Causalfactors

BY MARK LACAGNINA

ust as the captain relaxed rudder pressure seriously injured; 38 passengers, two flight atten-

while tracking the centerline for dants and the first officer sustained minor injuries; NTSB said that nothing in the captain’s him for experience or training had prepared at Denver. the winds he encountered takeoff on a tumultuously windy day, the 67 passengers and one escaped Boeing 737-500 was struck by a strong gust. injury. The airplane was substantially damaged JLike a weathervane, the airplane turned into the during the excursion and postcrash fire. crosswind and then ran off the side of the runway. In its final report on the accident, which Five of the 110 passengers and the captain were occurred at Denver International the

© David Zalubouswki/Associated Press Blown Away Blown

www.flightsafety.org | AeroSafetyWorld | October 2010 | 27 Causalfactors

afternoon of Dec. 20, 2008, the U.S. National rating and had about 8,000 flight hours, includ- Transportation Safety Board (NTSB) said that ing 1,500 hours as a 737 first officer. the probable cause was “the captain’s cessation When the first officer, the pilot monitoring, of rudder input, which was needed to maintain requested clearance for pushback at 1804, he directional control of the airplane, about four told the airport ramp controller that they had seconds before the excursion, when the airplane received automatic terminal information service encountered a strong and gusty crosswind that (ATIS) Information Charlie, which reported exceeded the captain’s training and experience.” surface winds from 280 degrees at 11 kt, 10 mi The report said that the following factors (16 km) visibility, a few clouds at 4,000 ft and contributed to the accident: “an air traffic con- a surface temperature of minus 6 degrees C trol system that did not require or facilitate the (21 degrees F). A notice to airmen advised that patches of snow, ice and/or slush were on the ramps and taxiways, but the runways were bare and dry, the report said. Denver has six runways, all at least 12,000 ft (3,658 m) long. Runways 25, 34L and 34R, all on the west side of the airport, were being used for departures. The airport ground controller instructed the crew to taxi to Runway 34R. At 1814, the airport traffic controller told the crew to taxi into position and hold for departure on Runway 34R. After the crew completed the “Before Takeoff” checklist, the captain remarked, “Looks like you got some wind out here.” The first officer replied, “Yeah.” The captain said, “Oh, © Brian Gore/.net yeah. Look at those clouds moving.” The 737 veered off dissemination of key, available wind information Shortly thereafter, the controller advised Runway 34R, the to the air traffic controllers and pilots; and inad- that the winds were from 270 degrees at 27 kt inner runway of equate crosswind training in the industry and cleared the crew for takeoff. “Although this the parallel pair on due to deficient simulator wind gust modeling.” wind was significantly stronger than the wind the northwest side reported by ATIS (280 degrees at 11 knots) 20 of the airport. ‘Some Wind Out Here’ minutes earlier, the wind was still within Con- The airplane was being operated as Continental tinental’s crosswind guidelines of 33 knots,” the Flight 1404 to Houston. The pilots ar- report said. rived at the airport about 1700 local time — one hour before the scheduled departure time. Elusive Centerline The captain, 50, was hired by Continental As the captain moved the thrust levers forward in 1997 and served as a first officer in Douglas at 1817:38, he said, “All right. Left crosswind, DC-9s, 737s, 757s and 767s before transition- twenty-seven knots.” He later told investigators ing as a 737 captain about 14 months before that when the airplane began to accelerate, he the accident. He had about 13,100 flight hours, shifted his attention from the engine gauges to including 6,300 hours in 737s. Before joining outside visual references and concentrated on Continental, he was a naval aviator. tracking the runway centerline. The first officer, 34, was a flight instructor “The first officer stated that after the power and regional airline pilot before being hired by was set, he shifted his attention to monitoring Continental in March 2007. He held a 737 type the airspeed so that he could make the standard

