Cleveland, OH, USA
N862RW
Embraer ERJ-170
The Safety Board's full report is available at http://www.ntsb.gov/publictn/A_Acc1.htm. The Aircraft Accident Report number is NTSB/AAR-08/01. On February 18, 2007, about 1506 eastern standard time, Delta Connection flight 6448, an Embraer ERJ-170, N862RW, operated by Shuttle America, Inc., was landing on runway 28 at Cleveland Hopkins International Airport, Cleveland, Ohio, during snow conditions when it overran the end of the runway, contacted an instrument landing system (ILS) antenna, and struck an airport perimeter fence. The airplane's nose gear collapsed during the overrun. Of the 2 flight crewmembers, 2 flight attendants, and 71 passengers on board, 3 passengers received minor injuries. The airplane received substantial damage from the impact forces. The flight was operating under the provisions of 14 Code of Federal Regulations Part 121 from Hartsfield-Jackson Atlanta International Airport, Atlanta, Georgia. Instrument meteorological conditions prevailed at the time of the accident.
The Safety Board's full report is available at http://www.ntsb.gov/publictn/A_Acc1.htm. The Aircraft Accident Report number is NTSB/AAR-08/01. On February 18, 2007, about 1506 eastern standard time, Delta Connection flight 6448, an Embraer ERJ-170, N862RW, operated by Shuttle America, Inc., was landing on runway 28 at Cleveland Hopkins International Airport (CLE), Cleveland, Ohio, during snow conditions when it overran the end of the runway, contacted an instrument landing system (ILS) antenna, and struck an airport perimeter fence. The airplane's nose gear collapsed during the overrun. Of the 2 flight crewmembers, 2 flight attendants, and 71 passengers on board, 3 passengers received minor injuries. The airplane received substantial damage from impact forces. The flight was operating under the provisions of 14 Code of Federal Regulations Part 121 from Hartsfield-Jackson Atlanta International Airport (ATL), Atlanta, Georgia. Instrument meteorological conditions prevailed at the time of the accident. According to weather observations, 15 inches of snow was on the ground at CLE at 0700 on February 17, 2007. Light snow fell from 0910 to 2156, with 1 inch of new snow reported during that period. Snow began to fall again from 0541 to 1201 on February 18, with 2 inches of new snow reported during the period, and from 1436 to 1538, with less than 1 inch of additional snow accumulation. On the day of the accident, the captain traveled as a nonrevenue passenger on a flight from Louisville International Airport-Standiford Field (SDF), Louisville, Kentucky, to ATL to report for a scheduled 2-day trip. The captain was scheduled to report to SDF at 0525, and the flight to ATL had a scheduled arrival time of 0733. The first flight leg, from ATL to Sarasota-Bradenton International Airport (SRQ), Sarasota, Florida, was delayed because of weather. The flight departed ATL at 0914 and arrived at SRQ at 1042. The second flight leg departed SRQ at 1108 and arrived at ATL at 1242. The third flight leg, the accident flight, departed on time (with a different first officer) from ATL at 1305 and had an expected arrival time at CLE of 1451. The accident flight was the first one in which the captain and the first officer had flown together. Shuttle America's common practice is for the captain to be the flying pilot for the first flight of any crew pairing. The captain reported that he received only about 1 hour of sleep during the night before the accident and, as a result, asked the first officer to be the flying pilot for the flight. The first officer reported that he would have preferred not to be the flying pilot because he had just completed a 3-day, 6-leg trip sequence but that he agreed to be the flying pilot because of the captain's references to fatigue and lack of sleep the night before. (The first officer did not verbalize this preference to the captain before the flight.) The flight dispatcher provided the crew with a weather update about 1310, via the airplane's aircraft communications addressing and reporting system (ACARS), indicating that visibility was unrestricted with no snow. The cockpit voice recorder (CVR) recording began about 1316:10. Shortly afterward, the captain stated, "so tired … had about an hours sleep last night. I just tossed and turned." The dispatcher provided the crew with another ACARS weather update about 1407, again indicating that visibility was unrestricted with no snow. About 1429:19, the flight crew received automatic terminal information service (ATIS) information Alpha, which indicated that the ILS runway 24R approach was in use, the landing runway was 24R, the glideslopes for runways 24L and 28 were "unusable due to snow build-up," and braking action advisories were in effect. The first officer then briefed the ILS procedure for runway 24R. About 1442:41, the crew received ATIS information Bravo, which indicated that the ILS