Laboratory experiments and numerical modelling of tsunamis generated by rock slides into fjords

The tsunami disasters following the huge earthquakes in the Indian Ocean (2004) and outside Japan (2011) shocked the whole world. These events also led to an increased awareness concerning potential Norwegian tsunamis. In Norway the major threat of this kind is due to sub-aerial rock slides into fjords and lakes. A slide is expected from the unstable rock slope at Åkerneset in Storfjorden, Sunnmøre. Such an event may produce a violent tsunami. In order to investigate the rock slope and assess the potential tsunami hazard the Åknes-Tafjord project (ÅTB) was started in 2004. Later this project has evolved into a monitoring project under the Norwegian Water Resources and Energy Directorate (NVE). Tsunamis in fjords are substantially different from oceanic tsunamis, such as those from 2004 and 2011. A rock slide may generate extreme waves locally and these will also retain a large wave height for a long propagation distance in the narrow waterways of a fjord. Hence, such waves require different tools for analysis than the oceanic tsunamis which, for instance, have vanishing amplitudes in deep water. The joint tsunami group at the Norwegian Geotechnical Institute and the University of Oslo has investigated tsunamis in fjords n the NFR (Norwegian Research Council) project "Laboratory experiments and numerical modelling of tsunamis generated by rock slides into fjords". The likely future incident at Åkerneset is an obvious starting point for the research in the tsunami group in Oslo. A series of experiments was performed in the facilities of the Coast and Harbour laboratory, Trondheim, where 1:500 model of the inner part of Storfjorden was available. To relate the model data to full scale all lengths (such as depths and widths of the fjord) must be multiplied by 500, whereas times (such as arrival times, periods of inundation) and velocities (of slide, waves and current) must be multiplied by 22.4. Wave heights and currents were measured in transects at a number of instrumentation bridges. Inundation of land, including flow depth and current velocities, were measured at the locations corresponding to the villages Hellesylt and Geiranger. In terms of full scale magnitudes the slide in the experiments had a volume of 40 million cubic meters and a speed at impact of 150 km per hour. At Hellesylt and Geiranger the wave runup was 40 meter and 30 meter, respectively. It is crucial that experiments is well controlled and yield repeatable measurements. For these experiments that was a challenge, in particular with respect to the slide motion. Additional experiments have been performed in simpler facilities, at the Hydrodynamic Laboratory at the University of Oslo, to compare granular and block slides. It is also important to assess the effect of the down-scaling of the model relative to the fjord. To this end, dedicated experiments on the evolution of the thin shear layers at the bottom have been performed and compared with theory. Little research has previously been made on these effects in the present context. It turns out that they may reduce runup heights on beaches considerably in model tanks. The experiments in the Storfjorden model do provide vital information on the potential tsunami. In addition they may be used for validation of computer models, which then may applied to cases elsewhere with increased confidence. The experimental data will be made available to the international research community through the European ASTARTE project, to which the Norwegian Geotechnical Institute is a partner. A range of computer models have been studied in the project. One group comprise general, but computationally demanding, models which may describe the wave generation by the slide. A systematic benchmarking of different models of this kind has revealed large variations in accuracy and the presence of a number of artifacts, even for models that conceived as well established in the research community. Another group of models have been tested with respect to wave propagation in the fjord system. The computation of waves in the narrow fjords, with steep beaches and complex geometries is generally more demanding than the computation of tsunamis in the open ocean. The work on combining the models, in order to obtain an integrated model for generation and propagation, is still in progress. Further down the road the outcome of this will be compared to the experiments. In short, the research in the project has provided a more profound insight in the limitations and performance of the different modeling tools. This is a solid basis for the remaining work. Hence, the tsunami research community in Norway is strengthened with respect analysis of tsunamis in general and participation in projects worldwide.

The primary objective of the project is to establish procedures and facilitate operational models for assessment and analysis of possible future rock slide tsunamis. Increased understanding and improved computational tools for tsunamis generated by rock slides are sought through a combination of laboratory studies and numerical modeling. Coupling of various mathematical models for rock slide generation, propagation, and run-up of waves will be emphasized. One of the possible tsunamigenic rock slopes in Norway today is located at Åknes along Storfjorden, western Norway. As a result of the need for risk assessment and management in the area, the Åknes/Tafjord project was initiated in 2004 (from 2009: Åknes/Tafjord Beredskap IKS) as a comprehensive R &D, monitoring, and early warning project related to large unstable rock slopes and their consequences in terms of tsunamis threatening local communities. The maximum volume of the potential Åknes rock slide is estimated to 54 million m3, and movements of up to 20 cm/year are recorded more than 900 m.a.s.l. The present findings of the Åknes/Tafjord project and the existing and unique 1:500 scale fjord model facilities form an excellent starting point for additional basic research required to establis h procedures and operational models. The deliverables will be a significant contribution to reduce hazard and risk related to rock slide tsunamis in fjords, lakes, and hydropower reservoirs. The proposed project will be a close cooperation between the Department of Mathematics at University of Oslo (UiO), the Norwegian Geotechnical Institute (NGI), and SINTEF Coast and Harbour Laboratory (CHL). Utilization of the 1:500 scale fjord model established by CHL in the Åknes/Tafjord project is conside red of uttermost importance. In the modeling part the project we will also collaborate closely with Cornell University. For the experimental part we have close links to Georgia Institute of Technology.