Rebuttal to Åström et al. Attributing mortality from extreme temperatures to climate change in Stockholm, Sweden., published in Nature Climate Change by Paul C. “Chip” Knappenberger, Patrick J. Michaels, and Anthony Watts
Last fall, the press pounced on the results of a new study that found that global climate change was leading to an increasing frequency of heat waves and resulting in greater heat-related mortality. Finally a scientific study showing that global warming is killing us after all! See all you climate change optimists have been wrong all along, human-caused global warming is a threat to our health and welfare.
Not so fast.
Upon closer inspection, it turns out that the authors of that study—which examined heat-related mortality in Stockholm, Sweden—failed to include the impacts of adaptation in their analysis as well as the possibility that some of the temperature rise which has taken place in Stockholm is not from “global” climate change but rather local and regional processes not related to human greenhouse gas emissions.
What the researchers Daniel Oustin Åström and colleagues left out of their original analysis, we (Chip Knappenberger, Pat Michaels, and Anthony Watts) factored in. And when we did so, we arrived at the distinct possibility that global warming led to a reduction in the rate of heat-related mortality in Stockholm.
Our findings have just been published (paywalled) in the scientific journal Nature Climate Change as a Comment on the original Oustin Åström paper (which was published in the same journal).
We were immediately skeptical because the original Oustin Åström results run contrary to a solid body of scientific evidence (including our own) that shows that heat-related mortality and the population’s sensitivity to heat waves was been declining in major cities across America and Europe as people take adaptive measures to protect themselves from the rising heat.
Contrarily, Oudin Åström reported that as a result of an increase in the number of heat waves occurring in Stockholm, more people died from extreme heat during the latter portion of the 20th century than would have had the climate of Stockholm been similar to what it was in the early part of the 20th century—a time during which fewer heat waves were recorded. The implication was that global warming from increasing human greenhouse gas emissions was killing people from increased heat.
But the variability in the climate of Stockholm is a product of much more than human greenhouse gas emissions. Variations in the natural patterns of regional-scale atmospheric circulation, such as the Atlantic Multidecadal Oscillation (AMO), as well as local impacts associated with urbanization and environmental changes in the direct vicinity of the thermometer are reflected in the city’s temperature history, and the original Oudin Åström et al. publication did not take this into account. This effect is potentially significant as Stockholm is one of Europe’s fastest growing cities.
But regardless of the cause, rising temperatures spur adaptation. Expanded use of air conditioning, biophysical changes, behavior modification, and community awareness programs are all examples of actions which take place to make us better protected from the dangers associated with heat waves. Additionally, better medical practices, building practices, etc. have further reduced heat-related stress and mortality over the years.
The net result is that as result of the combination of all the adaptive measures that have taken place over the course of the 20th century in Stockholm, on average people currently die in heat waves at a rate four times less than they did during the beginning of the 20th century. The effect of adaptation overwhelms the effect of an increase in the number of heat waves.
In fact, it is not a stretch to say that much of the adaptation has likely occurred because of an increased frequency of heat waves. As heat waves become more common, the better adapted to them the population becomes.
Our analysis highlights one of the often overlooked intricacies of the human response to climate change—the fact that the response to a changing climate can actually improve public health and welfare.
Which, by the way, is a completely different view than the one taken by the current Administration.
Knappenberger, P., Michaels, P., and A. Watts, 2014. Adaptation to extreme heat in Stockholm County, Sweden. Nature Climate Change, 4, 302-303.
Oudin Åström, D., Forsberg, B., Ebi, K. L. & Rocklöv, J., 2013. Attributing mortality from extreme temperatures to climate change in Stockholm, Sweden. Nature Climate Change, 3, 1050–1054.
Adaptation to extreme heat in Stockholm County, Sweden
Further detail by Anthony:
It should be noted that Nature Climate Change, which tends to be a fast track journal, took months to publish our correspondence, going through a longer than normal review process for such a short correspondence, and only did so along with a reply from Åström et al. Despite this uphill slog, we persevered.
Personally, I think the response from Åström et al. is ludicrous, especially this part:
“Our data indicate that there is no adaptation to heat extremes on a decadal basis or to the number of heat extremes occurring each year. “
Basically what they are saying is the people of Stockholm are too stupid to use an air conditioner or electric fan when it gets hot, and are incapable of any adaptation.
The other part of their response:
Our method of comparing the climate during two 30-year periods is valid for any two periods.
Well no, not really, and it is this flaw in their method that was a central point of our paper.
Variations in the natural patterns of regional-scale atmospheric circulation, such as the Atlantic multidecadal oscillation (AMO) as well as local impacts associated with urbanization and environmental changes in the direct vicinity of the thermometer are reflected in the city of Stockholm temperature history, and the original Åström et al. publication did not take this into account. By not looking at these factors, and by just taking the Stockholm temperature data at face value, assuming all of the heat extremes in it were “climate change” induced instead of being partially influenced by other effects, including the AMO and the city itself, allowed Åström et al. to become victims of their own confirmation bias.
For example, look at the GISTEMP record from Stockholm (which ends before 2000, not my fault). Note the 1900-1929 period.
Åström et al. compared two periods of Stockholm temperature data: 1900–1929 and 1980–2009, and used them as the basis for their entire paper. Here is their method from the abstract posted on the NIH website:
Methods: We collected daily temperature data for the period 1900-2009 and daily mortality data for the period 1980–2009 in Stockholm, Sweden. The relationship between extreme temperatures and all-cause mortality was investigated through time series modelling, adjusting for time trends. Attribution of mortality to climate change was calculated using the relative risks and baseline mortality during 1980-2009 and the number of excess extreme temperature events occurring in the last 30 years as compared to our baseline period 1900-1929. Results: Mortality from heat extremes doubled due to warming associated with climate change. The number of deaths attributable to climate change over the last 30 years due to excess heat extremes in Stockholm was estimated to be 323 (95% CI: 184, 465) compared with a reduction of 82 (95% CI: 43, 122) lives saved due to fewer cold extremes.
Only one problem, a big one, note that right after 1929 there was a big shift in the AMO data – what happens to the AMO in 1930 is essentially a “sea change”.
After 1930, the AMO was positive (warm phase) for over 30 years, went negative (cold phase) again around 1963-64, and stayed negative until a big uptick around 1998.
The AMO was primarily in its cold phase during the 1900–1929 period, and primarily in its warm phase during the 1980–2009 period — a difference likely to be responsible for some portion of the increase in extreme-heat events identified by Åström et al. and inappropriately attributed to global climate change. See Sutton and Dong 2012 for an explanation as to why the AMO affects the temperature record of Europe.
Then there were the changes/growth in the city itself, some movements and encroachments on the Stockholm observatory station, plus the fact that the mortality numbers they cited didn’t make sense when compared to other studies of trends in heat-related mortality across the United States and Europe which have reported declines in both total mortality and the sensitivity of urban populations to extreme heat,despite an increasing frequency of extreme-heat events.
Despite the long review, to the credit of Nature Climate Change, they recognized that we had a valid argument that mostly nullified the Åström et al. paper. Otherwise we’d never have gotten this published. Unfortunately, we can’t counter all the media hype from the original publication, but I hope readers will cite our rebuttal when appropriate.