But given that tornadoes mostly happen in one part of the country precisely because of temperature and humidity patterns (both of which will almost certainly be affected by global climate change), isn’t it reasonable to assume that “tornado alley” might shift it’s location? And isn’t the most likely shift given a warming Atlantic a shift to the north and east? And given that population density increases as one moves from, say, Oklahoma to the coast, isn’t it safe to assume that, even if tornado intensity doesn’t increase, global climate change could result in a larger number of tornadoes having an effect on a greater number of people?

Source:
http://en.wikipedia.org/wiki/U.S._state_temperature_extremes

I haven’t claimed it, either.

I wasn’t arguing that. I was just telling you about an interesting and potentially relevant bit of extreme value statistics. I was thinking that it might suggest to readers some less obvious questions that need to be asked and answered first before deciding what you ought to believe. Any further deductions or conclusions you draw as a result of reading it are entirely your own.

Maybe you already knew about it, and had already answered the relevant questions for yourself? If so, then well done! Feel free to continue the conversation with further discussion of your findings.

I’m intrigued at your honesty/cleverness claims. Are you saying that you’ve found an error in the mathematics presented? Or did you have another reason for saying it?

While I don’t expect you to like the implications of the mathematics, I would still expect you to be both honest and clever enough to follow wherever they lead, even so, and not reject them as lies and stupidity just because you don’t like them or they don’t match your preconceived opinion. It’s a matter of civilised respect. Not mutual, sadly.

An anomalously large number of records are being broken, again, and you’re arguing that this happens all the time?

More records were broken in many other years, including many in the 1930s.

I’ve never thought you were honest, but usually you’ve been cleverer than this.

So, for a slight change of pace, let’s have some fun with extreme value statistics!

Extreme value statistics are the statistical properties of extreme values – i.e. maxima and minima. They behave somewhat differently to the means everybody is used to.

First, let us imagine we have a hundred years of temperature records for a thousand weather stations across the country on every day of the month (25 days so far). Now, taking the temperature on a given day each year for a hundred years, at just one station, we know that (at least) one record must be the maximum. Suppose it has an equal chance of being in any year of the century – then it has a one in a hundred chance of being the last year, the same as any other year. (Higher, if you allow for ties, but we’ll ignore that.) In other words, the probability of a record-breaker this year is 1% per weather station per day. We look at this weather station every day for most of a month, and supposing temperatures are evenly distributed, find the probability of a record sometime this month to be 1 – 0.99^25 = 22%. If we examine 1000 weather stations, we expect about 220 of them to be record-breakers some time this month.

Is that what you would have expected?
Are there any questions you now want to ask?

Remember, we’re assuming no pattern, no trend, no relationship between temperatures at different locations on different days in different years, and we still get over 200 all-time records! If there are trends and correlations (and weather stations built on black tarmac next to aircon vents beside the barbecue pit) would we not expect even more?

You can play a game of spot-the-journo-article that explains all that. There are thousands of articles about the heatwave, but how many of them dive into the arcana of extreme value statistics in such an educational way? Thanks to comments, this one does! Well done! I should make a note of it so that next time the weather is really cold, you can trot it out as counter to all those “coldest day, evah!” stories we’ll be sure to put up.

I mean, sure, we can reply by pointing out the distinction between climate and weather.

This common insinuation–that somehow, human-caused climate change is refuted by the perennial occurrence of bad winter weather–puts us scientific rationalists in a bind.

Yet those same “scientific rationalists” believe that the perennial occurrence of bad summer weather, is ideal proof of human-caused climate change?

To summarize:

March to September: Weather is Climate!

October to February: Weather is NOT Climate!

They’re addictive, aren’t they? After the last time you swear off them forever, and really really mean it. But then a really juicy one comes along, succulent and delicious, gleaming like a chocolate-frosted donut filled with crack cocaine, and you just can’t resist.

Please, don’t let me interupt you… as Napoleon once said.