investigation in determining how the aircraft manoeuvred and its general flight path in the moments prior to its impact with the ground. The ACMI data was refined through the use of a high fidelity simulator and the resulting performance data was used to examine any role that potential flight control malfunctions may have had in this accident and to examine the expected output of the TAWS.

2.1.3. From a flight mechanics perspective, once the bank angle reached 90 degrees there was no longer any vertical lift vector (relative to the ground) being generated and the aircraft would start to descend. As the bank angle increased past 90 degrees, while still pulling approximately 5 g, the resultant total lift vector would now be below the horizon and as a result the nose of the aircraft would begin to increasingly track below the horizon, causing the aircraft to accelerate towards the ground. For this reason, once the aircraft reached 118 degrees of bank, the nose was well below the horizon and the aircraft was already rapidly descending towards the ground.

2.1.4. A review of the aircrew records indicate that both Swift 31 and Swift 32 were current and qualified for the planned mission, were motivated to fly it, and there was no evidence to suggest that fatigue was a factor in the accident sequence. As well, there was no evidence to support any environmental factors that could have caused the aircraft to manoeuvre as it did. The winds were relatively light, there was no reported turbulence or wind shear, the horizon was visible and well defined, there was no observed bird activity nor physical evidence of a bird strike. Consequently, these factors will not be analysed further.

2.1.5. Therefore, with no external factors affecting the flight path, the left roll to 118 degrees of bank, the application and maintenance of approximately 5g and the last second decrease in bank and g, required some type of flight control input(s), either commanded by the pilot, or otherwise. The analysis that follows will examine the available materiel and human factors evidence to attempt to explain the final flight path of CF188747 and then examine any latent conditions that may have played a role in the final outcome.

2.2. Mechanical Failure Scenarios

2.2.1. The briefed breakaway manoeuvre was to be flown as a steep turn using g, as required, to maintain level flight. The ACMI recorded a left roll from wings level to eventually 118 degrees of bank. Since the actual bank angle exceeded the bank angle that would normally be used, the investigation examined, to the extent possible, the likelihood that the aircraft flight controls may have acted in a manner contrary to the pilot’s commands due to some type of mechanical failure.

2.2.2. The aircraft had come out of a Periodic Inspection about one month prior to the accident, and had been performing well over the intervening time, with only a few minor maintenance write-ups. The pilot that flew the aircraft on the mission previous to the accident flight stated that it flew normally and he did not report any relevant discrepancies. The aircraft underwent the required daily and before flight inspections prior to the accident flight and was certified as airworthy. On the day of the accident there was nothing to indicate that Swift 32 was not satisfied with the condition of his aircraft.

2.2.3. The known final flight path was not representative of the way the aircraft would be expected to perform or how the pilot would react if there had been a reversion to MECH mode. In any event, the aircraft does remain controllable. Several flight control malfunction scenarios that may induce a rolling moment were analyzed in detail. These were: a LEF failure, a jammed aileron, a missing aileron, or an aileron lock down, all of which could cause an asymmetric lift situation between the two main wings and therefore introduce an undesired rolling moment. During the history of CF188 operations there have been several LEF failures reported, but only one case of an aileron lock down. The RARMs completed on these failures found that these were rare events and, that when they did occur, the aircraft remained controllable by the pilot. Nevertheless, several maintenance process related preventive measures were implemented to minimize the chance of their recurrence.

2.2.4. A detailed aerodynamic analysis using the ACMI data and known aircraft parameters was followed up with flight testing in a high fidelity CF188 simulator to further examine these possibilities. This analysis demonstrated that none of the examined failure scenarios would result in a flight path similar to that observed nor would they have made the stick forces so great that the pilot could not overcome them to maintain control of the aircraft. It would also be expected that if at any time the pilot was aware that the aircraft had become uncontrollable, especially in the low level environment, that an ejection would be attempted; however, the evidence showed that this was not the case.

2.2.5. In summary, the available evidence does not support a scenario in which a mechanical or materiel failure caused the aircraft to manoeuvre as it did and it is therefore concluded that the aircraft’s manoeuvres were the result of pilot inputs, deliberate or otherwise.

