NEWS & VIEWS SEISMOLOGY
Breaking the slab Two chains of seamounts on the Pacific plate subduct beneath central Japan. In the process, a fragment of the Pacific slab has become wedged in the subduction zone and may be the source of recurring deep-thrust earthquakes beneath Tokyo.
Meghan S. Miller Center for Computational Geophysics, Department of Earth Science, Rice University, Houston, Texas 77005, USA. e-mail:
[email protected]
arthquakes and volcanoes in subduction zone settings are traditionally described in a relatively simple model that involves one plate descending into the mantle beneath another. Many tectonic settings, though, are far more complex. Beneath central Japan, for example, the Philippine Sea plate and Pacific plate are both subducting beneath the Eurasian plate on which the metropolis of Tokyo is located1 (Fig. 1). Subduction in this region causes deep-thrust earthquakes, such as the destructive ones within the Kanto region in 1703, 1855 and 1923 (ref. 2). On page 771 of this issue, Toda and co-authors3 identify a discrete fragment of the Pacific plate beneath Tokyo using geophysical data: this fragment is jammed between the Pacific and Philippine Sea plates below the overriding Eurasian plate. Initial models of plate interaction and subduction processes overlooked the effects of topographic features on the subducting oceanic plate4. However, seamounts — volcanoes on the seafloor — have since been shown to affect the structure of the subducting slab as well as earthquake nucleation and propagation. They have also been linked to megathrust events5–8. Recent research on a portion of the Japan arc, north of Kanto, has suggested that subducted seamounts on the Pacific plate weaken interplate coupling and lead to the nucleation of repeating large, but not great (of moment magnitude Mw > 8), earthquakes8. Others have suggested that seamounts on the Philippine Sea plate subducted along the Nankai Trough are sites of earthquake nucleation — however, owing to entrainment of fluid-rich materials associated with the seamounts, these areas may actually inhibit rupture propagation6. Toda and co-authors3 examine seismic data, specifically the locations
E
730
Eurasian plate
Pacific plate
Tokyo
Philippine Sea plate
Figure 1. Schematic model of subduction in the Japan region. The figure shows the simplified configuration of the Philippine Sea plate, Pacific plate and Eurasian plate as viewed from the southeast. The seamounts on the Pacific plate and the trenches are clearly seen in the bathymetry map draped on the model. The black dots illustrate earthquake hypocentres beneath Tokyo and the Kanto region, and the red arrows show the direction of plate motion relative to the Eurasian plate.
of earthquake hypocentres, seismic tomography, and stress inversions, along with magnetic anomaly and bathymetry maps to understand the complex structure beneath Kanto in central Japan. Using three-dimensional visualization software, they identify an anomalous region beneath Kanto and interpret it as a broken fragment of the Pacific plate. In bathymetric and magnetic anomaly maps, two chains of seamounts on the Pacific plate off the coast of central Japan coincide approximately with the bounds of the slab fragment at depth. Toda and colleagues propose that these topographic features effectively inhibit subduction owing to their higher buoyancy compared with the
surrounding plate. They suggest that stresses induced by subduction of the buoyant seamounts on the westward subducting Pacific plate led to faulting that broke the slab and caused a fragment to become detached. This fragment eventually became jammed between the shallowly subducting Philippine Sea plate and the steeply subducting Pacific plate. The results of this study thus go beyond the links between subducted seamounts and seismogenesis, and suggest that the effects of seamount subduction may be more profound. The position of this slab fragment coincides with the location of recurring deep-thrust earthquakes beneath Tokyo. Key features of the earthquake
nature geoscience | VOL 1 | NOVEMBER 2008 | www.nature.com/naturegeoscience © 2008 Macmillan Publishers Limited. All rights reserved.
NEWS & VIEWS occurrence under Kanto can be explained by the frictional contact between the slab fragment with the Pacific plate below it, the Eurasian plate at its eastern end, and the Philippine Sea plate at its southern extent. However, although Toda and co-authors3 agree with previous studies9–11 regarding the estimated location and geometry of the Pacific plate at depth, their interpretation of the geometry and position of the Philippine Sea plate differs. Previously, the seismic anomaly has been interpreted as the bottom end of the northwestward subducting Philippine Sea plate2. Toda and colleagues3, contend that the Philippine Sea plate does not, in fact, extend as far north as previously interpreted2, and suggest that the anomaly originally associated with the Philippine Sea plate is actually a fragment of the Pacific plate. The identification of this fragment as a piece of the Pacific plate is based on its similarity with characteristics of the Pacific slab beneath Japan, as documented elsewhere. These include a double zone of seismicity and a thick zone of high–velocity perturbations of 5%, which is unlike the thinner, slower Philippine Sea slab. Plate reconstructions for this region also suggest that the
we obtain more seismic and geophysical data and as computations become faster, we can add more complexity to models of subduction-zone processes. The unique image of a slab fragment beneath Kanto was produced from dense, high-quality seismic data and extensive three-dimensional interpretation of a series of different geophysical datasets. As Toda and co-authors suggest, similar slab fragments are likely to exist in other regions. As equally extensive geophysical datasets become available, three-dimensional visualization similar to that used by these researchers may help locate such fragments.
