P21     Central U.S. WRF composite radar verification using MODE-Time Domain

 

Flemke, Jason, University of Saskatchewan

 

A 4-km convection-allowing Weather Research and Forecasting (WRF) model was ran for the contiguous United States from October 2000 to September 2013 (Liu et al. 2016). Downscaling to a higher resolution of 4-km permits the model to simulate deep convection without parameterization. Here, a verification of the model's simulated composite radar reflectivity in comparison to the Weather Surveillance Radar-1988 Dopplers (WSR-88Ds) national mosaic is performed using the Method for Object-Based Diagnostic Evaluation (MODE) with the included Time Domain (MODE-TD). The comparison focuses on the Central Plains of the U.S. for convective systems from March through August.
Using MODE-TD allows additional features of a storm systems lifetime to be analyzed such as initiation, development, and dissipation. MODE-TD compares each storm system identified based on the certain thresholds that are specified and compares different attributes such as object area, center, axis angle, and intensity. Pair attributes can also be defined such as area ratio, center distance, axis angle difference, and intensity ratio. Other biases that are examined include speed, shape, and storm track/duration. In a capsule summary of this model, it is described that the model is able to create realistic propagating convection, but the model underestimates the occurrence of storm systems in the central U.S. This study assesses the accuracy of the simulated composite radar data in the model.