Retired senior lecturer in the Department of Meteorology at Penn State, where he was lead faculty for PSU's online certificate in forecasting.
By: Lee Grenci , 2:57 PM GMT on February 19, 2013
I have one last observation to make about NEMO before I move on, and it's a lulu. There were some 50+ dBZ reflectivities in Connecticut the evening of February 8, and, in my 40 years as a forecaster, I've never observed higher reflectivities associated with snow. For confirmation, check out the 0218Z base reflectivity from the radar at New York City / Upton, NY (KOKX), on February 8 below (larger image). Courtesy of NOAA.
The 0218Z image of base reflectivity from the radar at New York City / Upton, NY (KOKX) on February 9, 2013 (the evening of February 8). Larger image. The 50+ dBZ reflectivity of snow is the highest I've observed in my 40 years of forecasting.
Of course, I've seen high reflectivities associated with bright bands (wet snow near the top of a melting layer; stay tuned), so I'm referring exclusively to high reflectivities associated with snow that reaches the ground. Before I discuss Nemo and the storm features that led to such high reflectivities over Connecticut on the evening of February 8, 2013, I think a slight digression on bright banding is in order (I've heard too many erroneous explanations for relatively high reflectivity in wintry situations that never even consider bright banding as a possible cause). For the record, bright banding was discovered during military operations in World War II, and there were papers written in the 1940s that explained the underpinning science (example 1946 paper).
Here's how I explain bright banding. As falling snowflakes reach the top of the melting layer, where the temperature is 0 degrees Celsius, they began to melt (again, 0 degrees Celsius = 32 degrees Fahrenheit is NOT the "freezing mark"). Melting snowflakes soon become covered with a film of meltwater, and they look like large raindrops to the radar. Thus, the base reflectivity abruptly (and dramatically) increases. In rather quick fashion, however, snowflakes melt completely, shriveling in size as raindrops take shape. Since radar is very sensitive to particle size, reflectivity decreases rapidly once the melting process is complete. Moreover, raindrops now quickly accelerate earthward (raindrops fall faster than snowflakes). This rather abrupt acceleration after the water-covered snowflakes melt completely into raindrops decreases the number of radar targets just below the melting level (like the flow of traffic quickly opening up after cars accelerate away from a crowded toll booth on a super-highway). The combined effect of decreasing both the size and number of drops causes radar reflectivity to decrease just below the melting layer, creating a band of higher reflectivity (a bright band) above the ground.
To summarize my discussion on bright banding, check out this nifty flash animation of an idealized bright band, copyrighted by Penn State's online Certificate of Achievement in Weather Forecasting.
During the evening of February 8, 2013, the tops of the melting layer lay south of New England (check out the 02Z Rapid-Refresh model analysis of the height of the melting level, in meters. Note the bulge in the contours toward eastern Massachusetts, a sign of warmer air pressing northward on the eastern flank of the low-pressure system (02Z Rapid-Refresh model analysis of MSL isobars). Just to be sure, I grabbed the 02Z Rapid-Refresh model skew-T at New Haven, Connecticut, which shows the entire temperature sounding (in red) to the left of (lower than) 0 degrees Celsius, which I highlighted in yellow. So snowflakes were the dominant hydrometeor indicated by the 50+ dBZ shown on the 0218Z images of base reflectivity (near the top of this blog) and composite reflectivity.
The 0218Z image of Hydrometeor Classification from the radar at New York City / Upton, NY, on February 9, 2013. Note the red and dark pink, which indicate hail and graupel, respectively. To be fair, the Hydrometeor Classification algorithm identified some unknown precipitation in southern Connecticut (purple). Larger image. Courtesy of NOAA.
Reflectivity is much more sensitive to particle size than number, so I can only deduce that wet snow, which is more likely to form large aggregates (agglomerations) than dry snow, generated such large reflectivities. Like I said before, I don't think I've ever observed 50+ dBZ associated with snow. On the other hand, I've never heard of snowfall rates as high as 8 inches per hour that were sustained for two hours. The rarity of the reported snowfall rates seemed to fit the rarity of the radar data for this event. I think the radar reflectivities for snow were so large that the hydrometeor classification algorithm, a utility of dual-polarization radars, simply failed in this case. Indeed, the hydrometeor classification (HC) algorithm "concluded" that the heaviest snow had to be hail or graupel (ice). For confirmation, check out the 0218Z HC product (above; larger image) from the dual-polarization radar at KOKX). Presumably, the red (hail) and dark pink (graupel) shown on the 0218Z image of hydrometeor classification were erroneously selected because there must be a red flag in the HC algorithm that automatically eliminates snowflakes as the possible dominant hydrometeor when radar reflectivities are 40 to 50 dBZ or higher. In effect, these very high radar reflectivities likely "overruled" other dual-pol metrics (such as differential radar reflectivity) that probably indicated otherwise.
A quick look at the synoptic set-up during the evening of February 8, 2013, starts with the 02Z Rapid-Refresh model analysis of 850-mb heights, which showed a large gradient in 850-mb heights off the New England Coast. In response, a robust low-level jet stream, with wind speeds exceeding 80 knots (see the 02Z Rapid-Refresh model analysis of 850-mb isotachs and streamlines below; larger image), rapidly transported Atlantic moisture inland, setting the stage for unprecedented radar reflectivities and snowfall rates over Connecticut and other parts of New England.
The 02Z Rapid-Refresh model analysis of 850-mb isotachs (color-filled in knots) and 850-mb streamlines. Larger image. Courtesy of Penn State.
For me, this snowstorm will make my lifetime top-ten list. And I'm pretty old.
Comments will take a few seconds to appear.