Return to top of page

2.3. Human Factors Scenarios

Incapacitation

2.3.1. The possibility of some form of sudden pilot incapacitation that precluded either the proper manipulation of the controls and/or a loss of situational awareness was considered. Evidence contrary to this hypothesis is that the accident pilot was relatively young, was reported to be in good health, and had a valid unrestricted aircrew medical category with no known pre-existing health problems. Although the autopsy provided no evidence that either supported or eliminated a sudden incapacitation scenario, the toxicology testing found no substances that would have impaired Swift 32’s ability to fly the aircraft safely and effectively.

2.3.2. Swift 32 was wearing a g-suit. Although its serviceability could not be confirmed, a detailed analysis of the flight and g profile by an experienced aviation physiologist from Defence Research and Development Canada concluded that the pilot did not pull enough g over a long enough period to cause either a g induced loss of consciousness or even a g induced narrowing or degradation of his visual field.

2.3.3. Finally, the increase in g after the turn was initiated, the maintenance of that g and longitudinal stick force, followed by the large degree decrease in bank angle and change in pitch attitude prior to impact, all strongly suggest that the pilot was making active control inputs throughout and therefore was not incapacitated.

Pilot Actions

2.3.4. If the aircraft was operating normally, and the pilot was conscious and capable of controlling the aircraft’s flight path, then the observed performance must have been the result of pilot inputs and a concurrent loss of situational awareness.

2.3.5. High-speed flight in the low level environment is a regime of aircraft operation that involves both high workload and high risk. Manoeuvring in this regime increases both the workload and the risk level. The primary focus when operating in the low level environment must be flight path monitoring and taking action, as required, to maintain terrain clearance. Any unintended nose-low attitude must be corrected immediately. CFIT can result when a pilot prioritizes mission, or other non-critical tasks, over terrain clearance tasks at a critical point in the flight. With terrain obviously being the greatest threat, no mission task should ever take precedence over time critical tasks – i.e., aircraft control, altitude awareness and flight path monitoring.

2.3.6. Swift 32 deviated from the briefed post-drop manoeuvre plan on both his Mark 83 drop and his LGTR drop, but notably did not when watching Lead’s drops on his Sniper Pod. On his Mark 83 drop, he manoeuvred his aircraft initially into a right turn, followed by a reverse into a steep left bank and then reversed again into a right turn. During these manoeuvres the aircraft descended to as low as 270 ft AGL. He was well clear of Lead so discounting that Swift 32 was just being indecisive about his turn direction, the manoeuvring is suggestive that he was trying to manoeuvre the aircraft to sight something on the ground below or behind the aircraft, either visually or in his Sniper Pod display inside the cockpit. The only thing of interest to observe at that time would have been the impact or post impact results of his inert Mark 83 bomb. Supporting this theory are his three previous radio transmissions to Lead in which he expressed excitement and satisfaction as he watched the impact and ricochet of Lead’s Mark 83 weapon on the Sniper Pod imagery displayed in his cockpit. There is no tactical reason to attempt to spot your weapon impact.

2.3.7. The overbank following the LGTR drop, which initiated the accident sequence of events, is again suggestive of either an aircraft being deliberately manoeuvred to improve the line of sight visibility to something on the ground behind the aircraft (such as the LGTR impact) or an unintentional overbank that developed as he was distracted from the primary task of monitoring the aircraft attitude and flight path vector. Whether the pilot intended to reach a bank angle of 118 degrees to improve his visibility towards the target, or rather intended some lesser but still steep bank angle for the same reason but, while looking inside at the cockpit displays, was distracted from monitoring the roll movement/attitude, is not knowable. However, based on the available evidence, it is deduced that, contrary to LLAT principles, once the aircraft began its turn the pilot’s attention was not focused on the time-critical task of monitoring the aircraft’s flight path and he did not notice that the aircraft had entered an overbank situation while maintaining the initial 5g, which combined to induce a very rapid descent. This situation, if recognized, should have cued him to immediately initiate corrective actions as per his LLAT. That this was not done is convincing evidence that he was distracted by non-critical tasks until the TAWS warning was generated and the situation quickly became non-recoverable.

