Range dependent errors in the convective and stratiform partitioning of a radar precipitation estimation algorithm

Abstract

The Tropical Rainfall Measuring Mission (TRMM), designed to assist climate and atmospheric circulation modeling efforts by providing accurate information on tropical precipitation and latent-heat release, will require ground based validation of satellite derived rainfall totals and type classification. A key element in ground validation is the precipitation classification method authored by Steiner and Houze. The method partitions precipitation by type using a radar reflectivity threshold technique. Steiner and Houze report a lack of any significant range dependency, despite the normal problems associated with beam spreading and increasing altitude with range. This study will examine the Steiner and Houze classification method to quantify the effect of range dependent problems in discerning precipitation type. Radar data for this study were collected on January 10 and 18, 1993 during the Tropical Ocean Global Atmosphere (TOGA) Coupled Ocean-Atmosphere Response Experiment (COARE), by the National Oceanic and Atmospheric Administration (NOAA) N42RF WP-3D Orion aircraft's tail Doppler radar as well as the Massachusetts Institute of Technology's (MIT) Doppler radar deployed on board the United States research vessel Vickers. The data were chosen so the PW-3D tail radar data set maintained low ranges to precipitation of interest while the MIT radar data set possessed a wide variation of range. Possible error sources such as position error, radar calibration error, and temporal data collection differences, were minimized so that range dependency could be focused on. Range dependent beam spreading influenced precipitation classification over the largest span of ranges, and the greatest portion of the data. Also, beam spreading was found to increase the horizontal extent of precipitation areas in a direction perpendicular to the central axis of the radar beam. Overall, using the Steiner and Houze classification method, 24.0% and 21.6% of the precipitation with reflectivity greater than 15 dBZ was classified differently by the two radar data sets at the 1.5 km and 3.0 km analysis levels respectively. The percentage of total rainfall from all precipitation classified differently was 26% for the 1.5 km analysis level, and 28% for the 3.0 km level.