Climate Change: Tornado myth update
Wikipedia: One of several tornadoes observed on May 3, 1999 in central Oklahoma
Austin, June 10, 2014 — A popular global warming myth persists saying that “extreme weather” tornadoes are increasing in numbers and intensity. The myth is fueled by news media hype. NOAA’s Storm Event Database disagrees.
The 2014 tornado season is on track to be the weakest on record, according to Dr. Harold Brooks, a long-time senior research analyst at the National Severe Storms Laboratory in Norman, Oklahoma. According to Brooks, this is the 3rd exceptionally weak year in a row.
NOAA records show, as of June 3rd, there have been 35 tornado fatalities in 2014. That compares to 55 in 2013, 70 in 2012 and a whopping 553 in 2011 by June 3rd in those years.
My best guess is that this is the slowest start since 1915, and maybe even 1900
-Dr. Harold Brooks, National Severe Storms Laboratory, 4/22/2014
This year is not just unusual, according to Dr. Brooks, it’s unprecedented. Through May 22nd the fewest number of U.S. tornadoes on record have been recorded.
From Aug. 7, 2013, through April 13, 2014, the Norman, Oklahoma NWS office set a new record for its longest tornado-warning-free stretch, at 316 days. The previous record was 293 days, set in 1991.
Just the facts
Returning to “extreme weather”, tallies of all major F3/F4/F5 and F4/F5 only tornadoes since 1950 show that intense tornadoes have decreased, not increased as the news media would have us believe.
The last bastion of hope for global warming theorists believing that “extreme weather” tornadoes are increasing, rides on the shoulders of the biggest, bad-boy tornado of them all, the frightening F5!
In the movie “Twister”, storm chaser Helen Hunt is the only researcher who’d ever seen an F5. It killed her father when she was a child; sucked him right out of their storm cellar while he valiantly struggling to hold the doors closed. That childhood tragedy drove her to become a tormented tornado hunter who risked her life seeking a way to increase tornado warning times by just a few minutes.
An F5 is the most terrifying tornado of all.
The F5/EF5 tornado record
NOAA lists deadly F5/EF5 tornadoes two ways. One way is a simple list of individual tornadoes. The other way is by tornado events from the storm events database. The difference is a tornado event occurs in only one county. If a tornado crosses county or state lines then it is entered as multiple tornado events, one for each crossing.
Combined, these two listings provide a more accurate assessment of the numbers and intensity of F5 tornadoes over time.
Author/NOAA data: Tornado event tallies shows relative strength of F5/EF5 tornadoes
NOAA’s complete record of F5/EF5 tornadoes (blue) and associated tornado events (orange) since 1950 is shown.
The number of tornado events above the number of individual tornadoes gauges the total energy release of the F5 tornadoes involved. The higher the number above, the stronger and more long-lasting the F5 tornadoes involved.
The above graph shows many more tornado events than tornadoes before 1975 compared to after. The trend line (linear regression) that accounts for the relative strength of tornadoes shows a 70 percent decrease since 1950, and there were zero F5s the first three years!
Some have suggested that 63 years isn’t long enough to establish a meaningful trend. They correctly suggest that a longer record with larger numbers of F5s would be more accurate. The pre-1950 work of Thomas Grazulis, documented in the July 1993 text “Significant Tornadoes 1680-1991: A Chronology and Analysis of Events” is cited as a source of additional data.
Additions or updates to the record are unlikely, according to Dr. Brooks. Brooks calls pre-1953 tornado records “sketchy”. Pre-1953 data is less reliable because the SELS program, precursor to the NOAA’s Storm Prediction Center, wasn’t started until 1952.
Brooks suggests that even NOAA’s records are unreliable. He says, “F2-F5 are overrated pre-75, F3-F5 probably underrated 2000-2007”.
According to Brooks, that means the official record of F5 tornadoes shows a steeper decline than it should because of inaccuracies in data collection.
Bottom line, though, no matter how you slice and dice the numbers, deadly “extreme weather” F5 tornadoes have not increased since 1950. If anything, they have decreased.
