Object-Orientated Filter Design in Spectral Domain for Polarimetric Weather Radar

Abstract

Aiming at removing stationary and moving clutter while retaining precipitation for dual-polarization weather radar, a new clutter suppression method, named the object-orientated spectral polarimetric (OBSpol) filter, is put forward in this paper. Based on the spectral polarimetric feature and the range-Doppler continuity of precipitation, the OBSpol filter generates a filtering mask implemented on the raw range-Doppler spectrogram to mitigate the clutter and noise. The procedures are as follows. After the spectral polarimetric filtering and the mathematical morphology method, one binary mask where "1" indicates the precipitation is obtained. Then, the contiguous bins of the same value "1" in the range-Doppler domain are grouped in areas termed as objects. Whether the produced objects are precipitation or not will be further judged based on appropriate weather radar observable. Thus, combining all the filtered separate objects, a filtering mask can be obtained. In this paper, data collected by the polarimetric Doppler IRCTR drizzle radar are used to demonstrate and assess the performance of the proposed technique in the case of ground clutter, narrowband moving clutter, and noise. Two precipitation cases are examined: 1) moderate precipitation with large scale and 2) light precipitation with severe clutter contamination. In the range-Doppler spectrogram, both stationary and narrowband moving clutters are mitigated, while maintaining nonoverlapping precipitation signal. This helps to solve the problem when clutter and precipitation overlap in the time domain. In addition, the OBSpol filter is proven to be effective with different Doppler velocity resolutions. This technique can be applied in real-time due to its low computation complexity. Moreover, the spectral polarimetric filtering can be designed using the measurements of dual-polarization radar systems which do not have cross-polar measurements. Hence, the proposed clutter mitigation technique can be implemented for operational dual-polarization weather radar.

Clutter Mitigation Based on Spectral Depolarization Ratio for Dual-Polarization Weather Radars

“…3(b) and (c), precipitation exhibits the smallest values compared with WTC and/or noise. It is verified that the larger SNRs produce the smaller sLDR values for precipitation [25], [30]. The signature of sLDR for weak precipitation is similar to that of WTC in this case, however, the spatial continuity of precipitation can be used to design the filter to retain the precipitation and remove the WTC.…”

“…These artifacts not only affect the reflectivity, but also the Doppler and polarimetric measurements. The exact phenomena responsible for producing the artifacts in the radar data remain unknown at this point, but the artifact features have been exhaustively studied in [11], [24], [25], and artifacts can be referred to as narrow-band moving clutter. It is noted that such phenomena are not unique to IDRA, but also for the high-resolution polarimetric Doppler X-band radars such as MESEWI [26], and the Bonn X-band radar systems [27].…”

Quantitative Evaluation of Polarimetric Estimates from Scanning Weather Radars Using a Vertically Pointing Micro Rain Radar

“…In this event, the Z e profiles were interrupted by several dry periods while the melting layer was not well identified. Also, Z e values were found mainly between 10 and 30 dBZ in which the IDRA radar is sensitive to the noise level and interference from nearby telecommunication systems (Yin et al 2019). Such interference is characterized by the flashy appearance, in light rain or no rain, of multiple radials with constant Z e values that vary between 210 and 5 dBZ.…”

Clutter-Contaminated Signal Recovery in Spectral Domain for Polarimetric Weather Radar

“…P OLARIMETRIC Doppler weather radar is a very effective remote sensing tool for gaining insights into the dynamics and microphysics of precipitation [1]. The prerequisite of using weather radar data is sufficient measurement accuracy, which needs to calibrate the radar system [2] and remove the clutter [3], [4]. Spectral polarimetric signal processing also appears to be very promising for clutter identification and mitigation.…”

“…Spectral polarimetric signal processing also appears to be very promising for clutter identification and mitigation. Two examples of this are the recently proposed double spectral linear depolarization ratio (MDsLDR) filter [3] and the object-orientated spectral polarimetric (OBSpol) filter [4]. Compared with the MDsLDR filter, the OBSpol filter has the advantage that it does not require cross-polar measurements (often not available), thus making it more broadly applicable.…”