Wind shear: a danger vanquished, or, one waiting in the wings?

“Lightning coming out of that one…”

The following conversation happened between 18:04:21 and 18:05:56, August 2, 1985, on the final approach course to Runway 17L at DFW airport in Texas. It was captured on the cockpit voice recorder of Delta Flight 191, a Lockheed L-1011.

Delta 191 crash site
Delta Flight 191 crashed after encountering wind shear on approach to DFW.

First Officer – Lightning coming out of that one

Captain – Where?

FO – Right ahead of us

Flight Engineer – You get the good legs don’t ya

FO – Garbled

Capt – I don’t have a DME on mine

FO – I don’t know, you haven’t had it for the last five minutes

Capt – A thousand feet

Capt – Seven sixty two in the baro

Capt – I’ll call ‘em for you

FO – Aw right

Capt – Watch your speed

Sound of rain begins

Capt – You’re going to lose it all of a sudden, there it is

Capt – Push it up, push it way up

Capt – Way up

FE – Way up

Capt – Way up

Sound of engines high rpm

Capt – That’s it

Capt – Hang on to the (expletive)

FO – What’s Vref

GPWS – Whoop whoop pull up

Capt – TOGA (takeoff/go around, ed.)

GPWS – Whoop whoop pull up

Unidentified – Push it way up

GPWS – Whoop whoop pull up

GPWS – Whoop whoop pull up

Sound of noise similar to landing, sound of takeoff warning horn, continues for 1.6 seconds

Unidentifed – (expletive)

Unidentied – Oh (expletive)

Second impact

Tower – Delta go around

By that time, Delta 191 was beyond going around. The airplane touched down 6,000 feet short of the runway just after emerging from a rain shaft. The airplane bounced across a highway, an engine hit a car and a wing hit a light pole. Then the airplane touched the ground again in a left wing low attitude and careened to the left, hitting two water towers (tanks really), and exploding.

The flight crew didn’t talk much about what was going on with the airplane when on final approach. Instead, they were trying to deal with it, and, “it” was a lot.

The airplane had penetrated a thunderstorm just as it spawned a strong downburst or microburst. A Learjet had flown the approach just a minute earlier and had no problem. Then things changed, drastically.

The gyrations and airspeed fluctuations were remarkable. For example, the input on the control column went from a 22 pound push force to a 25 pound pull force in about four seconds. Full lateral control was required to keep the wings level. Power varied from flight idle to full as the pilot flying tried to make the airplane track the ILS final approach course and glideslope while maintaining airspeed. At one point the airspeed dropped by 44 knots in ten seconds.

In addition to huge airspeed fluctuations caused by changing wind, there were strong up and downdrafts. When the airplane was but 200 feet above the ground, the rate of descent was about 5,000 feet per minute. It was speculated that once the airplane descended below 800 feet above the ground it was doomed.

Wind shear diagram
That increasing headwind won’t last long…

It was a classic case of being in the wrong place at the wrong time. I don’t think many pilots of that day would have figured it out in advance, either. The weather was typical of an August day with air mass thunderstorms around. It was not a condition that would dictate halting air commerce and indeed flight operations were pretty normal except for that one flight. There was no mention of severe weather.

The accident prompted a great flurry of hardware, software and educational activity and while there have been wind shear problems since, and will probably always be wind shear problems, we did go from having a lot of trouble with the phenomenon to having much less trouble. In other words the smoke and flames driven FAA activity seems to have helped.

My first in-depth exposure to this came when we published, in the January, 1962 issue of AIR FACTS, an article that was lifted from a TWA technical bulletin, prepared for the pilots of that airline. The title, “Wind Shear Effects on Airspeed,” pretty well summed up meteorologist J. A. Browne’s effort.

The airlines were transitioning from pistons to jets at the time and everyone knew that the power response of a jet would be slower. Also, there would be no big props blowing lots of air over the wings. In other words, wind shear would be more challenging to big jets than to big pistons.

What everyone didn’t know, or at least didn’t acknowledge at the time, is that wind can and does affect the airspeed of an airplane in flight, drastically in some situations. Many pilots didn’t, and some still don’t, think that wind can be a big factor in this regard. A steady wind can’t, but wind that changes in direction or velocity over altitude or distance can have a profound effect on airspeed. The spotty requirements for meteorological knowledge had, over the years, done little to educate pilots on this. Delta 191 changed all that.

