Tsunami Generation

In history, even much smaller earthquakes in the region, like the 1611, 1896, and the 1933 events, caused significant tsunamis along Sanriku coast, with run-up heights of up to 38 meters, causing considerable damage and about 22,000 fatalities (1896). Paleoseismic analysis suggest that about 1100 years ago a comparable tsunami to the one observed on March 11 could have occurred on July 13, 869 in the Sendai area.

Tsunamis seem to be caused by relatively shallow asperities along the trench, leading to significant vertical movement of the shallow part of fault zone, close to the ocean bottom. First models, like the one by G. Hayes ( USGS, see also Figure below, taken from the homepage on March 17), indicate a slip in the shallow part of the fault of about 15 meter.

Provided the average subduction angle of 14 degrees along the fault, this translates into more than 3 meters of sudden vertical uplift. The movement along the shallow patches happened within a very short time span (note: grey isolines indicate time steps of 15 sec), once the asperity failed.

Figure 1: Cross-section of slip distribution. The strike direction of the fault plane is indicated by the black arrow and the hypocenter location is denoted by the red star. The slip amplitude is shown in colour and motion direction of the hanging wall relative to the footwall is indicated by black arrows. Contours show the rupture initiation time in seconds. (Source: Gavin Hayes @ USGS/NEIC)

As a result of the earthquake, the North American plate and the overlying water was pushed upwards along a stretch of approximately 300 km- some 4 meters vertically into the sea at two hotspots and some 2 meters throughout the length of the fault. This can be calculated using DHI tools as QuakeGen and MIKE Zero that translate the movement of theseafloor into the vertical movement of water along the fault. To most accurately assess the seafloor movements the resulting sea surface movement was calculated directly from the results of Hayes (Figure 1). Figure 2 below illustrates the uplifted "mountain" of water that was generated our on the open sea.


Figure 2: The initial water level generated by EQ, one minute after the earthquake

The uplifted water started its horrific journey travelling downhill - practically without feeling any friction from the seabed due to the very long wave length travelling with a speed of up to 800 km per hour reaching the Japanese shoreline about 30 minutes after the quake. At shallower water depths, the travelling speed of the wave decreases - as a result, this speed varies depending on where the wave is headed. Our simulation clearly shows that the tsunami did not spread uniformly in all directions due to the bathymetry.