Climate-Security Linkages Lost in TranslationSeptember 13, 2010 By Marc LevyA recent news story summarizing some interesting research by Halvard Buhaug carried the headline “Civil war in Africa has no link to climate change.” This is unfortunate because there’s nothing in Buhaug’s results, which were published in the Proceedings of the National Academy of Sciences, to support that conclusion.
In fact, the possibility that climate change might trigger conflict remains very real. Understanding why the headline writers got it wrong will help us better meet the growing demand for usable information about climate-conflict linkages.
First, the headline writer made a simple mistake by translating Buhaug’s modest model results — that under certain specifications climate variables were not statistically significant — into a much stronger causal conclusion: that climate is unrelated to conflict. A more responsible summary is that the historical relationship between climate and conflict depends on how the model is specified. But this is harder to squeeze into a headline — and much less likely to lure distracted online readers.
Buhaug tests 11 different models, but none of the 11 corresponds to what I would consider the emerging view on how climate shapes conflict. Using Miguel et al. (2004) as a reasonable representation of this view, and supported by other studies, there seems to be a strong likelihood that climatic shocks — due to their negative impacts on livelihoods — increase the likelihood that high-intensity civil wars will break out. None of Buhaug’s 11 models tested that view precisely.
If we are going to make progress as a community, we need to be specific about theoretically informed causal mechanisms. Our case studies and statistical tests should promote comparable results, around a discrete number of relevant mechanisms.
Second, a more profound confusion reflected in the headline concerns the term “climate change.” Buhaug’s research did not look at climate change at all, but rather historical climate variability. Variability of past climate is surely relevant to understanding the possible impacts of climate change, but there’s no way that, by itself, it can answer the question headline writers and policymakers want answered: Will climate change spark more conflict? For that we need to engage in a much richer combination of scenario analysis and model testing than we have done so far.
We are in a period in which climate change assessments have become highly politicized and climate politics are enormously contentious. The post-Copenhagen agenda for coming to grips with mitigation and adaptation remains primitive and unclear. Under these circumstances, we need to work extra hard to make sure that our research adds clarity and does not fan the flames of confusion. Buhaug’s paper is a good model in this regard, but the media coverage does not reflect its complexity. (Editor’s note: A few outlets – Nature, and TIME’s Ecocentric blog – did compare the clashing conclusions of Buhaug’s work and an earlier PNAS paper by Marshall Burke.)
The stakes are high. This isn’t a “normal” case of having trouble translating nuanced science into accessible news coverage. There is a gigantic disinformation machine with a well-funded cadre of “confusionistas” actively distorting and misrepresenting climate science. Scientists need to make it harder for them to succeed, not easier.
Here’s how I would characterize what we know and we are trying to learn:
1) Economic deprivation almost certainly heightens the risk of internal war.To understand how climate change might affect future conflict, we need to know much more. We need to understand how changing climate patterns interact with year-to-year variability to affect deprivation and shocks. We need to construct plausible socioeconomic scenarios of change to enable us to explore how the dynamics of climate, economics, demography, and politics will interact and unfold to shape conflict risk.
2) Economic shocks, as a form of deprivation, almost certainly heighten the risk of internal war.
3) Sharp declines in rainfall, compared to average, almost certainly generate economic shocks and deprivation.
4) Therefore, we are almost certain that sharp declines in rainfall raise the risk of internal war.
The same scenarios that generate future climate change also typically assume high levels of economic growth in Africa and other developing regions. If development is consistent with these projections, the risk of conflict will lessen over time as economies develop and democratic institutions spread.
To say something credible about climate change and conflict, we need to be able to articulate future pathways of economics and politics, because we know these will have a major impact on conflict in addition to climate change. Since we currently lack this ability, we must build it.
Marc Levy is deputy director of the Center for International Earth Science Information Network (CIESIN), a research and data center of the Earth Institute of Columbia University.