Should tsunami simulations include a nonzero initial horizontal velocity?

Abstract

Tsunami propagation in the open ocean is most commonly modeled by solving the shallow water wave equations. These equations require initial conditions on sea surface height and depth-averaged horizontal particle velocity or, equivalently, horizontal momentum. While most modelers assume that initial velocity is zero, Y.T. Song and collaborators have argued for nonzero initial velocity, claiming that horizontal displacement of a sloping seafloor imparts significant horizontal momentum to the ocean. They show examples in which this effect increases the resulting tsunami height by a factor of two or more relative to models in which initial velocity is zero. We test this claim with a "full-physics" integrated dynamic rupture and tsunami model that couples the elastic response of the Earth to the linearized acoustic-gravitational response of a compressible ocean with gravity; the model self-consistently accounts for seismic waves in the solid Earth, acoustic waves in the ocean, and tsunamis (with dispersion at short wavelengths). Full-physics simulations of subduction zone megathrust ruptures and tsunamis in geometries with a sloping seafloor confirm that substantial horizontal momentum is imparted to the ocean. However, almost all of that initial momentum is carried away by ocean acoustic waves, with negligible momentum imparted to the tsunami. We also compare tsunami propagation in each simulation to that predicted by an equivalent shallow water wave simulation with varying assumptions regarding initial velocity. We find that the initial horizontal velocity conditions proposed by Song and collaborators consistently overestimate the tsunami amplitude and predict an inconsistent wave profile. Finally, we determine tsunami initial conditions that are rigorously consistent with our full-physics simulations by isolating the tsunami waves from ocean acoustic and seismic waves at some final time, and backpropagating the tsunami waves to their initial state by solving the adjoint problem. The resulting initial conditions have negligible horizontal velocity.

Modeling the Sources of the 2018 Palu, Indonesia, Tsunami Using Videos From Social Media

“…Furthermore, the approach does not consider dynamic contributions from the seafloor movement; in particular, mass is conserved, but the injection of momentum from the horizontal seafloor motion is ignored. While existing studies show that momentum is indeed transmitted to the oceans by the horizontal coseismic deformation (Song et al, ; ), this is mostly carried away by ocean acoustic waves rather than tsunami waves (Lotto et al, ).…”

Learning from crustal deformation associated with the M9 2011 Tohoku-oki earthquake

“…7C). For tsunami generation, it is the vertical motion of the seafloor relative to seawater that is important (Lotto et al, 2017). If the seafloor slope angle is α, and fault dip near the trench is β, the rise of the near-trench seafloor beneath a fixed water column due to a trench-breaching slip s is u = ssin(α + β) / cosα (Fig.…”

Meteorological Tsunami Generation Due to Sea‐Surface Pressure Change: Three‐Dimensional Theory and Synthetics of Ocean‐Bottom Pressure Change

“…This drastically changed tsunami research because the generation cannot be described by 2‐D equations that were often used in past tsunami studies. Three‐dimensional (3‐D) fluid theories play a fundamental role in the generation process (Lotto & Dunham, 2015; Lotto et al., 2017; Maeda et al., 2013; Takahashi 1942). Saito (2013) investigated the 3‐D tsunami generation due to sea‐bottom deformation, providing a theoretical basis for the initial tsunami height distribution for the 2‐D tsunami simulations.…”