United States of America

According to both pilots, the takeoff, climb, and cruise portions of the flight were uneventful. The No. 1 (left) thrust reverser was not operational and deferred for the flight in accordance with the airplane’s minimum equipment list. The captain was the pilot flying for the accident flight, and the first officer was the pilot monitoring. The captain was also performing check airman duties for the first officer who was in the process of completing operating experience training. During the approach to Jacksonville Naval Air Station (NIP), the flight crew had two runway change discussions with air traffic controllers due to reported weather conditions (moderate to heavy precipitation) near the field; the pilots ultimately executed the area navigation GPS approach to runway 10, which was ungrooved and had a displaced threshold 997 ft from the threshold, leaving an available landing distance of 8,006 ft. As the airplane descended through 1,390 ft mean sea level (msl), the pilots configured it for landing with the flaps set at 30º and the landing gear extended; however, the speedbrake handle was not placed in the armed position as specified in the Landing checklist. At an altitude of about 1,100 ft msl and 2.8 nm from the runway, the airplane was slightly above the glidepath, and its airspeed was on target. Over the next minute, the indicated airspeed increased to 170 knots (17 knots above the target approach speed), and groundspeed reached 180 knots, including an estimated 7-knot tailwind. At an altitude of about 680 ft msl and 1.6 nm from the threshold, the airplane deviated further above the 3° glidepath such that the precision approach path indicator (PAPI) lights would have appeared to the flight crew as four white lights and would retain that appearance throughout the rest of the approach. Eight seconds before touchdown, multiple enhanced ground proximity warning system alerts announced “sink rate” as the airplane’s descent rate peaked at 1,580 fpm. The airplane crossed the displaced threshold 120 ft above the runway (the PAPI glidepath crosses the displaced threshold about 54 ft above the runway) and 17 knots above the target approach speed, with a groundspeed of 180 knots and a rate of descent about 1,450 ft per minute (fpm). The airplane touched down about 1,580 ft beyond the displaced threshold, which was 80 ft beyond the designated touchdown zone as specified in the operator’s standard operating procedures (SOP). After touchdown, the captain deployed the No. 2 engine thrust reverser and began braking; he later reported, however, that he did not feel the aircraft decelerate and increased the brake pressure. The speedbrakes deployed about 4 seconds after touchdown, most likely triggered by the movement of the right throttle into the idle reverse thrust detent after main gear tire spinup. The automatic deployment of the speedbrakes was likely delayed by about 3 seconds compared to the automatic deployment that could have been obtained by arming the speedbrakes before landing. The airplane crossed the end of the runway about 55 ft right of the centerline and impacted a seawall 90 ft to the right of the centerline, 9,170 ft beyond the displaced threshold, and 1,164 ft beyond the departure end of runway 10. After the airplane came to rest in St. Johns River, the flight crew began an emergency evacuation.

The pilot landed the seaplane into an easterly wind, then noticed that the surface wind was greater than forecast. Unable to taxi to the beaching location, he elected to return to his destination. He maneuvered the airplane into the wind and applied takeoff power. He described the takeoff run as "bumpy" and the water conditions as "rough." The pilot reported that the left float departed the airplane at rotation speed, and the airplane subsequently nosed into the water. The pilot and passengers were assisted by a nearby vessel and the airplane subsequently sank into 50 ft of water. Inclement sea and wind conditions prevented recovery of the wreckage for 52 days, and the wreckage was stored outside for an additional 13 days before recovery by the salvage company. Extensive saltwater corrosion prevented metallurgical examination of the landing gear components; however, no indication of a preexisting mechanical malfunction or failure was found.

