Record high and low temps: An interesting trend

A fascinating paper about to be published in Geophysical Review Letters compares the number of record highs and lows at temperature stations across the U.S. since the 1940s. The authors found that we're getting more record highs and fewer record lows, in a pattern that yet again confirms that climatologists know what they're talking about. They also extrapolate that trend into the future, with some interesting results, but first let's deal with the past.

Gerald A. Meehl*, Claudia Tebaldi, Guy Walton, David Easterling and Larry McDaniel analysed millions of U.S. temperature records for the last six decades and counted the number of times one of some 1800 stations recorded a record high or low (to that point). As you might expect, if there was no long-term warming trend, the ratio between those two would be close to 1:1.

In that case, if you plot the number of record highs or lows on a graph, you'd get a line that falls gradually as time progresses toward an asymptote, which is mathematically expressed as 1/n, where n is the number of years in your data set. The longer the set, the less likely it is that any given year will break a record.

Of course, there is always a chance that a new record will be set, it's just that the chance of new record gets vanishingly small as time goes on. If there's no warming trend.

But Meehl and his colleagues did not find the number of records approaching an asymptote. Instead, over the last decade, for example, there were twice as many record highs as record lows. It turns out that we're getting few record lows because night-time temperatures are gradually getting warmer.


Meehl et al 2009

But here's what's really interesting, if the current trend toward more record highs holds:

For later in the 21st century, the model indicates that as warming continues (following the [IPCC business-as-usual] A1B scenario), the ratio of record highs to record lows will continue to increase, with values of about 20 to 1 by mid-century, and roughly 50 to 1 by late century.

And furthermore:

Presumably at some point after 2100 there would come a time when there would be no more record low minima being set, and only record high maxima would be recorded, though the model indicates that this has not yet occurred in the A1B scenario over the U.S. by 2100.

So although we can expect to get plenty more record highs over the next few decades, the natural variability of the climate means that even as late as 2100 there's a still a good chance of the odd record low, even while we approach global average temperatures several degrees above pre-industrial norms.

All of which means, we can probably continue to expect the hard-core pseudoskeptics to turn every instance of a record low temperature into more evidence global warming is a hoax for the rest of our lives.

And here's the lead author himself:

More like this

Ha! I tried to do this about ten years ago and could not achieve any reasonable degree of statistical significance. I suppose the last ten years have helped make that possible.

This is something I have wondered about for a long time. Here in Las Vegas, we haven't had a record low for about two years (I think) but regularly have record highs and especially record high low temps (you know, high night time temps.) However, this is generally presumed to be due to increasing urban heat island effects due to a rapidly growing city. Now that growth has essentially stopped, it will be interesting to see if the trend continues. I imagine it will, but it may be hard to keep the heat island effect separate from the global warming effect at least here.

If I were naive I'd think this should convince a lot of doubters: it's an easy to understand metric (no modeling), based on a lot of data points from robust thermometer dataset. But I'm not naive (anymore), so now I'm waiting for the first "heat island!" shouter.

Yes but is 60 years a long enough period to make the jump to predicting a trend and what the cause is ? I beleive that we are always going to be in a trend just like the stock market as soon as we know the direction of the trend then it will change at random. There are way too many factors to take into account.

The bar for the 1950's does not fit in the trend line. Is that an artifact caused by it being the first decade in the series? Or...?

The 1950s actually saw global cooling, probably due to increased industrial activity in the post-war period with all the associated aerosol-heavy emissions. So one would expect the trend to be reversed there. And after the Clean Air Act removed a lot of dirty emissions, global warming resumed in the 1970s, just as you'd expect.

A couple of years or so ago I got to thinking that with climate change we should be getting more new weather records than in the past when the climate was relatively stable. I turned on the TV and there it was, on the weather channel, showing the recent increase in records, exactly as I thought. They did not separate highs and lows.

As to the comment about aerosols, there was an interesting TV program titled "Global Dimming" about the cooling effects of air pollution. It was suggested that about 30% of global warming was mitigated by dimming, and that global warming would increase with cleaner air.

By Jim Thomerson (not verified) on 14 Nov 2009 #permalink

I always thought that such a plot would be very illustrative. We are still getting a few record low every year here in Calgary, but then that is to be expected since there are maybe only 120 years of data. We even manged to have a couple of record highs this fall even though it has been a cold year.

I also saw the Global Dimming documentary - the trend in the graph matches nicely with that hypothesis.

By Militant Agnostic (not verified) on 18 Nov 2009 #permalink

"And after the Clean Air Act removed a lot of dirty emissions, global warming resumed in the 1970s, just as you'd expect."

the solution is clear...repeal the clean air act.

bring back bell bottoms *and* cooler temperatures!!!