Asks RP Sr’s paper in GRL (or rather, ask Thomas N. Chase, Klaus Wolter, Roger A. Pielke Sr and Ichtiaque Rasool). Interestingly, they conclude “not really”. This of course is contrary to what everyone knows, so their paper has been ignored, to RP’s annoyance. And if I had demonstrasted conclusively that a well known thing was wrong, and everyone just steamed ahead and ignored this inconventient fact, I’d be annoyed too. But has he indeed demonstrated this?
I thought I’d have a closer look at the data.
Their main result is that the fraction of the globe (or rather, the fraction of 22N to 80N, which is a fair restriction, for reasons they explain) that is exceptional (measured not in terms of direct anomaly, but in terms of exceeding the mean by a given standard deviation) is not particularly high in 2003. I can replicate that, see inset picture. The four panels are for exceeding 2, 2.5, 3, 3.5. In none of them is 2003 exceptional; interestingly, the 1991 anomaly was clearly small but intense because it takes over from 1998 as you tighten the SD criterion. Red are warm exceedance; blue bars are for cold. This pretty well replicates the relevant parts of RP’s fig 2 so I think my data+methods are probably correct.
[Click on the plot to get a larger version]
If/when I get round to it,it would be interesting to plot out what the anomalies look like in the high years.
So far so good. But those results are for the 1000-500 hPa thickness anomaly (at least *mine* are; RPs are the same, but re-scaled into a temperature anomaly), which is a measure of the atmospheric temperature anomaly integrated between 1000 hPa (the surface, more or less) and the mid-troposphere. But as we all know, we live at the surface. What happens if you do the same analysis with near-surface (1.5m) temperature? The picture this time is rather different. For moderate SD (2 and 2.5) 1998 is biggest; for more exceptional anomalies (3, 3.5) 2003 is clearly largest.
the previous plots go back to 1979; which makes sense for the trop because its when the MSU comes in. For the sfc in the NH you can arguably go back further; see here for 1948 to 2006, if interested. 2003 wins again.
So what does this mean (apart from a possible Comment to GRL?). It means that whether 2003 was wildly exceptional or not depends on how you look at it. Its widely publicised impacts were all in terms of the surface, so its perhaps unsurprising that when you measure something other than the surface, it becomes less exceptional.