The estimation of X-band attenuation due to wet ice in the mixed phase region of convective storms and correction of LDR at X-band using the CP-2 radar

The CP-2 radar has been operational near Brisbane since late 2007 as a collaboration between the Bureau of Meteorology and NCAR yielding unique dual-wavelength and dual-polarized data in severe storms. The radar, with its new signal processor developed at NCAR, now measures the differential propagation phase and copolar correlation coefficient at S-band which prior was not possible with the CP-2 radar with its old processor. However, the measurement of the original fields have been maintained, i.e., dual-wavelength reflectivity ratio (S/X-bands), LDR at X-band, as well as Zdr, mean Doppler velocity and spectral width, the latter three fields all at S-band. Here, we make use of differential propagation phase and the dual-wavelength reflectivity ratio to estimate the specific attenuation in the mixed phase region of rapidly developing convective storms. The term ‘mixed phase' refers to the region of supercooled rain drops and graupel/hail in the ~ +5 to -10 C temperature zone where hail can grow rapidly via accretion prior to descent within the main precipitation downdraft.

The dual-wavelength reflectivity ratio has been used since the early ‘70s to detect hail larger than about 1 cm (Mie scattering at X-band) as well as to estimate the attenuation due to rain. Here we use an iterative FIR range filtering technique (developed for estimating Kdp) to estimate the total specific attenuation (Ah) at X-band due to rain and wet ice hydrometeors along the path. The rain component of this attenuation is estimated using the S-band Kdp data as the latter is not sensitive to the nearly isotropic wet ice hydrometeors. Hence, by simply differencing the total attenuation from that due to rain, the residual attenuation due to wet ice alone can be estimated, particularly in the mixed phase zone. The evolution with time of the total and wet ice attenuation in the ~ +5 to -10 C region should be correlated with storm intensity and hail production. Our iterative FIR range filter method also detects the Mie scattering at X-band due to hail (in the same way that the backscatter differential phase can be estimated from range profiles of differential phase). Prior radar simulations using the two-moment scheme of the Regional Atmospheric Modeling System (RAMS) of a supercell storm showed that the procedure for estimating attenuation due to wet hail in the mixed phase zone could be accurately determined even in the presence of random measurement errors in the reflectivities at S/X-bands and in the Kdp. However, data from CP-2 radar shows that such simulated accuracy is difficult to achieve because of large clutter power to signal ratio (CSR) as well as artifacts due to mis-matched sidelobes between the S and X-band antennas. Also, one has to deal with loss of the X-band signal after propagation through an intense precipitation core. These factors make it difficult to estimate the wet ice attenuation separately from the supercooled rain attenuation along single beams as this involves the difference between the total specific attenuation and that estimated from the Kdp. One way to overcome some of these problems is to do a spatial average over the storm core which reduces the error and gives a smoother vertical profile of the wet ice attenuation through the mixed phase region. We show examples of separately estimating the total, supercooled rain and wet ice attenuation from several severe convective storms. In addition, the other variables such as Zdr, HDR, Mie hail signal, and LDR give additional significant information on the vertical structure.

The X-band LDR suffers from one-way differential attenuation (Adp) between the H and V-polarizations which is apparent as a rapid increase in measured LDR along the beam in moderate-to-intense rain. The system LDR limit (vertically pointing in light rain) was determined to be between -30 to -32 dB. From theory Adp=Ah*g in rain but in practice assuming a constant value of g for the beam leads to either over or under correction of the measured LDR. Hence, we propose a scheme where we scale the total specific attenuation by a factor (g), this factor being the ratio of the difference in measured LDR (between the ‘end' and ‘begin' ranges along the beam) to the corresponding total PIA . Then we do a gate-by-gate correction of the measured LDR by assuming that Adp = Ah *g. We demonstrate the technique by showing vertical profiles of measured and corrected LDR especially when high values of LDR (e.g., due to hail) exist along the path. 129.82.229.2 on 5-1-2009-->