Recent Heat Waves as a Special Subset

Reading by the pool? Underrated. Really, the end of summer is an opportune time for looking at recent heat events in a longer-term context, and thinking about what can be learned from using their differences as a window into the broader categorizations of heat waves.

2015, like 2014 before it, is on track to nominally be the overall warmest year in the observational record (due to incomplete spatial sampling, adjustment uncertainties for urbanization and elevation, etc, this can never be completely certain, so several other years have a nonzero probability of actually having been the warmest, all of them since 1998; in other words, of all the years since 1880, 2014 wins first place by a plurality). Of course, as can be seen in the map at the first link, there is always large spatial variability around the globe, so not everyone feels this warmth firsthand; indeed, it does not rule out 2015 being the coldest year at certain sites (chart is for Ithaca, NY, with the black line representing 2015 through Aug 28). 

But it does, naturally, predispose a greater severity (intensity+duration) of heat waves overall. A recent paper examined this truism with several models to see how projected warming would affect the severity of heat waves over the course of the 21st century in regions that have recently experienced record-setting heat events. One figure from that paper is shown below: using a custom-defined "heat-wave magnitude index" that is essentially the sum of consecutive hot maximum temperatures, a few K of mean warming translate to heat waves every year in the late 21st century much more severe than any observed up to this point. Many other studies using different definitions and models have come to similar conclusions. Some notes of caution are due, however. There are general irreducible uncertainties about the models -- though in this case, with averaging over large regions, these standard (CMIP5) models hew closely to observations when tested over the historical period. There are synoptic kinds of uncertainty, such as whether blocking (stalled pressure systems) turns out to behave somehow differently under similar governing conditions than it does at present. And there is the question of the representativeness of this particular heat-wave magnitude index for human-health impacts and ecosystem vulnerability. Minimum temperature has generally been found more important than the maximum, for instance, and impacts are always a combination of duration, intensity, and frequency in unknown proportion, all interacting with cultural and physiological factors. Interestingly, there has also been some indication that acclimatization on multi-year timescales has already occurred in France (in addition to the perennial seasonal kind discussed in an earlier post), suggesting that this may help to pull down the impacts from the ballooning predicted severity of future heat waves.

Projections for a maximum-temperature-defined "heat-wave magnitude index" over the globe and in several regions with well-studied record-setting events that were large, intense, and long. Red and blue in the observational record indicate different datasets; green, light blue, and gray in the future indicate the RCP 8.5, 4.5, and 2.6 climate-change scenarios. Source: Russo et al. 2014.

So what were the major stories of the summer, heat-wise? A blog post by Jeff Masters summarizing the events to date as of July 1 already had a lot to talk about. The India/Pakistan heat wave probably received the most attention, particularly centered around the death toll in Karachi, with similar stories from India as well (though less exacerbated there by the water-forgoing strictures of Ramadan). High temperatures but minus the large cost in lives continued to be observed in parts of western Europe in July and August, with July temperatures in Geneva rivaling those of the infamous August 2003 heat wave. 

Perhaps most memorably, the Iranian city of Bandar Mahshahr on the Persian Gulf saw its heat index (the version as defined by the National Weather Service) soar to 165 F/74 C, pushed along by an exceptionally high dew point of 90 F/32 C (for a modest-sounding relative humidity of 47%). A Weather Channel graphic from that evening is shown below. These conditions aren't even on the NWS heat-index chart; after all, temperatures almost always can only get that high with relative humidities below 20% (witness Baghdad in the figure) -- because with more moisture in the air, solar heat input goes more to heating water vapor than to raising the temperature. Whether or not the standard formulation of this heat index is a reasonable representation of the physiological stress under those exceptional conditions is debatable (it's not valid under 80 F either, and there were no reports of mass casualties); nevertheless, a purely maximum-temperature-oriented approach would not capture the surely suffocating experience of 115-dewpoint-of-90. These kinds of humid extremes are poorly understood but clearly result from a rare degree of interaction between hot air from the inland deserts and moist air from the Persian Gulf along a coastal strip. In Dhahran, eastern Saudi Arabia, location of a world-record 95 F dewpoint in 2003, summertime maximum temperatures have increased over the last few decades, as have minimum temperatures, although the changes are spatially variable within the Arabian Peninsula. Dewpoint temperatures have modestly decreased (but sharply increased just to the south; see Fig. 15 in last link). Trends contributing to humid extremes: inconclusive.

Weather Channel graphic from the evening of Jul 30, showing the combination of residually high temperatures with a band of exceptionally humid air along the immediate Persian Gulf coast.

Overall, summer 2015 did see an abnormally high intensity and spatial coverage of extreme heat relative to the observational record, though on longer timescales we can only say probabilistically that with lower mean temperatures in the past we expect there to have been fewer and/or less-severe heat waves. Don't even try to look for trends in extreme heat indices -- there are too many definitions and too few data to come anywhere near a clear conception of that. And for impacts, it is surely not the case that Europe 2003 and Russia 2010 actually stood out as the deadliest, or even deserve a spot on the all-time top 10; like with temperatures themselves, changes in definitions, preparedness, and awareness dominate over the environmental conditions. The more angles we look at heat waves from the fuzzier our simple mental conception of 'high temperatures' gets, but conversely the more accurate it becomes.