Convective Addiction

Surprise New Years Eve Tornado Outbreak – Case Study

Posted on 03. Jan, 2011 by Ryan in Case Studies

Days after one of most impressive winter tornado outbreaks in recent memory we are beginning to grasp the severity and longevity of the New Years Eve Tornado Outbreak. At the time of this writing, Dec 31 was the 9th largest tornado outbreak of 2010 with 40 tornadoes. That ties April 22nd, 2010 which for many was the first big chase day of the year.From 6am to 7:15pm on December 31st, there was a tornado warning in effect somewhere in the United States. Impressive for any season.

Sadly, 7 people have died (As of Jan 3) as a result of these storms: 4 in Cincinnati, AR and 3 in Rolla, MO. There is no doubt in my mind that these storms would have killed many more people if not for the recent (relative term) developments in radar and warnings. The severe storms spawned at least 26 tornadoes according to the National Weather Service (NWS). The most significant damage was found in Cincinnati, Arkansas; Sunset Hills, Missouri; and Pearl, Mississippi.

EF-0: 9

EF-1: 7

EF-2: 4

EF-3: 5

(At least 1 more tornado is still under investigation near Roodhouse, IL by St. Louis NWS)

Below are damage pictures from various Tornadoes from the New Years Eve Outbreak:

Damage from Brandon, MS Tornado

Damage from the Cincinnati, AR Tornado

Damage from Sunset Hills, MO Tornado

CASE STUDY

Let’s take a look at what made this forecast so difficult, not just a few days out but even a few hours away.

Various models did a good job of showing the potential for severe weather and tornadoes. But variations in the models that were subtle but significant were enough to keep confidence low. While not shown, ensemble data from the SREF did continue to show low end capabilities of this storm producing tornadoes in the southern half of Missouri and south. Using the NAM and 4km WRF, you can see that both showed convection occurring Dec 31st and have it focused on the morning hours. But neither made forecasts think that the largest NYE tornado outbreak in at least 50 years was right around the corner.

Looking back, severe weather parameters from the 00z/30 NAM should have been a strong indicator to forecasters that tornadoes in close proximity to the surface low was a possibility. Most impressive was the helicity shown by the NAM. But even this was UNDERDONE compared to what actually transpired. A testament to how truly dynamic this system was.

18z NAM 0-1km Helicity

While helicity was forecasted to be strong…Instability was a large limiting factor (or was forecasted to be). While strong instability never materialized, it was also non-zero. More on the actual parameter evolution in a bit. One of the best case studies found on the subject of low-instability/high shear severe weather outbreak can be found here, by Jon Davies.

By using the above mentioned case study and RUC soundings, nowcasters would have been able to forecast that significant tornadoes were likely throughout the rest of the day on December 31st. Many of the near environment soundings from where significant tornadoes occurred later in the day had ‘bottom heavy’ CAPE profiles.

If a forecaster were to modify forecast soundings from various locations, they would have seen that CAPE profiles (although very weak) were very favorable in conjunction with high shear for tornadic supercells.

Modified 12z NAM STL Sounding,.Creates a bottom heavy sounding.

As an important side note — A big pat on the back needs to go to the NWS and their job of keeping on top of this quickly changing and evolving situation. Through the use of Mesocale Discussions they were able to not only keep the public aware of the severe weather threat on the warm side of the system but also the Blizzard conditions on the cold side.

Back to the tornado outbreak though…

As mentioned above, this was a high shear-low instability set-up. The amount of ‘turning winds’ in the low levels of the atmosphere may very well be the single most impressive parameter of this outbreak.

Very impressive couplets from the various storms show the strength of the low level meso-cyclones. Most of these couplets were only strong for 2-3 scans or even less as is often the case in high shear, low instability environmental storms.

Food for thought: Forecasting a tornado outbreak accurately and without worry of the ‘cry wolf’ syndrome can be a difficult task at any time of the year for a forecaster. But that request is magnified when the potential event is in the dead of winter. Understandably, many forecasters were not outwardly stating tornadoes were possible, much less an outbreak. Simply due to climatology and such a dynamic system, overall evolution of the storms were in question to the point that tornadic wording was not justified in medium range forecasts.

In the  near future I am not sure if it will be entirely possible to better forecast these type of events due to how rare they truly are. A saving grace may be the fact that the science of meteorology is advancing so quickly (it has only been a modern science since WWII). One can work and hope that by the time another one of these, twice a century, outbreaks occurs our understanding may be far ahead of Dec 30th forecasts.