Birgenair 301 was a charter flight for Alas Nacionales Airlines. At 0342, Birgenair flight ALW-301, operated with a Boeing 757-200 aircraft, took off from Gregario Luperon International Airport, Puerto Plata, Dominican Republic, with a final destination of Frankfurt, Germany. The flight was originally scheduled to be flown with a Boeing 767. Maintenance problems on the original 767 prevented its use, so approximately 2 ½ hours prior to the flight a 757 was substituted, and a new crew was assigned.
During the takeoff roll, the captain’s airspeed indicator appeared to have failed. However, the captain, who was the pilot flying, ordered the first officer to call out reference airspeeds based on his own indicator, and the takeoff was continued, Once airborne, the captain’s airspeed indicator appeared to begin functioning. The autopilot and auto throttle systems were engaged, and the airplane appeared to enter a normal climb. Within seconds of engaging the autopilot, both the captain and first officer recognized that a problem was occurring. The first officer commented that his airspeed was indicating 200 knots, and decreasing, while the captain’s indicator showed an increasing airspeed. At this point, altitude was 5344 feet, and the captain’s airspeed indicator read 327 knots. Pitch attitude was 15.1 degrees. The captain commented that both airspeed indicators were in error and asked the first officer, “what do we do,” before asking for a check of the circuit breaker panel. The reason for checking the circuit breakers was never established by the investigation.
One minute after autopilot engagement, at an altitude of 6688 feet, and a captain’s indicated airspeed of 352 knots, with a pitch attitude of 15.1 degrees, the over speed warning sounded, and the captain commented, “this is not important,” while ordering that the over speed warning system circuit breaker be pulled in order to silence the warning. Altitude had reached 7040 feet, and the captain’s indicated airspeed had slowed slightly to 349 knots while pitch had reduced to 14.8 degrees.
Twenty four seconds after the occurrence of the over speed warning, the stick shaker (stall warning) began. The autothrottle and selected autopilot mode (VNAV) disengaged, but the autopilot itself remained connected. At an altitude of 7132 feet, and a captain’s indicated airspeed of 323 knots, thrust on both engines was momentarily reduced, then after five seconds was restored to near takeoff thrust. The airplane pitched up to 21 degrees, and the autopilot disconnected. Pitch began to oscillate between 5 and 21 degrees.
Thirty-nine seconds after initiation of stall warning, altitude had decreased to 5984 feet, and captain’s airspeed had reduced to 194 knots. Pitch had further reduced to 14.4 degrees, and thrust was again reduced to near idle. After an additional 21 seconds, the captain again commanded a thrust advance in an attempt to arrest the descent. Within a few seconds, thrust on the left engine was reduced to a low level, while thrust on the right engine remained high. Pitch, bank angle, and heading began to oscillate, characteristic of a spin. Asymmetric thrust was maintained, and pitch, altitude, and heading continued to vary until impact with the Atlantic Ocean. All 189 occupants were killed, and the airplane was destroyed on impact.
Investigators determined that, in fact, the captain’s pitot tube was blocked, and the captain’s airspeed indication was incorrect, becoming more in error as the airplane climbed. Due to the pitot blockage, airspeed appeared to be continuously increasing, which resulted in the autopilot continuously increasing pitch in order to slow the airplane to the selected climb speed. Investigators determined that this pitch increase actually resulted in a slowly increasing angle of attack, and eventually caused the airplane to stall.
Just prior to the accident, the accident airplane had been on the ground for maintenance for over 20 days. During this time maintenance personnel conducted an engine inspection which required an engine ground test run before the next flight. Maintenance records reviewed by the investigators did not indicate any airplane discrepancies. However, investigators believed the airspeed system pitot covers had not been installed before, or after, the engine ground test. These covers protect the pitot tubes against the presence of any foreign bodies, or materials entering the pitots. The pitot tubes measure airplane total pressure, and provide information to the air data computer in calculating and indicating airspeed.
Analysis of the cockpit voice recorder (CVR) and the flight data recorder (FDR) indicated that the captain did not have any indication of airspeed until after takeoff. Once airborne, airspeed indications began, but the data presented were inaccurate. Investigators believed the captain did not understand the indications were incorrect. Recorded errors were consistent with indications that would be seen with a blocked captain’s pitot tube.
Investigators felt the probable cause of the obstruction was likely to have been mud and/or debris from a small insect that built a nest inside the pitot tube during the 20 days it was on the ground in Puerto Plata. This was not an uncommon occurrence for this area, which is why it was critical to cover the pitot probes whenever the plane was not in service.
The investigation learned that the aircraft was returned to service without perfoming all of the manufacturer’s prescribed return-to-flight maintenance tasks. After being on the ground for an extended period (20 days), the manufacturer’s maintenance manual recommended a functional check of the pitot-static system. If this check had been completed, the blocked pitot tube would most likely have been discovered and corrected prior to the next flight.