The pilot began the takeoff roll in visual meteorological conditions. The airplane was airborne about 1,300 ft down the runway, which was about 75% of the normal ground roll distance for the airplane’s weight and the takeoff environment. About 2 seconds after rotation, the airplane rolled left. Three seconds later, the airplane had reached an altitude of about 80 ft above ground level and was in a 90° left bank. The nose then dropped as the airplane rolled inverted and struck the ground in a right-wing-low, nose-down attitude. The airplane was destroyed. Postaccident examination did not reveal any anomalies with the airframe or engines that would have precluded normal operation. The landing gear, flap, and trim positions were appropriate for takeoff and flight control continuity was confirmed. The symmetry of damage between both propeller assemblies indicated that both engines were producing equal and high amounts of power at impact. The autopsy revealed no natural disease was present that could pose a significant hazard to flight safety. Review of surveillance video footage from before the accident revealed that the elevator was in the almost full nose-up (or trailing edge up) position during the taxi and the beginning of the takeoff roll. Surveillance footage also showed that the pilot did not perform a preflight inspection of the airplane or control check before the accident flight. According to the pilot’s friend who was also in the hangar, as the accident pilot was pushing the airplane back into his hangar on the night before the accident, he manipulated and locked the elevator in the trailing edge up position to clear an obstacle in the hangar. However, no evidence of an installed elevator control lock was found in the cabin after the accident. The loss of control during takeoff was likely due to the pilot’s use of an unapproved elevator control lock device. Despite video evidence of the elevator locked in the trailing edge up position before the accident, an examination revealed no evidence of an installed control lock in the cabin. Therefore, during the night before the accident, the pilot likely placed an unapproved object between the elevator balance weight and the trailing edge of the horizontal stabilizer to lock the elevator in the trailing edge up position. The loss of control was also due to the pilot’s failure to correctly position the elevator before takeoff. The pilot’s friend at the hangar also reported that the pilot was running about one hour late; the night before, he was trying to troubleshoot an electrical issue in the airplane that caused a circuit breaker to keep tripping, which may have become a distraction to the pilot. The pilot had the opportunity to detect his error in not freeing the elevator both before boarding the airplane and again while in the airplane, either via a control check or detecting an anomalous aft position of the yoke. The pilot directed his attention to the arrival of a motorbike in the hangar alley shortly after he pulled the airplane out of the hangar, which likely distracted the pilot and further delayed his departure. He did not conduct a preflight inspection of the airplane or control check before the accident flight, due either to distraction or time pressure.

Two instrument-rated commercial pilots and one passenger were conducting a cross-country flight in instrument meteorological conditions when they began discussing an electrical malfunction; they then reported the electrical problem to air traffic control. The airplane subsequently made a descending right turn and impacted wooded terrain at a high speed. Most components of the airplane were highly fragmented, impact damaged, and unidentifiable. Based on the limited discussion of the electrical problem on the cockpit voice recorder and the damage to the airplane, it was not possible to determine the specific nature of the electrical malfunction the airplane may have experienced. While it was not possible to determine which systems were impacted by the electrical malfunction, it is possible the flight instruments were affected. The airplane's descending, turning, flight path before impact is consistent with a system malfunction that either directly or indirectly (through a diversion of attention) led to the pilot's loss of awareness of the airplane's performance in instrument meteorological conditions and subsequent loss of control of the airplane.

The two commercial pilots were conducting a personal, cross-country flight. A video surveillance camera at the airport captured their airplane’s approach. Review of the video revealed that, as the airplane approached the approach end of the landing runway, it began to climb, rolled left, became inverted, and then impacted terrain. The left thrust reverser (T/R) was found open and unlatched at the accident site. An asymmetric deployment of the left T/R would have resulted in a left roll/yaw. The lack of an airworthy and operable cockpit voice recorder, which was required for the flight, precluded identifying which pilot was performing pilot flying duties, as well as other crew actions and background noises, that would have facilitated the investigation. Postaccident examination of the airplane revealed that it was not equipped, nor was required to be equipped, with a nose landing gear (NLG) ground contact switch intended to preclude inflight operation of the thrust reverser (T/R). The left T/R door was found unlatched and open, and the right T/R door was found closed and latched. Further, electrical testing of the T/R left and right stow microswitches within the cockpit throttle quadrant revealed that the left stow microswitch did not operate within design specifications. Disassembly of the left and right stow microswitches revealed evidence of arc wear due to aging. Based on this information, it is likely that the airplane’s lack of an NLG ground contact switch and the age-related failure of the stow microswitches resulted in an asymmetric T/R deployment while on approach and a subsequent loss of airplane control. Also, there were additional T/R system components that were found to unairworthy that would have affected the control of the T/R system. Operational testing of the T/R system could not be performed due to the damage the airplane incurred during the accident. Toxicology testing results of the pilot’s specimens indicated that the pilot had taken diazepam, which is considered impairing at certain levels. However, the detected amounts of both diazepam and its metabolite nordiazepam were at subtherapeutic levels, and given the long half-life of these compounds, it appears that the medication was taken several days before the accident; therefore, it is unlikely that the pilot was impaired at the time of the accident and thus that his use of diazepam was a not factor in the accident.

