A Controversial Claim: Climate Change = War & Violence

By Keith Kloor | August 2, 2013 11:19 am

In the late 2000s, the notion that climate change could trigger wars and geopolitical instability gained currency in military and intelligence circles. Security scholars gave credence to the possibility, think tanks were debating it and the media had another angle into the climate story. (I covered this news at the time–see here, here, and here.) One book captured the zeitgeist.

As The Economist reported in 2010:

The forecast is close to becoming received wisdom. A flurry of new books with titles such as “Global Warring” and “Climate Conflict” offer near-apocalyptic visions. Cleo Paskal, at the Royal Institute of International Affairs in London, predicts that floods, storms, the failure of the Indian monsoon and agricultural collapse will bring “enormous, and specific, geopolitical, economic, and security consequences for all of us…the world of tomorrow looks chaotic and violent”. Jeffrey Mazo of the International Institute for Strategic Studies, also in London, calls climate change an “existential threat” and fears it could usher in “state failure and internal conflict” in exposed places, notably Africa.

[UPDATE: See below for note on Cleo Paskal] This piece, unlike most of the coverage at the time, was skeptical:

Yet surprisingly few facts support these alarming assertions. Widely touted forecasts such as for 200m climate refugees in the next few decades seem to have been plucked from the air. Little or no academic research has looked at questions such as whether Bangladeshis displaced by a rising sea would move a series of short distances over a long period, or (more disruptively) a greater distance immediately.

There were some respected scholars, such as Geoff Dabelko, trying to tamp down the hype. “Don’t oversell the link between climate change and violent conflict or terrorism,” he advised several years ago in the Bulletin of the Atomic Scientists.

But it was too late. A cottage industry was born and it built on the climate/conflict meme. In 2011, researchers published a study in Nature that got wide media play. Ed Carr, a geographer and development expert, wondered how it made it through peer review. He was blunt:

Look, the problem here is simple: the connection between conflict and the environment is shaky, at best…The simple fact is that for interstate conflict, there are more negative cases than positive case . . . that is, where a particular environmental stressor exists, conflict DOES NOT happen far more often than it does.  Intrastate conflict is much, much more complex, though there are some indications that the environment does play a triggering/exacerbating role in conflict at this scale.

Carr goes on to lay out an extensive rebuttal to the Nature study and concludes:

This paper is a mess. But it got into print and made waves in a lot of popular outlets (for example, here and here).  Why?  Because it is reviving the long-dead corpse of environmental determinism…people really want the environment to in some way determine human behavior (we like simple explanations for complex events), even if that determination takes place via influences nuanced by local environmental variation, etc.  Environmental determinism fell apart in the face of empirical evidence in the 1930s.  But it makes for a good, simple narrative of explanation where we can just blame conflict on climate cycles that are beyond our control, and look past the things like colonialism that created the foundation for modern political economies of conflict.  This absolves the Global North of responsibility for these conflicts, and obscures the many ways in which these conflicts could be addressed productively.

Well, it so happens that the same researchers are it again, and Carr will probably be none too pleased to see that they are still seemingly intent on reviving the long-dead corpse of environmental determinism.  The researchers (from Princeton and Berkley) have published a new paper in Science that is generating splashy headlines like this:

Violence will rise as climate changes, scientists predict

And my personal favorite:

Climate change EVEN WORSE than you thought: It causes WAR and MURDER

But once you get past the headlines, there are good news stories that take a critical look at the new study, such as this one by Lauren Morello in Nature, who reports that

the lack of causal mechanisms [between climate change and conflict] leaves many political scientists sceptical about the environment’s role in conflicts, which they say are driven by a complex array of social factors.

Excellent context on the controversial nature of this research is also provided by Peter Aldhous at the New Scientist, who writes:

this provocative attempt to quantify the influence of climate on human conflict is itself setting off clashes among researchers who study the issue. “I would take their projections with a huge grain of salt,” says Halvard Buhuag of the Peace Research Institute Oslo in Norway.

I doubt that will happen, since any study published in a prestigious journal that associates climate change with war and violence is bound to be taken very seriously. Fortunately, there are scholars like Carr and Dabelko who plead for a more a nuanced discussion of a field with serious research gaps.