Mud Dauber Wasp
The “Black and Yellow Mud Dauber” wasp is suspected to have been the culprit in the blockage of the captain’s pitot tube. This wasp is notorious for quickly filling open holes, such as pitot tubes, with nests. The wasp builds a simple, one-cell nest made of mud, that is attached to crevices, cracks, corners or inside holes, and then lays a single egg. Mud daubers have been implicated in several aircraft accidents and numerous incidents.
The investigation cited several areas of breakdown in crew operations. The pilot in the left seat (the captain) was the pilot flying, and the first officer (FO) in the right seat was the pilot monitoring. There was also a third pilot, sitting in the flight deck “jump seat” (aft of the center console) acting as a relief pilot due to the expected flight duration. The relief pilot was not acting as a crewmember during the accident sequence, but was technically on duty, and could be called upon in case of emergency.
Investigators determined that the crew did not follow standard operating procedures (SOP) during the takeoff roll. There are several key airspeeds that trigger actions during the takeoff sequence:
- The first checkpoint occurs at 80 knots. The pilot monitoring (in this case, the first officer) calls out “80 knots.” The pilot flying is supposed to verify that airspeed indicators agree, and usually responds by saying “check.” Standard procedure is to abort the takeoff for any warning lights or anything unusual up to that point.
- The second checkpoint is “V1,” or “Engine failure speed.” The pilot monitoring calls out “Vee-One.” This is the highest speed at which an engine failure can occur and the takeoff can be safely rejected and the airplane stopped on the runway. Prior to reaching V1, the crew should reject the takeoff for any safety of flight issue such as engine failure, fire or any control problems. Once past V1, takeoff should be continued, even in the event of an engine failure or other mechanical problems. The only reason to stop after V1 would be an indication that the airplane would be unable to fly.
- The final checkpoint during takeoff ground roll is “VR,” or “Rotation Speed.” At VR, the pilot monitoring calls out “Vee-R” at which point the pilot flying pulls back on the control column and rotates to the takeoff attitude. Following liftoff, the pilot monitoring sees a positive rate of climb on the altimeter, and calls out, “positive rate.” The pilot flying then calls, “gear up,” and the pilot monitoring raises the landing gear.
During this accident after the “80 knots” call by the FO, the captain first responded with, “Check,” but within 2 seconds stated, “My airspeed indicator’s not working.” At this point, per standard procedures, the takeoff should have been rejected. Rather than call for a rejected takeoff however, the captain then conferred with the FO for about 7 seconds to confirm that the FO’s airspeed was working and remarked, “you tell me,” indicating that the FO should call out the remaining takeoff speeds (V1 and VR) using his own indicator.
During climb, static pressure decreases as altitude increases. In this accident, this created an increasing difference in pressure between the sensed static pressure and the constant (sea level) pressure in the plugged pitot tube. This difference resulted in an increasing airspeed indication on the captain’s display. Shortly after liftoff, because the airplane had climbed, the captain’s airspeed indicator appeared to begin functioning, but was in fact only indicating the increasing pressure difference between the “trapped” pitot pressure and the correctly functioning static pressure source. An airspeed indication was confirmed by the captain, as he remarked, “It began to operate,” about 22 seconds after liftoff. Investigators believed that whatever airspeed indication problem had manifested itself on the ground was perceived by the crew to have been resolved.
The captain then engaged the autopilot VNAV mode (which used the captain’s airspeed source) to maintain the selected climb speed. In this mode, with the throttles at a fixed climb thrust, the autopilot pitches the aircraft up or down to maintain the selected climb speed. During normal operation, to accelerate the airplane the autopilot commands the airplane to pitch down. If a deceleration is necessary, the autopilot commands the airplane to pitch up. In this case, with the speed appearing to continuously increase as the airplane climbed, the autopilot continued to command increasing pitch in order to slow the airplane to the commanded climb speed.
Activation of an aural alarm and illumination of a master caution light alerted the crew to EICAS (Engine Indication and Crew Alerting System) messages. In the center of the instrument panel are two vertically placed electronic displays that show EICAS information. Caution and advisory messages appear on the left hand side of the upper display. Both a “RUDDER RATIO” and “MACH AIRSPEED TRIM” message appeared, and were the first indication that an airplane anomaly existed. Both crew members remarked, “..something’s crazy,” and attempted to diagnose the problem. The FO stated, “…right now mine is only 200 and decreasing sir” (referring to his airspeed indicator). This statement is an accurate reflection that the autopilot is pitching the nose up in response to the increasing indicated airspeed on the Captain’s display. This was the point at which the captain stated, “Both of them are wrong, what can we do?”
