Climate cycles drive civil wars

A child in a rebel camp in the north-eastern Central African Republic, one of the countries where internal conflict has been shown to be twice as likely to occur during a year when the El Niño weather pattern occurs than in La Niña years. Credit: Pierre Holtz/UNICEF CAR/Flickr

Civil war is much more likely during the warmer phase of a global climate cycle, seemingly as political tensions get literally overheated. That’s according to researchers from Columbia University, New York, who say that conflicts within a single country are twice as likely to occur during warmer El Niño years as cooler La Niña years. This is the first indication that modern societies’ stability relates strongly to climate, though the scientists warn that their findings might not be applicable to human-caused climate change. However some of the more dramatic changes in climate and society during humanity’s history have successfully been tied together. Columbia’s Mark Cane says his team’s latest findings build on those results. “What it shows beyond any doubt is that even in this modern world, climate variations have an impact on the number of civil conflicts,” Cane said Tuesday. “It’s frankly difficult to see why that won’t carry over to a world that is disrupted by global warming.”

Previous studies on whether modern climate has influenced war found only weak links between temperature over long periods, while studies on year-to-year local changes have disagreed and been criticised for having too narrow a focus. Consequently, together with Solomon Hsiang and Kyle Meng, Cane turned to the El Niño-Southern Oscillation, or ENSO, that affects weather patterns where half the world’s people live. El Niño originates from around 1°C warming in the tropical Pacific every three to seven years, bringing hotter, drier weather to the tropics. That alternates with cooler La Niña phases that provide more tropical rain, but can dry out more northern areas, as in East Africa and the southwest US this year. Consequently, the Columbia scientists were working with changes between two states on a worldwide scale that happened relatively regularly. This comes close to the “ideal but impossible” experiment of studying two Earths with different climates, they write in top science journal Nature.

“Teleconnected” countries

The scientists used Columbia’s NINO3 ENSO data from 1950 to 2004, and compared it against data, compiled by researchers in Norway, on conflicts within countries that killed more than 25 people in a given year. That covered 175 countries and 234 conflicts – representing the majority of all large-scale violence seen in this time – over half of which each caused more than 1,000 battle-related deaths. “We imagined a scenario where the world remains in the most peaceful La Niña, to predict what the conflict level would have looked like had the world stayed like that,” said Hsiang. “We then compared that with the actual number of conflicts as the global climate oscillates back and forth. The difference between those two worlds sums up to 48 conflicts.” That’s equivalent to just over one fifth of all civil conflicts since 1950.

This figure shows which countries have a climate that is strongly linked to El Nino. Countries colored red are considered “teleconnected” in our study. This means that a majority of the population in red countries lives in locations that become much warmer during El Nino years. Countries colored blue are considered “weakly-affected.” This means that either (1) temperatures in these countries do not rise substantially during El Nino years and/or (2) regions where temperatures rise during El Nino are not heavily populated. Grey countries are not included in the study because they have no data. Credit: Hsiang et al. Nature

But beneath that overall figure, there is a group of countries that is more affected by El Niño-related changes in average surface temperature, rainfall, and farming output. Hsiang, Meng and Cane separated these “teleconnected” countries from other, more weakly affected ones, finding that the difference between the number of civil conflicts between El Niño and La Niña was nearer to one in three. During La Niña, the chance of civil war breaking out in teleconnected countries was about 3 percent; during El Niño, the chance doubled, to 6 percent. Countries not affected by the cycle remained at 2 percent no matter what. With each El Niño, the countries in this group experience two-thirds of a year less peacetime than they would have enjoyed in a La Niña state on average .

A link, or a cause?

In the battle between Peruvian government and Shining Path rebels, triggered after a powerful El Niño destroyed crops in 1982, the military trained peasants as auto-defense groups to defend themselves. Credit: Jessica Varat, Museo de la Nacion, Lima, Peru/Flickr

Looking more deeply within that group, nations’ wealth also has a role to play. Australia’s climate is controlled by ENSO, but has never seen a civil war. At the opposite extreme, Peru and Sudan have experienced long-term violence closely linked to El Niño’s impact, with Sudan’s conflicts often flaring up as the warmer phases hit. This begins to hint at ways the ENSO cycle could exerts its effects, although a full explanation is not yet possible. Meng noted that the worst-affected countries earned most of their money from farming, which is often badly hit by El Niño. So, it could be that poor countries don’t have the power to respond to the effects of El Niño, and consequences like shortages of food tip them into violence. “When crops fail, people may take up a gun simply to make a living,” Hsiang said. Alternatively, it could be that these countries are more easily damaged by weather, and that damage starts violence, which then makes them poorer. Finally, that another factor, like poor government, could make them both poor and vulnerable to ENSO-triggered violence.

Not knowing exactly how ENSO is linked to violence makes Halvard Buhaug, who had previously found virtually no link between climate and the frequency of civil wars over the past 50 years in sub-Saharan Africa sceptical of the new findings. “The study fails to improve on our understanding of the causes of armed conflicts,” he said. Hsiang conceded that this uncertainty limits the usefulness of the findings, but said that work is in process to understand the link. “If we really understood the detailed mechanisms, we might be better able to make projections about future climate change,” he said. “Because that’s such an important question, we are currently pulling together new datasets and conducting new analyses now that we’ve established this basic empirical fact, that climate is linked to patterns of violence around the world.”