News & Views | Published:

Plate tectonics: Lost sea floor

Nature Geoscience volume 5, pages 165167 (2012) | Download Citation

The configurations of ancient tectonic plates are difficult to reconstruct. Seismic images of deep subducted plates, combined with data from ancient volcanic arcs, help to derive a tectonic map of the Pacific Ocean as it was 200 million years ago.

The system of plate tectonics on Earth has probably existed for billions of years. As the tectonic plates move, new sea floor is formed at mid-ocean ridges and the old oceanic plates are recycled back into the mantle through subduction. Because subduction eventually destroys most of the sea floor, our knowledge of plate configurations declines back in time; the distribution of oceanic plates before the Cretaceous period, about 140 million years ago, is largely unknown. Writing in Nature Geoscience, van der Meer et al.1 report a reconstruction of the global plate tectonic environment during the Triassic to Jurassic periods, and demonstrate the existence of ancient subduction zones within the palaeo-Pacific ocean.

The theory of plate tectonics gained wide acceptance in the 1960s with the discovery that sea floor is created at mid-ocean ridges and then spreads away2. The subsequent identification of the main tectonic plates further grounded the theory3. Recently, with a detailed understanding of the plate tectonic boundaries4 and the age variation of the sea floor5, reconstructions of past plate positions became possible. This is important because configurations of tectonic plates during the geological past bear information on how the interior of Earth evolved with time.

The simplest case is to reconstruct the migration of continents on both sides of the Atlantic Ocean. In fact, a glance at the outline of the South American and African continents clearly shows that the two must have once been joined. When the Atlantic Ocean started to form with new oceanic crust created at the Mid-Atlantic Ridge, the continents were separated and progressively forced farther apart. The evolution of the oceanic plates and the migration of these continents can be reconstructed backwards in time from a start point at the present-day Mid-Atlantic Ridge, where the oceanic lithosphere is youngest. In such a reconstruction, the youngest oceanic lithosphere from either side of the ridge is sequentially removed, providing a crude method for simulating a reversal of the force that originally drove the motion of continents. The Atlantic Ocean decreases in size and the continents are brought back together, as they were during the Triassic period.

Plate reconstructions for the Atlantic Ocean are relatively simple because the surrounding continents are joined directly to the Atlantic Ocean floor, with no interjecting subduction zones. Reconstructions in the Pacific Ocean, however, are more complex. The Pacific Ocean floor is separated from the surrounding continental plates by subduction zones, and large amounts of old oceanic lithosphere have been subducted into the mantle (Fig. 1). To reconstruct past positions of the plates that make up the Pacific Ocean floor with the same method described above, currently subducted lithosphere needs to be — virtually — drawn back out of the subduction zone and returned to the surface. With no direct measurements of their age and geometry, these virtual sea floors form a tectonic gap in the reconstruction, and plate configurations in this gap are often conjectured (Fig. 1). To produce a tectonic map of the Pacific Ocean, especially before about 140 million years ago, therefore requires constraints that are independent from seafloor ages.

Figure 1: Schematic reconstructions of Pacific-style oceanic plates.
Figure 1

a, In the present-day Pacific, the oceanic plates are not directly attached to the surrounding land masses. To reconstruct older plate configurations, the youngest oceanic crust from two plates (A and A') that share a common mid-ocean ridge (orange lines) can be sequentially chopped off in a process that simulates forcing of the youngest lithosphere back into the mantle. This process draws the oldest, subducted lithosphere out of the subduction zones and back to the surface. b, For plate reconstructions beyond Cretaceous times, more than 140 million years ago, all of the lithosphere whose age is currently known is forced back into the mantle and the reconstructions consist of an ocean whose entire floor has no age constraints, creating a tectonic gap. Within the tectonic gap, reconstructions based on seafloor ages will fail. Inference of the palaeo-subduction zone between plates A and B' therefore requires independent constraints, such as fossil volcanic arcs and mantle seismic images. Van der Meer and colleagues1 use these methods of reconstruction to demonstrate that north–south-oriented subduction zones may have existed within the palaeo-Pacific Ocean about 200 million years ago. Red ovals represent volcanic arcs. The grey line in b indicates the boundary of sea floors with known age.

