Wednesday, November 9, 2016

“Pence Plane” EA3452 Overrun

While thankfully no fatalities, having three airplane accidents in less than 24 hours is stunning.

1) 737 overrun at LGA
2) 767 uncontained engine failure and fire – Hull loss
3) MD-10 landing gear collapse and fire – Hull loss

All three of these accidents are under investigation by the NTSB, from which probable cause will be determined in due course. My purpose is to gather the reported information and offer some analysis of what may have happened.  The information is limited, leaving the potential for misunderstanding and mistake. Why something happened, and finding probable cause, are with the safety board and their exhaustive investigation.

Part 1: EA3452 overrun at LGA (the Pence plane)

The analysis of the overrun at LGA relies extensively upon ADS-B data reported by ADS-B reports offer position, heading and speed data. While significant shortcomings hold back ADS-B in comparison to flight data recordings, there can be enough information to yield insight.

Just after sunset Thursday evening, 27 October 2016, a charter 737 appears to have landed long and failed to stop before entering the Engineered Material Arresting System (EMAS) at the end of runway 22.

Flightradar24 shared ADS-B reports which offered a glimpse at the conditions leading to the overrun. The positon information has been converted to distance from the threshold to runway 22. Runway 22 is 7001 feet long.

Ground speed plotted against distance down runway 22 displays an alarming lack of deceleration until about 5,000 feet traveled, entering the EMAS at over 16 knots.

The last reported position was 7,197 feet from runway 22 threshold, or about 196 feet overrun.

The airplane appears to have encountered an upset on short final pushing its ground speed up a bit.

Forecast winds indicate a tailwind of about 7 knots on final approach to runway 22. It is possible that the tailwind picked up at low altitude – by the limited measurements about 1 knot in three seconds, very easily offset with a bit of power.

KLGA 272330Z AUTO 09009KT 3SM +RA BR BKN008 OVC013 13/12 A3011

KLGA 272335Z AUTO 07008KT 3SM RA BR BKN008 OVC013 13/12 A3011

KLGA 272340Z AUTO 08009KT 3SM RA BR OVC008 13/12 A3011

KLGA 272345Z AUTO 09007KT 4SM +RA BR BKN008 OVC012 13/12 A3011

Altimeter 30.11 offers a correction of +190 feet to the ADS-B reported altitude.

Report from NTSB (Bloomberg) preliminary data.

The plane’s crash-proof data recorder showed the plane touched down at 140 miles (225 kilometers) an hour, Sumwalt said 
The preliminary data released by the safety board may offer some clues hinting at why the plane didn’t stop. Devices known as spoilers, panels on the top of the wings that come up after touchdown to help slow a plane, were broken, requiring pilots to set them manually, Sumwalt said. They deployed four seconds after touchdown, which is several seconds later than if they had come up automatically, he said. 
The plane’s thrust reversers, devices on the engines that reverse airflow and also help slow the aircraft down, were engaged about seven seconds after touchdown, he said.

140 mph = 122 knots, which is estimated to be at around 5,000 feet from runway 22 threshold, and more than 18 seconds since crossing it. This is well beyond mid-field. This where the touch-down is estimated.

Four seconds after touchdown is about 5,400 feet from runway 22 threshold. This is where spoilers were applied.

Seven seconds touchdown is about 6,000 feet from runway 22 threshold, and only 1,000 from runway end. This is where thrust reverser is applied.


  • There appears to be a mild, increasing tailwind shear on short final. Otherwise there is no data to show any wind changes.
  • The landing was made with about a seven knot tailwind.
  • The touchdown appears to be made about 5,000 from the threshold of runway 22, with only 2,000 feet remaining.
  • The spoilers came on late, and that may have contributed to less braking in the first few seconds after touchdown.
  • It may have been possible to reverse thrust sooner, but it seems unlikely to have enough contribution to avoid the overrun.
  • There does not appear to be any attempt to miss the approach. The flight crew had over 15 seconds to initiate a go-around, and floated well beyond mid-field, but instead chose to continue the landing. 

Traffic Flow

Here are some depictions of the traffic flow into LGA with EA3452 highlighted.  Take note of the day/night discriminator passing through at the time.

Engineered Material Arresting System (EMAS)

Follow the following link to read all about EMAS.

Stay tuned!

Peter Lemme
peter @

Follow me on twitter: @Satcom_Guru

Copyright 2016     All Rights Reserved

Peter Lemme has been a leader in avionics engineering for 35 years. He offers independent consulting services largely focused on avionics and L, Ku, and Ka band satellite communications to aircraft. Peter chairs the SAE-ITC AEEC Ku/Ka-band satcom subcommittee developing PP848, ARINC 791, and PP792 standards and characteristics. 

Peter was Boeing avionics supervisor for 767 and 747-400 data link recording, data link reporting, and satellite communications. He was an FAA designated engineering representative (DER) for ACARS, satellite communications, DFDAU, DFDR, ACMS and printers. Peter was lead engineer for Thrust Management System (757, 767, 747-400), also supervisor for satellite communications for 777, and was manager of terminal-area projects (GLS, MLS, enhanced vision).

1 comment:

  1. I learned to fly at LGA long ago but I remember something that may
    apply here. It was common practice for pilots to land long if they were heading to the Marine Air Terminal. If you did a 'by the book' landing, you would touch down on the numbers or a bit farther, turn off the active, and spend a bit more time taxing to MAT. In my case, flying a Cherokee 140, it was not an issue, but I routinely saw large jets land long for MAT. We know Pence was running late and I can't help but wonder if the crew was pushing it a bit and things got out of hand. Flight conditions were certainly not ideal and are likely a contributing factor.