Attenuation correction of X-band polarimetric Doppler weather radar signals: application to systems with high spatio-temporal resolution

Abstract

In the last decade the atmospheric science community has seen widespread and successful application of X-band dual-polarization weather radars for measuring precipitation in the lowest 2 km of the troposphere. These X-band radars have the advantage of a smaller footprint, lower cost, and improved detection of hydrometeors due to increased range resolution. In recent years, the hydrology community began incorporating these radars in novel applications to study the spatio-temporal variability of rainfall from precipitation measurements near the ground, over watersheds of interest. The University of Iowa mobile XPOL radar system is one of the first to be used as an X-band polarimetric radar network dedicated to hydrology studies. During the spring of 2013, the Iowa XPOL radars participated in NASA Global Precipitation Measurement's (GPM) first field campaign focused solely on hydrology studies, called the Iowa Flood Studies (IFloodS). Weather radars operating in the 3.2 cm (X-band) regime can suffer from severe attenuation, particularly in heavy convective storms. This has led to the development of sophisticated algorithms for X-band radars to correct the meteorological observables for attenuation. This is especially important for higher range resolution hydrology-specific X-band weather radars, where the attenuation correction aspect remains relatively unexamined. This research studies the problem of correcting for precipitation-induced attenuation in X-band polarimetric weather radars with high spatio-temporal resolution for hydrological applications. We also examine the variability in scattering simulations obtained from the drop spectra measured by two dimensional video disdrometers (2DVD) located in different climatic and geographical locations. The 2DVD simulations provide a ground truth for various relations (e.g., AH-KDP and AH-ADP) applied to our algorithms for estimating attenuation, and ultimately correcting for it to provide improved rain rates and hydrometeor identification. We developed a modified ZPHI attenuation correction algorithm, with a differential phase constraint, and tuned it for the high resolution IFloodS data obtained by the Iowa XPOL radars. Although this algorithm has good performance in pure rain events, it is difficult to fully correct for attenuation and differential attenuation near the melting layer where a mixed phase of rain and melting snow or graupel exists. To identify these regions, we propose an improved iterative FIR range filtering technique, as first presented by Hubbert and Bringi (1995), to better estimate the differential backscatter phase, δ, due to Mie scattering at X-band from mixed phase precipitation. In addition, we investigate dual-wavelength algorithms to directly estimate the α and β coefficients, of the AH = αKDP and ADP = βKDP relations, to obtain the path integrated attenuation due to rain and wet ice or snow in the region near the melting layer. We use data from the dual-wavelength, dual-polarization CSU-CHILL S-/X-band Doppler weather radar for analyzing the coefficients and compare their variability as a function of height, where the hydrometeors are expected to go through a microphysical transformation as they fall, starting as snow or graupel/hail then melting into rain or a rain-hail mixture. The S-band signal is un-attenuated and so forms a reference for estimating the X-band attenuation and differential attenuation. We present the ranges of the α and β coefficients in these varying precipitation regimes to help improve KDP-based attenuation correction algorithms at X-band as well as rain rate algorithms based on the derived AH.