Independent constraints that help fill the tectonic gap may come from geological records6, among which are fossilized volcanic arcs. These are volcanic chains that formed above ancient subduction zones. Rocks within the arc contain information, such as the age and palaeo-latitude of the subduction zone that they formed above. Most of the arcs that formed on oceanic lithosphere would eventually be consumed at subduction zones. However, some arc rocks, due to their buoyancy, can resist subduction and instead accrete to the overriding continent. Thus, information on the original subduction zone is preserved within the accreted arc rocks. Seismic images of the deep mantle, on the other hand, can also be used to identify remnants of ancient subducted slabs7. Broken-off sections of subducted oceanic lithosphere can linger in the mantle for millions of years and thus provide a marker for the position of ancient subduction zones.

The study by van der Meer and colleagues1 makes use of such independent constraints to propose a model for the Pacific Ocean as it would have looked 200 million years ago, then called the Panthalassa Ocean. All of its original sea floor has been consumed by subduction, so that the tectonic gap covers the entire ocean floor. Without any sea floor whose age and position are known, reconstruction of the Panthalassa Ocean is based on far-travelled fossil volcanic arcs and seismic images of the mid-Pacific lower mantle. Data from these fossil arcs indicate that they formed above an ancient subduction zone that would have been located within the Panthalassa Ocean about 200 million years ago. In addition, seismic images of the mid-Pacific lower mantle8,9 identify potential remnants of the Panthalassa ocean floor that now linger in the mantle. This oceanic lithosphere could have entered the mantle at the subduction zone where the arcs were formed. Thus, the two data sets are complementary. What's more, both point towards a north–south-oriented series of intra-oceanic subduction zones that must have once divided the Panthalassa Ocean into two (Fig. 1).

The improved reconstruction of oceanic plate geometries helps to identify the position and evolution of ancient supercontinents, such as Gondwana and Pangaea. The reconstructions also shed light on changes in mantle flow that have occurred in the past. Subduction zones coincide with regions of downwelling in the mantle. Therefore, the identification of ancient north–south-oriented subduction zones implies that large-scale downwelling occurred in this part of the mantle, beneath the Panthalassa, 200 million years ago. Existence of several major subduction zones at this time requires a re-evaluation of the hypothesis that during supercontinent times Earth may have had only one zone of mantle downwelling and one of mantle upwelling — a configuration known as degree-1 convection10. Convection within the mantle acts to redistribute Earth's mass. Transitions between different degrees of convection — from one to multiple zones of upwelling and downwelling — could cause Earth to shift on its spin axis. However, the proposed intra-oceanic subduction system implies a pattern of convection during Triassic to Jurassic times that is not too different from the present-day situation. A long-term, stable pattern of mantle convection could provide a mechanism for generating the long-term stability of Earth's rotational axis11.

The tomographic reconstruction by van der Meer et al.1, which indicates that the palaeo-Pacific Ocean was once divided into two sub-basins, is strongly dependent on the seismic images of the deep mantle below the Pacific Ocean8,9. The resolution here is not very high and efforts to refine these images are needed. If the proposed intra-oceanic subduction zones are confirmed, work can focus on identifying the exact timing and mechanisms driving the subduction.


  1. 1.

    , , , & L ., Nature Geosci. 5, 215–219 (2012).

  2. 2.

    Science 154, 1405–1415 (1966).

  3. 3.

    J. Geophys. Res. 73, 1959–1982 (1968).

  4. 4.

    Geochem. Geophys. Geosys. 4, 1027 (2003).

  5. 5.

    , , & Geochem. Geophys. Geosys. 9, Q04006 (2008).

  6. 6.

    Geol. Soc. Am. Bull. 115, 655–668 (2003).

  7. 7.

    , & Nature 386, 578–584 (1997).

  8. 8.

    Global Travel Time tomography with 3-D Reference Models PhD thesis, Univ. Utrecht (2007).

  9. 9.

    , , & Geophys. J. Int. 184, 1223–1236 (2011).

  10. 10.

    & Geophys. Res. Lett. 32, L01301 (2005).

  11. 11.

    & Nature 452, 620–623 (2008).

Download references

Author information


  1. Lijun Liu is at the Institute of Geophysics and Planetary Physics, Scripps Institution of Oceanography, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093-0225, USA

    • Lijun Liu


  1. Search for Lijun Liu in:

Corresponding author

Correspondence to Lijun Liu.

About this article

Publication history



Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.