28 | flight safety foundation | AeroSafetyWorld | October 2010 Causalfactors

airspeed callouts, the first of which was 100 kt,” The captain later the report said. told investigators that Wind and Rudder Deflection Recorded flight data indicated that as the he had “felt the rear 1818:15 airplane accelerated, the captain applied right end of the airplane rudder pedal inputs of increasing amplitude slip out hard to the 20 while holding the control wheel and control col- right and the wheels 19 umn in their neutral positions. The 737 was ac- lose traction.” He 41 celerating through about 55 kt at 1818:07, when perceived that the 44 it began to move left, away from the runway airplane had encoun- 43 centerline. The captain responded two seconds tered a slippery patch 36 later by moving the right rudder pedal almost of runway, a strong 35 all the way forward, displacing the rudder nearly gust of wind, or both. to its maximum deflection of 26 degrees (Figure The first officer 32 30 1). “Almost simultaneous with the onset of this recalled that there was 1818:10 large rudder pedal input, the FDR [flight data “a slight deviation left 37 recorder] began to record a left control wheel of centerline [at about 40 input,” the report said. 90 kt], but we seemed 35 The airplane began to head back toward the to be correcting back 35 runway centerline. However, as it accelerated to the right.” He said 36 through about 85 kt at 1818:10, “the airplane’s that the airplane then 36 nose reversed direction and began moving back “abruptly swung ap- 36 to the left at a rate of about one degree per sec- proximately 30 degrees 39 ond,” the report said. “The leftward movement left with the tail to the 39 of the nose continued for about two seconds and right, and we were 33 was accompanied throughout its duration by heading for the left 30 1818:05 100 20 30 another substantial right rudder pedal input.” side of the runway.” Relative wind Rudder Time This right rudder input slowed the left- direction and de ection speed (kt) (degrees) ­turning motion momentarily, but the nose again ‘Very Painful Bumps’ began moving rapidly to the left at 1818:13, The CVR recorded Source: U.S. National Transportation Safety Board, Susan Reed about the same time that the pilot relaxed pres- an expletive voiced by sure on the right rudder pedal, returning the the first officer just Figure 1 pedal to its neutral position. before the 737 ran Shortly thereafter, the voice recorder off the left side of the runway at 1818:17. The (CVR) recorded an exclamation by the captain, captain called “reject” twice, announcing that and “the FDR recorded the beginning of a tran- he was rejecting the takeoff. “FDR data showed sition from left control wheel input (consistent engine power reductions, as well as activation with crosswind takeoff technique for a left cross- of the brakes,” the report said. “Thrust reverser wind) to right control wheel input (crossing the deployment began about three seconds after the control wheel’s neutral point at 1818:14),” the airplane left the runway.” report said. “The FDR did not record any more Groundspeed was about 110 kt when the substantial right rudder pedal inputs as the airplane veered off the runway about 2,600 ft airplane continued to veer to the left.” (792 m) from the approach threshold, on a mag- As the airplane neared the edge of the run- netic heading of about 330 degrees. The pilots way, the captain tried unsuccessfully to use the began reducing power, which also activated nosewheel-steering tiller to regain directional the autobrake, about three seconds later. The control. The report noted that the tiller typically airplane crossed a taxiway and an airport service is used only during low-speed taxiing. road, and came to a stop on a heading of about www.flightsafety.org | AeroSafetyWorld | October 2010 | 29 Causalfactors

airplane stopped and made no immediate at- tempts to get up or leave the cockpit.” Unable to communicate with the pilots, the flight attendants initiated an evacuation when they saw a fuel-fed fire erupt on the right side of the airplane. The flight attendants, assisted by two deadheading pilots, were able to evacuate all the passengers through the three left exits before the fire entered the cabin. Although injured, the captain and first officer exited without assistance. ARFF personnel arrived about five minutes after the evacuation was completed and extinguished the fire. The most serious injuries during the excursion involved back and/or spi- nal column damage.

Variable Winds Denver International Airport is at an elevation of 5,431 ft in the foothills just east of the Front Range of the Rocky Mountains. Weather condi- tions on the day of the accident were influenced by a stationary front extending through Colo- rado. Analyses performed by the U.S. National Center for Atmospheric Research (NCAR) indi- cated that the airport was affected by significant mountain wave activity. “The undulating motion of these waves as they moved eastward across [the airport] result- ed in strong, very localized, intermittent gusts,” The airplane had a 315 degrees just north of an rescue and the report said. The NCAR analyses indicated rough ride on uneven fire fighting (ARFF) station located between that a 45-kt gust was moving across the runway terrain after leaving Runway 34R and Runway 34L. The FDR and when the captain made the remark about cloud the runway at 110 kt. CVR recordings ended at 1818:27. movement while awaiting takeoff clearance. “Postaccident interviews with passengers The flight crew had received wind information and crewmembers, as well as evidence from the from different sources and locations on the huge crash site, indicated that as the airplane crossed airport. The ATIS wind information — 280 de- the uneven terrain before coming to a stop, it grees at 11 kt — was derived from the automated became airborne, resulting in a jarring impact surface observing system (ASOS) sensor located when it regained contact with the ground,” the near the center of the airport, about 2.5 mi (4.0 report said. km) southeast of the approach end of Runway 34R. The captain told investigators that he was The wind information — 270 degrees at 27 just “along for the ride” after the airplane veered kt — provided about 20 minutes later by the off the runway. “Both pilots stated that there airport traffic controller who cleared the crew were a couple of ‘very painful’ bumps before the for takeoff was based on readings from a low airplane came to a stop,” the report said. “They level wind shear alert system (LLWAS) sensor indicated that they were somewhat dazed or located near the departure end of Runway 34R.

