Initial condition and forecast uncertainty for the 24 December 2009 blizzard

On 24 December 2009 a powerful and damaging blizzard struck central Oklahoma. This event was difficult to anticipate given the wide range of forecasts produced by numerical models. To test whether initial condition (IC) uncertainty played a role in the forecast uncertainty, two single-deterministic forecasts that are in every way identical except for the ICs, which are either from the North American Mesoscale (NAM) or Rapid-Update-Cycle (RUC) models, are performed. The position and intensity of synoptic-scale features, including the surface low and upper-level jet are different in the two forecasts, leading to significantly different precipitation patterns. The RUC IC has a slightly greater baroclinicity, but differences between the ICs are within the range of uncertainty used for perturbing ICs in many ensemble systems. To test whether an ensemble would show a similar IC sensitivity, the above is repeated using two 36-member ensembles with initial perturbations whose maxima are within the range of differences noted between the NAM and RUC. The results from these will be presented.