More about “extreme weather” tornadoes
There are other less direct ways to determine if severe tornadoes are increasing or not.
NOAA Storm Prediction Center: Tornado fatalities from 1875 to 1995
Fatalities caused by tornadoes are decreasing. That should not surprise anyone.
Weather forecasting, building construction and tornado warning systems are much improved. People have more time to seek shelter. You’d expect deaths to decrease whether tornadoes are becoming more intense or not. We are simply better prepared for them.
The impact of improved tornado warning systems is plainly seen in a change in the damage/fatality ratio after the SELS program started in 1952.
NOAA Storm Prediction Center: Levelized cost of tornado property damage by year from 1950-2011
This graph shows that property damage caused by tornadoes before 1980 was much higher than after 1980. Costs are levelized using three different methods so that an apples-to-apples comparisons can be made between the years.
A notable exception is the super outbreak of 2011 with very high property damage. However, since then 2012, 2013 and now 2014 are exceptionally weak tornado years with low property damages reported.
The graph comes from a scientific paper titled, “Normalized tornado damage in the United States: 1950–2011” published by Kevin M. Simmons, et. al, in the journal Environmental Hazards, 12/5/2012.
Again, part of the decrease in property damage over time is explained by better warning systems and more tornado-resistant building construction. Forewarned is forearmed.
Conclusions
NOAA’s yearly report of strong to violent tornadoes, which includes all F3/F4/F5 tornadoes, shows that damaging tornadoes have decreased since 1953. Remove the weaker F3 tornadoes from the tally and the results are still the same.
The last glimmer of hope for global warming proponents claiming that “extreme weather” tornadoes are increasing in numbers and intensity rested with the biggest of the big tornadoes, F5. But F5s show a 70 percent decrease since 1950, though the numbers aren’t solid according to Dr. Brooks, tornado researcher at the severe storms lab in Norman, Oklahoma.
Truth is, with only 138 tornado F5 events in the entire NOAA database, a trend line has low statistical value given the wide scatter in F5 yearly tallies.
Even combined with decreases in both fatalities and property damage over time, the numbers do not conclusively prove that “extreme weather” tornadoes are decreasing.
But what it does prove is that “extreme weather” tornadoes are not increasing! The tornado myth, hyped by the media and pushed by global warming theorists, is just that – a myth. It’s a myth that refuses to die.
——————————————————————————————————————————–
Update 6/14/2014: More tornado count data added to F5/EF5 counts above that further discredits the myth, commonly believed, that “extreme weather” tornadoes are increasing in numbers and intensity.
——————————————————————————————————————————–
Posted on Jun 10, 2014, in Business, Climate, climate change, economics, environment, Global Warming, Government, news, Politics, science and tagged tornado. Bookmark the permalink. 30 Comments.
Inform The Pundits! 
OK, great. Now you are being factual and impartial!!!!!
I’ve always been factual. Impartiality is a value judgement made by others.
I’m partial to standing up for underrepresented positions and will always speak out against misrepresentations of fact.
Btw… didn’t we have an agreement as to what you would do if the number of F5 tornadoes decreased by more than 50%? 😉 The actual decrease is 70%.
Btw… thanks for apparently agreeing with the main point of this article, that “extreme weather” tornadoes are not increasing as we are frequently told in the media.
In the previous ‘agreement and disagreement’ article that generated this discussion in the first place, one of the identified disagreements is on the occurrences of “extreme weather” events. Skeptics believe they are less frequent than do AGW theorists and they question a causal connection with human carbon dioxide emissions linked to them.
This evidence lends support to that skeptic position. Skeptics can’t be right agreeing with 90 percent of AGW theory, but then be 100 percent wrong everywhere else.
I owe you thanks. What you called “factual and impartial” is now even more factual and impartial. I updated this article.
I added the latest data you found and dovetailed it into the F5 events graph. Now that graph combines both F5 events and individual F5 tornado counts onto one graph.
