Examination of False Alarm and Missed Mesoscale Convective Systems Predicted by the High-Resolution Rapid Refresh Model

Although in general the use of Convection-Allowing Models has improved the forecasting of convective systems, these models still often struggle to predict some Mesoscale Convective Systems (MCSs) well. The present study investigates the 06, 12, and 18 UTC runs of the High-Resolution Rapid Refresh (HRRR) Model to study occurrences of both false alarm MCSs and missed MCS predictions from June through September of 2018-2021. In a preliminary analysis comparing 40 observed MCSs to HRRR forecasts, it was found that 10, 12, and 16 of the 06, 12, and 18 UTC runs, respectively, generated little to no convection when MCSs were observed, and thus missed the systems. When analyzing HRRR model runs on days with zero observed MCSs, it was determined that there were many instances where the 06, 12, and 18 UTC runs all generated an MCS. To further examine this problem, five missed events and five false alarm MCS cases were selected. Each case was simulated using the Weather Research and Forecasting (WRF) model with various reanalysis products serving as Initial Conditions. The objective of this analysis is to determine mechanisms responsible for both modes of model failure and to understand what near-storm environmental conditions are most likely to be associated with the failure modes.