Abstract

Tsunami waves pose a threat to coastal communities through overland inundation and overtopping of coastal defence structures and river walls. Tsunami wave overtopping in rivers and in straight or converging channels is yet to be studied in detail. The present study evaluates the performance of current numerical models that are employed in various tsunami warning centres across the world, in terms of their accuracy for the prediction of tsunami wave characteristics in converging channels. The study evaluates each model performance against a benchmark dataset obtained from novel physical model experiments of representative tsunami waves (solitary waves and bores) overtopping a river wall in a converging channel. The complex process of simultaneous steepening and overtopping was observed to be different for solitary waves and bores. The experimental data were then used to compare results from a suite of different numerical models to predict wave heights and overtopping volumes in straight and converging channels. Different numerical models ranging from nonlinear/nondispersive in two-dimensions (2D) to fully nonlinear/dispersive in three-dimensions (3D) were used. The widely used models based on the nonlinear shallow water equations ( ANUGA , BASEMENT , and TELEMAC 2D) fail to capture the wave profiles and heights, and consequently do not accurately predict the overtopping volumes in channels for both breaking and non-breaking waves. A Boussinesq model ( TELEMAC 2D) was found to produce results consistent with experimental measurements for non-breaking waves only, with an error of up to 8% and 20%, for wave heights and overtopping volumes, respectively. On the other hand, the wave heights and overtopping volumes are underpredicted and overpredicted, for solitary waves and bores, respectively, using a three-dimensional nonhydrostatic model ( TELEMAC 3D). When the overtopping occurs in the converging channel, reflection of the incident waves is minimal in both the numerical and the physical models. However, in the absence of overtopping, both the numerical models and the classical analytical long-wave model fail to correctly predict the observed minimal reflection of non-breaking solitary waves in the converging channel. The reasons for this remain to be investigated. • Novel experiments of representative tsunami waves overtopping river wall were conducted. • 2D and 3D numerical modelling results were compared with experimental data. • Nonlinear shallow water models fail to predict solitary wave/bore heights, shapes and overtopping volumes in channels. • Boussinesq model produces satisfactory results for non-breaking solitary waves only. • Numerical models and classical analytical long-wave model fail to predict solitary wave reflection in a converging channel.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call