Wind · Assessing a site
One number decides whether a wind turbine on your property is a power plant or an ornament: the average wind speed at the height the rotor will actually turn. Everything else on this page exists to help you find that number honestly, before any money moves.
It is the cheapest work in the whole project and the work most often skipped.
Start here
A stream carries roughly twice the energy when twice as much water flows. Wind does not behave that way. The energy in moving air rises with the cube of its velocity, which means doubling the wind speed makes eight times the power available, and a wind speed one fifth lower than expected takes away nearly half the energy.
That single relationship is why wind assessment deserves an entire page, and why wind is the source on this site most often oversold. Every other decision in this section, the machine, the tower, the wiring, the permits, is downstream of a number that most people estimate carelessly and then build on for twenty years.
Suppose a site is assessed at 11.8 miles per hour average at hub height, and a turbine is bought against that figure. The site actually averages 9.4 miles per hour, which is 20 percent lower. The energy available does not fall by 20 percent. It falls by the cube of the ratio.
0.8 × 0.8 × 0.8 = 0.512
The machine produces 48.8 percent less energy than planned. Nothing is broken. Nothing was installed badly. The site was simply read wrong by an amount that would be unremarkable in almost any other kind of estimate.
Read that box again before you read anything else in this section. It explains most of the disappointed owners, most of the abandoned towers, and most of the reason the second guide in this hub exists.
Where the wind actually is
Air moving across the earth drags against everything it touches. Grass, crops, hedgerows, barns, and above all trees. The result is a layer near the surface where wind is both slower and messier than it is a hundred feet up, and the rougher the ground cover, the thicker and more disturbed that layer becomes.
The relationship between height and speed is described by the wind shear power law. It takes a wind speed measured at one height and projects it to another.
V2 = V1 × (H2 ÷ H1)α
V1 is the known speed at the known height H1, V2 is the projected speed at your hub height H2, and α is the wind shear exponent, which stands in for how rough the ground is.
The National Laboratory of the Rockies uses a shear exponent of 0.14 across residential, commercial and large distributed wind classes in its cost modelling[1]. That figure is a reasonable default for open ground. Rougher sites run higher, and a higher exponent is not good news even though it makes the multiplier larger, because it means the wind at the bottom was badly suppressed to begin with.
Steep shear carries a second cost that never appears in an energy estimate. When the wind at the top of the rotor circle is meaningfully faster than the wind at the bottom, every blade is loaded and unloaded once per revolution, for every revolution of a twenty year service life. That is a fatigue problem, and fatigue is what actually ends most small turbines.
| Terrain | Typical α | 30 ft to 100 ft multiplier |
|---|---|---|
| Open water, smooth coastline | 0.11 | 1.14 |
| Open field, flat prairie, cropland | 0.14 to 0.16 | 1.18 to 1.21 |
| Forested land, closed tree canopy | 0.21 | 1.28 |
| Suburban, scattered buildings and trees | 0.25 | 1.35 |
| Dense built environment | 0.29 | 1.42 |
The multiplier column is the arithmetic only. It does not mean a forested site produces more than an open one. The forested site starts from a much lower speed near the ground, and a larger multiplier applied to a smaller number still gives a smaller number.
Department of Energy guidance in the Small Wind Guidebook is that the bottom of the rotor blades should sit at least 30 feet above any obstacle within 300 feet of the tower[2]. The Small Community Wind Handbook, from the same program, states the same 30 foot clearance measured against anything within a 500 foot horizontal radius[3].
Both figures are official and both are rules of thumb rather than engineering limits. The honest reading is that 300 feet is the minimum anyone should consider and 500 feet is the better target, particularly on the upwind side and particularly where the obstacles are trees, which grow.
Plan for the site you will have in twenty years, not the site you have today. A row of saplings on the western fence line is a future turbulence source with a known growth rate, and it is far cheaper to buy tower height once than to discover the shortfall in year twelve.
The other half of the problem
A wind resource has two properties, and the assessment methods below only measure one of them well. Average speed sets how much energy passes the site. Turbulence sets how much of it a machine can convert, and how long the machine survives doing so.
Turbulence is the rapid, disorganised variation in speed and direction that ground clutter creates downwind of itself. A rotor sitting in it faces air that changes velocity and bearing faster than the machine can respond. A horizontal-axis turbine spends its time yawing to chase a direction that has already changed, and every one of those movements is load on a bearing.
The number that matters here is not published for most sites, and estimating it from a desk is not realistic. What is realistic is recognising the conditions that produce it: obstacles upwind, complex terrain, and any mounting height that puts a rotor inside the layer the obstacles disturb.
A manufacturer's power curve is produced under smooth, low-turbulence conditions. A rotor in disturbed flow does not reach those figures, and the gap is not the manufacturer being dishonest. It is the site being different from the test.
The disturbed air behind an obstacle extends well downwind of it, and in complex terrain a ridge or a dense shelterbelt can affect flow far beyond the distance that looks reasonable from the ground.