Philippine Sea plate could not extend as far north as previously thought. The explanation of the seismicity and its link to recurring earthquakes beneath Kanto is dependent on the interpretation of the slab fragment. The authors attribute the cluster of seismicity, and the occurrence of the large devastating earthquakes that occur under the Kanto region, to frictional contact between the slab fragment and the Pacific plate. The bounds of the seismicity are consistent with the position of the piece that has broken off the Pacific slab. This mechanism, however, differs from previous interpretations that invoked the interaction of the Philippine Sea slab with the Pacific slab8–11. The interpretation of tectonic processes and seismic hazard analysis in this region is inherently complex due to several interacting structures at depth and at the surface. These include the Philippine Sea plate, the Pacific plate, possible slab fragments, subducted seamounts, as well as large subaerial features such as the Izu-Bonin island arc and Sagami Trough that are colliding with the Kanto region of Japan (Fig. 1). The ongoing debate regarding the link between seamounts, tectonic evolution, and seismogenesis will inevitably continue. As
References 1. Seno, T., Stein, S. & Gripp, A. E. J. Geophys. Res. 102, 53–84 (1993). 2. Utsu, T. http://iisee.kenken.go.jp/utsu/index.html (2006). 3. Toda, S., Stein, R., Kirby, S. & Bozkurt, S. Nature Geosci. 1, 771–776 (2008). 4. Isacks, B., Oliver, J. & Sykes, L. J. Geophys. Res. 73, 5855–5899 (1968). 5. Cloos, M. Geology 20, 601–604 (1992). 6. Bangs, N., Gulick, S. & Shipley, T. Geology 34, 701–704 (2006). 7. Bilek, S. L. & Engdahl, E. R. Geophys. Res. Lett. 34, L20311 (2007). 8. Mochizuki, K., Yamada, T., Shinohara, M., Yamanaka, Y. & Kanazawa, T. Science 321, 1194–1197 (2008). 9. Wu, F., Okaya, D., Sato, H. & Hirata, N. Geophys. Res. Lett. 34, L18301 (2007). 10. Ishida, M. J. Geophys. Res. 97, 489–513 (1992). 11. Sato, H. et al. Science 309, 462–464 (2005).
NATURAL DISASTER On 26 December 2004, a massive tsunami inundated the coastline of southeast Asia with no warning. The wave left behind death and destruction on a scale virtually unprecedented in living memory. Devastation crossed political borders and was worsened by a lack of preparedness: in many of the hardest-hit regions there were no historical records of a tsunami of similar magnitude. Government attention soon turned towards better early-warning systems for future tsunamis. These efforts, however, were hampered by a lack of knowledge about the history of tsunami activity in the region. Now, two teams have reconstructed the millennial-scale history of tsunamis in Thailand (Nature 455, 1228–1231; 2008) and the Indonesian island of Sumatra (Nature 455, 1232–1234; 2008). As the flood waters receded, they left behind a thick layer of sand on many beach plains and ridges. A team led by Kruawun Jankaew of
USGS
Flood of evidence
Chulalongkorn University, Thailand, dug and drilled into the marshy sediments of the Thai island of Phra Thong, hoping to identify similar sand layers. Their efforts were rewarded with two such deposits. Using radiocarbon dating of the shell and organic material that sandwiched the layers, they conclude that the older tsunami occurred approximately 2,800 years ago, and that the younger layer was deposited sometime around or immediately after ad 1300–1450. Meanwhile, a team led by Katrin Monecke of Kent State University,
Ohio, was searching the depressions between beach ridges in Aceh, Sumatra, where they uncovered a thick sand layer. Their radiocarbon measurements placed the timing of deposition between ad 1290 and 1400, roughly corresponding to the Thai deposits. The team’s transects also revealed another large tsunami event from the period ad 780–990. The two groups’ findings suggest that tsunamis such as the 2004 event have a recurrence interval of approximately 600 years. However, the studies were not designed to detect tsunamis of smaller magnitudes, such as the one that followed the 28 March 2005 earthquake, which still caused extensive damage to buildings and infrastructure. Both teams stress that despite the apparently low risk of a large tsunami event in the near future, there is still a need to maintain adequate warning systems and evacuation plans to protect coastal residents from the unpredictable seas. Alicia Newton
nature geoscience | VOL 1 | NOVEMBER 2008 | www.nature.com/naturegeoscience © 2008 Macmillan Publishers Limited. All rights reserved.
731
