1:45 PM - 2:00 PM
[HDS10-01] Generation mechanism of the 1929 Grand Banks tsunami caused by a deep-ocean landslide.
Keywords:1929 Grand Banks Earthquake, submarine landslide tsunami, tsunami, Tsunami Squares
The Grand Banks earthquake-induced submarine landslide tsunami, which occurred on April 1, 1929, in a region far from volcanic and plate boundary activity zones, represents a rare case where the fault movement itself did not generate the tsunami. Instead, seismic shaking triggered a submarine landslide, leading to tsunami excitation and significant damage. During this event, submarine cables installed near the source and in the deep-sea region were progressively severed both immediately after the earthquake and over a period of 13 hours (denoted by red crosses, red circles, and red question marks in Fig. a). Initially, the cable-break was confined to the source area (red crosses), but subsequently, it expanded southward along the gradient of the ocean topography (red circles). The annotated times at each location indicate the time elapsed from the earthquake occurrence to cable-break. This progressive break of submarine cables is attributed to the motion of landslide masses and the transformation of the submarine landslide into turbidity currents.
In this study, we developed a numerical simulation method to model the generation process of deep-ocean submarine landslide tsunamis and applied it to the 1929 Grand Banks event. Our objective was to explain both the observed tsunami waveform at Halifax, Nova Scotia, Canada (red rhombus in Fig. a), and the sequence of submarine cable-breaks through landslide mass motion modeling.
First, we modified the Tsunami Squares method, which was originally designed for landslide tsunamis in coastal regions, to incorporate deep-water effects, enabling the numerical simulation of tsunamis generated by submarine landslides at great depths.
Next, we preliminarily defined the landslide mass distribution within the area identified in previous studies based on ocean-floor acoustic surveys as the source area (shaded area in Fig. a). Using the improved simulation method, we calculated landslide mass motion and estimated the friction parameters required to reproduce the observed cable breakage timings.
Based on the estimated friction parameters, we refined the submarine landslide mass distribution that explains the observed tsunami waveform at Halifax. This was achieved by dividing the initial source area into four longitudinal areas (denoted by the red solid lines in Fig. b) and estimating the landslide mass volume in each area. The landslide mass volume within each area was adjusted sequentially, starting from the westernmost area (Area1), which is closest to Halifax, to fit the computed tsunami waveform to the observed waveform (Fig. c). Our results indicate that the total mass volume of landslide required to reproduce the first pulse of the observed tsunami waveform was approximately 450 km³. The estimated initial thickness distribution of the landslide masses is presented in Fig. b, with a maximum initial thickness of 79 m.
The estimated friction parameters obtained from preliminary simulations showed consistency with values derived from laboratory experiments in previous studies. Furthermore, the total estimated landslide mass volume falls within the range proposed by multiple prior studies. The submarine landslide mass distribution derived in this study explains both the first pulse of the observed tsunami waveform at Halifax and the timing of submarine cable-breaks.
In this study, we developed a numerical simulation method for submarine landslide tsunamis at great depths by improving an existing coastal landslide tsunami computation method. This allowed for a more detailed modeling of the tsunami generation and landslide mass motion of the 1929 Grand Banks submarine landslide tsunami, which had previously only been represented by simplified models. The improved numerical simulation method developed in this study is expected to be applicable to modeling other deep-ocean submarine landslide tsunamis and assessing the potential impact of future submarine landslide tsunamis.
In this study, we developed a numerical simulation method to model the generation process of deep-ocean submarine landslide tsunamis and applied it to the 1929 Grand Banks event. Our objective was to explain both the observed tsunami waveform at Halifax, Nova Scotia, Canada (red rhombus in Fig. a), and the sequence of submarine cable-breaks through landslide mass motion modeling.
First, we modified the Tsunami Squares method, which was originally designed for landslide tsunamis in coastal regions, to incorporate deep-water effects, enabling the numerical simulation of tsunamis generated by submarine landslides at great depths.
Next, we preliminarily defined the landslide mass distribution within the area identified in previous studies based on ocean-floor acoustic surveys as the source area (shaded area in Fig. a). Using the improved simulation method, we calculated landslide mass motion and estimated the friction parameters required to reproduce the observed cable breakage timings.
Based on the estimated friction parameters, we refined the submarine landslide mass distribution that explains the observed tsunami waveform at Halifax. This was achieved by dividing the initial source area into four longitudinal areas (denoted by the red solid lines in Fig. b) and estimating the landslide mass volume in each area. The landslide mass volume within each area was adjusted sequentially, starting from the westernmost area (Area1), which is closest to Halifax, to fit the computed tsunami waveform to the observed waveform (Fig. c). Our results indicate that the total mass volume of landslide required to reproduce the first pulse of the observed tsunami waveform was approximately 450 km³. The estimated initial thickness distribution of the landslide masses is presented in Fig. b, with a maximum initial thickness of 79 m.
The estimated friction parameters obtained from preliminary simulations showed consistency with values derived from laboratory experiments in previous studies. Furthermore, the total estimated landslide mass volume falls within the range proposed by multiple prior studies. The submarine landslide mass distribution derived in this study explains both the first pulse of the observed tsunami waveform at Halifax and the timing of submarine cable-breaks.
In this study, we developed a numerical simulation method for submarine landslide tsunamis at great depths by improving an existing coastal landslide tsunami computation method. This allowed for a more detailed modeling of the tsunami generation and landslide mass motion of the 1929 Grand Banks submarine landslide tsunami, which had previously only been represented by simplified models. The improved numerical simulation method developed in this study is expected to be applicable to modeling other deep-ocean submarine landslide tsunamis and assessing the potential impact of future submarine landslide tsunamis.