It is not uncommon top be slightly low on the glideslope when crossing the threshold. Used to do it intentionally on the occasional short runway where we were in tight confines in the old days before the aircraft were so closely monitored. On the turboprops, we really used to "dip the PAPI's" close in. I remember one particularly short runway where the PAPIs had you landing halfway down the runway. Unsafe. Still did it in the narrowbody jets at one location that was max landing weight and runway limited weight together. Haven't felt the need to on the widebodies....yet.

Is there really any difference between a smaller jet and a larger jet in the threshold crossing height? As far as I know there is an adjustment in the avionics in order to ensure a similar gear clearance over the threshold.

For example, the Boeing FCTM's have a table for "main gear over threshold" with pilot eye height over threshold/main gear height over the threshold. For the 737-600 the two numbers for a 3.0 degree glide path are 49'/33'. For the 747-400 the numbers are 66'/31'. So being a dot low on either is practically the same reduction in main gear clearance. The real difference between the widebody and the narrowbody is the pilot sitting higher for a given glideslope location.

Who hasn't seen near full scale glideslope deflecton below 100 feet? You might need that extra couple of hundred feet of rollout on a slippery runway that was supposedly "braking action reported as fair/medium".

You might need that extra couple of hundred feet of rollout on a slippery runway that was supposedly "braking action reported as fair/medium".

Especially in the non-normal landing direction, on a rainy day, with all the rubber at the stopping end.
I used to adjust flaps & thrust for takeoff in a similar situation; in case of an RTO to gain a few more meters. It caused various interesting reactions from the other seat when we 'didn't just use the numbers'. But that's another topic and I do to wish to 'creep the thread'.

Years ago a lot of these personal techniques were common in my experience. Folks would add knots to the approach speed for the wife and kids. Non-standard additives would be made for wind with autothrottles in use.

Seems like some of the Air Force guys would brief 'I'm going to duck under a dot on short final since the viz is good' while landing a widebody. Was it taught in the T-38, the C-141 or the C-5 perhaps?

One of the regional carriers decades ago had a non-approved B-737 short field technique of pulling up the speedbrake handle before touchdown so that the boards would fully deploy the moment there was weight on the wheels. When a newer version of the 737 arrived, ground spoiler mode was triggered by radar altitude, the boards fully deployed before the wheels were on the runway and a hard landing mishap occurred.

In recent years, thankfully, most of these cowboy techniques seem to be less common where I've worked. However, with all the airline 'mergers' I'm sure non-standard 'we did it this way at Brand X' procedures will give the feds and the training department a lot of job security for years to come.

Doors to Automatic asked if there is a set threshold crossing height or if discretion can be used by the PF.

There is a published threshold crossing height yes. Judging from the replies there's no consensus about discretion of going low. Could it be the ones who fly into short/slippery runways in small(er) jets and turboprops take it lighter than those flying widebody jets?

Crossing the threshold how many feet lower does being a half dot low equate to? Or a full dot?

"near full scale glideslope deflection below 100 feet"

"how many feet lower does being a half dot low equate to? Or a full dot"

Depending on the type and location of GS ground installation the aircraft indications will progressively show below the ideal path because of the parabolic nature of a reflected beam. Thus the GS scale is not a suitable indication of height or deviation at lower altitudes.

Crossing the threshold a half dot, or a dot, is a relatively small displacement. IMO a bigger issue is demanding "don't do this, don't do this" so guys make relatively large changes in aimpoint/touchdown point because of an over concern about being slightly low.

At 100' a half dot is roughly 2.5 feet. A dot is five feet(+/-). If your aimpoint/flight path is correct you'll cross the threshold 1-2' from ideal. That's not the end of the world. Making a significant correction and losing your aimpoint/flight path, and perhaps requiring a secondary correction, puts you at risk of a bigger deviation from acceptable performance.

Hmm. Rule of thumb for stabised approaches (1 dot high max, no lower than 0,5 below) is the error becomes +20 % and -10% of the height. Your numbers would be wrong by factor of four. Anybody care to do the math?

