期刊名称：Proceedings of the National Academy of Sciences
印刷版ISSN：0027-8424
电子版ISSN：1091-6490
出版年度：2014
卷号：111
期号：47
页码：16712-16717
DOI：10.1073/pnas.1411899111
语种：English
出版社：The National Academy of Sciences of the United States of America
摘要：SignificanceA robust debate about the effects of climate change on conflict occurrences has attained wide public and policy attention, with sub-Saharan Africa generally viewed as most susceptible to increased conflict risk. Using a new disaggregated dataset of violence and climate anomaly measures (temperature and precipitation variations from normal) for sub-Saharan Africa 1980-2012, we consider political, economic, and geographic factors, not only climate metrics, in assessing the chances of increased violence. The location and timing of violence are influenced less by climate anomalies than by key political, economic, and geographic factors. Overall, the temperature effect is statistically significant, but important inconsistencies in the relationship between temperature extremes and conflict are evident in more nuanced relationships than have been previously identified. Ongoing debates in the academic community and in the public policy arena continue without clear resolution about the significance of global climate change for the risk of increased conflict. Sub-Saharan Africa is generally agreed to be the region most vulnerable to such climate impacts. Using a large database of conflict events and detailed climatological data covering the period 1980-2012, we apply a multilevel modeling technique that allows for a more nuanced understanding of a climate-conflict link than has been seen heretofore. In the aggregate, high temperature extremes are associated with more conflict; however, different types of conflict and different subregions do not show consistent relationship with temperature deviations. Precipitation deviations, both high and low, are generally not significant. The location and timing of violence are influenced less by climate anomalies (temperature or precipitation variations from normal) than by key political, economic, and geographic factors. We find important distinctions in the relationship between temperature extremes and conflict by using multiple methods of analysis and by exploiting our time-series cross-sectional dataset for disaggregated analyses.
