The Study About Numbers of Tornadoes That Didn't Count Tornadoes [see correction below]

As clearly stated below, this blog posting was based on Seth Borenstein's story about the Gensini and Brooks study. The story states (link below):

Because tornadoes sometimes go undercounted, especially in the past and in less populous areas, scientists don't like to study trends by using counts of tornadoes.

I have since heard from Dr. Gensini and I have read the paper itself. They do count number of tornadoes and the story is clearly incorrect. My objection as to the conclusions of the paper itself stand. They are:
  1. Back to Dr. Ted Fujita in the 1970's, there has been speculation about an east-west cycle regarding tornado maxima. It wish this study had referenced his work and commented on it.
  2. Forty years is too short a period of time in and of itself and, because, 
  3. From 1979 to 1996 (release of Twister) tornado chasing was restricted to the Great Plains which might have had the effect of inflating tornado numbers versus the areas farther east where tornado chasing was much more rare. 
  4. The STP (see below) is often not a good proxy for number of tornadoes. 
I have stricken out the incorrect information in the original posting. 

-- original posting --
Tornado near Tescott, Kansas, May, 2018
A reader brought to my attention and asked me to comment about this article pertaining to a study (based on a computer model) that alleges tornadoes are becoming less common in the Great Plains and more common farther east. However, they didn't actually count tornadoes to see if their model was producing realistic results.

"Because tornadoes sometimes go undercounted, especially in the past and in less populous areas, scientists don't like to study trends by using counts of tornadoes. Gensini and tornado scientist Harold Brooks of the National Severe Storms Lab looked at "significant tornado parameters," a measurement of the key ingredients of tornado conditions." 


Unfortunately, the "significant tornado parameter" is not a good proxy for the total number of tornadoes. For example, on a given day the "STP" can be high but because the parent thunderstorms are elevated, no tornado occurs. There is also the fact that a study of less than forty years is too short in climatological terms to draw any real conclusions.

This is yet another example of weather science being dominated to an unhealthy extent by computer models in lieu of doing the detective work needed to do fully validated and replicable science. We need more weather scientists in the mold of the late Dr. Ted Fujita and fewer computer models.  [Not a fair comment in this context]

Comments

Popular posts from this blog

Hilary's Forecast Path Shifts West; Updated 9:20am PDT

Dangerous Travel Conditions - People Reportedly Stranded

The East Coast Severe Weather Threat is Over