‘knocked out’ for one or two minutes after the Those readings, as well as readings from sensors Board Safety Transportation National U.S.

30 | flight safety foundation | AeroSafetyWorld | October 2010 Causalfactors

associated with the ends of the other The report said that the “unusually be unmanageable for a line pilot,” the runways in use, were displayed on a large” rudder inputs that the captain report said. monitor at the controller’s station. made twice during the takeoff roll likely The participants also said that the The flight crew did not, however, increased the difficulty he encountered flight simulator did not accurately receive wind information from a closer in maintaining directional control. “To reflect lateral forces or provide a good source — an LLWAS sensor about 3,300 avoid overshooting the baseline head- “seat of the pants” feel for wind gusts. ft (1,006 m) from the approach end of ing after each large right rudder pedal Investigators found that Continen- Runway 34R. Readings from that sen- input, the captain had to compensate tal’s annual simulator recurrent training sor also were displayed on the control- by relaxing the right rudder pedal more included takeoffs and landings with a ler’s monitor and were designated as than he would have had to for a smaller 35-kt crosswind. “However, the com- “AW,” for “airport wind.” When the crew rudder pedal advancement,” the report pany’s 737-500 flight simulators were was cleared for takeoff, the monitor said. “Furthermore, because of slight not programmed to simulate gust effects showed the airport wind as from 280 delays in the effect each rudder pedal below about 50 feet above the ground degrees at 35 kt with gusts to 40 kt. adjustment had on the airplane’s rate and, therefore, were not capable of rep- The controller did not provide and of heading change, the captain had to licating the complex disturbances that was not required to provide the “airport anticipate the effect of each adjust- pilots would experience during takeoffs wind” information to the crew. “It was ment ahead of time. This task was very and landings in gusty surface winds,” common practice for [airport] control- difficult for the captain because of the report said. “Further, takeoff data lers to issue departure runway end winds the highly variable and unpredictable obtained from Continental indicated to departing aircraft,” the report said. nature of the crosswind gusts.” that the company’s pilot rarely, if ever, Based on the wind information The captain’s full-right control encountered crosswind components that was provided, the crew’s decision wheel movement and use of the nose- greater than 30 knots during actual flight to depart on Runway 34R rather than wheel steering tiller three seconds be- operations.” requesting Runway 25 for takeoff was fore the excursion “likely resulted from Based on the findings of the inves- “reasonable,” the report said. “Further, acute stress stemming from a sudden, tigation, NTSB issued several recom- other airplanes departed on Runways unexpected threat, perceived lack of mendations to the U.S. Federal 34L and 34R before the accident pilots’ control and extreme time pressure,” the Administration, including more research departure; the pilots of those departing report said, noting that these actions on mountain waves and downslope airplanes did not report any crosswind- were ineffective and delayed the initia- winds; a requirement that controllers related issues or difficulties.” tion of a rejected takeoff. provide pilots with information on Investigators estimated that the 737 the maximum wind components they encountered direct crosswind com- Insufficient Simulation might encounter on takeoff or landing; a ponents ranging from 29 kt to 45 kt Postaccident flight simulator tests with requirement that operators of air carrier, during the takeoff roll. The peak gust pilots holding 737 type ratings showed air taxi and fractional ownership aircraft of 45 kt occurred about the same time that when they removed their feet incorporate “realistic, gusty crosswind that the captain relaxed pressure on from the rudder pedals while encoun- profiles” in their simulator training; the right rudder pedal. The report said tering a 35-kt crosswind at an airspeed and a requirement that manufacturers that the captain likely would have been of 90 kt, the “airplane” veered off the of transport category airplanes develop able to maintain directional control if runway within five seconds. They were type-specific crosswind limitations that he had maintained or rapidly reap- able to continue or reject the takeoff account for wind gusts.  plied right rudder input. “Performance successfully if they resumed corrective calculations indicated that the airplane’s rudder inputs within two seconds after This article is based on NTSB Accident Report rudder was capable of producing releasing pedal pressure; but three NTSB/AAR-10/04, “Runway Side Excursion During Attempted Takeoff in Strong and Gusty enough aerodynamic force to offset the seconds was too late. “Participants Crosswind Conditions; Continental Airlines weathervaning tendency created by the agreed that a three-second delay in Flight 1404; Boeing 737-500, N18611; Denver, winds the airplane encountered during reapplication of corrective rudder in- Colorado; December 20, 2008.” The full report is the accident takeoff roll.” puts resulted in a situation that would available at . www.flightsafety.org | AeroSafetyWorld | October 2010 | 31