Attribution of Extreme Events to Global Warming

The public, when contemplating global warming, includes those who question whether warming underlies the occurrence of extreme weather and climate events.  The U.S. National Academies of Science, Engineering and Medicine recently issued a report, “Attribution of Extreme Weather Events in the Context of Climate Change”.  It emphasizes that the appropriate format for questions concerning this issue can be phrased in terms such as “Are events of this severity becoming more or less likely because of climate change?” 

The report emphasizes that because of rapid advances in data analysis and climate modeling the severity of many extreme events can be attributed to contributing factors arising from warming.  It includes a schematic image, shown here, characterizing how the increasing degree of understanding of various types of events leads to increasing confidence in the degree to which we can attribute severity to contributions from global warming.

This post concludes with a selection of examples from the climate literature where such analyses have been made, leading to affirmative statements of attribution such as considered here.

Recently weather news has reported on extreme events in many regions of the world. 

• The American states of Texas and Louisiana, and their neighbors, had intense rain for over a week in early 2016, triggering record-breaking flooding in the area.    

• The American West has suffered severe drought conditions for four years, straining water supplies and lowering yields of important crops. 

• Over the past winter England and Wales have had prolonged bouts of heavy rain and flooding, including hundreds of flood alerts and warnings, more than 30 of which were labeled severe.  

• Perth, Australia had a heat wave four days long with temperatures over 40ºC (104ºF) in February 2016, and has had seven days over this temperature this season.  

• Pakistan has had increasing numbers of serious flooding events over the last century, including 40 events in the last 15 years.  Most of these are accompanied by loss of life and social dislocation, and significant damage to structures and agricultural lands.

These few anecdotes are potentially significant indicators of the effect that global warming is having at regional and local levels around the world.  Indeed, increased rates of occurrence of such phenomena are predicted with high confidence in recent global warming reports.  But in discussing the role of warming at the scale of such geographically small areas it is important to evaluate whether each one in fact is related to warming.  For example, if few or none are, climate science would be hard pressed to persuade policymakers and the public to undertake efforts to minimize further warming and adapt to its effects.  On the other hand, if it can be shown that there is indeed a contribution from warming, then climate science is justified in actively pursuing appropriate policies such as those agreed to at the United Nations conference in Paris in December 2015.

Here we discuss the question of whether, and how, extreme events can be attributed to global warming.

The U. S. National Academies of Sciences, Engineering, and Medicine issued the report “Attribution of Extreme Weather Events in the Context of Climate Change” (2016); Washington, DC: The National Academies Press (the Report) on March 10, 2016.  The Report emphasizes that extreme events, just as all our weather every day, results from the interactions of many atmospheric forces and processes.  For this reason, it makes the very important point that we can not justifiably ask a question such as “Was this extreme event caused by [man-made] climate change, yes or no?”  Such a simplified black-or-white question overlooks the complexity in climate and weather. 

Rather the Report offers different questions phrased as “Are events of this severity becoming more or less likely because of climate change?”; or, for single events such as local storms, the question could be “To what extent was the storm intensified, or its precipitation increased, because of climate change?”  Note that the questions deemed appropriate seek expressions of likelihood, rather than the certainty demanded by the yes-or-no question.  To any viewer of a televised weather forecast this should hardly be surprising.  After all, these daily predictions always provide likelihoods for sunshine (clear, partly cloudy, mostly cloudy, etc.) as well as percent probabilities of rain, sleet or snow.  In the same way, and for the same reasons, the Report concludes that statements of attribution should be phrased in terms of the probability or likelihood that global warming was a climatic factor contributing to the occurrence of the extreme event.

The Report summarizes its detailed deliberations about attribution of events to known or understandable contributing factors with a simplified conceptual diagram, shown here

 

Schematic diagram showing how our confidence (vertical axis) in attributing different kinds of extreme events to man-made climate change depends on our understanding of the extent to which climate change affects the particular type of event (horizontal axis).  Different types of extreme events are shown by the differently colored labeled circles.  The dashed line represents where a particular type of extreme event would fall if we had an ideal ability to attribute how climate change affects it.  In reality all the circles fall below the dashed line because our ability to attribute the specific events to man-made climate change is less than perfect.  Note that the understanding/attribution level is high for temperature events (upper right), and that these become lower as we pass through extreme rainfall (near the center) to severe convective storms (e.g., thunderstorms and tornadoes; lower left).

Source: U. S. National Academies of Sciences, Engineering, and Medicine,  “Attribution of Extreme Weather Events in the Context of Climate Change” (2016); Washington, DC: The National Academies Press. doi: 10.17226/21852.

