A recently-released study by Jan Selby and colleagues analyzes existing research on the intersection of climate change and conflict in Syria. The article, published in the Journal of Political Geography, includes a critique of a 2015 study published by the Center for Climate and Security’s (CCS) Caitlin Werrell, Francesco Femia and Troy Sternberg (and a short briefer by CCS from 2012), as well as two other studies by Colin Kelley et al (2015) and Peter Gleick (2014). More research into the climate-conflict nexus in pre-civil war Syria is certainly welcome for better understanding the risks and informing future policies for addressing them. In this study, Selby et al. point to some important gaps in the data on the connection between displaced peoples and social and political unrest, and the possible role of market liberalization in the Syrian conflict. However, the study does nothing to refute the role of climate change in Syrian instability in the years before the war, while muddying the waters on the subject through a few mischaracterizations that are worth addressing at some length.
On the surface, this may seem to be low-stakes ivory tower drama. However, the significance of robust research on the connections between climate and security should not be understated. As we’ve noted previously, it is important not to oversimplify the links between climate change and conflict, as that can lead to bad policies and maladaptation. However, it is also important not to underestimate these links, as that can also lead to bad policy and a failure to adapt.
In this context, there are three elements of the study that contribute to an underestimation of climate-related risks to Syria and the broader region: a mischaracterization of the climate science, a conflation of causality and contribution, and an underestimation of the human toll.
Mischaracterization of the Climate Science
First, the study does not adequately substantiate its assertion that there is “no evidence of a long term drying trend” in the Fertile Crescent over the past 10-25 years, or its assertion that this drying is due to natural variability. Both claims run contrary to clear evidence in the scientific literature. Regarding the former claim, Colin Kelley (full disclosure, he is a Senior Research Fellow with CCS) and his colleagues noted in their commentary on the article published in the same issue of the Journal of Political Geography:
“…an analysis of a new gridded tree ring dataset of winter/ spring surface moisture availability for all of Europe, North Africa and the Middle East (Cook et al. Science Advances, 2015) concluded that 1998-2012 was the driest 15-year period in the Levant in the last 900 years (Cook, Anchukaitis, Touchan, Meko. Journal of Geophysical Research, 2016). This new evidence confirms that recent drying is outside the range of what would be expected due to natural variability. “
Further, Selby and his colleagues conclude that this exceptionally dry 10-25 year period was due to “natural decadal to multidecadal variability,” but again cite no evidence to substantiate the claim. Indeed, as Colin Kelley and his colleagues note, there is a compelling body of literature supporting the case that this drying trend is very likely due to a changing climate. This includes Kelley et al. 2015, Zappa et al. 2015, Hoerling et al. 2012 and Kelley et al. 2012.
Why This Matters: How specifically climate change relates to other socio-economic-political factors is difficult to disentangle, as Selby et al. and previous researchers have correctly acknowledged. However, artificially downplaying the climate change factor in the region, in the face of significant scientific evidence to the contrary, does a disservice. This approach can serve to encourage political leaders to accept a rosier picture of the MENA region’s climate future, and thus contribute to inadequate preparation. Further, given that water stress has historically often served as a foundation for cooperation between conflicting parties, minimizing the contributions of climate change to water security could close potential avenues of conflict resolution and peacebuilding.
Conflation of Causality and Contribution
Second, the study seems heavily focused on building up and batting down the strawman that climate change caused, or was a primary cause of, the conflict in Syria, which no serious researchers – including those who conducted the three main studies critiqued in the article – have claimed. The research to date is clear that climate change, which according to Kelley et al made the extreme drought in Syria 2-3 times more likely, “contributed to” social unrest in Syria prior to the civil war by multiplying risks to agricultural lands and rangelands, particularly but not exclusively in the northeast. In this context, the drought was one of a number of other environmental, economic and governance factors – including natural resource mismanagement by the Assad regime – contributing to the mass displacement of a significant number of Syrians, and thus potentially increasing the “likelihood” of conflict – a probabilistic, not a causal, claim. This is the claim made in every article published by CCS experts on the subject (2012, 2013, 2014 and 2015), none of which assert that climate change “caused” the conflict in Syria, or even was a “primary cause” of that conflict.
Why This Matters: This is a very important clarification to make, particularly in terms of avoiding the narrative that government actors, such as the Assad regime at the time, are blameless in the face of natural forces. Poor governance on a number of levels – including unsustainable agricultural and pastoral practices – are significant contributing factors in the displacement that occurred during the period of the drought. Climate change, and the extreme drought, simply made the underlying environmental conditions in a poorly-governed state worse, thus increasing the likelihood of social unrest. In this context, the research to date has mostly avoided the language of causality in favor of the language of probability. As Cullen Hendrix wisely notes in his published commentary on the article (full disclosure again, Cullen is a Senior Research Advisor with CCS):
“…scholars must avoid the siren’s song of using causal language as applied to particular cases when the evidence supports more probabilistic relationships. Both the public and the policy community are keen to link abstract, probabilistic mechanisms to particular cases, and thus scholars face implicit encouragement to frame their results in terms of cases that seem to fit the causal processes they seek to model. However, most work in this area finds climate shocks raise the probability of a large-scale event (like conflict onset) occurring relative to some baseline or increases the frequency with which smaller-scale events (protests, individual battles or skirmishes, cattle raids) occur. When this evidence is marshalled to explain any particular event, however, it often takes on the air of a necessary condition – if but for the climate shock, the event would not have occurred. This claim is almost always impossible to substantiate and invites significant criticism – to wit, the exchange here. Doing so undermines an already strong case for considering climate change a human and national security issue.”
