How accurate does a wind sensor need to be?

The level of accuracy depends on the application. Environmental monitoring, aviation, wind-assisted propulsion and industrial control systems all require reliable, repeatable measurements that remain stable over time. Equally important is long-term performance, as mechanical wear can reduce measurement quality. Gill’s ultrasonic sensors provide stable wind speed and direction measurements without moving parts, helping maintain confidence throughout the instrument’s operating life.

Read the reference speed, not just the percentage

An accuracy figure without a reference speed is not a specification. Gill 2-axis sensors are stated as 2% RMSE for speed and 2 degrees RMSE for direction at 12 m/s. The 12 m/s is part of the claim.

Two other things belong in the same sentence and are often left out. RMSE is a statistical measure across a population of readings, not a worst-case bound. And on some instruments the figure changes across the range: WindSonic 60 is stated at under 2% RMSE up to 20 m/s and under 3% RMSE from 20 to 60 m/s, which is the honest way to publish it.

Three properties, not one

For a crane or bridge wind cut-out, repeatability at the threshold matters more than absolute accuracy: the consequence of a sensor that trips at a slightly different speed each time is either a needless shutdown or an unsafe lift. For a wind resource assessment, stability across the campaign is what makes the campaign valid.

What the application usually needs

The error that usually dominates

On most installations the sensor is not the largest source of error. Siting is. An instrument in the wake of a mast, a handrail, a superstructure or a building reports the wake accurately, and no accuracy specification protects against it.

Before tightening a sensor specification, check the exposure, the height and the distance from obstructions. A 2% sensor badly sited is worse than a 4% sensor well sited, and the second one is cheaper.