UPDATE: I should have mentioned that I favorably reviewed Cleo Paskal’s book, which I think the Economist has unfairly characterized in the quote I excerpted above. In fact, I would count Paskal as one of those scholars who has a nuanced take on the climate/conflict issue.

UPDATE: Ed Carr dives into the Science paper. His longish post is respectfully critical.

UPDATE:  For additional recent scholarly perspective (not behind a paywall), check out the 2012 PNAS study that examined the relationship between climate change and war in East Africa over the last 30 years. From the press release:

While a new study led by the University of Colorado Boulder shows the risk of human conflict in East Africa increases somewhat with hotter temperatures and drops a bit with higher precipitation, it concludes that socioeconomic, political and geographic factors play a much more substantial role than climate change.

Indeed, as one news analysis put it:

Unlike previous studies on the climate-conflict nexus, this new study paints a more nuanced picture of the linkages between climate, resources and conflict.

Also, I want to point out the work of other journalists who have covered this issue in a larger context. Brad Plumer examined the climate security argument in a 2009 TNR piece. And John Horgan at Scientific American has dissected the related “water wars” meme, some of the “overheated” resources war rhetoric, and more recently, the “Deep Roots Theory of War, which holds that war is ancient and innate.”

Finally, Andrew Holland of the American Security Project takes note of the contested debate among scholars, but finds it “exhausting.”  He writes:

In the end, this academic debate has largely been overtaken by events. Militaries and governments around the world overwhelmingly do see climate change as a threat to their national security. ASP’s Global Security Defense Index on Climate Change shows that more than 70% of the countries in the world see it as a threat to their security.

That begs an interesting question: Are important nuances of this debate irrelevant to the larger specter of climate change, which as Holland points out, is being taken seriously by militaries and governments around the world?


CATEGORIZED UNDER: climate change, global warming
  • scientist

    “…where a particular environmental stressor exists, conflict DOES NOT happen far more often than it does.”

    I don’t follow this reasoning. As an example, wearing sunscreen reduces your chances of getting skin cancer from small to very, very small. But the benefits of sunscreen are not insignificant.

    • Joshua

      Seriously – that is odd logic. If environmental stressors move the likelhood of conflict from 1% to, say 15%, we’d say that environmental stressors don’t increase the likelihood of conflict because “conflict DOES NOT happen far more often that it does?”

      I must be missing something, because Keith seems to think that Carr’s logic is powerful – certainly far more powerful than those scary wacko neo-environmental determinists.

  • scientist

    Also, “This absolves the Global North of responsibility for these conflicts…”
    I’m pretty sure a widely understood fact among environmentalists is that the Global North IS responsible for global warming. Plus most environmentalists I know are aware of colonialism.

  • Ramez Naam

    I do think the current paper is overstated. Yet environmental stressors can be a source of violence, particularly in least developed nations where the economic resources to find alternatives to local environmental resources are lacking.

    This is clearest, IMHO, in the conflict in Darfur, where drought is generally believed to have been a significant contributor (though not the only factor).

    Those conflicts exist already. Fresh water depletion in particular is an existing challenge across much of the developing world. To the extent that climate change will increase drought and shift rainfall patterns, it may exacerbate conflict over water resources and (in a domino effect) grazing and farming lands.

    But the real issues are the underlying local resource scarcity coupled with poverty that makes it difficult for locals to find an alternative to those local resources. Rich countries facing drought or crop failures can import food. Poor countries facing the same issues have far fewer options.

    • http://blogs.discovermagazine.com Keith Kloor

      Actually, Darfur, while being the poster child for the climate/conflict link, is not the ironclad example that many think it to be. Or at least you can make a good argument that the climate angle there is also way overstated:


      • Ramez Naam

        Yes, I think it’s clear that resources were only one factor there. And I actually hesitate to use the term “climate” to describe the issue in Darfur at all. To the extent that it was an ecological issue it was a combination of drought and local overpopulation, but I’m not aware of a linkage between that particular drought and climate change.

        • jh

          So, when is climate change just climate change, and when is it CLIMATE CHANGE?

          Any drought is surely a climate change event. Does it become a CLIMATE CHANGE event if Scientist Bob does a study and claims it to be so? If, as is often the case, Scientist Bob claims that, say, 10% of the drought intensity is a result of CLIMATE CHANGE, then where are we?