TAWS Warning Reaction

2.3.8. As described in section 1.6.19, the CF188 is equipped with a TAWS. The purpose of the TAWS is to prevent CFIT accidents by comparing the aircraft trajectory with the known terrain profile around the aircraft and providing an appropriate visual and aural warning to the pilot when it calculates that a collision with the ground is imminent. The investigation sought to determine why the TAWS was ineffective in preventing a collision with the terrain. Given the nature of its role and the dynamic manoeuvring capability of fighter aircraft, the CF188 TAWS algorithms are adjusted to minimize nuisance warnings and take into account the possibility that the aircraft may be deliberately manoeuvring in the low level environment. It only provides for a couple seconds of recovery time and has limited ability to deal with extreme attitudes at low level.

2.3.9. The lack of RMM data precluded a definitive determination of whether or not there was a TAWS generated warning in the accident aircraft. However, based on historical maintenance records, TAWS failures are very rare. Simulation trials using the ACMI data to replicate the actual flight profile indicated it should have provided TAWS visual and audio warning cues shortly before the predicted ground intercept. The backup GPWS audio and visual warning cues should have provided the pilot with a similar warning period.

2.3.10. The accident flight profile was flown in the simulator by experienced CF188 pilots and these trials demonstrated that, even knowing the profile beforehand, they were unable to consistently recover the aircraft once the TAWS warning was generated. It was found that an effective recovery was unlikely if there was any delay in the pilot’s reaction, if the pilot’s attention was diverted away from the HUD, and/or if the rolling attempt was impeded by not decreasing the g first.

2.3.11. Therefore, if Swift 32 did receive a TAWS/GPWS alert, to have any chance of success it required an immediate and correct response. Recovery following the TAWS alert may have been possible but only if he executed a near perfect recovery by unloading the g, rolling rapidly to near wings level, and rapidly pulling back on the stick to avoid the ground. It can be speculated that the last second roll back towards level flight may have been indicative of his initial reaction to a TAWS “ROLL RIGHT” aural warning. However, and not unexpectedly, given the situation, the ACMI data indicated that the pilot rolled while pulling over 4 g, which is not the recommended technique and in the simulator trials this proved to be a detriment to quickly attaining the near wings level attitude required for recovery. Furthermore, the ACMI data suggests that the pilot did not utilize the maximum roll rate available to him.

2.3.12. A TAWS warning requires “instinctive and immediate pilot action, with no hesitation, until situational awareness is re-established” and a vector away from the ground is achieved. This system is the pilot's last defence between life and death and a correct and timely reaction is absolutely critical. An effective TAWS recovery, particularly involving descents involving high bank angles, may not be intuitive to pilots as it requires the pilot to fly an unloaded roll while in a descent approaching the ground, counter to the pilot’s survival instinct of wanting to pull-up to avoid the ground. Given the time critical response required, the potential for a delay due to the element of surprise and the importance of proper technique, pilots should be well practiced in the recovery technique so that it becomes an immediate and automatic response. In civil aviation it is common for crews of TAWS/EGPWS equipped aircraft to regularly practice their response to TAWS/EGPWS warnings in the simulator. TAWS recoveries are not currently part of the CF188 simulator training syllabus. While most CF188 training and operational flights are flown above 5,000 ft AGL, occasionally, missions are flown at low altitude for LLAT, therefore TAWS simulator training would be prudent to ensure CF188 pilots are trained to immediately and effectively react to and recover from a TAWS warning.

Return to top of page

2.4. Mission Planning and Weather

2.4.1. Both pilots were motivated by the opportunity to drop weapons and practice level flight deliveries. They actively planned the mission together using a delivery altitude of 600 ft AGL. At least one other pilot observed they were planning 600 ft deliveries but did not view this as contrary to the regulations. Swift 31 and 32 were of similar CF188 experience and both were qualified as element leads, but for this particular training mission, Swift 31 assumed the role of Lead and Swift 32 as wingman. While there are clearly defined roles and responsibilities for the Lead and the Wingman, in the case of these two particular pilots there was not a large authority gradient between them. It is expected that Swift 32 would have felt free to express any concerns that he may have had about how the mission was being planned; however, there was no evidence to indicate that he was uncomfortable with the mission as planned.

