A complete meteorological analysis of the loss of AF447

June 2, 2009 12:01 PM

Tim Vasquez, former U.S. Air Force meteorologist, author, software engineer, and head honcho of the storm chaser hangout Stormtrack Forums has done a complete meteorological analysis of the weather conditions which may have resulted in the loss of Air France Flight 447.
posted by spock (26 comments total)

I mentioned this in the existing thread on AF 447.
posted by exogenous at 12:15 PM on June 2, 2009

Sorry if I came off harshing on your post, spock. Meteorology is certainly facinating and worthy on it's own.
posted by exogenous at 12:31 PM on June 2, 2009

Very cool analysis (not like I understood any of it!). Had to look up MCS (Mesoscale convective system).
posted by i_am_a_Jedi at 12:35 PM on June 2, 2009

Good stuff here. Thanks for posting it spock.
posted by zerobyproxy at 12:47 PM on June 2, 2009

it's own

its own

I've been learning more recently about skew-T diagrams, which have tremendous amounts of information but don't really make sense on first glance. They are based on atmospheric soundings of wind, temperature, and dewpoint from weather balloons. A good tutorial on them here (free registration required), and others here and here. Actual or forecast model soundings can be found here.
posted by exogenous at 12:53 PM on June 2, 2009

I didn't understand this comment in the last link of the FPP:

G day Tim

I am an Air France Captain, found your post very interesting and if true, will be hugely damaging for Air France.
Best regards

B-----

Any reason why this would be damaging for AF?
posted by KokuRyu at 1:05 PM on June 2, 2009

Those with an interest in learning more about weather can go to the NWS JetStream online school or (the one all storm chasers will point you to and tell you to read in its entirety) theweatherprediction.com including Haby's Hints (parts 1 and 2).
posted by spock at 1:10 PM on June 2, 2009 [2 favorites]

Good information.

There was a couple of incidents on Qantas with A330s where the planes became convinced that they were stalling, and dropped the nose. At cruise, this is disconcerting, but with 30,000 or more of altitude before you hit the ground, there was plenty of time to recover.

There's are lots of important velocities in piloting, most of them referred by the letters V_something. One of the most important is V_ne, which is Velocity, Never Exceed. The reason you never want to go faster than V_ne is that the manufacture thinks that if you do, drag will rip things off the plane. In most of the high-altitude disasters, like TWA800 and Challenger, it isn't an explosion that tears the craft apart, it's drag, caused by airspeeds well over V_ne.

So: For some reason, the nose goes down. They're at cruise, so they're moving fast -- .8 to .85 Mach. They descend, picking up speed, then hit a mighty updraft, which, in terms of the airframe, add to the already high airspeed -- which goes beyond V_ne, something fails to stay attached (usually, the wings) and they fall out of the sky.

So: Thunderstorm windshear, combined with the Coffin Corner. The Coffin Corner is named after the shape on the performance charts. As you fly higher, stall speed increases, and it increases faster than V_ne. So, at the Corner, you find that Stall is very close to V_ne. Go faster, your wings fall off. Go slower, you fall out of the sky for lack of lift. The U2 spent a great deal of time here, and a number of them were lost. Airliners stay the hell away from the Corner.

So, they're at cruise, high and fast -- and V_ne and Stall aren't that far away. They hit a burst of tail wind, which drops the airspeed quickly, and they head rapidly towards stall speed. The plane dives and powers up to get away from stall, and they hit the shear line, the winds become headwinds, adding to the effective airspeed, and between the dive, the extra power, and the sudden headwind, they go over V_ne and the plane breaks.

Windshear kills on landing, but usually, in the other way -- you get the headwind first, which adds to your airspeed, which makes you climb. You're trying to land, you find the plane going high, you pull back on power and lower the nose to hold the glideslope, hit the shear line, and find yourself with engines at low power and a sudden tailwind stealing your airspeed, you stall and crash.

