About 4 to 5 million years ago, the Earth was warmer than today. Now that greenhouse gas pollution has the planet's temperature rising again, researchers want to know more about why this early Pliocene period was so warm, with the hopes of improving future climate predictions.

A new study in the journal Nature concludes that it is difficult to model the exact conditions behind the pattern of warming in the early Pliocene. None of the proposed mechanisms-from high carbon dioxide levels to changes in global ocean circulation patterns-can explain why the ancient warm period looks the way it does.

The findings raise the question of whether climate models for the early Pliocene might be missing key processes. If researchers can uncover these missing processes, they can apply them to models of modern climate and improve future climate predictions, says San Francisco State University Assistant Professor of Geosciences Petra Dekens, a co-author on the Nature study.

"It's very hard to look at a climate record from the past and say this directly applies to modern climate," Dekens says. "But what it does do is help us think about what the gaps might be in our models, what are the uncertainties in our current models, and whether those uncertainties could be important."

While the early Pliocene has attracted the interest of researchers looking to understand today's warming climate, the planet was a markedly different place 4 to 5 million years ago. In particular, while the highest sea surface temperatures were relatively stable, there were only small differences in sea surface temperature moving from the equator to the poles, or moving east to west.

Things began to change after the early Pliocene. The pool of warm water spreading out from the equator began to shrink toward lower latitudes, and east-west differences in sea surface temperature began to develop. Overall, the planet's climate shifted toward cooler temperatures.

The Nature authors were able to see this broad shift in climate by examining a wealth of already-published data on sea surface temperatures. "Very few of these records existed 10 years ago, but we're now at this point where we have records in high latitudes and low latitudes," Dekens said.

Ancient sea surface temperatures can be reconstructed in a variety of ways. Dekens studies sea surface temperatures by looking at the ratio of minerals like magnesium and calcium in the shells of tiny single-celled sea animals, found in sediment cores drawn from the deep sea. These ratios reflect sea temperature at the time the shells were deposited.

The records allowed the researchers to see that the early Pliocene climate was "structurally different" from today's climate, Dekens said. "It's not just that the absolute temperature in any one location is different, it's that the patterns are different."

Dekens' colleagues constructed several models to try and recreate the sea surface temperature conditions of the early Pliocene, but none of the expected "drivers" of climate that they tested could account for all the major features of the ancient climate.

For instance, the researchers found that increases in greenhouse gases and changes in ocean circulation could not reproduce the early Pliocene climate in their models.

Other tweaks to the models--reducing the reflection of sunlight by tropical clouds, for instance-did bring the models closer to matching the early Pliocene. But they still fell short of explaining the full pattern.

The paper, "Patterns and mechanisms of early Pliocene warmth," was published today in the 4 April issue of the journal Nature.