A building is an obstacle. Mounting a turbine on top of one places the rotor in the worst air on the property, and couples vibration into the structure people sleep in. This is covered directly in the guide on why residential wind disappoints.
Finding your number
They are listed in ascending order of trustworthiness and ascending order of effort, which is not a coincidence. Use the early ones to decide whether to bother. Use the last one before you spend money.
National and state wind resource maps published through the Department of Energy's WINDExchange program show broad patterns at stated heights. They are free, they take a minute, and they will tell you whether you live in a part of the country where wind is worth a second thought.
Department of Energy guidance is explicit about the limitation. State wind maps cannot include complex terrain, ground cover, wind speed distribution, direction distribution, turbulence intensity or other local effects, and their resolution is not high enough to identify local site features[2].
Verdict: use it to rule a site out, never to rule one in. Also check what height the map is drawn at, because adjusting from map height to your hub height adds error through the shear exponent you choose, and that error grows with the height difference.
Long records, freely available, and measured with real instruments. That combination makes airport data tempting and it is genuinely more useful than a map, because it shows seasonal shape and year to year variation rather than a single average.
The problem is where the instrument sits. Standard meteorological measurement is taken at 10 metres, roughly 33 feet, in a wide open paved corridor kept deliberately clear of obstructions. Your farmyard has a barn, a treeline and a hill. The two places are not the same place, and the difference does not average out.
Verdict: excellent for understanding the shape of a normal year and how bad a calm one gets. Poor for the absolute number.
Trees exposed to persistent wind grow into a permanent record of it. Small branches bend downwind first, a condition called brushing. Larger limbs follow, and the upwind side grows short or is stripped away, which is flagging. In stronger regimes the trunk itself curves downwind, called throwing, and at the extreme the crown grows along the ground.
The Griggs-Putnam index of deformity grades this on a scale from 0 to VII and associates each class with a wind speed band. Brushing with slight flagging corresponds to roughly 7 to 9 miles per hour, and clear flagging to roughly 9 to 11 miles per hour[2]. The effect reads most reliably on coniferous evergreens, whose profile to the wind stays constant through the year.
Verdict: the most undervalued method on this page. It costs nothing, it reflects decades rather than one season, and it describes the exact spot you are standing in rather than a regional average. Sample several trees and average them. Treat it as an indicator for identifying candidate sites, not as a measurement.
An anemometer and data logger on a temporary mast at the planned hub height, left in place through a full annual cycle. It is the only method on this page that produces a number you can responsibly build against, because it is the only one measuring the actual air the actual rotor will meet.
A year matters. Wind has a strong seasonal shape in most of the country, typically stronger in winter and spring and calmer in summer. A three month campaign run over the windy season produces a number that flatters the site, in the same way that measuring a creek only in April flatters a stream.
Good logging equipment records more than an average. Speed distribution, direction distribution and turbulence intensity all come out of the same campaign, and all three change what machine suits the site.
Verdict: the standard. On a project costing tens of thousands of dollars, a year of measurement is a rounding error against the cost of being wrong, and this is the step that separates the installations that work from the ones photographed rusting.
Look at the map. If the region is plausible, pull the nearest long weather record and look at its seasonal shape. Walk the property and read the trees, paying attention to which direction they have been shaped from. If all three still point the same way, put up a mast and measure for a year while you read the rest of this section. If any of the three says no, you have saved yourself a great deal of money in an afternoon.
The screening equation
Once you have an average speed at hub height, the Department of Energy publishes a screening equation for annual energy output from a small wind system. It takes two inputs and returns kilowatt-hours per year[2].
AEO = 0.01328 × D2 × V3
AEO is annual energy output in kilowatt-hours per year. D is rotor diameter in feet. V is the annual average wind speed in miles per hour at your site, at hub height.
Two things are worth noticing about its shape. Rotor diameter appears squared, because the rotor sweeps a circle and area rises with the square of diameter. Wind speed appears cubed, for the reason the first section of this page laboured. A machine one third larger across gathers about 78 percent more energy. A site one third windier delivers about 137 percent more.
Which is to say: given a choice between a bigger machine and a better site, take the site every time. Given a choice between a bigger machine and a taller tower, the tower usually wins too, for the same reason.
No machine converts all the energy in the wind it meets. Betz's law, which follows from the physics of a slowing airstream rather than from any engineering limitation, caps extraction at 59.3 percent of what passes through the rotor circle. Real machines fall well short of that.
Department of Energy figures put the maximum power coefficient of practical turbines between 0.25 and 0.45[2]. The 0.01328 constant carries a representative conversion efficiency inside it, which is what makes the equation usable without a power curve, and also what makes it an estimate rather than a specification.
It is not a substitute for a manufacturer's power curve run against your measured speed distribution, and it takes no account of cut-in speed, cut-out speed, rated power clipping, or the turbulence losses described above. Its job is to answer one question: is there enough here to keep going. Answer that first, and buy nothing until it comes back yes.
Worked example
Open farmland, no obstruction within 500 feet of the proposed tower site. The figures below are illustrative and chosen to demonstrate the sequence. This is the same property the rest of the wind section returns to, so these numbers reappear in the later guides.
A year of on-site logging at 30 feet returns an annual average of 10 miles per hour. That is a real but unremarkable resource. It is better than most of the country and well short of the plains.
The plan is a 100 foot hub height. Open cropland, so a shear exponent of 0.14, matching the value the National Laboratory of the Rockies applies to distributed wind.
10 × (100 ÷ 30)0.14 = 10 × 1.18 = 11.8 miles per hour at hub height.
A 12 foot rotor, which is a machine of roughly two and a half kilowatts rated capacity.
0.01328 × 122 × 11.83 = 0.01328 × 144 × 1,643 = 3,142 kilowatt-hours a year, which is about 8.6 kilowatt-hours a day averaged across the year.
The property worked through the solar section consumes 9,100 watt-hours a day, or 3,322 kilowatt-hours a year. This machine screens at 95 percent of that, on an annual average, at a site better than most people have.
3,142 kilowatt-hours from a 2.5 kilowatt machine running all year implies a capacity factor of 14.3 percent. Capacity factor is simply what a machine actually produced divided by what it would have produced running flat out continuously.
The Department of Energy's Distributed Wind Market Report puts the average net capacity factor for a sample of 100 small wind projects at 13 percent, with the range running up to 28 percent[4]. The screening estimate lands within about a point of the observed national average, which is a reasonable indication that the method and the inputs are both behaving.
The same 12 foot rotor, on the same property, at four hub heights. Nothing changes except the length of the tower.
| Hub height | Speed at hub | Screened annual output | Against 30 feet |
|---|---|---|---|
| 30 ft | 10.0 mph | 1,912 kWh | baseline |
| 60 ft | 11.0 mph | 2,545 kWh | +33% |
| 100 ft | 11.8 mph | 3,142 kWh | +64% |
| 140 ft | 12.4 mph | 3,646 kWh | +91% |
Raising the same machine from 30 feet to 100 feet increases its annual output by about 64 percent. No larger rotor, no better generator, no different site. Only steel.
This is the single most useful table on the page, and it is why the tower is treated in this section as the power plant rather than as the thing the power plant sits on. It is also why short towers and rooftop mounts produce the results they produce.
Before you celebrate
The worked example covers 95 percent of a household's annual consumption, and a reader who stopped there would be badly misled. Averages hide the property that actually governs an off-grid system, which is when the energy arrives.
Because output follows the cube of speed, a wind machine spends much of its life producing very little and a small number of hours producing a great deal. Calm weeks happen. They often happen in high summer, and they happen in still, cold high pressure in midwinter, which is exactly when a household most wants power.
Wind is therefore rarely a sole source at property scale. It works as one contributor in a system with storage, usually alongside solar, whose good and bad periods are partly opposite to its own. Solar peaks in summer daylight, wind commonly peaks in winter and at night. That complementarity is the real argument for wind on a property that already has panels, and it is a better argument than annual totals.
Work out what your household actually consumes, from a device list. The four checks do not change at property scale.
Run the numbersThe source most properties should build first, and the section this worked example borrows its load from.
Read the sectionIf the property also has moving water with real drop, this is where to look before buying a turbine for the air.
Read the sectionCommon mistakes
The most common error and the most expensive, because the cube relationship multiplies it. A speed figure without a stated height attached to it is not a figure at all.
Three months over the windy season produces a number that flatters the site. Measure across a full annual cycle, for the same reason a stream is measured in August rather than April.
The exponent is a description of ground roughness, not a tuning parameter. Picking a high value for a forested site to make the hub-height projection look better ignores that the same roughness suppressed the reference speed underneath it.
Trees grow, neighbours build, and a shelterbelt planted for the livestock becomes a turbulence source in fifteen years. The clearance rule applies to the mature site, not to today's.
The equation carries an assumed conversion efficiency and knows nothing about cut-in speed, rated power or turbulence. It decides whether to continue. It does not size or select a machine.
Ninety-five percent coverage on paper still leaves calm weeks in which the machine contributes almost nothing. Storage and a second source are part of the design, not an upgrade to consider later.
Next
Most sections of this site move from assessment straight into equipment. This one does not, because the gap between what small wind is expected to do and what it typically does is larger than in any other part of the Energy domain, and a reader who meets that gap after buying a machine has been failed.
The next guide sets the screening figure you just produced against measured performance, real installed costs, and the regulatory obstacles that stop many projects before they start. Read it before pricing anything.
Measured capacity factors, installed costs, and the four failure patterns that account for most of the gap between expectation and outcome.
Read the guideThe six property-scale generation areas, and how they compare against each other for a given site.
Back to the hubSources
Wind speed, rotor and output figures in the worked example are illustrative and chosen to demonstrate the method. Measure your own site at hub height across a full year before designing anything, and confirm what your jurisdiction permits before committing money.