Tornado Alley Shift or Just Normal Variability

Updated 2/14/2020

A few weeks ago, a study was published saying that tornado alley - N TX thru S ND into E CO, W AR, SE WY & SW MO - is seeing fewer tornadoes and the climatological (30 yr) epicenter of tornadoes is shifting eastward. Several rumors/ innuendos/ stories/ and hyped-up statements have emerged. But I think several factors are at play here.

Errors?

The first thing we need to look at to dispute this is potenital for error. Thirty-year climatology for tornadoes for the 1960-1990, 1970-2000, 1980-2010 and soon-to-exist 1990-2020 periods will capture most tornadoes thanks to advancements in radar capabilities. Some tornadoes are still missed due to population density - cited by Dr. Nese & Dr. Forbes (both were at Penn State) in their 1998 article on PA tornado climatology, an intense study of PA tornadoes in 1996 found at least one tornado that would have been classified as straight-line wind damage. However, these missed tornadoes, even on a national scale, do not provide a huge error to the results (at least not as huge as before 1960 or especially before 1990). NWS now surveys the worst damage whether or not a tornado was sighted and pictures on social media also help NWS (the claims can be a problem). For instance:

Lingonier, MA in Aug. of 2015 was hit by a supercell and a very ominous cloud - sent through social media. The supecell had a hook on radar but a survey found no tornadic damage. Why didn't one form? Not enough moisture. I was watching the radar live from the AMS Mesoscale Conference in nearby Boston at the time alongside Howard Bluestein. Yep, the Boston native and Univ. of OK professor frequently cited in the news, professional journals and on video documentaries.

So, with errors out of the way, what is my problem?

Recent Data: Dixie Alley & 2011

The results of the study are 1985-2015 and there were several recent outbreaks in the Southeast states, known as Dixie Alley (think Civil War). Most notable is April 27th, 2011: Alabama was hardest hit. And 2012 was quiet overall but even more so in the Plains states - again, traditional tornado alley. Last year (no reflected in the report) was quiet too but unfortunately for 2-3 dozen people so far, this year is starting active. My condolences to the families.

This is a consideration.

Combining Errors & Recent Data

We also have to consider that the study didn't compare data in large enough increments - change in 25-yr climatology over time: '60-'85, '65-'90...'90-'15. With so much year-to-year variability, I conclude that their study is flawed. It is a trend worth watching.

Further: ENSO, Plains vs. Southeast, and Climate Change

Studies also found a connection between the epicenter and teleconnections. You'll frequently hear that term in seasonal forecasting or you may be familiar with PNA, AMO, SOI (a big factor this winter), QBO, MJO, AAO, NAO, AO, PDO and particularly ENSO (El Nino "Southern" Oscillation) - region 3.4 affects the USA. ENSO's has two phases: warmer than avg (El Nino) or cooler than avg (La Nina). These teleconnections tell us where the shift will occur: somewhat but not exactly where. Using both climate and NWP models, all the scenarios (back to the Plains, staying in the SE and shifting further north & east) do play out in the climate guidance - and it doesn't matter if the future is cooler or warmer (whether human-influenced or natural variability). Again, all scenarios still play out this year and in future years.

Preparation

There are two points I'd like to emphasize. In terms of tornadoes, they have occurred in the history of each of the lower 48 states because the USA has access to all the ingredients for tornadoes. Second, if you are not in tornado or Dixie alley (and even if you are in some of those areas), you are still prone to other disasters that everyone should be prepared for. That's all any of us can do is to be hope for the best and prepare for the possibility of the worst (whether imminent or not). There's no need to panic and it is irresponsible to cause panic when something isn't imminent (severe weather) nor predictable (exact time of major earthquakes for instance). The disaster kits I have you put together (and check when the time changes - just now BTW!) apply to any disaster to which your community is prone. And even if there is something predicted, moving calmly yet steadily will increase your chances of executing your plan (evac, shelter in place, etc.) successfully. Thank you for your time.

The NCEP Suite: An Update

Good afternoon and hope most of you East of the Rockies kept warm yesterday. We in Philly had our second big storm (first since Nov - second snowfall since last week) and it took the northernmost route possible. Now we look toward the weekend into next week for our next event. More info can be found on Facebook & Twitter toward the end of the week but a summary is posted on social media. I am writing for two reasons but really one big one: the shutdown.

First, to all the forecasters, great job forecasting these last events and issuing the watches and warnings. You have my deepest gratitude. The reason the NWS spends lots of time forecasting - for those that favor consolidation - is to prepare the warnings: most of us know that the NWS issues severe storm, tropical, flood and winter storm alerts but did you know they issue temperature-, marine-, aviation-, and wind-related alerts too? They are just as critical because all those types lead to lives threatened - exposed skin can become frostbitten and heat stroke requires hospitalization - while the others threaten both lives and property. It can be stressful when the models don't agree (Euro is doing well, GFS is ok but UKMET & CMC didn't do so well the last few days) and that's doesn't take into account the NWS's mission - add that and no paycheck and the stress levels go through the roof. Forecasting is not just monitoring the models but they can tell us what's coming in the next five days with great confidence and - if you follow them daily - sometimes seven or even ten days.

According to Capital Weather Gang (Washington Post's Weather Team), there is only one NCEP employee (no staff either, so thank you!) monitoring the models and while the government is shuttered, they are only authorized to prevent & fix crashes - like loss of data or frozen components. Any other changes are on hold, including the implementation of the 13-km GFS and constantly-running now-4D-VAR-EnKF-hybrid GFS Data Assimilation (initialization) System (GDAS). It was originally scheduled for last fiscal year but has been delayed to early FY2020. If the shutdown persists, it will be pushed back even further. In testing, the new GFS reduced error in TC Track and the skill score (unitless) at 500 mb/ 18,000 ft - a key layer of the atmosphere - improved by 0.01. Taking yesterday into account, that would have told us - earlier - that the storm would most likely turn NW over Arkansas.

Please, Mr. Trump, end the shutdown. These disasters need better predictions, especially hurricanes and tornadoes - it was a former hurricane (Sandy) that led to accelerating this effort. Oh yes, (1) this low produced a strong tornado in Alabama before coming up the mid-Atlantic and (2) damage and especially disruption can occur from any storm. Our predictability falls further behind Europe everyday and that is without the shutdown.

Good job to the forecasters. Good luck to the others once this fiasco ends.