Is Global Warming significantly affecting atmospheric circulation extremes?

Although the anthropogenic influence on 20th century global warming is well established, the influence on the atmospheric circulation, especially on regional scales at which natural variability is relatively large, has proved harder to ascertain. And yet assertions are often made to this effect, especially in the media whenever an extreme warm or cold or dry or wet spell occurs and is tied to an apparent trend in the large-scale atmospheric circulation pattern. We are addressing this important issue using the longest currently available global atmospheric circulation dataset, an ensemble of 56 equally likely estimates of the atmospheric state within observational error bounds generated for every 6 hours from 1871 to the present in the 20th Century Reanalysis Project (20CR; Compo et al, QJRMS 2011). We previously presented evidence that long-term trends in the indices of several major modes of atmospheric circulation variability, including the North Atlantic Oscillation (NAO) and the tropical Pacific Walker Circulation (PWC), were weak or non-existent over the full period of record in the 20CR dataset. We have since investigated the possibility of a change in the probability density functions (PDFs) of the daily values of these indices, including changes in their tails, from the first to the second halves of the 20th century and found no statistically significant change. This was done taking into account the generally skewed and heavy-tailed character of these PDFs, and using both raw histograms and fitted “SGS” probability distributions (whose relevance in describing large-scale atmospheric variability was demonstrated in Sardeshmukh and Sura, J. Climate 2009) to assess the significance of any changes through extensive Monte Carlo simulations. We stress that without such an explicit accounting of departures from normal distributions, detection and attribution studies of changes in climate extremes may be seriously compromised and lead to wrong conclusions. Our finding of no significant change in the PDFs of the NAO and the PWC has important implications for how global warming is influencing atmospheric circulation variability and extreme anomaly statistics, and to what extent the CMIP5 models are correctly representing those influences.