Photogrammetric Analysis of Rotor Clouds Observed during T-REX

One of the key objectives of the T-REX field campaign, which took place in 2006 in the lee of the Sierra Nevada in eastern California, was to explore the origin, dynamical structure and temporal evolution of atmospheric rotors. On several days during T-REX, a largely stationary rotor cloud was observed over Owens Valley, in the lee of the Sierra Nevada. Small cloud fragments often formed just upstream of this stationary cloud, moving towards the main cloud and eventually being subsumed by it. Studying the evolution of these upstream cloud fragments could provide significant new insights into the structure and dynamics of the airflow upstream of and within the rotor. However, because of the clouds' small spatial scale and high temporal variability they are difficult to investigate with traditional tools such as radar and lidar.

In this study, we make use of a stereo photogrammetric data set, which was collected during T-REX but little used thus far. Using photogrammetric data is a little known but highly valuable tool for studying smaller, highly ephemeral clouds. The T-REX data set consists of matched stereo pairs of photographic images obtained at high temporal (on the order of seconds) and spatial resolution (limited by the pixel size of the cameras). To this data set, we have applied computer vision techniques to develop algorithms for camera calibration, automatic feature matching, and ultimately reconstruction of 3D cloud scenes. Using the photogrammetric imagery, we have been able to track the 3D path of cloud fragments at the upstream edge of the rotor cloud and monitor their growth and dynamics until their merger with the main cloud. The excellent temporal and spatial resolution of the images allows us to demonstrate the existence of small-scale features in the highly turbulent airflow at the leading edge of rotors and quantify their spatial and temporal scales.