It visually displays relative strength of F5s as well as their numbers.
Every time I included things you find, it bolster the skeptic position on tornadoes.
I will admit those things, but where are you getting your F5 tornado data? They are not comparable to the data given by NOAA in here: http://www.spc.noaa.gov/faq/tornado/f5torns.html For example, there are only five F5 tornadoes in 1953, you have seven or eight.
I identified the source… It’s NOAA’s storm events database:
http://www.ncdc.noaa.gov/stormevents/choosedates.jsp?statefips=-999%2CALL
As I explained, in that database if an F5 tornado crosses state or county lines, it is entered once for each crossing. It doesn’t link individual events to individual tornadoes. I even explained how that was a better measure for determining whether “Extreme tornadoes” are increasing in number and intensity.
Dr. Harold Brooks of the National Severe Storms Laboratory has left two comments on this same article in the Community Digital News version and he didn’t seem to have any problem with the F5 graph itself or my article conclusions. He may know more about tornadoes than anyone alive.
His main concern is the inaccuracy of tornado counts themselves in the first place. Because of that he took issue with my suggestion, made mostly for your benefit, that their records might be updated with pre-1950s data in the future. I only included that because NOAA documentation identified that it could be and sometimes is done.
When I plot with the correct data, there is a 27 % reduction.
I’m considering writing an article about the great lengths to which a global warming bigot will go to deny overwhelming evidence in order to maintain an untenable belief that even the experts at NOAA don’t hold true.
That article, if written, will be about you.
We had an agreement… I expect you to hold up your end of the bargain, but know you will not!
Why am I not surprised that your percentage is off… AND it is lower. lol!!!!!
Here is the “correct” graph with the “correct” numbers showing the “correct” percentage decrease… which you failed to get “correct”:
My formula and calculations are shown on the graph
Things to notice:
1-The F5 percentage decrease from 1950-2013 this time is 36.3%, lower than the 70% decrease shown in storm events database records.
2-This latest graph, combined with the storm event database records clearly indicates that tornadoes were longer lasting and traveled further in the past than they do now.
3-This new graph nicely supplements the storm events data graphed earlier.
Every time you come up with something new that you think invalidates previous conclusions, you only manage to bring in more evidence that strengthens the skeptic position.
Put into plain English… the often repeated claim that “extreme weather” tornadoes are increasing in numbers and intensity is a fabricated myth!
There are several mistakes in your analysis. (1) There was no F5 tornado in 1969, you have one in your plot ( the real data is here: http://www.spc.noaa.gov/faq/tornado/f5torns.html ) to bias the data in your favor, I assume you would not have made that mistake for say year 2005, (2) The regression analysis you have is WRONG, the correct regression formula is y = 23.724945120145 – 0.010989769573782*x when x is the year number. In 1953 that gives you 2.261925143 and in 2013 it gives 1.602538968, the difference is 0.659386175, the % change is 0.659386175*100/2.261925143 that gives 29 % (my earlier calculation was approximate, this is exact, but the answer is not that different) – your 49% reduction calculation is WRONG. [use some online sites like http://www.xuru.org to do the regression] (3) Any neutral person would tell that your way of double (or multiple) counting a tornado is the skeptics way of manipulating data – it works for you since there is a slight reduction in tornadoes and you want to amplify that. You won’t do that if tornadoes were slightly increasing. And what is so special about state boundaries, they are man made boundaries. So if one day Oklahoma, Kansas and Illinois or some other states decide to unite to form a new larger state, that makes a reduction in the effect of global warming on tornadoes? Or if a tornado occurs in small states like Delaware or northeastern side, that will have a big effect? Again, why don’t we count whenever it crosses a county, or a city? Why states are so special? Amazing the length to which so called skeptics will go to in order to mislead the public, they will make mistakes in data, they will make mistakes in data analysis (regression analysis), and they will make deliberately wrong counting of events in order to help their beliefs. I have shown examples of all three above. That is why the skeptics have no credibility.
