Convective Addiction

12/8/09 Midwest Blizzard Case Study

Posted on 02. Feb, 2010 by Ryan in 12/8/09 Midwest Blizzard, Case Studies

Starting on December 8th and lasting through December 9th a major to historic winter storm impacted nearly all portions of the upper Midwest. Snowfall totals in the 10-15 inch range was reported in a large swath through the heart of Iowa, into Minnesota and Wisconsin. What made this event so dangerous to the general public were the intense winds that occurred during/followed the snow. The afternoon and evening hours of the 8th, the storm started to lay down snowfall with intense rates approaching 2 inches per hour in some spots. Des Moines and several stations in central Iowa reported visibilities at a quarter mile for several hours during the first intense banding of snowfall. In addition the dynamics of the storm created at least one report of thunder-snow in eastern Iowa during the event  (See Below)

*** If any of these images are too small, click on them to enlarge**

As the cyclone deepened as it progressed into the Great Lakes, the created very strong winds over the same areas that had seen the heaviest snowfall. Blizzard conditions were reported throughout the day over areas that had seen an snowfall from the previous days storm.

0205 PM     BLIZZARD         1 N BURLINGTON          40.82N 91.12W

12/09/2009                   DES MOINES         IA   TRAINED SPOTTER

            52 MPH RECORDED WIND GUSTS WITH BLIZZARD CONDITIONS IN

            BLOWING SNOW.

Synopsis:

An intense and well advertised cyclone ejected out of the western Rockies and into the Oklahoma panhandle during the morning of December 8th, already sub 1000mb at the surface. Pressure falls in the two to three millibar/hour range were recorded in central Missouri during the early evening hours and these continued into the Great Lakes,  signaling the storms intense deepening during this period.

Surface winds across much of the lower Mississippi valley had shifted to the southeast in response to this low pressure, creating warm-air advection precipitation across much of the area. Radar returns and METAR data in Iowa, Nebraska and northern Kansas showed precipitation in those areas were mainly snowfall. However, south of roughly a BRL to OTM to STJ line the precipitation quickly switched from a snow/sleet mix to pure liquid for much of the evening. (On the surface charts, the dark blue line denotes the 32 degree isotherm.

This system also had a lot of powerful dynamic’s simultaneously occurring in the upper portions of the atmosphere. Taking a look at the 250mb chart from the morning sounding on December 8th, you see a 155kt jet streak still rounding the four corners.

It wasn’t until a cold front sweep through the Eastern Missouri/Illinois region that much of those areas were able to switch over precipitation production from rain to snow. By that time the system had occluded and lost it source of rich moisture in the Gulf to lay down significant snowfall totals in most areas.

Forecasting the Event:

This was a complex forecast to say the least. Models showed signs of a significant winter storm to impact the mid-section of the United States several days out, but as always, the trick to this forecast lied in the details. A shift of 50-75 miles northward shifted the southern extent of snowfall in the Missouri/Illinois regions further northward, thus affecting the winter headlines issued by the National Weather Service. Specifically here in West Central Illinois, subtle changes ended up making significant differences in forecasted snow totals and verification. Below is a map of just how sharp the snowfall line on the southern edges of the storm was.

For the most part models did a fantastic job of preparing forecasters for this event, with evidence of a significant winter weather event as far out as one week from impact. But as the event drew close (within 48 hours) closer range and higher resolution models were able to pick out subtle details. Many of the longer range models attempted to hang onto the rather shallow artic near the surface and underestimated the amount of warm air advection that the system would generate. Compare the GFS and the NAM thirty hours from verification (Click on these images for higher quality):

Both are comparable on their two meter surface temperatures. But notice how much further north and west the NAM takes the surface low, despite not deepening the low as much as the GFS. At this point the NAM was the outlier amongst model ensembles and thus largely disregarded by forecasters, despite the models great track record with previous winter storms. As the event drew closer, the GFS did move back to the northwest to come more in line with the NAM and actually started to produce a more accurate low-level thermal profile of the actually event, by bringing warm air further into the system.

While we maybe talking about only a few grid points of change, it is very important as closer range models get a better handle of the event to take note of any change from previous long-range models runs.

While it is hard to adjust your snowfall totals and still call yourself accurate, doing now casting during an event can really allow you to reassess what is going on in the storm. For example take a look at two soundings from Topeka during the event, notice how much warmer the 800-600mb layers warmed in a rather short time frame. This goes to show how fast warm air can really be drawn into a powerful system. The first  sounding is from the morning of the 8th and the second is from the afternoon of the 8th.

So what did I learn in all of this?

-It is important during significant weather events to pay attention to models which have proven reliable in the past, even when they diverge from consensus.

-With such a powerful and dynamic winter weather system, don’t under estimate the amount of warm air that the system will draw northward. And that warm air is hard to be pushed south until a cold front move through or other strong forces.