          It seems that the best way to approach the climate-security nexus is to forget about CLIMATE CHANGE altogether. It’s not relevant to determining a course of action, since cc occurs regardless of CC and in much larger proportions in the near and midterm. Also, whether a climate event is natural or anthropogenic, it can’t be predicted with any precision, for better or worse. In other words, events are random regardless of their cause.

      • scientist

        Keith, do you think the climate-conflict connection is overstated by the authors of the paper, or by various media outlets/commentators? If it is the latter, do you also take issue with the actual analysis done in the papers? It seems like the authors are only arguing this is a statistical difference in the probability of conflict during extreme conditions. The first author on the Nature paper submits a fairly extensive takedown of Carr’s piece in the comments section.

  • jh

    Can we stop for just a second and think about things a bit more carefully?

    Significant changes in climate have historically caused substantial disruption in human societies. Keith, didn’t you study Anasazi culture? Isn’t it true that a substantial change in rainfall patterns co-occurred (i’m avoiding “caused”) with the disappearance of that culture? I think it’s patently obvious that, if cultures are dramatically disrupted by climatic events, conflicts are likely to follow.

    However, I think it’s fair to be extremely skeptical of the probabilities published in this paper. Take, for example:

    “for each 1 standard deviation (1σ) change in climate toward warmer
    temperatures or more extreme rainfall”

    What is the quality of data that we have on temperature and rainfall for conflicts that occurred prior to the last 200 years? Terrible. The data quality is terrible. best, we have stable isotope or dendrochronological data (the latter of which is having many of it’s basic assumptions severely tested) and some general observations from the geologic record. Temp data is slightly – but not much – more robust. This data tells us almost nothing about interannual variations – and that’s exactly what climate is. Not just annual rainfall, but how much rainfall comes in what part of the year. Not just annual average temp, but daily variations and seasonal variations. This interannual variability is what most impacts essentials of life, such as food production.

    And of course there are a host of other very basic problems: defining, for example, what constitutes a “conflict” or “violence,” and many other definitions that would be, because of their very nature, not too objective.

    Moreover, there’s a huge difference between the modern world and the world of a few hundred years ago. Today, most of the leading nation states (US, China, Europe, Russia, Brazil, India) occupy numerous climate zones. Their economies are partly insulated from potentially devastating effects of climate events. The most likely place for major impacts is in areas comprised of many small nation states that comprise one or just a few biomes within their borders – such as Africa.

    • http://blogs.discovermagazine.com Keith Kloor

      There is a strong archaeological argument for climate change being a substantial factor in the “collapse” of numerous ancient societies, including the Anasazi (or ancestral Puebloans, which is the more politically correct term increasingly in use).

      However, when we say climate change, we have to think more than just mega-droughts. And even then the debate is still a complex one. Both of these points are discussed in a long 2007 feature I wrote for Science. It’s behind a paywall, but I’ve located a reprint of it (with typos, tho) at a university site.


      • jh


        Thanks for the link. A good read.

        After reading the PNAS paper you linked above, I see the distinction between the “major climate event” and the “incremental climate change” arguments. That major climate events impact security and cause social upheaval is, in my opinion, irrefutable, while the idea that incremental climate change causes violence is open to serious question.

        We might also add another distinction: “minor” vs “major” climate events: 1-2yr duration events vs 5-10 yr duration events. The fall of the “ancestral Puebloans” seems to land on the “major event” end of the spectrum.

        The PNAS paper is clearly interested in incremental climate change or minor climate events, since the period covered (30 yrs) can’t contain enough major climate events to provide a basis for comparison. The new Science paper seems to land somewhere near the “major climate event” end of the spectrum, since it relies in part on archeological evidence, which usually can’t tie 1-2yr climate events precisely to archeological events, so in that sense the two approaches aren’t comparable. Andrew Holland also seems to be thinking along the major climate event spectrum bcz he refers to events like the failure of the Indian monsoon.

        But if you get back to the basic reasoning perspective – the one that should be relevant in a security context – I think there are some clear and safe conclusions. First, I doubt we can safely draw any conclusions about the impact of incremental warming. However, for climate events, it seems safe to say that the larger the scope (spatial and temporal) of a climate event, the more likely it is to cause significant social disruption – if all other factors are equal. Thus a minor climate event presents dramatically higher risks in a small, poor country (Pakistan, Mali) with an unstable government than it does in a large country with many different climate regimes and a strong government (US, Canada, EU).

