A few years ago geologist Carlos Jaramillo stood in a man-made canyon in Panama staring at rocks he knew to be 20 million years old, and shook his head in confusion. According to conventional geologic theory, the Panamanian Isthmus didn't emerge from the sea until just a few million years ago. So what was a 20 million-year-old fossilized tree doing there?

A new body of data emerging from such questions threatens to upend what geologists thought they knew about our planet. The Isthmus of Panama plays an outsized role in ocean circulation and may be a reason that our planet currently undergoes ice ages, so the new theory could rewrite not just the history of continents and biology, but also global climate.

Science owes this research to an unlikely source: a public works project. Panama is expanding its namesake canal, which has required monumental excavations to accommodate the world's growing fleet of ships that are too large for the original channels. Those digs have exposed an abundance of ancient rocks across a land normally choked by jungle. Jaramillo, a staff geologist at the Smithsonian Tropical Research Institute in Panama City, is leading a team of about 40 scientists who are taking advantage of this brief opportunity to study the rocks before they again surrender to plants or water.

Until recently, scientific theory has dictated that up to about three million years ago, the Atlantic and Pacific formed a single wide and deep sea between the American continents. As continental plates collided, a chain of islands between the two rose up, forming the Isthmus of Panama and ending what preeminent 20th-century paleontologist George Gaylord Simpson famously described as South America's "splendid isolation."

But Jaramillo and his colleagues have proposed a new model: most of Panama existed as it does today 12 million years ago, with shallow, narrow channels connecting the two oceans periodically after that. The results are detailed in a recent issue of the Geological Society of America Bulletin (pdf), with more details in press in the Journal of Geophysical Research (pdf).

The discrepancy between the two theories is no small matter. The three-million-year time frame neatly accounted for an important sequence of events that began about the same time. The current global cycle of glaciation dates to this period and might have been triggered by a transformation of the world's ocean currents, which a slender rib of land separating Atlantic and Pacific would naturally explain. New currents began carrying warmth to northern Europe and precipitation to the Arctic. The Atlantic grew saltier and warmer; the Pacific grew more nutrient-rich. Flora and fauna began traipsing between the two American continents, often extinguishing each other. In Africa a savanna formed, which may have nudged forward the evolution of our species.

So if the new theory is right, and the oceans were separated much earlier—then what triggered all of those epochal events?

"This is the most interesting for me, because if you tell somebody living in Nepal that the isthmus rises three million years ago versus 10 million—who cares?" Jaramillo says. "But think about this: having ice in the Arctic is the reason we're in the climate we are right now, and we still don't have a clear mechanism for it. That's very interesting, no? How can we even think about modeling the climate of the next 100 years if we cannot model how to produce such a big feature of our climate today?"

Climatologists will have to begin taking seriously alternative theories about why the ice ages began, says geologist Peter Molnar of University of Colorado at Boulder's Cooperative Institute for Research in Environmental Sciences. His favored alternative theory involves precipitation in Indonesia. "I think this will turn the field on its head," Molnar says.

Not everyone buys the new time frame. Geologist Tony Coates, a former director of scientific research programs at the Smithsonian Institution who fathered the original theory 30 years ago, says Jaramillo's team "has done some beautiful geology," but has ignored earlier lines of evidence, such as the genetics of fossil organisms on either side of the isthmus. Coates is mostly retired, but plans to write a rebuttal. Nevertheless, he says, these new questions are important to answer because the rise of the isthmus was the "last big episode of global change."

The effect of a new land mass on ocean currents is remarkably tricky to understand. Supercomputers in France have been examining how close the continents had to be before currents changed. The preliminary results, analyzed by paleoclimatologist Pierre Sepulchre of the Climate and Environment Laboratory, suggest that with any channel deeper than 200 meters currents behave as though there's an entire ocean there. But at shallower depths currents abruptly hit a wall. The surprise finding: at less than 50 meters currents arise that have never before been simulated.

As for Panama's role as a transcontinental bridge, biologists in Colombia and Sweden have conducted a meta-study examining migration between the Americas. They've found a bump in animal migration three million years ago, but a spike in flora migration 10 million years ago, in support of the revisionist time frame. It is not so unusual for plants to cross where animals dare not tread. The famous 19th-century naturalist Alfred Russel Wallace discovered that animals do not cross the strait separating Bali and Lombok, as if observing an invisible line.

Jaramillo's team aims to gain more insight into these subtleties in the next two years, as Canal Zone excavations continue. But they have a longer lens in mind as well: They have been setting aside hundreds of rock samples in hopes that future scientists can study them with yet-to-be-developed geologic dating methods. After all, it could be decades before geologists have another chance to delve physically into Panama's ancient history, Jaramillo says.

"Most of the rocks we have seen in the past three years have already gone," Jaramillo says. "Some entire mountains that we used to collect from two years ago are now 50 meters below water. So we know we're in a race against time."