Single-Pilot Overload: A Black-Hole Departure Into Lake Erie
---------- ADS -----------
A Cessna 525C (Citation CJ4, N614SB) crashed into Lake Erie moments after takeoff from Runway 24R at Cleveland’s Burke Lakefront Airport (KBKL) on Dec. 29, 2016, at 2257 EST. Darkness, windy and snowy conditions, marginal VMC existed at the airport at the time.
The pilot, his wife, two sons, a neighbor and the neighbor’s daughter were on board returning to Ohio State University Airport (KOSU) in Columbus where the airplane was based. The group had initially departed KOSU about 1730 and arrived at KBKL about 1800. The visit to Cleveland would be quick. Their mission was to attend a basketball game and head home. All six were killed and the airplane destroyed when it crashed, broke up and sank into the lake. The entire flight lasted less than 90 sec.
The two-week search and recovery efforts for the airplane and its occupants captured national media attention. The weather during the effort was miserable, the lake wind-blown and icy and the water murky. Nevertheless, local, state and federal public safety agencies ultimately recovered the victims and some of the airplane.
The pilot’s family and his guests were well-known in Columbus. He was CEO of a beverage distribution company and an enthusiastic private pilot. He had accumulated 1,205 hr. total time, 427 of them in multiengine turbine airplanes. He had logged 121 hr. of nighttime flying and 145 hr. of actual instrument time.
The pilot earned a Cessna 525 (CJ4) single-pilot type rating on Dec. 8, 2016, after completing an FAA check ride in the accident airplane. The rating was added to his private pilot certificate that already included ratings for single-engine and multiengine airplanes, instrument and helicopter. Two weeks later, he completed a CJ4 simulator-based recurrent training course at FlightSafety International.
At the time of the accident, he had a total of 56.5 hr. in the CJ4. Of that time, 8.7 hr. were as pilot-in-command, which included the practical test. His most recent logged flight was on Dec. 17 from Orlando International Airport (KMCO) in Florida to KOSU. His night experience in the CJ4 totaled 16.8 hr.; actual instrument in make and model totaled 2.5 hr., and simulated instrument totaled 16.1 hr.
The pilot had owned a Cessna 510 Mustang for about two years before purchasing the CJ4. He had logged 372.9 hr. total time in that aircraft. He was in good health with no significant medical issues.
On Dec. 29, the day of the accident, the pilot arose about 0600 and was in the office by 0800. He visited a construction site and returned to the office after lunch. He left for the day around 1600. By the time he started the engines for the return flight to KOSU, he had been up and about for 17 hr.
The pilot had received weather briefings before leaving Columbus and they were reasonably accurate — IMC but not especially difficult for the CJ4’s avionics and systems although perhaps challenging for a tired pilot in a new airplane.
The observations from KBKL and Cleveland Hopkins International (KCLE) indicated that marginal visual conditions prevailed at the time of the accident. Precipitation was reported in the 1-min. observations at KBKL until 2251, with none reported at the surface until 2342. While the surface temperature remained above freezing after the airplane landed at the lakefront facility and about the accident time, the dew point temperature remained below freezing the entire time with precipitation occurring on and off in the snow shower activity.
At 2253, the observed conditions at KBKL were: wind, 260 deg. at 25 kt., with gusts to 31 kt.; 8 mi. visibility in light snow; scattered clouds at 1,200 ft., broken clouds at 2,200 ft., overcast clouds at 3,200 ft.; temperature, 1C, dew point, -2C; and altimeter, 29.74.
At 2300, the conditions at the airport were: wind, 260 deg. at 22 kt., with gusts to 31 kt.; 9 mi. visibility; scattered clouds at 1,500 ft. AGL, broken ceiling at 2,300 ft. AGL, overcast skies at 3,900 ft. AGL; temperature, 1C, dew point, -2C; and altimeter, 29.74.
There were no SIGMET advisories valid for Burke Lakefront at the accident time; however, AIRMET advisories Zulu, Tango and Sierra, issued at 2145, were valid and forecast IFR conditions due to precipitation and mist, moderate icing conditions below 10,000 ft. MSL, and moderate turbulence below 10,000 ft. MSL. The possibility of a trace to light icing was present at low altitudes at the time of the accident.
How It Happened
NTSB investigators reviewed ATC communications, the CVR transcript, ADS-B data and FADEC unit data to create the following timeline:
At 2247, the pilot contacted the KBKL tower controller and requested an IFR clearance.
At 2250, he requested a taxi clearance, and 5 min. later informed the tower controller he was holding short of Runway 24R and ready for takeoff. The controller subsequently cleared him for takeoff and instructed him to turn right to a heading of 330 deg. and maintain 2,000 ft. MSL after departure. The pilot acknowledged the clearance.
At 2256:33, the engine power increased, and 15 sec. later the airplane became airborne.
At 2257:09, an automated voice annunciated “altitude.” A second “altitude” annunciation followed 14 sec. later. (Safety Board investigators noted that in normal operations, the altitude preselect mode will provide an annunciation passing ±1,000 feet from the preselected altitude. Once tracking the selected altitude, the system will provide an alert if the airplane deviates more than 200 ft.)
At 2257:25, a sound similar to a decrease in engine power was recorded. Two seconds later, the EGPWS annunciated an excessive bank angle warning. (The EGPWS will provide a warning if the bank angle exceeds 50 deg. when the airplane is operating 210 ft. above ground level or higher.)
At 2257:29, about 2 sec. after the bank angle warning, the tower controller instructed the pilot to contact departure control. The CVR recorded “to departure six one four sierra bravo;” however, the tower controller did not receive that communication. (The CVR will record any audible sound that is picked up by the cockpit area microphone or a connected microphone such as on a pilot’s headset. However, any sound picked up by a pilot’s headset will not be transmitted unless the push-to-talk switch is simultaneously depressed. The presence of the pilot’s response on the CVR recording in conjunction with the absence on the ATC recording is consistent with the pilot not having the push-to-talk switch depressed.)
At 2257:37, the controller again attempted to contact the pilot. However, 2 sec. after the controller’s transmission, the EGPWS provided a “sink rate” warning to the pilot. The pilot again responded, “six one four sierra bravo,” but this was not received by the tower controller. (The EGPWS will provide a sink rate warning when the aircraft is within 2,450 ft. of the terrain. At 2,450 ft., the triggering descent rate is 5,007 fpm. This varies linearly to a descent rate of 964 fpm at 10 ft.)
Beginning at 2257:43, the EGPWS provided “pull up” warnings at 1.6-sec. intervals until the end of the CVR recording. During that time, a sound similar to the overspeed warning began, which continued until the end of the recording at 2257:58.
The tower controller made several additional attempts to contact the aircraft, then initiated search and rescue procedures.
ADS-B position data helped investigators re-create the flight profile:
The airplane had crossed the departure end of the runway and began a climbing right turn. (The bank angle steadily increased until 2257:31, when it reached about 62 deg. (right wing down). Over the next 14 sec., the bank angle decreased to about 40 deg. right wing down. The bank angle decreased further to about 25 deg. right wing down shortly before impact.)
About 2257:28, the airplane became established on a magnetic course of 310 deg. During that time, the airplane reached an altitude of approximately 2,925 ft. MSL. About 5 sec. later, the airplane entered a descending right turn that continued until the final data point that was recorded at 2257:52 and was located 1.83 mi. northwest of the airport. The altitude was 775 ft. MSL, about 205 ft. above the lake.
A Safety Board performance study indicated that after takeoff the airplane’s pitch attitude was about 5 deg. nose up for approximately 8 sec. The Citation accelerated to about 215 kt. The pitch attitude increased to about 16 deg. nose up and the rate of climb reached over 6,000 fpm during the initial climb.
Beginning about 2257:25 and continuing over the next 12 sec., the pitch attitude began to steadily decrease until reaching about 15 deg. nose down. The airplane accelerated to about 300 kt. and the rate of descent reached about 6,000 fpm once it became established in the descent. The maximum operating limit speed (Vmo) below 8,000 ft. is 260 KIAS.
The ADS-B data included information related to the altitude preselect and heading bug settings. The altitude preselect setting was consistent with the 2,000-ft. altitude assigned to the accident flight. The heading bug was set to 240 deg. until 2257:11. Over the following 12 sec., the heading bug was reset to 329 deg., where it remained for the duration of the flight. The CVR recording ended at 2257:58.
No visual cues were available to the pilot during this departure. KBKL is located along the south shoreline of Lake Erie within the metropolitan area of the city. The airport elevation is about 584 ft., which is approximately 14 ft. above the level of the lake. Visual cues are available from the city lights south of the airport. However, there is a lack of visual cues north of the airport due to the lake’s expanse.
Two local pilots discussed with investigators the lack of visual cues at the airport at night. One described the conditions as a “black hole” during a dark night, VFR departure. The second noted that turning toward Lake Erie and away from the lights of the city may result in “absolute darkness” for a pilot.
Two weeks of bad weather made recovery difficult. The airplane was fragmented. The nose section was not recovered other than the baggage compartment doors and the nose landing gear assembly, which had separated from the airframe. The cockpit window assembly — pilot and copilot windshields and side windows — was structurally intact. Each window was shattered but remained in place, except for portions of the copilot windshield outer pane, which was missing. The copilot’s windshield exhibited a red impact mark near the center of the windshield. (The mark was examined by ornithologists from the Smithsonian Institution — National Museum of Natural History; no evidence of bird feathers or DNA.)
No portions of the cockpit controls or instrumentation were recovered other than the center pedestal assembly that was separated and damaged. The right throttle lever, flap handle, engine run-stop switches, and friction adjustment handle were missing.
Sections of the left- and right-side fuselage structure were recovered but badly deformed. Smaller sections of the airframe structure including wing carry-through spar sections and structural splice assemblies also were recovered.
The aft, right-side cabin seat frame and the pilot and copilot seat frames were not recovered. Several seat cushions, including the pilot and copilot seat-back cushions were located.
The left and right wings had separated near the wing roots and showed extensive fragmentation and deformation.
The left engine was never found; however, the deformed left exhaust assembly was located. The right engine and its inlet and exhaust assemblies were recovered, but its forward bypass duct and fan case assemblies had separated and were lost. The fan blades remained attached to the hub. Each blade exhibited gouges, scraping and deformation (curling) at the blade tips consistent with rotation at impact.
The right fuel delivery unit was separated and recovered. The unit exhibited damage consistent with impact.
No anomalies consistent with a pre-impact failure or malfunction of the airframe structure, flight controls or engines were observed.
Interviews with his instructor determined that the pilot was trained to consistently use the autopilot after takeoff after reaching at least 300 ft. AGL. The instructor also told investigators the pilot had inadvertently pressed the autopilot transfer button instead of the autopilot engagement button on two occasions during training without recognizing the error.
Another possible factor — the CJ4’s attitude presentation was different from the Mustang’s. The attitude indicator presented by the PFD on the Cessna 525 was an egocentric (“inside out”) type display. An “inside out” perspective involves a fixed aircraft symbol and moving horizon similar to what a pilot sees when looking outside of the aircraft. On the other hand, the Cessna 510 utilizes an exocentric (“outside in”) display. An “outside in” perspective involves a fixed horizon and a moving aircraft symbol.
Studies have demonstrated, said the Safety Board, that pilot performance with sole experience using either of these types of displays is similar; however, performance degrades when experienced pilots switched between the two types of displays. This, of course, can contribute to spatial disorientation.
The flight guidance panel (FGP), located on the glareshield, allows the pilot to select manual or autopilot guidance for airplane control. The autopilot button is located on the upper row of button controls near the right side of the panel. Autopilot engagement is indicated in the flight control system display area along the upper portion of the primary flight display (PFD). There is no indication of the autopilot status on or near the autopilot button on the flight guidance panel.
Investigators said a comparison of the Citation 525 systems and those of the Citation 510 revealed that the autopilot engagement button on the latter is located in a slightly different location on the AFCS panel. Autopilot engagement is indicated along the upper portion of the PFD in both airplanes. In addition, an indicator light adjacent to the autopilot button on the AFCS panel is illuminated when the autopilot is engaged.
Spatial disorientation is always possible in black-hole departures (and many other situations.) The FAA Civil Aeromedical Institute’s publication, Introduction to Aviation Physiology, defines spatial disorientation as a loss of proper bearings or development of a state of mental confusion as to position, location or movement relative to the position of the earth.
Factors contributing to spatial disorientation, says the FAA, include changes in acceleration, flight in IMC, frequent transfer between VMC and IMC, and unperceived changes in aircraft attitude.
The Safety Board’s investigators also pointed to the FAA’s Airplane Flying Handbook (FAA-H-8083-3A), which describes some hazards associated with flying when the ground or horizon is obscured. The handbook states, in part: “The vestibular sense (motion sensing by the inner ear) in particular tends to confuse the pilot. Because of inertia, the sensory areas of the inner ear cannot detect slight changes in the attitude of the airplane, nor can they accurately sense attitude changes that occur at a uniform rate over a period of time. On the other hand, false sensations are often generated, leading the pilot to believe the attitude of the airplane has changed when in fact, it has not. These false sensations result in the pilot experiencing spatial disorientation.”
The Accident Chain
This CJ4 black-hole departure lasted less than a minute and a half. The pilot had less than 10 hr. command time in an airplane whose instrumentation and controls were slightly different from a similar airplane in which he had several hundred hours. During training, those slight differences had led the pilot to mis-select autopilot modes. He was also seeking guidance from a PFD that displayed dynamic attitude differently from that with which he was familiar. There were no visual cues from the black night outside the cockpit once the right turn began. Even routine departure communications went awry, perhaps because he failed to finger the push-to-talk button correctly.
A quick exploration of the NTSB accident database shows that black-hole departure accidents are not uncommon. Black holes have trapped all manner of pilots from solo pilots in single-engine trainers to airline crews in commercial operations.
Black-hole departures are essentially zero-zero takeoffs once the airplane rotates. They require planning, concentration and alertness. The abrupt loss of visual cues followed by low-altitude pitch changes and turning maneuvers is disorienting.
Despite all the hype popular media gives “multitasking,” humans can only attend to one thing at a time. (While we can move our attention quickly among a number of things, the results are not always positive.) Successful black-hole departures depend on the flight crew performing the basics one step at a time — rotate, establish climb, configure the aircraft, engage the FD/AP and keep your attention inside. Until all that is squared away, don’t even think about turning maneuvers. If something in the sequence isn’t working as planned, level the wings, climb and (when you can) let ATC know what you are doing.
At one time or another we’ve all been taught that the proper sequence of cockpit duties is to aviate, navigate and communicate. This sequence is nowhere more important than during the execution of a black-hole departure, and especially as a single pilot.