Presque Isle, ME, USA
N14171
Embraer EMB145
The flight crew of an Embraer EMB145XR airplane was attempting to land on runway 1 at Presque Isle International Airport (PQI), Presque Isle, Maine. Snow was falling at the time. A Notice to Air Mission issued 2 hours 19 minutes before the accident indicated that runway 1 was covered with 1/4 inch of dry snow, and the PQI maintenance foreman stated that the airport had been conducting snow removal operations to maintain that condition on the runway surface. Accident Sequence 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 seein’,” 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. Flight Crew Performance The first officer was relatively new to the EMB145; he received his type rating for the airplane about 7.5 months before the accident. Even though the first officer did not see the runway at the decision altitude, he might have continued the second approach to a landing because he trusted that the captain had the runway in sight. Also, the captain had instructed the first officer to “stay in” multiple times as the airplane descended through 100 ft agl. The captain had apparently intended for the first officer to focus on the flight director and not look outside for the approach lights or the runway. Company policy stated that the pilot flying should monitor the instruments until the callout “runway in sight” and then transition to outside references no later than 100 ft above the touchdown zone elevation. Company policy also stated that a pilot should call for a go around if either the runway environment was not in sight by the decision altitude or the successful completion of the approach was in doubt. Thus, when the first officer looked outside after the captain’s “runway in sight” callout and did not see the runway, one or both flight crewmembers should have called for a go around. Further, the captain reported that, during the second ILS approach to runway 1, she saw the tower again but explained that the airplane had leveled off to clear the tower before continuing to descend. However, the appearance of the tower should have prompted the captain to call for a go-around, just as she did during the first approach. The captain should have recognized that an airplane that was on the proper vertical and horizontal paths of an ILS approach would not have flown that close to a 30-ft tower while descending below the decision altitude. The National Transportation Safety Board considered why the flight crew might have continued the second approach rather than call for and perform another go-around, especially since the CVR recorded the captain telling the first officer that, if they did not see the runway during the second approach, they would go to their alternate airport. The CVR also recorded the first officer stating, “if there’s nothing there then we’ll go.” Confirmation bias is a type of cognitive bias that involves a tendency to seek information to support one’s belief instead of information that is contrary to that belief. In this case, the localizer and glideslope information indicated that the airplane was aligned with the runway centerline, and the captain stated that she had the runway in sight. However, the first officer did not have the runway in sight, and both flight crewmembers expressed confusion about what they were seeing outside the airplane when it was below 100 ft agl. Also, during his previous flight to PQI, the first officer noticed an “incongruency” between the pink needle (providing guidance from the airplane’s flight management system) and the green needle (providing guidance from the ILS localizer signal). Although the first officer shared this information with the captain during the predeparture briefing, neither flight crewmember considered that a navigational air error could be occurring, even though the captain saw a tower (first and second approaches) and the first officer saw a structure with an antenna (first approach). Thus, the crewmembers discounted their confusion about the runway environment and continued the approach likely because of confirmation bias. The captain was ultimately responsible for the flight. However, she demonstrated poor judgment and decision-making when she instructed the first officer to stay on the flight instruments as the airplane descended below the decision altitude. By the time that the first officer transitioned outside the airplane, not enough time remained for him to determine the airplane’s position in relation to the runway. Although cognitive biases, including confirmation bias, can affect judgment, decision making, and behaviors, a review of the captain’s training records revealed deficiencies regarding her piloting abilities. For example, the CommutAir vice president of flight operations stated that, while the captain was a De Havilland Canada DHC-8 first officer, she received a disciplinary letter and agreed not to pursue captain upgrade training and be monitored for 9 months. Also, in September 2017, the captain received a notice of disapproval from the Federal Aviation Administration (FAA) for her EMB145 type rating. In addition, twice in September 2017, CommutAir placed the captain under “increased scrutiny” due to training failures, including a failed proficiency check. Even though the captain received her EMB145 type rating in early October 2017 and upgraded to captain afterward, her repeated training problems indicated an inadequate foundation for being a captain, which CommutAir did not effectively address. In addition, although a review of the flight crew’s recent activities determined nothing noteworthy about the captain’s activities and her sleep opportunity (7.5 hours) on the night before the accident, evidence indicated that the first officer was