Wet Ice Specific Attenuation Retrieval and Correction for X-band using a Piece-wise Forward Correction Method

Wet ice, which is common in convective storms, can significantly contribute to attenuation in radar signals. Wet ice attenuation is usually observed at high frequencies such as 10GHz (X-band). Correcting for such signal degradation becomes very important when using X-band radars to observe atmospheric events such as severe convective storms. Current attenuation correction methods make use of the relationship between differential phase (ϕ_dp) and specific attenuation (A_h) in rain. Knowing that the specific attenuation A_h=αK_dp(r) (where K_dp is the specific differential phase and α is the coefficient) and K_dp=[ϕ_dp(r₂)-ϕ_dp(r₁)]/2(r₂-r₁) the attenuation due to rain can be corrected by simply using, Z_h(atten)=Z_h(un-atten)-α[ϕ_dp(r)-ϕ_dp(0)] .

However, this relationship breaks down when wet ice is present along the radar's path because it contributes little to the differential phase while highly contributing to reflectivity. This makes it difficult to correct for the contribution of wet ice to attenuation along a radar path.

A methodology to estimate and correct for such attenuation was presented before, where a variation of the TRMM Surface Reference Technique (SRT) was used. When using this technique, un-attenuated frequency reflectivity values (such as S-band) were assumed to be available at the end of each of the X-band radar beams. With this reference value the SRT-like technique was use to adjust the α value in the A_h=αZ^β relationship (with a starting value of 0.00048, while β was fixed to 0.6) to make the reflectivity values match at the end of the beam (ΔZ=Z_h(S-band)-Z_h(X-band)). The adjusted alpha α_adj was used to calculate an estimated specific attenuation to apportion backward the reflectivity values. This technique was first tested with a 2-moment scheme Supercell simulation using the Regional Atmospheric Modeling System (RAMS) Supercell model where the attenuation due to wet ice was up to 2.5dB km^-1. The technique was also applied to real X-band radar data such as that from International H20 Project (IHOP), the CASA (Collaborative and Adaptive Sensing of the Atmosphere) first generation system Integrative Project 1 (IP1) radar network in Oklahoma operating at X-band and the dual-frequency CP2 radar deployed in Brisbane, Australia. During IHOP, an X-band radar (XPOL) and S-band radar (SPOL) were available to apply the SRT-like technique. Analysis of data from June 16^th, 2002, where XPOL and SPOL data were matched, shows the wet ice attenuation was as large as 1.2dB km^-1. Over the X-band CASA radar network, un-attenuated reference data was provided by the NSSL Polarimetric WSR-88D (KOUN) radar. Data from a convective event on June 10^th, 2007 showed up to 5dB km^-1 of attenuation in areas of wet ice. Data from the dual-wavelength (S-band and X-band) CP2 system have also been studied. The SRT-like methodology was applied to intense thunderstorm event on March 26^th of 2008. The wet ice attenuation was found to be around 3.5dB km^-1. Comparisons of range profiles between un-attenuated S-band reflectivity and attenuation-corrected X-band reflectivity showed the SRT-like technique had reasonable results in all the analyzed datasets. However, this method requires a reference value (such as from S-band), which is not always available.

A new technique that does not require a reference is presented. First, rain induced attenuation in the X-band data is corrected using the A_h-K_dp relationship. Then the relationship coefficients (α and β) between the wet ice specific attenuation A_h and X-band radar reflectivity Z_h is found using a large CP2 dataset from the March 26^th event. Here the A_h due to wet ice was calculated by subtracting the rain specific attenuation (calculated from the K_dp(S-band)) from the total specific attenuation (which is estimated using the Dual Wavelength Ratio). To determine the locations where the wet ice is in each beam, a hydrometeor identification (HID) algorithm is used. Since the CP2 radar can only measure  and LDR at X-band, the S-band polarimetric information necessary to determine the locations of wet ice (such as Z_h, Z_dr, ρ_hv and K_dp) are used as a proxy. Once these location ranges of wet-ice are established, a forward Hitschfeld-Bordan (HB) method is applied to the rain attenuation corrected X-band profiles using , Z_h(corr)(r)=Z_h(un-corr)-A_HB(r), where A_HB(r)=1-qβ º(αZ^βdS), where q=0.2ln(10), the integral limits are determined by the HID results, and the coefficients from the extensive CP2 dataset. This way the HB method corrects for the excess attenuation caused by wet ice in a piece wise way. The attenuation corrected results are compared to those obtained with the SRT-like methods.

This new method intends to address the need to correct the excess attenuation induced by the wet ice without using a reference as presented before. X-band radars, such as the CASA radar network, might not have a reference available. This method can be applied to such systems when the polarimetric variables needed for the HID are available making it possible to correct such attenuation independently.