Read our COVID-19 research and news.

Weak spot. As a tectonic plate (red) moves and rotates, it “inherits” previously formed weak zones (in blue) that then continue to grow to ultimately form a plate boundary.

Weak spot. As a tectonic plate (red) moves and rotates, it “inherits” previously formed weak zones (in blue) that then continue to grow to ultimately form a plate boundary.

David Bercovici

How Earth Became a Jigsaw Puzzle

Plate tectonics—the worldwide face-lift that continually, but slowly, reshapes the surface of Earth—is apparently unique to our planet; at the very least, it occurs nowhere else in the solar system. This process is responsible for volcanoes, earthquakes, and mountain building, and possibly for helping harbor early life on Earth.  Yet there is much we still don’t know about what drives it and when it began. Now, a new study may help resolve one question: when and how Earth’s rigid outer shell, or lithosphere, first divided into plates and their global dance began.

Evidence of Earth’s earliest geologic history is scant, thanks to the constant recycling of our planet’s surface. But geologists do have a few clues. One line of evidence comes from hardy crystals called zircons, found primarily in granite—the formation of granite requires subduction, the sinking of a lithospheric slab into the mantle where it partially melts to produce so-called granitic magma. Based on the very existence of ancient zircons, some geologists surmise that subduction occurred, at least intermittently, sometime around 4 billion years ago. Other evidence to bolster this claim includes rock sequences from the deepest point on Earth, the Mariana Trench in the Pacific Ocean. The trench, formed where the Pacific Plate is sinking into the mantle, contains 4.4-billion-year-old lavas that may have resulted from the earliest subduction zone on Earth

But a wealth of other geochemical and geologic data—from rock assemblages within ancient continental crust to studies of varying chemical contents of granitic rocks—suggest that plate tectonics really went global about 2.7 billion to 3 billion years ago. By then, the lithosphere, rather than forming a solid shell around the planet, had divided into dozens of thick plates. Driven by circulation in the underlying mantle, the plates slid past each other, pulled apart, or collided. The jigsaw didn’t fit quite the way it does now, but the pieces were already moving around.

This discrepancy has produced some heated debate, but there’s a way to reconcile the two dates, says David Bercovici, a geophysicist at Yale University. The billion-year time lag between the earliest, “proto-subduction” and the full onset of plate tectonics can be explained by the slow, painstaking development of weak zones within the plates, he proposes. The deformation produced by the rocks bending and shearing as the slab sinks into the mantle “leaves behind a scar, a weak spot.” Over time, the zones weakened again and again, until, like a runner’s overstressed ankle, they broke.

To better understand the development of these weak zones and how they lead to plate boundaries, Bercovici and geophysicist Yanick Ricard of the University of Lyon in France investigated what kind of damage happens to lithospheric rocks under intense deformation. Earth’s upper mantle, which makes up much of the lithosphere, is a relatively simple mix, consisting primarily of two rock types, olivine and pyroxene. Smaller crystals of rock are more vulnerable to deformation, and the shear stress placed on the rocks during the bending and twisting of subduction tends to metamorphose the rocks and reduce the crystal sizes, increasing their vulnerability. Although the heat of the upper mantle might help the olivine or pyroxene crystals within the rocks grow larger again, or “heal,” the two rock types are competing for space: Each is actually inhibited in its growth by the other’s presence. And once the crystals are damaged enough and small enough, Bercovici and Ricard reported online yesterday in Nature, that ancient inheritance of weakness becomes a plate boundary.

This process not only explains how Earth’s plate boundaries could lie apparently dormant for a while but still evolve, but it also highlights why plate tectonics wouldn’t occur even on Earth’s so-called twin planet, Venus, which is of similar size and mass. The surface of Venus is more than 400°C hotter than the surface of Earth—and at those extreme temperatures, the rock crystals can grow more quickly, healing themselves so that the boundaries never form.

Lithospheric weak zones with tiny grains have long been suspected to be important to facilitating plate tectonics—but this study explains for the first time how they form, and how “the lithosphere can ‘remember’ these zones of weakness for a geologically long time,” says geophysicist Paul Tackley of Swiss Federal Institute of Technology in Zurich. Regardless of when plate tectonics began, he adds, this grain-size-reduction mechanism could have been very important for causing weak zones in the lithosphere at all times in Earth’s history.

But whether this really reconciles the different proposed dates for the onset of plate tectonics is less clear, Tackley says. The researchers’ model of early Earth is extremely simplified, he adds: Temperatures in Earth’s interior were much hotter billions of years ago and the planet was geologically more “active,” with more volcanism at the surface and more churning in the mantle. So, rather than a uniform layer that grew weak zones over time, the lithosphere “was always heterogeneous, with weaker and stronger parts.”