Southwest B737 Overrun @ Chcago MDW

[the_hawk]

Somethings either don't add up / equal the problem

The only possibility I see at the moment how the statements you quoted fit is that dispatcher and crew agreed on a better rwy condition (and braking action) than occured (than the base for the NTSB calculations leading to 5300 ft).

Anyway, the 18 seconds probably will be the primary cause, let's wait and see how this happened.

[RatherBeFlying]

Approximately 20 minutes prior to the accident... temperature [+] 3C...
Approximately 23 minutes after the accident... temperature -4C

i.e. that's wet snow/slush turning into icy snow/slush

[Ignition Override]

1) Regarding the law suit filed by Kreindler & Kreindler, no matter what the merits of the case, the constant US obsession (no matter what the business) with liability concerns has, for many years, placed limitations on the information which might quickly be shared between a given airline's Fleet Training/Standards and its pilots. But this has also limited the speed, accuracy and potential distribution of helpful, even critical information between Flt. Ops and other pilot groups whose daily jobs on the same aircraft type(s) control the safety of thousands of passengers (and expensive cargo).

Very valuable info can be learned, without reading the misleading, edited versions in the press (most don't bother to learn the difference between a flap and an aileron, or whether the aircraft has automation etc), which is the only aviation source for the public, or waiting many months for the final results from the NTSB.

2) Do pilots out there ever land the simulator and immediately do a balked landing go-around?

My deepest condolences to the family of the child who was lost in the tragedy.

[ni7irs]

Response to NTSB Report

Folks,

I have watched this post with interest. Is there any doubt that the approach should have been discontinued because of the calculations posted by the NTSB? Also, how is it possible that 18 seconds could elapse before the co-pilot deployed the reversers. Finally, why did the co-pilot need to move forward? Shouldnt the copilot be in position to assist the pilot in landing?

[Halfnut]

For the sake of argument when landing on a sheet of clear polished ice the discussion of which braking system will work better autobrake vs. manual brakes is mute. A 1965 Volkswagen Beetle’s brakes could lock up all four wheels on a 737. Antiskid would use the Volkswagen’s brakes to bring the 737 main wheels up to brake release point and then cycle the brakes until the aircraft came to a stop many miles later or the brakes melted.

Using this theory at MDW I seriously doubt there was enough coefficient of braking for the argument of which system is better matters. Once the tires reach the antiskid release point you are at max braking and I’m sure this night the autobrakes at MAX were doing just as good as job if not a better job then the two guys up front who were busy fighting the T/Rs.

On the DC-9-82 (which I do know about) the autobrakes use the right brake system only for landing and both brake systems for takeoff. When the pilots use the brake peddles for manual braking they get both brake systems all the time. MED & MAX brakes will apply two seconds after the main wheels are down.

[Astrosfan]

Stu Bigzorst:

I'm a little confused.

If a 737-700 needed 7000+ feet to land at MDW as you have calculated or else an overrun was inevitable -- and we know the runways at MDW are shorter than that -- then how do you explain all the previous safe landings of WN's 737s immediately preceding this accident? Shouldn't they all -- or at least all of the 737NG landings before WN #1248 -- have wound up parked in the street if your calculations were correct? Or am I missing something?

[Stu Bigzorst]

I just pulled the figures out of my 737-700 QRH. Boeing calculated them, not me.

Maybe the snow fell heavily and the wind shifted a bit. I wouldn't have made the approach, not least because it's not allowed (t/w + contam) in our operation.

I'm sure the NTSB will let us know everything in due course.

[DingerX]

Just a clarification on reading skills:

Preliminary calculations also show that, for the runway conditions and use of brakes and thrust reverser that occurred, the stopping distance without hitting obstructions would have been about 5,300 feet (the actual stopping distance was about 5,000 feet). In addition, had the airplane landing into the wind, rather than with a tail wind, the stopping distance for a landing would have been about 1,000 feet less.

The key phrase is "for the runway conditions and use of brakes and thrust reverser that occurred".
All this NTSB calculation tells us is that, given the conditions of MDW at the time, if a 737 touches down on 31C with max autobrake, then later applies full manual braking and 18 seconds into the landing roll the thrust reversers pop out, it will need 5300 feet of runway to stop.

So either A) the conditions as understood by the crew were different then those actually existing, or B) they fully understood the conditions and pursued a landing that depended upon the timely and effective operation of thrust reversers, or C) the NTSB's Preliminary Calculations are somehow flawed, and something else (such as the Global Conspiracy of Hot Dog Stand-Operators) is involved.

I'll now duck out and let you continue bickering about autobrake vs. manual braking.

[the_hawk]

A) and B) I assume, while the crew probably knew of the t/w they might have expected less contamination / better braking action, in any case the crew certainly knew that there was not much room for error and B) is correct as well

[Willie Everlearn]

westhawk

Thanks for posting the NTSB update. It says alot about the result and I'm sure the final report will reveal even more. It sounds to me like the crew performed to standard with an unexpected (as well as disappointing) result.
=====

Stu Bigzorst

In light of your calculations (undoubtedly acccurate) I think as pilots we need to also be reminded that the QRH for contaminated, ice covered and snow covered surfaces are "Advisory" and not regulatory. So I'd say Boeing (and others) would prefer at least to ensure the numbers 'look good' or 'better' than they should (at times). This seems to be a case where those figures indicated a strong potential for an overrun. If your calculations are accurate then the actual result in landing distance is probably greater than your calculations because of this.
We have to work with what we're given.