Scholars studying the linkages between climate change and conflict in Syria have largely followed Cullen’s advice, which is why it seems that Selby et al. primarily have an issue with the discourse in the media, rather than the discourse behind the academic journal paywall. Certainly, there is no shortage of headlines that have oversimplified the links between Syria, climate change and conflict. And yes, these oversimplifications can trickle up to the highest levels of government. We share the frustrations of Selby et al. about this dynamic. However, it is important to not conflate how research is characterized by the media, with what the research actually concludes. Research, and the public dissemination of that research, are two very different phenomena. Ideally, the way research is characterized by the media is accurate, but the audience for academic journals is not the same as the broader public. The role of the media is to take complex research and communicate it to a non-expert community accurately but clearly. This is a perennial challenge that will likely never be fully resolved. That said, responsibility for how an issue is covered is a shared one, and both researchers and journalists should aim to avoid oversimplification.
Underestimation of the Human Toll
Third, the study underplays the scale of livelihood decimation and internal displacement in Syria during the period of the extreme drought, selectively choosing data from a shorter time period that excludes the later stages of the extreme drought, and selectively dismissing some journalistic, government and UN sources that are important for filling out a generally hazy picture. This is a worrying assertion, as it suggests a more stable picture in Assad’s pre-conflict Syria than may have actually existed. As Colin Kelley and colleagues note in their commentary:
“…in 2010 the UN Office for the Coordination of Humanitarian Affairs found that about 300,000 families were driven to Damascus, Aleppo, and other cities in one of the “largest internal displacements in the Middle East in recent years” due to drought (Ababsa, 2015). During a mission to Syria that ended in September 2010, Olivier de Schutter, UN special reporter, stated that 1.3 million people had been affected by the four-year drought, 800,000 of whom had their livelihoods devastated. Not included in these estimates is the considerable displacement that occurred in the 6-18 months directly preceding the uprising, the time when we would expect that most of the displacement occurred given the accrued stress associated with the drought’s persistence. The authors of S2017 [Selby et al] provide little evidence that “excess migration” due to the drought was only a small proportion of the total displaced. Also cited by Ababsa was a 2010 report (IRIN, 2010) by IRIN, an independent, non-profit media service specializing in ground reporting on humanitarian crises. This report, titled “Syria drought pushing millions into poverty,” stated: “A top UN official warns that Syria’s drought is affecting food security and has pushed 2 to 3 million people into ‘extreme poverty’.”
Why This Matters: We of course cannot expect perfectly accurate numbers on internal displacement during the extreme drought period of 2007-2010, as the Assad regime made it difficult for journalists, researchers and humanitarian assistance workers to access those displaced peoples. Further, this period of drought was immediately followed by a violent civil war, thus making analysis of the primary reasons for the pre-war displacement, and a survey of how many of those displaced by the drought contributed to social and political unrest, very difficult to conduct. However, that does not imply that researchers should only accept official Syrian government and UN estimates of displaced peoples from one governorate during one year of a three-year extreme drought (2008-2009) to arrive at an estimate, as Selby et al. have done, and dismiss reports of mass displacement from the tail end of the extreme drought period (2009-2010). Journalistic sources, government sources, UN sources and even policy institute sources are all important in trying to cobble together a picture of vulnerability in an otherwise difficult-to-penetrate political environment. Collectively, these sources support estimates of around 1.5-2 million people having been displaced during the broader drought period of 2006-2011.
Let’s Focus on Probabilities and Prepare
In conclusion, Selby and his colleagues expand the discourse on climate change and conflict simply by further exploring the issue, and by pointing to key gaps in the research related to the motivations of displaced peoples for contributing to political unrest, and the relative role of market liberalization during the period of inquiry. However, dismissing much of the scientific evidence, conflating causal claims with probabilistic claims, and downplaying the scale of livelihood decimation and human displacement by narrowing the scope of analysis to one year, can damage the truth of the drought’s impact. This also, albeit unintentionally, risks sending signals to policy-makers in the region (and internationally) that the risks are more easily managed than they are, and will likely be in the future.
We share the study’s aim of better understanding the dynamics associated with climate change, political unrest, state fragility and conflict, and probing the blank spots in the data. On matters of peace and security, however, it is important that we not wait for ironclad causal links in the literature before advancing policies to anticipate, prevent and respond to risks that are probable. Climate change, by placing strains on basic natural resources, is very likely to exacerbate dynamics that can increase the vulnerability of populations, and the fragility of states. Those dynamics can increase the probability of conflict and instability, but do not guarantee it. Focusing on this broader picture, and emphasizing a probabilistic rather than causal paradigm for assessing these links, will lead to more fruitful – and useful – areas of future research.
UPDATE (9/10/2017): A core element of Selby and his colleague’s article is the assertion that the existing literature more robustly supports the thesis that it was economic liberalization policies beginning in 2007, and “major structural transformations of the [agricultural] sector,” that were at the heart of internal migration during the 2007-2010 period. Previous studies on the role of the drought did indeed acknowledge that the existing literature focused on these economic drivers of migration, with little exploration of the underlying environmental conditions during the period. This absence of exploration of the drought’s impact on unrest in Syria is precisely what drove the research critiqued in Selby’s study. The question of how the drought interacted with other social, political and economic drivers of migration, and political unrest, was very under-explored in the literature, and deserved a closer look. Our study, and the studies conducted by Kelley et al and Gleick, began to scratch the surface of that question, and we encourage additional research to explore it further.
Thank you for the very deep and informative text. Especially – having carefully studied the relationship between individual statements and logical constructions – I was convinced of a very high and leading level – practically – of the presented verbal model. Admired!
PS. I sympathize … with your country, which has been attacked by hurricanes and suffered tremendous damage … including human casualties.