I think that airline pilot understanding of this phenomenon has probably done more to eliminate wind shear accidents than have the hardware and software solutions. Once a pilot understands that an increasing tailwind or decreasing headwind can result in airspeed loss and a real sinking spell, then all else that is required is a basic understanding of the atmosphere to keep the old guard up. One airline advises pilots to execute a missed approach if an updraft or increase in airspeed is experienced on final approach because the situation will almost surely reverse.

Thunderstorms huff and puff and it doesn’t take a rocket scientist to see that, when flying through one, you’ll pass through up- and downdrafts and a change from a headwind to a tailwind as you move into the outflow from the storm. Delta 191 was an extreme and classic example of this. I think that the study of it prompted pilots to carefully consider any interaction between the airplane and the various wind flows around a thunderstorm.

When we ran the TWA paper in 1962, we noted that while this was written for pilot of heavy airplanes, it was also applicable to light airplanes. On one hand, our airplanes have less mass and far better acceleration characteristics; on the other hand we fly approaches at fewer knots above the stall than do heavier airplanes so a 44 knot wind shift will be working on a smaller number.

Wind shear can be a factor in more benign conditions, too.

When the Ercoupe came to the fleet in numbers after World War Two, a number of pilots of that airplane were victims of a mild version of wind shear. The airplane was stall resistant and spin proof, yet a number of them impacted the ground nose low and with great force. They didn’t spin in but if an Ercoupe was flown too slow and the headwind decreased rapidly when, for example, the airplane descended below a tree line around the field, the airplane would pitch nose-down to regain the lost airspeed and there would be no elevator authority until the nose-down attitude generated some airspeed. The ground often got there first.

When I first heard about Dr. Richard Rockefeller’s tragic accident on departure from HPN, White Plains, New York, one the first things that came to my mind related to wind shear and my experience with the hairy instrument departure off Runway 16 at that airport.

Because LaGuardia airport is not far to the south, with a high volume of airline traffic, they require HPN traffic off 16 to turn back to the northwest ASAP. The departure calls for a right turn when about 400 feet a.g.l. The initial assigned altitude is low, too, at 3,000 feet, so if you are flying a high performance turboprop Piper Meridian, as Rockefeller was, not much of that high performance will be used on this departure.

Take off, turn, begin working the power back, level off, and do it all in a short period of time. It is demanding. If there is an increasing tailwind as you climb and turn, as is often the case in weather like it was for Rockefeller’s departure, it is just one more factor working against success. The increasing tailwind can have a big effect on acceleration and climb. I have flown the departure many times and there is little margin. If the increasing tailwind is strong, it feels weird, as if something is not right.

I have no idea whether or not that had anything to do with the Rockefeller crash, and the NTSB’s preliminary report suggests that the airplane was not far enough into the departure for it to have been a factor, but that is an example of how understanding the wind aloft, and, the air you are about to fly into, can help avoid surprises.

GPS did wonders for pilots who want to understand everything about what is coming next on approaches. Say, for example, the ATIS shows a calm surface wind at the airport you are approaching, and as you come up on the final approach fix the GPS shows a 20 knot tailwind, that outlines something that will have to happen as the approach unfolds.

When a tailwind decreases as you descend, the airplane will trend toward high and fast. The airplane will have to decelerate, by 20 knots in this case, to keep the airspeed in the right place as the tailwind goes away. Knowing that is going to happen makes that easier to handle.

Also, if you have, say, ten degrees of drift correction on final and the surface wind is calm, at some point you will have to take that correction out to remain on the final approach course. That usually starts to happen when descending through about 500 to 600 feet a.g.l. and again, if you know it is coming, it is easier to handle. GPS makes this easy.

In studying overrun accidents, there is often a common thread: a decreasing tailwind during the descent on final. When there is an overrun, it is usually because of a long and hot touchdown or an excessive period of floating while excess airspeed is dissipated. A pilot who can anticipate that this is going to happen has the advantage – as long as he does something about it.

Jet core
Wind shear isn’t just a low altitude problem.

When dealing with big wind shear, as in a thunderstorm or at altitude when nibbling at the edges of a jet core, or jet streak as some call it, there is a lot of turbulence. The reports you hear of folks getting tossed about in the airliner cabins during cruise flight are directly related to wind shear.

Like so many other things related to meteorology, there are no absolutes. I flew my pressurized 210 a lot in the high teens and low twenties and got to play with the occasional jet streak. One day it got so rough I just said to heck with the big tailwind and went back lower for a smoother ride. Another day I punched into a strong jet streak that eventually reflected a tailwind of over 140 knots and after flying in it a while, I punched out the other side – all with no more than light turbulence. When I was losing the tailwind, I had to work to keep the airspeed in the green because it wanted to increase by a lot as the tailwind rather slowly went away.

The wind shear we see on an IFR day, when descending or climbing into a stronger head or tailwind usually does not result in turbulence of note but you can tell when the wind change is being experienced by some light bumps and, of course, the effect on airspeed.

The wind shear that is most obvious comes on a gusty day with the wind highly variable in both direction and speed. This results in airspeed and height fluctuations. It is short term but can be challenging.

Some airports are more challenging than others on this. Runway 34 at White Plains was always a special challenge to me. A look at the airport diagram suggest that it will be a piece of cake to land on 34 and make the first left turn off for a straight shot to Panorama, the friendliest FBO for light airplanes.

On gusty days, with a crosswind from the left, there was often some wild turbulence that made the airspeed jump around and caused a sinking spell when on short final. Where I had been thinking about landing and making that first turn off, my thoughts turned to getting the airplane onto the runway in one piece and still able to turn off the runway anywhere.

When I got my ATP (ATR then) in 1958, I hired a Trans-Texas Airways DC-3 captain to help me prepare for the check-ride because it would be given by an FAA air carrier inspector.

At that time there were not a lot of ILS approaches around but the TTA pilot passed on a word of wisdom about this. Most ILS approaches were to the northeast. I think there were more Runway 4 ILS approaches than any other. His word was that when flying one in actual weather the atmospheric conditions would almost always dictate a decreasing tailwind on final that would cause the airplane to trend high and fast on the approach.

Why did he tell me this? He told me because he said you don’t see that phenomenon when flying ILS approaches in simulated conditions in good weather and you have to see it to understand it. Then, we didn’t have GPS to give us an idea of how much wind shear there would be on the approach but forewarned was still forearmed. As a teacher he wanted me to learn from his real experience as opposed to simulated experience. He also said that pilots who only fly VFR probably never notice this because it is virtually unique to inclement weather. There can be strong wind shear in a cloudless frontal zone or in a dry microburst, however.

Wind shear is still out there. Let us hope we have learned enough lessons the hard way to avoid future smoke and flames FAA action.

13 Comments

  • I’m not so sure that light planes are horribly affected by windshear, like a heavy jet is.
    If the change in windspeed is relatively slow, it’s pretty much a non event. Even with rapidly changing wind speed, if one is reasonably attentive, a piston plane can react pretty quick.

    To prove this out, just go up in smooth air reduce or add throttle (as appropriate) and see how fast the airspeed changes. Then do the same in your jet… which is a fair amount slower.

    Not to minimize the risk, but it’s much more controllable in the piston plane.

    • Steve – It’s true that light aircraft can respond more quickly to a given windshear (velocity change) than a heavy jet, just due to its lower inertia as compared to the inertia of a heavy aircraft, and also due to the faster response of a prop vs. turbine engines.

      However, the same inertial effect also works AGAINST the light aircraft, because being lighter, light aircraft have less inertia to carry them through a sudden short term tailwind or downdraft that would have little effect on a high speed heavy jet.

      Don’t kid yourself – wind shear can most definitely have a huge impact on light aircraft.

      Don’t know how much mountain or canyon flying you’ve done, but windy days in rugged country creates horrific wind shears that routinely kill even experienced light aircraft pilots. I personally knew two extremely experienced mountain pilots, both of whom were lifetime professional pilots and decades long CFIs and mountain flying instructors, with tremendous experience in mountain flying. Both these guys were killed in stall-spin accidents in very light (LSA) aircraft on very windy days in or near canyons … in both accidents it is thought that wind shear was the principle culprit. Renowned mountain flying expert, instructor, and author Sparky Imeson also crashed and died a few years ago in a similar accident in similar windy conditions in the mountains. Ditto with Steve Fossett, also killed in windy mountain flying.

      Personally, I have experienced huge wind shears in or near mountains in the Rockies that took all I could do to recover from without crashing … in once instance I got caught in a mountain wave when in cruise flight 20 miles downwind of a crest and (fortunately) 3-4 thousand AGL, and for several minutes at full power and at Vx, I was still doing over 1,000 fpm descent until I finally got past the wave!

      Another time I was on final approach at about 500 AGL, with a headwind of probably 35-40 knots right down the runway leading to a mountain airstrip on the edge of a steep canyon face … as I approached within the final quarter mile to the runway and intentionally relatively high, suddenly I had to apply full power and down elevator (which might seem counterintuitive to some pilots!) to get back to Vx to avoid stalling out and crashing on the runway. Once past the canyon, the wind shear ended as suddenly as it started.

      (moral of the story – don’t fly in the mountains on windy days!)

