The federation
Intranet
The Japan Earthquake: Key Scientific Issues
Résultat de Takeo Ito, Kazuhiro Ozawa, Tsuyoshi Watanabe et Takeshi Sagiya de l’université de Nagoya.
The earthquake that struck Japan on 11 March 2011 had a magnitude of 9, which makes it the 4th largest observed earthquake, and the largest in Japan, including historically if one is to believe the estimates. It concerned an area of 500 to 600 km of extension on the Pacific subduction off Japan.
You can find a lot of information on the internet.
Does the earthquake correspond to the expected scenarios?
In 2005, the scientists in charge of estimating seismic hazard in Japan (National Research Institute for Earth Science and Disaster Prevention) estimated the potential magnitude of earthquakes in northeast Japan and the probability of occurrence of these earthquakes within the next 30 years (2005-2035). A 7.5 magnitude earthquake was well planned over a portion of the area that broke on Friday, March 11, 2011 with a 99% probability of occurrence before 2035.
A subduction earthquake of magnitude 8.2 was also considered with a probability of 20%.
This type of probabilistic hazard map is used for the realization of regulatory zoning related to the current building (other than nuclear or chemical). We do not currently have all the information on the scenarios considered to estimate the Tsunami and seismic risks of nuclear facilities in this part of Japan.
The first available information and the review of technical and scientific documents suggest that the magnitude 9 earthquake of 11/03 could be greater than the scenarios considered. This underestimation seems to be mainly due to the fact that the scenarios considered, generally based on the study of past earthquakes, considered that this portion of the plate should break into several pieces, by several earthquakes and not into a single event.
Feedback from the 11/03 earthquake (and other less dramatic examples) shows that earthquakes are capable of breaking several segments of the same fault and that it is therefore very difficult to estimate the maximum possible magnitude on a fault network.
The analysis of past historical earthquakes thus gives only a minimum limit
to potential earthquakes in a given region.
How can we explain the large increase in damage?
Although the size of the earthquake (500km by 100km) was exceptional, the buildings were well resistant to seismic vibrations. Precautions to protect the conventional frame have been very useful. The damage is mainly due to the large-scale tsunami, which was similar in amplitude to the tsunamis triggered by earthquakes slightly further north, such as theMeiji-Sanriku earthquake (1896) with wave heights of more than 30m [temporary rise in sea level due to tsunami passage] due to high tides or the earthquake of Sanriku (1933) with a water rise of more than 25m.
This large extension of the affected areas most likely explains the difficulty of rehabilitating critical power and water networks.
Have seismic design levels of nuclear facilities been exceeded?
The seismic rules of Japanese nuclear installations were tightened in 2006 following work begun in the aftermath of the Kobe earthquake in 1995. The last acceleration level taken into account following this revision for the Fukushima power plant seems to be 600 cm/s2. This calculation was linked to the scenario of a 7.1 magnitude short-distance earthquake, which is different from last Friday’s earthquake. This level was 370 cm/s2 before 2006.
The measurements of Japanese accelerometer networks show median values in the range of 500-600 cm/s2 over the entire northeast coast, indicating that the acceleration levels used for sizing were probably exceeded.
However, three points must be considered to understand the uncertainties and discussions related to the effects of the earthquake on this type of facility:
– Feedback from the Niigata earthquake (July 2007) showed that Japanese power plants were able to withstand higher levels of acceleration than those used for their design. The resistance of an installation to an earthquake is therefore linked to the seismic design level (characteristics of the vibrations considered during construction) and the existing margins (difficult to assess) beyond this design level. [It should be recalled that design allows three levels of protection, the first with an immediate return of the installation to operation, the second with a return after verification and repair and the third ensuring the containment of radioactivity but not allowing the return to operation in a permanent way].
– The problems currently encountered (level 6 incident, the most important level to date for the Japanese nuclear industry) are also related to the Tsunami (destruction of cooling facilities). It is therefore important to highlight the difficulties associated with cumulative risks.
– The maximum acceleration generated by vibrations (a parameter traditionally used to describe the amplitude of vibrations) is not the most relevant parameter for assessing the harmfulness of vibrations. Indeed, such an earthquake is characterized mainly by a very long duration of vibrations more than by very strong accelerations.
Why exceptional damage in a country renowned for its earthquake preparedness?
The images we can see->http://www.slate.fr/grand-format/le-japon-devaste-35509] show terrible effects of the tsunami surge beyond the protection dikes and the problems that followed for the nuclear power plants on the coast.
