Flying circuits this Friday with a student at CYTZ, I noticed an interesting phenomenon – that our progress on the downwind leg seemed unusually slow. We had a lot of time to level off from the climb out, pick a distant landmark to help fly a straight leg, contact ATC and ask for a touch-and-go, and then monitor our progress to the point where we began our descent. So I set up the GPS in the aircraft to display our ground speed, which turned out to be 55 knots. A quick cross reference to our airspeed (85 knots) showed a 30 knot headwind.

A second interesting phenomenon: on the base leg, we had to maintain a significant crab angle away from the runway to fly a square leg. We were using runway 08 (winds reported as 060° at 10 knots) yet we had to point the aircraft significantly to the west to compensate for the drift in the winds occurring on the base leg.

A third interesting phenomenon: on the crosswind leg (climbing out) my student remarked that he needed to place the nose significantly higher than normal to maintain the best-rate-of-climb speed of 80 knots.

A fourth interesting phenomenon: our approaches were consistently on the high side – we had to go around a couple of times too.

Obviously all these things are linked. What we were noticing were the effects of a significant wind shear between circuit altitude at 1000 feet above the ground, and the winds at the surface. After landing, we checked the NOAA forecast soundings for CYTZ and were not surprised to see a strong south westerly wind at 1000 feet (giving us the headwind in downwind leg and a quartering tailwind in the base leg) backing rapidly to about 10 knots down runway 08 at ground level.

What is wind shear?

Wind shear is often thought of as a change in wind speed or direction over a short period of time. You might imagine someone experiencing the effect of wind shear would say to themselves the wind changed suddenly.. But in reality wind shear is most often a change in the winds over a short distance. As the airplane moves from point A at time A to point B at time B – a difference in the wind at A compared to B is felt as a change over time, even though the wind at A is steady and the wind at B is steady.

Point A and point B can also be located over the same point on the ground, at different heights. If the plane is in level flight then there’s no change in the winds as it moves – no shear effects. But if in a climb or descent then as it moves through different layers of air with different wind speeds, wind shear effects can become very noticeable.

Going back to our circuits on Friday – the fact of having a strong headwind on the downwind leg should have warned us to expect wind shear effects. If the wind is blowing in the same direction from the ground to circuit altitude you’d expect to have a tailwind on the base leg – you’re flying opposite to the direction you’ll be landing, on that leg. A headwind – slow progress on the downwind – tells you the wind is going to shift on your descent. (And if it doesn’t, you’re using the wrong runway.)

Now, obviously, on any regular day, the wind at 1000 feet isn’t going to be identical to the wind on the ground. You’d expect stronger winds at altitude, and typically the wind veers as you climb. But on Friday we had at least 45 knots of wind speed difference between the 35 knot headwind on the downwind and the 10 knot headwind when we were facing the other way at touchdown. A change of that magnitude is quite unusual.

When flying at low altitude and in strong winds students are taught to anticipate and correct for the otherwise unexpected ground path the airplane they’re flying will take. We have a whole air exercise devoted to it – the one called “Illusions caused by drift.” But we don’t spend very much time looking at the effects of strong wind gradients. Perhaps we should. We could call it “difficulties caused by shear”.

Effects of wind shear

So why do we need to be concerned with a change in windspeed, above and beyond being concerned with flying in a steady wind?

Let’s do a thought experiment. Imagine we were in a small airplane, set up to fly in level flight at, say, 70 knots (so, perhaps with the flaps extended, and a moderate power setting.) Let’s pretend we’re flying into a headwind of 70 knots, too. So we’re stationary over the ground. Now let’s imagine that the great-weather-controller-in-the-sky instantaneously switches off the 70 knot headwind. What happens to the airplane? The airplane has momentum, so in that instant it remains stationary over the ground but now it’s airspeed has dropped to zero – the wind that was blowing over its wings just stopped. So it’s going to do what any airplane with no airspeed will do – nose down, and descend rapidly, at least until it has time to accelerate forwards through the now stationary air, rebuild its airspeed, and recover lift on the wings. Once that has happened, assuming no change in power settings, the plane will have recovered to level flight, but now moving foward over the ground at 70 knots.

Fine. That was wind speed changing in time. But how does that relate to steady winds but different in different places? Now imagine that the wind above 1000 feet was blowing steadily at 70 knots, and the air at 999′ feet and below was stationary. If your plane was descending slowly from some height above 1000′ it would begin being stationary over the ground feeling like it was almost hovering, until it descended across the boundary between 1000 and 999 feet at which point it’s airspeed would instantaneously drop to zero and it would start to fall out of the sky. After a second or two in fall it would quickly rebuild airspeed until it was moving forward at 70 knots instead of making a vertical descent. From the pilot’s point of view, there was a very sudden change in the wind at a particular instant that made the plane “drop”. It would instead be more accurate for the pilot to consider there was a sudden change in the wind at a particular place. We are apt to think of the airplane we’re sitting in as somehow stationary, and that outside influences occur at different times. But it’s just as likely that outside influences occur at specific places, and we simply take note of them happening when the plane arrives at that place. Those places are just as likely to be different altitudes than locations over the ground too.

The wind dropping from 70 to zero over a vertical distance of 1 foot is an extreme and unphysical example. Change the thought experiment, so the wind drops from 70 knots to zero over a vertical height of 100′, or 200′. You’ll get the same overall effects, smoothed out: a rapid descent, followed by an increase in ground speed. The suddenness of the change depends on the range of height over which the wind changed and the rate of descent of the airplane: the small the former, and the bigger the latter – the more likely the pilot is to be startled by the airplane’s behaviour.

