13 October 2026
What Slope Triggers CFD Over a Linear Wind Flow Model

You run WAsP on a ridge, the energy numbers come back clean, and then the met mast data doesn't match the model within the margin you'd want to defend in a bankability review. Before you blame the mast, check the slope. Linear flow models have a point where they stop being linear flow models and start being wrong in a specific, predictable direction.
Where a linear model starts lying to you
WAsP and similar tools solve a linearized version of the flow equations (the IBZ model, if you want the textbook name). That linearization assumes the flow stays attached to the terrain. It doesn't recirculate, doesn't separate off a ridge crest, doesn't form a lee-side eddy. On rolling hills and gentle escarpments that assumption holds well enough for a bankable assessment.
It stops holding once the terrain gets steep enough that air can't follow the surface anymore. Past that point you get flow separation on the lee side, speed-up errors on the windward face, and a wake structure downwind that no amount of roughness-length tuning will fix. The model keeps producing a number. The number just stops meaning anything.
The slope number people actually use
There's no single hard cutoff written into WAsP's code that flips a switch at a specific percent. What the industry leans on instead is the ruggedness index, RIX: the fraction of terrain inside your assessment radius where the slope exceeds a critical value, conventionally set at 30% (around 17°). A site with RIX close to zero is the gentle terrain WAsP was built for. A site where a meaningful share of the surrounding terrain sits above that 30% slope threshold is where linear results start drifting from reality, and the drift tends to get worse with distance from the mast.
So the practical answer to "what slope triggers CFD" is: once your candidate ridge has slopes regularly crossing 30%, especially on the lee side relative to your dominant wind direction, you're past the point where a linear model's output can be trusted without a CFD cross-check. Below that, on a windward slope that's a steady 15-20%, WAsP is probably fine. A lee slope at 35% with a crest, that's a different conversation.
It's rarely just one number
Slope alone doesn't tell the whole story. A steep but smooth, convex ridge behaves differently than a steep ridge with a forest edge partway up it, or an escarpment with a sharp crest that forces separation right at the point you planned to put a turbine. Surface roughness changes near the crest, isolated obstacles below the hub, even the orientation of the slope relative to the prevailing direction all shift where the linear model's errors show up and how big they get.
That's the part that's easy to miss working from a contour map alone. Two ridges with the same average slope can need completely different modeling approaches depending on where the steep sections actually sit relative to your turbine string and your dominant wind rose. A desk review built on slope and roughness pulled from stereo imagery across the whole candidate line, not just the handful of points you were planning to survey, is what lets you make that call before a crew drives out. Running that screen over the full ridge before committing to a mast location is the cheaper place to find out your site needs CFD, not after the data comes back strange.
If a ridge is borderline, that's exactly the kind of candidate worth screening properly before anyone packs a truck for a met mast install.