Starting to Suss Out Connections Between Extreme Weather and Climate Change

by Bill Chameides | September 10th, 2013
posted by Erica Rowell (Editor)

Permalink | 7 comments
Hurricane Sandy was a massive "superstorm" that struck the U.S. eastern seaboard in late October 2012, breaking many a record. Was it linked to climate change? A new study looks at it and 11 other events and connects the climate dots where appropriate. (NOAA)
Hurricane Sandy was a massive "superstorm" that struck the U.S. eastern seaboard in late October 2012, breaking many a record. Was it linked to climate change? A new study looks at it and 11 other events and connects the climate dots where appropriate. (

Scientists start to meet a “grand challenge” and none too soon.

Whenever an extreme weather event occurs these days, the question almost inevitably asked is: Was it caused by global warming? For years, the less-than-satisfying scientific response went something like this: We don’t know; even though global warming will increase the likelihood of extreme events, directly linking a specific event to global warming is not possible.

What Had Been Impossible Approaches the Possible

But that was then and much has changed of late. While for some the words “not possible” mean don’t even bother trying, for others it is merely a challenge to go about proving the perceived impossible possible.

And so the World Climate Research Programme — a joint project of the World Meteorological Organization and the International Council for Science whose “main objectives … are to determine the predictability of climate and to determine the effect of human activities on climate” — has identified predicting extreme weather events and attributing them to climate change (or not) as one of six “grand challenges” in climate research.

In answer to that challenge, “Explaining Extreme Events of 2012 from a Climate Perspective,” the second [pdf] in a series of papers edited by Thomas Peterson of NOAA’s National Climatic Data Center and colleagues, was released by the Bulletin of the American Meteorological Society on September 5, 2013.  (A similar analysis of extreme events from 2011 was released last year.)

Connecting the Dots to Climate Change or Ruling Out a Connection in a Given Event

So how do scientists do the “impossible”? How do they establish a causal link between an extreme weather event and climate change? The methodology is probabalistic rather than deterministic, similar to the way weather predictions provide probabilities of rainfall (i.e., a 30 percent chance of rain) rather than “rain” or “no rain” absolutes.

In this case, the analyses determine how much more likely a given event is as a result of global warming. This is done using a model (most often a climate model) to simulate the extreme event for two cases: one with preindustrial conditions including preindustrial greenhouse gas concentrations and the other for current enhanced concentrations of greenhouse gases. Each case is run repeatedly, and the number of times the model produces an event like the one being studied is recorded. If the model with current conditions produces the event more often than that of preindustrial conditions, it is concluded that global warming made the event more likely and probably contributed to its occurrence.

Probably the first instance of this methodology was the study of the 2003 heat wave in Europe that led to the premature deaths of more than 20,000 people. Writing in the journal Nature, Peter Stott of the Hadley Centre for Climate Prediction and Research and co-authors concluded that “it is very likely (confidence level >90%) that human influence has at least doubled the risk of a heatwave exceeding” a mean summer temperature for 2003.

There is a caveat to all this: these analyses are based on computer model simulations and as a result are only as good as the models’ ability to simulate the processes that led to the extreme event in the first place. This probably does not represent a huge caution for events involving extremes in temperature, but I am a lot less sanguine when it comes to events involving more complex phenomena like heavy rainfall.

A Look at Some of 2012′s Most Extreme Events

The most recent study looked at 12 events analyzed by 18 different scientific teams (some events had multiple analyses). The choice of events studied was neither random nor pre-determined; it looks like scientists were encouraged to do analyses of events of their choosing from 2012 and submit them for review and publication. The ones that got included therefore reflect a combination of scientific interest and success in passing the peer-review bar.

