Merging Doppler velocity spectra in time and height to overcome mismatch in radar pulse volumes from collocated radars

 

Author

Christopher R Williams — University of Colorado, Boulder

Category

Precipitation

Description

The DOE ARM Climate Research Facility has combined scanning cloud radars into dual-frequency configurations designed to sample the three dimensional structure of cloud systems. One radar configuration consists of W-band (95 GHz) and Ka-band (35 GHz) radars with matched beamwidth antennas mounted on a single steerable platform. This configuration allows both radars to simultaneously observe the same radar pulse volume. Since both radars are observing the same cloud particles, differences in measured reflectivity and Doppler velocity are due to non-Rayleigh scattering features of the cloud particles and can be used to retrieve cloud properties. Another radar configuration consists of Ka-band (35 GHz) and X-band (9.5 GHz) radars mounted on a steerable platform, but the X-band beamwidth is larger than the Ka-band beamwidth (1 vs. 0.3 degrees). Since both radars are not observing the same cloud particles, differences in measured reflectivity and Doppler velocity cannot be directly used to retrieve cloud properties. Observations from multiple scans and multiple range gates need to be combined to scale the mismatched radar observations to a common spatiotemporal resolution.

This study addresses the pulse volume mismatch in collocated vertically pointing radar observations by combining Doppler velocity spectra over multiple profiles and multiple range gates forming statistically homogenous regions. The size of a homogenous region is determined by pre-defined correlation lengths. When studying precipitation features, the correlation lengths can be scaled to reflectivity (Z) and mass-weighted mean diameter (Dm). By setting Z and Dm tolerances, the uncertainty estimates of homogenous regions are consistent from region to region. Relaxing the tolerances generates larger spatiotemporal homogenous regions. The tolerance level is dictated by the scale of the modeling being performed.

Since the initial data from collocated Ka- and X-band radar systems are still being collected, the upscaling procedure was validated using 35-GHz (KAZR) radar (0.5 degree and 1-s dwell) and 2.8-GHz (S-band) radar (2.5 degree and 7-s dwell) observations during the MC3E field campaign. During stratiform rain and using tolerances of 0.5 dBZ and 0.1 mm, homogenous regions were observed to extend over 21 seconds in time and 500 m in height. The analysis technique as well as the vertical and temporal structure of homogenous regions will be shown at the ASR Science Team Meeting.