An examination of the performance of two high-resolution numerical models for forecasting extended snow bands during the DTC Winter Forecast Experiment

Two high-resolution numerical models were run daily on the CONUS scale for the 2005 winter season at the Developmental Testbed Center during the DTC Winter Forecast Experiment (DWFE). Two different dynamic cores of the WRF model were used, at a horizontal grid resolution of 5 km over the CONUS domain. The models were initialized at 00 UTC, with forecasts out to 48 h. The high resolution enabled the models to be run without using a convective parameterization scheme. This test of potential operational models of the future provided an opportunity to determine if some of the banded structures that are often a difficult part of the forecast problem for winter storms could be forecast directly by a numerical model. This paper will compare a number of events, contrasting the forecasts of the two models with observed snowfall for the various events, as well as with predictions from the current operational models.

A substantial sample of snow band cases were collected during the rather active DWFE period. While the Northeast had an exceptional winter, the focus of this paper includes events elsewhere across the CONUS, often typified by a relatively narrow band (or bands) of snow that extended for hundreds of km. Events were saved for storms east of the Rockies so as not to focus on smaller scale topographically induced snowfall, and additionally, we did not consider lake effect snow bands. Typically, the type of events that were saved were those associated with long-track "Alberta Clipper" type storms, or in other cases with considerably less organized surface lows but a well-defined jet streak and low to mid-level organized warm advection. Some of these more innocuous storms can be especially tricky to predict, and although they do not produce the hefty snows of a well-developed Nor'Easter, they can result in an extensive band of 1-4" type snowfall affecting several WFOs. Among the questions we will address are whether the DWFE models were any more accurate (or not) than the current operational model suite in predicting the location of the snow and snowfall amounts. For those cases with somewhat smaller scale embedded snow bands, were the high-resolution models able to resolve these structures? In addition, some of the fields typically examined to determine the potential for snow bands (for example, mid and upper level forcing mechanisms in the presence of small moist symmetric stability) will also be contrasted for the different events.