You got it right… http://www.spc.noaa.gov/faq/tornado/f5torns.html is where the individual tornado data comes from, exactly as was identified as “Source 2” in the article graph above.
Now your graph is a lot better – instead of 49 % reduction you claimed earlier, now it is only 36.3 % reduction in your own calculation. You can see, just one tornado in 1969 made a 13 % change. If there are couple of tornadoes in 2014, the trend will be completely different. So there is no statistical significance in these data. I do not think we can get any pattern from these. I still think your regression analysis is wrong, the actual reduction from 1950 to 2013 is 31 % – I do not even understand why you are getting 1.22 and .66 at two ends – just qualitatively looking at the plot, those numbers are wrong. The equation is wrong – you have y(1950) = -0.0093*x + 1.22, that is not correct. It should be y(x) = ********. I think you are doing something wrong in Excel input. I will see whether I can post some figures, but the moderation may delay or prevent posting links.
I put my graphs in http://www.filedropper.com/globalwarming
Please look at the graphs in http://www.filedropper.com/globalwarming and the R^2 value since you brought it up. The linear fit is worthless, a 6th order polynomial is better, but not much. In any case, if you use a linear fit, the percentage reduction is less than 30 %, something like 27%. There is no pattern here.
You have GOT to be kidding about filedropper.com! But thanks for proving to me that I was right that the problem with your graph is that you failed to include all the years from 1950 to 2013. It is a timeline graph, you have to include all years, including the ones without any F5s. Please find someone to explain that to you.
Would you please, PLEASE get an independent third person who actually knows something about spreadsheets to help you.
I’ll send them my spreadsheet. Let them tell you it is correct. You are obviously incapable of figuring it out yourself.
I told you… I did not… I repeat… I DID NOT create the regression… Microsoft Excel did. Let an expert verify it for you.
If you knew the trends were worthless in the first place, then why the heck are you making a federal case out of it? You’re wasting time nit-picking it.
You already implicitly agreed long ago with the main theme of this article – that tornadoes are NOT increasing in number and frequency!!!!!
Oh… btw… bad news… the 36.3% is a simple graph title heading typo… the actual excel generated percent is 49.7%. That calculation, clear as the nose on your face, is also included in the spreadsheet I’ll send to your expert:
1 – ((1.2232 – (0.0093*63))/1.2232) = .47898954 = 47.9% (rounded)
Just to give you warm fuzzies over what to expect when I write about sea level rise…
Item 1 – Sea level rise is real, it will continue and it will not reverse
Item 2 – Claims of over-flooding of low-lying lands are greatly exaggerated
One of the many claims of AGW alarmists is that Pacific Islanders are losing their lands to man-caused global warming and that – GASP! – they will all be under water before we know it.
I found this 1941-era (73 years ago) map included in a project done at American University in Washington DC:
http://www1.american.edu/ted/ice/kiribati.htm
It is included in one of many case studies documenting the effects of global warming put out by the university:
http://www1.american.edu/ted/ice/kiribati.htm
Here are the GPS coordinates of that island:
1°25’9.40″N, 172°59’48.71″E
Look it up on Google Earth. Resize and compare what we see today with the 1941 map.
You will see the island has more land area overall today than 73 years ago, mostly at the southeastern elbow. No discernible land area has been lost. Keep in mind that this island was specifically selected as proof that global warming is causing Pacific atoll islands to sink into the ocean.
By my approximations, sea level rise in 73 years is between 5.6 inches and 8.5 inches, where the higher number represents the inclusion of that caused by global warming, assuming the IPCC is right.
Heck, those folks living in New York City, already ravaged by the effects of “extreme weather” sea level rise, should pack up and move right now before they all drown!
