WPR16FA158

A departure flight path that consisted of several unexplained turns during the initial climbout, and terminated in a high-speed descent and ground impact. The reason(s) for the turns and descent could not be determined due to lack of definitive evidence.

The airline transport pilot regularly used his twin-engine airplane to conduct volunteer flights for a non-profit organization dedicated to transporting medical cargo. On the day of the accident, he departed his home airport about 0945, made a planned interim stop at one airport, and then flew to a second airport, where he waited for his next cargo pickup. The cargo was delivered to him, as scheduled, about 2100, and he departed shortly thereafter on a night visual flight rules (VFR) cross-country flight. The air traffic control tower was closed at the time of the pilot's departure. The airport was situated in semi-rural, mountainous terrain, and both the sun and the moon had set about 2 hours before takeoff. Ground-based tracking radar data indicated that the airplane departed to the southwest, turned west, then south, and then north before it descended rapidly and impacted trees and terrain about 2.5 minutes after takeoff. The first segment of the radar-derived trajectory was consistent with a normal takeoff and initial climb. About 1 minute after takeoff, some undetermined occurrence(s) or circumstance(s) interrupted the climb and resulted in the course deviations and the extreme descent. Examination of the accident site indicated that the airplane was in a banked attitude when it impacted the trees. The available evidence indicated that both engines were developing significant power at impact, and that the propellers were operating normally. No evidence of a bird strike or an in-flight fire was observed. With the exception that one of the two instrument air pressure pumps were inoperative, the investigation did not discover evidence of any pre-impact mechanical deficiencies with the airplane or its equipment. However, the wreckage was highly fragmented, which could have masked or destroyed such evidence. The artificial horizon and the directional gyro were two of the flight instruments that were driven by the air pressure pumps. Although the two air pressure pumps and associated valving were designed to provide automatic continued normal system operation in the event of a single air pressure pump failure, the severity of the damage precluded determination of the pre-accident functionality of the instrument air system. Therefore, it is possible that a failure of the valving system could have resulted in the loss of valid artificial horizon and directional gyro information, which in turn, due to the dark night and scarcity of ground lights, could have resulted in the accident. Damage also precluded testing of the autopilot system. However, the pilot typically hand-flew the airplane to cruise altitude before engaging the autopilot; therefore, the accident was likely not due to an autopilot malfunction. Although the departure airport automated weather observation reported 10 miles visibility with no clouds less than 10,000 ft above the airport, the heavy rain earlier in the day, combined with nighttime cooling and the lack of a temperature-dew point spread, suggested the potential for localized low altitude clouds or fog. Given the dark night and the lack of significant ground lighting, pilot disorientation for a variety of reasons could not be discounted. The most likely possible reasons included: - Distraction from some unknown event inside or outside the airplane - Malfunction or failure of one or more flight instruments - Inadvertent encounter with localized instrument meteorological conditions The medical investigation of the pilot was significantly limited by the degree of injury, and it could not be determined whether the pilot experienced an acute medical event during the flight. Due to the pilot's age, his reported pre-existing medical conditions, and the paucity of autopsy findings, physiological impairment or incapacitation of the pilot could not be eliminated as a possible reason for the accident. Because it has no direct psychoactive effects, it is unlikely that the pilot's use of metoprolol contributed to the accident. The absence of ethanol in the kidney tissue indicates that the identified ethanol in muscle tissue was from post-mortem sources, and therefore did not contribute to the accident. Although the pilot was the president of the organization for which he was flying, and that the organization published flight operations guidance intended to ensure a minimum level of operational safety, the pilot's history demonstrated repeated deviations from that guidance. One of those deviations concerned the number of pilots on board. On the accident flight, as well as on most of his other night missions for the organization, the pilot flew solo, even though the guidance recommended that night flights use two pilots. Although the investigation was unable to determine the specific reason for the accident, it is possible that a second pilot might have been able to assist in some way to prevent the accident.

