Nature: El Niño tied to civil conflicts in tropical countries

An interesting summary of recent research published in the journal Nature has appeared today in an article inside Ars Technica Nobel Intent section:

Political conflicts are extremely complex, and we almost never understand all the factors that are involved in their timing, the course they take, and their eventual outcome. In this week’s Nature, a paper suggests a new variable to consider: the climate. According to researchers from Columbia and Princeton, there is good evidence that global climate variations can play a role in the onset of civil conflicts.

The El Niño-Southern Oscillation cycle, known as ENSO, is a climatic pattern that repeats about every 5 years. Variations in water temperatures and air pressure in the Pacific Ocean cause El Niño years to be warm and dry, while La Niña years are cooler and wetter. These oscillations are felt most strongly in tropical countries, while mid-latitude areas are much less affected.

The authors of the study hypothesized that these cycles might play a role in the onset of civil conflicts. Although some research has been done on the correlation between climate and conflict, methodological difficulties and inconclusive results have caused lots of confusion. Climate studies are difficult because there is no Earth we can control to experiment with; we are stuck examining global patterns with few controls. However, in this study, ENSO provides a convenient experimental setup: volatile El Niño years provide a “treatment” group, while calmer La Niña years serve as a control.

175 countries were included in the dataset, including 93 tropical countries that are highly prone to the effects of El Niño and 82 other countries that are weakly affected by ENSO. Countries were classified as experiencing "conflict onset" in a given year if more than 25 people had died as a direct result of a new political dispute between two groups. Conflicts from 1950 to 2004 were included. Then the researchers calculated the annual conflict risk (or ACR) for both the tropical countries and the weakly affected countries. The ACR is the probability that a randomly selected country from the group experienced conflict onset in a particular year.

For the countries that aren’t highly affected by ENSO, such as Greece, Afghanistan, and Sweden, the ACR was about 2 percent in both El Niño and La Niña years, indicating that these climate cycles are unlikely to affect civil conflict in these countries. However, the ACR for tropical countries such as Australia, Sudan, and Trinidad doubled during El Niño years, increasing from 3 percent to 6 percent. From their analyses, the researchers concluded that the ENSO cycle may have affected 21 percent of civil conflicts since 1950.

The results are remarkably robust; the researchers repeated the analysis with various types of statistical models and with different ENSO indices, and their conclusions remained the same. Furthermore, the results held up even when high-conflict countries were excluded from the dataset, and when other variables, such as a country's age structure, income growth, and agricultural reliance, were included in the analyses.

Two additional findings were particularly intriguing. Many of the conflicts affected by ENSO are particularly deadly recurring conflicts. By changing the requirement for the length of "peaceful periods" between conflicts, the researchers found that the relationship between ENSO and large conflicts decreased. Additionally, low-income countries were the hardest hit by El Niño years, indicating that poorer countries are particularly sensitive to ENSO patterns.

Although the relationship between ENSO and civil conflicts is quite clear, the reasons behind this correlation are still not understood. Warm, dry El Niño years might decrease agricultural output, stressing a country’s resources and increasing food prices. ENSO patterns affect the frequencies of natural disasters, such as hurricanes and cyclones, which put countries at risk for upheaval. Extreme conditions can also cause psychological stress and alter human behavior. Finally, ENSO is a widespread phenomenon, and may cause suboptimal conditions on a large scale.

 In order to use the results of this study to predict or alleviate conflicts, we need to better understand the mechanisms driving the relationship between climate patterns and political conflict.

Civil wars, of course, have been the primary mode of fighting globally since the 1960s at least.... This research may help to illuminate one other aspect of underlying motivations and triggers for this type of conflict.