To better understand how large tsunamis are generated, it is important to be able to accurately model the sea level signals they generate. A number of researchers have identified systemic discrepancies between observed and modelled tsunami wave speeds for two recent major tsunamis, the 2010 Maule and 2011 Tohoku events. To account for these discrepancies, we developed a numerical tsunami propagation code solving the shallow water equation and including the effects of elastic loading of the sea-floor by the tsunami as well as a linear density profile in the sea water column. We found that both effects are important to explain the commonly observed difference between observations and simulations. We conclude that the density variation in the sea water column affects the wave speed without changing the waveform, whereas the loading effect has an effect on the wave speed and the waveform showing a negative phase before the main arrival due to the depression of the sea floor surrounding the tsunami wave. The combination of both effects is needed to achieve a better match between observations and simulations.

Snapshot of the tsunami generated by the 2010 Maule after 16 hours of propoagation. (Left) Numerical results from the ‘standard’ shallow-water wave equations (rigid substrate); (middle) Numerical results from the new equaitons tat account for tsunami self-loading and density stratification; (right) the vertical seafloor dislacement caused by tsunami loading of the sealfoor. (from Allgeyer S., Cummins P.R. (2014) Numerical tsunami simulation including elastic loading and seawater density stratification, 
Geophysical Research Letters 41 (7), 2368-2375).