Did a submarine landslide contribute to the 2011 Tohoku tsunami?

Transcription

1 Press Release Did a submarine landslide contribute to the 2011 Tohoku tsunami? 1. Key points Large tsunami amplitudes in Northern Tohoku (Sanriku) suggest that a secondary tsunami source lies offshore this region. The best candidate for the secondary tsunami source is a submarine landslide near the trench. The earthquake source alone cannot explain the tsunami waveforms and amplitudes, but the earthquake plus landslide source can. 2. Summary The source of the exceptionally high tsunami runups of up to 40 m recorded along the central Sanriku coast between 39.2 N and 40.2 N on March 11, 2011 has not previously been satisfactorily explained. Based on an analysis of seismic and geodetic data, together with recorded tsunami waveforms, we propose that, while the primary source of the tsunami was the vertical displacement of the seafloor due to the earthquake, an additional tsunami source is also required. An analysis of the travel times of higher frequency tsunami waves suggests that the additional source is located near the trench at about latitude 39.4 o N and longitude 144 o E. We propose that the additional tsunami source was a submarine mass failure (SMF i.e., a submarine landslide). We find that: (1) A comparison of pre- and posttsunami bathymetric surveys reveals that there were tens of meters of vertical seafloor movement at the proposed SMF location, and a slope stability analysis confirms that the horizontal acceleration from the earthquake was sufficient to trigger an SMF. (2) Forward modelling of the tsunami generated by a combination of the earthquake and the SMF reproduces the recorded on-, near- and offshore tsunami observations well, particularly the high frequency component of the tsunami waves off Sanriku, which were not well simulated by previous models. The conclusion that a significant part of the 2011 Tohoku tsunami was generated by an SMF source has important implications for estimates of tsunami hazard in the Tohoku region as well in other tectonically similar regions. 3. Details Fig. 1 shows the general setting of this study. Particularly important tsunami waveform data were recorded at the buoys in the near offshore of Iwate Prefecture (nos. 4, 5, and 6) and at the far offshore buoy DART# A backward travel-time analysis of the high frequency component of the waveforms recorded at these buoys identified the most likely SMF tsunami source as being located at about latitude 39.4 o N and longitude 144 o E, and a slope stability analysis confirmed that the SMF could have failed due to earthquake loading.

2 A comparison of bathymetric data before and after the earthquake validated the presence of a large SMF at this location, with post-earthquake vertical motions on the order of 100 m. This is consistent with an SMF failing as a rigid slump with a short runout. Our research demonstrates that the high runups in Sanriku can be accounted for by the combination of the earthquake source and an SMF located almost directly east of the central Sanriku coast (Fig. 2), but cannot be accounted for by the earthquake source alone. Our modeling of tsunami waveform data at offshore buoys 4, 5, and 6, and the DART#21418 buoy shows that the observed higher frequency waves can be best explained by a combination of high-frequency waves generated by an SMF located off the Sanriku coast (triggered with a 135 s delay, consistent with the propagation time of seismic waves from the main rupture area), and long-period waves generated by the earthquake (Figs. 3 and 4). Comparison of our simulations for the combined earthquake plus SMF source with those for other studies demonstrates that our source model is more successful in reproducing the high frequency components of the observed waveforms. 4. Publication Journal: Marine Geology Title: Did a submarine landslide contribute to the 2011 Tohoku tsunami? Authors: David R. Tappin, Stephan T. Grilli, Jeffrey C. Harris, Robert J. Geller, Timothy Masterlark, James T. Kirby, Fengyan Shi, Gangfeng Ma, K.K.S. Thingbaijam, P.M. Mai doi: /j.margeo

3 5.Figures Figure 1. (a) Seismotectonic setting of the March 11, 2011 Tohoku earthquake. Small black dots are aftershocks (from March 11, 2011 to May 6, 2011), which approximate the surface projection of the main rupture. Green and blue circles are locations of centroids for solutions derived from seismic inversions and tsunami inversions, respectively, and the black dot is the USGS centroid. White star is the Global GCMT Project centroid, and red star is the average of the centroids obtained by the tsunami waveform inversions. Purple square is DART Buoy #21418, and purple diamonds labeled 4, 5, and 6 respectively are North Iwate, Central Iwate, and South Iwate buoys, respectively. Other offshore GPS buoys are shown as black diamonds. The brown polygon approximates an area whose bathymetry we studied in detail. The black rectangle show the region shown in panel b. (b) Enlarged view of the rectangle in panel a. 3

4 Figure 2. Tsunami runup and flow depth measured by Mori et al. (Coastal Engineering Journal, 2012) in field surveys (black dots) (a) runup. (b) inundation/flow depth at the shoreline. Simulations using the earthquake source alone (blue line) cannot explain data between about 39 N and 41 N, while the dual earthquake plus SMF source (red dots) fits the observations better. 4

5 Figure 3. Observed and calculated tsunami waveforms (surface elevations). Horizontal axis is time after the origin time of the earthquake, in minutes. Panels (a) to (i) are for buoys numbered 1 through 9 in Fig. 1, respectively, and panel (j) is the DART#21418 buoy. The observed data (black trace) is compared to calculated waveforms for the earthquake-only source (light blue). The red trace is the residual (the difference between the observed and calculated waveforms). 5

6 Figure 4. Observed and calculated tsunami waveforms (surface elevations). Horizontal axis is time after the origin time of the earthquake, in minutes. Panels (a) to (i) are for buoys numbered 1 through 9 in Fig. 1, respectively, and panel (j) is the DART#21418 buoy. The observed data (black trace) is compared to calculated waveforms for the earthquake-only source (light blue), and for the dual earthquake plus SMF source (green). The calculated waveforms for the dual source fit the data better than the earthquake-only data in panels d, e, f, and j (i.e., for buoys 4, 5, 6, and the DART#21418 buoy). 6