Detecting Waves in Doppler Radial Velocity Observations

Waves are ubiquitous in the atmosphere. Previous work on both summer and winter storms has shown that waves can trigger and/or enhance convection with impacts on cloud and precipitation. Historically, atmospheric waves have been identified using wind profilers, vertically-pointing lidar, and arrays of pressure sensors. Examination of sequences of Doppler velocity scans often reveal perturbations in the background velocity field in the form of moving parallel sets of linear features. We describe a method to identify these "velocity waves" in scanning operational radar Doppler radial velocity observations.

To isolate the waves seen visually in sequences of radial velocity plots, we subtract two sequential radial velocity fields to get a difference field. The difference field is then converted to a binary field to provide better visual distinction between nearby velocity waves. Sequences of binary images are used to estimate wavelength, phase speed, direction of propagation, wave depth, and the nature of the temporal evolution of the waves. We also characterize the limitations of this wave identification method in the context of the likely range of wave characteristics, radar hardware, and scan strategies.

Based on a sample of over 100 winter storms along the northeast U.S. coast, we found that velocity waves occurred in half of the storms. The velocity waves have wavelengths between 10 and 30 km and typically move several m/s faster than nearby mesoscale snow bands and frequently move in a different direction. The wave features are consistent across adjacent radar domains and appear to originate outside of the precipitation echo.

Our easy to implement method will provide both the operational and research communities with a robust wave detection tool in regions with precipitation echo.