We show a good match to estimated wave heights, but these might be further refined by adjusting the slide parameters further, as per Bondevik et al. (2005). The Fluidity modelling presented here assumes one particular type of slide movement as a single rigid block. It is unclear how somewhat more realistic slide behaviour would affect tsunami magnitudes and inundation heights around surrounding coastlines. More work is required in order to attempt

to improve the veracity of the model by altering the slide initiation and shape and to study the effects of such BKM120 supplier changes and how they compare to the changes described here with respect to resolution. The effects of bathymetric and coastline resolution are important in determining accurate simulated run-up heights of tsunamis. We have shown that the higher resolution coastline and bathymetric simulations produce simulated wave heights that are in closer agreement to inferred wave heights from observational data and have some sense of numerical convergence. Overall numerical resolution is important to minimise numerical errors and for this simulation a fixed mesh of 12.5 km is sufficient with coarse coastlines to reproduce the work of Harbitz (1992). However, as along-coastline resolution

increases, commensurately higher mesh resolution is required around the coasts. Assumptions of the slide AZD1208 acting as a rigid block, accelerating to 35 m/s, are similar to previous studies, but as the Storegga slide is thought to be retrogressive and disintegrate as it moved, more work is required to ascertain the effects of this on wave run-up heights. In establishing the spatial resolution of coastlines and palaeobathymetry required to adequately model the Storegga slide-generated tsunami, this work provides a foundation on which simulations examining the effect of complex slide parameters can build. Given the simplicity of our slide model and the absence of an inundation model, our multiscale models of the Storegga submarine slide generated tsunami shows remarkable agreement with inferred wave-heights from sediment deposits along the Norwegian and Scottish coasts.

The agreement within the Faroe Islands is less good, with a simulated wave height that is around a 6 m too small, but consistent with previous studies (Bondevik et al., 2005). Our multiscale not model simulates the Storegga tsunami for 15 h, tracking the wave propagation into the southern North Sea, predicting wave heights of less than 1 m for the northern coast of mainland Europe. The addition of palaeobathymetric information, neglected in previous studies, aids the match to observed data within the region where our data is valid and makes a substantial difference in the southern North Sea region and around the Shetland Islands. However, the use of realistic palaeobathymetry makes little difference along the Norwegian coast, which was the primary focus of previous studies.