2.4.2. Although the mission was scheduled and authorized in accordance with the existing orders, Swift Flight’s decision to transit and drop weapons at 600 ft AGL was contrary to the intent of the 1,000 ft minimum enroute altitude stated in the FFTR at the time. Nevertheless, the occurrence pilots, and at least a number of other Squadron pilots, believed that their choice of altitude was acceptable. They apparently interpreted the 1,000 ft minimum enroute altitude as applying only to flights outside the military terminal control area or the CLAWR. They believed that since their A/G mission, except for the initial transit, was to be flown in an area approved for low level flying and because they had briefed LLAT principles, they were cleared or authorized to fly low level. They had also participated in other training missions with other pilots in which a similar profile had been flown, re-enforcing their view that this was an acceptable practice. In contrast, there were other CF188 pilots who were very surprised to learn that the mission had been flown at altitudes below 1,000 ft and it was clear in their mind that this was contrary to the FFTR. Overall, the evidence suggests that over time there had developed a subset of pilots who had come to interpret the 1,000 ft minimum as something that did not apply to A/G missions in the CLAWR, and the occurrence pilots were among that group. In their mind, the mission profile was in accordance with the existing rules. However, just because a certain area is authorized for low level flight operations, it does not necessarily follow that all flights in that airspace are authorized to conduct low level flight simply because it is operating in that airspace. Following this accident, the FFTRs were amended to explicitly prohibit weapon deliveries below 1,000 ft AGL.

2.4.3. The actual weather at the airport was suitable for a VFR departure. However, the GFA indicated that the weather in at least half of the CLAWR would be unsuitable for their mission (forecasting extensive ceilings of 400 ft to 1200 ft AGL and 2 miles visibility in drizzle). If they had encountered a very low ceiling, this would have caused them to cancel but the actual cloud ceiling, although below the required limit, was apparently such that they felt they could safely continue the mission. The actual weather conditions in the operations area were below the FFTRs 1,500 ft ceiling requirement. Swift 32’s report to Lead via radio that the ceiling was 800 to 900 ft (above ground) should have cued both pilots immediately that the cloud celling was below the required minimum. However, Swift 32’s radio transmissions to Swift 31 did not indicate any reluctance to continue and this can be taken as a tacit agreement, if not encouragement, particularly considering the relatively flat authority gradient between the two, that he was content with continuing the mission in those weather conditions. One factor influencing the decision to continue with the mission, despite the weather, was the pressure they felt to not return with their weapons.

2.5. Pressure to Drop Weapons

2.5.1. Swift Flight felt pressured to expend their ordnance (i.e. not to come back with weapons still on the aircraft). If they delayed for weather, there would be a schedule risk to Swift 32’s afternoon simulator training mission and the planned afternoon flying schedule. If they took off but did not drop their weapons and returned to base with unexpended ordnance, the aircraft would have to be shut down, the weapons downloaded prior to refueling and the aircraft then re-inspected prior to flight. These activities could take considerable time and delay or jeopardize the follow-on air-to-air mission planned using these same aircraft. This air-to-air mission was part of a pilot upgrade trip and involved extensive planning and coordination with external agencies. Finally, the pilots of Swift Flight were directly told by a member of the Squadron operations staff not to return with their weapons. This pressure influenced their decision to continue the mission and drop their weapons, even when it became evident to the pilots that the weather was below the required minimums.

2.6. Summary

2.6.1. An analysis of the available evidence leads to the conclusion that misplaced motivation and a lack of appreciation for the very real risks associated with manoeuvring in the low level environment set up a situation where Swift 32 disregarded LLAT principles and either flew his CF188 aircraft into, or allowed it to enter, an overbanked condition while at low altitude. The specific reason for the overbank and the pilot’s distraction from the primary task of terrain clearance during the turn cannot be known with certainty. The evidence from his Mark 83 drop and his comments to Swift 31 over the radio strongly suggest that he was attempting to visually acquire his weapon impact. The investigation could not determine any other logical reason for the aircraft to be manoeuvred in such a manner. Nevertheless, once overbanked, the pilot’s lack of focus on the primary task of terrain clearance created a situation from which the aircraft could not be safely recovered. A TAWS warning should have been triggered at the last moment, but it would have required an immediate and textbook recovery action on the part of Swift 32 to avoid ground contact.

