5:15 PM - 7:15 PM
[HDS10-P09] Generation mechanism of the large tsunami caused by the 1946 Aleutian tsunami earthquake
Keywords:The 1946 Aleutian Tsunami, Deep-sea Landslide Tsunami
The 1946 Aleutian earthquake (Ms=7.2, Mt=9.3) generated a huge tsunami, a maximum surveyed tsunami height of 40 m, along the coast of Unimak Island in the Aleutian Islands (Okal et al., 2003). This event is categorized as the most unusual tsunami earthquake in the world. Previous studies indicated that the huge tsunami near the source region should be generated by a landslide along the continental slope initiated by the strong shaking. The geomorphological study along the continental slope indicated that the landslide depositional trace near the source area was created by the landslide caused by the 1946 Aleutian earthquake (Fryer et al., 2004).
In this study, we developed a deep-sea landslide tsunami computation method by modifying a previously developed Tsunami Squares method (Wang et al., 2015) with a different friction term and an effect of deep water. To compute a deep-sea landslide, an initial landslide mass distribution is necessary to be estimated. At first, the landslide depositional trace which was previously found was eliminated by using GMT surface command. Next, the same amount of landslide mass eliminated from the recent topography was put back in the region of the steep continental slope as the initial landslide mass distribution. The landslide computations were carried out by changing a frictional factor to fit the observed landslide depositional trace. Then the tsunami computation with the landslide computation was carried out using the estimated frictional factor. After 15 minutes, the tsunami generation computation with the landslide model using the modified Tsunami Squares method was switched to the non-linear long-wave tsunami computation with the tsunami inundation (JAGRUS) to compute tsunami heights along the coast of the Aleutian Islands. In this tsunami computation, four levels of the nested grid systems (27, 9, 3, and 1 sec.), were used to compare with the observed tsunami heights along the coast. In addition to the tsunami generated by the landslide, the tsunami generated by the earthquake source was also computed using the fault model previously estimated.
The result showed that the surveyed tsunami height distribution along Unimak Island including the maximum tsunami height of 40 m was well explained by the computed one. We also computed the tsunami height distribution by using only the earthquake source. The result showed that computed tsunami heights were much less than the observed ones. This suggested that the main tsunami generation mechanism along the coast near the source is the landslide near the continental slope, although the tsunami generation due to the earthquake faulting is also necessary. Finally, we conclude that our method developed to compute deep-sea landslide tsunamis is useful for studying the landslide tsunami generation mechanism.
References
Fryer et al., 2004, Source of the great tsunami of 1 April 1946: a landslide in the upper Aleutian forearc, Marine Geology, 203, 201-218, doi: http://doi.org/10.1016/S0025-3227(03)00305-0.
Okal E.A., et al., 2003, Near-field survey of the 1946 Aleutian tsunami on Unimak and Sanak Island, Bull. Seis. Soc. Am., 93, 1226-1234, doi: 10.1785/0120020198.
Wang et al., 2015, Numerical modeling of rapid, flow-like landslides across 3-D terrains: a Tsunami Squares approach to El Picacho landslides, El Salvador, September 19, 1982, Geophys. J. Int. 201, 1534-1544, doi: 10.1093/gji/ggv095.
In this study, we developed a deep-sea landslide tsunami computation method by modifying a previously developed Tsunami Squares method (Wang et al., 2015) with a different friction term and an effect of deep water. To compute a deep-sea landslide, an initial landslide mass distribution is necessary to be estimated. At first, the landslide depositional trace which was previously found was eliminated by using GMT surface command. Next, the same amount of landslide mass eliminated from the recent topography was put back in the region of the steep continental slope as the initial landslide mass distribution. The landslide computations were carried out by changing a frictional factor to fit the observed landslide depositional trace. Then the tsunami computation with the landslide computation was carried out using the estimated frictional factor. After 15 minutes, the tsunami generation computation with the landslide model using the modified Tsunami Squares method was switched to the non-linear long-wave tsunami computation with the tsunami inundation (JAGRUS) to compute tsunami heights along the coast of the Aleutian Islands. In this tsunami computation, four levels of the nested grid systems (27, 9, 3, and 1 sec.), were used to compare with the observed tsunami heights along the coast. In addition to the tsunami generated by the landslide, the tsunami generated by the earthquake source was also computed using the fault model previously estimated.
The result showed that the surveyed tsunami height distribution along Unimak Island including the maximum tsunami height of 40 m was well explained by the computed one. We also computed the tsunami height distribution by using only the earthquake source. The result showed that computed tsunami heights were much less than the observed ones. This suggested that the main tsunami generation mechanism along the coast near the source is the landslide near the continental slope, although the tsunami generation due to the earthquake faulting is also necessary. Finally, we conclude that our method developed to compute deep-sea landslide tsunamis is useful for studying the landslide tsunami generation mechanism.
References
Fryer et al., 2004, Source of the great tsunami of 1 April 1946: a landslide in the upper Aleutian forearc, Marine Geology, 203, 201-218, doi: http://doi.org/10.1016/S0025-3227(03)00305-0.
Okal E.A., et al., 2003, Near-field survey of the 1946 Aleutian tsunami on Unimak and Sanak Island, Bull. Seis. Soc. Am., 93, 1226-1234, doi: 10.1785/0120020198.
Wang et al., 2015, Numerical modeling of rapid, flow-like landslides across 3-D terrains: a Tsunami Squares approach to El Picacho landslides, El Salvador, September 19, 1982, Geophys. J. Int. 201, 1534-1544, doi: 10.1093/gji/ggv095.