Mars has many signs of a watery past, yet today the surface is almost entirely dry. The question of what happened to the water has haunted planetary scientists for decades, and could have big implications for where we search for life, there and elsewhere. One team of scientists think the water isn't all gone, its just reacted with the Martian rocks and drawn into the mantle.

Mars does not have much of a magnetic field, and this absence means the atmosphere is exposed to ionizing radiation. This is thought to have split water molecules at the surface into hydrogen and oxygen atoms, with the hydrogen escaping Mars' low gravity while the oxygen reacted with surface materials.

Although Dr Jon Wade of Oxford University agrees that this explains a lot of Mars' missing water, he, like many others, does not accept it is the whole story. Instead, Wade and colleagues propose, in a study published in Nature, that some of the water remains trapped in rocks deep inside. “People have thought about this question for a long time, but never tested the theory of the water being absorbed as a result of simple rock reactions,” Wade said in a statement.

Wade noted that meteorites blasted off Mars in asteroid impacts, and eventually making their way to Earth, are chemically different compared to the rocks studied by Mars rovers.

Mars has more iron in its mantle than the Earth, and of course is much colder. In consequence, Wade argues, basalt rocks on the red planet should not be treated as exactly matching those on Earth. When he took these and other differences into account, Wade found Martian lavas can hold 25 percent more water than their nearest Earth equivalents.

Importantly, hydrated rocks float on Earth, and only become absorbed into the mantle after their water has been forced out, but under Martian conditions these saturated rocks will sink. Consequently, the paper claims, more than 9 percent of the Martian mantle (by volume) might be made up of water-storing minerals, compared to 4 percent on Earth.

With most of the water at great depth, the task of colonizing Mars, or even finding life there, looks harder.

Wade argues we need to add extra factors to our questions of what makes a planet suitable for life. If Earth had as much iron in its mantle as Mars, even with all its other advantages, it might also have become a desert world.

Wade's work is unlikely to be the final word on Martian water, however. Just this month another Nature paper proposed we had been misreading the signs and Mars had little water in the first place. This is a debate that may run for some time yet.