A set of 4437 severe weather outbreaks from 19792010 is analyzed. North American Regional Reanalysis (NARR) data are used for evaluation and as input data in WRF mesoscale model simulations. Using a multivariate, linear-weighted index (developed in previous work) that ranks the severe weather outbreaks based on their perceived severity, the 4437 outbreaks are separated into training and testing datasets. Frequencies with which training cases exceed outbreak ranking index scores for a given magnitude of areal coverage are determined and evaluated on independent cases. Additionally, the frequencies are bootstrapped to assess their uncertainty as a function of areal coverage magnitude. Initial findings suggest that (1) the frequencies with which areal coverage magnitudes are associated with major severe weather outbreaks exceed 50%, despite the rare-events nature of major outbreaks, for most of the SWDVs tested, (2) the frequencies generalize well with independent outbreak cases, (3) the uncertainty increases if either the ranking index threshold used to identify major severe weather outbreaks or the areal coverage magnitude used to diagnose the outbreaks increases, owing to small sample size, and (4) frequencies derived from the WRF model simulations initialized at 0000 UTC on the nominal date of the outbreaks are similar to those derived from NARR data available at the outbreak valid time.