Validation of CloudSat-Based Snowfall Profiles using C-band Ground Radar

While recent studies have shown that the 94 GHz CloudSat radar is sensitive to falling snow, retrievals of snow rate or snowfall profiles are difficult to validate. Typical Z-S relations show wide variability depending upon hydrometeor shapes, size, composition, and ground-based snow gauges are sparsely located or suffer from undercatch due to blowing snow. We present an analysis of CloudSat-based retrievals of snow water content profiles, for cases where the CloudSat satellite passed within 150-km of the C-band King City (WKR) polarimetric ground radar operated by Environment Canada. 28 cases were identified where the CloudSat overpass occurred during precipitation events during December-February 2006, 2007 and 2008, and where the associated ECMWF temperature profile never exceeded -2 C. The tabulated scattering properties for dendrite shapes developed by Liu (2008) were used to vary the snow size distribution parameters, for different assumptions on the supercooled cloud water fraction, such that the overall 94 GHz reflectivity profile from each candidate profile exactly matches the observed CloudSat radar profile.

Rather than attempting to verify in retrieval space (i.e, retrieved water content or snow rates), we use the WKR reflectivity profile to compare against each of these candidate profiles directly in observation (reflectivity) space. Using the nearest-time 10-minute WKR volume scan, the C-band radar profile was reconstructed underneath the CloudSat transect. Attenuation between the WKR radar and the snow event can be neglected owing to the minimal attenuation of frozen water at C-band. From each of the candidate profiles above, the vertical C-band reflectivity profile was forward modeled and compared against the associated WKR radar profile. To examine overall agreement between these two radar profiles, we examine cumulative contoured frequency by altitude (CCFAD) diagrams for different snow events. Additionally, upwelling brightness temperatures at 91, 150 and 183 GHz are computed for each of the candidate hydrometeor profiles, and compared with observations at these same frequencies from the Microwave Humidity Sounder (MHS) onboard NOAA-18 (in near time coincidence with CloudSat).