California Institute of Technology, USA

A planetary scientist foresees a shift in the debate about Earth's heat flow.

Measurements of the heat coming out of Earth's interior have long posed a puzzle for understanding the planet's history.

Earth's heat output is estimated to be around 44 terawatts, about twice that expected from radioactive decay. The difference can be attributed to cooling of the deep Earth, implying a present-day cooling rate of 100 kelvin per billion years. But simple models with this much cooling 'blow up' when they are run back in time, predicting ridiculous temperatures for the early Earth. Acceptable models rely on unconventional deviations from the usual simple scaling laws for mantle convection. This is an attractive but untested idea.

I and many others have wondered whether an alternative explanation is that today's heat flow is higher than the average for the past half a billion years. Such fluctuations could arise as a result of the dispersal and accumulation of continental land masses.

A recent paper (J. Korenaga Earth Planet. Sci. Lett. 257, 350–358; 2007) assessed this possibility by taking advantage of a long-known connection between sea level and the heat flow from sea-floor spreading. It finds little room for more than a few percent fluctuation in heat flow around its long–term decline.

I think this pushes the problem back into the realm of models, focusing attention on plate tectonics, the deep water cycle (because water affects how rocks flow), and perhaps even the long-standing question of whether Earth's mantle is well mixed from top to bottom.

On a decadal timescale, we can hope that better measurements of heat generation and flow will be combined with more realistic theory. Like many central Earth science questions, the heat-flow problem resists quick resolution.