Here’s a quick summary:

Event Finding Authors
1. Extreme Temperatures, Drought in the United States
1.1 Severe drought, central United States “[H]uman alteration … may have had little effect on the frequency of low-precipitation periods.” D. Rupp et al.
1.2 July heat, United States The “likelihood … is greater in the current forcing than in the preindustrial forcing.” N. Diffenbaugh and M. Scherer
1.3 Spring-summer heat, United States “The record hot spring …can be explained mainly by atmospheric dynamics.” Its connection to climate change “requires further research.” J.Cattiaux and P. Yiou
1.4 March-May heat, eastern United States “The anthropogenic contribution … leads to a factor of 12 increase in the risk of such an event.” T. Knutson et al.
2. Mid-Atlantic Coast
2.1 Hurricane Sandy “[W]e estimate … a one- to two-thirds decrease in Sandy-level event recurrences.” i.e., more frequent.* W. Sweet et al.
3. Arctic Sea Ice
3.1 Sea ice minimum, Arctic “Our modeling results indicate that the exceptional 2012 sea ice loss was primarily due to the sea ice memory [partly related to long-term warming] and to the positive feedback of the warm atmospheric conditions [partly related to long-term warming], both contributing approximately equally.” Guemas et al.
3.2 Sea ice minimum, Arctic “The … [Arctic sea ice extent] is extremely unlikely to have occurred due to internal climate variability alone, … and has a much greater likelihood of occurrence in the ‘forced plus internal variability’ scenario.” R. Zhang and T Knutson
4. Europe
4.1 Not enough ice in Holland for 11-City Tour despite frigid February temperatures “Snowfall on the thin ice that had just formed is shown to limit the ice growth more strongly than the effect of warming.” (See related post.) H. de Vries et al.
4.2 Extreme summer rain and drought, Europe “Both anthropogenic forcing and natural variability may have influenced [the event].” B. Dong et al.
4.3 Extreme summer rain in northwestern Europe “We find no evidence that declines in sea ice can explain these recent wet summers” and “the possibility remains that recent European summer precipitation anomalies are due to other drivers in the climate system rather than chance.” S. Tett
4.4 Summer rain, United Kingdom “[T]he model distributions studied suggest that any anthropogenic influence on these patterns was minimal.” S. Sparrow
4.5 Summer rain, northern Europe “This suggests a contribution of climate change to precipitation rate in northern Europe.” P. Yiou and J. Cattiaux
4.6 Winter drought 2011-2012, Iberian Peninsula “[T]he return period of such an extreme drought has decreased [i.e., the likelihood has increased] between the 1960s and 2000s decades. …[S]uch a trend has been partially driven by the anthropogenic emissions of greenhouse gases and aerosols.” R. Trigo et al.
5. Africa
5.1 Spring drought; eastern Kenya, southern Somalia “It seems likely that other factors, such as warming in the western Pacific … a recent transition in Pacific decadal variability.., and stronger western-to-central Pacific SST gradients may be contributing to the recent dryness.” C. Funk et al.
6. East Asia
6.1 Summer flooding, northern China “We are unable to confirm or reject the role of climate change in the 21–22 July 2012 rainfall event due to the inability of the models to accurately replicate observations in this region of China.” T. Zhou et al.
6.2 Summer rain, Japan “The extremeness of the Japanese heavy rainfall of 2012 was mainly caused by the oceanic natural variability and probabilistic atmospheric natural variability rather than by anthropogenic climate change.” Y. Imada et al.
7. Australia, New Zealand
7.1 2011-12 Extreme rain, southeastern Australia “[W]e detect limited evidence of a change in extreme rainfall … that may be attributed to anthropogenic climate change.” A. King et al.
7.2 Extreme rain March 2012, eastern Australia “[W]e find that the overall long-term effect of human influences on the climate increases the chances of above-average rainfall by 5%–15% (best estimate), although .… the extreme magnitude of the event appears to arise mainly from unforced internal climatic variations.” N. Christidis et al.
7.3 Extreme rain Dec. 2011; Golden Bay, New Zealand “This analysis indicates that the total moisture available for precipitation in the Golden Bay/Nelson extreme rainfall event of December 2011 was 1%–5% higher as a result of the emission of anthropogenic greenhouse gases.” S. Dean et al.


In their conclusion, Peterson et al. point out that the attribution business with the current methodology is a lot more complex than a pat yes or no answer. About half the papers found a global warming connection. Some, as I do, may find that to be a large percentage — it’s a lot harder to establish a positive result than a negative one. Perhaps more telling is the fact that in some cases different teams looked at the same event (such as the extreme summer rainfall in Northern Europe) and reached differing conclusions.

