Blogs and Stories
Behind the Buffalo Crash
David Duprey / AP Photo
It’s too early to tell what caused last night’s crash of a Continental Airlines plane near Buffalo. But Clive Irving points to some troubling instances in which other turboprop planes failed in icy weather.
There is something hauntingly familiar about the crash of the Bombardier Dash 8 commuter plane in Buffalo. While a lot more information is needed before investigators can be sure of the cause, there are precursors that suggest a pattern.
It begins with the crash of American Eagle Flight 4184 at Roselawn, Indiana, in October 1994. It was a different type of plane, a European-built ATR-72 (the Bombardier is a Canadian plane). All 49 people on board died. The cause was ice on the wings.
The ATR-72 had been in a holding pattern, flying through sleet that caused a build-up of ice on the wings. When this happens at a critical point the ice effectively destroys the power of the wings to provide lift. Flight 4184 lost lift, rolled over, and crashed.
What urgently needs to be looked at is whether the basic configuration of all turboprop commuter planes, and particularly their flying characteristics in ice-making weather, leaves them unusually exposed to ice build-up on the wings and control surfaces.
This has been a hazard since the first days when planes encountered ice. And usually it can be disregarded because all airliners have what are called de-icing boots—vulnerable parts of the wings are heated to prevent ice forming.
The Roselawn crash led to investigations that revealed that the ATR-72 had a history of being vulnerable to ice formation. ATR-72s were later modified to correct the problem, and flight crews were alerted to the dangers of flying in weather that could cause rapid ice build-ups.
In 1997, a Comair commuter flight, this time involving the Brazilian-built EMB-120, crashed near Detroit, killing 3 crew and 26 passengers. Again, ice was the cause and the National Transportation Safety Board discovered that the plane maker, Embraer, had warned the airline that the EMB-120 could be uncontrollable if flown too slowly in icy conditions. The airline had not passed this warning to the crew.
As recently as January 27 this year, a FedEx ATR crashed on landing at Lubbock, Texas, in a freezing mist. Officials are still disputing the role of ice in that incident. The crew was able to walk away, but the crash has every appearance of ice build-up affecting the controllability of the plane at a critical moment.
So here there are at least three different types of commuter planes—European, Brazilian and Canadian—involved in crashes where ice was almost certainly a factor. What do they have in common?
They are all turboprops—each with two engines in which a gas turbine, as in a jet, drives propellers. When turboprops descend for a landing, as the Buffalo flight was, they fly more slowly than a jet, which means that they spend longer at the altitude where sudden, unpredictable bursts of ice can appear.
The Bombardier Dash 8 Q400 is state of the art as turboprops go, with extremely sophisticated equipment and, until now, a good safety record. However, what urgently needs to be looked at is whether the basic configuration of all turboprop commuter planes, and particularly their flying characteristics in ice-making weather, leaves them unusually exposed to ice build-up on the wings and control surfaces. In other words, whether those heated “boots” can be so suddenly overwhelmed by sudden flashes of rapid ice formation that the pilots have no chance of retaining control.
This comes at a time, incidentally, when these turboprops, which are more fuel efficient for short commuter flights than jets, are experiencing a big boost in demand from cash-strapped airlines.
Clive Irving is Senior Consulting Editor at Conde Nast Traveler, specializing in aviation. He has flown the A320 in a simulator and earned an “Honorary Captain” certificate for making a perfect approach and landing at JFK—thankfully, he says, not for real.







vankuyk
Strange, they say plane crashes come in three's. I hope not to be on number three. Better stop flying for a while.
workonit
Most of us will die of heart disease, stroke or cancer, but no one is as afraid of the causes of those diseases as of flying. We should run screaming from cheesburgers and fries if we were rational.
steveissak
The issue of icing on turbo-prop planes is not a small one - particularly the Bombardier and other high wing models. Even with operable de-icing boots, the slightest buildup of ice on the leading edges throws this type of aircraft out of balance and due to the high wing configuration - to the point that the horizontal stabilizer is unable to compensate, thus causing the aircraft to go nose down.
randall234
these "de-icing boots" are not exactly what Mr Irving means to say. A "de-icing boot" is an inflatable rubber balloon like device placed on the leading edges of wings. When ice forms the boot is expanded to break the ice and cause it to fall away. Airlines and complex aircraft, like our dear turbo-prop, use a different system. Heated air taken from the turbine engine is run through a pipe on the leading edge of the wing allowing no ice to form. If the turbo-prop did indeed have a boot device what can happen if the boot is inflated too early ice can form around the boot and take it's shape, thus destroying the lift capability of the wing
otooleco
Low and slow on approach. It will be interesting to see if they find evidence of tailplane icing. More than one experienced pilot has lost it on approach when flaps were extended with excessive ice buildup on the horizontal stabilizer.
Yailbloor
The boots aren't heated as you state. They are actuated by air to infalte, which causes the ice to break away from the leading edges of the control surfaces. A sad day when an hour or so in a sim makes one an expert on flying and icing in the journalistic community, this is pure sensationalist, self serving rubbish
Fentro
If Man were meant to fly, he'd have jets.
gus140160
For it to have iced up and gone verticle...i'm thinking tail stall, ice built up on the tail and during approach, flap settings mess up the configuration of the plane and it just falls outta the sky....the usual procedure to break a stall have to be done backwards to get out of a tail stall, and since they'd be doing the normal procedure...they'd be just increasing the stall. Maybe the boots were inop?
JWGotsch
I have done a lot of short flights on these turbo-props: Philly to Erie, Toronto, Albany, Montreal; St Louis to Chicago, Springfield, MO; Chicago to Lacrosse, Minneapolis. In all kinds of weather including ice, snow and tornadoes.
This analysis doesn't fixate me. Sorry.
spinozareader
To steveissak--Thank you for your insight.
To workonit--I concede that most of us eat some stuff that's not good for us. But THAT shouldn't detract from the possibility that there seem to have been some foreseeable icing issues surrounding this flight that might've rendered it inadvisable? If someone you cherished had been on this flight, wouldn't you want the circumstances of its crashing to be investigated fully? Especially if it's crash could've been predicted??
allen-o
This article is BS. I participated in an analysis of the Roselawn flight, including full access to the flight data recorder, flight voice recorder, and weather radar records. Yes, there were severe icing conditions. But, the pilots were simply not paying attention. They chatted up a flight attendant for over a half the time that they were in a holding pattern, an hour-long pattern that had them repeatedly passing through freezing rain. Also, at the request of the other flight attendant, they had put the flaps down while holding, which was not a recommended operating condition, and which altered the airflow such that ice formed behind the leading edge de-icing boots. In my experience analyzing this, and many other aviation accidents, there is rarely any one thing that brings down a plane. It is generally a series of events, usually involving human error.
vankuyk
What did I tell you? Well crash no 3 just happened in London, I guess we can start flying again.
Thank you.
As a first time user, your comment has been submitted for review. It can take anywhere from a few hours to a day or two for your comment to be reviewed, depending on the time of week and the volume of comments we receive.