WPR23MA113

The pilot’s loss of control due to spatial disorientation while operating in night instrument meteorological conditions, which resulted in an in-flight breakup. Contributing to the accident was the disengagement of the autopilot for undetermined reasons, as well as the operator’s insufficient flight risk assessment process and lack of organizational oversight.

The pilot, two medical crew members, and two passengers departed on the medical transport flight, which was operating on an instrument flight rules (IFR) flight plan in night instrument meteorological conditions (IMC). Onboard data and ADS-B flight track information showed that, between 1 and 3 minutes after takeoff, the autopilot disengaged and then reengaged; however, the airplane continued to fly a course consistent with the published departure procedure. About 11 minutes after takeoff, the airplane turned about 90° right, away from the next waypoint along the departure procedure, and remained on that heading for about 47 seconds. Around this time, the airplane’s autopilot was disengaged again and was not reengaged for the remainder of the flight. Also, about this time, the airplane’s previously consistent climb rate stopped, and the airplane maintained an altitude of about 18,300 ft mean sea level (msl) for about 20 seconds, even though the pilot had been cleared to climb to 25,000 ft msl. The airplane subsequently turned left to a northeasterly heading and climbed to about 19,400 ft msl before entering a descending right turn. Shortly after entering the right turn, the airplane’s rate of descent increased from about 1,800 ft per minute (fpm) to about 13,000 fpm, and the rate of turn increased before ADS-B tracking information was lost at an altitude of about 11,100 ft msl, in the vicinity of the accident site. The distribution of the wreckage at the accident site was consistent with a low-altitude in-flight breakup. Examination of the airframe and engine revealed no evidence of mechanical malfunctions or failures that would have precluded normal operation, and data obtained from onboard recording devices showed that the engine was developing power at the time of impact. Postaccident examination of the autopilot, trim servos, and trim actuators did not reveal any evidence of mechanical malfunction or failures that would have precluded normal operation. Recorded weather conditions at the departure airport about the time of departure included 1 ¾ statute miles visibility and a cloud ceiling 1,700 ft above ground level (agl). The departure airport and surrounding areas had been impacted by significant winter weather throughout the day of the accident, and the pilot who was on call for the accident operator earlier that day turned down a flight request due to the weather conditions. Another air medical operator, who operated the same make and model airplane as the accident airplane, also turned down a request for a flight in the area due to the low visibility, turbulence, and icing conditions. The accident airplane was equipped for flight into known icing conditions. Review of weather information indicated that the airplane was likely operating in IMC above 6,000 ft msl, and likely did not accumulate much, if any, structural icing. Icing conditions would have been present between 5,000 and 10,000 ft msl. Although turbulence was likely present in the area, there was no evidence to support that the accident airplane encountered hazardous turbulence during the flight. The airplane was equipped with several sources of recorded data, including a central advisory and warning system (CAWS) computer. The CAWS computer, which captured autopilot status, among other parameters, was significantly impact damaged and missing one of the memory chips that stored time information; therefore, the two autopilot disengagements could only be identified as occurring in two-minute windows after elapsed takeoff time, with the first about 1 to 3 minutes after takeoff, and the second between about 2 and 4 minutes before the accident. There are several ways in which the autopilot could have been automatically or manually disengaged during the accident flight; however, based on the available CAWS data and examination of the airplane and system components, the reason for the two autopilot disengagements during the accident flight could not be determined. Following the second autopilot disengagement, the pilot would have been required to manually maintain control of the airplane while operating in IMC, which increased his susceptibility to spatial disorientation. The airplane’s subsequent flight path was consistent with a phenomenon known as a “graveyard spiral,” a sensory illusion in which a pilot believes they are flying in a wings-level descent; however, the airplane is actually in a descending turn. Attempts to arrest the descent by pulling back on the control yoke have the effect of tightening the turn and losing altitude at an increasing rate until the airplane’s structural limits are exceeded, resulting in an in-flight breakup, or until the airplane impacts the ground. Graveyard spirals are most common at night or in poor weather conditions where no horizon exists to provide a visual reference to correct misleading inner-ear cues. Autopsy of the pilot revealed a 3 cm fibroblastic meningioma (tumor) in the right parietal section of