Hazy days. Ash from Mount Kelud drove 100,000 people from their homes, but long-term, gradual changes in climate are more likely to make people move permanently.

Hazy days. Ash from Mount Kelud drove 100,000 people from their homes, but long-term, gradual changes in climate are more likely to make people move permanently.


A volcanic eruption probably wouldn't make you move

When Mount Kelud, a volcano on the Indonesian island of Java, erupted in February, nearby villages were blanketed with ash and gravel and an estimated 100,000 people fled their homes, filling the news with images of displaced people. But when it comes to the decision to permanently relocate, natural disasters like this one don’t matter as much as less dramatic changes in local climate and weather. A new study shows that the main factors influencing the families’ decisions to permanently relocate between Indonesian provinces are changes in the average temperature of a region and, to a lesser extent, changes in the annual rainfall. The findings may help scientists predict the likely number and location of climate change refugees.

Several previous studies have examined the effects of temperature change, rainfall variation, and natural disasters individually on migration, but scientists have never managed to accurately compare the effects of these factors. One reason is that it can be hard to separate the effects of each variable; for example, if the temperature in an area changes, the rainfall may change as well, so it is difficult to find out which one caused a particular migration. What’s more, many studies collect data for only a short period of time and don't track people after they move, making it impossible for researchers to establish whether displaced people stay in their new homes permanently or return after a few months or years.

In the new research, social scientists from Princeton University; the University of California, Berkeley; and the National Bureau of Economic Research in Cambridge, Massachusetts, looked at migration between provinces in Indonesia. The country, the world's largest archipelago, has a population of more than 250 million, about 40% of whom depend on agriculture and most of whom live near the coasts. Those factors put Indonesia at extreme risk from the changes in global weather patterns and rising sea levels wrought by climate change. Furthermore, it is on a plate boundary, which exposes it to major earthquakes and volcanic eruptions. (Indonesia has more active volcanoes than anywhere else on Earth.) Fortunately for the team, however, better data are available for Indonesia than for most developing countries. The Indonesia Family Life Survey, conducted by the RAND Corporation based in Santa Monica, California, follows more than 7000 families from 13 of the most populous of Indonesia's provinces, contacting them every few years to see how their lives have changed.

The researchers looked for a relationship between the number of permanent relocations of entire households between Indonesian provinces and the average annual temperature and the average annual rainfall in each province. Lead author Pratikshya Bohra-Mishra of Princeton explains that considering temperature and rainfall separately helped the researchers disentangle the effects of the two, as in one area the temperature might change without any difference in rainfall, whereas in another rainfall alone might change and in a third both could change simultaneously. Determining how many people moved out of areas where each of those patterns occurred allowed the team to work out the relative importance of each.

The researchers found that average temperature had by far the greatest impact on the number of people moving and staying in their new homes, they report online today in the Proceedings of the National Academy of Sciences. In places where the average temperature is below 25ºC, any increase in temperature made people less likely to move away. In areas with an average temperature above 25ºC, however, rising temperatures drove people to move. And the hotter it got, the more likely people were to relocate. For example, increasing the temperature from 26ºC to 27ºC raised the probability of migration by 0.8%, whereas increases from 27ºC to 28ºC raised it by 1.4%. Bohra-Mishra says one possible reason for the difference is that higher temperatures decrease crop yields, slashing farmers’ income and driving them to look for new homes. Rainfall had a similar, but much smaller, effect. Below an annual total of about 2.2 meters, increases in rainfall reduced the number of households leaving a province, whereas above this threshold, increased rainfall led to more migration.

The average temperature in Indonesia is about 25.1ºC. Climate forecasts suggest that, depending on emissions, this value is likely to increase to between 26.9º and 27.4º by 2100. If current trends continue, the researchers calculate, this could increase the number of people leaving warmer provinces of Indonesia like Jakarta, whereas provinces like West Sumatra, which now has a temperature of about 22ºC, could see more people staying put.

The team also looked at the frequency and severity of natural disasters such as earthquakes and volcanic eruptions. In most cases, the effects of such events on permanent relocation were statistically insignificant, suggesting that people who left their home provinces after disasters tended to return as soon as conditions were back to normal. The only natural disasters that consistently increased migration were landslides, but even here the increase was very slight. A 1% increase in the number of people killed by landslides, for example, increased migration by 0.0006%.

Environmental economist Valerie Mueller of the International Food Policy Research Institute in Washington, D.C., says that the study provides a valuable insight into what causes people to permanently relocate and uses an unusually thorough data set, adding that she can count the number of rigorous studies on this subject  “on the fingers of my one hand.” However, she says more detail is needed for the research to be useful in guiding policy. “Where are the people going?” she asks. “Where are the migration hot spots, for example, and how might that influence the allocation of resources?” If people start to suffer from illnesses, for instance, they may migrate to be near a hospital, whereas appropriate local investment in health care might allow them to stay put, she says. One thing is clear, however: As seemingly subtle changes in temperature increase the number of climate change refugees around the world, scientists like Bohra-Mishra will have a lot more data to work with.