By the way I'm in the "aim to land as calculated using as much muscle memory/visual drill as possible". The humble observation being, that "out of the box" / "this one time only" techiques make for longer landings, at least on 50% of the attempts. Uglier ones for about 90%. (A320)

As I said, we used to land turboprops(no reverse by the way on the engines) on a short runway less than 3000 feet(and some that were a bit longer). Was 100% hard pack snow all winter.

Now some folks might call you a cowboy if you don't follow the official flight path all the way down and land at about 13-1500' down the runway(which has a downslope in both directions for the latter half of the rollout) because you followed the PAPI and had your 50' wheel clearance....

...but in reality, you would be a fool if you were anything but a so-called cowboy. And yes, it was done during a check flight. Some 10,000 foot plus runway guys have a hard time understanding stuff like this.

Why the gripe? Different OPS, different standard techniques... Carry one across to the other field, and you're set for trouble.

BTW it is not wheels at 50', but G/S antenna, the wheels are at 33 over the greens. LOT of difference. 33' wheel clearance also illustrates, why "going low" brings little benefit.

Airbus: pilot eye / antenna / gear
320: 56 / 50 / 34
330: 59 / 50 / 28

There seems to be an excessive amount of 'this is what I did' versus 'this is what they did'.

How do you recognize an Airline Pilot at a party, he'll tell you comes to mind.
Enough about me let's talk about my boat.

What you did 30-40 years ago in your 3-holers has no relevance to this accident.
Why they were doing 130kts GS 4600' down the runway does.

FlightDetent, re calculating deviation in feet from dots above/below the glideslope and, "Anybody care to do the math?"

I looked this up some time ago on PPRuNe when I was working on FDM Program determination of touchdown points to ensure the correctly-calculated distance from 50' to touchdown.

The matter was discussed, along with the mathematics, at:

http://www.pprune.org/questions/3622...ml#post4719685

That post suggest's one dot below equals to 0,2 degress angular displacement. Some other sources on the web state 0,15 and 0,35. Difference too large to draw conclusions. Off the topic, anyhow.

Thanks still.

There isn't a "standard dot". Some instruments have 5 dots per side (full deflection), while others might have only 2 dots or 3 dots for the same deflection. At the extremes I've seen older instruments with no dots at all and one with 10 dots per side.

Assuming a 3 degree GP, on a 5-dot display one dot is approx. 0.15 degrees. On a 2-dot display, one dot is approx. 0.35 degrees.

Has anyone closed in on factual truth behind this over-run. One assumes the landing performance was adequate, so what lined up a few holes? All this talk of dots & degrees & slopes & displacements is giving me a headache. KISS. Keep the crash point fixed in the window at the correct speed and you should have a modicum of success more often than not.

B2N2, re, "Why they were doing 130kts GS 4600' down the runway . . . ", where'd you get the "130kts" & "4600ft" from?

Firstly, with as little personal input as possible:

https://aviation-safety.net/wikibase/wiki.php?id=190959

https://www.flightradar24.com/data/a...n278ea#b70e06f

https://s19.postimg.org/nj5b30wq7/pence_chart.png

https://s19.postimg.org/674yhl38v/pence_raw.png

my notes:
. FR24 claim to use "FAA data", what it means I do not know, however the recorded altitude seems to be corrected for QNH (which is not the case with ADS-B raw values)
- For ADS-B raw data, speed = ground speed
- Typically for ADS-B raw data, altitude = 0 means "positive weight on wheels", not the actual value
- ALT = 0 at time 23:40:34 should be interpreted with extreme caution as the speed for that frame is invalid.

Based on the above, the time lapse to slow down on ground from 135 >> 132 kt is 18 seconds, which translates to 1200 meters of distance covered. This is ignoring the suspicious reading at 23:40:34.

Fact: LDA = 2100 m
Observation: B2N2's quoted figure of 4600 ft = 1400 m.

Link provided by PJ2 shows 1 dot = 6.6' at 100' AGL.

So one dot = roughly 3' crossing the threshold.
1/2 dot = 2'(rounding error).

