Roughness lengths in complex terrain derived from sodar wind profiles

Wind profiles have been measured using monostatic sodar in more than 20 locations in the Northeast US and Canada, in support of wind energy resource assessment programs. Most of the locations have been in complex terrain, and often they have been partly or mostly forested. A better understanding of roughness lengths derived from wind profiles to 200 m would be an aid to modeling the wind resource on a meso- or micro-scale. This is particularly important because modern utility-scale wind turbines have hub heights of 80 m, reaching a layer of the lower boundary layer that is not commonly addressed by tower measurements. Furthermore, many present-day wind farms are large, covering a range of surface types, so that a single roughness parameterization is not adequate for modeling the wind resource for a given project.

Sodar data include the 3 orthogonal wind speed components and their standard deviations. The standard deviation of the vertical velocity was used to make a conversion from the vector speeds to scalar speeds more commonly provided by the reference tower (cup anemometer) measurements. In flatter more homogeneous terrain, agreement between sodar and tower wind speeds is quite good, with slopes near 1.0 and R2 values greater than 0.9. However, significant variation in the wind speed profile is found over short distances in complex terrain or more heterogeneous cover.

For this paper, sodar wind profiles from 30 to 200 m at 5 sites are compared, representing a range of landscape and cover types, as well as stability classes. The roles of heterogeneity in surface roughness (derived from high-resolution aerial photos), as well as stability and terrain to the observed z0 are examined. The accuracy of the commonly used power-law description of the wind profile is evaluated for various terrain types and conditions.