runway 28 approach was in use and that the landing runway was 28. Also, this ATIS repeated that the glideslopes for runways 24L and 28 were unusable and that braking action advisories were in effect. Neither flight crewmember discussed the information in each ATIS broadcast about the unusable glideslopes. The weather information in the flight crew's preflight paperwork included a notice to airmen (NOTAM) for runways 24L and 28 that stated, "due to the effects of snow on the glide slope minimums temporarily raised to localizer only for all category aircraft. Glide slope remains in service. However angle may be different than published." During postaccident interviews, both pilots indicated that they had not read this NOTAM. About 1450:14, the captain contacted Cleveland approach control, and the approach controller provided vectors for the ILS runway 28 approach. About 1453:06, the first officer briefed the ILS procedure for that runway, stating the location of the glideslope, descent altitude, minimum safe altitudes, and missed approach procedure. The first officer did not brief the runway length, and the captain did not request this information. The approach controller then notified the flight crew that ATIS information Charlie was current and that the winds were from 290º at 18 knots, visibility was 1/4 mile with heavy snow, and the runway 28 runway visual range (RVR) was 6,000 feet. The captain then stated, "one-quarter mile visibility … well we got the RVR. So we're good there." According to the Jeppesen March 24, 2006, ILS approach chart for CLE runway 28, the minimums for the precision (ILS) approach required an RVR of 2,400 feet or 1/2-mile visibility, and the minimums for the nonprecision localizer (glideslope out) approach required an RVR of 4,000 feet or 3/4-mile visibility. About 1458:46, the approach controller informed a Jet Link flight crew that the flight was cleared for an ILS runway 28 approach and that the glideslope was unusable. The Shuttle America flight crew heard this transmission, and the crew began to discuss how that flight could be cleared for an ILS approach if the glideslope were unusable. About 1459:10, the approach controller instructed the Shuttle America flight crew to descend from 6,000 to 3,000 feet, and the captain acknowledged this instruction. Afterward, the captain stated, "it's not an ILS if there's no glideslope," to which the first officer replied, "exactly, it's a localizer." During postaccident interviews, both pilots stated that they were confused by the term "unusable," but the CVR indicated that neither pilot asked the controller for clarification regarding the status of the glideslope. About 1500:04, the approach controller instructed the flight crew to turn left onto a new heading and intercept the runway 28 localizer. The captain acknowledged this instruction. The first officer then stated, "wonder why they put it on two eight without a … glideslope if it's … ILS." About 1500:30, the controller instructed the crew to maintain 3,000 feet until established on the localizer and indicated that the flight was cleared for the ILS runway 28 approach and that the glideslope was unusable. The captain acknowledged the approach clearance and the altitude restriction but did not read back that the glideslope was unusable. About 1501:09, the captain contacted the tower controller, stating "localizer to two eight." The controller then cleared the airplane to land on runway 28 and reported that the winds were from 310º at 12 knots and that the braking action was "fair." The captain acknowledged the landing clearance. About 1502:01, the first officer stated that the glideslope had been captured. During a postaccident interview, the first officer stated that he and the captain did the "mental math" for a 3º glideslope and that, on the basis of this calculation, they assumed that the glideslope was functioning normally. Also, the captain stated that the cockpit instrumentation showed the airplane on the glideslope with no warning flags. Because the flight crewmembers assumed that the glideslope was working properly, they used the ILS decision height, which was 227 feet above ground level (agl), instead of the localizer (glideslope out) minimum descent altitude, which was 429 feet agl. About 1502:25, the tower controller announced to all airplanes under his control that the runway 28 RVR was 2,200 feet. The controller did not ask the Shuttle America flight crew to acknowledge this information, and the crew did not provide an acknowledgment. About 1502:39, the captain stated, "we're inside the [outer] marker, we can keep going." The first officer then briefed the procedure to go around in case it became necessary to do so. About 1503:04, the first officer stated that the localizer and the glideslope were captured. Afterward, the tower controller announced to all airplanes under his control that the runway 28 RVR was 2,000 feet. Again, the controller