Paper Reviewed
Sardeshmukh, P.D., Compo, G.P. and Penland, C. 2015. Need for caution in interpreting extreme weather events. Journal of Climate 28: 9166-9187.

Introducing their work, Sardeshmukh et al. (2015) note there is great scientific and public interest in discerning the influence (if any) of global warming on extreme weather events, writing that "it is tempting to seek an anthropogenic component in any recent change in the statistics of extreme weather." What is more, they note that, for many people, "the occurrence of any extreme event not previously observed 'within living memory' or 'since records began' (in both cases, about 100 years) immediately becomes a candidate for attribution to global warming."

Sardeshmukh et al., however, are quick to caution that such efforts may well "lead to wrong conclusions if the distinctively skewed and heavy-tailed aspects of the probability distributions of daily weather anomalies are ignored or misrepresented." And it was against this backdrop that they began their study, which involved the development of a protocol to adequately detect and attribute changes in extreme weather events. Thereafter, they tested this protocol in an effort to assess changes "in the observed distributions of daily wintertime indices of large-scale atmospheric variability in the North Atlantic and North Pacific sectors over the period 1872-2011."

With respect to their protocol (see the original paper for details), the authors presented a series of mathematical and statistical procedures that ultimately produced, in their words, "a sharper tool for investigating the statistical significance of observed changes in extremes over the twentieth century and of projected changes over the twenty-first century." And in applying that protocol to two indices of atmospheric variability (North Atlantic Oscillation Index and North Pacific Index), Sardeshmukh et al. report they found "no significant changes either in the mean or in the entire probability density functions of these indices over the last 140 years" despite "an apparent upward trend in the NAO index and a downward trend in the NP index during much of the second half the twentieth century."

In discussing the significance of this finding, Sardeshmukh et al. say it "has important implications for understanding the atmospheric circulation response to global warming, and casts doubt on inferences about this response drawn in studies that focus only on the second half, or other subsets, of the full record."