I am reposting this since the other one is waiting for moderation, since I added some websites as references. There are several mistakes in your analysis. (1) There was no F5 tornado in 1969, you have one in your plot to bias the data in your favor, I assume you would not have made that mistake for say year 2005, (2) The regression analysis you have is WRONG, the correct regression formula is y = 23.724945120145 – 0.010989769573782*x when x is the year number. In 1953 that gives you 2.261925143 and in 2013 it gives 1.602538968, the difference is 0.659386175, the % change is 0.659386175*100/2.261925143 that gives 29 % (my earlier calculation was approximate, this is exact, but the answer is not that different) – your 49% reduction calculation is WRONG. (3) Any neutral person would tell that your way of double (or multiple) counting a tornado is the skeptics way of manipulating data – it works for you since there is a slight reduction in tornadoes and you want to amplify that. You won’t do that if tornadoes were slightly increasing. And what is so special about state boundaries, they are man made boundaries. So if one day Oklahoma, Kansas and Illinois or some other states decide to unite to form a new larger state, that makes a reduction in the effect of global warming on tornadoes? Or if a tornado occurs in small states like Delaware or northeastern side, that will have a big effect? Again, why don’t we count whenever it crosses a county, or a city? Why states are so special? Amazing the length to which so called skeptics will go to in order to mislead the public, they will make mistakes in data, they will make mistakes in data analysis (regression analysis), and they will make deliberately wrong counting of events in order to help their beliefs. I have shown examples of all three above. That is why the skeptics have no credibility.
Thanks for your input. I’ll take your well intentioned critique under advisement.
—————————————————————————-
There aren’t any F5 tornadoes in 1969. I checked both databases. Where did you get that idea?
If your linear regression is for tornadoes and not tornado events then it doesn’t match the list you brought to my attention. If you show me a graph or share your data I can help you find the problem.
Do you remember when I suggested that you weren’t asking the right question? You still aren’t. I’ll give you a hint. Do have the vaguest notion what “R2” means?? I think not.
Btw… why would I need an online linear regression tool if I already have probably the most used tool in the world (MS-Excel) for doing that sort of thing?
The source where I got the F5 tornado data is given in my post at 10:12 AM, June 21. It is the NOAA F5 tornado summary. I did that regression with Excel and other tools, all gave the same result. I would hope you won’t prejudge people: on RSS (75.8 in this case) and R2 on measuring the deviation from the trend line, does not help much since the analysis is faulty to begin with, there is relation between years and tornadoes (if we plot with heating rate, or average global temperature, there may be some meaningful pattern). With a variation between 7 tornadoes and zero, I do not give much statistical significance on the result. But that is the trend line, and a 29 % reduction (from 1953 to now, or 30 % if you count from 1950) doesn’t give us any pattern. In 2011 there were 6 F5 tornadoes, and if somehow there are four or five tornadoes in 2014, the trend pattern will change significantly. I am putting the data I used below, you can copy it and paste and find the regression.
1953 5
1955 2
1956 1
1957 3
1958 1
1960 1
1964 2
1965 1
1966 3
1968 4
1970 1
1971 1
1973 1
1974 7
1976 3
1977 1
1982 1
1984 1
1985 1
1990 3
1991 1
1992 1
1996 1
1997 1
1998 2
1999 1
2007 1
2008 1
2011 6
2013 1
Ahhh… I see the problem…
I don’t have to cut and paste anything because I already have those exact same numbers in an Excel spreadsheet.
The difference is that I include all years… you don’t. That is why, I believe, your regression is off. Remember, it is a timeline plot, you have to include the years when there were no F5 tornadoes in order to be accurate.
Your listed data, including the zero years, is what I already used in Excel to draw the graph(s) and let Excel calculate the regression formula.
For your data, correctly including all years, Excel generates this regression formula:
y = -0.0093x + 1.2232
If you want to see for yourself, and have access to Excel or someone who has it, then I can send you the spreadsheet which has the numbers plus the Excel generated chart, trend line and regression formula. You are welcomed to nit-pick it to death.
You didn’t take the hint and find out what “R2” means did you? You should. It will put this whole F5 tornado trend line business into proper perspective. Basically, nit-picking the trend line is a scientifically meaningless exercise.
I should have said something about R2 earlier and saved you time and effort, but didn’t. Just for you, I put a big R2 value in the spreadsheet graph I can send, but don’t plan on doing anything further with it.