I would not be surprised if the black box shows a sudden drop in airspeed, a pitch down, engine power up, then a huge spike in airspeed -- right up to and past V_ne.
posted by eriko at 1:13 PM on June 2, 2009 [38 favorites]

The Professional Pilots Rumor Network thread on the incident (picking up on Page 4 where a link to this meteorological analysis is given) might provide some insights into why this would be damaging to Air France. (Off to read it myself)
posted by spock at 1:16 PM on June 2, 2009

Just to give credit where credit is due (heh) I'd like to point out to anyone having trouble reaching the weathergraphics.com site that it is hosted on a Dreamhost server.
posted by spock at 1:26 PM on June 2, 2009

Better thread from the PPRN.
posted by spock at 1:59 PM on June 2, 2009

It certainly looks like a plausible and well thought out scenario for this particular tragedy and possibly preventative information for pilots flying that part of the planet.

it emphasizes that the aircraft was certainly within the bulk of an extensive cumulonimbus cloud field for a significant amount of time and that storms could indeed have been a contributing factor to the crash...significant turbulence and thunderstorm activity for about 75 miles (125 km), lasting about 12 minutes of flight time

Sounds like it must have been a really bumpy ride before the crash.

YAY creative and communicative scientists. Inspiring work by Tim Vasquez. Thanks for the post, spock.

And my condolences to the friends and families of those who died.
posted by nickyskye at 3:04 PM on June 2, 2009 [1 favorite]

Why are people so fascinated by this plane crash or plane crashes in general?
A few people died. People I did not know. So what? No people speculate what could have happened, what tremendous work we could do to prevent such accidents in the future etc.
Every day thousands of people die of hunger and/or disease that could have been prevented for a few dimes but nobody gives a f.
This "plane crash porn" starts annoying me. Maybe Taleb is right and I shouldn't read the news anymore.
posted by yoyo_nyc at 3:53 PM on June 2, 2009 [2 favorites]

Why are people so fascinated by this plane crash or plane crashes in general?

This crash is interesting as it is such a mystery.

Every day thousands of people die of hunger and/or disease that could have been prevented for a few dimes but nobody gives a f.

Oh God.....
posted by caddis at 5:04 PM on June 2, 2009

yoyo_nyc: Your post confuses me on several levels. Given you view the deaths of people in airliners is no big deal (even this, the largest airline fatality in a decade), I don't understand how in the next sentence preventing such accidents could be considered "tremendous work".

Secondly, did you look at the link in the post? By doing a thorough analysis of the meteorological conditions that plane was in, Vasquez is attempting to speculate on what could have happened. By learning from it, perhaps pilots/airlines perhaps can learn something to prevent this from happening in the future.

Thirdly, you are assigning the worst possible motives on those who are interested in the story. For some, there may be a "rubbernecking" quality to their interest in the news... but I think most of us feel a psychological kinship to these people because most of us have (and will again) fly in an airliner ourselves... or we have friends/family/coworkers that do (perhaps often). Therefore, we do feel empathy for the victims and, perhaps more importantly their mourning and suffering family and friends.

To say "A few people died. People I did not know. So what?" does not succeed in putting you on some morally round higher than those who are interested in this story.
posted by spock at 5:31 PM on June 2, 2009 [2 favorites]

^ last line correction: moral ground higher
posted by spock at 5:34 PM on June 2, 2009

A few people died. People I did not know. So what?

does not work with

Every day thousands of people die of hunger and/or disease that could have been prevented for a few dimes but nobody gives a f.
posted by Nauip at 6:21 PM on June 2, 2009 [3 favorites]

The reason you never want to go faster than Vne is that the manufacture thinks that if you do, drag will rip things off the plane.

My layman's understanding is that Vne is often limited by flutter rather than by drag loads— the limiting factor is how rigid you can make the plane, or its skin, not how strong. Of course it depends on the plane.
posted by hattifattener at 7:49 PM on June 2, 2009

eriko's comment was great, but didn't he mean 'tailwinds' when he said 'headwinds' and vice versa?
posted by lukemeister at 10:29 PM on June 2, 2009

but didn't he mean 'tailwinds' when he said 'headwinds' and vice versa?

