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 , 3:35 PM GMT on मार्च 16, 2017
Just a short blog about a few of my takeaways from the winter storm in the Northeast this week...I had no doubt that precipitation type would be an issue in some of the major cities, and I thought forecasters were aware of the difficulties with regard to predicting snowfall.
However, all the forecasts I saw were deterministic rather than probabilistic. And many forecasts were framed in the context of the European and GFS models (essentially, choosing the model of the day). I'm certain many forecasters looked at the ensembles, but I didn't see anybody frame the forecast in probabilities (probability of snowfall exceeding six or twelve inches, for example, or even probabilities of precipitation type). Instead, it was the same old song...although some forecasters talked about uncertainty, they refused to issue probabilistic forecasts and universally defaulted to deterministic maps of snowfall.
It was pretty clear early on that some major metropolitan areas might not receive all snow. And probabilities, especially early on, conveyed this uncertainty. Check out the forecast for probabilities of snowfall greater than 12 inches from WPC's super ensemble issued 00 UTC on Monday (Sunday evening) and ending 00 UTC Wednesday (Tuesday evening). And yet, the some media still ran with huge snowfalls in Philadelphia and New York City. In fairness, probabilities increased toward the onset of the storm, but there was still enough uncertainty for some forecasters to be a bit more cautious. That's what happens when media starts publishing hyperbolic numbers two to three days in advance of the storm. Yes, a swing for the fences, hoping that the pitcher throws a fast ball right down the middle of home plate. When are we going to learn that the atmosphere is a knuckle-ball pitcher?
A loop of successive SREF forecasts (36-, 30-, 24-, 18-, 12-, and 6-hour forecasts) for the probabilities of snow, all valid at 21 UTC on March 14, 2017. Courtesy of SPC.
Yes, I know. The public wants deterministic forecasts. Why do we, as a profession, always acquiesce to this demand? Like I always say...maybe this isn't too difficult to understand in light of this country never adopting the metric system.
Nothing ever seems to change these days, including some of the bad science used to convey the meteorology of the storm. For example, I saw water-vapor imagery being used to generally quantify the moisture feeding into the storm's circulation. For the millionth time, water vapor imagery cannot routinely detect moisture in the lower troposphere (below 700 mb), where most water vapor resides. If that's the message you want to convey, you should not use water vapor imagery; you should use charts of precipitable water (PWAT).
The loop of GFS model analyses of 500-mb heights (in meters) and 500-mb absolute vorticity (color-filled) from 00 UTC on March 14 to 12 UTC on March 15. Only vorticity values greater than or equal to 16 x 10-5 sec-1 are shown in order to emphasize vorticity maxima. Courtesy of Penn State..
After the storm, I read media blogs / discussions that claimed that the prominent northern and southern 500-mb short-wave troughs actually phased during the storm. In my opinion, such phasing did not occur, Check out (above) the loop of GFS model analyses of 500-mb heights and 500-mb absolute vorticity from 00 UTC on March 14 to 12 UTC on March 15. Note that I only present contours of absolute vorticity (color-filled) of 16 x 10-5 sec-1 or greater in order to emphasize the vorticity maxima.
I don't see any phasing...just a Fujiwhara near the end of the loop. Why am I raising such a fuss? Well, the lack of phasing probably resulted in the storm jogging a bit more westward, allowing warmer air to gain more ground inland and to knock snow totals down in places. Yes, you gotta know your science.
All in all, however, I believe forecasters had a pretty good handle on this storm. Their deterministic tenor of their message, however, likely gave some people the wrong impression about the lingering uncertainty of the storm.
As for some of the media's explanations after the storm (phasing, use of water-vapor imagery, etc.), I can only shake my head.
It's the same old song.
Many thanks to Jon Nese and Steve Seman of Penn State's Department of Meteorology for their helpful input.
The views of the author are his/her own and do not necessarily represent the position of The Weather Company or its parent, IBM.