The picture that has emerged is of a planet being driven in towards its star during the late stages of planet formation by the gravitational force between the planet and the remnant of the swirling disc of dust out of which the planets formed. But, if this could happen to a gas-giant planet such as Jupiter, why could it not happen to an ice-giant planet like Uranus or Neptune?

“Exactly,” says Christophe Sotin of the University of Nantes in France. If such a planet migrated towards its star, the heat would cause its huge “mantle” of ice to melt, resulting in a planet with a gigantic ocean. “Observational techniques are not good enough to detect such ‘ocean planets’ yet,” Sotin says. “But we think they could be very common in our galaxy.”

The possibility of ocean planets is not exactly new; they were first proposed by David Stevenson of the California Institute of Technology in Pasadena. However, Sotin and his colleagues are the first to work out the details and publish them.

Sotin learned about the possibility of ocean planets from colleagues. He began to figure out what they might look like when he discovered that his colleagues thought they could have oceans thousands of kilometres deep. “I realised that was totally preposterous,” Sotin says. He is an expert on ice that forms under high-pressure conditions; he has used such expertise to deduce the internal structure of Jupiter’s giant ice-moon, Callisto. Usually, the melting point of ice goes down at high pressure. This would mean that deep down in an ice planet, where the ice was being squeezed by the weight of all the material bearing down from above, it should be liquid. “This is what led my colleagues to think that all the ice in an ice planet would melt, making an ocean thousands of kilometres deep,” says Sotin.

However, he knew that at ultra-high pressure – the kind found deep in an ice giant – ice enters a different “phase” known as “Ice II.” Instead of its melting point dropping as the pressure goes up, it increases, so the ice is prevented from melting. “Consequently, you only get an ocean about 100 kilometres deep, not thousands of kilometres deep,” Sotin says.

Sotin and his colleagues have calculated the depth of liquid water for ice-giant planets less than 10 times the mass of the Earth. The depth depends on the surface temperature of the planet, which in turn depends on the atmosphere, which may contain heat-trapping greenhouse gases such as carbon dioxide and water vapour. Since the atmospheric composition of such a planet is not well known, Sotin has calculated the ocean depth for different surface temperatures.

If the surface temperature is 7C, the ocean will be 72 kilometres deep and the seabed temperature and pressure 35C and 11,000 atmospheres respectively. For a higher surface temperature, the ocean will be deeper, and vice versa. For instance, a surface temperature of 30C leads to an enormous ocean depth of 133 kilometres.

A planet about eight times the mass of the Earth would have about twice the Earth’s diameter and gravity at the surface about 50 per cent higher. An ocean 100 kilometres deep is 10 times the depth of the deepest ocean trench and 40 times the average depth of the ocean. With the ocean covering an area six times that of the Earth, we are therefore talking about a volume 240 times greater. “The Earth’s ocean is a mere puddle by comparison,” Sotin says.

Others researchers have also suggested the existence of water worlds, but ones actually born close in to the star, rather than ones that have migrated there from far away. According to the simulations of a team led by Sean Raymond of the University of Washington, Seattle, it is possible to make worlds with more than 100 times the Earth’s water.

On such ocean planets, there would be no possibility of any continents poking above the water, as the ocean would be 10 times deeper than the height of Everest. The climate on such a world would depend on things such as the orbit of the planet and the orientation of the axis about which it spins, so in general it is not predictable.

But the planet’s weather is likely to be extreme. “On Earth, hurricanes run out of steam when they encounter land, but, on an ocean world, there would be no land to encounter,” Sotin says. “There might be super hurricanes.” The winds would pile up enormous ocean swells, which would never run in to land. “We’re talking about the kind of mountainous waves you get in the Pacific, only more so,” Sotin says.

What about the prospects for life? Well, on Earth, we believe life got started in nutrient-rich waters close to volcanic vents on the sea floor. The crucial factor, therefore, is how much ice at the foot of the ocean separates the water from the planet’s volcanic interior. “If it is a thin layer, the prospects for life are good,” says Sotin. “If it’s thick, they’re not.”

However, a minority of scientists think that volcanism isn’t necessary to get life started. These are the people, led by Chandra Wickramasinghe of the University of Wales College, Cardiff, who believe that life is “seeded” from space, with micro-organisms spread from star system to star system in cometary dust. If this “interstellar panspermia” idea is correct, then ocean planets will certainly contain life. Essential nutrients could rain down from meteorites continually bombarding the ocean from space.

Ocean planets are also exciting to planetary scientists because they ought to be relatively easy to detect. Currently, our observational techniques have difficulty detecting even a giant mass such as Jupiter, and such bodies are not detected directly, only by the effect their gravity has on their parent star.

The race is on to detect an extrasolar planet directly – by the light it reflects from its sun. For an Earth-like planet – the Holy Grail of planetary searches – this is immensely difficult. “However, for an ocean planet, far bigger and with an ocean shining like a giant silvered mirror, it should be much easier,” Sotin says.

Currently, several space-based observatories are being built to detect extrasolar planets directly. They include the European Space Agency’s project Darwin, a flotilla of six spacecraft flying in formation, each carrying a two-metre infrared telescope; and NASA’s Terrestrial Planet Finder, an eight-metre visible-light telescope. If ocean planets really do exist, then the odds are that it will be one of their number, rather than an Earth-mass planet, that will be discovered first. “If so,” Sotin says, “We’d better start learning dolphin!”