Hurricanes, Variability and Climate Change

Since this has come up repeatedly in the recent discussions of hurricanes and global climate change, I think it's worth revisiting the issue of extreme events and the extent to which it's reasonable to attribute them to global climate change.

David Appell articulates this well over at Technology Review:

The right way to describe these kind of events seems to be that they "are consistent with" the consequences of global climate change, because no storm outside of The Day After Tomorrow can definitely be ascribed to a warmer world. It’s a peculiar phrase, but the best scientists are likely to ever be able to do.

As has been discussed here and elsewhere, hurricanes are not a terribly good example - the connection between greenhouse climate change and more or stronger hurricanes is tenuous at best.

But there are situations in which it's reasonable to make strong assertions about the connection between a single event and the predictions made by the greenhouse warming hypothesis. David's right on the epistemological point - the best science can do is assert consistency with the hypothesis. But there are situations in which the consistency, even in the case of a single event or cluster of events - can get mighty strong.

Europe's 2003 heat wave is a good case in point. I've written about this before (and also here), but it's worth revisiting.

Most weather and climate phenomena tend to be normally distributed, along a typical bell-shaped curve with most cases falling near the middle and fewer cases out along the more extreme tails.

When you have an extreme event, it's useful to not only look at how much it differs from the norm (say, for example, how many degrees warmer than average it is), but where it fits along that curve. If it's so far outside the curve that it doesn't make sense, then you're in all likelihood onto something interesting in statistical terms. (Well, to be careful, the most likely thing is that you've got bad data, but once you've ruled that out then things get interesting.)

That's what Christoph Schär and colleagues did with the European summer of 2003, publishing their data in January in Nature (Nature \ 427, 332 - 336 (22 January 2004); doi:10.1038/nature02300). They concluded that the only way to make sense of this single extreme event - the hot summer of 2003 - was in the context of greenhouse climate models that predict not only increased warming, but also a substanial widening of the bell curve.

We find that an event like that of summer 2003 is statistically extremely unlikely, even when the observed warming is taken into account. We propose that a regime with an increased variability of temperatures (in addition to increases in mean temperature) may be able to account for summer 2003. To test this proposal, we simulate possible future European climate with a regional climate model in a scenario with increased atmospheric greenhouse-gas concentrations, and find that temperature variability increases by up to 100%, with maximum changes in central and eastern Europe.

In other words, it not only could be expected to get warmer, but also more events at either extreme - a wider standard distribution - should be expected.

That makes the summer of 2003 a great example of a single climatological event that's, to paraphrase David's careful wording, "frighteningly consistent with the consequences of global climate change."

Posted by John Fleck at September 04, 2004 10:08 AM