MD-80 skidded off runway at LGA

[elaw]

I think in many cases that's true and helpful. But you're still going to have that case once in a blue moon where all external signs point to things being a certain way, but somehow in someone's mental image it still ends up the other way. And once in many blue moons, you'll get two or more people in the cockpit having that problem simultaneously.

Witness the AF90 crash where during the takeoff checklist one pilot said "Engine anti-ice?", the other said "Off" and neither caught the problem. That's in spite of the fact one pilot checked the setting, saw it was wrong, verbalized that fact, and both heard it.

[MCM]

Brian,

On many modern aircraft there's not much point reaching out and touching because system condition is based on ECAM/EICAS response, not on the switch position. The presence/absence of a message, or a light, or a display on a screen is what determines the checklist response rather than where the switch is. Looking at the switch is actually a really tough habit to get out of, and one instructors love to tell you off for.

One way to look at it is that the switch is a request button - the FMA/other indicators tell you if that request has been granted.

That's in spite of the fact one pilot checked the setting, saw it was wrong, verbalized that fact, and both heard it.

Thats an interesting point. Its clear, really, that the pilots didn't 'hear' it. It may have been audible, but it wasn't processed by the brain. We all know the concept - similar to when you meet someone for the first time, and then 20 seconds later say 'what was their name again?'.

In a slightly different way, hearing is one of the first things to go in an overload scenario, which is why a pilot flying will often not respond to the guidance of the PM when things are getting a bit pear-shaped. Its not that he's wilfully ignoring him, but that he is overloaded and simply does not hear him. 'I didn't hear you say that' is a common phrase after a very high work load approach.

Its all good and well to have a procedure to do - for example, check that the speed brake is up. But when we get overloaded we will do the task (looking), but won't necessarily process what we actually see.

A very common one is a landing clearance - if it comes at the wrong time (high workload in the cockpit) the brain will perform the automated task (read back the clearance), but then 20 seconds later neither pilot will remember if they'd received a landing clearance or not! The brain has done its job, and performed its automated task, but the conscious memory doesn't recall it.

If you're ever particularly bored listen to some Tower ATC tapes, and listen for how many times pilots ask for the landing clearance to be repeated despite being read back perfectly the first time. Its not frequent, but its certainly not rare.

Its really easy to say 'I would always check it'... all pilots would. But in the heat of the battle, when there's a dozen distractions, all of which could be as critical, the 'check' isn't as robust as it should be. And thats not through lack of intent, but the way the brain functions in load shedding.

[Evan]

Its really easy to say 'I would always check it'... all pilots would. But in the heat of the battle, when there's a dozen distractions, all of which could be as critical, the 'check' isn't as robust as it should be. And thats not through lack of intent, but the way the brain functions in load shedding.

What are the distractions the PM faces after the MLG has touched down that would make it difficult to focus on the the spoiler deployment?

It seems to me that forgetting to do this is like forgetting to rotate on takeoff, it's just such a basic, critical and simple task that should be a core instinct.

[ATLcrew]

I know that if I were a pilot I would always, religiously, check that lever upon touchdown.

No, you wouldn't. Not "always, religiously". I'm not calling you a liar (tempting as it is), I'm only saying you're incorrect in this particular instance.

[guamainiac]

Even, even you to would have a number of "failed to perform" checks (per say .. 1,000), against your scorecard.

Just like taking the SAT's you may score perfect one time but the chances of repeating that task get slimmer over time.

[3WE]

What are the distractions the PM faces after the MLG has touched down that would make it difficult to focus on the the spoiler deployment?

It seems to me that forgetting to do this is like forgetting to rotate on takeoff, it's just such a basic, critical and simple task that should be a core instinct.

Forgetting to rotate gives you numerous, immediate, extremely-obvious, direct and indirect, life vs. death feedbacks right in front of your eyes pretty much 100% of the time. If you want to live, you better rotate.