Nope, because we're talking air speed, not ground speed.

A tailwind increases your ground speed, because the plane is moving at X speed with respect to the air it is flying in, but that air is moving at Y, so your speed, relative to the ground, is X+Y.

A headwind decreases your ground speed, you're still flying at X, relative to the air you are flying in, but that air is moving backwards, so your ground speed is X-Y.

This is why it typically takes 6.5 hours to fly ORD-LHR, but 8 to fly LHR-ORD.

Now, let's look at airspeeds. The plane is flying at X, relative to the ground, into a headwind of speed Y. What's the airspeed over the surfaces? The plane is moving into the air at X, the air is moving towards the plane at Y, thus, airspeed is X+Y. Thus, if you're flying at 100kts into a 30kt headwind and you stall at 110kts, you're going to keep flying. So: Ground Speed 100kts, Airspeed 130kts, Stall 110kts, Life Good.

With a tail wind, you're moving at X through the air (relative to the ground) , but the air is moving with you at Y. Thus, airspeed is X-Y. If you're flying at 100kts with a 30kts tail wind, your airspeed is 70kts. Ground Speed 100kts, Airspeed 70kts, Stall 110kts, Life Bad.

This is why planes take off and land into headwinds. Well, modern airliners on long runways at busy airports will occasionally find themselves dealing with a slight tailwind, but they have the power to do it and the long runway needed to stop after they touch down. In general, pilots care about airspeed, right up the moment they touch down. Then, it's groundspeed that tells them how long they have to stop -- landing with a big tailwind means you land fast, and you have that much more speed to get rid of before you run out of runway.

Pilots live by airspeed -- what matters to them is that there is enough air moving by the airfoils to fly, so when you're dealing with the "can it fly?" or "why didn't it fly?" questions, you need to work the problem with speeds relative to the airfoils. Navigators live by groundspeed -- you need to know when you're going to arrive at X, you can't just use airspeed, because you don't know how the air is moving. Without modern naviaids, figuring out ground speed at altitude was a very difficult problem.

So, in this storm -- a sudden strong tailwind reduces airspeed. Strong enough tailwind, and your airspeed drops to stall. To avoid that, you need to go faster, the fastest way to do so it to dive and increase power. You get the airspeed back up, and the wind spins around to a strong headwind at the wind shear line, and now, instead of X-Y being your airspeed, it become X+Y. If your dealing with 100kt high-altitude gusts, you can suddenly find yourself with an almost instant increase in airspeed of 200kts when the vector flips -- and if you are 100kts from V_ne when that happens, you're suddenly exceeded, by 100kts, the Velocity, Never Exceed.

And this can make for a very bad day.
posted by eriko at 6:22 AM on June 3, 2009 [2 favorites]

LUkemeister, no he didn't. For example, when he said "Headwind adds to your airspeed" he means something like: the ground speed of the plane is for a short time, constant. With a headwind, the lift surfaces experience an effective speed greater than the ground speed, so you have a surplus of lift.

Basically, even though headwinds scrub speed overall on the aircraft, from a lift standpoint, they *increase* effective speed.
posted by notsnot at 7:40 AM on June 3, 2009 [1 favorite]

Why in hell didn't that post for an hour and a half?
posted by notsnot at 7:40 AM on June 3, 2009

notsnot,

Heavy headwinds on the Internet ... or should that be tailwinds?
posted by lukemeister at 7:50 AM on June 3, 2009

eriko, thanks very much for the technical details, I barely understand them but they offer a fascinating insight into the intricacies of modern aviation!!
posted by supermedusa at 7:19 PM on June 3, 2009

Interesting page that elaborates on (so far as I can tell) and illustrates Eriko's suggestion.
posted by Rumple at 1:00 AM on June 4, 2009

"Debris 'not from Air France jet'"
posted by b1tr0t at 7:20 PM on June 4, 2009