Forecasting of intense mesoscale bands of snow is one of the more difficult problems facing operational forecasters in the northern United States. Because of the extreme gradients of snowfall which normally occur with these bands, forecast errors in location of 50 to 100 km can result in snowfall forecast errors greater than 15 cm. While high resolution mesoscale models have allowed operational forecasters to anticipate these events, many of these models still show large errors in locating the heaviest snow. A case study will be used to show that the production of intense snow bands appears to be the result of a favorable interaction between the upper level (synoptic scale) wave and the mid-level (mesoscale) front.

By using a case with a heavy snow band from 14 March 2002, a methodology for forecasting banded precipitation will be proposed. This proposal follows the recommendation of Roebber et al. (2002) for convective weather – use of low resolution grids to examine the synoptic scale environment, the use of mesoscale resolution grids to examine the frontal forcing, and then high resolution grids to examine the distribution of QPF. It will be shown that for this event the synoptic scale forcing displayed little run-to-run or model-to-model variation. The higher resolution MesoEta consistently showed the development of an intense mesoscale band. However, the frontal scale forcing showed significant variation between different models and different times. As the event drew closer, it was observed that the model mesoscale forcing and QPF adjusted closer to the synoptic scale forcing.

The implications of this methodology indicate that forecasters will continue to need access to both low resolution and high resolution gridded data sets. The lower resolution data sets allow the forecasters to calculate derived quantities and quasi-geostrophic diagnostics in order to determine the synoptic scale dynamic environment. Higher resolution grids (< 20 km) are needed to examine the mode of the response within this large scale environment.