Good… I hope this means we are done playing silly games. I’m sure you have better things to do with your time.
One last attempt to get through to you, Susan.
Try to open your mind just a crack, you’ll be surprised how good it feels.
Plot this:
1950 0
1951 0
1952 0
1953 5
1954 0
1955 2
1956 1
1957 3
1958 1
1959 0
1960 1
1961 0
1962 0
1963 0
1964 2
1965 1
1966 3
1967 0
1968 4
1969 0
1970 1
1971 1
1972 0
1973 1
1974 7
1975 0
1976 3
1977 1
1978 0
1979 0
1980 0
1981 0
1982 1
1983 0
1984 1
1985 1
1986 0
1987 0
1988 0
1989 0
1990 3
1991 1
1992 1
1993 0
1994 0
1995 0
1996 1
1997 1
1998 2
1999 1
2000 0
2001 0
2002 0
2003 0
2004 0
2005 0
2006 0
2007 1
2008 1
2009 0
2010 0
2011 6
2012 0
2013 1
Thanks. But again, it is a linear regression analysis, and adding a lot of zeros instead of no data will not change the slope of the trend line by any reasonable amount. It is the slope that determines the rate at which the numbers are increasing or decreasing. Using your own regression formula, the slope is -0.0093, and the actual reduction in the number of F5 tornadoes in (2013-1950) years is (2013-1950)*(-0.0093) which is -0.5859. When there were 6 F5 tornadoes in 2011, this reduction is not very meaningful, in addition to R2 indicating the same.
Now, if you read my comments from the other blog post where you first posted comparing skeptics and proponents of global warming, you brought up the trend line saying that this is where those two groups disagree and that chart was an evidence to suggest that tornadoes are decreasing. If you read my comments there, I said that I have no reason to believe that scientifically there is any relation between global warming and increased tornadoes; may be the opposite is true. These are random events with few data points, which cannot be taken with any trends. So please don’t put that on me, with respect to R2.
Tornadoes are unpredictable – but sea level rise is a sure thing. And one can see how skeptics help to change scientific predictions with their hope. http://www.washingtonpost.com/business/economy/ncs-outer-banks-got-a-scary-forecast-about-climate-change-so/2014/06/24/0042cf96-f6f3-11e3-a3a5-42be35962a52_story.html?hpid=z6
I’ll take on the “extreme weather” threat of sea level rise in another article. I’ll read whatever references you suggest before I explain what is behind it in detail.
According to the IPCC AR5 report, under its absolute worst case scenario, RCP8.5, sea level will rise from 0.53–0.98M by 2100. That is 21-39 inches.
And that assumes global warming returns with a vengeance. It’s already 15 years behind the IPCC schedule. The State of North Carolina is assuming the worst and isn’t keeping up with current events.
Speaking of rising sea levels…
http://oregonstate.edu/ua/ncs/archives/2014/jun/study-links-greenland-ice-sheet-collapse-sea-level-rise-400000-years-ago
http://oregonstate.edu/ua/ncs/archives/2014/may/new-study-finds-antarctic-ice-sheet-unstable-end-last-ice-age
Sure, heating and cooling happened in the past and will happen again naturally. On a long term trend, we are still in the middle of a cooling trend that started three million years ago. But that is not the issue – 400,000 years ago we did not have such permanent structures and economic centers that could be threatened by sea level rise. Also, sea level rise is not just due to melting of polar ice or global warming. A lot of heat is absorbed by the ocean, and that will expand water causing the sea level to rise. Many scientists say that the absence of atmospheric warming is caused by water mass taking a lot of heat deep into the sea. That will have a significant effect on ocean levels.
I’m researching a series on the Clean Power Plan at the moment, but you’ll have ample opportunity to tell me how wrong I am about sea level rise when I have time to write about that topic. lol!!!!
With all those mistakes, I am thrilled to learn that you might write an article on me. I would think it should be on you. But thank you.
Thanks again for your input.