likely fatigued on the day of the accident. Although the first officer stated that he felt rested that day, he had been home with the flu for several days before the accident. Further, the first officer was prescribed a continuous positive airway pressure (CPAP) machine, but he did not consistently use his CPAP between February 26 and March 3, 2019. The daily-use graphic provided in the CPAP download indicated that the first officer used the device for less than 1 hour on February 26, did not use the device between February 27 and March 2, and used the device for about 1 hour 30 minutes between 1200 on March 3 and 1200 on March 4. Due to his illness and lack of CPAP use, the first officer was likely not obtaining adequate sleep during that period. The first officer commuted to Newark, New Jersey, on the night before the accident. He arrived at a local hotel about 0000 on March 4 due to flight and shuttle delays. The first officer went to sleep about 0100 and awoke about 0600, resulting in a sleep opportunity of 5 hours. The first officer normally slept 7 to 8 hours; thus, he had a sleep debt of about 2 to 3 hours. The quality of his sleep would also have been compromised because he was still coughing and did not use his CPAP. The first officer’s fatigue likely exacerbated the cognitive bias that he experienced during the flight. Localizer Misalignment Both flight crewmembers reported that the localizer and glideslope needles were centered during the first and second approaches, indicating that the airplane was aligned with the runway centerline. However, postaccident flight testing of the ILS localizer and glideslope revealed that the localizer was out of tolerance by about 200 ft to the right. After the accident, the airport conducted snow removal operations in the area around and in front of the localizer array; the snow depths (on the day before removal operations began) ranged from about 2 to 5 ft. After the snow was removed, a flight check determined that the localizer signal was in alignment. About 36 hours before the accident (the night of March 2, 2019), a CommutAir flight crew noted the localizer misalignment while on approach to PQI. After landing, the crew reported the misalignment to a controller at the Boston Air Route Traffic Control Center. The controller reported the localizer misalignment to FAA technical operations personnel, including the center’s operations manager-in-charge, who then informed the center’s National Airspace System operations manager. FAA procedures stated that, for reports of a navigational aid malfunction, air traffic control personnel should request a report from a second aircraft. Because a second pilot report had not yet been received to confirm the localizer misalignment, the National Airspace System operations manager did not act on the initial report about the misalignment. The accident flight was the first instrument flight rules flight to arrive at PQI after the initial report about the localizer misalignment. FAA procedures for air traffic control personnel also stated that, “in the absence of a second aircraft report, activate the standby equipment or request the monitor facility to activate.” However, PQI did not have an air traffic control tower, and air route traffic control centers, including Boston Center, do not have the capability to activate standby equipment. Airport personnel at PQI stated that they had no means to determine the alignment of the localizer signal and had to depend on pilot reports. However, the March 2, 2019, pilot report about the localizer misalignment was not provided to airport personnel. In addition, although FAA winter operations guidance contained specific criteria for the allowable snow depth around a glideslope antenna, the guidance did not specify similar information for the area around a localizer antenna array. After the accident, the FAA revised its winter operations guidance to state that snow around a localizer array could affect its radiated signal and that a snow accumulation level of 2 ft was the point at which an FAA ILS specialist would need to begin observing the condition of the localizer signal. Company Reports of Localizer Misalignment At least six pilots who flew into PQI during the 5 days before the accident (including the accident first officer) encountered issues with the ILS localizer. However, none of those pilots submitted a company aviation safety action program (ASAP) report before the accident. (Four of the pilots submitted an ASAP report after the accident.) During a postaccident interview, the CommutAir managing director of safety stated that he did not know why the reports were not filed before the accident. According to the CommutAir director of operations, a company flight data analyst reviewed ASAP reports “right away,” and the analyst provided time-critical information to the appropriate company managers and directors when necessary. Because the reports of the PQI localizer misalignment were submitted after the accident, CommutAir missed an opportunity to make this information available to company flight crews flying to PQI and employ strategies to mitigate any potential threat that the misalignment posed. For example, CommutAir could have alerted those pilots to maintain a heightened awareness of the localizer alignment, restricted the use of the runway 1 ILS approach to higher weather minimums, or prohibited the use of the approach. In addition, if the Boston Air Route Traffic Control Center had received an earlier report of a localizer misalignment, center personnel could have had the opportunity to confirm that report with a second report or take another action to designate the approach as unusable until the localizer signal could be assessed for proper alignment.
