Microphysical Evolution Within the Comma Head of an Extratropical Cyclone

Forecasting snowfall accumulation is challenging due to limitations and uncertainties in the snow-liquid ratios and the model bulk microphysical parameterizations (BMPs). The source of these errors is often unknown, since there have been relatively few in situ observations of the microphysics (ice habit, degree of riming, and snow density) during Northeast U.S. winter storms. This study investigates the microphysical evolution and model validation within 15 winter storms observed at Stony Brook, NY (SBNY located on north shore of east-central Long Island) during the 2009-2010, 2010-2011, and 2011-2012 winter seasons. Surface microphysical measurements were taken every 15 to 30 minutes using a stereo microscope and camera, and snow depth and snow density were also recorded. During these storms, a vertically-pointing Ku-band radar was used to observe the vertical evolution of reflectivity and Doppler vertical velocities. A Particle Size and Velocity (PARSIVEL) disdrometer was also used to measure the surface size distribution and fall speeds of snow at SBNY.

This work has resulted in a unique conceptual model of the ice habit and riming spatial distributions at the surface within the cyclone comma head. Changes in the height of the maximum vertical motion relative to the favored growth temperatures led to changes in ice habit throughout the comma head and two heavy snow bands. Cold type ice habits with a few plates and dendrites were observed with light riming as the surface low was located along or east of the Mid-Atlantic coast. As the cyclones move northward towards Long Island, moderately rimed dendrites, plates, and needles are generally observed. Heavily rimed needles and graupel are often observed near the warm front and cyclone center. Mainly needles with light riming and a snow-liquid ratio from 8:1 to 9:1 are observed 2 to 4 hours before two heavy snow bands. With the strongest frontogenetical ascent during snow band maturity, moderately rimed dendrites were observed with snow-liquid ratios from 11:1 to 13:1. Lightly rimed plates and a snow-liquid ratio of 8:1 were observed after the heavy snow bands. The WSM6, MORR, THOM2, and SBU-YLIN BMPs in the Weather Research and Forecasting model at 1.33-km grid spacing were validated in this study. A non-spherical snow assumption (THOM2 and SBU-YLIN) simulated a more realistic distribution of reflectivity than spherical snow assumptions in the WSM6 and MORR schemes. In heavier riming, the Doppler velocity in the WSM6, THOM2, and MORR schemes were ~0.25 m/s too slow with the SBU-YLIN was 0.25 to 0.5 m/s too fast.