Snow-Level Estimates Using Operational Polarimetric Weather Radar Measurements

Scanning polarimetric measurements from the operational Weather Surveillance Radar-1988 Doppler (WSR-88D) systems are evaluated for the retrievals of snow-level (SL) heights, which are located below the 0 degrees C isotherm and represent the altitude within the melting layer (ML) where snow changes to rain. The evaluations are conducted by intercomparisons of the SL estimates obtained from the Beale Air Force Base WSR-88D unit (KBBX) during a wet season 6-month period (from October 2012 to March 2013) and robust SL height measurements h(SL) from a high-resolution vertically pointing Doppler snow-level profiler deployed near Oroville, California. It is shown that a mean value height measurement h(L3) between the estimates of the ML top and bottom, which can be derived from the WSR-88D level-III (L3) ML products, provides relatively unbiased estimates of SL heights with a standard deviation of about 165 m. There is little azimuthal variability in derived values of h(L3), which is, in part, due to the use of higher radar beam tilts and azimuthal smoothing of the level-III ML products. Height estimates h(rho) based on detection of the ML minima of the copolar cross-correlation coefficient rho(hv) calculated from the WSR-88D level-II products are slightly better correlated with profiler-derived SL heights, though they are biased low by about 113 m with respect to h(SL). If this bias is accounted for, the standard deviation of the rhv minima-based SL estimates is generally less than 100 m. Overall, the results of this study indicate that, at least for closer radar ranges (up to similar to 13-15 km), the operational radar polarimetric data can provide snow-level estimates with a quality similar to those from the dedicated snow-level radar profilers.