Surface roughness length for wind resource screening

If you've sat through a resource assessment review where someone asks "what z0 did you use for that sector" and the room goes quiet, you know why this term causes more arguments than it should. Surface roughness length is a single number, in meters, that stands in for how much the terrain upstream of your mast drags on the wind before it reaches your anemometer height.

Get it wrong in a screening model and your extrapolated hub-height wind speed drifts off by a margin that matters when you're ranking ridgelines against each other. Get it right and you've got a defensible basis for picking which candidate position deserves a met mast instead of a desk pass.

What z0 actually represents

z0 is the height above ground where, mathematically, the mean wind speed in the logarithmic wind profile would hit zero. Nobody measures wind at z0 directly, it's a fitted parameter, but it correlates tightly with what's physically sitting on the ground: grass, crop stubble, scattered farmsteads, shelterbelts, a treeline, a village. Taller and denser obstacles upwind mean a larger z0, a thicker boundary layer, and more shear between your 10 m reference and your 100+ m hub height.

The log law puts it plainly:

u(z) = (u*/κ) × ln(z/z0)

where u* is friction velocity and κ is the von Kármán constant (0.4). You don't need to run that by hand for a screening pass, but it's worth knowing that z0 sits inside a logarithm, not a linear term, so small changes in roughness class compound more than people expect once you're extrapolating up past 80 or 100 meters.

Roughness class table you'll use

Most siting shops still work off the Davenport/WAsP-style classes rather than raw z0 values, because it's faster to eyeball terrain cover and assign a class than to argue over the third decimal place.

Class	z0 (m)	Terrain description
0	0.0002	Open water, mudflats
0.5	0.0024	Open sea, lake, airport runway
1	0.03	Open farmland, no hedges, scattered buildings
1.5	0.055	Farmland with some windbreaks (>1 km apart)
2	0.1	Farmland with many windbreaks, low crops
2.5	0.2	Farmland with closer windbreaks, orchards
3	0.4	Villages, small towns, farmland with many trees
3.5	0.8	Larger towns with tall trees
4	1.6	Dense forest, large cities

For a ridge screen, you're rarely dealing with one class across the whole fetch. Treelines on the windward flank, open pasture on the crest, a cluster of farm buildings tucked in the lee. A screening tool that derives roughness from imagery rather than a single lookup assumption catches that variation sector by sector instead of flattening it into one number for the whole site.

Where the shear exponent comes in

The power law version of the same idea is what most people use for quick hub-height extrapolation:

u(z) / u(zr) = (z / zr)^α

α, the wind shear exponent, isn't an independent measurement, it's a stand-in for roughness and stability combined. Open water sits around 0.10-0.12. Open farmland lands closer to 0.14, which is why that value shows up as a default in so many IEC-flavored templates. Push into sheltered farmland with scattered trees and you're in 0.18-0.22 territory, and once you're near forest edges or built-up terrain α can climb past 0.25, which means your 10 m reference data badly undersells what's happening at 100 m.

This is the part that bites people doing a desk screen off a single nearby reference mast: if the reference site's roughness doesn't match your candidate ridge, borrowing its α and calling the extrapolation close enough can rank two positions in the wrong order.

Why this matters before you mobilize

None of this requires a trip to the field to sort out an initial ranking. Roughness class, obstacle height near the proposed mast location, and the resulting shear profile are all things you can screen from stereo imagery and terrain derivatives before a crew drives out. Wind Farm Siting builds that terrain and obstacle picture per project so you're sending the survey team to the ridge that already looks like it clears the obstacle and roughness checks, not the one that turns out to have a windbreak sitting right where the mast would go.

If you're staring at three candidate ridges and only have budget for one mast this season, run the roughness and obstacle screen first and let it tell you which one is worth the drive.