Operational Considerations of the December 26, 2004 Snowstorm across Hampton Roads Virginia and Northeast North Carolina

On 26 December 2004, low pressure systems moved northeast off the North Carolina coast, producing 6 to 14 inches of snow over Hampton Roads and Northeast North Carolina. Winter storm warnings were issued for much of the area with snowfall forecasts expected in the 4 to 6 inch range over interior areas, and mostly rain predicted for coastal areas. Temperature forecasts were critical for this event, with morning forecasts calling for temperatures to rise to the mid to upper 30s over the eastern part of the area. Instead, temperatures dropped back into the 20s inland, and into the lower 30s around the southern Chesapeake Bay. These lower temperatures, combined with the presence of mesoscale banding, produced a substantial snow storm for Hampton Roads and northeast North Carolina.

An analysis of the processes involved in the formation of the mesoscale snow bands will be applied in an operational setting. This will include examination of the potential for Conditional Symmetric Instability (CSI), as well as assessing frontogenesis. These parameters will be examined in operational models such as the ETA and GFS, as well as the local Workstation ETA. This analysis will show in the model runs, 36 to 48 hours prior to the snow beginning, the potential for a major snowfall.

In addition, model surface temperatures were initialized too warm for several days prior to the event. The models also indicated warmer air being advected into the area from the north, including southern New Jersey through the Chesapeake Bay during the event. We will examine operational methods to improve these temperature forecasts. These warmer temperature forecasts played a role in forecasting less snowfall for the area close to the southern Chesapeake Bay. Model initialization of surface temperature fields, particularly along marine areas, will be examined. Methods will be discussed and demonstrated to improve temperature forecasts along marine areas.

The models had two problems with this storm. One problem was an under forecast of precipitation amount caused by an absence of mesoscale banding in the models. The second problem dealt with too warm of boundary layer temperatures near the water. The purpose of this paper is to evaluate means of improving our forecasts in these types of situations, which are frequent over Hampton Roads and northeast North Carolina.