A dual-wavelength radar method to measure snowfall rate

A dual-wavelength radar method to estimate snowfall rate has been developed. The method suggests taking simultaneous and collocated reflectivity measurements at two radar wavelengths. Snowfall backscattering at one of these wavelengths should be in the Rayleigh regime or sufficiently close to this regime, while backscattering at the other wavelength should be substantially outside this regime for typical snowflake sizes. Combinations of K-a-band (for a shorter wavelength) and X-, C-, or S-band (for a longer wavelength) radar measurements satisfy this requirement. The logarithmic difference between reflectivities at these two wavelengths provides an independent estimate of snowflake median size D-m, which exhibits a very low sensitivity to snowflake density and details of the size distribution. The estimates of D-m and radar reflectivities Z(e) at the longer wavelength are then used to obtain snowfall rate R from the Z(e)-R-D-m relationships, which have a snowflake effective density rho(e) as a "tuning" parameter. The independent information about snowflake characteristic size accounts for much of the improvement of the dual-wavelength method over traditional, single-parameter Z(e)-R relationships. The paper also presents experimental data collected during January-March 1996, near Boulder, Colorado, with the National Oceanic and Atmospheric Administration's K-a- and X-band radars. The radar data were supplemented by simultaneous ground measurements of snow accumulation. Comparisons of the ground and dual-wavelength radar measurements indicate that a tuning value rho(e) of about 0.03-0.04 g cm(-3) provides a good match with surface-observed snow accumulations. Differences in dual-wavelength radar estimates of accumulation for rho(e) between 0.03 and 0.04 g cm(-3) are usually within 25%, while existing X-band, single parameter Z(e)-R relationships yield accumulations that differ by as much as a factor of 4.