Our exaggerated though experiment reveals the effect of descending into a decreasing headwind: the plane will descend more rapidly, and its ground speed will increase. If you reverse the experiment and climb into an increasing headwind, the airplane will show a reducing groundspeed and an increased rate of climb, until the wind remains constant with further altitude changes Then, there are also the combinations of climbing and descending into a decreasing headwind to be considered. I’ve put all four combinations with into this table, with a summary of what you can expect in that situation:

 

	…into an increasing headwind or decreasing tailwind leads to…	…into a decreasing headwind or increasing headwind leads to…
Climbing…	increased rate of climb and reducing ground speed	reduced rate of climb and increasing ground speed
Descending…	reduced rate of descent and reducing ground speed	increased rate of descent and increasing ground speed

 

What the Aeronautical Research Council had to say

The effect of wind gradients (“shear”) on climbing and descending has been recognized for more than 100 years. This is a research paper from Britain’s Aeronautical Research Council from 1954, written by one W. A. Mair of the Cambridge University Aeronautics Laboratory, which itself cites a previous paper from 1917 discussing the phenomenon and how to calculate its effects:

I can’t find the paper from 1917, also called The effect of a wind gradient on the rate of climb of an aeroplane, and I suspect it has been lost to history. (One can imagine its author: clipped tones, stiff upper lip, handlebar moustache and probably a monocle too.) W. A. Mair on the other hand is Professor William Austyn Mair, who was the Professor of Aeronuatical engineering at Cambridge University from 1952 to 1983. In 1954 he had a few conclusions that might be still useful to us here in the 21st century.

On Friday we had the wind changing by 45 knots over 1000 feet. 45 knots is 76 feet per second, which means an average wind gradient of 76 feet per second per 1000 feet, or 0.076 per second. (It may seem funny to measure something in “per seconds” but it’s correct.)

Mair writes that the maximum values of v likely to be encountered are roughly (below) 100′ – 0.1/s or more, and from 100-1000 feet 0.05/s. Our wind shear on Friday was close to the maximum we should be likely to encounter.

He also derives the (simplified) result that when climbing into a changing headwind at steady true airspeed the rate of climb increases by a proportion equal to -Vw / g where V is the airspeed, w is the headwind gradient, and g is gravity. On friday we had an average value of w of 0.076/s, and if were were maintaining a steady airspeed of 80 knots (135 feet per second) the rate of climb would have been higher by 32% compared to a no wind-shear day. That rather explains interesting phenomenon number 3, the higher-than-usual nose attitude to main the correct airspeed on the climb-out.

Similarly, when descending on the base and final legs on Friday, with the decreasing tailwind and increasing headwind, at and airspeed of 70 knots we could have expected a rate of descent 28% lower than usual with a regular power setting. We also had the wind swing behind us at the end of the base leg as it transitioned from south west to east, pushing us into the base to final turn. Those effects go some way to explaining unusual effect number 4.

If you’re really on the ball, you can expect to have a lower power setting than usual, to maintain a regular rate of descent, when descending into an increasing headwind. On the second circuit I did suggest using a lower-than-normal power setting for the initial descent, which helped.

Effects on approach

What other practical effects does the pilot notice? Mair investigates rates of climb but when approaching to land we are mostly interested in angles of approach, the better to answer the question with present aircraft configuration, are we going to reach the selected touchdown zone on the runway?

The effect on rate of a descent of a smooth wind gradient is constant: the rate of descent is changed to a new, lower, value, when descending into an increasing headwind. But the effect on ground speed gets bigger and bigger: imagine descending through an increasing headwind for thousands and thousands of feet. Eventually the headwind will get so strong you’ll end up with a negative ground speed, going backwards over the ground.

So with an increasing headwind on approach your approach will steepen over time, as you descend. Although initially you had to set a lower power setting to achieve a regular rate of descent, if you’re not careful by the time the tail wind drops off and the headwind increases you may find that you’re coming in short, and have to pile in the power at the end of the approach.

Effects on departure

The effects on climb-out can be startling too. Climbing into an increasing headwind gives an increased rate of climb, as discussed. But I do recall one recent flight climbing out to the east, also on a day of strong wind shear, with the tailwind behind us picking up. Even though my student was doing a great job maintaining 80 knots airspeed for the climb, the nose was way lower than usual and the rate of climb very poor – only about 300 feet per minute. I was watching the GPS ground speed, and over the course of a few seconds I watched it wind up: 75 knots, 80 knots, 85, 90, 95, 100…. very soon we were in a layer of air providing us a 25 or 30 knot tailwind. That experience on the departure to the east was clear evidence that we were going to have some interesting performance issues caused by wind shear when we came back to the circuit for landing, later – and so it proved to be.

Effects on the controls

Finally: let’s talk about how wind shear feels when you’re holding the controls. In this respect it’s quite similar to turbulence. A loss of airspeed due either to turbulence or to entering (by climbing or descending) a layer with decreasing headwind, will have the aircraft try to nose down to maintain airspeed. If you’re flying trimmed (and you probably should be) then you’ll feel the yoke pull forward against you as the nose drop and the plane lurches downwards. Recall that the yoke is not just something you use ot control the airplane, but it’s one way for the airplane to feed back to you what’s happening. A sudden increase in headwind due to shear will see the plane try to nose up and the yoke will push back towards you.