How Does Earth's Interior Work?

See allHide authors and affiliations

Science  01 Jul 2005:
Vol. 309, Issue 5731, pp. 87
DOI: 10.1126/science.309.5731.87

The plate tectonics revolution went only so deep. True, it made wonderful sense of most of the planet's geology. But that's something like understanding the face of Big Ben; there must be a lot more inside to understand about how and why it all works. In the case of Earth, there's another 6300 kilometers of rock and iron beneath the tectonic plates whose churnings constitute the inner workings of a planetary heat engine. Tectonic plates jostling about the surface are like the hands sweeping across the clock face: informative in many ways but largely mute as to what drives them.

A long way to go.

Grasping the workings of plate tectonics will require deeper probing.


Earth scientists inherited a rather simple picture of Earth's interior from their pre-plate tectonics colleagues. Earth was like an onion. Seismic waves passing through the deep Earth suggested that beneath the broken skin of plates lies a 2800-kilometer layer of rocky mantle overlying 3470 kilometers of molten and—at the center—solid iron. The mantle was further subdivided at a depth of 670 kilometers into upper and lower layers, with a hint of a layer a couple of hundred kilometers thick at the bottom of the lower mantle.

In the postrevolution era, the onion model continued to loom large. The dominant picture of Earth's inner workings divided the planet at the 670-kilometer depth, forming with the core a three-layer machine. Above the 670, the mantle churned slowly like a very shallow pot of boiling water, delivering heat and rock at mid-ocean ridges to make new crust and cool the interior and accepting cold sinking slabs of old plate at deep-sea trenches. A plume of hot rock might rise from just above the 670 to form a volcanic hot spot like Hawaii. But no hot rock rose up through the 670 barrier, and no cold rock sank down through it. Alternatively, argued a smaller contingent, the mantle churned from bottom to top like a deep stockpot, with plumes rising all the way from the core-mantle boundary.

Forty years of probing inner Earth with ever more sophisticated seismic imaging has boosted the view of the engine's complexity without much calming the debate about how it works. Imaging now clearly shows that the 670 is no absolute barrier. Slabs penetrate the boundary, although with difficulty. Layered-earth advocates have duly dropped their impenetrable boundary to 1000 kilometers or deeper. Or maybe there's a flexible, semipermeable boundary somewhere that limits mixing to only the most insistent slabs or plumes.

Now seismic imaging is also outlining two great globs of mantle rock standing beneath Africa and the Pacific like pistons. Researchers disagree whether they are hotter than average and rising under their own buoyancy, denser and sinking, or merely passively being carried upward by adjacent currents. Thin lenses of partially melted rock dot the mantle bottom, perhaps marking the bottom of plumes, or perhaps not. Geochemists reading the entrails of elements and isotopes in mantle-derived rocks find signs of five long-lived “reservoirs” that must have resisted mixing in the mantle for billions of years. But they haven't a clue where in the depths of the mantle those reservoirs might be hiding.

How can we disassemble the increasingly complex planetary machine and find what makes it tick? With more of the same, plus a large dose of patience. After all, plate tectonics was more than a half-century in the making, and those revolutionaries had to look little deeper than the sea floor.

Seismic imaging will continue to improve as better seismometers are spread more evenly about the globe. Seismic data are already distinguishing between temperature and compositional effects, painting an even more complex picture of mantle structure. Mineral physicists working in the lab will tease out more properties of rock under deep mantle conditions to inform interpretation of the seismic data, although still handicapped by the uncertain details of mantle composition. And modelers will more faithfully simulate the whole machine, drawing on seismics, mineral physics, and subtle geophysical observations such as gravity variations. Another 40 years should do it.

View Abstract

Stay Connected to Science


Navigate This Article