the brain. The parietal lobe is one of the four major components of the cerebral cortex and has a primary role in the integration of sensory information, including spatial and navigational information. The parietal lobe is also primarily responsible for the integration of visual and vestibular information. The presence and location of the tumor may have impacted the pilot’s ability to synthesize and respond to sensory interpretation from the conditions under which he was flying; however, it is also possible that the tumor may have been an incidental finding without any significant symptoms, and the pilot’s next of kin indicated no changes in his behavior or health before the accident. Based on the available information, whether the effects of the pilot’s undiagnosed brain tumor contributed to the accident could not be determined. The accident pilot was not permanently assigned to the base from which the accident flight departed; rather, he was classified as a “float” pilot, who rotated across the operator’s bases throughout the country. The operator did not have any formal training or mentoring procedures in place to ensure that local area knowledge was passed along to pilots new to a specific operating area; the investigation was unable to establish the pilot’s experience operating in night IMC over mountainous terrain. All three crewmembers of the accident flight were relatively new in their respective roles. The pilot was hired by the operator about five months before the accident, while both clinicians had been assigned to the airplane for about six months. The company’s website highlighted the “Three to say go, one to say no” protocol as a best practice among air ambulance providers that empowers any member of the flight team, for any reason, to raise a safety concern. For rotorcraft flights, the operator required that clinicians with less than one year of experience be paired with clinicians with more than one year of experience, a practice that leveraged the flight team’s collective experience to benefit flight safety. However, fixed-wing operations were not subject to this requirement. The operator’s procedures also required company dispatchers to inform flight crews if a flight had been turned down by another operator. Although the weather can change throughout the course of a given day, the fact that other pilots and operators turned down flights due to weather in the area on the day of the accident should have been relayed to both the pilot and medical crew as part of their decision-making process. However, review of communication logs did not indicate that the company’s dispatchers made the accident crew aware of the turndowns earlier on the day of the accident. The operator also required that a flight risk assessment be completed before each flight; however, no such assessment was located for the accident flight. Even if a risk assessment had been conducted, the crew’s

The Current Icing Potential imagery for the accident region provided by the National Center for Atmospheric Research (NCAR) showed that a 10 to 20% chance of trace to light icing existed between 14,000 ft and 16,000 ft at the time of the accident. A High-Resolution Rapid Refresh (HRRR) model sounding for near the accident site at 2100, using an elevation of 4,418 ft, was retrieved from the National Oceanic and Atmospheric Administration Air Resources Laboratory and analyzed by the RAwinsonde OBservation (RAOB) program. Clouds were identified by RAOB between about 5,000 and 28,000 ft msl. The freezing level was at the surface, and RAOB identified the presence of light icing below about 15,000 ft, with moderate icing between about 9,500 and 11,500 ft msl. RAOB identified the potential for “severe” turbulence between about 11,000 and 12,000 ft msl and between about 15,000 and 16,500 ft msl. The wind at about 15,000 ft msl was from the southeast about 30 knots. NCAR-provided HRRR data identified positive values of liquid water content between 5,000 ft and 10,000 ft along the accident airplane’s flight path, and RNO observations about the time of the airplane’s departure recorded overcast ceilings near 6,100 ft above mean sea level. Geostationary Operational Environmental Satellite (GOES)-18 infrared data imagery from 2111 showed temperatures over the accident site were about -41°C, which corresponded to cloud top heights about 24,000 ft msl. There were no Convective SIGMET advisories active at the time of the accident. At 1829, SIGMET Victor 2 was issued by the National Weather Service Aviation Weather Center (AWC) for an area to the south through east of the accident location that was valid until 2228 and advised of occasional severe turbulence below 15,000 ft due to strong low-level winds, mountain wave activity, strong updrafts, and low-level wind shear. Data from pilot reports, model data, and weather-reporting aircraft showed a turbulent environment at various altitudes. Analysis and additional data provided by NCAR showed the presence of turbulence due to wind shear and the possibility of mountain wave action; however, the derived turbulence severities were not considered hazardous, and there was no evidence of significant turbulence despite the environment, though severe turbulence could not be ruled out. Around the time of the accident, an airline crew was descending into RNO in the vicinity of the