Evaluation of WRF Simulated Multi-bands over the Northeast U.S. Using Varied Initial Conditions and Physics

Mesoscale precipitation structures within Northeast U.S. winter storms result in heterogeneous spatial and temporal snowfall throughout the region during any one particular storm. There have been many studies of single-banded snowbands in the comma head, but fewer studies of multi-banded events. Multi-bands are defined as > 3 finescale (5–20 km width) bands with periodic spacing and similar spatial orientation, with intensities > 5 dBZ over the background reflectivity maintained for at least 1 h. While multi-bands have been observed to be more transient and shorter-lived structures than single bands, they are capable of producing similar snowfall rates and wind speeds.

The typical 3-4 km grid spacing used in some operational mesoscale models is thought to not to be enough horizontal resolution to resolve multi-bands. The Weather Research and Forecasting (WRF) mesoscale model is used to test the sensitivity of the fine-scale precipitation structures within Northeast U.S. winter storms to horizontal grid spacing, initial and lateral boundary conditions, and physics parameterization schemes including planetary boundary layer (PBL) and microphysical (MP). WRF temperature and moisture output is verified using radiosonde observations from Upton, NY (KOKX) and Chatham, MA (KCHH) and precipitation structures are verified using WSR-88D radars (KOKX, KDIX, KBOX).

The Northeast U.S. blizzard of 26-27 December 2010, also known as the “Boxing Day Storm,” was an exemplary case of multi-banding. Preliminary results show that multi-bands exhibiting linear structures and organized spacing are not resolved by the simulations with 1.33 km grid spacing. Further results from additional cases will be shown that suggest that the locations of the simulated primary band and multi-bands are most directly tied to the initial and lateral boundary conditions used, but the magnitude and duration is most directly tied to the microphysical parameterization scheme used. Additional simulations that test the effects of adding a < 500 m nest, doubling vertical levels, and adjusting forecast lead time will be discussed.