Reflections on climate-conflict research: More confusion than knowledge

A guest post from Prof Halvard Buhaug, Research Director and Research Professor at the Peace Research Institute, Oslo (PRIO).

Does climate change constitute a threat to peace and security? Many agree that it does. The US administration’s new National Security Strategy, launched last month, portrays climate change as ‘an urgent and growing threat.’

And this week, a new study appears to add scientific credibility to this concern, suggesting human-caused climate change contributed to the drought that preceded the Syrian civil war.

So does the Syrian case represent a general pattern, where climate changes and extremes are systematically increasing conflict risk? The short answer is no. But if scientists want to explore these links more closely, there are a few steps they need to take.

Cacophony of different findings

Recent research has reported a strong effect of climate extremes on violent conflict, yet many researchers question the robustness of such a link. Some even argue the relationship between climate and conflict is so complex that it can never fully be captured and understood.

There are legitimate reasons to be concerned about the impacts of climate change on security. However, a decade of research into the area appears to have produced more confusion than knowledge. But the cacophony of different findings and inadequate scientific evidence could be the result of poor data and simplistic research designs, rather than because no relationship exists.

In trying to establish links that can be observed and quantified, I see five key challenges that need to be addressed.

Poor data and vague hypotheses

The first challenge is to move away from grouping all impacts of climate change under the same banner. We should specify which climate condition or event we consider a security threat, and what the implications could be.

There is no reason why we should expect a flood to have the same impact on social behavior as the early arrival of the rainy season or a particularly wet year. Yet poor data and vague hypotheses seem to suggest that these distinct phenomena often look similar, and are treated as such.

Second, researchers should avoid the naive assumption of a simple and direct cause-and-effect link between climate and conflict. We should specify how a climatic condition could cause social unrest directly, or how they could contribute to existing problems.

In general, there is a tendency to underestimate or outright ignore the importance of the role of government or other institutions. For example, could local food price shocks constitute an intermediate step between drought and conflict? Possibly, but a failed harvest might not lead to food shortages or price rises if the situation is managed by a competent government.

Proponents of violence

Conflicts do not materialise out of thin air. So a third challenge is to consider who are the proponents of violence and what they are fighting for.

A conflict might not necessarily be provoked by the people most frustrated by a situation. Affected people at the margins may simply be too weak to protest, or they may lack the organisational skills needed to mobilise and act as a collective. A government can exploit a difficult environmental situation to excuse their own poor management – an accusation levelled at the Syrian government.

Fourth, researchers should consider not just who might be affected, but how they are likely to react.

Urban residents are unlikely to employ guerrilla warfare in response to rising cost of bread, for example, but may well turn to the streets in large numbers to protest. While organising a rebel movement needs planning, resources, know-how and motivated recruits. And a full-blown civil conflict requires a state government that is willing to respond with military force.

Until recently, researchers were limited to studying civil wars, which are perhaps the least climate-sensitive type of conflict one can imagine. But recent advances in data collection now make it possible to investigate a much broader range of conflict types.

Time and distance

The final challenge is how to consider the time and distance between the climate-related disaster and a conflict. Flooding will have a much more instant impact on a society than drought or rising sea levels, for example, and a dwindling water supply might be made worse by a neighbouring country upstream.

The Syrian uprising, for example, was preceded by a multi-year drought, which itself was preceded by another severe drought. At the same time there was an influx of refugees fleeing Iraq. Where do researchers draw the line for contributing factors?

Unfortunately, researchers are often restricted into using data averaged over a whole county and across whole years, which isn’t sufficiently specific.

Certainty and uncertainty

Let’s be honest; research on social processes always comes with a considerable margin of uncertainty, and even if the five challenges discussed above are addressed, there will be limits to what we can say about climate and conflict.

For example, we will never be able to conclude with certainty that a given conflict would never have occurred in the absence of climate change, or that a conflict would certainly have occurred if a society hadn’t adapted to that change.

Likewise, despite advances in conflict  forecasting methodologies, we will never be able to predict the onset of a new violent conflict with certainty based on data alone.

But it should be remembered that the climate-conflict link can work the other way as well. The destructive nature of armed conflict means it will increase a society’s vulnerability to the impacts of climate change. Civil war is development in reverse.

Therefore, in unstable corners of the world, ending violent conflict may be the most efficient and cost-effective way to improve resilience to climate change.

Main image: Unidentified Turkish soldier guard along the Turkish/Iran border.