accident site. That crew reported that they were in “complete IMC throughout our entire descent.” The crew also reported that they encountered light to moderate turbulence and some light rime ice in the area. The pilot was hired on September 6, 2022, as a float pilot who rotated across Guardian Flight’s bases throughout the country. The accident pilot had requested extra shifts, and was originally given an assignment out of Yuma, Arizona; however, due to staffing issues, he was assigned to the RNO base for a week rotation. According to an assistant chief pilot who oversaw the float pilot program, the accident pilot had previously flown two shifts in Ely, Nevada, with one of those shifts within the 4 to 6 weeks before the accident. The assistant chief pilot reported that the accident pilot was familiar and comfortable with the RNO area. The accident pilot arrived on Monday, began with a day shift on Tuesday, and then a night shift on Wednesday before the accident occurred on Friday. The following information about the pilot’s training was provided by a representative of Guardian Flight, LLC: A review of the pilot’s training record from the operator indicated that he began his initial PC-12 flight training on October 5, 2022, and concluded on October 10, 2022, for a total of 8.1 hours. On the flight dated October 10, 2022, of the 53 flight subject training areas, 33 subject areas were graded “S” or “satisfactory” and 20 were graded “W” or “waived.” Of those 53 flight subject training areas, one was titled “autopilot system” and another was titled “Nav and Avionics System”; on two previous separate flights (October 5th and 7th), the pilot received a grade of “U” or “unsatisfactory” for those areas and subsequently graded satisfactory on October 9, 2022. Training records also indicated that, between September 19, 2022, and September 23, 2022, the pilot received a total of 7.1 hours of simulator training. All grades were marked as either a “1” or a “2.” A grade of “1” was considered “Proficient” and a grade of “2” was considered “Normal Progress.” Before being hired by the accident operator, the pilot worked as a Cessna 208 pilot for a cargo operator based in Michigan. A review of training records indicated that the pilot was initially hired by that operator on November 14, 2021. His most recent CFR 135.293, 135.297, 135.299 checks were completed on June 9, 2022, during his employment with that operator. Of the 30 entries, all indicated “Satisfactory”; however, the entry for item #24 “Approaches: GPS” indicated that the first attempt was unsatisfactory, and the second attempt was satisfactory. The remarks stated: “Retrain and retested items #24 GPS. Satisfactory.” The accident pilot was employed from May 10, 2021, until August 12, 2021, by a Part 121 airline. The airline reported that the accident pilot was unable to satisfactorily complete the training program. The reasons provided were pre-departure, climb, descent, and approach procedures. Autopilot The accident airplane was equipped with a Bendix/King KFC 325 Digital Flight Control System (AFCS). The KFC 325 Digital AFCS had three axis controls for pitch, roll, and yaw. The Bendix/King Digital Flight Control System Pilot’s Guide provided, in part, the following information about the KFC 325 flight control system: The KFC 325 monitors autopilot operations continuously through sensors that monitor the aircraft's pitch attitude and acceleration, as well as servo motor operation. If monitors in the KFC 325 detect a problem, the autopilot will disconnect, illuminate a flashing AP annunciation, and provide an aural disconnect tone. If an autotrim failure is detected, the TRIM annunciator on the mode controller illuminates and the trim fail tone sounds. If a manual electric trim failure is detected, the TRIM annunciator illuminates and the trim fail tone sounds. The malfunction continues until the pilot takes action to stop it. In event of autopilot or flight director malfunction pay primary attention to basic aircraft control prior to attempting to diagnose the exact nature or cause of system failure. Once aircraft control is assured, the crew may attempt to reengage the affected autopilot or flight director mode by pressing the related mode pushbutton. Autopilot Emergencies The KFC 325 Pilot’s Guide stated, in part, in the event of an autopilot malfunction, the flight crew should immediately execute the following procedures: 1. Airplane Controls - GRASP FIRMLY AND REGAIN AIRCRAFT CONTROL. 2. Simultaneously PRESS AND HOLD the Autopilot Disconnect/Trim Interrupt Pushbutton located on the yoke. Autopilot and yaw damper will disconnect and trim power is interrupted. 3. While HOLDING the Autopilot Disconnect/Trim Interrupt Pushbutton, pull the autopilot circuit breaker. 4. After the autopilot has been disengaged, DO NOT REENGAGE. Resume normal manual flight operations. 5. Refer to aircraft flight manual supplement for procedures. There were seven means by which the pilot could manually disconnect the autopilot, including the autopilot mode pushbutton, the autopilot disconnect/trim interrupt switch, the manual electric trim switch, the go-around pushbutton, the autopilot circuit breaker, the autopilot power switch (if installed), and the avionics master switch. CAWS The CAWS integrates the display functions of aircraft systems. The CAWS co