  • Andrew Holland

    Keith – thanks for the link. I think you’re final question there is a good one, but maybe we need to step back a bit.

    I think its not a stretch to ask all of the authors on both sides of this debate would agree with this statement: “Climate change is real and is caused by humans. Prudent measures to both mitigate emissions and adapt to the effects could reduce threats of conflict.”

    The multiple regression analyses that they all run purporting to show a 4% causal link between one degree of temperature rise and the likelihood of conflict are therefore somewhat irrelevant to actual debate. This is really “how many angels can dance on the head of a pin?” stuff.
    I think I tend to come down more hawkish on this issue than you – for instance, I see important links between the 2010 food price spikes (caused by fires in Russia) and the Arab Spring – you’ve questioned that. However, there’s still some outsized claims that cannot be supported by facts (200 million climate refugees! Wars across the Sahel! Arizona going to war with California over water!). Overheated rhetoric is sometimes used by advocates as a sales technique because the public and the media don’t seem to care otherwise – but that’s a whole different can of worms.

    But – on the other hand, it doesn’t take much thinking to see that a repeated failure of the monsoon season along the Indus River, devastating Pakistan’s ability to feed itself could have some serious national security implications (i.e. riots in cities, claims that India is stealing Pakistan’s water, etc.).

    As I say in my post, let the academics continue to puzzle out the direct or indirect links between climate and conflict – we’ve got more than enough evidence to get to work.

    • jh

      Andrew –

      I’m not sure what a person’s stance on human-caused climate change has to do with the issue. Historically there have been repeated instances of climate events that have caused social disruption (by “climate event” I mean any climate event – drought, changes in rainfall etc – that impact civilization. The Dust Bowl is one example).

      With or without human caused climate change, climate disruptions will occur and will cause social tension. How that tension plays out depends largely on the governments of the affected states, and it really doesn’t matter if the disruption is in some way human induced or not.

      A major gripe of many AGW skeptics (J Curry, for example) is that resources are better spent by building broad resilience to all climate events than by reducing emissions, the latter of which protects us only from human-caused events. Since even the most extreme estimates of human contribution to the intensity of extreme events are around 10% or less, it does seem rather stupid to expend massive resources to address this small proportion.

  • Tom Fuller

    The climate has warmed by 0.8C over the past 30 years.

    War has declined significantly–actually, dramatically–during that period.

    During the Little Ice Age, for those areas for which we have good historical records, war was pretty much constant.

    • Joshua

      “War has declined significantly–actually, dramatically–during that period.”

      If you don’t control for other influential variables, that in itself tells you nothing.

      Similarly, you wouldn’t be “cherry-picking” time periods, would you? What if we considered slightly longer time periods. Say we started counting at 1914, or 1939?

      This is an example of why unscientific speculation on blogs is often (if not always) not a very meaningful rebuttal to well-researched analysis. It seems that the results of well-researched analysis on this issue returns somewhat ambiguous results. Maybe you should go with that, Tom, instead of just throwing shallow reasoning around to see what sticks.

      • harrywr2


        War has decreased significantly in the last 30 years.

        According to our good friend James Hansen, almost all of the the global warming that has occurred in the last 200 years occurred in the last 30 years.

        Getting to a causality of human caused global warming causing war when we don’t even have correlation seems to be a great stretch.

        Have extreme ‘weather events’ cased social pressures…absolutely…do prolonged droughts cause social pressures sure.

        Has humanity been migrating since the the dawn of humanity…absolutely.

        • Joshua

          Harry –

          The study, as they say, is what it is and it seems that other analyses produce contrasting results…

          But they look at the association between weather and violence over some 12,000 years of data. And they did find a correlation.

          To respond by saying that there is no correlation over the last 30 years doesn’t seem to me like a particularly meaningful response.

          They never suggested that weather is the single causal factor, but that it can exacerbate existing tensions – to increase the risk of conflict some 14% beyond what it would be otherwise. Responding by saying “but it’s been hotter lately along with a decline in wars” is actually pretty much a non-sequitur.