However, the images broadcast do not show these examples of large collapsed buildings that have claimed many victims in Mexico City or Haiti.
This needs to be explained. Despite the very high magnitude of the earthquake,
the vibrations produced did not cause massive destruction. Several elements can explain this:
– the quality of construction, particularly for large buildings where binding rules are applied. This is only an element of explanation because in Japan as elsewhere there are a large number of old buildings that were not built according to the rules in force today.
– the position of the sliding zone that generates seismic waves: at depths of about 30 to 50 km and at distances exceeding 100 km from major urban areas. Distance and depth are important elements because the high frequency waves, to which constructions are sensitive, attenuate quickly and are less strongly excited for deep sources. We would therefore be in a much more favourable situation than in the case of Haiti (surface source at close range) but the magnitude is much higher in Japan: 9 instead of 7.2 in Haiti (which makes an energy ratio of about 500 anyway).
The important point is therefore to understand the meaning of these magnitude values. To put it simply, high frequency waves (0.1-10Hz, those that destroy buildings) are emitted on the fault in proportion to the stress that is released. However, this stress, which depends on the maximum resistance that friction exerts to slip, is a characteristic of rocks and can be considered a first-order constant. It is not the released constraint that increases with magnitude, it is essentially the size of the earthquake. A very small emission zone for a low magnitude (1 km² for magnitude 4) and a huge one for the Japanese earthquake. But the waves attenuate with distance and those emitted by the different regions of the fault are not composed in a coherent way. As a result, the maximum accelerations of the ground do not increase as much as the power of the moment. Fortunately.
It is important to note that observations of high-frequency ground movements, which have a high intrinsic variability, do not increase significantly between a magnitude of 8 and 9, but it is the area concerned that is expanding, with obviously a greater accumulation of damage.
In the case of the East Coast of Japan, Japanese seismologists had considered that the failure scenarios involved several segments that would break independently. This idea was based on the complex geometry of the subduction plane, geodetic indications that show the heterogeneous nature of the coupling between the plates at the scale of the region and especially historical observations that indicate repeated earthquakes of magnitude between 7 and 8. This misinterpretation (since it is indeed several of these segments that broke simultaneously) and magnitude value error probably does not have a dramatic consequence for the level of high-frequency waves. This is suggested by the generally good condition of the buildings. On the other hand this reference to several segments has a dramatic consequence on the sizing of tsunami protection.
A word on the genesis of tsunamis: tsunamis are generated by the vertical displacement that affects the seabed. Indeed, a subduction earthquake produces a permanent deformation with a strong vertical component. It is this topographical change, and the rebalancing of water masses, that produces the tsunami.
The images showing the sea breaking over the dikes clearly indicate that they were undersized compared to the earthquake of 11 March. So why is the question of magnitude so important this time?
If we accept that the stress release is more or less constant, the slip over the fault increases linearly with the lateral extension of the earthquake. It is a law of scale. Intuitively this expresses the idea that it is easier to force a large slip over a large crack than over a small one. With a magnitude of 9, the earthquake that actually occurred corresponds to total landslides much larger than those expected for an 8 magnitude earthquake, and a simple first order calculation suggests a sliding ratio of 3. The mass of water set in motion by vertical movements per unit of subduction length is therefore also in a ratio of 3 for earthquakes of magnitude 8 or 9. The height of the water blade on the coast will be proportional to the vertical movement.
The images show the overflow of protective dikes by a rise in water levels that reached around 10 m. If the magnitude scenario less than or equal to 8 had been the correct one, the sizing would have been sufficient for a 3 m rise.
This dramatic example shows how the basic assumption of the scenarios is decisive. Unfortunately, it is likely that the sizing required to protect against this tsunami would have been considered extravagant for arguments that seemed reasonable. But our historical observations only give us minimal magnitudes to take into account, mega-seisms like this one, probably very rare, very "improbable", remind us painfully.
Scientific links useful for understanding
Other analyses
– Richard Monastersky in the Scientific Amercian
– on the landslide blog (agu)
Earthquake information
– http://www.jma.go.jp/jma/en/News/20...
– http://earthquake.usgs.gov/earthqua...
– http://www.emsc-csem.org/
– https://sites.google.com/site/cellu...
– http://www.insu.cnrs.fr/a3738,seism...
Earthquakes and nuclear safety
– http://www.irsn.fr/
– http://www.iter-consult.it/Meetings...
Seismic hazard maps of Japan (before the earthquake) for conventional buildings