Extended Abstract (192K)P5.6
A technique for automated selection of multiple Z-R relationships within a single domain
George Limpert, University of Nebraska, Lincoln, NE; and A. L. Houston
Estimating rainfall using radar observations is highly parameterized, employing Z-R relationships to calculate rainfall rate from reflectivity. Radar reflectivity depends on the refractive index of scatterers, the amount of scatterers, and the size of the scatterers. However, only a single quantity, reflectivity, is measured. To account for these additional unknown parameters, the scatterers are assumed to be liquid water and the drop size distribution is parameterized through use of a Z-R relationship.
A new technique is proposed to directly account for each of the three parameters. RUC model temperatures will be used to attempt to identify the type of hydrometeors that are observed by radar, thus accounting for potential variations in refractive index. Additionally, multiple Z-R relationships may be selected within a single domain to account for variations in drop size distribution. The first distinction is made between convective and stratiform precipitation and is based on the scale of precipitating structures, with stratiform precipitation occurring at relatively large scales and exhibiting weak horizontal reflectivity gradients. Convective rainfall is further classified as tropical or extratropical based on RUC model parameters that have been found to be statistically significant predictors of convective type using discriminant analysis.
Verification of the technique is done by comparing radar estimates to weighing-type rain gage measurements. Five minute data are used from gages deployed in the United States Climate Reporting Network. These data are chosen because weighing-type rain gages are not prone to some sources of error that affect tipping bucket gages. The performance of the technique employing multiple Z-R relationships will be compared against using only the Marshall-Palmer Z-R relationship or the WSR-88D convective Z-R relationship. A detailed description of the algorithm and preliminary results of applying the technique will be presented.
Poster Session 5, Novel Instrumentation and Data Processing Techniques Posters
Tuesday, 12 October 2010, 3:00 PM-4:30 PM, Grand Mesa Ballroom ABC
Previous paper Next paper