Suppression of precipitation bias in wind velocities from continuous-wave Doppler lidars

Abstract

In moderate to heavy precipitation, raindrops may deteriorate the accuracy of Doppler lidar measurements of the line-of-sight wind velocity because their projected velocity in the beam direction differs greatly from that of air. Therefore, we propose a method for effectively suppressing the adverse effects of rain on velocity estimation by sampling the Doppler spectra faster than the time taken for a raindrop to transit through the beam. By using a special averaging procedure, we can suppress the strong rain signal by sampling the spectrum at 3 kHz. A proof-of-concept field measurement campaign was performed on a moderately rainy day with a maximum rain intensity of 4 mm h−1 using three ground-based continuous-wave Doppler lidars at the Risø campus of the Technical University of Denmark. We demonstrate that the rain bias can effectively be removed by normalizing the noise-flattened 3 kHz sampled Doppler spectra with their peak values before they are averaged down to 50 Hz prior to the determination of the speed. In comparison to the sonic anemometer measurements acquired at the same location, the wind velocity bias at 50 Hz (20 ms) temporal resolution is reduced from up to −1.58 m s−1 for the original raw lidar data to −0.18 m s−1 for the normalized lidar data after suppressing strong rain signals. This reduction in the bias occurs during the minute with the highest amount of rain when the focus distance of the lidar is 103.9 m and the corresponding probe length is 9.8 m. With the smallest probe length, 1.2 m, the rain-induced bias is only present at the period with the highest rain intensity and is also effectively eliminated with the procedure. Thus, the proposed method for reducing the impact of rain on continuous-wave Doppler lidar measurements of air velocity is promising and does not require much computational effort.