The images of Pakistan's 2010 devastating flood still haunt. Women up to their necks in the waters of the swollen Indus River, carrying children to safety. Men wading through the brown currents with bags of rice on their heads and young ones on their backs.

The numbers, too, were staggering. Triggered by unusually heavy monsoon rains, the fast-rising waters along the heavily populated Indus River Valley affected some 20 million people, leaving about 11 million homeless. More fierce floods followed with the next monsoon season and then the next, each year with headline-grabbing figures of millions left homeless.

Scientists called climate change a major contributing factor to Pakistan's increasingly intense flooding, and reports from environmental groups as well as global development agencies cited the floods as early examples of the mass migration that that global warming could induce.

But now a major new study that surveys family data over the course of two decades finds that flooding is actually not the most serious driver of long-term migration. Instead, heat stress—a phenomenon largely overlooked by international aid programs and winning scant attention in the national media—is far more likely to drive a male wage-earner in Pakistan out of his village in search of wages to replace lost agricultural yields.

Published yesterday in the journal Nature, the study found no "robust" effect of rainfall on migration patterns, with men slightly more likely to move out of a village in heavy rains. But, the researchers found, the results consistently showed that men do leave villages in response to extreme temperatures—and are 11 times more likely to do so when exposed to both heavy rains and unusually high temperatures.

"We are left with an overall picture that heat stress, not high rainfall, flooding or moisture, is most strongly associated with migration," the study notes. "This approach reveals a complex migratory response that is not fully consistent with common narratives of climate-induced migration."

Valerie Mueller, lead author and co-leader of the International Food Policy Research Institute's Development Strategy and Governance Division, added: "This is not to say people aren't moving in response to floods. Some of these villages are going to be inundated, and they will have to move. But that seems short-term."

Once purely the realm of climate scientists and environmentalists, the study of global warming-induced displacement has attracted a growing level of academic rigor from migration economists and others who focus on human movement. While early reports simply assumed that poor communities in vulnerable coastal areas would be forced en masse to move, more sophisticated examinations have started to factor in the sometimes surprising ways people adapt to disasters, what types of environmental stresses drive people to seek livelihoods elsewhere, and whether those moves help or hurt a migrant's well-being.

Getting underneath the drama and rhetoric
"There is an increasing body of research that is being done... that has been trying to dig down below the high-level rhetoric we had before in understanding the dynamics and trying to understand what the impacts are," said Susan Martin, director of the Institute for the Study of International Migration at Georgetown University.

Martin said the new study shares common ground with other emerging research that finds acute natural disasters like floods and typhoons may shatter villages, but many ultimately rebuild. People leave for far longer stretches of time in response to slow-onset and recurring situations like drought, desertification or sea-level rise that destroy fishing, agriculture and other natural resources that communities depend on for their livelihoods.

"What's interesting with this study is it takes the next step," she said.

Mueller and co-author Clark Gray, a geographer at the University of North Carolina, Chapel Hill, used a 21-year longitudinal survey conducted in rural Pakistan between 1991 and 2012. That survey was based on a previous one IFPRI conducted in the previous decade as well as two tracking studies. The team resurveyed heads of households to track the movement of family members over time. Then, using what the researchers call "discrete-time event history models," they measured the response of individuals based on wealth and other demographics to different weather patterns.

In Pakistan, catastrophic flooding regularly triggers outpourings of international aid and usually sets in motion long-practiced U.N. and Red Cross emergency response efforts. At the same time, she said, Pakistan has some social protections built in for victims, like a damage compensation program instituted in 2010. Those mechanisms both help villagers cope with losses and eventually even move back home.

Heat stress, on the other hand, rarely attracts global aid. Yet the ramifications are serious. In 2010 alone, the report notes, uncharacteristically high temperatures in Pakistan led to wheat grains maturing early. That year ended in a 13 percent decline in wheat yields with no social protection or international relief to help offset those damages.

Meanwhile, the study also examines whether extreme rainfall and heat affect agricultural income as another avenue of exploring the possible impact of migration. To do that, the researchers looked at how temperature fluctuations affected farm and non-farm wages and incomes.

The result: "Agriculture suffers tremendously when temperatures are extremely hot... wiping out over a third of farming income." The blazing heat also puts a crimp in non-farm income, but to a lesser extent.

"Interestingly, high rainfall increases all sources of income substantially. This analysis suggests one possible reason why heat stress drives migration, whereas extreme rainfall does not. Heat stress (unlike rainfall) provides a negative income shock," the report finds.

Finally, the researchers found that villagers who did not own land or other assets were more likely to move. Mueller associated the phenomenon in part to tenuous land rights in some areas that would leave families fearful that if they left their land, they would essentially forfeit it.

Poverty raises tendency to move
"Everyone is affected, but people with lower assets have higher odds of moving than people with land in response to extreme temperature," she said.

Lori Hunter, an associate professor of sociology at the University of Colorado, Boulder, who reviewed the study said she was impressed with the ways the researchers used existing data, noting that obtaining migration data is one of the hardest parts of studying the phenomenon.

Meanwhile, she added, the results could help reshape how policymakers approach disasters into thinking about a broader swath of stresses beyond "emergency relief," like helping households diversify their livelihoods and longer-term agriculture development.

Mueller agreed. "We need to think more carefully about how heat is affecting agriculture," she said. The next step, she noted, is trying to understand whether migration due to climate stress is in fact as negative as environmental groups and others paint it.

"We don't know if it's for better or for worse," she said. "Understanding why people move and when they're going if they're better off or worse is a missing puzzle."

Michael Kugelman, senior program associate for South and Southeast Asia at the Woodrow Wilson International Center for Scholars and an expert on Pakistan, called the study eye-opening and noted that water shortages also are a looming disaster there. But he also maintained that losing sight of the migration impact of natural disasters would be a mistake.

"Natural disasters are going to continue to be a huge spark for migration in Pakistan," Kugelman said. He noted that emergency responses remain poor in the country, while heavy deforestation is making flooding even worse.

"These things are not going to change overnight, and climate change is going to increase the chances of these big weather events happening," he said. "If there was no warfare in South Asia ... there would still be IDPS [internally displaced persons], and that will be because of natural disasters that are going to be there and probably intensify."

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