We now know as well why they landed with a tailwind.
=====

As curiosity seekers, it's easy for us to say or suggest what should have been done or not even considered following a known result. This is one way in which we acquire experience. Learning from the misforturne of others. Perhaps it's safe to admit many of us would have likely handled this landing in the same manner.

If you've been in aviation long enough, you know there are times when "there but for the grace of God" applies and of those times you may have gotten away with it.

[Huck]

I've read that NTSB update twice, and I can't find any mention of whether the spoilers deployed....

[boofhead]

When looking at the figures posted here, remember that most of them are "Net" numbers, not "Gross". that makes a difference when considering what should have happened compared to what did happen.

Nigel, the company will select a suitable V1/Vr (called V1, Vr ratio) based on averages for the expected operation, or will accept a standard figure for the entire fleet operations. Certainly using a variable ratio will result in better numbers and a greater payload carried, but most pilots are not trained in this procedure, with a greater risk of error.

[hopharrigan]

I can't help myself; I have to make a comment about procedures.
What makes any airline think it has more knowledge about operating airplanes than does the manufacturer? Why does every airline have procedures that are different from each other, and different from the recommended procedures as published by the manufacturer, who has access to all the data out there, and is in the best position to decide what works and what does not?
I worked for many airlines, and only one (Royal Brunei) used Boeing procedures, (which worked marvellously well). Every other airline used locally designed procedures, changed every time a new Chief Pilot took up residence. One of the more aggravating ways procedures were changed was to "standardise" them with other airplane types in the fleet. Thus if there was a mix of Airbus and Boeing, or an airplane that did not have recent technology (auto brakes?), all the airplanes had to use procedures that were workable on the oldest or weakest airplanes. It dumbs down the operation and begs the question of what will happen when the old airplanes are retired, leaving the newer generation airplanes still being operated in a degraded manner. Is this a factor here?

[Halfnut]

http://abcnews.go.com/Business/Flyin...ory?id=1412384

Column by John J. Nance

New Report Only Deepens Mystery of Midway Runway Accident

Investigators Find Southwest Flight Needed 800 More Feet to Land Safely

Dec. 16, 2005 — - The National Transportation Safety Board has just released some preliminary information on last week's runway overrun accident at Chicago's Midway Airport, but those hoping for a quick explanation will be disappointed. The report raises new issues about what went wrong aboard Southwest Airlines Flight 1248.

This, however, is the very nature of accident investigation. Let me explain.

First, the most important consideration is that there is never, ever just one cause behind a major airline accident.

An airliner's landing sequence is a series of complex steps and decisions, all made against the background of thousands of established procedures, the company and professional culture, and legal and regulatory structures.

A single pilot decision may be based on company and Federal Aviation Administration decisions and practices, information provided (or omitted) by airfield authorities, decisions (such as the dispatcher's contribution) made in distant places, or even a sudden change in the weather. In fact, virtually all of such considerations may have been involved at Midway when the Boeing 737-700 overran Runway 31C and rolled onto a highway.

The plane crashed into a car, killing 6-year-old Joshua Woods.

The NTSB's interim report raises several issues, each of which may (or may not) be a crucial contributing factor:

1. Air-traffic control authorities chose to use Runway 31C even though there was a tail wind, probably because the instrument approach procedures to guide pilots to Runway 13C, which runs in the opposite direction, were unusable due to low visibility.

This means that if Midway's tower switched to 13C, the airport would effectively be unusable. The FAA does not shut down an airport, by the way, but it can render the airfield unusable when the weather drops below legal minimums for whatever runway is in use.

2. The pilots, in conjunction with Southwest's dispatcher (who shares legal authority with the flight captain), apparently communicated several times regarding the ability to land on Runway 31C, given the difficult conditions reported by the airport. The question is whether those conditions were reported correctly and, if not, were any significant deviations the result of human error, equipment failure, or some other systemic problem (such as delayed reports).

3. The pilots listened several times to the ATIS -- the automated radio signal broadcasting current conditions at Midway -- and used the reported figures in an onboard laptop system authorized by Southwest to determine when a landing would be both safe and legal. In aviation terms, legal means in compliance with regulations and procedures.

If the crew's conclusions were wrong, did it result from operator error (i.e., someone entering the wrong data), untimely or bad information from the ATIS, a computer program glitch, or some very human failure such as misreading the result? The NTSB will need to pay close attention to why this process designed to enhance safety apparently indicated the landing would be safe and legal.

4. The NTSB reports the crew elected to use maximum autobrakes, which would seem to be a very good and conservative decision. The report found the autobrakes began delivering maximum braking as soon as the main wheels were on the ground and, in addition, the report indicates that the aircraft touched down inside or very close to the appropriate landing "zone" and had about 4,500 feet in which to decelerate and stop. Yet, post-flight NTSB analysis shows that with the tail wind, snow conditions and slipperiness -- as well as an apparent delay in deploying the thrust reverse system on the two engines -- the aircraft would have needed more than 5,300 feet to stop. Here we have a difficult problem.