2.6.2. The degree to which the low level flying aspect played in the final outcome is speculative, at best, and a direct causal link between the low level flying and the accident sequence cannot be drawn. The cloud ceiling, although lower than that required by the regulations was sufficiently high that it did not materially affect their flight profile; they had planned to fly at 600 ft AGL anyway. If Swift 32 had flown the breakaway manoeuvre as planned and briefed, i.e., a level 5g turn with proper adherence to the principles of LLAT, the aircraft would not have struck the ground despite the low altitude. Intuitively, a higher entry altitude would allow more time and space for a recovery, decreasing the overall inherent risk. However, it cannot be known with any certainty whether Swift 32, had he been at 1,000 ft or higher, would have recognized the situation in time and whether there would have been enough altitude remaining to effect a safe recovery. In summary, the low altitude flying did increase the risk level but did not necessarily lead to the accident. Ultimately, it was the unnecessary manoeuvring and a disregard of the LLAT principles that placed the aircraft in a position from which recovery was unlikely.

Return to top of page

3. CONCLUSIONS

3.1. Findings

3.1.1. CF188747 was airworthy prior to departure. It was loaded with one inert Mark 83, one LGTR, a Sniper Pod, one CATM, one AIS pod and three external fuel tanks. The weight and balance was within limits. [1.6.4, 2.2.2]

3.1.2. Swift 31 was the formation Lead and Swift 32 was the wingman. Swift 31 and Swift 32 were both qualified as element leads and both were qualified and current for the planned A/G mission. [1.5.1, 1.5.2, 1.5.3]

3.1.3. The mission was properly scheduled and authorized in accordance with the 401 Squadron Orders. [1.17.4]

3.1.4. Swift 31 and Swift 32 were motivated to fly the air-to-ground mission and planned to drop their weapons from an altitude of 600 ft AGL. The briefed post-drop fragment clearance or breakaway manoeuvre was to be flown as a steep turn away from the drop point. [1.1.1]

3.1.5. The forecast weather for departure from Cold Lake was for temporarily VFR conditions, with the graphical area forecast calling for ceilings of 4,000 to 5,000 ft ASL (2,000 to 3,000 ft AGL) in the range area with good visibility below the clouds and extensive ceilings of 400 to 1200 ft AGL and locally visibility reduced to 2 SM in freezing drizzle and mist. [1.7.1, 1.7.3]

3.1.6. While some pilots understood the 1,000 ft minimum enroute altitude to apply to all flights except LLAT, other CF188 pilots at 4 Wing did not consider that the 1,000 ft minimum enroute altitude in the FFTR applied to flights in the CLAWR, and that weapons deliveries from altitudes below 1,000 ft was acceptable as long as minimum fragmentation clearance altitudes were respected. The orders were subsequently modified to prohibit weapons drops at altitudes below 1,000 ft AGL. [1.18.3, 2.4.2]]

3.1.7. The pilots of Swift Flight believed that the FFTR 1,000 ft minimum enroute altitude did not apply to flights operating in the CLAWR for which LLAT principles had been briefed. [2.4.2]

3.1.8. Other 4 Wing pilots had flown similar weapons delivery profiles/altitudes as the accident mission profile and did not perceive that this was contrary to the FFTR. [1.18.6]

3.1.9. The pilots of Swift Flight felt pressure to expend their weapons and to complete the mission in a timely manner to avoid disruptions to the Squadron’s planned flight and training schedule. [2.5.1]

3.1.10. Enroute to the target Swift 32 climbed to check the cloud ceiling and reported to Swift 31 that the ceiling was between 800 and 900 ft AGL. Visibility was good below the cloud layer. [1.1.5]

3.1.11. The actual cloud ceiling in the target area was below the minimum required in the FFTR and National Defence Flying Orders specified weather limits for such a mission. [1.18.7, 2.4.3]

3.1.12. Following his Mark 83 drop, Swift 32 banked steeply right, then reversed the roll and banked steeply left, and then reversed the turn back to the right and completed the breakaway manoeuvre using 60 to 74 degrees of bank. The overall altitude loss during this manoeuvring was approximately 225 ft and the lowest altitude recorded was 2,360 ft ASL (270 ft AGL) [1.11.7, 2.3.6]

