Climate change isn’t just affecting the natural world. Researchers have long understood that rising levels of greenhouse gas emissions will also have cascading ramifications on the dynamics of human society, whether by forcing refugees to flee from newly flood-prone areas or arid regions, by causing spikes in the prices of food crops, or by reducing the productivity of livelihoods based on fishing or grazing in certain regions.
Recently, studies and journalistic investigations have focused on one particularly chilling potential social consequence of climate change: an increased frequency of armed conflicts around the world. By studying the link between various climactic factors and rates of historical violence, researchers have speculated that the climate trends we’ll experience over the next century—hotter overall temperatures, more erratic rainfall patterns and a rising sea level—could make conflict and war more common in the future.
Now, in the most comprehensive analysis of the work on climate change and armed conflict to date, a team from UC Berkeley and elsewhere has found that these climate trends are indeed likely to significantly increase the incidence of armed conflict overall. Their paper, published today in Science, examined 60 studies to aggregate sets of data on events spanning 8000 B.C.E. to the present that examined climate variables and incidences of violence in all major regions of the globe. For example, one of the source papers focused on temperature changes and violent crime in the U.S. from 1952 to 2009, while another looked at the number of conflicts in Europe per decade from 1400 to 1999 as a function of precipitation.
Cross-comparing these studies with the same statistical methods revealed patterns that, when projected into future, suggest that by 2050 we could see 50 percent more instances of mass conflict due to the effects of climate change.
The team, led by Solomon Hsiang, specifically looked the historical relationship between climatic factors (temperature and rainfall fluctuations) and the incidence of all sorts of conflicts detailed in their source studies, which they grouped into the categories of personal crime (murder, domestic violence, rape and assault), intergroup violence (civil wars, ethnic violence and riots) and institutional breakdowns (collapses of governing bodies or even of entire civilizations such as the Mayan empire). They examined this relationship on a variety of spatial scales, ranging from countries to regions to even warmer areas within a large building or stadium, and on varying time scales, from months to years to centuries in duration.
To standardize data from many different climates and regions, the researchers calculated the number of standard deviations away from baseline averages that temperatures and rainfall rates shifted in the areas studied by the previous papers, based on the time periods covered. A standard deviation is a statistical tool used to examine how data is clustered about an average—the more standard deviations away from the average you go, the more the observation in question is an outlier.
They found that when temperatures or precipitation patterns in an area strayed from the norm, all three types of violence tended to increase, with intergroup conflict in particular surging the most during hotter periods. Specifically, a region that experienced a period of warming that fell beyond one standard deviation of average conditions saw 4 percent more personal crime and 14 percent more intergroup conflict over the period studied. In other words, assuming the variables fall in a bell curve around from average conditions, life became more violent for the roughly 32 percent of regions that significantly deviated away from average temperatures and precipitation rates.
This level of deviation, to put it into perspective, is equivalent to a country in Africa going through an entire year of temperatures averaging 0.6°F warmer than usual or to a county in the U.S. experiencing average temperatures of 5°F warmer than normal in a given month. “These are moderate changes, but they have a sizable impact on societies,” explained Marshall Burke, the study’s co-lead author and a doctoral candidate at Berkeley’s Department of Agricultural and Resource Economics.
Extrapolating to the future, these rates mean that if the entire planet went through an average of 3.6°F of warming by 2050—an optimistic limit set at the 2009 Copenhagen conference—we’d see personal crime rise by 16 percent and intergroup conflicts surge by 50 percent. The distribution of violence wouldn’t be equal, either, as climate models indicate that some areas will be hit with warming periods that fall outside two, three or even four standard deviations of the norm (and thus experience more conflict), as shown in the map below:
But what characteristics of these climate changes—heat and erratic rainfall—cause people or institutions to become violent? The mechanisms that link climate trends with violence are varied and, in many cases, unclear.
Statistics show that in cities, hotter temperatures lead to more arrests for violent crimes, and some researchers believe our basic physiological stress response to heat is to blame someone or something for the heat—but it’s unclear whether the data represent causation or correlation. On a broader level, it’s believed that reductions in agricultural productivity—especially in largely agrarian societies—can drive intergroup conflict, as can extreme weather events and reductions in resources such as potable water (due to erratic rainfall) and arable land (due to sea level rise). All of these factors are likely to come into play as the climate changes.
Of course, there are a few caveats to the finding. For one, the researchers are extrapolating from historical data, so it’s possible that even though humans have previously become more violent as temperatures increased, we could behave differently in the future. Additionally, these hypotheses can’t be rigorously tested in a lab, so it’s impossible to entirely rule out all confounding factors and establish that the climate trends cause more conflict, rather than coincidentally occurring at the same time.
The researchers, though, say that they conducted the most rigorous analysis possible. The fact that the climate-violence relationship was consistently found among a wide range of time periods, cultures and regions, they argue, indicates that there is a substantial link between the two.
If warmer temperatures and erratic precipitation really do drive violence, what can we do? The researchers say that we need to engage in research to better understand the mechanisms by which this occurs—so that eventually, just as we’ll build infrastructure to anticipate and defend against the brunt of climate change’s most dire effects, we can also create innovative social institutions and policies that might minimize violence in a warming world.