Matrosov, S. Y., K. A. Clark, and D. E. Kingsmill, A Polarimetric Radar Approach to Identify Rain, Melting-Layer, and Snow Regions for Applying Corrections to Vertical Profiles of Reflectivity. J. Appl. Meteor. Clim., 46, 2, 154-166.
Abstract
This article describes polarimetric X-band radar-based quantitative precipitation estimations (QPE) under conditions of low freezing levels when, even at the lowest possible elevation angles, radar resolution volumes at longer ranges are in melting-layer or snow regions while it rains at the ground. A specifically adjusted vertical-profile-of-reflectivity (VPR) approach is introduced. The mean VPR is constructed based on the range–height indicator scans, and the effects of smoothing of brightband (BB) features with range are accounted for. A principal feature of the suggested QPE approach is the determination of the reflectivity BB boundaries and freezing-level heights on a beam-by-beam basis using the copolar correlation coefficient ρhv, which is routinely available from the X-band radar measurements. It is shown that this coefficient provides a robust discrimination among the regions of rain, melting hydrometeors, and snow. The freezing-level estimates made using ρhv were within 100Ð200 m from the corresponding estimates of the 0° isotherm heights from radiosonde soundings. The suggested VPR approach with the polarimetric determination of the reflectivity BB boundaries was used for QPE during the wintertime deployment of the NOAA X-band radar as part of the 2006 Hydrometeorological Test Bed (HMT-06) field experiment in the California Sierra Nevada foothills. It is shown that this approach noticeably improves radar-rainfall accumulation estimates. The use of the HMT-06 mean X-band reflectivity–rain-rate (Zeh–R) relation resulted in an approximately 65% relative standard deviation of radar estimates from the surface rain gauges if no VPR adjustments were made. Applying the VPR approach with polarimetric detection of the melting layer resulted in reduction of the corresponding relative standard deviation by about a factor of 2.