Evaluation of the Six-Beam Lidar Scanning Strategy at the Boulder Atmospheric Observatory

Lidars have recently emerged as a valuable tool for wind resource assessment. Unlike cup anemometers on a traditional meteorological tower, which have fixed measurement heights and are limited by the tower height, lidars can measure wind speeds across an entire turbine rotor disk and can be easily deployed at different locations around a wind farm to examine spatial variability of the wind resource. The ability of lidars to accurately measure mean wind speeds has already been well-documented in the literature. However, several questions remain regarding the measurement of turbulence with lidars. Turbulence has profound effects on the amount of power produced by a turbine and can also impact loads on the turbine blades. Thus, lidars must be able to accurately measure turbulence in order to be seen as a viable alternative to meteorological towers.

One major difference between lidar-measured turbulence and turbulence measured by a cup or sonic anemometer on a tower results from the volume averaging inherent in remote sensing technology. While a sonic anemometer measures wind speeds in a small volume of air, lidars measure wind speeds across a probe volume that is typically 10–30 m in length at all measurement heights for a pulsed Doppler lidar, and increases quadratically with distance for a continuous wave lidar. This volume averaging tends to decrease the lidar-measured turbulence, as smaller scales of turbulence are averaged out in the probe volume. In contrast to volume averaging, two factors increase the lidar-measured turbulence: instrument/atmospheric noise and variance contamination induced by the use of a Doppler beam-swinging (DBS) scanning strategy. The examination of these latter two factors was a primary focal point of the Lower Atmospheric Thermodynamics and Turbulence Experiment (LATTE).

LATTE was carried out at the Boulder Atmospheric Observatory in Erie, Colorado in February and March 2014. A Halo pulsed Doppler scanning lidar, a WindCube v2 pulsed Doppler lidar, and a ZephIR 300 continuous wave lidar were deployed at the site, and a 300-m tower at the site was instrumented with sonic anemometers at several different levels. Additional instrumentation included an unmanned aerial vehicle (UAV) and a 449-MHz wind profiling radar. For most of the experiment, the Halo scanning lidar was used to evaluate the six-beam lidar scanning strategy, a technique developed at The Technical University of Denmark to mitigate variance contamination caused by the DBS strategy. The WindCube and ZephIR lidars employed DBS and velocity-azimuth display strategies, respectively, throughout the experiment, allowing for the comparison of turbulence measurements from three different lidars and scanning techniques.

The primary focus of this paper is the comparison of turbulence measured by the WindCube lidar with the DBS technique to the turbulence measured by the Halo lidar with the six-beam technique. Preliminary results indicate that the variance measured by the six-beam technique more closely matches variance measured by the sonic anemometers on the tower, particularly under unstable conditions when variance contamination tends to be most prominent. However, several drawbacks emerged for using the six-beam technique, including the tendency for the u and v variance to become erroneously negative when the vertical variance is large as a result of the six-beam variance solutions. Lidar noise correction techniques are also briefly discussed in this paper and covered in more detail in a companion presentation.