FLG was situated at an elevation of 7,014 ft msl, about 3 miles south of the city in semi-rural, mountainous terrain. The surrounding terrain for several miles to the east, south, and west was sparsely populated, with little ground lighting. FLG was equipped with a single paved runway, designated 3/21, which measured 8,800 ft by 150 ft. The runway was lighted. The airport was equipped with an air traffic control tower, but the tower had closed for the night at 2100. Flights for Life Information According to its website, FFL is a non-profit organization dedicated to providing free air transportation to transport blood for UBS. FFL works in cooperation with hospitals, blood banks, health-care agencies, and private individuals, and flies scheduled and on-demand missions, primarily within Arizona. The pilot was a well-known, long serving, and active member of FFL. The pilot became a member in 2006. He spent 4 years as the FFL Mission Safety Officer and then became the FFL President. At the time of the accident, the pilot had been the president for 3 years. The January 2016 edition of the FFL Member Handbook, which was current at the time of the accident, specified the particular conditions, requirements, limitations, and recommendations for the FFL pilots and the conduct of the mission flights. The pilot's status as the FFL President did not relieve him from complying with any of the guidance in the FFL Member Handbook. Comparison of the Member Handbook guidance with the known circumstances of the accident flight revealed several instances of non-compliance. The following two paragraphs, excerpted from the FFL Member Handbook, list the pilot requirements for day and night cargo flights. Pilot Requirements for Day Cargo Flights. 1. 100 Hours minimum [pilot-in-command] PlC time. 2. 5 Hours minimum time in aircraft make and model to be flown, or 10 Hours minimum if in high performance aircraft. 3. 25 Hours minimum time flying cross-country. 4. Be in an "Active" pilot status. Pilot Requirements for Night Cargo Flights. 1. Same requirements as for day Cargo flights, except as listed below: a. 300 Hours minimum PlC time, or 200 hours minimum with Instrument Rating. b. 50 Hours minimum time flying cross-country. c. 30 Hours minimum night flight, or 15 hours with an Instrument Rating. The handbook did not specify any instrument currency requirements for either day or night flights. The handbook stated, "All FLIGHTS FOR LIFE, INC., flights will be conducted in accordance with Visual Flight Rules (VFR), except [that] Marginal VFR and IFR [instrument flight rules] to VFR on Top flights are permissible, but are restricted to daylight hours only." The forecast and observed weather conditions were consistent with the potential for IFR conditions enroute to FFZ. The handbook also stated, "It is strongly recommended that flight plans be filed for all flight distances greater than 50 nautical miles." The straight-line distance between FLG and FFZ was about 101 nautical miles (nm), but the pilot did not file a flight plan for the accident flight. The handbook recommended two pilots for night flights. In contrast, the pilot was the only pilot planned for, or onboard, the airplane. The accident flight was not the first occurrence of such a situation for this pilot; review of FFL flight records as far back as 2008 indicated repeated occurrences of solo night flights. Gyroscopic Flight Instruments The airplane manufacturer's Pilot's Operating Handbook (POH) included the following information about the airplane's flight instrumentation. "The directional gyros and attitude indicators are driven by positive air pressure. The pressure system consists of a pressure pump on each engine, plus plumbing and regulating equipment. Check valves and a pressure air manifold…allow pressure instruments to function during single-engine operation, or in the event of malfunction of one of the pressure pumps.…Operation of the gyro pressure system can be monitored via a pressure gauge on the instrument panel, to the left of the copilot's control wheel shaft. The two warning indicators mounted on the gauge serve to alert the pilot should one of the engines be producing less than sufficient pressure to operate the gyro instruments. Additional warning of a possible malfunction in the gyro pressure system is provided by a light in the annunciator panel." There was no specific POH procedure for failure of an instrument air pressure pump; the only related POH procedures were for a decrease in instrument air pressure below the minimum allowable limit of 4.5 inches of mercury, which would not typically occur with the failure of one pump. The POH guidance stated that, if the pressure decreases below the minimum limit, engine speed should be increased to 2,575 rpm, and the airplane should be descended to an altitude at which a pressure of 4.5 inches of mercury can be maintained. The investigation did not locate any records of any recent pre-accident problems or anomalies with the flight instruments. Damage to the flight gyroscopic instruments precluded determination of their functionality, or whether they were operating during the flight or impact. The instrument air pressure pumps were recovered and shipped to the manufacturer (Tempest) for examination. The pump installed on the left engine was model number AA3215CC, and the pump installed on the right engine was model AA442CW-6. The pumps were examined in late June 2017 at the Tempest facility, with FAA oversight. According to the manufacturer's report, the vanes in the left pump were found to be "worn well beyond" their service limits. Powdery residue found inside that pump indicated that the left pump was inoperative at the time of impact. The internal condition of the right pump was consistent with it operating at the time of impact. Impact damage to the air system lines, check valves, and other components precluded determination of the pre-accident functionality of the system. Detailed information regarding the instrument air system and the pump examinations is contained in the NTSB public docket for this accident. Flagstaff Information The 2057 FLG automated weather observation included calm winds, visibility 10 miles, scattered clouds at 10,000 ft above ground level (agl), temperature 15°C, dew point 14°C, and an altimeter setting of 30.35 inches of mercury. The 2157 observation included winds from 240° at 3 knots (kts), visibility 10 miles, a broken cloud layer at 11,000 ft, with unchanged temperature, dew point, and altimeter setting. The National Weather Service (NWS) surface analysis chart station model for Flagstaff for 2000 indicated a light westerly wind about 5 kts, clear skies, temperature 16°C, and dew point 12°C. The NWS national composite radar image at 2125 for the Flagstaff area depicted only very light intensity echoes associated with a dissipating area of echoes. Data from the NWS Flagstaff Weather Surveillance Radar-1988, Doppler (WSR-88D), which was located 41 miles southeast of the accident site, indicated the absence of any significant weather echoes. Geostationary Operational Environmental Satellite number 15 (GOES-15) data indicated that a broken to overcast layer of altocumulus to altostratus type clouds extended over the Flagstaff area and the accident site. The measured cloud top temperatures corresponded to cloud tops near 29,000 ft above mean sea level (msl) in that region. The terminal area forecast (TAF) for FLG, which has a coverage area radius of 5 statute miles, called for light and variable winds at 5 kts or less, visibility better than 6 miles, scattered clouds at 6,000 ft agl, broken ceiling at 10,000 ft agl, with a temporary period (between 1700 and 2000) of light rain showers and a broken ceiling at 7,000 ft agl. The area forecast for northern Arizona for the period of the accident was for a broken ceiling between 12,000 and 14,