 

 

The Report identifies two ways to label an event as “extreme”.  We have to be able to distinguish a supposed extreme event, potentially caused in part by the effects of global warming, from normal behavior.  First, data characterizing the event could be shown to differ from comparable data for a reference period extending over a sufficiently long period of time, with a high level of statistical significance.  Alternatively, data characterizing the event could be successfully modeled in a climate model that includes man-made global warming factors, whereas it could not be modeled by a similar model run excluding those factors, again, with a high level of statistical significance.  An event that fulfils either of these conditions can then be labeled as an event to which global warming contributes to its extreme character.
 

Extreme events had already been attributed to global warming before this Report. 

 

• Analysis of data over the 49 year period from 1951-1999 shows that global warming is responsible for extremes of rainfall over the Northern Hemisphere.  Min and coworkers (Nature, 2011, Vol. 470, pp. 378-381, doi:10.1038/nature09763) showed that “human-induced increases in greenhouse gases have contributed to the observed intensification of heavy precipitation events found over approximately two-thirds of … Northern Hemisphere land areas.”  This work is significant because it covers a period of time in which greenhouse gas levels, and global average temperatures, were much lower than at present.

 

• Extreme rainfall and catastrophic flooding in England and Wales in 2000 was very likely due to human-induced global warming.  Pall and coworkers carried out a probabilistic analysis of weather patterns and likelihood of flooding in the region (Nature, 2011, Vol. 470, pp. 382-385, doi:10.1038/nature09762). During October and November 2000 England and Wales had the heaviest rainfall since records began in 1766, leading to severe flooding. The authors concluded “it is very likely that global [human-induced] greenhouse gas emissions [occurring during the twentieth century] substantially increased the risk of flood occurrence…in autumn 2000.”

 

• Kelley and coworkers (Proc. Natl Acad. Sci., published online before print March 2, 2015) report on the worst drought in recorded history in Syria and neighboring countries just prior to the “Arab Spring”.  The drought was serious enough that large numbers of farmers left their villages and migrated to Syria’s cities.  This caused major social and political turmoil and is considered to be a contributing factor to Syria’s ongoing civil war.  The authors found that human-derived greenhouse gases contributed to the drought.

 

• Moore and Lobell (Proc. Natl. Acad. Sci., vol. 112 no. 9, pp. 2670–2675, 2015) analyzed changes in crop yields and climate change across Europe between 1989 and 2009.  Large scale decreases in yields were found in many localized regions, which correlated with increased temperatures and decreased precipitation over the 20 year period studied.

 

• Diffenbaugh and coworkers (Proc. Natl. Acad. Sci., published online before print March 2, 2015) examined the drought in California from December 2012 to September 2014, likely the worst in 1000 years.  By simulating the region’s climate in model calculations the authors found that the extra amount of greenhouse gases added by human activity likely resulted in higher temperatures and reduced precipitation in the region.  This factor also contributes to a high risk of continued severe droughts.

 

• Williams and coworkers also conducted detailed analyses of climate-related data for California, from 1901 to 2014 (Geophys. Res. Let. 2015; DOI: 10.1002/2015GL064924).  From rigorous statistical analysis the authors estimated that global warming was responsible for 8-27% of the observed excess drought conditions for 2012-2014, and for 5-18% for 2014 alone.  These findings indicate that although drought conditions may originate from various climatic factors operating cyclically over many years, its full extreme extent in the current drought has been worsened by global warming, producing the current record conditions.

 

• Cook and coworkers (Sci. Adv. 1, e1400082 (2015); published electronically 12 February 2015) assessed drought conditions in the American Southwest and Central Plains.  Assuming that unrestrained emission of greenhouse gases would continue, the risk of severe droughts in these regions is projected to be extremely high, by various measures between about 69% and 97% in the second half of this century.

 

Conclusion

 

The Report by the National Academies makes clear that climate science has advanced remarkably in the last decade, so that attributions of global warming as a contributing factor to extreme events can reliably be made after appropriate analysis.  This depends on use of one or both of the methods, either data based or model based, mentioned above.  The progress making this possible includes the availability of far more observational data, on the one hand, and constant improvements in climate models on the other.  As a result statements of attribution, when merited based on analysis, now authoritatively support the role of man-made global warming, due to humanity’s continued and increasing use of fossil fuels, in the increasing severity of various types of extreme climate and weather events.  We all, as citizens of countries around the world, must marshal our efforts to reduce emissions as aggressively as possible.  The goal should be achieving near zero annual emission rates within a few decades, say by mid-century.

 

© 2016 Henry Auer