3.1.13. The post Mark 83 drop manoeuvring of Swift 32 was suggestive of a pilot trying to visually acquire his weapon impact. [2.3.6]

3.1.14. Immediately following his LGTR drop, Swift 32 entered a 5g left turn and while maintaining the g, the bank angle reached 118 degrees of left bank, resulting in a nose slice and rapid descent towards the ground. [1.11.8]

3.1.15. Just before impact, Swift 32 began to roll to the right while maintaining over 4g. [1.11.8]

3.1.16. An ejection was not initiated. [1.15.2]

3.1.17. The aircraft struck the ground at over 440 knots indicated airspeed, left wing low at a flight path angle between minus 10 and minus 15 degrees, destroying the aircraft and causing immediately fatal injuries to the pilot. [1.12.2, 1.13.1]

3.1.18. The evidence did not indicate that there had been a mechanical or materiel failure that would have caused the aircraft to manoeuvre as it did. [2.2.5]

3.1.19. The evidence did not indicate that there had been a bird strike or that the pilot had become incapacitated. [1.12.3, 2.1.4, 2.3.3]

3.1.20. The investigation concluded that following the LGTR drop, Swift 32’s attention was distracted from the time-critical task of monitoring the aircraft’s flight path during the turn and he did not notice that the aircraft had entered an overbank situation while maintaining approximately 5g, inducing a very rapid descent. [2.3.7]

3.1.21. A successful TAWS recovery at the likely time of warning activation would require an immediate reaction and textbook recovery technique. [1.16.3, 2.3.11]

3.1.22. A TAWS recovery, if attempted by Swift 32, was ineffective in preventing the aircraft from striking the ground. [2.3.11]

3.1.23. TAWS reaction/recoveries are not part of the CF188 simulator training syllabus. [1.18.13]

3.2. Cause Factors

Active Cause Factors

3.2.1. Following the LGTR drop, Swift 32’s attention was distracted from the time-critical task of monitoring the aircraft’s flight path during the turn and he did not notice that the aircraft had entered an overbank situation while maintaining approximately 5g, inducing a very rapid descent. [3.1.20]

3.2.2. A TAWS recovery, if attempted by Swift 32, was ineffective in preventing the aircraft from striking the ground. [3.1.22]

Latent Cause Factors

3.2.3. CF188 pilots do not specifically practice TAWS recovery techniques in the simulator. [3.1.23]

Return to top of page

4. PREVENTIVE MEASURES

4.1. Preventive Measures Taken

4.1.1. The Fighter Force Training Rules were amended to clarify that weapons drops from altitudes less than 1,000 ft AGL are not authorized. [3.1.6]

4.2. Preventive Measures Recommended

4.2.1. Recommend all CF188 squadrons re-enforce the principles of LLAT to their pilots to further emphasize both the inherent risks of manoeuvring in the low-level environment and the recommended strategies for ensuring that a safe flight path is maintained. [3.2.1]

4.2.2. Recommend that the CF188 simulator syllabus be amended to include training in TAWS limits and recovery techniques. [3.2.2]

4.3. Other Safety Measures Recommended

Nil.

4.4. Airworthiness Investigative Authority remarks

4.4.1. Based on the evidence, this was a preventable accident. Prior to this accident, it had been 24 years since the last low level flying related fighter (CF116) CFIT and 26 years since the last CF188 low level flying related CFIT. While that time span is notable, this accident once again demonstrates that old but hard learned lessons can sometimes be forgotten and that the low level environment is an inherently hazardous and unforgiving region where only a few seconds of distraction can mean the difference between life and death.

4.4.2. Finally, it is concerning that something as basic as a minimum flying altitude could be interpreted differently by pilots within the same community. It appears the intent of the minimum altitude published in the FFTR was lost over time and, to at least a few pilots, new interpretations eventually crept in. This exemplifies the requirement for clear, unambiguous direction and the importance of strong, well communicated standards. Nevertheless, I am satisfied as the AIA that the actions taken to date, together with those recommended, will minimize the chance of a similar accident in the future.

 

J.J. Alexander
Colonel
Airworthiness Investigative Authority

Return to top of page

 Notes

[1] Unless otherwise specified, all times in the report are local Mountain Standard Time, which is seven hours behind Universal Coordinated Time (UTC or Z).

