Scientists have long conjectured that climate change would spur families in poor countries to migrate as ever-fiercer storms, floods and other disasters made rural life unbearable. But understanding what specific weather elements would cause people to leave has remained elusive.

Until now. A small but growing body of evidence is finally pointing to rising temperatures—and not headline-grabbing natural disasters—as the main environmental force permanently ousting people from their homes.

A new study published this week in the Proceedings of the National Academy of Sciences follows more than 7,000 households in Indonesia over 15 years to conclude that sudden disasters in fact have a much smaller impact on provincial migration than heat stress. Rainfall, it finds, also affects decisions to move, but far less so than rising temperatures.

"We can now say something intelligent about the conditions under which people move," said Michael Oppenheimer, a professor of geosciences and international affairs at Princeton University and a co-author of the study.

Oppenheimer said the analysis shows for the first time that when researchers looked at the same group of households over time, permanent moves to new locations "tended to be among the people responding to long-term climate changes, particularly temperature." He and others noted that declining crop yields resulting from the heat were likely a major cause of moving.

A dearth of reliable data
The study is the second major empirical examination of how climate change might affect migration at a local level. But efforts to replicate such studies, researchers say, are hampered by a dearth of reliable, long-term data in affected countries.

Indonesia was chosen in part because the country is the world's largest archipelago and is exposed to both geologic and climatic hazards. It also is densely populated with 40 percent of the labor force engaged in agriculture and more than 60 percent living in coastal areas—factors that contribute to Indonesia's climate vulnerability.

Researchers also were able to find household panel data with what the study called an "exceptional tracking rate" that allowed them to study migration before and after disasters, as well as climate variations from 1993 to 2007. The survey represents 83 percent of the Indonesian population from 13 of its 27 provinces.

Rainfall had what the authors called a "quadratic effect" in that where conditions were initially dry, a further decline in rainfall tended to spark migration. Likewise, in wetter conditions, an increase in rainfall also increased migration. But the moves were usually not permanent, "suggesting that rainfall may have a more profound impact on temporary and shorter-distance movement."

Meanwhile, disastrous events like floods, volcanos, landslides and earthquakes—the four events posing the highest risks in the provinces—led to "no large, consistent or significant effects on migration."

"Typically when you talk about environmental migrants, everybody has these images of helpless refugees," said Pratikshya Bohra-Mishra, a postdoctoral scholar at Princeton University's Woodrow Wilson School of Public and International Affairs and lead author of the report.

"What is important to understand is that environmental migration can really take other forms," she said.

The unknown factor of economics
Valerie Mueller of the International Food Policy Research Institute, who led the first major study published earlier this year, looking at the effects of heat, rainfall and natural disasters on migration in Pakistan, called the new research significant. But, she noted, the study has a limitation in that it does not look at the income levels of those migrating.

"It's always sort of implicit that it's the poor making the moves, but that's not always true," she said.

Mueller said the next challenge for researchers will be determining whether people fared well or suffered economically from their moves. She also noted the need to discover if the strong link between temperatures and migration is merely particular to Indonesia and Pakistan, or is relevant elsewhere.

"We have isolated studies throughout the world, but we haven't shown this systematically," she said. "We'd also like to be able to say whether people are better off or worse off when they respond in this way. It's a really hard question to study, because it is hard to identify the effects on migrants."

The absence of answers to those question so far, researchers acknowledge, makes it difficult for policymakers to draw strong conclusions about how to deal with climate-induced migration.

For now, said Oppenheimer, "People need to think about climate migration not with a disaster mode in their head."

Reprinted from Climatewire with permission from Environment & Energy Publishing, LLC., 202-628-6500