HISTORY OF FLIGHTOn March 4, 2019, about 1129 eastern standard time, CommutAir flight 4933, an Embraer EMB145XR, N14171, was attempting to land on runway 1 at Presque Isle International Airport (PQI), Presque Isle, Maine, and impacted terrain to the right of the runway. The first officer and 2 of the 28 passengers sustained minor injuries, and the captain, the flight attendant, and 26 passengers were not injured. The airplane was substantially damaged. The scheduled passenger flight was operating under the provisions of Title 14 Code of Federal Regulations Part 121. Instrument meteorological conditions prevailed at the time of the accident. The first officer’s most recent flight to PQI before the accident was on February 27, 2019. As part of the predeparture briefing for the accident flight, the first officer mentioned that, during the previous flight, the instrument landing system (ILS) localizer for runway 1 was offset when the airplane was aligned with the runway during a visual approach. (A localizer uses a radio beam to provide pilots of landing aircraft with lateral navigation information to align with the runway and is one of the two main components of an ILS; the glideslope is the other main component.) During a postaccident interview, the first officer, who was the pilot monitoring for that flight, stated that both he and the captain of that flight noticed an “incongruency” between the pink needle (which provides guidance from the airplane’s flight management system) and the green needle (which provides guidance from the ILS localizer signal). The accident flight departed from Newark Liberty International Airport (EWR), Newark, New Jersey, about 1004. The captain was the pilot monitoring, and the first officer was the pilot flying. The en route portion of the flight was uneventful. According to CommutAir, once the airplane was in range of PQI, the flight crew received an updated airport weather report at the time—a special weather observation at 1031. The observation indicated that the wind was from 090° at 5 knots, visibility was 1/2 mile in moderate snow and freezing fog, and the cloud ceiling was broken at 1,100 ft and overcast at 1,800 ft. According to the cockpit voice recorder (CVR), at 1101:42, a controller from the Boston Air Route Traffic Control Center (ARTCC) cleared the flight for an ILS approach to runway 1, and the captain acknowledged the instruction. At 1105:35, the controller terminated radar services and instructed the flight crew to change to the PQI common traffic advisory frequency (CTAF); PQI did not have an air traffic control tower. The captain acknowledged this instruction and then announced, over the CTAF, that the flight was 2 miles from FEROG (an approach waypoint) and inbound for the ILS approach to runway 1. A PQI maintenance staff member (later identified as the maintenance foreman) contacted the flight crew at 1105:59, and the captain responded that the flight was 4 minutes away from the airport. At 1106:14, the first officer stated to the captain, “he said…he was out of the way now, so he’s clear,” indicating that the runway was clear of snow removal vehicles. Flight data recorder (FDR) data indicated that, at 1106:58, the airplane began its first approach to PQI. At 1107:50, the captain announced over the CTAF that the airplane was nearing the final approach fix for the approach. The ILS approach to runway 1 at PQI had a decision altitude of 678 ft, which was 200 ft above ground level (agl), and a visibility requirement of 1/2 mile. The captain made the 1,000-ft callout (indicating that the airplane was 1,000 ft above the decision altitude) at 1108:38 and stated that the approach was stable and that she had “ground contact.” At 1109:03, the captain asked the first officer if he wanted the airport lights to be turned on. (The runway 1 edge lights, runway end identifier lights, and the approach lighting system were pilot controlled on a published frequency.) The first officer’s reply of “yeah. Turn them on” was preceded and followed by a sound similar to five microphone clicks. The captain made the 500-, 400-, and 300-ft callouts between 1109:20 and 1109:33. The first officer then stated, “autopilot’s coming off,” which the captain acknowledged. At 1109:40, the captain made the 200-ft callout, which was followed by the “approaching minimums” and “minimums” aural annunciations from the enhanced ground proximity warning system. The CommutAir EMB 145 Aircraft Operations Manual stated that pilots should call for a go-around if the runway environment was not in sight by the decision altitude or if the successful completion of the approach was in doubt. At 1109:54, the captain stated, “runway in sight. See it?” to which the first officer responded, “yeah” and “well I got somethin’ [that] looks like a runway up there.” The CVR recorded the aural annunciation “one hundred [ft]” at 1109:59. About 3.5 seconds later, the captain stated, “watch your speed,” which was followed by sounds similar to the stickshaker. At 1110:09, the captain stated, “go missed” twice; less than 1 second later, the first officer stated, “yeah we’re goin’ missed.” During a postaccident interview, the captain stated that she saw the approach lights but that she also saw a tower that looked “very close” to the airplane’s position. The first officer stated that, when he transitioned from looking at the instruments to looking outside, he expected to see the runway but