      Thunderstorm cells are much worse than mountain winds, and are known to result in downdrafts in excess of 6,000 fpm … meaning if a light plane runs into one of those at traffic pattern altitude or lower, you’re pretty much dead, no matter how responsive your aircraft may be.

      • Duane,

        You make some great points.

        However, there’s a difference in a down draft and windshear, but can be related. You may have windshear without a downdraft.

        Yes a light plane is affected a lot with a downdraft, but can usually handle moderate windshear, which was my point.

        >>>However, the same inertial effect also works AGAINST the light aircraft, because being lighter, light aircraft have less inertia to carry them through a sudden short term tailwind or downdraft that would have little effect on a high speed heavy jet.<<<

        I will agree to disagree with you on the above statement. The jet will have a much more difficult time. When the jet loosed airspeed, it's hard to get back, unlike a piston plane, regardless of how short term the wind change is.

        I don't have quite the experience you have, as I have strict limits when it comes to mountain/canyon flying and winds. But with 25000 hours, I've done my fair share of mountain flying in light planes from a Cessna 172 to a Beech Baron (and a few others). I've never had sheer or turbulence to a point where I didn't think I'd make it. (And plan on keeping it that way).

        Agree, thunderstorms "can" be much worse than mountain winds, however, they are easily avoidable and there's almost always a way around them.

    • Rob – as you know, with a tailwind the aircraft will have a higher groundspeed, relative to any given airspeed (which are functions of the power and elevator trim settings for any given aircraft). As the tailwind decreases, it is equivalent (with respect to the aircraft’s relative momentum thru the air) to an increasing headwind. That causes a temporary increse in aerodynamic performance … meaning the aircraft will momentarily increase its indicated airspeed and pitch up.

      The end result on an approach at short final is the aircraft will tend to be high and fast.

      Given sufficient time and distance the aircraft will eventually return to a stabilized pitch and airspeed that correspond to whatever the original power and trim settings were before the wind shear … but unless there is a relatively long runway ahead when this happens on short final, it may be too late to salvage a landing approach unless the pilot takes immediate action on the power and pitch settings.

      Fortunately for us typical GA pilots, light aircraft respond more quickly to changes in power and pitch than do heavy aircraft, and most airport runways are plenty long enough to recover and land for most light aircraft. With the caveat being that the aircraft otherwise is on a correct flight path to touchdown with correct pitch and power settings just prior to the wind shear.

      To get a feel on how to sense and properly respond to this phenomenon, practice some touch and go approaches and landings on a gusty day with the wind more or less aligned with the runway – preferably with a flight instructor, if you’re not yet comfortable with such conditions. As the wind gusts up and down, it will be a challenge to maintain a consistent three-degree glide path and constant airspeed without overcontrolling the aircraft.

  • Ahhh. Got it.

    Makes sense now.

    I thought the decreasing ground speed would leave you short but I was not thinking of the planes reaction to the change in ‘current’.

    Thanks

  • With the decreasing headwind, to get the “high and fast” effect, it’s a rather significant wind change… probably a condition that we wouldn’t want to be in from the start, if we knew.

    However, once things stabilize again one needs to adjust the power to keep from getting low, as the ground speed slows. Goal for speed should be to remain constant.

  • Wind shear and momentum.
    Momentum = Mass X Velocity
    To best see this in real life, toss a popsicle stick into a moving stream of water against the current, it will almost immediately stop and flow with the water. Then toss in a 2 X 4. The 2 X 4 will enter the opposing current, go below the surface as it slows and then begin to flow down stream.
    There you have a light airplane with low wing loading and a heave beast with higher wing loading.
    Every fighter pilot knows energy management is everything (as in do not get behind, you know when you have that sinking feeling) and speed is life! (as in do not slow down to you normal stabilized approach in strong shifty winds)
    After the TWA L1011 accident, our airline gave the same wind profile without warning during simulator checks. Using high speed to ground cushion and taking a cut over the fence I landed just fine….and never got chewed out so much in my life….and the guys that crashed were just debriefed! They were teaching go arounds of course. (But what if have nowhere to go with your fuel remaining?)
    Once in the South China Sea we had a deck pitch well over 20 feet with very gusty winds well above normal recovery limits for the deck crew. Very rough. The angle of attack was all over and useless. Only 2 of us (out of six) got aboard. (Took a cut in a single seat jet.) They launched a couple tankers and all 6 planes went to Japan (and the bar)! (Now that was real go around training with a reward.)
    Anyway, when it is really, really bad, maybe everything you have be taught for landings within procedural limits needs rethinking.

Leave a Reply

Your email address will not be published. Required fields are marked *