Tappin, D.R., Grilli, S.T., Harris, J.C., Geller, R.J., Masterlark, T., Kirby, J.T., Shi, F., Ma, G., Thingbaijam, K.K.S., Mai, P.M., (2014). Did a submarine landslide contribute to the 2011 Tohoku tsunami? Marine Geology 357, 344-361. doi:10.1016/j.margeo.2014.09.043
On March 11, 2011 a magnitude 9.0 earthquake with an epicenter located off of the east coast of Japan generated a tsunami with localized run up heights of 40 meters. This was one of the most powerful earthquakes on record, and the most powerful earthquake to impact Japan. Besides an impressive list of devastating statistics, more than 18,000 people lost their lives or are missing. Additionally, the environmental impacts due to the destruction of the Fukushima Nuclear power plant, such as the discharge of radioactive water to the Pacific Ocean, are still being realized today.
Offshore of Japan exits a subduction zone style tectonic plate boundary, where a dense oceanic plate collides and sinks below a continental plate. The subducting plate can become locked to the overriding plate leading to an enormous amount of stress building up. Over time, the locked portion of the overriding plate deforms downward in the direction of the locked subducting plate. The rapid release of stress along the locked portion causes an earthquake. The upward rebound of the overriding plate displaces the overlying water column, generating a tsunami (See “Resource 1” below for an animation of this process).
The Tohoku tsunami has been very well studied due to an abundance of data recorded by Japan’s abundant seismic and geophysical networks. As soon as seismic and tsunami waveform data became available, researchers began studying what exactly triggered the devastating tsunami. These studies employed techniques in which data recorded by buoys, wave run up and inundation heights and other seismic information, that resulted from the tsunami are used to trace the tsunami back to its source, much like a ballistics expert would take fragments of bullets from a crime scene to recreate the shooter’s position. Study after study presented similar discrepancies. First, computer modeling studies which used seismic and buoy wave data to recreate a tsunami were unable to predict the extraordinarily high (>40 m) tsunami run up heights. Additionally, the timing and the short wavelength of the tsunami waveforms recorded by nearshore buoys could not be explained by a single tsunami source.
The authors of this study argue that the vertical displacement of the seafloor along a fault due to the 2011 earthquake alone cannot explain the aforementioned discrepancies, and propose an additional tsunami source: a submarine landslide. Similar to a typical landslide occurring above sea level, a submarine landslide is the rapid movement of large volumes of rock and sediment down a sloped surface underwater.
Methods and Results
To test the hypothesis that a secondary tsunami source explains the discrepancies in tsunami run up height and high frequency waveform observations at nearshore buoys, the authors first outlined the evidence supporting an additional tsunami source. The authors compiled published reconstructions of the source locations of the tsunamis by a technique called tsunami inversions. This technique uses wave data to trace tsunamis back to their source location. The average source location is represented by the red star, shown in Figure 1. The authors also show the earthquake centroid location as determined by the Global Centroid-Moment-Tensor (GCMT) project, represented by the white start on Figure 1. The centroid differs from an earthquake epicenter, in that the former best estimates the point source of the earthquake and not the first point of rupture, where the latter is the location on the sea floor, below which the earthquake focus marks the initiation of rupture along a fault. The GCMT project is considered to be a best estimate of earthquake source location as it uses a compilation of data from around the world, benefitting from decades of refinement. The 60 km offset between the GCMT source location and the numerous source locations determined through tsunami inversion is evidence that the tsunami produced was not solely a response to vertical movement of the seafloor along a fault.
Next, the authors determined where a submarine landslide was likely to have occurred. This was done by comparing high-resolution bathymetric maps (maps of seafloor depth) from surveys before and after the 2011 earthquake. A submarine landslide large enough to act as a secondary tsunami source would change the bathymetry beyond any associated error between the pre and post-earthquake surveys. Numerous landslide features were identified in the survey, but one in particular fit the bill for both magnitude and location. The identified feature measured 40 km wide and 20 km long, with changes in vertical elevations on the order of 100 m. To add to the evidence of a submarine landslide, the authors performed a slope stability analysis. This analysis takes into consideration the sediment characteristics (such as grainsize [size of particles] and water content), estimates of density, bathymetry prior to the earthquake, and slope angles; all which can contribute to a landslide. The results of the slope stability analysis suggest that the identified region was very likely to fail prior to the earthquake.
The waveforms produced by a tsunami generated by a submarine landslide differs significantly from one produced by seafloor motions. Submarine landslides displace water over a much smaller region than vertical motions along a fault, and therefore produce tsunamis with much shorter wavelengths. These shorter wavelengths (and thus higher frequency) were recorded by nearshore buoys. The addition of a short wavelength tsunami source accounts for the discrepancies in waveform observations.
The authors lastly validated tsunami simulations by comparing information from the simulations to tsunami wave characteristics from buoy observations and field surveys of tsunami run up height and inundation (or how far inland became covered in water). Figure 2 shows run up heights and inundation from field surveys (black dots) as compared to model simulations of inundation (red line). Simulating the single source earthquake based tsunami and the dual source scenario including the submarine landslide yield very similar buoy observations, except at the buoys nearest to maximum run up heights. The dual source scenario predicts the highly focused waves that are consistent with short wavelength submarine landslide generated tsunamis (Figure 3).
Conclusions and Significance
A tsunami generated solely from the vertical movements of the seafloor due to a strong earthquake cannot completely explain the localized maximum tsunami wave run ups as well as the observation buoy data indicating higher frequency waves. The authors were able to locate and characterize a submarine landslide that when used in simulations, accounted for the discrepancies in observational data.
The unpredictable nature of earthquakes and tsunamis combined with their devastating effects necessitate rapid alert systems, accurate observational networks and proper engineering of infrastructure. Better understanding of how tsunamis are generated, especially in complicated multi-source scenarios can only benefit coastal societies which rely on early warning systems and evacuation plans.