For many decades, thrust reversal was never taken into account when computing how much runway a pilot needed to legally and safely stop. Yet the NTSB says its post-crash conclusion that 5,300 feet of runway was needed is based on partial usage of thrust reversers. That would mean the only safe landing of Flight 1248 required full and rapid use of thrust reversers. This is a troublesome inconsistency.

5. Perhaps the most worrisome and puzzling indication from this report is that the aircraft was on the runway for 18 seconds before the thrust reversers deployed, and while the captain could not deploy them, the first officer was able to do so. Even if the landing was only legal with the use of thrust reversers, they are normally out and functioning within four to seven seconds of touchdown -- at least 10 seconds sooner than what happened on Flight 1248.

What the report does not mention, however, is that the aircraft itself has to mechanically "know" that it is on the ground before an interlock will permit the reversers to open. That air-ground switch system is located on one of the main gear struts, and yet another NTSB focus will be whether anything was wrong, even momentarily, with that switch. A significant delay in sending an "on the ground" message to the thrust reversers could explain why the first officer succeeded. Then again, it might not.

So the information presently available about Flight 1248 raises many more questions, but it also helps focus the investigation. It is important to remember the NTSB's job is not to assign blame -- which is counterproductive to improving air safety -- but to discover every single contributing factor that could produce a future accident.

Every such contributing element has to be found and fixed, and it is clear in the case of Flight 1248 the complexity of accident means that just pointing a finger -- at the pilots or anyone else involved -- will accomplish nothing.

Copyright © 2005 ABC News Internet Ventures

[alf5071h]

Comments and questions.
hopharrigan some very valid points, well said.

Not a 737 driver, but is wheel speed used in any of the spoiler / reverser deployment logic? On a previous type it was with interesting results on slippery runways.

Is there enough elevator power to raise the squat switches out of contact if an inadvertent forward stick input were made? With / without spoilers deployed. This has happened on other types; humans often make inadvertent inputs in marginal conditions; surprise, stress, etc.

For more information on runway reports and the increase in landing distance required on contaminated runway see the thread Good - Fair - Poor – Nil.
Note the statement “it is not practicable at the present time to determine aeroplane performance on the basis of an internationally accepted friction index measured by ground friction devices”. i.e. “There is no overall accepted certification / operational correlation between mu meters and airplanes”

Also note “On a smooth, wet concrete runway surface close to the minimum maintenance standard, the Falcon 20 tire-to-ground effective braking coefficient was found to be less than the aircraft certification requirement for a fully modulating anti-skid system. Based on this finding, the current operational dispatch factor of 1.92 for turbojet aircraft landing on wet runways at destination or alternate airports would have to be increased to a value of 2.2 to 2.4 in order to achieve the same level of safety as that which is currently accepted for dry runway operations. (Transport Canada Publication No. TP 14273E).

JAR-OPS operators are required to use manufacturers ‘certificated’ AFM data on contaminated runways; no tailwind landings. Is the US still operating on advisory data; are there differences between Boeing and Airbus operators in the US? Presumably, European standard landing data is available for both types, but do US operators request it?

Beware the human factors involved with presenting ‘advisory’ data on computers. Human nature is more biased towards a technology based solution, which may inhibit careful thought about the basis of the data and its implications.
--------------------
Unless specifically authorized everything else is forbidden.

[Chiller]

A notable omission in the above NTSB report is the mention of the activatin of auto-spoilers; probably the most critical element, after brakes, in getting the bird stopped.

[RRAAMJET]

DEFPOTEC - what I should have stated is that legally there is no credit on calculation; ie, you can have one locked out and still be legal on a slippery runway. They do not have to be operative in your landing dist. calcs.
I believe that's still the case.

[just.n.av8r]

rrAAmjet, FYI:
In my company's operation both t/r's must be operative on 737NG for landing with braking action less than good. There is also a penalty for landing distance calculation with inop t/r.

[bubbers44]

It looks like the reversers were the primary method of stopping for all of the aircraft that night. I was the first flight into Reno one morning and realized after coming out of reverse if I didn't go back into reverse on a 737 I would be off the end. I was empty. If they couldn't get the reversers out for 18 seconds, they had little chance of stopping. Midway had shut the airport down due to poor breaking action and plowed the runway prior to this crash and got a satisfactory braking action from a truck but reports were there were hundreds of feet of glare ice at the end of 31C.
FAA has pulled FDR info out of previous arrivals checking their deceleration rate on landing.

Something to think about considering a close to V1 abort on ice with one engine out.

[Ignition Override]

To repeat what someone else highlighted earlier, just when will "our friends" at the FAA use a NOTAM format which will split various groups of info into more readable paragraph? For how many decades has this garbage been printed as such?

Can this not be justified by another dead child? Or does the FAA's "cost/benefit analysis" require even more, no matter what the primary cause is?