Two billion year-old water pockets and a revised deep hydrogen content are good news for Earth's vast subsurface biosphere, and could offer clues to life on Mars and much further beyond.

Excitement over the Curiosity rover's recently reported detection of a 'spike' in localized atmospheric methane - persisting over a couple of months - is well founded. It's possible that this represents a very real clue to past or present life on Mars. Or rather, life in Mars.

The great majority of methane that we find here on Earth (whether in the air or in subsurface deposits) has a biological origin, clearly marked by the preference biological systems have for lighter isotopes - carbon-12 over carbon-13 for example. This methane is generated by methanogenesis, a metabolic process that seems to be confined to members of the domain of single-celled organisms called the Archaea.

There's more than one chemical pathway for making methane, but the most obvious is the combination of carbon dioxide with molecular hydrogen, and that's precisely the reaction that a slew of methanogenic archaea latch onto. Molecular hydrogen is a potent source of chemical energy, and other organisms such as sulfate-reducing bacteria also gobble it up. So where do they find that hydrogen?

One source is where rock and water sit together. Radioactivity from rocks laced with elements like uranium can break up water molecules (the process of radiolysis), and the geochemical process of serpentization also spits out molecular hydrogen in abundance. Active hydrothermal vent systems on ocean beds are one environment where hydrogen is readily made, and methanogenic organisms thrive there. But what about the subsurfaces of Earth's continents, the most ancient parts of the lithosphere?

Recent discoveries of isolated pockets of highly saline water at extraordinary depths of 1-2 miles in South African and Canadian mines have revealed these reservoirs to be tens of millions to billions of years old - the current record holder having formed somewhere between 1.5 and 2.6 billion years ago. Places like these are, in relative terms, packed with chemical energy that life can exploit, and does. Except it's not been clear that extrapolating the local production of molecular hydrogen in such reservoirs to a global total yields anything worth getting excited about.

Now a new study by Lollar, Onstott, Lacrampe-Couloume, & Ballentine reported in Nature, suggests that the deep (5km) continental zones could indeed be a major producer of hydrogen. Specifically, the most ancient Precambrian continental subsurface (rock older than about 540 million years), could generate molecular hydrogen at a rate 40 to 250 times higher than previously thought - a production on par with that associated with the much younger marine lithosphere. This Precambrian material exists in about 70% of the Earth's continental area, and could contain as much water as all surface rivers, swamps, and lakes.

The conclusion is that the global production of molecular hydrogen needs an upward revision and, most critically, at least half of that is coming from the deep, ancient continental subsurface - which is not dry, not inert, and seems to be filled with life.

Connecting these discoveries to methane on Mars is, at present, a stretch. But it's not unreasonable to suppose that the ancient martian subsurface could resemble Earth's present Precambrian subsurface environment - with long undisturbed water sitting in fractures, and a healthy molecular hydrogen production. Extinct or extant life, however it operated, would surely exploit an energy source like this, and the products could eventually find their way to the surface.

If time, power, and the luck of another methane spike allows Curiosity to examine the isotopic ratios in the gas, we might get a critical hint of any biochemical origins - with a ready explanation on hand from our very own planetary depths.

On a far grander scale, the idea of purely geophysical 'habitability' in distant exoplanets - biospheres only ever driven by the inner workings of a world, without surface life - is an important one. It may be the height of conceit to imagine that the externally visible biosphere of the Earth is a template for the majority of life in the universe.

Figuring out how detectable signatures of troglodytic life could bubble up from below would go a long way in encouraging us to target worlds that we might otherwise ignore. So too would a better understanding of molecular hydrogen generation in a planetary crust, where the original ingredients of a world - from the elemental mix to the availability of radioactive nuclei - are set by still deeper cosmic rhythms of stellar life and death.

Traces of methane on Mars, and deep hydrogen on Earth may seem remote and tenuous, but the implications could be enormous.