The captain then requested that the FO to check the circuit breakers, though it was not clear to the investigation which circuit breakers he intended to be checked. Both the captain and FO agreed that the alternate (standby) system was correct, however, it appeared to the investigation that they did not revert to use of the standby airspeed indicator, or compare all three indicators. Investigators believed that only the captain’s display was incorrect, and if control of the airplane had been transferred to the FO, or if the captain had begun using the standby indicator as his primary airspeed reference, the accident might have been averted.
At this point, faced with conflicting warnings – an overspeed warning generated from the captain’s airspeed system, and a stall warning based on an excessive angle of attack, investigators believed the crew became confused. Both crew members were apparently aware that an airspeed indication problem of some kind existed, but they were unable to correctly diagnose the problem, and seemed to investigators to be unaware that the stall warning system functions completely independently from the airspeed system. Lack of a proper response to the stall warning was cited by the investigation as the probable cause of the accident.
At the time of the accident, there were no identifiable EICAS messages or applicable procedures related to unreliable airspeed. Eight months prior to the accident (June of 1995), Birgenair had added a paragraph to their Flight Crew Training Manual regarding “Flight with Unreliable Airspeed,” however, the investigation could not determine if the crew had been briefed or trained on this change. Even if they had, it appears they did not follow the advice. Following the accident, Birgenair added an “Airspeed Unreliable” procedure to their pilot checklist.
The German Federal Bureau of Aircraft Accidents Investigation (German: Bundesstelle für Flugunfalluntersuchung, BFU) produced an animation of the accident sequence. Versions in both German (the original version), and English (a translation) are available at the following links:
Prevailing Cultural / Organizational Factors
The accident board cited Birgenair’s lack of a formal Crew Resource Management (CRM) training program as an additional factor in this accident. It is not known if this was a requirement of the Turkish regulators at the time of the accident, but it was a formal requirement of the FAA.
Further, manufacturer’s maintenance procedures specified that for extended out of service periods (such as extended maintenance), pitot covers should be installed. Additionally, the same maintenance procedures specified that following extended maintenance periods, the pitot-static system should be checked for proper operation prior to airplane return to service. It was established by the investigation that the pitot covers were not installed, and that the pitot-static system was not functionally checked, per the manufacturer’s recommended maintenance procedures, before the airplane was returned to service.
Accident Board Findings
The Dominican Republic Air Accidents Investigation Board (JIAA) concluded that the probable cause of the accident was the failure on the part of the flight crew to recognize the activation of the stick-shaker as an imminent warning of an entrance to aerodymic stall and their failure to execute proper procedures for recovery of the control loss. Before activation of the stick-shaker, confusion of the flight crew occurred due to the erroneous indication of an increase in airspeed and a subsequent overspeed warning.
The following were also cited as contributors:
- Flight crew training, actions taken in the cockpit, use of proper procedures and basic aeronautical abilities.
- Lack of knowledge of the aircraft on the part of the flight crew: including aircraft systems, airspeed indications, automatic pilot, aircraft procedures, selection of alternate sources of data and flight with an untrustworthy airspeed indicator.
- Maintenance practices – not installing the pitot system covers while the aircraft was on the ground, the failure ot perform tests for the return to service of the pitot/static system after a lengthy time on the ground.
The following were cited as additional factors:
- It is possible the flight crew were not physically or mentally rested and prepared to fly the trip due to the unexpected call of the crew during scheduled free time.
- The age of the Captain (62 years) did not allow him to act as pilot in command in certain countries.
- Birgenair’s training did not include Crew Resource Management (CRM) and there was a combination of training from outside sources that failed to provide continuity or an integrated approach to attaining the maximum efficiency of the flight crew.
- The Operations Manual of the 757/767 did not contain detailed information to provide the flight crew with a list of appropriate verifications, to signal a discrepancy in the indication of airspeed, simultaneous activation of rudder/mach trim and other EICAS warnings and a flight with an airspeed indicator that may not be trustworthy.
- The EICAS system of the Boeing 757/767 aircraft did not include an alert of “caution or warning” when a signal of erroneous airspeed is detected.
Click here to see the full accident report.
Investigators determined that the probable cause of the accident was the flight crew’s improper response to the stall warning, and confusion on the part of the flight crew resulting from the airspeed indication anomaly. The investigation discussed the lack of flight crew response while the airplane was still on the ground during takeoff roll, when the captain recognized that his airspeed indicator was not functioning, but did not list it as a contributing factor.
Three memorials were erected to commemorate the accident. All three are similar obelisks carved from the same piece of granite. The monument on the left was placed in the main cemetery in Frankfurt, Germany, the city where the flight was to terminate. The middle photo is the monument placed in San Felipe, on the seafront of Puerto Plata, to mark the crash site. The photo on the right is the monument placed in the town church of Schonefeld, Germany, near the Berlin airport where the flight was to have made an intermediate stop.