To aid readers’ understanding of this tricky business, the authors point to a driving analogy used by scientists at the University Corporation for Atmospheric Research in 2012:

“Adding just a little bit of speed to your highway commute each month can substantially raise the odds that you’ll get hurt some day. But if an accident does occur, the primary cause may not be your speed itself: it could be a wet road or a texting driver.” Similarly, while climate models may indicate a human effect is causing increases in the chances of having extremely high precipitation in a region (much like speeding increases the chances of having an accident), natural variability can still be the primary factor in any individual extreme event.”

Stay tuned for more studies like this one. It’s a safe bet that we’’ll see another series of papers appearing in the Bulletin of the American Meteorological Society about this time next year looking at extreme events in 2013. And there’ll be some fodder for scientists to chew on. Granted it’s been an especially quiet hurricane season so far (with only six named storms and no hurricanes) and thankfully nothing even approximating a Sandy this year. But the year has not been without its doozies:

And we still have some months to go before we close the books on 2013. So, as I said, stay tuned.


End Note

* A new study by Elizabeth Barnes of Colorado State University et al. suggests Sandy’s westward trajectory will be less likely to occur with climate change.

filed under: Africa, Arctic, Asia, climate change, ENSO, Europe, faculty, global warming, heat waves, rainfall, science, temperatures, weather
and: , , , , , , , , , , ,


All comments are moderated and limited to 275 words. Your e-mail address is never displayed. Read our Comment Guidelines for more details.

  1. MattN
    Oct 7, 2013

    You know the IPCC has a special report on extremes, right?

    Here’s what they say:

    ■“There is medium evidence and high agreement that long-term trends in normalized losses have not been attributed to natural or anthropogenic climate change”
    ■“The statement about the absence of trends in impacts attributable to natural or anthropogenic climate change holds for tropical and extratropical storms and tornados”
    ■“The absence of an attributable climate change signal in losses also holds for flood losses”

    So that takes care of flooding, hurricanes and tornados. Very little evidence of a connection.

    • Bill Chameides
      Bill Chameides
      Oct 24, 2013

      MattN: It’s always a pleasure to receive comments but less so when the commenter is providing a knee-jerk response without seemingly having read the post. The post was about a new methodology for testing the connection between climate change and extreme weather. And the studies discussed all postdate the IPCC report you reference. Which begs the question: Can we assume by your citing of an IPCC report, that you now accept the findings of all IPCC reports, or is this another example of cherry-picking?

      • MattN
        Oct 25, 2013

        Can we assume you now reject the findings of all IPCC reports? Keep hiding the decline, Doc.

  2. JimBerry
    Sep 12, 2013

    It would seem that it would make more sense to determine how many similar extreme events occurred in time periods of the same length in the past to see if these extreme events have been taking place at a higher rate in then last 40 years compared to the past rather than use models. And next to compare warm periods with cool periods.
    While hurricane Sandy may have been an “extreme event”, the overall occurrence of extreme hurricane events is significantly lower in the last 40 years than in most 40 year periods of record.

    My humble opinion is that this is a grabbing at straws effort to keep the funding flowing.

    Jim Berry

  3. Louis Hooffstetter
    Sep 12, 2013

    Without a doubt, determining the effect of human activities on climate is the “grandest” challenge faced by climate research.

    To my knowledge, no one has yet been able to separate the effect(s) of human activities on climate from natural variation. Please provide references to the paper or papers that have successfully done so. These of course would be peer reviewed papers with data available for replication. Please refrain from listing papers based on models, as the projections of every climate model drifts further from reality with each passing day.

  4. Hostpay
    Sep 11, 2013

    Climate change is a challenging issue for the global society as a whole and especially for global leaders, who must find solutions together and urgent action to prevent global warming from reaching dangerous levels and irreversible for humanity.
    Unambiguous data produced by scientists on the temperature rise of the planet and the effects of climate change on the environment, agriculture, human health and other sectors called the world to action. Science urged the world to revisit standards of accountability, control and security, as well as the consequences of the damage caused by traditional economic patterns to society and the environment.


  1. Another Week in the Ecological Crisis, September 15, 2013 – A Few Things Ill Considered - [...] 2013/09/10: GreenGrok: Starting to Suss Out Connections Between Extreme Weather and Climate Change [...]

Nicholas School of the Environment at Duke University | Box 90328 | Durham, NC 27708
how to contact us > | login to the site >

footer nav stuff