First author

One often-cited concern about global warming is that it could cause Greenland's ice sheet to melt. To better understand what controls the ice sheet's volume, researchers want to know how it first grew to cover much of Greenland around 3 million years ago. Existing theories implicate differing drivers of glaciation — including tectonic closure of the Panama seaway, uplift of the Rocky and Himalayan mountains, changes in El Niño–Southern Oscillation and variations in atmospheric carbon dioxide concentrations. Dan Lunt at the University of Bristol, UK, and his colleagues modelled each theory, and report that a significant fall in atmospheric CO2 concentrations (and, subsequently, cooler summers) drove ice-sheet growth (see page 1102). Lunt explains why he needs new clues to reconstruct past climates.

Was there a favoured theory at the outset of this work?

One popular hypothesis is that the closure of the Panama seaway caused a rise in the salinity gradient between the Pacific and Atlantic oceans, changing ocean circulation and ultimately enhancing snowfall over Greenland. It's popular because data from the Caribbean and North Atlantic indicate that the closure coincided with glaciation.

What is new about your approach?

First, we used a state-of-the-art climate model that takes months to simulate ocean, atmosphere and vegetation components. But climate models do not include variations in topography, which are important to glaciation dynamics. So we also used a high-resolution ice-sheet model to resolve the dynamics of ice formation at high altitudes.

Was there a Eureka moment?

Yes. I was blown away when the models produced a map of a huge Greenland ice sheet under low CO2 conditions, which we hadn't expected. Now we need more indicators of the range of past CO2 concentrations to better model past climate.

Don't ice cores provide that information?

Ice cores are amazing records, but those we have stop about 650,000 years ago. There is no direct record further back in time, and proxy indicators of CO2, such as marine sediment isotopes, are open to interpretation. We need multiple indicators to confirm past CO2 values. The search is on!

Which past climate era most interests you?

This period of the Pliocene, about 3 million years ago. Because we are now approaching the same atmospheric CO2 levels seen then — about 400 parts per million — it provides a potential analogue of future climate change.