Forgetting to double check that the spoilers auto-deployed, and that you pulled the lever all the way back and that it's interlocked with auto-brakes gives you relatively little, not all that critical feed back, well behind your eyes nearly 100% of the time... And, if the spoilers don't pop, so what there's several thousand feet of runway, still a fair bit of weight on the wheels and reverse thrust is supplementing things nicely on 99% of your landings anyway.

...along with the fact that the pilots are almost always expected to provide the rotation but almost never expected to provide the spoiler activation.

[3WE]

A very common one is a landing clearance - if it comes at the wrong time (high workload in the cockpit) the brain will perform the automated task (read back the clearance), but then 20 seconds later neither pilot will remember if they'd received a landing clearance or not! The brain has done its job, and performed its automated task, but the conscious memory doesn't recall it.

If you're ever particularly bored listen to some Tower ATC tapes, and listen for how many times pilots ask for the landing clearance to be repeated despite being read back perfectly the first time. Its not frequent, but its certainly not rare.

...and a reflection of the broad, fundamental rule, Aviate, Navigate, Communicate.

[Evan]

Final report out. If you had 'rudder blanking', step up receive your plaque.

http://avherald.com/h?article=482b659f/0004&opt=0

[Gabriel]

The accident that wasn't:

There was a plane already cleared to land behind the MD-80. The tower could not see the MD-80 because of the weather conditions. When the tower failed to contact the MD-80, they assumed that they had already switched to the ground frequency, and didn't cancel the landing clearance of the other plane.

We have 2 issues here:
- Landing clearance issued before the preceding plane lands and vacates the landing. In most of the world (but not in USA) that is forbidden. The tower cannot issue a landing clearance until the preceding airplane vacated the runway. That's what "expect a late landing clearance" is for. Especially if the tower cannot see the runway!!!
- The tower assuming....

We had 2 slices still available, that prevented the accident:
- Whiteness that saw the accident and advised the tower, who then finally instructed the other plane to go-around.
- The MD-80 stopped far enough to the side that a plane keeping more or less the center line would not have crashed against the MD-80.

[Evan]

The accident that wasn't:

There was a plane already cleared to land behind the MD-80. The tower could not see the MD-80 because of the weather conditions. When the tower failed to contact the MD-80, they assumed that they had already switched to the ground frequency, and didn't cancel the landing clearance of the other plane.

We have 2 issues here:
- Landing clearance issued before the preceding plane lands and vacates the landing. In most of the world (but not in USA) that is forbidden. The tower cannot issue a landing clearance until the preceding airplane vacated the runway. That's what "expect a late landing clearance" is for. Especially if the tower cannot see the runway!!!
- The tower assuming....

We had 2 slices still available, that prevented the accident:
- Whiteness that saw the accident and advised the tower, who then finally instructed the other plane to go-around.
- The MD-80 stopped far enough to the side that a plane keeping more or less the center line would not have crashed against the MD-80.

I think maybe we've got our wires crossed. Did you read the final report I linked to?

Vnav nailed it on page 1. Maybe I get an honorable mention for speculating a ground spoiler deployment failure, which did happen here, but the F/O caught it immediately and manually deployed them. Pilot Material.

[Gabriel]

I think maybe we've got our wires crossed. Did you read the final report I linked to?

Vnav nailed it on page 1. Maybe I get an honorable mention for speculating a ground spoiler deployment failure, which did happen here, but the F/O caught it immediately and manually deployed them. Pilot Material.

Evan, we are talking about the same event. You are talking about what happened in the cockpit which, together with other factors, led to the accident.
I am talking about what happened in the control tower, which could have led (but didn't, partially thanks to luck) to another accident in the same event, with the airplane following the MD-80 landing and crashing against the MD-80 on / next to the runway.

[3WE]

...Blah blah blah landing clearance blah blah blah...

Extremely valid, albeit a shred off topic.

And an excellent example of blind procedure having a detrimental effect.

Cleared to land does not mean clear runway to land on... we have defined procedures, and official traffic reports in many instances and read backs and even probably have unwritten SA where the landing plane ALMOST always monitors the runway status...

