The Amelioration of Uncertainty

The Decline of Science and Science (Magazine)

According to an article in Science magazine titled “Quantifying the Influence of Climate on Human Conflict” global warming may result in “amplified rates of human conflict could represent a large and critical impact of anthropogenic climate change”. The article is gated, so I can’t look at the details buts it’s worth doing a sanity check on the main conclusion.

In the abstract the authors claim:

We find strong causal evidence linking climatic events to human conflict across a range of spatial and temporal scales and across all major regions of the world. The magnitude of climate’s influence is substantial: for each 1 standard deviation (1σ) change in climate toward warmer temperatures or more extreme rainfall, median estimates indicate that the frequency of interpersonal violence rises 4% and the frequency of intergroup conflict rises 14%.

I don’t find this hard to believe. If you’re going to have intergroup conflict then it’s easier in the summer than the winter. The war in Afghanistan has a strong seasonal component to it, usually referred to as the “summer fighting season”, which results in easily modeled oscillations in kinetic activity. See for example IED trends:


The trend is a result of the number coalition troops in Afghanistan and superimposed on that is an oscillation which troops on the ground have no trouble identifying even without the use of graphs or statistics.

So it’s no surprise the authors found a strong correlation between rising temperatures and rising conflicts. But the key here is “if you’re going to have intergroup conflict”. During my lifetime, the global risk of a military aged male dying in a conflict reached an all time low. On a per capita basis, we’ve easily been living in the most peaceful time in man’s history and this seems to be part of a long term trend (the German rate of combat deaths during the world wars are supposedly lower than those seen in primitive tribal societies). See for example this:


and yet you can find endless graphs like this one showing a global rise in temperatures over the same time period:


So if the authors really did find a strong correlation “across a range of spatial and temporal scales” then either these global trends in temperature change are wrong, or the authors of that science study error when they infer “anthropogenic climate change” would have a “large and critical impact” on “human conflict”.

August 2, 2013
  • August 2, 2013dab

    I haven’t read the paper either, and as such don’t have an opinion on it, but according to appendix B in the freely available supplementary material,

    The goal of our analysis is to identify studies that estimate the causal relationship of climate on some conflict outcome, and to compare the direction and magnitude of the estimated effects across studies. Where possible, this latter goal is accomplished by calculating standardized effects for each study.

    So, it appears that they are looking at multiple “conflict outcomes,” not just combat deaths. Now, you’re clearly more expert in this area than me, but it seems that an obvious confounder for your chosen metric (i.e., combat deaths) is the advances in medical care over the time period you show. Is there a similar graph showing a decline in casualties over time? And it seems that another confounder could be the scale of the conflicts. If there are fewer people doing the actual fighting in a given conflict (due to advances in military technology requiring fewer personnel to wage a war), then the absolute number of deaths will be smaller. Now, again, given your background, I’m sure you know far more about this than I do. But your critique seems kind of shallow to this mildly informed layman.

  • August 2, 2013Joseph


    Ha! no I’m not an expert, although in a few cases I’ve been knee deep in some data and analysis not readily available.

    Those are all valid points and a real political scientist or gentleman scholar could easily generate dozens more. For example, perhaps conflicts have shifted from warlike situations towards gang violence.

    Moreover, I have zero doubt that temperature and violence are strongly causally related. Violence tends to occur during the day and during the summer, both of which have higher temperatures. In the case of Afghanistan, I can say with certainty that we understand in intimate detail the practical considerations that affect insurgent’s planning. The evidence for a causal connection is overwhelming and goes way beyond any ratty old statistical analysis.

    The smell test though involves looking at the big picture. I think it’s well established that the world has been getting less violent. And I’m guessing the authors have no problem with the claim that the world has been getting warmer. That directly contradicts their assertion “amplified rates of human conflict could represent a large and critical impact of anthropogenic climate change”.

    So the question is which do you think is more likely: that this big picture is fundamentally wrong in some way, or that there’s some subtle issue with the statistics?

    From my experience, I’d say the latter is more likely, but would gladly accept, and find it very interesting, if the big picture was wrong.

  • August 2, 2013dab

    Thanks for your reply. I get your argument and it is quite reasonable, assuming one accepts that “it’s well established that the world has been getting less violent.” Not having any social science background, I was unaware that that had been established and I guess my point was that your post did not establish that fact convincingly. However, if it has been well established (i.e., if the result is “well-known” as they say in the math and physics literature), then there is no reason why you should need to re-establish it. So, I suppose that’s where I’ll have to defer to your greater level of expertise.

Leave a Reply or trackback