A Chilean surprise

Article metrics

Initiation of the great 2010 Chile earthquake occurred within the rupture zone of the 1835 event experienced by Charles Darwin. However, the peak fault slip was to the north of the epicentre — not where it was expected to occur.

On 20 February 1835, Charles Darwin was resting in the woods near Valdivia, Chile, when the ground suddenly trembled for about two minutes, although “the time appeared much longer”1. Two weeks later, at Concepción, Darwin saw the devastation wrought by the ground shaking and tsunami, and observed coastal uplift that he inferred to have “accompanied or caused the earthquake”1. This great earthquake, with an estimated2 seismic magnitude (Mw) of about 8.5, is now understood to have occurred on a megathrust fault, where the Nazca plate underthrusts the South American plate with a plate-convergence rate of about 70 millimetres per year.

After 175 years, the region ruptured again, in the great Mw-8.8 earthquake of 27 February 2010, again causing extensive shaking and tsunami damage. Studies by Moreno et al.3 and Lorito et al.4 have respectively quantified the fault-zone processes before and after the 2010 earthquake, revealing surprising contrasts to the event witnessed by Darwin.

Repeated stick–slip sliding of plate-boundary faults occurs as a result of the frictional build-up and release of stress driven by relative plate motions. In 1939, Concepción was again struck by a devastating earthquake, but the faulting occurred within the sinking Nazca plate, not on the plate boundary2. The plate boundary north of the 1835 rupture experienced large underthrusting earthquakes in 1906, 1928 and 1985, whereas to the south the immense Mw-9.5 earthquake of 1960 accommodated plate convergence in that region (Fig. 1a). Between 1835 and 2010, some 12 metres of relative plate convergence had occurred near the Darwin earthquake zone, presumably with frictional resistance causing extensive build-up of elastic strain in the rock around the fault zone that was released in the great 2010 Chile earthquake. Such a build-up of strain can now be detected by modern geodetic methods.

Figure 1: Earthquake history and the rupture zone of the great 27 February 2010 Chile earthquake.

a, Source regions of large earthquakes over the past century along the central Chile subduction zone, shown as outlined areas with corresponding event dates. This subduction zone is where the Nazca plate (to the left) thrusts under the South American plate (to the right). The white arrow indicates the plate-convergence direction. The approximate rupture lengths for events before 1900 are indicated by white dashed lines, with the 1835 event being the Darwin earthquake. b, Inferences of fault-zone processes for before3 and after4 the 2010 event. The blue curves indicate regions of >75% locking of the megathrust, as estimated3 from a decade of GPS observations prior to 2010. The red curves depict post-event analysis of the earthquake's behaviour4. Solid red, the 2010 earthquake aftershock zone; dotted red, the region with unexpected slip greater than 15 m north of the 2010 epicentre (red star); dashed red, a region with slip of 5–10 m south of the epicentre.

The work of Moreno et al.3 consisted of analysis of a decade of ground-deformation data prior to 2010, recorded by a dense network of Global Positioning System (GPS) sensors along the length of Chile. These data indicated that the western margin of the South American plate north of the 1960 rupture zone had generally eastward ground velocities relative to the stable interior of the plate — a manifestation of strong frictional coupling between the converging Nazca and South American plates, and of accumulating strain in the overriding plate.

The authors' finite-element modelling3 of the geodetic velocities as the result of interseismic 'shortening' of the Andean subduction zone yielded estimates of the spatially varying interplate coupling of the plate-boundary megathrust fault (Fig. 1b). They found that the southern half (35.5° S–37.5° S) of the 1835 rupture zone was 'locked' (not slipping). The northern region of the rupture zone seemed to be 'partially locked' between 34.8° S and 35.5° S, meaning that some interplate sliding was occurring by non-seismic slow slip, although not keeping up with the total plate convergence rate, effectively lessening the accumulated strain and slip deficit by up to 50%. The megathrust was not slipping farther north, from 33.8° S to 34.8° S, along the 1985 rupture zone and the northern portion of the 1928 rupture zone.

With the megathrust regions both north and south of the 1835 rupture having experienced relatively recent large ruptures, the accumulated stress on those areas would presumably be lower than in the strongly locked portion of the 1835 zone. On the basis of plate-tectonic convergence rate, the history of large underthrusting earthquakes, and geodetically inferred interplate locking for the previous decade, one might reasonably expect the slip distribution of the 2010 event to have been confined to the region between the 1960 and 1928 ruptures, with maximum slip in the southern portion. However, determinations of the actual slip distribution, including that by Lorito et al.4, are largely incompatible with these expectations.

The great 2010 Chile earthquake did initiate in the central deeper portion of the 1835 rupture zone. But the rupture expanded bilaterally over a total length of about 500 km, with the aftershock zone overlapping the 1928 and 1985 rupture zones and the northernmost part of the 1960 rupture zone (Fig. 1b). The event was thus much larger than the 1835 event, possibly resembling an earlier earthquake in 1751. Persistent segmentation of megathrust ruptures along the coast is clearly not pronounced in this subduction zone.

Comparison of initial rupture models for the megathrust slip distribution, made on the basis of remotely recorded seismic waves, showed a strong correlation with the geodetically inferred regions of strong locking3,5. But refined models — which involved matching data from GPS displacements, tsunami recordings and radar-remote-sensing observations, and measurements of coastal uplift4,6,7— reduce the correlation by repositioning the main slip patches by 30–50 km.

These models place the largest slip, of up to 20 m, north of the epicentre at 34.5° S–35.5° S, overlapping the region of previously partial (50–75%) locking in the 1928 rupture zone, where the pre-stress was presumably relatively low. Co-seismic slip of 5–10 m did occur in the southern region of the 1835 rupture zone that previously had strong locking, but the strongly locked region of the megathrust west of the epicentre has only a few metres of slip in all of the refined models4,6,7. Lorito et al.4 thus infer that there have been large increases in shear stress on the locked megathrust in this region, and so there is continuing potential for a near-future large earthquake in the 1835 rupture zone. However, our lack of understanding of why the region did not rupture during the 2010 earthquake casts doubt on our ability to project its future behaviour.

The surprising size and slip distribution of the 2010 Chile earthquake has now come into focus, highlighting the shortcomings in our ability to anticipate the rupture properties of great earthquakes, even along plate boundaries where we can quantify interseismic locking. Like the 26 December 2004 (Mw-9.2) Sumatra earthquake8, the 2010 rupture extended far beyond conventional expectations. The northward expansion of rupture beyond the 1835 rupture zone may reflect the intrinsic tendency of ruptures to continue as long as no acute barrier or strain deficit causes them to arrest.

The patchiness of the co-seismic slip in the 2010 rupture emphasizes the importance of quantifying large-rupture slip distributions, because they will influence the pre-stress conditions for future earthquakes. It seems that the region of Darwin's earthquake still has the potential to make “the earth tremble”1, but when it will next do so remains uncertain.


  1. 1

    Darwin, C. Journal of Researches into the Natural History and Geology of the Countries Visited During the Voyage of H.M.S. Beagle Round the World 2nd edn (Murray, 1845).

  2. 2

    Beck, S., Barrientos, S., Kausel, E. & Reyes, M. J. South Am. Earth Sci. 11, 115–129 (1998).

  3. 3

    Moreno, M., Rosenau, M. & Oncken, O. Nature 467, 198–202 (2010).

  4. 4

    Lorito, S. et al. Nature Geosci. 4, 173–177 (2011).

  5. 5

    Lay, T. et al. Geophys. Res. Lett. 37, L13301, doi:10.1029/2010GL043379 (2010).

  6. 6

    Tong, X. et al. Geophys. Res. Lett. 37, L24311, doi:10.1029/2010GL045805 (2010).

  7. 7

    Delouis, B., Nocquet, J.-M. & Vallée, M. Geophys. Res. Lett. 37, L17305, doi:10.1029/2010GL043899 (2010).

  8. 8

    Lay, T. et al. Science 308, 1127–1133 (2005).

Download references

Author information

Correspondence to Thorne Lay.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lay, T. A Chilean surprise. Nature 471, 174–175 (2011) doi:10.1038/471174a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.