...except when they don't or can't.

In IMC and night, the tower is supposed to monitor ground radar.

But, 'we' have run over a metroliner in CA, ran over a conquest at Flyover, and who knows how many more not-especially-ideal cases like this one.

Could TCAS assist and 'we' make sure that pilots ALWAYS check for a clear runway at some milestone like aMM / 250 feet / whatever?

PS-I LOVED MCM's comments at how often the crew of two does not process their landing clearance (a few hours with a scanner at a busy airport is all it takes to hear one first hand)... and it is a super busy time with lots of stuff to do like monitoring the ILS and double checking height vs distance, arming spoilers, remembering the wet vs dry reverse EPR, dealing with thunderstorm gusts, New Yark ATC, turning off warnings, knowing if you have one vs two autopilots to do the go around, dealing with underslung engines, and the one time Evan will let you use the rudder pedals...

[Evan]

Evan, we are talking about the same event. You are talking about what happened in the cockpit which, together with other factors, led to the accident.
I am talking about what happened in the control tower, which could have led (but didn't, partially thanks to luck) to another accident in the same event, with the airplane following the MD-80 landing and crashing against the MD-80 on / next to the runway.

Ah, now I see it...

About 14 seconds later (and 33 seconds after the initial notification that the runway was closed), the controller instructed the flight crew of Delta flight 1999 (the next arriving airplane for runway 13) to go around.

The danger wasn't coming from Delta 1086 at that point, as it was well off the runway, but from the snow coordinators vehicle's presence on the runway (as he had already reported to ATC twice that the runway was closed).

However, this assumption led to a situation in which the snow coordinator’s vehicle was on the runway for about 23 seconds while Delta flight 1999 was on final approach. At the time of the controller’s go-around instruction, the flight 1999 airplane was only about 30 seconds from landing.

[Gabriel]

Ah, now I see it...

About 14 seconds later (and 33 seconds after the initial notification that the runway was closed), the controller instructed the flight crew of Delta flight 1999 (the next arriving airplane for runway 13) to go around.

The danger wasn't coming from Delta 1086 at that point, as it was well off the runway, but from the snow coordinators vehicle's presence on the runway (as he had already reported to ATC twice that the runway was closed).

However, this assumption led to a situation in which the snow coordinator’s vehicle was on the runway for about 23 seconds while Delta flight 1999 was on final approach. At the time of the controller’s go-around instruction, the flight 1999 airplane was only about 30 seconds from landing.

Yes, in this specific case your description is accurate.
But in a more generic case... Can't you envision a weakness in the system?
A plane landing in IMC can be disabled on the runway and the plane trailing 2 minutes behind can be already cleared to land, in IMC, and land on top of the first one.

In this accident the plane was already cleared to land. It required an active instruction from the tower for the plane to go around. If, as in most of the world, a plane was never cleared to land until the preceding airplane is confirmed to have cleared the runway, then the plane will (or at least shall) initiate a go around on it's own if they don't get a landing clearance in time.

[3WE]

...Can't you envision a weakness in the system?
A plane landing in IMC can be disabled on the runway and the plane trailing 2 minutes behind can be already cleared to land, in IMC, and land on top of the first one...

Evan only sees procedural violations devoid of humanity.

...it would seem some sort of confirmation procedure by the tower WAS skipped... by the same ole mechanism... the plane 'ALWAYS' exits and sometimes forgets to check in.

I guess I want the pilots to have their own mechanism to 'see' the runway and confirm that the runway is clear for landing.

ATL and Bobby may scoff... and I doubt the officials will give this thread much attention, but I will at least waste a few bytes on the server...

[Evan]

Yes, in this specific case your description is accurate.
But in a more generic case... Can't you envision a weakness in the system?
A plane landing in IMC can be disabled on the runway and the plane trailing 2 minutes behind can be already cleared to land, in IMC, and land on top of the first one.

