Elsevier

Coastal Engineering

Volume 56, Issue 9, September 2009, Pages 897-906
Coastal Engineering

Experimental investigation of impact generated tsunami; related to a potential rock slide, Western Norway

https://doi.org/10.1016/j.coastaleng.2009.04.007Get rights and content

Abstract

Two-dimensional experiments of wave generation from the possible Åkneset rock slide have been performed using solid block modules in a transect with a geometric scaling factor of 1:500. The width of the slide model was kept fixed at 0.45 m. The length of the blocks spanned from 1 to 2 m, the thickness was either 0.12 or 0.16 m and the front angle was 45°. Maximum water depth was 0.6 m with the slide plane having an angle of 35°. Three different scenarios were studied. Only the run out side was modelled.

Surface elevations at three locations outside the sloping region were measured with ultra sonic wave gauges and discussed in light of hydrodynamic wave theory. Particle Image Velocimetry (PIV) was used to extract instantaneous velocity fields. Comparison is made between experimental velocity profiles and profiles consistent with a Boussinesq theory. High speed video of the impact was recorded and used to determine qualitative aspects of the forward collapse of the impact crater (backfill wave).

Introduction

Slides are today recognised as an important tsunami source. Local mass gravity flows and slumps are believed to be regularly triggered by earthquakes. In some cases, such as for the 1998 Papau New Guinea event, (Bardet et al., 2003, Lynett et al., 2003), the landslide generated waves are the main hazard. Large scale slides in the ocean are rare, but a series of pre-historic events has been detected, (Masson et al., 2006). Historic examples of larger slides producing tsunamis include the Shimabara event, Japan 1792 and the slide at the Ritter Island Volcano into the sea northeast of New Guinea in 1888, which is the largest lateral collapse of an island volcano to be recorded in historical time, (Ward and Day, 2003). A major future collapse of the Cumbre Vieja volcano at La Palma on the Canary Islands has been suggested, (Ward and Day, 2001), but is regarded as unlikely by many geologists, (Wynn and Masson, 2003, Masson et al., 2002, Masson et al., 2006).

In confined water bodies, such as lakes, fjords and dams, large amplitude tsunamis may be generated from even moderate sized subaerial slides with disastrous consequences for near-shore settlements. An example of such a disaster is the 1934 rock slide in Tafjord, Western Norway, when a 1.5·106 m3 rock formation plunged into the fjord and presumably released another 1.5·106 m3 of submerged masses. The run-up heights were up to 60 m and 41 people perished. Famous further examples are Lituya Bay, Alaska, where an earthquake caused a subaerial rock slide into Gilbert Inlet on July 8, 1958, yielding a maximum run-up of 524 m, (see Fritz et al., 2001) and the Vaiont reservoir disaster, 1963, where the waves over-topped the dam and claimed 2500 casualties, (see e.g. Semenza and Ghirotti, 2000).

The flow dynamics of large masses entering a water body are highly complex. Topography and bathymetry as well as the shape, velocity, density and composition of the rock slide combine to determine the wave generation. Murty (2002) studied the volume dependency for waves from submarine landslides, and fitted a linear regression between wave amplitude and slide volume based on observations. The fit of the regression is rather poor, indicating that other parameters could be equally important. Fritz (2002), Fritz et al., 2003a, Fritz et al., 2003b, Fritz et al., 2004 studied experimentally the waves created by a deformable landslide in a 2D wave tank. Their extensive work resulted in classifications of the waves as either weakly nonlinear oscillatory, nonlinear transition waves, solitary-like or dissipative transient bores. Based on Froude number and dimensionless slide height, they found a criterion to determine the collapsing direction of the impact crater if separation at the impact occurred. Zweifel et al. (2006) also studied experimentally the non-linearity of impulse waves. Huber and Hager (1997) looked at both 3D and 2D impulse waves. Their work studied the importance of the different factors controlling the amplitudes, and found that the impact angle and volume of the slide were the governing parameters. Raichlen and Synolakis (2003) performed experiments with a freely sliding wedge representing a land slide. They measured the wave elevation and run-up. Liu et al. (2005) used the same type of experiments to validate a numerical model, based on the large-eddy-simulation approach.

The work herein considers the potential Åkneset rock slide, (see e.g Deron et al., 2005 and Norem et al., 2007). Åkneset is a rock formation in Storfjorden/Synnulvsfjorden, located in the Stranda municipality in Western Norway (see Fig. 1), that is the fjord branch next to the Tafjord where the 1934 event occurred. At Åkneset a volume of 15–80·106 m3 of unstable rock, shaped as large blocks, have been detected, (Blikra et al., 2005). Such a landslide could have devastating consequences for the settlements in the vicinity, especially at locations like Hellesylt and Geiranger. In addition, during summertime the narrow Geiranger fjord is one of the most popular cruise ship destinations in the world.

