A recent paper in PNAS by Hansen et al. (there’s also a recently released discussion paper on the topic) has caused quite a stir. The essential result is that extreme heat (beyond the 2-sigma and even 3-sigma level) has become so much more commonplace, that the only plausible explanation is global warming.
In a sense, this paper doesn’t tell us much we didn’t already know. What it does accomplish is to show in practical terms the observable result of man-made global warming, which is not just to make the average temperature hotter, but to make extreme heat so much more common. What was once 3-sigma heat — which at any given time we would expect to cover less than 1% of the globe — is now at least 10 times more prevalent. There have been 3-sigma events before, it’s true, and because of that we know that such extremes have consequences. When they’re as rare as they should be, life can recover from those consequences. When they’re 10 times more common …
I have two criticisms of this paper. First, the exposition is not always as clear as it could be — but that’s a matter of style more than substance. Second, it gives the impression that variability of temperature has increased recently. I’m not convinced that’s the case when considering local temperature because part of the increased variability in “standardized” (i.e., scaled by the local standard deviation) temperature anomaly is due to spatial rather than temporal changes — different amounts of overall warming in different locations (i.e., different trends) — as I stated here. I admit I haven’t analyzed hemispheric data, nor did I (in the previous post) consider seasonal (principally summertime) temperature specifically.
But in another sense, temperature variability has increased precisely because of spatial as well as temporal variability. The point of Hansen et al. 2012 is that what used to be rare extreme heat is now much more common. Much. This is made even more true by the fact that some regions have warmed (trend-wise) more than the global or hemispheric average, so they’re even more susceptible to extreme events (“extreme” by the standard prior to 1980).
Even hot times in earlier years don’t stack up to what we’re seeing today. In their more recent discussion paper, they show standardized anomalies (i.e., anomalies divided by the local standard deviation) for summertime in the northern hemisphere, using a longer baseline period than in the original paper (in response to some critics). Here’s the color legend (units are standard deviations):
Here’s the map for the very hot summer (in the U.S.) 1936:
Note the strong heating over much of the American midwest, with a small region even showing 3-sigma (or more) heat (dark brown color). It was hot back then in the USA, but only 1% of the northern hemisphere was in the 3-sigma or more extreme range. Now look at what happened in the summer of 2010:
Not only is there a region of 3-sigma heat along the east coast of the USA and another along the north coast of South America, there’s a giant area from Russia down through the middle east. Fully 18% of the hemisphere is in the 3-sigma range. That’s not “natural variation.” It’s global warming.
This is, I believe, an important way to characterize the simple temperature effect of global warming because it puts it in the context of what we’ve seen before, of the conditions on which we have based building our modern civilization. The baseline period for their recent analysis is 1931 to 1980. That’s when we layed out the infrastructure which drives modern high-tech civilization. Those are the conditions from which we derived our expectations. But because of global warming, conditions today exceed expectation much more often than they used to. Much more often than we’re prepared to deal with. Much.
I did a similar analysis for the lower 48 states of the US only, using summertime-mean data for climate divisions from the National Climate Data Center. I compared the distribution of standardized (i.e., scaled by the local standard deviation) temperature anomalies prior to 1980, to those since 2000 (using 1930-1980 as a baseline period):
The most important aspect of this comparison isn’t the higher mean value for the most recent temperatures. It’s the fact that extreme high values — above 2-sigma and especially above 3-sigma — are so much more frequent. Much. And that’s the real problem. When a 3-sigma event happens, it’s a problem but we can deal with it and recover from it. When 10 (or more) times as many 3-sigma events happen … we have a problem.
That means we’re already in trouble. The really bad news is that we’re already in trouble from just the warming we’ve already experienced, but it’s going to get worse because it’s going to get hotter. You think the 2011 Texas-Oklahoma heat wave was bad? You think this year’s corn-belt heat wave was bad? You think the 2010 Russian heat wave was very very bad? You ain’t seen nothin’ yet.
That’s why Hansen et al. is so important. From a purely scientific perspective it doesn’t really add to our knowledge. But from a human perspective, it lays it on the line. We’ve had bad heat events in the past but now they’re so much more common they’re vastly more difficult to deal with, so stop kidding yourselves, it’s already a bad problem and it’s just gonna get worse.
All this reveals the utter foolishness of Cliff Mass’s distorted view that global warming has little to do with the extreme heat witnessed in recent years in many places. His argument is that global warming has raised temperature in the U.S. by about 1 degree F, but last year’s Texas-Oklahoma heat wave was 7 to 8 deg.F over large portions of TX and OK, so global warming is only responsible for a small portion of that heat wave.
Even if his result were correct (which it is not), he utterly misses the point. Rather than sum up the situation the wrong way as he does, Hansen et al. did it right, showing that global warming doesn’t just make heat waves hotter. What’s much much much more important is that it makes heat waves more frequent. Cliff Mass gives the impression that there’s nothing to worry about because our “3-sigma” events — the real killers — will only be one degree hotter, quite ignoring the fact that we’ll get 10 times as many of them.