What do you think of when you see posted in an article title, “There has been a tornado outbreak…”? Do you think of a set number of tornadoes, an idea of what a tornado outbreak is, supercells, a place in the world, etc? The list of things associated with tornado outbreaks can be expanded upon and discussed for a long time; however, that isn’t the point I am making. The term “Tornado Outbreak” actually has a legitimate, numerical definition. The Storm Prediction Center uses six tornadoes in a 24 hour timespan (between 7am and 7 am) as a benchmark for tornado outbreaks.
I can’t say that I find this number low, but I think defining tornado outbreaks by one set number is incredibly limiting. Six tornadoes, for much of the Great Plains, is too low; while in the Northeast, that is well above what it should be. There has been discussion about the ridiculousness of this recently, but no ideas have been thrown out about it changing. What do I think? I mentioned this last week, but I truly believe that tornado outbreaks should be defined by tornado climatology.
You know, why don’t we take into consideration region/climatology when describing a tornado event as an “outbreak”?
— Pierce Larkin (@tornadolarkin) June 1, 2016
Yes, tornado climatology would absolutely work. Whether it be on a state level, or on a regional level or whatever your favorite division of geography is, this would work. While it may seem that I am against defining this numerically, I absolutely am not. Meteorological events inherently need numerical benchmarks to be defined by, but that should not be a blanket number. I’ll give you some examples.
June 1, 2011 was a day in which 4 tornadoes formed in central Massachusetts, and wouldn’t be considered a tornado outbreak under normal circumstances. However, when you look at the tornado climatology of the state (on average, thy see one tornado per year), this is put into a different light.
This is an example of why this desperately needs to change, as this was a significant event, but isn’t technically defined a tornado outbreak in that state. Another example of where this is useful was during the May 10th severe weather event across the western half of KY recently.
There were seven tornadoes across the western half of the state that day, including an EF-3 that struck Marshall Country in far western KY. When the 21 is the annual tornado count, a single day of seven tornadoes is one that receives the “Outbreak” tag, but for the wrong reason. It receives it because of a set number, as opposed to a climatological anomaly.
The typical number of six gives us no rhyme or reasoning as to why that number is considered a tornado outbreak. This number isn’t backed up by climatology, nor is it backed up by logic. It includes tornado days that are way too small in states that have larger climatologies, and doesn’t include days that have less tornadoes in states with smaller climatologies. It just simply doesn’t make sense. That needs to be removed, and backed with climatology. Basing the outbreak on climatology gives us a solid numerical constant per state that can be conglomerated with the states in its region to get numbers for a regional outbreak. It also gives us an easy explanation as to why our numbers are the way they are.
So what percentage of climatology makes sense? It needs to be constant throughout the country and in each state. There is no other way to do this. In theory, we could up the percentage to a wildly high number like 50% of the yearly average in one day to keep hype from surrounding events several days out. However, that becomes way too high to justify.
Ideally, the number should be set at either 15% or 20% of the average tornado count. This allows for states like Texas or Kansas, who have high average numbers, to have a much better representation of outbreaks, and it gives a better comparison as to what is a major outbreak and what is just a regular outbreak. Say we used 15%. In Texas, that would be 23 tornadoes in one day. Here in Kentucky, that would be around three. This percentage is reasonable to have set, as these numbers aren’t easy to meet, but they aren’t absolutely wild either.
This could be used in regions as well. The number for regional outbreaks can be set at 15% as well. This can be separated using the Census defined regions, or some other way. No matter what, the 15% would be the constant. The SPC could take all of the average tornado counts for each of the states within that region and combine them. If the 15% threshold is met for that region, then it can be defined as a tornado outbreak.
I’ll give you an example. Adding up the average yearly counts from the East South Central region on the above map, you get 134 tornadoes on average. 15% of that is 20 tornadoes in one day. Climatologically, this is significant enough to classify as an outbreak, but isn’t too high to be considered unattainable.
Additionally, going by percent of yearly climatology allows us to adjust to climatological trends. This gives an adjustment to climate change and allows for trends to be picked out relatively easily.
Overall, I don’t know how realistic implementing this idea is. However, I think it makes logical sense for the SPC to do. This definition will allow for an easier benchmark to understand what an outbreak is. It also preserves the word “outbreak” for relatively extreme events, that are actually relative to something. Climatology gives all of this a base, and just makes sense.