Following the recent (and very normal even for January) tornadoes in the southeast US, the media seems to be recycling the oft-repeated claim that human-induced global warming is increasing the threat of tornadoes, most recently, the tornadoes in Florida this past week (the Guardian):
In a dramatic illustration of stronger storms, at least 14 people were killed as severe thunderstorms and at least one tornado flattened homes and a church in Florida today.
Does the claim of a correlation (not to mention causation) between global warming and the threat posed by tornadoes carry any water? After the tragic May, 2003 outbreaks, Patrick Michaels, a climatologist who focuses his work on climate change, published a piece (Tornado Spin) directly disputing this claim. Michaels quoted an editorial cartoon that stated “These super powerful tornadoes are the kind of storm we’re likely to see more of with global climate change.” That cartoon was just one of many in a long string of unsubstantiated claims that deserves analysis.
Even more recently, Roger Edwards, a highly-respected meteorologist who focuses on severe weather and is formerly based out of the Storm Prediction Center wrote a guest post on Earth & Sky entitled Will Global Warming Cause More Tornadoes? His expert answer: we don’t know.
As of this writing, no scientific studies solidly relate climatic global temperature trends to tornadoes. I don’t expect any such results in the near future either, because tornadoes are too small, short-lived, hard to measure and count, and too dependent on day to day, even minute to minute weather conditions.
Edwards makes the case that the combination of many factors is working against any attempt to correlate tornado frequency with global warming. Among them, the dynamics of tornado formation (that are still not well understood), the scale differences (spacial and temporal) between average global temperatures and phenomena as minute as tornadoes is simply impossible to compare, and finally; tornado frequency data is tenuous, at best.
Before you can even attempt a thorough analysis of the correlation between global climate (specifically, temperatures) and tornado frequency, you must take stock of the available data. It is widely accepted that tornado frequency statistics data must be taken with a grain of salt. As the population of the country has grown, so has the reporting of tornadoes. Let’s face it - many of the tornadoes that likely went completely unreported (likely even unseen) in the heart of tornado alley (rural regions of the plains) are statistically much more likely to be spotted and reported in 2007 than they were in 1970 or 1930 or 1500. First, a look at the raw numbers:

If one hopes to complete a statistical analysis of the torando frequency data shown above, one must at least attempt to normalize the data for population metrics, much the way economic figures are often normalized for inflation or crime data is normalized for population changes over time. Of the numerous, ambiguous claims made in the mainstream media regarding tornadoes and global warming, never have I found any that claim to have normalized the tornado frequency data for population trends (Found one? Email me!). Without such normalization, how can the analysis be robust, how can it carry any validity?
Raw data on Tornadoes by Year 1950-1998 (SPC) does indeed show a statistically significant increase, but how much of that can be attributed to the increase in overall population? How many of the tornadoes reported in recent years would have only been spotted by a stray cow or windmill? How do we account for the undeniable impact of an increase in telephone service, cell phones (from under 10 million in 1990 to over 150 million in 2005!), media coverage, the ease of storm reporting (much of which is imperfect), the growing storm spotter network, computer-assisted reporting, not to mention the booming tornado chaser community, ad infinitum?
So maybe it isn’t tornado frequency that we can sound the alarms on, but tornado severity? Nope, that doesn’t work either. Even without normalizing for population changes (which would only act to lower the slope of any trend), (particularly, the frequency of violent (F3-F5) tornadoes) shows a statistically significant decline over recent decades (The full tornado climatology data is available through the NCDC Tornado Climatology Page):

So where does this leave us? Right back where we started: we simply cannot draw any trends between global warming and tornado frequency or severity. Edwards hit the nail on the head: There simply does not exist enough data to draw any scientifically sound conclusions. This does not play well with those who have an agenda to promote via global warming alarmism. To that end, I doubt Edwards’ expert opinion (which is bolstered by many others) will be disputed, if not cleanly ignored by the mainstream media, as exemplified in the opening quote of this post.
02/21/07: Crossposted to Newsbusters