I'm not saying ignore the vertical deviation but IMO aim point is a significant factor in the entire process.

FlightDetent;

Oh, okay, thank you. I just took it for granted that with a number that precise, that accuracy was also implied from an NTSB communication or something that I'd missed.

That kind of accuracy can't be done from FR24 etc., for the reasons you've stated so we'll see what the Interim Report says.

Other than ADS-B, sites like FR24 get a "near real-time" data feed from the FAA (called the ASD feed). This feed contains flight plan and position data (including lat / long, ground speed) for all IFR aircraft in the US air space.

Don't confuse precision with accuracy. :O

FWIW, that's almost certainly ADS-B-derived data, it's too granular to have come from ASDI.

PJ2: The FR24 data is not authoritative, but still coming from the A/C's data buses, precise.

The position recorded at time frame 23:40:49 is here https://www.google.co.uk/maps/place/...8!4d-73.879684. Distance measured from the threshold 4650 ft, GS = 132 kt.

From that point the distance remaining to DERA is 710 m, figure just about enough to stop with max braking effort corresponding to AFM. Already tested in real life: http://avherald.com/h?article=463e0584

DaveReidUK, re, "Don't confuse precision with accuracy. http://cdn.pprune.org/images/smilies/embarass.gif",

Yes, agree - if the writing came across that way, it wasn't intended - I'm aware that one can be both precise and inaccurate, etc.

FlightDetent;
My larger point is not to rely upon FR24 for any serious data work, particularly lat/long. The system isn't designed for it and we don't know the concurrence between a commercial, non-specialized system such as FR24 data and a system engineered for such data recording like the flight-recorder/QAR data.

Another way to look at it is to consider that ADS-B data is designed principally for surveillance (i.e. as a radar substitute) - that's what the "S" stands for - so it's perfectly adequate for use by ATC to ensure separation.

But attempting to use it as if it was a mini-FDR is fraught with difficulties, not least because all those FR24 data points before touchdown that purport to show both a lat/lon and a velocity are fabricated from two separate transmissions from the aircraft at different points in time, one with each of those two parameters.

Incidentally the QNH at the time of the incident, according to the METARs, was about 1020 mb so the height readouts are likely to be about 200' adrift from true AMSL values.

P2J, your point is valid and agreed with. No official conclusions to be drawn on data recovered from FR24 amateur/DYI channels. It is not designed for such purpouse neither certified.

Yet you know well, that on your A/C, the DMC feeds ADIRS data to QAR, FDR and the ADS-B out unit. Unless the transport layer - FR24 - corrupts or adjusts that data, all readouts will eventually turn out to be the same.

Case study:

A) Unofficial source "FR24" overlayed by an anonymous internet enthusiast (me) over non-purpouse built resource "Google Maps":
Attachment 1256

B) NTSB findings:
Pending

C) Unofficial source "photography" taken by a non-licenced, commercial journalist:
Attachment 1253
Attachment 1252

Now :), everybody's welcome to not draw their own conslusions.

Overlay of FR24 recorded positions where ALT = 0.
Attachment 1255

Thanks FlightDetent. I understand your pov as well. The data is there, "why not use it to see what happened."

If I may...

The industry is recognizing that it is becoming a significant challenge in flight safety and investigative work that unofficial sources like everyone's cellphone, Twitter and FR24 can influence widespread understanding and the drawing of conclusions of an incident or accident. It is true that both can be helpful in the "official" investigation but only in the hands of trained and experienced investigators.

I know the tide cannot be stopped from coming in and that today, "official" & "unofficial" have become maleable concepts. However, from a retired airline pilot's and safety investigator's pov, the above cautions, which apply to all aviation incidents, ensure that as solid a set of conclusions are derived from the sources designed for the work and that appropriate corrective actions are enabled.

The photograph of the flaps I referenced at the beginning of the thread, showed them in the takeoff position. Why were they in that position and not the 30 or 40 position? The question can't be answered by FR24 or anyone here until the NTSB releases their interim work. In the meantime, we dont' know the flap position's contribution to the overrun, if any. Did they have a hydraulic issue - likely not, but we don't know.