          • Tom Fuller

            As should be obvious to the most casual observer, the real cherry pick is climate. War has been a commonplace since the advent of agriculture brought the invention of property. It has been endemic in hot climates and cold. This whole line of attack is, if anything, stupider than Xtreme Weather. I didn’t know that was possible.

          • Joshua

            “…the real cherry pick is climate.”

            The study is what it is and the data are what they are. They researched data for 12,000 years and found a statistical correlation with a higher prevalence of violence. It’s funny that you want to just wave it all away with cherry picking because you think the correlation is too “alarmist.”

            Data aren’t partisan. Is their methodology flawed? Did they make some calculation errors? Was their sampling in error? Did they omit data? Without any of that, your doing nothing other than arguing by assertion. If so, make an argument – don’t cherry-pick dates with no consideration of relevant variables and argue that somehow you’re making a cogent analysis in contrast to theirs.

            Certainly, we can explain some statistical correlations as a matter of chance or methodological flaws – but for me when investigating that question it seems useful to consider whether there might be a plausible mechanism that would explain causality. It certainly seems logical to me that the stresses of environmental pressures or extreme weather could plausible exacerbate existing tensions.

            “War has been a commonplace since the advent of agriculture brought the
            invention of property. It has been endemic in hot climates and cold”

            That is all, essentially, irrelevant. This is basic logic. All of what you said could be true (and of course, it is true), and weather extremes and environmental pressures could still increase the likelihood of conflict.

            Do you really not understand that?

          • kdk33

            Fascinating. So we now have an accurate temperature history of the last 12,000 years? And an accurate history of human warfare over the last 12,000 years. And heatwaves inspire robberies?

            I love the climate wars. You can’t make this stuff up.

            Meanwhile among the sane…

          • Joshua

            By whatever means they used to measure temperature, which was no doubt imperfect, they found a correlation.

            Unless you can point out their statistical/methodological errors that lead to a bias in one direction of the other, probability indicates that the flaws in their methods for evaluating past temperatures would be even distributed, and thus not invalidate their findings.

            The same goes for how they measured conflict.

          • Joshua

            But perhaps you’re right. Perhaps they authors just aren’t sane.

            Just because you’re paranoid doesn’t mean that people aren’t out to get you, eh?

          • kdk33

            Wow. Fascinating defense: All the data might be garbage, but if there is a statistical correlation then the conclusions are justified.

            The statistics may be perfect. The conclusions absolute garbage.

            Meanwhile among the sane. No we don’t have an accurate enough temperature history of the last 12,000 years to correlate with anything. Ditto human warfare.

            What we know with absolute certainty is this: Human welfare has never been better. Ever. And carbon fuel is the reason why. And over the period where we do have accurate data, the correlation is perfect: human prosperity increases with increasing temperature.

            But, hey, don’t let the facts dissuade you. Otherwise I’d have noone to argue with.

        • Joshua

          You are suggesting a bias, but you have no argument in support of your suggestion.

          You are indicating that the data are flawed. All data are flawed. You couldn’t draw any kind of conclusions on anything if you insist only on perfect data. If you want to point out that the flaws with their data are so flawed as to render invalid the association they find between the pressures of weather and climate changes and increased risk for violence – have at it; but then you’d need to actually do an analysis to support that argument. Simply saying that their data are flawed is holding them against a meaningless standard.

          But yeah – better to insist that their conclusion that environmental pressures exacerbate existing tensions is “garbage” without actually making an argument. Better to rely on argument by assertion, and to just conclude that those who agree with you are sane and those who disagree with your argument by assertion are “insane.”

          And better, without any study, analysis, or evidence, to conclude that no other variables are relevant, and carbon fuel explains why “human welfare is better than ever.” Freedom, civil rights, social safety net, more democratic representation, the threat of nuclear weapons? All irrelevant. And we know that is true simply because you make that assertion.

          That’s truly a work of art and a thing of beauty.

          • kdk33

            I’m led to believe the Superbowls correlate with presidential elections and stock markets. We have perfect data for these. I’m sure the statistics are methodologically correct.

  • Buddy199

    Fortunately, there are scholars like Carr and Dabelko who plead for a more a nuanced discussion of a field with serious research gaps.


    That mentality should be applied to the field of climate change in general.

