Southwest Flight 345 Accident: NTSB reports the aircraft pitched downward in the final four-seconds, landed on the nose gear

NTSB is continuing their outstanding effort to quickly share investigative information with the public. In a July 25th Press Release, NTSB provided the following ‘facts’ related to the July 22nd nosegear collapse that injured three at New York’s LaGuardia Airport:

  • The airplane’s front landing gear collapsed on landing.
  • Evidence from video and other sources is consistent with the nose-gear making contact with the runway before the main landing gear.
  • The flight data recorder on the airplane recorded 1,000 parameters and contained approximately 27 hours of recorded data, including the entire flight from Nashville to New York.
  • The cockpit voice recorder contains a two-hour recording of excellent quality that captures the entire flight from Nashville to New York and the accident landing sequence.
  • Flaps were set from 30 to 40 degrees about 56 seconds prior to touchdown.
  • Altitude was about 32 feet, airspeed was about 134 knots, and pitch attitude was about 2 degrees nose-up approximately 4 seconds prior to touchdown.
  • At touchdown, the airspeed was approximately 133 knots and the aircraft was pitched down approximately 3 degrees.

    The nosegear was jammed up into the lower fuselage, damaging this electronics bay.

  • After touchdown, the aircraft came to a stop within approximately 19 seconds.
  • A cockpit voice recorder group will convene tomorrow at NTSB laboratories in Washington to transcribe the relevant portion of the accident flight.

When Asiana 214 crashed at San Francisco, the evident pilot error was a failure to notice that airspeed had decayed to an imminent stall. But, with this Southwest crash, it appears the opposite problem may have been involved: the failure of the flight crew to recognize they had too much speed.

The table below is taken from Flightaware. The data, from left to right, is: time, latitude, longitude, course, direction, groundspeed (KTS), groundspeed (MPH), altitude, rate of descent, down arrow, and radar data source. This data is computer-generated, using digital data obtained from FAA radar. The column in red is the calculated compass heading, while the column in blue is calculated groundspeed, in knots. This data shows clearly that the Boeing 737 was aligned for a landing on Runway 4, and tracking inbound from the time when they were descending through roughly 8,000′ altitude. Data at Flightaware indicates that the accident flight was running an hour and ten minutes late.

02:35PM 40.3000 -74.3667 66° East 318 366 8,700 -1,140 Descending  New York TRACON
02:36PM 40.3667 -74.3000 37° Northeast 290 334 7,900 -960 Descending  New York TRACON
02:37PM 40.4333 -74.2333 37° Northeast 273 314 6,800 -1,560 Descending  New York TRACON
02:38PM 40.4833 -74.1667 45° Northeast 274 315 4,800 -1,860 Descending  New York TRACON
02:39PM 40.5500 -74.1000 37° Northeast 270 311 2,900 -1,740 Descending  New York TRACON

Contrast these numbers with a few other flights. In particular, note that the airspeed on the accident flight (highlighted in yellow above) at 3,000′ altitude was slowing from 274 knots to 270 knots. Flight #3208 and Flight #435 fly BNA to LGA everyday. Today’s Southwest Flight 3208 landed at 7:13AM; when it was at 3,000′ altitude, the groundspeed was 199 knots, and that was on a downwind leg, heading southeast to land Runway 31. When Southwest Flight 3208 landed at 7:03AM on the day of the accident, it was almost on-time (it landed just 5-minutes behind schedule) on a straight-in approach to Runway 4, and showing a 201 knot groundspeed at 3,000′ altitude. The same flight as the accident, but the day before, also landed Runway 4, but it landed only 13-minutes behind schedule; at 3,000′ altitude it was slowing from 230 knots to 203 knots. Translation: the accident flight was in a hurry, and likely still getting their speed under control at the moment of impact.

Some news articles are suggesting pilot error, but there is more to this incident than simple human mistake. Pilots all know that winds can vary considerably, and this can become a problem while flying an approach. Windshifts at the surface are closely monitored and recorded, creating a solid data history. But, these official numbers tend to lag behind unrecorded changes at altitude, going on as the different air layers mix. Thus, if the LGA weather on July 22nd was unsettled and shifting (and it was), recorded windshifts would tend to happen after those more significant shifts happened during the approach of an aircraft like Southwest Flight 345.

The table below is the weather data for that day, as located at Weather Underground. A METAR is the hourly weather observation, and it is typically done at or shortly after the :50-minute mark; note how the METARs above are all time-stamped at 51-minutes past the hour. A SPECI is a special weather observation, which is done shortly after an accident, or for specified significant changes in weather conditions, such as a wind-shift. In this case, the winds prior to the accident were directly aligned with the landing runway Runway 4 (i.e., the first three-digits in the wind data are ‘050’ and/or ‘040’, meaning the wind was from those compass directions, which parallel Runway 4). But, the winded shifted a full ninety degrees with the next regular METAR, at 6:51PM; beginning with that weather observation, winds were clearly favoring Runway 13, thus the need for a runway change by ATC. (Note, too, given the configuration of airports around New York City, it appears Runway 13 may be the least favorite landing runway for ATC). The 5:58PM SPECI was generated due to the accident, which happened at 5:45PM (according to NTSB’s first Press Release, issued on July 23rd). The 7:12PM SPECI was triggered by the development of a low ceiling, a broken cloud layer at 1,300′ (see the code BKN013, highlighted in light green).

4:51 PM METAR: KLGA 222051Z 05012KT 8SM FEW028 SCT050 BKN070 25/22 A2986 RMK AO2 SLP111 60000 T02500222 56016
5:51 PM METAR: KLGA 222151Z 04008KT 7SM FEW030 SCT050 BKN075 OVC130 25/22 A2985 RMK AO2 SLP106 T02500222
5:58 PM SPECI: KLGA 222158Z 04007KT 8SM SCT020 BKN080 OVC130 26/22 A2984 RMK AO2
6:51 PM METAR: KLGA 222251Z 13008KT 10SM SCT020 BKN080 OVC130 27/23 A2984 RMK AO2 SLP105 T02670228 […accident was at 7:03PM, on Runway 4; pure crosswind…]
7:12 PM SPECI: KLGA 222312Z 12008KT 10SM -RA BKN013 BKN018 BKN050 27/23 A2984 RMK AO2 RAB08 P0000

So, was this pilot error? Maybe, maybe not. The most important fact revealed thus far is that the aircraft pitched downward and landed nose-first, in a span of just four seconds. The airspeed was somewhat high at 270 knots while  at 3,000,  but the flightcrew fully reduced this to a normal landing speed in the next couple minutes.  And, the weather data definitely shows significant shifting winds that would eventually necessitate runway changes by ATC. So, while the cause of this late and dangerous pitch was most likely windshifts, a contributing factor may have been that the crew was rushed.

One detail remains not yet clear: when did the windshift actually occur? The criteria for doing a SPECI weather observation include a windshift of 45-degrees or more. Between 5:58PM and 6:51PM, the wind shifted a full 90-degrees. Such a shift would produce a SPECI, so what are the odds this shift occurred in the ten minutes just prior to 6:51PM METAR? Another question: what kind of pilot reports were being collected by ATC, from other arrivals on final to Runway 4? Were any reporting windshifts or loss/gain of speed?

Obviously, NTSB will be looking closely at when the shift happened and if the SPECI was missed. And, NTSB will be factoring in just how resistant ATC may have been to changing runways, during windshifts in some of the most complex airspace in the world.