The pilot departed on a short cross-country flight in the twin-engine airplane. Instrument meteorological conditions (IMC) were present at the time. While en route at an altitude of 3,000 ft mean sea level, the pilot reported that the airplane was "picking up icing" and that he needed to "pick up speed." The controller then cleared the pilot to descend, then to climb, in order to exit the icing conditions; shortly thereafter, the controller issued a low altitude alert. The pilot indicated that he was climbing; radar and radio contact with the airplane were lost shortly thereafter. The airplane impacted a field about 7 miles short of the destination airport. Examination of the airplane was limited due to the fragmentation of the wreckage; however, no pre-impact anomalies were noted during the airframe and engine examinations. Extensive damage to the pitot static and deicing systems precluded functional testing of the two systems. A review of data recorded from onboard avionics units indicated that, about the time the pilot reported to the controller that the airplane was accumulating ice, the airplane's indicated airspeed had begun to diverge from its ground speed as calculated by position data. However, several minutes later, the indicated airspeed was zero while the ground speed remained fairly constant. It is likely that this airspeed indication was the result of icing of the airplane's pitot probe. During the final 2 minutes of flight, the airplane was in a left turn and the pilot received several "SINK RATE" and "PULL UP PULL UP" annunciations as the airplane conducted a series of climbs and descents during which its ground speed (and likely, airspeed) reached and/or exceeded the airplane's maneuvering and maximum structural cruising speeds. It is likely that the pilot became distracted by the erroneous airspeed indication due to icing of the pitot probe and subsequently lost control while maneuvering.

The commercial pilot was conducting an aerial observation (surveying) flight in a piston engineequipped multiengine airplane. Several hours into the flight, the pilot advised air traffic control (ATC) that the airplane had a fuel problem and that he needed to return to the departure airport. When the airplane was 8 miles from the airport, and after passing several other airports, the pilot informed ATC that he was unsure if the airplane could reach the airport. The final minutes of radar data depicted the airplane in a descent and tracking toward a golf fairway as the airplane's groundspeed decreased to a speed near the single engine minimum control airspeed. According to witnesses, they heard an engine sputter before making two loud "back-fire" sounds. One witness reported that, after the engine sputtered, the airplane "was on its left side flying crooked." Additional witnesses reported that the airplane turned to the left before it "nose-dived" into a neighborhood, impacting a tree and private residence before coming to rest in the backyard of the residence. A witness approached the wreckage immediately after the accident and observed a small flame rising from the area of the left engine. Video recorded on the witness' mobile phone several minutes later showed the airplane engulfed in flames. Examination of the wreckage revealed no evidence of any preimpact mechanical malfunctions or failures of either engine. The fuel systems feeding both engines were damaged by impact forces but the examined components generally displayed that only trace amounts of fuel remained; with the exception of the left engine nacelle fuel tank. Given the extent of the fire damage to this area of the wreckage, and the witness report that the post impact fire originated in this area, it is likely that this tank contained fuel. By design, this fuel in this tank was not able to supply fuel directly to either engine, but instead relied on an electric pump to transfer fuel into the left main fuel tank. Fire damage precluded a detailed postaccident examination or functional testing of the left nacelle fuel transfer pump. Other pilots who flew similar airplanes for the operator, along with a review of maintenance records for those airplanes, revealed at least three instances of these pumps failing in the months surrounding the accident. The other pilots also reported varying methods of utilizing fuel and monitoring fuel transfers of fuel from the nacelle fuel tanks, since there was no direct indication of the quantity of fuel available in the tank. These methods were not standardized between pilots within the company and relied on their monitoring the quantity of fuel in the main fuel tanks in order to ensure that the fuel transfer was occurring. Had the pilot not activated this pump, or had this pump failed during the flight, it would have rendered the fuel in the tank inaccessible. Given this information it is likely that the fuel supply available to the airplane's left engine was exhausted, and that the engine subsequently lost power due to fuel starvation. The accident pilot, along with another company pilot, identified fuel leaking from the airplane's left wing, about a week before the accident. Maintenance records showed no actions had been completed to the address the fuel leak. Due to damage sustained during the accident, the origin of the fuel leak could not be determined, nor could it be determined whether the fuel leak contributed to the fuel starvation and eventual inflight loss of power to the left engine. Because the left engine stopped producing power, the pilot would have needed to configure the airplane for single-engine flight; however, examination of the left engine's propeller found that it was not feathered. With the propeller in this state, the pilot's ability to maintain control the airplane would have been reduced, and it is likely that the pilot allowed the airplane's airspeed to decrease below the singleengine minimum controllable airspeed, which resulted in a loss of control and led to the airplane's roll to the left and rapid descent toward the terrain. Toxicology results revealed that the pilot had taken doxylamine, an over-the-counter antihistamine that can decrease alertness and impair performance of potentially hazardous tasks. Although the toxicology results indicated that the amount of doxylamine in the pilot's cavity blood was within the lower therapeutic range, review of ATC records revealed that the pilot was alert and that he was making necessary decisions and following instructions. Thus, the pilot's use of doxylamine was not likely a factor in the accident.

