7A.2 Snow microphysical retrieval from dual-wavelength radar measurements

Tuesday, 27 September 2011: 9:15 AM
Monongahela Room (William Penn Hotel)
Isztar Zawadzki, McGill University, Montreal, Quebec, Canada; and W. Szyrmer
Manuscript (1.2 MB)

An algorithm was developed to retrieve profiles of some microphysical properties of snow by using ground-based dual-wavelength observations. The measurement system consists of two collocated vertically pointing Doppler radars operating at X- and W-band. The three observables used in the retrieval are: the difference of the non-attenuated reflectivity at the two wavelengths [DWR], the X-band Doppler velocity, UD(X), as well as the equivalent reflectivity factor at X-band Ze(X), The microphysical quantities retrieved are: the parameter controlling the mass-size relationship [α] for individual particles and two parameters defining the particle size distribution (PSD) within a double-moment normalization framework. The two normalizing moments are the 2nd and 3rd for PSD with particle maximum diameter as coordinate (or 3rd and 4th for PSD defined for the equivalent melted diameter). These moments are related closely to the ice water content (IWC) and the PSD characteristic size representing the sizes being the main contributor to the IWC, as well as having important contribution to the reflectivity factor. Moreover, they are both not very sensitive to the two PSD ends that are very uncertain in general. The parameter α and the PSD characteristic diameter are determined by searching for their combination that describes the measured DWR and UD(X) with the smallest error. Then, the value of the IWC is deduced from Ze(X) corresponding to the obtained α and PSD characteristic size. This retrieval is repeated using different possible mass-size power law exponent, assuming different mass-projected area relation (used in the terminal velocity calculations) and for different forms of the normalized PSD complete or truncated, taken from the published literature. A few of the possible combinations have been excluded as not consistent with the measurements on the basis of the observed difference of the Doppler velocity at the two wavelengths, not used directly in the retrieval. The average value for each deduced quantity and its standard deviation represent the retrieval result and its uncertainty. The developed algorithm has been tested in the case of moderate snowfall without larger snowflake aggregates. The calculations are done at each range gate independently. The data processing of the radar measurements used in the retrieval includes: time averaging, calibration and correction for atmospheric (gaseous) attenuation for W-band reflectivity as well as air density correction for Doppler velocity.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner