Uncertainty Analyses of the Operational WSR-88D Hydrometeor Classification Product

The United States National Weather Service (NWS) has recently implemented polarimetric upgrade to the weather surveillance radar – 1988 Doppler (WSR-88D) network. Polarimetirc radars can provide information about hydrometeors' sizes, shapes and orientations, and such information helps discriminating between different precipitation types such as rain, hail, and snow. A fuzzy logic based hydrometeor classification algorithm (HCA) developed at the National Severe Storms Laboratory was implemented in the upgraded radar network and provided operational HCA products. The objective of the current study is to analyze uncertainties associated with various hydrometeor categories in the HCA. The uncertainty information can be useful for HCA product users such as the NWS forecasters. The operational HCA utilizes all polarimetric variables and discerns ten difference classes of radar echoes. Some of these classes have similar polarimetric characteristics, which lead to uncertainty in the class assignment. Consistences between HC products from adjacent radars are analyzed through case studies, and reliabilities of the HC decisions are evaluated both qualitatively and quantitatively. A confidence index was developed for each HC decision through statistical analyses and the contributing factors included the quality of input polarimetric variables and the fuzzy logic aggregation scores in the HCA.