The first instrument landing system (ILS) approach to runway 1 appeared to be proceeding normally until the first officer (the pilot flying) transitioned from instrument references inside the flight deck to outside references. During a postaccident interview, the first officer stated that he expected to see the runway at that time but instead saw “white on white” and a structure with an antenna that was part of the runway environment but not the runway itself. The captain (the pilot monitoring) stated that she saw a tower and called for a go-around. (Both flight crewmembers were most likely seeing the automated weather observing system wind sensor pole, which was located about 325 ft to the right of the runway 1 centerline and about 870 ft beyond the runway threshold, and the damage to the lightning arrester at the top of the wind sensor pole was likely due to contact with the accident airplane as it flew over the pole.) According to the cockpit voice recorder (CVR), after the go-around, the first officer asked the captain if she saw the runway lights during the approach. The captain responded that she saw the lights but that “it’s really white down there that’s the problem.” Airport personnel stated that snow plowing operations on the runway had finished about 10 minutes before the first approach. The CVR recorded the flight crew’s discussion about turning on the pilot-controlled runway lights and sounds similar to microphone clicks before and after the discussion. However, the PQI maintenance foreman stated that, after the first approach, the runway lights were not on. Thus, the investigation could not determine, based on the available evidence, whether the flight crew had turned on the runway lights during the first approach. The captain thought that the airplane had drifted off course when the first officer transitioned from flight instruments to the outside, so she instructed the first officer to remain on the instruments during the second approach until the decision altitude (200 ft above ground level [agl]). The second approach proceeded normally with no problems capturing or maintaining the localizer and glideslope. During this approach, the captain asked airport maintenance personnel to ensure that the runway lighting was on, and the PQI maintenance foreman replied that the lights were on “bright”(the high-intensity setting). Thus, the flight crew had a means to identify the runway surface even with the reported snow cover at the time. As the airplane approached the decision altitude, the captain instructed the first officer to disconnect the autopilot, which he did. About nine seconds later, the airplane reached the decision altitude, and the captain called, “runway in sight twelve o’clock.” This callout was followed by the first officer’s statement, “I’m stayin’ on the flight director ‘cause I don’t see it yet.” A few seconds later, while the airplane was below 100 ft agl, the captain and the first officer expressed confusion, stating “what the [expletive]” and “I don’t know what I'm see in’,” respectively, but neither called for a go-around. The airplane subsequently impacted the snow-covered grassy area between runway 1 and a parallel taxiway. During a postaccident interview, the first officer stated that, when he transitioned from flight instruments to the outside during the second approach, he again saw “white on white” as well as blowing snow and that the airplane touched down before he could determine what he was seeing. The maintenance foreman estimated that, at the time of the accident, the runway had about 1/8 inch of snow with about 20% to 25% of the runway visible.

