Climate change: The tornado myth
Austin, June 5, 2014 — The 2014 tornado season started off slow, but is now in full swing. That makes this an opportune moment to challenge the climate science claim that “extreme weather” tornadoes are increasing in numbers and intensity – the “tornado myth”.
A new analysis of the most intense EF4/EF5 tornadoes strengthens the skeptic claim that “extreme weather” tornadoes are decreasing, not increasing as AGW theorists believe. Furthermore, matching the new results to the IPCC’s HadCRUT4 global land-sea temperature database further bolsters the skeptic position.
In the wacky world of climate science when empirical data contradicts an entrenched belief there are a number of immediate naysayer reactions:
- The bringer of bad news is biased or lying
- The data is tainted or cherry picked
- The selected data isn’t the right data to prove the point
Related to tornadoes, those reactions became painfully obvious in the comments section of a recent article by yours truly titled “Climate Change: Where theorists and skeptics agree and disagree” published May 30th.
It’s intended purpose is to clarify the areas of agreement and disagreement between human-caused global warming (AGW) theorists and skeptics. It was naively hoped that it might serve as a catalyst for informed, constructive dialog.
It includes 10 sets of diverse information used for illustrative purposes.
One of the illustrations, produced by the U.S. National Oceanic and Atmospheric Administration (NOAA), shows the occurrence of “strong to violent” tornadoes:
In the interest of full disclosure, the author added the guesstimated trend line. The guesstimate isn’t far off, as will soon become apparent.
The first allegation leveled against the above graph is that it is cherry picked data. It got selected for inclusion in the original article because it is the only one the author could find on NOAA’s website that tallies intense tornadoes by year.
That isn’t cherry picking. Cherry picking is when an author cuts out a small subset of data from a larger dataset that isn’t representative of the whole, and then uses it to make false claims.
What is an “extreme weather” tornado?
Often times, inspiration and progress come from the strangest places. The spirited debate over the above graph, driven by AGW criticisms, led to seeking further data on the subject.
The graph above includes all EF3, EF4 and EF5 tornadoes. This author didn’t give it a second thought since that same selection of extreme tornadoes has been applied elsewhere.
A skeptical AGW commenter disagreed with the selection. They suggested that including EF3 tornadoes diluted the results. According to them, EF3 tornadoes are not intense enough to be considered “extreme weather”. Therefore, by their reckoning, the chart is cherry picked, evidently by NOAA itself.
The skeptic defined an “extreme weather” tornado as EF4 or EF5 only.
The skeptic further suggested that if the author, me, refused to look at tallies of EF4/EF5 tornadoes that it would prove I was biased all along.
Interesting logic, a skeptic asking me to do data analysis for them to discredit a NOAA graph!
After much discussion it was agreed every EF4/EF5 tornado in the entire NOAA online database would be downloaded and tallied by year. That was to avoid cherry picking.
Then the results would be graphed. The graph would include a software generated trend line, something NOAA did not do in their graph.
As it turned out, downloading NOAA data was the hardest part. They have one of the flakiest download utilities around. It took probably 20 different selections, to successfully piecemeal together the data into one spreadsheet. With that amount of finagling a few tornadoes could have been missed.
There are 1,220 violent EF4/EF5 tornadoes in the selection. 2014 data isn’t included because the season isn’t completed yet. The depressing part is the download includes the number of deaths, injuries, and damages inflicted by each tornado.
For the record, the COUNTIF(Cell1:Cell1220,YYYY) command was used to do the adding for each year from 1950 through 2013. There was no human intervention.
As one might expect, excluding EF3 tornadoes from the tallies didn’t fundamentally alter the results. The most extreme EF4/EF5 tornadoes have been declining since 1950.
However, there is a NOAA record keeping technique that needs explanation. The biggest EF4/EF5 “extreme weather” tornadoes last longer and move further than normal ones do. Each time a tornado traverses from county-to-county or state-to-state NOAA enters it as a new tornado into its database specific to each locale. That artificially inflates the totals.
However, it has the unexpected benefit of acting as a barometer of the intensity of a tornado. That has a direct bearing on the claim there are more extreme tornadoes and they are more intense than in the past. After all, a short lived EF4 is less extreme than an hour’s long EF4 that traverses five counties and two states.
In terms of loss of life, the worst tornado of all time was in 1925 when 695 lives were lost to the infamous tri-state tornado. That is more than twice the death toll of the multi-state “super outbreak” of 1974 and another similar outbreak in 2011. 2011 was widely blamed on human-caused global warming in the media.
Extreme tornadoes vs. global warming
It’s fortuitous that NOAA’s dataset begins in 1950. That is the same year that the newest IPCC AR5 report says that human-caused global warming became the dominant driver of climate change.
According to AGW theorists, human-caused global warming is behind an increase in the numbers and intensity of violent tornadoes. The above graph shows when most extreme tornadoes struck after 1950, compared to earth’s warming period.
From 1950 through 1977, the 27-year period outlined in a blue shaded box, there were 725 extreme EF4/EF5 tornadoes. That period was marked by modest cooling. A famous 1974 Time Magazine article, published just two months after the biggest outbreak of EF4/EF5 tornadoes in recorded history, suggested we might be headed for another ice age!