[2] Royal Canadian Air Force Fighter Force Training Rules, AL7, issued 14 January 2014

[3] Hot turn/hot crew swap means that the aircraft is not fully shutdown between sorties. It involves shutting down the left engine to allow the pilot to safely disembark and a new pilot to board. This practice is conducted regularly to avoid potential maintenance problems that can occur with a cold start and to maximize aircraft availability for the planned flying program.

[4] The SNIPER XR targeting pod (Sniper Pod) is used for long-range target detection/identification and continuous stabilized surveillance as well as laser tracking and laser designation of targets. Real-time imagery from the pod is presented on cockpit displays.

[5] Authority Gradient refers to the established, and/or perceived, command and decision-making power hierarchy in a team, crew or group situation, and also how balanced the distribution of this power is experienced within the team, crew or group. Concentration of power in one person leads to a steep gradient, while more democratic and inclusive involvement of others results in a shallow gradient.

[6] See paragraph 1.11.5 for a description of the Air Combat Manoeuvring Instrumentation system.

[7] Publication C-12-188-NFM/MB-003, Version 2010-10-01, Change 5 - 2016-04-01

[8] Link 16 is a military tactical data exchange network used by aircraft as well as ships and ground forces to exchange their tactical picture in near-real time.

[9]  Reference Chapter 4 of the 2017 Weather Update and Review, published by 1 Canadian Air Division Headquarters

[10] B-GA-100-001/AA-000, National Defence Flying Orders Book 1 - Flight Rules, dated 1 October 2013. These are the highest level flying orders in the CAF and have the greatest precedence.

[11] C21XB1.1 Grey book section 2.18

[12] Canadian Aviation Regulations Part 724.115 Training Programs, paras (9) and (32)

 

Return to top of page

Annex A - List of Abbreviations

Abbreviation - Meaning

A/G - air-to-ground

ACMI - Air Combat Manoeuvring Instrumentation

AGL - above ground level

AIS - airborne instrumentation subsystem

ASL - above sea level

CATM - captive air training missile

CFIT - controlled flight into terrain

CLAWR - Cold Lake Air Weapons Range

D-CVRS - digital video recording system

DEL - direct electrical link

DFS - Directorate of Flight Safety

DOO - Duty Operations Officer

ELLAT - electronic low level awareness training

FCC - flight control computer

FFTR - Fighter Force Training Rules

FPA - flight path angle

FSIMS - Flight Safety Information Management System

GFA - graphical area forecast

(E)GPWS - (enhanced) ground proximity warning system

HMD - helmet mounted display

HSD - horizontal situation display

HSI - horizontal situation indicator

Annex A1010-CF188-171345 (DFS 2-7) 12 January 2018

Abbreviation - Meaning

HUD - head-up-display

INS - inertial navigation system

JHMCS - joint helmet mounted cueing system

KIAS - knots indicated airspeed

KM - kilometer

LLAT - low level awareness training

LGTR - laser guided training round

LEF - leading edge flap

LVDT - linear variable directional transducer

M - magnetic

MAC - mean aerodynamic chord

MDI - multi-purpose display indicator

MST - Mountain Standard Time

NM - nautical mile

NOTAM - notices to airmen

NRC - National Research Council

PER - periodic inspection

RARM - Record of Airworthiness Risk Management

RCAF - Royal Canadian Air Force

RMM - removable memory module

SAR - search and rescue

SAR Tech - Search and Rescue Technician

SIM - simulator

TAF - terminal aerodrome forecast

TAWS - terrain awareness and warning system

TFS - Tactical Fighter Squadron

UTC - coordinated universal time

VFR - Visual Flight Rules

Return to top of page

Annex B - Figures

Figure 1: Cold Lake and CLAWR with Swift Flight departure route depicted

Figure 2: Aerial depiction of the target area and final flight path

Figure 3: ACMI screenshot cockpit view taken at maximum bank angle

Figure 4: Aerial view of the crash site

Figure 5: Right wing

Figure 6: Engine compressor section

Figure 7: Right vertical stabilizer

Figure 8: Removable memory module as found in the debris field

Figure 9: Parachute as found in the debris field

Return to top of page

Date modified: