Agusta AW139 accident near Big Grand Cay, Bs, July 4, 2019

The NTSB's narrative final · quoted from the NTSB record

The pilot-in-command (PIC) and second-in-command (SIC) were conducting a personal flight from the Bahamas to Fort Lauderdale, Florida, with five passengers onboard. The night flight was conducted under visual flight rules. About 2324 the day before the accident, the helicopter and company owner contacted the PIC, who was his friend and confidante, and told him that he needed him to conduct the flight to transport his daughter and her friend from Big Grand Cay, Abaco, Bahamas, to the United States for medical treatment. About 20 minutes later, the PIC contacted the SIC telling him he needed him to conduct the flight with him. The flight from Florida landed in Big Grand Cay at 0142. At 0145, the PIC filed an instrument flight rules flight plan, but it was not activated. While the flight crew was on the ground, the cockpit voice recorder (CVR) did not record them conducting a formal preflight instrument flight briefing. The flight crew’s pretakeoff conversation was limited to discussing flight plan information, including altitude, heading, and navigation; programming the flight computer; and the number of passengers expected on board. They did not discuss how to take off in night, visual meteorological conditions over water or their roles and responsibilities. The flight crew had a short discussion about the use of the flight controls and their automated functions during takeoff. Thus, their limited planning and communication for the takeoff from Big Grand Cay was indicative of inadequate crew resource management (CRM). According to flight data recorder data, the helicopter departed about 0152. The helipad from which they departed was brightly lit with floodlights, but then the helicopter proceeded over water in dark night conditions with no visible moon, likely zero ambient illumination, and no visible horizon, which would necessitate the pilots’ reliance on the instruments in order to fly because of the very limited outside cues. After takeoff, the PIC, who was the pilot flying, manipulated the cyclic and antitorque control pedals, engaged the collective pitch trim, and began the helicopter’s first climb to about 190 ft. The cyclic force trim release (FTR) switch was engaged and remained engaged for the entire flight, indicating that the pilot was controlling the cyclic motion. Subsequently, the helicopter began to descend and the airspeed increased, all while the cyclic’s position continued to move forward to a more nose-down attitude. The first of numerous enhanced ground proximity warning system (EGPWS) warnings began and continued during the descent. About 0152:50, while at an altitude of about 110 ft descending about 1,380 ft per minute (fpm), one of the pilots engaged the autopilot in the altitude acquire (ALTA) mode with indicated airspeed hold, which set a vertical speed reference target of +1,000 fpm and an airspeed reference target of about 110 knots. Nearly simultaneous to the ALTA mode activation, the collective FTR switch was momentarily activated. Because the helicopter was descending at that time and the target altitude for ALTA was above the helicopter’s current altitude, the ALTA rate of climb was reset to +100 fpm (per system design), where it remained for the rest of the flight. Despite the repeated EGPWS warnings, the PIC continued commanding forward cyclic and the helicopter continued to descend. About 0152:51, with the helicopter about 52 ft above the water, the PIC pulled the cyclic back and initiated a second climb. He then asked the SIC for the altitude, and, not receiving a response, stated that the helicopter was at 300 ft, and the SIC advised him that the helicopter was not at 300 ft and that it was "diving." It is likely that the PIC confused the vertical speed indication with the altitude indication, as the helicopter was at 116 ft radio altitude but was climbing about 300 fpm at the time. Subsequently, multiple EGPWS warnings annunciated until the helicopter climbed above 150 ft and the warnings stopped. Although the PIC and SIC each made comments during the remainder of the flight, there did was no apparent coordination or troubleshooting between them, further indicative of a lack of CRM. When near the top of the climb, the collective pitch trim increased about 5% per second, with a corresponding increase in engine torque and power index (PI) values. After activation of ALTA mode, the PI levels began to increase to a point where the PI limiting function, as part of the flight director, began restricting collective movement, which prevented the ALTA mode from maintaining a positive vertical speed and climb to the set altitude. Because the PIC was manually controlling the cyclic, the flight director was unable to compensate for the high PI levels, such as reducing airspeed; thus, the flight director had to reduce collective to prevent a PI level exceedance. Given the lack of discussion about the negative vertical speed or any attempts by the PIC to manually manipulate the collective, it is likely neither pilot was adequately monitoring the vertical speed and altitude trends, which led to a loss of altitude. About 0153:13, as the helicopter began to descend from 212 ft because the cyclic was moved forward again to command a nose-down attitude and the EGPWS warnings began to annunciate again, the SIC stated that “this is exactly what happened” in a fatal accident in the United Kingdom in which the accident was caused by somatogravic illusion and subsequent spatial disorientation. The PIC did not respond to the SIC, likely due to his continued confusion about the helicopter’s position in space and his misunderstanding of the information on the helicopter’s flight instruments. The helicopter then entered a left descending turn in a nose-down attitude with airspeed and engine torque increasing, significant forward cyclic being applied, the descent rate increasing, and EGPWS warnings continuing. The PIC repeatedly asked for a heading and once for altitude, but the SIC did not respond. As the helicopter continued descending toward the water, the flight crew did not communicate the helicopter’s attitude, energy state, and steps needed to recover from the descent. Given that postaccident examination indicated the helicopter’s flight instruments were operational (and they were operational for the flight to the Bahamas), they had information available to them to understand the helicopter’s flightpath. However, about 0153:22, the helicopter impacted water at high speed while in a nose-down, left-bank attitude. As the pilot transitioned the helicopter to forward flight by commanding forward cyclic, the flight crew appeared initially unaware of the helicopter’s first descent until multiple EGPWS warnings annunciated. The PIC likely perceived that the accelerations associated with the helicopter’s increasing forward airspeed was the helicopter pitching up and he provided control inputs that caused the helicopter to descend. These improper control inputs during the second descent were consistent with the onset of a type of spatial disorientation known as somatogravic illusion, and the PIC likely did not effectively use his instrumentation during the departure to recognize the helicopter’s flightpath and orientation. The CVR indicated that the SIC recognized and announced the helicopter’s first descent to the PIC. In response, the PIC likely selected ALTA, which contributed to the recovery of the altitude lost from the first descent. However, the PIC continued to command forward cyclic (using the FTR switch), leading to the helicopter’s second descent. Again, numerous EGPWS warnings annunciated, but the PIC continued decreasing the helicopter’s pitch attitude while the airspeed and descent rate increased; these inputs were also consistent with spatial disorientation and a failure to rely on the helicopter’s instruments. Based on the sequence of events and the flight crew’s actions and comments, they lost awareness of the helicopter's flightpath after takeoff over water during dark night conditions, which likely led to spatial disorientation and the subsequent collision with water. The PIC’s night flight experience and instrument currency could not be determined. The SIC was reportedly night current but it could not be determined if he was night current in the helicopter make and model. Further, the PIC and the SIC had never flown to Big Grand Cay at night. Given both pilots’ many hours of flight experience, it is likely the PIC recognized the risk associated with the intended flight and contacted the SIC to make the flight with him. The PIC’s comfort flying with the SIC likely contributed to his decision to take the flight. Further, the urgency of the mission and the direct communication from the helicopter owner likely created external pressure on the flight crew, which can affect decision-making and create a sense of pressure to complete a flight. However, no records were found that the flight crew evaluated or planned for the impact of external pressure on their flights to and from Big Grand Cay in dark night conditions to transport ill passengers to a hospital. It is likely that they allowed the external pressure to affect their decision to conduct the flight even though neither of them had ever flown to Big Grand Cay at night. Examination of the helicopter’s flight control system including autopilot system, structures, main and tail rotor system, and engines revealed no evidence of any preimpact mechanical failures or malfunctions that precluded normal operation. Although one of the four separated sections of tail rotor blades was not recovered, analysis of the recorded flight data as well as the CVR showed no evidence of anomalous operation of the tail rotor prior to impact. All observed damage was consistent with the helicopter’s impact with the water.

The complete narrative as the NTSB published it. The NTSB's docket holds the report as a PDF and any photographs, statements and other documents from the investigation.