Geological surveys have revealed a number of deposits from previous incidents in Storfjorden, (Blikra et al., 2005). These deposits are laterally confined and are terminated by the bottom slope at the facing side of the Fjord. Hence, they are very different in nature compared with wide fans. This suggests that a block model may be more appropriate than a granular one in the present case. Therefore, a block slide in a 2D transect of fjord is investigated experimentally with a high speed camera and PIV. The facing slope of the fjord is not modelled and the waves are allowed to propagate freely to an absorbing beach at a distance of 10.5 m from the slope. The present experiments aim to provide support for the ongoing Åknes/Tafjord project as well as a general study of impact generated impulse waves. Both the near field and far field are investigated by high speed video and particle image velocimetry (PIV) and the outcome is compared to related investigation for granular slides by Fritz (2002), Fritz et al., 2003a, Fritz et al., 2003b, Fritz et al., 2004. We observe different characteristics both in the generation phase and in the far field.

Section snippets

Experimental setup

The experiments were performed in the wave tank at the Hydrodynamics Laboratory at the University of Oslo. The tank is 25 m long, 0.51 m wide and 1 m deep. The bottom and sides are made of transparent glass, facilitating observations of the flow from the outside.

The model of the Åkneset site was comprised of a 1:500 geometrical scaled cross section of the run-out side of the fjord, with the topography provided by the Åkneset project (Norem et al., 2007). The water depth (d) in the model was

Far field

For all scenarios, the leading wave had the largest amplitude (Fig. 7). The most important wave characteristics for the 3 first waves in the wave train are summarized in Table 2.

As could be expected from the volumes and the findings of Huber and Hager (1997), scenario 1 with the largest volume had the largest leading wave. The difference in amplitude for the leading wave between scenarios 2 and 3 was small, despite the difference in slide height. Slide height S thus seems to be less important

Conclusions

Our investigation is focused on impact generated impulse waves, more specifically subaerial rock slides with solid slide models. The resulting waves are nonlinear and according to the criterion and classification by Fritz et al. (2004), all three scenarios create nonlinear oscillatory waves.

Qualitative investigation of the collapsing crater using high speed video shows that the resulting waves are forward collapsing. Comparison with the results of Fritz et al. (2003b) yields that their

Acknowledgments

This study was done as a part of the larger Åkneset project. We are deeply indebted to Svein Vesterby and Arve Kvalheim at UiO for providing technical expertise and building the slide model. Further, we wish to thank Sylfest Glimsdal at NGU for providing map and constructive discussions. The conveyor belt was made at Sintef Coast and Harbour Research Laboratory.

References (28)

  • MassonD. et al.

    Slope failures on the flanks of the western Canary Islands

    Earth-Science Reviews

    (2002)
  • BardetJ.-P. et al.

    Landslide tsunamis: recent findings and research directions

    Pure and Applied Geophysics

    (2003)
  • BlikraL.H. et al.

    Quantification of rock-avalanche and tsunami hazard in Storfjorden, western Norway

  • DalzielS.B.

    DigiFlow user guide

    (2006)
  • DeronM.H. et al.

    Analysis of a complex landslide with an airborne lidar dem

    Geophysical Research Abstracts

    (2005)
  • Fritz, H.M., 2002. Initial phase of landslide generated impulse waves. PhD thesis, Swiss Federal Institute of...
  • FritzH.M. et al.

    Lituya bay case: rockslide impact and wave run-up

    Science of Tsunami Hazards

    (2001)
  • FritzH.M. et al.

    Landslide generated impulse waves. 1. instantaneous flow fields

    Experiments in Fluids

    (2003)
  • FritzH.M. et al.

    Landslide generated impulse waves. 2. hydrodynamic impact craters

    Experiments in Fluids

    (2003)
  • FritzH.M. et al.

    Near field characteristics of landslide generated impulse waves

    Journal of Waterway, Port, Coastal and Ocean Engineering

    (2004)
  • HuberA. et al.

    Forecasting impulse waves in reservoirs

  • LiuP.L.-F. et al.

    Runup and rundown generated by three-dimensional sliding masses

    Journal of Fluid Mechanics

    (2005)
  • LøvholtF. et al.

    Oceanic propagation of a potential tsunami from the La Palma Island

    J. Geophys. Res

    (2008)
  • LynettP.J. et al.

    Field survey and numerical simulations: a review of the 1998 Papua New Guinea tsunami

    Pure and Applied Geophysics

    (2003)
  • Cited by (69)

    • Numerical simulations of water waves generated by subaerial granular and solid-block landslides: Validation, comparison, and predictive equations

      2022, Ocean Engineering
      Citation Excerpt :

      Zweifel (2004) studied subaerial landslide dynamics through physical modelling and reported that water depth and slide thickness are the governing parameters on the maximum wave amplitudes. Saelevik et al. (2009) focused on the effect of slide geometry on impulsive waves generated by subaerial landslides. The results reported by Saelevik et al. (2009) indicated that, by decreasing the slide length, a deeper void for the collapse of leading wave is created, and therefore a larger trailing wave is generated.

    • Study on solid block landslide generated tsunami using a modified δ-les-SPH model

      2022, Ocean Engineering
      Citation Excerpt :

      This poses new challenges for numerical simulation. In this section, the four-block landslide-tsunami experiment test of the Sælevik et al. (2009) is selected to detect the proposed SPH procedure. The geometric of the experiment in the research of Sælevik et al. (2009) is shown in Fig. 14.

    View all citing articles on Scopus
    View full text