TSUNAMIs induced by large LANDslides

The fatal destruction due to the tsunamis in the Indian Ocean 2004 and Japan 2011 has changed society's perspective on how to deal with such rare and high-consequence natural disasters. The spotlight is put on the risk imposed on critical facilities such as nuclear power plants that should withstand tsunamis of low probability. Earthquakes cause most tsunamis. However, landslides have a potential for generating larger local tsunamis. Landslides of extreme size are today evident from field surveys around the world. However, the power of such landslides to generate tsunamis is presently not well constrained. Therefore, we lack established methods for assessing the present day landslide-induced tsunami hazard. To reduce this scientific gap, it is crucial to understand the dynamics of submarine landslides and volcanic flank collapses and their tsunami generation. Different types of landslides may result in different tsunami generation. In this project, we have developed new approaches model the failure and flow processes of submarine landslides in different ways. Furthermore, we have focussed on reviewing and extending existing theories for tsunami generation from landslides. The models are now being put to use in order to understand better the processes that govern tsunami generation from landslides. Massive pre-historical submarine slide tsunamis are presently being studied, and we shed new light on how these landslides generated tsunamis. Since there are relatively few observations of landslide tsunamis, we have to resort to calculations, in order to unveil the mechanisms. In the first phase of the project we have therefore a set of simulations in simple geometries, to delineate how the main mechanisms in real and more complex events, such as the Storegga landslide outside the coast of Norway some 8,000 years ago, and Papua New Guinea in 1998, takes place, but also how they differ greatly. We look into how the velocity, acceleration, landslide dimensions and depths of generation combine to influence the shape and size of the tsunami. The generation can be compared to an airplane moving close the speed of sound. The speed of the tsunami propagation speed increases with the depth. A landslide moving fast in shallow waters will generate waves in an efficient manner, which is the case in with rock slides impacts in fjords as well as shallow landslides in Norwegian fjords. However, for the largest submarine landslides, the generation process is primarily governed by the acceleration of the landslide. An important emphasis in this project has been the development of a new submarine landslide model that take into account the effect of the terrain and changes in material properties during flow, as well as the resistance forces of the ambient fluid. This is essential to model the large run-out of submarine mass movements. By means of this model, we can for the first time realistically model both the landslide run-out and the tsunami generation from large submarine landslides. To this end, the model have been used to explain the link between the landslide dynamics and the tsunami generation of the largest observed landslide in recorded history, the 1929 Grand Banks landslide offshore Canada. A second example is the modeling of the combined landslide and tsunami of Statland, Norway 2014. Field evidence of past submarine landslide offshore Norway suggest that a multistage landslide development is essential for many types of landslides. Multistage landslide evolution strongly influence tsunami generation, but are rarely studied. For the first time, we have implemented a multistage landslide model into a tsunami model and investigated how it influences the tsunami generation. Preliminary findings show that the multistage failure process have important implications for the tsunami generation that have been previously unknown. We have now modelled two major events, namely the Trænadjupet Slide occurring 4500 years ago, and the Storegga Slide. These landslides have enormous volumes and slide areas. Example wise, Storegga involved a slide volume that exceed the area of Denmark. The Storegga Slide generated a huge ocean wide tsunamis, whereas we have not found traces of a tsunami due to Trænadjupet. By means of new models, it is shown that retrogressive landslides inherit an inefficient tsunami generation mechanism. Consequently, the giant Trænadjupet slide, with a volume of 500 km3, apparently only produced a moderately sized wave.

Flodbølgene som rammet det Indiske hav i 2004 og Japan i 2011 forandret samfunnets håndtering av slike hendelser. Disse hendelsene har også økt det globale fokuset på sårbarheten til kritiske installasjoner som kjernekraftverk som følge av flodbølger (tsunamis) og andre sjeldne naturkatastrofer. Selv om de fleste flodbølger forårsakes av jordskjelv, vil potensialet for ødeleggelser fra tsunamis generert av skred lokalt kunne være større. Flere ekstremt store undersjøiske skred er kartlagt ulike steder i verden, men hvilket potensiale slike store skred har for å generere flodbølger er ikke avgrenset. Økt kunnskap om mekanismene for generering av flodbølger fra store undersjøiske skred eller flankekollapser fra vulkaner er helt avgjørende for å kunne forstå faren og utvikle metoder for å tallfeste den. Prosjektet vil utvikle nye metoder for beregning av flodbølger fra undersjøiske skred og vulkankollapser. Hovedfokuset er å beskrive mekanismene for bølgegenering ved hjelp av regnemodeller. Feltundersøkelser av store undersjøiske skred utenfor norskekysten og Kanariøyene indikerer at skredene har løsnet i en trinnvis (retrogressiv) prosess. Koblingen mellom retrogressive skred og flodbølger er lite studert, og vi kjenner derfor ikke til hvordan den trinnvise løsnemekanismen påvirker bølgegenereringen. Materialegenskapene til skredene endrer seg under skredets bevegelse. Prosjektet vil derfor fokusere på å beskrive hvordan skredets materialegenskaper og friksjon mot omgivelsene påvirker bølgegenerering. Ved å kombinere nye modeller for skreddynamikk og flodbølger, skal det gjennomføres beregninger av tidligere hendelser med det formål å forklare hovedmekanismene knyttet til generering. For dette formålet vil beregningene knyttes til tilgjengelige data fra Atlanterhavskysten utenfor Europa og Nord-Afrika og i Middelhavet. Modellene skal også brukes til å beregne potensielle fremtidige hendelser.