Using the Existing Spectral Clutter Filter With the Nonuniformly Spaced Time Series Data in Weather Radar

Abstract

The National Weather Service (NWS) has a network of weather surveillance Doppler radars, WSR-88D, that is routinely upgraded to reflect current technological and engineering advances. During the last decade, the staggered pulse repetition time (SPRT) was proposed to mitigate range/velocity ambiguity in WSR-88D; however, it could not be used at lowest elevation tilts due to the inability to filter ground clutter from the nonuniformly spaced time series. A complicated spectral procedure was developed to address this issue, but the procedure was derived for a specific set of data acquisition parameters and used a lookup table to provide a value for the clutter spectrum width. Attempts to use SPRT with different sets of acquisition parameters always led to the degradation of filtering. It is proposed here that the clutter spectrum width be estimated by using Gaussian model adaptive processing (GMAP) that is currently used by NWS for uniformly sampled time series. GMAP is an adaptive ground clutter filter that performs an iterative fit of a Gaussian curve to the spectral coefficients identified as being due to ground clutter. GMAP cannot be used directly with the SPRT spectrum because of multiple clutter replicas that are due to staggered nonuniform sampling. However, the elements of GMAP can be exploited and used in the SPRT procedure. This letter presents how the elements of GMAP can be used with the SPRT data to provide successful clutter filtering at lowest elevation tilts.