Again, thanks for your clarifying response.

The feedback is my reward. :ok:

--

The thread had run off on two tangents, threshold crossing height and techniques, plus data usability from public domain. Back on track now?

Runway overruns are rather simple physics. Either the landing is deep, excess energy is carried, or braking performace is degraded. Typically any single one is not sufficient enough to cause an exursion. A 737 touched down in Jamaica, December 2009, with 162 kt GS, 4100 feet past the THR and thrust well above idle - then the result becomes inevitable.

In this case, however, the FR24 data (to be verified if accurate) show optimal if not shorter touchdown position, and quite normal, stable speed. What is little perplexing, is that for the 65% of overall ground run no speed reduction is recorded.

"What is little perplexing, is that for the 65% of overall ground run no speed reduction is recorded."

That may be a function of the ADS-B not transmitting after ground-mode is set by the touchdown.

Re, "Runway overruns are rather simple physics."
Agree of course, but then it becomes much more complex. The first key in data analysis of air distances from the threshold to touchdown is accurate determination of the touchdown point. That is nowhere nearly as easy as it sounds. It's slightly better if one has good, on-board GPS-derived lat/long data.

Perhaps the following is orthogonal to the thread topic but as long as we're determining causes and prevention of overruns, the following may be tangentially-useful.

The NLR provides some very good papers on such determinations. At the 2nd Conference of the European Operators Flight Data Monitoring forum (EOFDM), Gerard van Es was invited to present "Monitoring Landing Overrun Risk using FDM: Tips and Tricks". The presentation can be found here.
http://essi.easa.europa.eu/ecast/wp-...Excursions.pdf

This page gives you some general information about runway excursions. What are the statistics telling us, what initiatives are taking placing to prevent runway excursions etc.
NLR-ATSI Runway Excursions

The FAA Document, (DOT/FAA/AR-07/7) provides a very good discussion of all aspects of landing performance. It has a $30 title, but is well worth reading!
DOT-FAA-AR-07-7_A Study of Normal Operational Landing Performance on Subsonic Civil Narrow Body Jet Aircraft during ILS Approaches - ar077

I lead you astray ... the altitude is zeroed, yes. Both heading and speed are however alive:
Attachment 1258

Ah, okay - thanks for the screenshot by the way - the original FR24 data is no longer available to non-members.

Yes, it is.

CSV and GE/KML here:

Flightradar24 Data Regarding Runway Excursion at La Guardia Airport

Appreciate it, thanks.

On EMAS and ILS

Does anyone know if there is information about installation of EMAS and the effect of that ground-plane change on ILS LOC and G/S signals?

Did the installation of EMAS at LaGuardia require new modelling and is EMAS included in the latest reflection models? And were these used at LaGuardia?

Also, is it possible to widen the EMAS beyond the width of the runway without having a negative impact on ILS signals?

The reference to AA in Jamaica was only made to exemplify that (out of three) more than just one element needs to go wrong - and in a significant manner - for a runway overrun to happen. In this respect that accident report is like a box of chocolates. And if it were not for the contradictions there would be no excursion to report on. :suspect:

On the contrary, from the little that is known, the accident in KLGA seems somewhat unusual as both the touchdown point and energy at landing look quite allright.

"Ducking under" - illegal?
 

91.129(2) Each pilot operating a large or turbine-powered airplane approaching to land on a runway served by an instrument approach procedure with vertical guidance, if the airplane is so equipped, must:

(i) Operate that airplane at an altitude at or above the glide path between the published final approach fix and the decision altitude (DA), or decision height (DH), as applicable; or

(3) Each pilot operating an airplane approaching to land on a runway served by a visual approach slope indicator must maintain an altitude at or above the glide path until a lower altitude is necessary for a safe landing.

NTSB Issues Investigative Updates for LaGuardia 737 Runway Excursion, Fort Lauderdale MD-10-10F Gear Collapse, Fire
Nov. 22, 2016

WASHINGTON — The National Transportation Safety Board issued investigative updates Tuesday as part of the agency’s ongoing investigation of a Boeing 737 runway excursion at New York's LaGuardia Airport and a Boeing MD-10 landing gear failure accident in Fort Lauderdale, Fla. Both events happened in late October.
Eastern Air Lines Runway Excursion

A Boeing 737-700 (registration N278EA), a chartered flight operated by Eastern Air Lines Group, overran Runway 22 during landing at LaGuardia Airport, Flushing, New York, Oct. 27, about 7:42 p.m. EDT. The airplane veered to the right during the overrun and partially transited the Engineered Material Arresting System before it came to a stop on the turf about 200 feet from the runway end. Instrument meteorological conditions prevailed at the time. The 11 crew and 37 passengers, including then vice presidential candidate Mike Pence, deplaned via the airstairs. The flight originated at Fort Dodge, Iowa, about 4:23 p.m. CDT.

NTSB investigators document the B-737 involved in the runway excursion. Photo: Peter Knudson/NTSB

Initial findings include the following:
- The airplane was manufactured in 1998. It had accumulated 48,179 total flight hours with 17,098 total flight cycles at the time of the incident.
- Investigators retrieved the flight data and cockpit voice recorders shortly after arriving on scene. The recorders were transported to the NTSB recorders lab for download. Both recorders contained good quality data.
- The flight crew reported the landing followed a stable approach.
- The flight crew reported sighting the runway when the airplane was about 700 feet above the ground; they said that the airplane “floated” during the landing flare; the main landing gear touched down on the wet runway about 3,000 feet beyond the runway threshold.
- The speed brakes were manually engaged about four seconds after touchdown. Thrust reversers were deployed about seven seconds after touchdown.
- The flight crew did not report any mechanical irregularities or abnormal braking action, which was corroborated by the flight data recorder.
- Investigators interviewed flight crews of the four airplanes that landed immediately prior to the incident airplane; none reported any problems with braking action on the wet runway.

The on-scene investigation was completed Oct. 30, after which the airplane was released to the operator. The performance of the EMAS will be examined to determine its effect on the deceleration of the airplane.

Parties to the investigation are the Federal Aviation Administration, Eastern Air Lines Group and Boeing.

200 ft for every second of late lift dump. (4 x 200 ft)
200 ft for every second above 2 sec for late deployment of reverse. (~ 500 ft)
How much reverse, what level of braking was chosen for the reported runway condition?
Add that the the reported long flare (3000 ft?) and compare with the landing distance available.

Then consider at what point during the landing roll-out the crew could be aware that the aircraft might not stop in the distance remaining; would use of all available deceleration devices at this point be able to stop the aircraft in the remaining distance.

'Avoiding an overrun' http://www.faa.gov/documentLibrary/m.../AC_91-79A.pdf

SafetyP, I wondered the same things. They have all the data. Why not comment on deceleration efforts? Why the comment that the speedbrakes were "manually engaged", (thinking here that they weren't armed prior to touchdown, or there was some hydroplaning?).

Guess we'll have to wait for the report.

I believe that the auto function of the speedbrakes were MEL'd.

Pence?s skidding plane had inoperable speed reducer, NTSB says | Newsday

Which means that you want to get them and the reversers out quick in a situation like this. Can be done in 1-2 seconds after touchdown. Best to make sure of a 1500' touchdown or less as well.

Curious about the autobrake setting.

Just a comment
 

"- The speed brakes were manually engaged about four seconds after touchdown. Thrust reversers were deployed about seven seconds after touchdown."

It seems to me that 4 seconds is a long time to get the spoilers up. Just me, as soon as I had the thrust levers at idle and was on the ground, I was pulling the spoilers up (707/727 and 737 with inop auto-speedbrake)

3 seconds after speedbrake deployment might be OK for the T/R's to be fully open.

Maybe he was pulling on the reversers first, and being unable to get them up due to no weight on the squat switches, then remembering to pull the speedbrake handle?