Radar bright band correction using the linear depolarization ratio

The enhanced radar return from the melting snowflakes in the radar bright band remains one of the major obstacles when trying to provide accurate rainfall estimates from radar. We present analysis of high resolution observations of the linear depolarisation ratio (LDR), which demonstrate that the bright band is associated with large and unambiguous LDR returns. These large LDR returns can be used firstly for identifying and flagging occasions when the radar reflectivity return is contaminated by the bright band, and, secondly the magnitude of LDR provides some information to correct the reflectivity so that it is more closely related to that in the rain below the bright band. Observations made with the 0.25deg beamwidth S-band Chilbolton radar show that the bright band has a depth of about 700m and is associated with an LDR of -15dB, whereas in the rain below and the ice above the bright band the values of LDR are typically more than 10dB lower. This factor of ten change in LDR is much larger than the changes in the co-polar correlation from about 0.98 in the rain to about 0.9 in the bright band. The bright band is absent in regions of embedded convection because the melting ice is in the form of graupel; such events can be identified because the value of LDR associated with melting graupel is 10dB lower than that for melting snow. Most dual-polarisation radars now transmit at 45degrees, so LDR is not available, and special scans are needed for LDR; however, because LDR is only needed to about 2dB this LDR scan can be extremely rapid. We will present analysis of high resolution LDR data, which have been degraded to the one degree beamwidth of operational radars and provide an estimate of the improved accuracy of rainfall retrievals expected when LDR and Z observations are combined.