31 May 2011

Rupture mechanics of the Japan Tohoku-Oki Earthquake, and landslide problems in the aftermath

Posted by Dave Petley

Two articles have been published this week in Science Express, the rapid online version of the journal Science, on the mechanics of the rupture event of the M=9.0 Tohoku-Oki Earthquake in Japan earlier this year.  Both shed some light on the reasons why the earthquake was so immensely damaging.  In the first, Simons et al. (2011) (NB link is a pdf, as the two below) have investigated the magnitude of the slip on the fault.  The results are startling – in places the fault is thought to have displaced by as much as 50 metres or more.  Meanwhile, Sato et al. (2011) measured the  displacement of the seafloor during the earthquake, using five seafloor reference stations installed some years before the earthquake.  Again, the results are surprising in terms of magnitude of displacement, with one of the stations moving 24 metres horizontally and 3 metres vertically.  Finally Ide et al. (2011) examined the rate of slip on the fault.  This paper is accompanied by a very nice press release that explains the implications of the results.  Interestingly, this work suggests a smaller total slip component than Simons et al. (2011) (maximum of about 30 metres, which is still very large indeed), but the  most interesting aspect is the late stage, very high strain rate, slip event that occurred close the sea floor on the eastern margin of the fault.  This is illustrated beautifully by the animation of slip rate (unfortunately it is too large for me to embed!).  It could well be that it was this very rapid slip event that was responsible for the large tsunami that created such significant damage in eastern Japan.

Meanwhile, the remains of Typhoon Songda struck the earthquake-affected areas over the last 24 hours, brining heavy rainfall.  The first heavy rainfall event after an earthquake is a high risk period for any seismic event in a mountainous area.  In Japan this problem is accentuated by the unusual intensity and magnitude of tropical cyclone precipitation.  Early news reports suggest that the impact has been quite large, even though in typhoon terms this is a comparatively small event.  The Japan Times reports at least some flooding, whilst Xinhua (the Chinese state news agency) reports extensive flood and landslide damage:

Heavy rain has flooded the already-devastated Miyagi area of northern Japan, hindering its reconstruction after a massive earthquake and tsunami in March.  Torrential rains brought by typhoon Songda across the country caused landslides and floods, leaving at least 13 people dead and many more missing. Roads have been swept away in at least 200 places and some 19 bridges were damaged. Authorities in the Northeast of Tokyo urged more than 400,000 residents to evacuate their homes Friday following the flooding of a river.  In some areas in the North, 54mm of rain fell in just 12 hours. Up to 1,000 troops have been supporting rescue missions and strengthening flood protection. Songda were downgraded to a tropical storm in the southwest of Japan late on Sunday, but strong winds and rain continue to batter the north.

A Google translation of an article in Japanese at Yomiuri Online suggests that there may have been extensive landslide damage (I have tried to turn this into intelligible English):

In Midorigaoka  landslide damage was confirmed in 189 households out of 190 total units, of which 69 houses were completely destroyed.  Nearly 1 meter of road subsidence occurred. The city issued an evacuation advisory to 104 households.

I am sure that more information will emerge about these impacts over the next few days.  The key point may well be the very high vulnerability of this area to a direct hit from a full strength typhoon later in the summer.  Such an event would be associated with very high levels of hazard.  As an aside, this issue is a key focus of the “When the Shaking Stops” research programme at the Institute of Hazard, Risk and Resilience at Durham.  We post updates on this research on the IHRR blog.