The instrument-rated private pilot and passenger departed into instrument meteorological conditions with a 600-ft cloud ceiling in an airplane that was about 550 lbs over gross weight. Air traffic control data showed the airplane in a climbing left turn that continued beyond the assigned heading. After reaching 1,400 ft msl, the airplane continued turning left and its altitude and speed began to vary. The airplane continued in a left spiral, completing more than two full circles, then decelerated in a right turn and rapidly descended until impact with terrain. Examination of the flight control system revealed no evidence of mechanical malfunctions and downloaded engine data indicated normal engine operation. Downloaded data from the autopilot system revealed three in-flight error codes. The first error code, which likely occurred about 1 minute after takeoff, would have resulted in the autopilot, if it was engaged at the time, disengaging. The subsequent error codes likely occurred during the erratic flight profile, with the autopilot disengaged. Before the accident flight, the pilot had informed a mechanic, who is also a pilot, of intermittent issues with the autopilot system and that these issues were unresolved. The mechanic had flown with the accident pilot previously and assessed his instrument flying skills as weak. The flight instructor who provided initial flight training for the turbine engine transition stated the pilot's instrument flying proficiency was poor when he was hand flying the airplane. Toxicology testing revealed that the pilot had used marijuana, and his girlfriend stated the pilot would take a marijuana gummy before bedtime to sleep more soundly. However, given that no psychoactive compounds were found in blood specimens, it is unlikely that the pilot was impaired at the time of the accident. The instrument conditions at the time of the accident, the airplane's erratic flightpath, and the pilot's reported lack of instrument proficiency when flying by hand support the likelihood that the pilot experienced spatial disorientation sometime after takeoff. In addition, given the reports of the intermittently malfunctioning autopilot that had not been fixed, it is likely the pilot experienced an increased workload during a critical phase of flight that, in combination with spatial disorientation, led to the pilot's loss of airplane control.

On February 23, 2019, at 1239 central standard time, Atlas Air Inc. (Atlas) flight 3591, a Boeing 767-375BCF, N1217A, was destroyed after it rapidly descended from an altitude of about 6,000 ft mean sea level (msl) and crashed into a shallow, muddy marsh area of Trinity Bay, Texas, about 41 miles east-southeast of George Bush Intercontinental/Houston Airport (IAH), Houston, Texas. The captain, first officer (FO), and a nonrevenue pilot riding in the jumpseat died. Atlas operated the airplane as a Title 14 Code of Federal Regulations Part 121 domestic cargo flight for Amazon.com Services LLC, and an instrument flight rules flight plan was filed. The flight departed from Miami International Airport (MIA), Miami, Florida, about 1033 (1133 eastern standard time) and was destined for IAH. The accident flight’s departure from MIA, en route cruise, and initial descent toward IAH were uneventful. As the flight descended toward the airport, the flight crew extended the speedbrakes, lowered the slats, and began setting up the flight management computer for the approach. The FO was the pilot flying, the captain was the pilot monitoring, and the autopilot and autothrottle were engaged and remained engaged for the remainder of the flight. Analysis of the available weather information determined that, about 1238:25, the airplane was beginning to penetrate the leading edge of a cold front, within which associated windshear and instrument meteorological conditions (as the flight continued) were likely. Flight data recorder data indicated that, during the time, aircraft load factors consistent with the airplane encountering light turbulence were recorded and, at 1238:31, the airplane’s go-around mode was activated. At the time, the accident flight was about 40 miles from IAH and descending through about 6,300 ft msl toward the target altitude of 3,000 ft msl. This location and phase of flight were inconsistent with any scenario in which a pilot would intentionally select go-around mode, and neither pilot made a go-around callout to indicate intentional activation. Within seconds of go-around mode activation, manual elevator control inputs overrode the autopilot and eventually forced the airplane into a steep dive from which the crew did not recover. Only 32 seconds elapsed between the go-around mode activation and the airplane’s ground impact.