Ornithologist Henry Streby was happy enough that songbirds weighing less than two nickels managed to carry tiny electronic sensors from Tennessee to Colombia and back.
Then he looked at the flight paths taken by several of the birds.
The data showed that five of his recently returned golden-winged warblers fled their Appalachian Mountain breeding ground and winged back to the Gulf of Mexico a day or two ahead of a massive thunderstorm cell that would later spawn 84 tornadoes and kill at least 35 people.
Streby, a National Science Foundation visiting research scholar at UC Berkeley, thinks the birds may have been reacting to very low-frequency sound waves produced by the distant, approaching storm, according to a study published this week in the journal Current Biology.
"Everybody knows that birds can respond to changes in barometric pressure, wind speed and wind direction, and cloud cover — all the things that come with the front of a storm," Streby said. "But these birds left long before any of those things happened."
At first, Streby was skeptical that the birds had even left Appalachia. Maybe the batteries malfunctioned on the tiny sensors, which record sunlight and time so that researchers can extrapolate their global position.
"We're not quite to the point of handing the birds cellphones and getting pinpoint GPS on these tiny, tiny songbirds," he said. "The purpose of our study was just to see if birds this small could carry these things so we could start tracking their migration."
The more Streby thought about it, the more he believed the migrations were real. It defied the odds that all five sensors could have spit out false daylight and timing information in five different ways, he said. The data suggested each bird took separate paths to evade the storm — some flew east, some flew west, but four eventually hooked south to the Florida panhandle, while one traced the east coast of Florida and continued to Cuba.
Obligate migrators, which fly long distances seasonally, are known to detour around storms, the study notes. But combining a full round-trip migration with a separate lengthy flight sparked by environmental factors has not been observed among such species, according to the study.
Streby and his colleagues from the University of Tennessee and the University of Minnesota were stumped about why the warblers would have left their coveted mating grounds after such an arduous migration. Then he and his graduate student assistants remembered how, back in late April, ferocious thunderstorms sent them fleeing for shelter in a Waffle House in Caryville, Tenn.
Still, weather and sensor data showed the birds would have departed 24 to 48 hours before the storms arrived. That means the supercell storm was between 250 miles and 600 miles away, the researchers found. At that distance, none of the wind, barometric pressure or other meteorological changes would have been evident, they calculated.
Previous research, however, has shown that birds can sense infrasound. These sound waves travel at frequencies well below the human audible range and at such long wavelengths that they can propagate for thousands of miles without petering out. (Indeed, the military has studied them to detect distant weapons explosions.)
Birds also can detect shifts in sound waves caused by the motion of their source, known as the Doppler effect.
That put the puzzle together plausibly enough to publish, Streby said.
"It's possible that something that hasn't been discovered explains it, but this really makes sense," he said. "We know for sure that [the storm] makes the sound, we know for sure that they hear the sound, and we know for sure that it travels far enough. And we also know that they can sense Doppler changes."
Biologist John McCormack, director-curator of the Moore Laboratory of Zoology at Occidental College, said he found its hypothesis plausible.
"It's something that just wasn't possible to observe prior to geolocators," said McCormack, who wasn't involved in the study.
Streby is hoping other researchers might reexamine data that they may have discarded as anomalous. Conducting a field experiment to test his hypothesis, he said, would be next to impossible. But nature might help.
"It's not that I'm hoping that next year or the year after there will be a giant storm that pushes through our study," Streby said. "But at the same time, if that storm does happen, it would be nice to see if this migration does or doesn't happen again."