saw what he described as “white on white.” The first officer also stated that he saw a structure with an antenna that was part of the runway environment but not the runway itself, so he executed the go-around. The maintenance foreman stated, during a postaccident interview, that the runway lights were not on after the first approach. FDR data showed that the airplane had descended to a minimum pressure altitude of 703 ft (169 ft agl) before beginning to ascend. At 1110:33 and 1110:56, the captain stated over the CTAF that the flight was “going missed.” At 1111:05, the captain notified the Boston ARTCC that the flight “went missed…[and] we’re gonna give it another try.” The controller acknowledged the information and then instructed the flight crew to climb to and maintain 3,200 ft. At 1113:28, the captain contacted PQI maintenance and stated that the flight “went missed” and that she would call back again “for another try.” Between 1113:50 and 1114:08, the captain and the first officer discussed the previous approach. The captain asked the first officer whether he lost the localizer, and he stated, “I don’t think so…I went outside the airplane, too early, and I didn't have the runway.” The first officer continued, “I thought I had the runway then I was like that is not the runway,” to which the captain responded, “yeah I thought that too.” At 1114:13, the first officer stated that, for the second approach, he would “stay inside on the localizer,” and the captain agreed. At 1115:20, the controller provided vectors for the ILS approach to runway 1, which the captain acknowledged. At 1116:20, the captain contacted PQI maintenance about the second approach to the runway, and the maintenance foreman stated, “we’ll be all clear runway one.” The captain stated, “can you make sure those lights are on for us?” The maintenance foreman replied, “yes we will.” During a postaccident interview, the maintenance foreman reported that he turned on the lights to the high-intensity setting. At 1116:46, the first officer asked the captain, “did you ever see the lights at all last time?” The captain stated that she saw the lights but that “it’s really white down there that’s the problem.” The first officer agreed and stated, “everything is washed out.” The captain also stated, “if we don’t see it we’ll just go to, Vermont.” (The captain was referring to Burlington International Airport, South Burlington, Vermont, which was the alternate airport for the flight.) The first officer replied, “you got it.” Between 1118:36 and 1118:42, the first officer stated, “so this time I’ll stay on the flight director until things start screaming minimums…then I’ll look up…if there’s nothing there then we’ll go, if there is something there we’ll land.” The captain commented, “yup, sounds good.” The first officer also stated that he would specifically look for the lights that surround the runway and that, during the previous approach, “all I saw was the antennas at the end of the runway.” Between 1119:00 and 1120:41, the captain and the first officer began discussing the previous approach and the second approach. The captain repeated, “it’s really white down there,” and instructed the first officer to “stay inside and I’ll let you know when you can look up.” The captain also instructed the first officer on the actions to take if she commanded “go missed” again. The captain contacted PQI maintenance at 1121:46 and stated that the airplane was about 16 miles and 7 minutes away from the airport. The maintenance foreman replied, “we’ll be clear runway one and the lights are on bright.” The maintenance foreman stated he had again activated the lights to high intensity. At 1123:41, the controller told the flight crew that the airplane was about 8 miles south of the locator outer marker for the ILS runway 1 approach and cleared the flight for the approach. The controller also instructed the flight crew to report when the airplane was established on the localizer. The captain acknowledged this information. At 1125:03, the captain told the first officer, “localizer’s comin’ in alive,” and the first officer responded, “localizer is alive so is the glideslope.” The captain then informed the controller that the airplane was established on the localizer. The controller instructed the flight crew to change to the CTAF, and the captain acknowledged this instruction and notified local traffic that the airplane was inbound for runway 1. FDR data showed that, at 1126:51, the airplane started its final descent to PQI. At 1126:51, the captain announced over the CTAF that the airplane was 4 miles from runway 1. The captain made the 1,000-ft callout at 1127:21 and stated that the approach was stable. After the captain made the 500-ft callout at 1127:57, the first officer stated, “five hundred cleared to land. I’m inside you’re outside.” The captain made the 400-ft callout shortly afterward and instructed the first officer to keep the autopilot on until 200 ft agl. The first officer stated, “I will,” which was followed by the captain’s 300-ft callout. At 1128:22, the captain stated, “there’s two hundred, get the autopilot off,” and the first officer stated “off” about 6 seconds later. At 1128:30 and 1128:37, the enhanced ground proximity warning system announced “approaching minimums” and “minimums,” respectively; in between those annunciations, the captain made the 100 ft callout. According to the CVR, about 2 seconds after the “minimums” annunciation, the captain called, “runway in sight twelve o’clock.” During a postaccident interview, the captain stated that the localizer and glideslope needles were centered when she called the runway in sight; also, the CVR did not record any discussion between the flight crewmembers about a localizer or glideslope deviation. Federal Aviation Administration (FAA) automatic dependent surveillance-broadcast data showed that the airplane was aligned to the right of the runway 1 centerline during both approaches (with the last data point for the second approach recorded when the airplane was about 23 ft agl), and FDR data showed that the localizer and glideslope needles were mostly centered with only small deviations consistent with normal piloting. At 1128:42, the first officer stated, “I’m staying on the flight director ‘cause I don’t see it yet,” which was followed by the captain stating “stay in” several times within a 5-second period. At 1128:53, the captain stated, “what the [expletive],” and the first officer stated, “I don’t know what I’m seein’.” FDR data showed that, at 1128:56, the air-to-ground switch parameter changed from air to ground; about 1 second later, the vertical acceleration parameter reached its maximum value of 3.35 Gs. At 1129:14, the airplane’s groundspeed was 0 knots. During a postaccident interview, the first officer stated that, when he transitioned from the instruments to the outside during the second approach, he saw “white on white” again and blowing snow. The first officer also stated that it was difficult to comprehend what he was seeing outside the airplane because everything was covered in snow and that, before he could determine what he was seeing, the airplane touched down. The first officer added that he did not see the structure with the antenna during the second approach because the airplane had flown over it by the time that he looked outside. The captain reported that she saw the tower again but that the airplane had leveled off to clear the structure before continuing to descend. The captain recalled that, after touchdown, the airplane was “bouncing up and down a few times” before coming to a stop. The flight attendant reported that the landing was “rough and violent” with seat cushions and passenger belongings falling into the cabin aisle as the airplane came to a stop. The airplane came to rest in the snow-covered grassy area between runway 1 and a parallel taxiway located about 630 ft to the right of the runway. The airplane’s resting location was about 3,600 ft beyond the runway threshold, about 305 ft to the right of the runway centerline, and about 230 ft from the right edge of runway. Airport personnel estimated that, when the airplane landed, visibility was about 1/2 mile, and about 30 minutes had elapsed since the time that the runway was last plowed. Postaccident Events At 1129:18, the captain instructed the passengers to remain seated; 3 seconds later, the flight attendant provided the same instruction. At 1129:24, the captain instructed the first officer to run the engine shutdown checklist, which he did. The flight attendant reported that, after the airplane came to a stop, she called the flight deck and received no answer, but she could hear the flight crew from her aft-facing jumpseat and decided to wait for the captain to contact her because an immediate evacuation was not necessary. The flight attendant then moved out of her seat, looked outside, and walked through the cabin to check on the passengers and remove items that had fallen into the aisle. When the flight attendant returned to the front of the cabin, the captain opened the flight deck door and told her that an evacuation would be occurring, and the flight attendant relayed that information to the passengers. After the main cabin door was opened, the flight attendant noticed that the snow reached the bottom of the belly of the airplane. Firefighters came aboard the airplane; one attended to the first officer, who was injured, and one walked through the cabin to check on passengers. A snow plow created a walkway so that the crewmembers and passengers would not have to walk through the snow after exiting the airplane. Emergency personnel took the first officer off the airplane, and the flight attendant then directed the passengers to exit the airplane via a ladder that firefighters had brought to the airplane. After the passengers had exited, the flight attendant checked to make sure that no one was left behind, and the captain and the flight attendant then exited the airplane. A bus transported the captain, flight attendant, and passengers to the terminal. PERSONNEL INFORMATIONThe Captain The captain began working for CommutAir in March 2013 as a first officer on the De Havilland Canada DHC 8 airplane. The captain left the company in November 2015 and joined another air carrier. In May 2016, the captain left that air carrier and returned to CommutAir as a first officer on the DHC-8. The
The flight crew’s decision, due to confirmation bias, to continue the descent below the decision altitude when the runway had not been positively identified. Contributing to the accident were (1) the first officer’s fatigue, which exacerbated his confirmation bias, and (2) the failure of CommutAir pilots who had observed the localizer misalignment to report it to the company and air traffic before the accident.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
