Mars has lots of water, but future astronauts won't exactly be able to scoop it into bottles -- it's generally trapped in ice deposits below the surface. Scientists from Penn State think climate change lasting millions of years once warmed it enough to let the water flow free on the surface, however. That might have created large lakes in Gale Crater (above) and etched out channels and other water-based features on the Red Planet.

Researchers have long thought that canyons and valleys on Mars were caused by flowing water. However, those features were formed billions of years ago when the planet was frozen, and would require millions of meters of rain. Current Mars climate models predict that warming caused by volcanoes or meteor impacts could only account for hundreds, not millions of meters of rainfall. So how did the planet warm enough to create the amount needed?

The Penn State team figured out that the Mars could have experienced warming periods caused by a familiar culprit -- carbon dioxide. It was emitted in Mars' early years by volcanoes, cooling magma and seepage from the crust. As CO2 and hydrogen gradually built up, they warmed the atmosphere and caused large amounts of rainfall.

After up to 10 million years, the rain reabsorbed more CO2 than the volcanoes and magma could produce, storing carbon in the ground. The planet then settled back into an ice age, but the time-period was sufficient for the rain to carve out dramatic features. On Earth, the Grand Canyon "only" took around 16 million years to form for example, the team points out.

"Mars is in this precarious position where it's at the outer edge of the habitable zone," said Penn State grad student Natasha Batalha. "It's receiving less solar flux, so you start at a glaciated state. There is volcanic outgassing, but because you are colder, you don't get the same deposition of carbon back into the planet's surface. Instead, you get this atmospheric buildup and your planet slowly starts to rise in temperature."

If Mars' tectonic activity in its early years was similar to Earth's, the team's model would account for the relatively large amount of precipitation. "But that's a big debate," says the paper's co-author Jim Kasting. "A lot of people don't think Mars ever had [plate tectonics like Earth]."

However, there is a way to test the theory. When CO2 levels in the atmosphere were at a peak, they would've caused rain so acidic that it dissolved surface carbonate rocks and deposited them underground. "So if the next Mars mission was able to dig down deeper, you might be able to uncover these different carbonates," says Batalha. "That would be a sort of smoking gun for the carbon dioxide."