A new approach for analyzing and comparing TRMM PR and ground CP2 radar data for two convective storms near Brisbane, Australia

The TRMM precipitation radar (PR; Ku-band) uses a hybrid method based on Hitschfeld-Bordan and the Surface Reference Technique (SRT) to correct the measured reflectivity for attenuation and to derive the rain rate profile. The basic assumptions are a power law k=αZ_c^βand an assumption for the vertical model for the initial α values. The initial α is later adjusted by the factor ε to give α_adj. Even though DSD parameter retrievals were not part of the 2A25 products, the α-adjustment can be related to a constant Nw (normalized gamma intercept parameter) along the beam and varying Do.

The 2A25 vertical storm structure model in mixed phase precipitation is an important assumption along with the k-Z relations used. While the bright band structure model is based on a firmer microphysical footing, the convective structure model was presumably developed for ocean events and is less likely to apply to land events where updrafts are stronger along with mixed phase precipitation starting at temperatures as warm as 10 C (even in the sub-tropics) and extending to –20 C (by mixed phase we mean rain mixed with wet ice/graupel/hail). Existence of hail is not assumed in the storm structure model and the 2A25 algorithm in such cases (even assuming perfect PIA estimation) will tend to underestimate ε ; this in turn will cause the α-adjustment procedure to underestimate Nw and correspondingly overestimate Do leading generally to lower rain rates. More importantly, the k-Z relation strongly depends on the phase state of the particles; any mis-classification here will cause error in attenuation-correction and the resulting profile of corrected reflectivity.

The CP-2 dual-frequency radar (S/X-band) with new signal processor has been operational near Brisbane since late 2007 as a collaboration between the Bureau of Meteorology and NCAR. This unique radar measures traditional dual-polarization radar fields (e.g. Zh, Zdr, Φdp and so on) at S-band and additional Zh, LDR at X-band. Using the Dual Frequency Ratio (DFR; S/X-band) and applying iterative FIR technique, we can estimate the specific attenuation (k_h) at X-band. This specific attenuation is directly measured by radar, and only assumes that the S-band radar reflectivity is un-attenuated. The use of the CP2 radar data during TRMM overpasses allows us to evaluate several critical assumptions in the 2A25 PR products, namely the calculation of ε and the vertical phase state model especially for convective storms over land.

Our approach is as follows: after aligning and interpolating the CP2 and PR data to a common Cartesian grid, we use the X-band k_h from CP-2 and scale it to Ku-band according to the median volume diameter (Do) which is retrieved from CP-2 radar (using the Zdr data). With this ‘GR-based' Ku-band estimates of k, we re-do the attenuation correction of the measured PR reflectivity to arrive at a ‘GR-based' corrected Zc. This leads to a new ‘GR-based' PIA from which the ‘GR-based' ε is calculated. This leads to a ‘GR-based' retrieval of Nw and Do which is used to calculate a new rain rate profile which is compared to the 2A25 rain rates. We also compare the new rain rate profile with the CP2-based rain rate using S-band data (Zh,Zdr,Kdp)only. Using such an approach it is possible to evaluate some of the main assumptions in the 2A25 PR rain rate product.

In this study, we analyze two storms from Brisbane area for which we had both CP2 radar and TRMM PR coverage. The first one was the convective event of April 4, 2008 over the ocean, south east of the CP2 site. The second one was a convective storm event of November 5, 2008 over land west of the CP2 site (hilly terrain). In general, having a dual-wavelength, dual-polarized ground radar allows a much more detailed comparison with PR 2A25 products as opposed to conventional ground radars that offer only a comparison of the reflectivity profiles.