  • http://blogs.discovermagazine.com James Evans

    Calling for “nuance”, makes you sound kind of clever and such like.

    It would be great if you actually called out the BS instead of just acting as a “reasonable-sounding” link repository.

    Though, I know you’re trying to scrape a living from this.

  • Matt B

    I was listening to Living on Earth on NPR today with host “no matter what I talk about I can find a climate change angle” Steve Curwood interviewing study co-author Edward Miguel, and hear this exchange:

    CURWOOD: How about the United States?

    MIGUEL: We also looked at studies that look at crime in United
    States, and so if you look at cities, for instance, that have
    anomalously high temperatures in a given month, they tend to have more violent crime, more murder, more assault.

    So, peer-reviewed experts state that this effect manifests itself in the United States. A quick check of FBI statistics shows from 1992-2011, violent crime in the US dropped by almost 50%. Thus:

    1. Since higher temperatures increase violent crime, and

    2. Violent crime is way down in the US over the last 2 decades, then

    3. The US has been cooling the last 2 decades

    No worries for climate change in the land of the free & home of the brave!

  • hunterson1

    You are edging closer the obvious conclusion. It is a difficult journey, but it is impressive that you are willing to make it at as far as you have.

  • http://synapse9.com/signals Jessie Henshaw

    I share your view, technically, but you don’t seem to be thinking about the system producing the conditions you are talking about. You write as if we lived in a bubble, with separate words representing separate influences.

    From a more realistic systems view the combined crescendo of major environmental and economic crises that include climate change and the global societal traumas now erupting have one common origin. That’s our society’s addiction using technology to change the earth and reorganize our ways of doing it by ever bigger and faster changing steps. It produces an illusion of limitless growing wealth, that we like, and euphemistically call “growth”. It’s not “growth”. It’s “self-deception”.

    In nature, growth is actually a process of building something that you’ll then take care of. What our economy is doing is carrying out a process of blind over-building that if continued can only destroy itself, as a way of satisfying our addictions by feeding them… The appeal of reliving addictions by feeding them is obviously rather seductive, a good way to destroy the world you live in and care deeply about.

  • lngtrm1

    This argument sounds just like a climate change discussion between opposing camps. Someone says CO2 is the culprit and then someone says that water vapor has WAYYY more of an effect.

    So yes, political issues are still primary, but triggers are really important, just like a few more PPMs of CO2 are really important.

    We know there are large natural releases of methane, but isn’t it the incremental releases above and beyond that that are troubling?

    Isn’t the actual issue the incremental effect of climate related stresses on peace and stability?

    I’d like to see the State Department’s estimates on regime change prior to the Arab Spring. I’ll bet they completely missed Egypt and likely Tunisia because they were using all the “traditional” analysis and not including climate effects in those regions.

    • Buddy199

      Global temperatures dropped in the 1940’s during WW2 and the beginning of the Cold War (aptly named?).

    • jh

      Food prices may have been a catalyst for the Arab Spring. Nonetheless, they didn’t instigate revolution in most of the rest of the world, right?
      So in countries with poor governance – some might say extremely poor governance – climate may be a factor in violence.

  • jh

    Most of the argument about climate events depends on the idea that all climate events are bad – that is, we can have “normal” climate, or negative “climate events.” But, obviously, there are periods of better than average climate as well.

    So how do we, in a particular region, define “average” climate? Can that be done with a single index such as annual average temperature? If so, then shouldn’t we see beneficial effects from “better than average climate” just as we would see deleterious effects from “worse than average” climate? Where’s the baseline?


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


Collide-a-Scape is an archived Discover blog. Keep up with Keith's current work at http://www.keithkloor.com/

About Keith Kloor

Keith Kloor is a NYC-based journalist, and an adjunct professor of journalism at New York University. His work has appeared in Slate, Science, Discover, and the Washington Post magazine, among other outlets. From 2000 to 2008, he was a senior editor at Audubon Magazine. In 2008-2009, he was a Fellow at the University of Colorado’s Center for Environmental Journalism, in Boulder, where he studied how a changing environment (including climate change) influenced prehistoric societies in the U.S. Southwest. He covers a wide range of topics, from conservation biology and biotechnology to urban planning and archaeology.


See More

Collapse bottom bar