Compensating reflectivity for clutter filter bias in the WSR-88D

The presence of ground clutter, especially anomalous propagation (AP) ground clutter, presents serious problems for radar precipitation estimation and model initialization. Clutter filters are able to effectively cancel AP clutter in most situations. However, it is well known that clutter filters introduce a negative bias in reflectivity near zero radial velocity. This bias can range from 1 to 15 dB, significantly impacting radar precipitation estimates. Recently, numerous methods for identifying AP clutter have been developed. These methods will allow the clutter filters to be selectively used only in the regions containing ground clutter. However, it is not always possible to avoid application of the clutter filters to regions with weather echoes. In this case, a reflectivity compensation method may be used to correct for the negative bias introduced by the clutter filter. A simple Gaussian correction model (SGCM) was developed at NCAR by Frank Pratte for this purpose.

The compensation system has been tested on several data sets from the WSR-88D. Time series data, collected with NCAR's Archive 1 data acquisition (A1DA) unit, was processed using the pulse-pair method to obtain reflectivity (Z), radial velocity (V) and the velocity spectrum width (W). By processing the time series data we were able to obtain both the filtered and unfiltered data streams for comparison. This allowed us to apply the reflectivity compensation to the filtered data and compare the results to the unfiltered data, which in the absence of clutter contamination, provides the `truth' field.

The SGCM effectively removes the clutter filter bias in reflectivity (in most situations) using only the filtered Z, V and W estimates. However, the compensated reflectivity estimate has an increased variance, as well as isolated outliers with unrealistically high reflectivity values. Refinement of the SGCM method has resulted in elimination of the outliers. Various filtering methods have been tested in order to further reduce the variance of the compensated reflectivity values. In this paper we describe the SGCM compensation system, and present the results from the verification tests performed.