Research Briefs

Capturing the wind

New research is helping wind energy developers determine the best sites for capturing wind

Wind turbines on a hillside.

Wind turbines in action on a wind farm in Utah. As wind energy grows in importance, scientists at NCAR are studying how wind turbines and farms interact with the atmosphere and how their output can be better predicted and managed.

New research from NCAR is helping wind energy developers determine the best potential sites for capturing wind.

Energy companies can lose money if they install turbines where winds are either too low to generate much power or so high that the turbines often need to be shut down to avoid damage. At present, wind farm developers try to reconstruct historical weather conditions by downscaling results from computer weather models to proposed wind farm sites. Typically this is done for 365 days of wind data, with each date randomly chosen from the preceding 10 years. But this approach, even when combined with a year of onsite observations, can miss extreme wind events and does not guarantee that the selected dates represent a realistic sample of the historical wind conditions.

A new method developed by a team of researchers led by Daran Rife can help developers identify prime wind sites more quickly and accurately. Instead of randomly selecting a single set of 365 days of wind data, the new method creates thousands of such sets. It then uses a novel statistical technique that compares each set to a high-quality reanalysis of observations of wind in the general vicinity. The set that aligns most closely with the observations is selected as the basis for a detailed map of likely winds at the site.

Such a method, when combined with a year of actual observations, reduces the uncertainty around the winds by 15 to 40% compared with the current approach. Moreover, it is so effective that an energy company can analyze just 180 days of data and obtain results as reliably as the current 365-day approach with far less cost and effort.

Based on a statistical technique known as the Monte Carlo Method, the approach can also be used for efficiently downscaling results from global climate models to project changes in regional climate.