Spectrum Width Measured by WSR-88D: Error Sources and Statistics of Various Weather Phenomena.

Spectrum widths, one of the three moments measured and displayed by the Weather Surveillance Radar-1988 Doppler (WSR-88D), are categorized for various weather conditions showing both expected and unexpected results. Weather phenomena are classified into seven categories based on radar observations, and the statistics of the censored spectrum width fields for each of the categories are obtained. Daytime fair weather without birds, stratiform rain and snow, and isolated tornadic storms produce weather signals that have the smallest volumetric median values of spectrum widths (i.e., < 2 m s^-1). Surprisingly, the median spectrum width values in the isolated tornadic storms are as low (i.e., <2 m s^-1) as in the fair weather (without the presence of echoes from birds). The median spectrum width value from fair weather regions contaminated with bird echoes is larger (i.e., 3.0 m s^-1). The largest median spectrum width values, ranging from 4.0 to 5.4 m s^-1, are associated with embedded areal squall lines. Clusters of severe storms and storms along broken squall lines appear to have median spectrum width values between these two regimes. Spectrum width fields are also shown to be more prone to errors than fields of reflectivity and velocity. Errors mainly result from overlaid echoes, improper automatic gain control (AGC) settings, low signal-to-noise ratios, and incorrect estimates of noise power. Thus spectrum width data fields require extensive censoring. The most persistent errors appear to be those related to overlaid weather signals and Alow signal-to-noise ratios.