Two-component horizontal motion vectors from scanning eye-safe aerosol lidar

First results are presented from the application of two very different numerical techniques to derive two-dimensional, two-component, vector motion fields from eye-safe scanning aerosol lidar data. The first technique, known as the correlation method, calculates cross-correlation functions from block regions on the images. The method has been applied to other lidar data previously. The second technique is known as optical flow and is based on calculating spatial and temporal gradients and minimization of an energy function that has been modified to take into account the physical properties of fluid motion.

In this paper, we apply both techniques to ground-based scanning eye-safe aerosol lidar data collected during the 2007 Canopy Horizontal Array Turbulence Study (CHATS). The vectors closest to an instrumented tower are compared with sonic anemometer measurements. The algorithms are applied to two cases to investigate their performance during very different boundary layer states. In one case from evening hours, the atmosphere was weakly stable, nearly quiescent, with gravity waves and episodes of turbulence. In a second afternoon case, the atmosphere was unstable, turbulent, and featured an abrupt reversal of wind direction with the passage of a density current front.

The results from these two cases indicate that both algorithms perform better than expected for the weakly stable case and perform worse than expected for the unstable turbulent case. A hypothesis was that stratification and waves would create aerosol features that do not advect with the wind and that turbulence would. While the turbulent case is not bad (the means are in good agreement) there is substantially more scatter in the results. Overall, the results are encouraging because they suggest that two-component vector fields can be extracted from the images produced by a single, eye-safe, scanning aerosol lidar. This capability may be of particular value in applications where it is more important to have two wind components rather than only a single wind component with higher resolution as is available from current Doppler lidars.