> 
Paris, September 23, 2004
In highly active subduction zones, where earthquakes are numerous and frequent, understanding the history of past earthquakes makes it possible to take a "probabilistic" approach to identifying segments that may be expected to rupture in the near future. Thus, for example, an earthquake of a magnitude of the order of 8 is expected in the Tokai segment of the Nankai fault, where the Philippines Sea plate dips under central Japan, since apparently there was no rupture in the last century as elsewhere in the region. In less seismically active subduction zones, such as the Cascadia region, the Hellenic Arc, and the Antilles Arc, where records of past events are less reliable, one cannot exclude that such large earthquakes may also be in preparation.
The international research community has taken a growing interest in the high-risk zones over the last 20 years. Now new developments based on three distinct approaches are turning up unexpected findings. According to Alfred Hirn and Mireille Laigle, these are probably precursor signs indicating major earthquakes, which deserve further study and observation.
Very specific, weak seismic signals, known as "tremor," were discovered in the Nankai zone thanks to the development of a network of seismometers in Japan installed since the Kobe earthquake in 1995, and in the Cascadia region. Tremor is attributed to the contribution of fluids at the seismic source and seems to indicate the presence of water near the interplate boundary at a depth of several dozen kilometers. This water appears to come from the dehydration of rocks in the subducting plate. Depending on the depth at which the water is released, and depending on the structure and thermal properties of the plates, the water is either absorbed by the rocks in the mantle, or trapped under the crust of the overlying plate, thus contributing to the source of tremor.
In addition, ruptures too slow to create seismic waves, known as "silent earthquakes," were discovered in the subduction of the Cascadia region by a dense network of Global Positioning System (GPS) measures that continually record changes in the Earth's surface. Researchers then discovered that these silent earthquakes were accompanied by tremor. Similarly, a silent earthquake was observed in Japan, in precisely a part of the zone where a major earthquake is expected: in the Tokai segment of the Nankai subduction zone, along which tremor was discovered.
Lastly, a third approach was adopted to examine the details of the deep structure of the subduction zones, in particular the seismogenic part of this overlapping mega-thrust fault touching the two plates, and to observe the hydration areas. This approach is called Seismic Exploration Imaging (reflection-refraction-tomography) and is performed in Japan, or (by the authors of the article) off Greece and in the Antilles (article forthcoming in Tectonophysics).
For Alfred Hirn and Mireille Laigle, in the case of Japan and in light of the deep structure revealed by imaging, associating tremor and slow earthquakes can be interpreted as a signal announcing an expected major earthquake. They suggest the following model: the pressure of the fluid changes the plate coupling of a part of the seismogenic zone, which seems to influence pre-seismic behavior.
In the case of the Cascadia region, the observations of tremor and silent earthquakes are also consistent with the processes of liberating fluids from the plate dipping under the crust. However, their occurrence appears to be periodical, and various data, including seismic imaging, indicate that their source zone is located downdip from the seismogenic part, thus with a different relationship to an eventual mega-earthquake.
While the discovery of these new signals of the activity in certain subduction zones does not mean that forecasting is possible already, understanding these new phenomena can be furthered by deep structure imaging and setting up dense networks, on land and at sea, of instruments to allow in-depth observation of these signals over long periods of time. With such an approach, the possibility of developing forecasts cannot be ruled out.
In the other subduction zones where there is no reliable record of the past occurrence of megathrust earthquakes, it is not known yet if such signals of activity might be detected, but structural exploration using seismic imaging can help locate a seismogenic segment. Work has begun in this area for the Hellenic and Antilles subductions.
The ability to draw up forecasts will require an effort in terms of observation, exploration and the development of modeling to understand the events leading up to a major earthquake in such geodynamic settings.
Researcher contacts:
Department of Seismology (IPGP-CNRS)
Alfred Hirn: +33 1 44 27 39 14 hirn@ipgp.jussieu.fr
Mireille Laigle: +33 1 44 27 47 78 laigle@ipgp.jussieu.fr
INSU contact:
Christiane Grappin: +33 1 44 27 43 37
christiane.grappin@cnrs-dir.fr
Press contact:
Martine Hasler: +33 1 44 96 46 35
martine.hasler@cnrs-dir.fr
Latest press releases
All disciplines