did not ask the Shuttle America flight crew to acknowledge this information, and the crew did not provide an acknowledgment. The captain then stated to the first officer, "gotta have twenty four [hundred feet] to shoot … the ILS." About 1503:54, the captain indicated the he was "gettin' some ground contact on the sides" but "nothing out front." The CVR recorded the electronic callouts "approaching minimums" about 1504:46 and "two hundred [feet agl], minimums" about 1504:53. One second later, the captain stated, "I got the lights," which was followed by the electronic callout "minimums" and the first officer's statement, "and continuing." About 1504:58, the captain announced that the runway lights were in sight but then stated that he could not see the runway; this statement was immediately followed by "let's go [around]." The first officer then stated, "I got the end of the runway." About 1505:07, the CVR recorded the 50-foot agl electronic callout followed immediately by the captain's statements, "you've got the runway?" and "yeah, there's the runway, got it." During a postaccident interview, the first officer stated that, when the airplane was 10 feet agl, he momentarily lost sight of the runway because a snow squall came through and he "could not see anything." Flight data recorder (FDR) and CVR data showed that the airplane was about 1,050 feet past the runway threshold when it descended to a height of 10 feet agl. The CVR recorded the sound of the airplane touching down about 1505:29. According to the aircraft performance study for this accident, the airplane touched down about 2,900 feet down the 6,017-foot runway. During postaccident interviews, the captain stated that he thought the airplane had touched down closer to the runway threshold (somewhere between taxiway U and runway 24L), and the first officer stated that, during the landing rollout, he could not see the end of the runway or any distance remaining signs (which appeared every 1,000 feet). FDR data showed that the ground spoilers deployed automatically and that the thrust reversers were deployed shortly after landing (as further indicated by the captain's statement "two reverse" about 1505:33). Although the thrust reversers were initially selected to the full reverse position upon landing, engine reverse thrust reached a peak of only 65 percent N1 (low pressure rotor speed), compared with a maximum of 70 percent N1, for about 2 seconds before the commanded reverse thrust tapered off to reverse idle during the landing rollout. In addition, FDR data showed that the first officer's initial wheel brake application was about 20 percent of maximum and remained relatively steady for about 8 seconds before increasing to 75 percent of maximum. Braking then increased to about 90 percent of maximum when the captain applied his brakes. The antiskid system did not modulate the brake pressure until the captain and the first officer applied their brakes aggressively. The CVR recorded the sound of numerous impacts starting about 1505:50 and a sound similar to the airplane coming to a stop about 1505:57. The airplane came to rest on a snow-covered grass surface located southwest of the extended runway 28 centerline. Available airport movement area safety system video data showed that four flights (all transport-category airplanes, including two 737s) arrived without incident on runway 28 during the 10 minutes before the Shuttle America airplane landed. The airplane that directly preceded the Shuttle America airplane to the runway had arrived 2 minutes earlier. About 1506:04, the tower controller asked the flight crew about the flight's status, but the crew did not initially respond. About 1507:04, the tower controller asked the flight crew again about the flight's status, and the first officer responded, "we're off the runway through the fence … everybody seems to be okay on board." The controller then informed the flight crew that emergency equipment was on the way. The flight crew later reported to Shuttle America and the controller that braking action on the runway was nil. The CVR recording ended at 1519:16.
the failure of the flight crew to execute a missed approach when visual cues for the runway were not distinct and identifiable. Contributing to the accident were (1) the crew's decision to descend to the ILS decision height instead of the localizer (glideslope out) minimum descent altitude; (2) the first officer's long landing on a short contaminated runway and the crew's failure to use reverse thrust and braking to their maximum effectiveness; (3) the captain's fatigue, which affected his ability to effectively plan for and monitor the approach and landing; and (4) Shuttle America's failure to administer an attendance policy that permitted flight crewmembers to call in as fatigued without fear of reprisals.
Source: NTSB Aviation Accident Database
Aviation Accidents App
In-Depth Access to Aviation Accident Reports