In this accident the plane was already cleared to land. It required an active instruction from the tower for the plane to go around. If, as in most of the world, a plane was never cleared to land until the preceding airplane is confirmed to have cleared the runway, then the plane will (or at least shall) initiate a go around on it's own if they don't get a landing clearance in time.

The scenario you describe is certainly not defended by layers of procedure, and yes, I would like it to be so. But this accident was very unique in one aspect. When the nose of the plane breached the wall, it destroyed ALL battery power (APU not running), so ALL comms were lost. In just about any scenario that doesn't involve a rather noticeable fireball or thunderous crunching sound (such as a minor runway excursion), the crew is going to report their situation immediately to the tower. So this was a fluke, but it happened, and, while extremely unlikely, it could happen again.

I just think you have to consider the very unlikely prospect of a total (AC, DC and BAT) power loss after a minor runway excursion.

[Gabriel]

The scenario you describe is certainly not defended by layers of procedure, and yes, I would like it to be so. But this accident was very unique in one aspect. When the nose of the plane breached the wall, it destroyed ALL battery power (APU not running), so ALL comms were lost. In just about any scenario that doesn't involve a rather noticeable fireball or thunderous crunching sound (such as a minor runway excursion), the crew is going to report their situation immediately to the tower. So this was a fluke, but it happened, and, while extremely unlikely, it could happen again.

I just think you have to consider the very unlikely prospect of a total (AC, DC and BAT) power loss after a minor runway excursion.

I didn't say minor runway excursion. There are a number of ways a plane can be disabled on the runway and the pilots unable to communicate with the tower (including them being hurt, dead, or giving priorities to other matters like evacuating the plane filled with smoke).

Do you see any cons to withholding the landing clearance until the runway is clear and there are no other planes scheduled to occupy it? I only see pros.

[3WE]

Do you see any cons to withholding the landing clearance until the runway is clear and there are no other planes scheduled to occupy it? I only see pros.

Yes, I see a big con: Last minute clearance is a distraction during the criticalest of the criticalest times.

Instead of a last second black and white clearance, the pilots just need to know THAT the runway is clear and maybe have more data to truly track the situation, as opposed to a black and white radio call with legal ramifications.

Making any sense?

[Evan]

I didn't say minor runway excursion. There are a number of ways a plane can be disabled on the runway and the pilots unable to communicate with the tower (including them being hurt, dead, or giving priorities to other matters like evacuating the plane filled with smoke).

Sure, but what are the odds of those things happening on an aircraft disabled on (or partially off) the runway, with no comms, and no obvious fire or indications of a crash to tip off operations? And if comms are working, I think in poor visibility a quick ( two-second) request for emergency services should always be part of the pre-evacuation procedure.

Do you see any cons to withholding the landing clearance until the runway is clear and there are no other planes scheduled to occupy it? I only see pros.

No, not offhand. As long as it comes before DH/MDA. You want go-arounds from a safe altitude. I've been on a flight that went around due to separation issues just aeround that point. No big deal.

[Gabriel]

Sure, but what are the odds of those things happening on an aircraft disabled on (or partially off) the runway, with no comms, and no obvious fire or indications of a crash to tip off operations?

If there is no fire (or in certain conditions even if there is), what are the obvious indications of a crash in very poor visibility?

No, not offhand. As long as it comes before DH/MDA. You want go-arounds from a safe altitude.

That makes no sense for various reasons:
1- DH/MDA can be anything from 0 ft to 400 ft depending on the type of approach.
2- There is no unsafe altitude for a go-around. Last minute wind gust, too long flare... you go around.
3- It is perfectly routine that airplanes expect the preceding airplane to clear the runway just on time for their landing, and that they go around at almost the last second if it didn't. The go-around can be initiated by the crew (provided that they have enough visibility) or by the tower (provided that they have enough visibility or that they receive a communication that the plane is disabled on the runway or that they proactively act when the landed plane did not call clearing the runway).

This happens quite often today. The only difference is that if the tower fails to cancel the landing, if the visibility is poor the next plane can land on an occupied runway. While in most of the world the crew would be required to go-around because they would have not received the landing clearance in the first place.