In the remaining 35 years after 1977, during all the global warming, there were 485 tornadoes, according to NOAA records.
The two biggest outbreaks of tornadoes since 1950 were in 1974 and 2011. The 1974 outbreak, the largest on record, was at the end of a cooling phase. The 2011 outbreak, the largest since 1974, occurred 13 years into the current “hiatus” period where there has been no statistical warming.
The straw man argument
AGW non-scientist theorists submit this graph from “Tornadoes – 2012 Annual Review” as proof that tornadoes are increasing. It includes every tornado in the NOAA database. None have been excluded.
Unfortunately, this graph has a recording bias. NOAA explains:
This disparity between tornado records of the past and current records contributes a great deal of uncertainty regarding questions about the long-term behavior or patterns of tornado occurrence. Improved tornado observation practices have led to an increase in the number of reported weaker tornadoes, and in recent years EF-0 tornadoes have become more prevalent in the total number of reported tornadoes.
– NOAA, Historical Records and Trends
As a result, NOAA doesn’t include weak EF0 tornadoes in its official long-term historical graphs. It includes the following graph instead:
To minimize yet another recording bias, this graph starts with 1954 instead of 1950. The Severe Local Storms Forecasting Unit (SEL) and the SKYWARN volunteer spotter programs weren’t started until 1952.
Remove reporting bias, as NOAA did, and tornado increases disappear.
Unfortunately, because it includes all F1+ tornadoes, it doesn’t directly address the AGW theorist claim that “extreme weather” tornadoes are increasing in numbers and intensity. The earlier graphs above do that.
As fate would have it, a lively exchange of opposing viewpoints over a May 30th, 2014 climate story led to a closer look at extreme EF4/EF5 tornadoes.
In addition, comparing the results to IPCC temperature data suggests the opposite of the tornado myth may be true. Cooling trends, not warming ones, may cause increases in “extreme weather” tornadoes. The warmer the temperatures, the fewer EF4/EF5 tornadoes there are.
In the end, for tornadoes at least, the skeptic position holding that increases in “extreme weather” events is weakly supported by empirical evidence is vindicated.
Posted on Jun 5, 2014, in Climate, climate change, environment, Global Warming, Government, IPCC, news, Opinion, Politics, science. Bookmark the permalink. 2 Comments.
We made progress, I am glad. Let me nit-pick your article first, and then I will make my opinion. So here are the issues:
(1) Doppler radars did indeed help to locate tornadoes even when they do not touch down, so there is a bias there. But Doppler came in 1990s – and there should be an increase in the number from 1990s because of that. But if you look at all tornadoes from 1950 to 1990, you will still see an increase.
(2) I have no problem in using EF-1 + tornadoes as in the last figure, but please put a linear regression line there so we can find whether it is increasing or decreasing.
(3) We have every figures, except the most “strongest” tornadoes. If you plot the EF5 tornadoes with a regression line, then the pattern study would be complete. Without that, it is hard to find the real pattern.
Tornadoes come from the collision of warm moist air with colder air, neither cold or hot air by itself make a tornado. It needs both. So I am not sure why global warming should have any effect on that. If the entire globe is warming at a similar rate, that temperature difference may not change much. On the other hand, if the colder regions are warming more than hot regions, as AGW suggests, then the temperature difference on average will be smaller, which will result in fewer tornadoes.
“As you wish, m’lord” (Darth Vader-Star Wars)
I’ll do an F5/EF5 only graph… with pre-conditions…
Tally and graph all F5/EF5 tornadoes by year from NOAA database; include computer generated trend line. The trend line equation will be displayed.
The trend line (linear regression) equation will then be used to calculate the number of tornado occurrences for both 1950 and 2013 and the numbers provided. The percentage increase or decrease is calculated from those numbers.
Conditions 1 (acceptance):
The NOAA numbers are accepted without question. The generated trend line (linear regression) equation will be accepted, as is, without question.
Condition 2 (inconclusive):
Should the increase or decrease in the number of tornadoes be less than 50% higher or lower one way or the other, then that result will be considered inconclusive.
Condition 3 (level of sufficient evidence):
Should the linear regression equation show an increase or decrease of 50% or more one way or the other, then that is defined as a sufficient level of evidence to validate or invalidate the claim that “extreme weather” tornadoes are increasing.
Condition 4 (tornadoes increasing):
Should the linear regression show a 50% or more increase in the number F5 tornadoes then I, Steve Davidson, will be required to add a comment here and to the “Climate change: The tornado myth” article on the Community Digital News website endorsing the AGW position that “extreme tornadoes” are increasing!
Condition 5 (tornadoes decreasing):
Should the linear regression show a 50% or more decrease in the number F5 tornadoes then you, Susan Clipper, will be required to add a comment here and to the “Climate change: The tornado myth” article on the Community Digital News website endorsing the skeptic position that “extreme tornadoes” are decreasing!
Agreed, or no? You do seek the truth, right?
(I’ll also post this at the CDN webside where we’ve been discussing this subject)