Source Representation and Earthquake-Ocean Coupling in Tsunami Generation by the 2020 Samos Normal-Faulting Earthquake

The 2020 Mw 7.0 Samos normal-faulting earthquake produced damaging ground shaking and a destructive local tsunami. We construct a three-dimensional kinematic rupture model from a finite-fault inversion and use its shear stress changes to inform two dynamic rupture models. We initialize a non-linear shallow-water tsunami simulation with instantaneous Okada displacement and source one-way linked simulations with evolving seafloor deformation from the kinematic and dynamic rupture models. Fully coupled three-dimensional earthquake-tsunami simulations resolve rupture and elastic, acoustic, and gravity waves together. The coupled sea surface wavefield contains the developing tsunami, dispersive oceanic Rayleigh waves, first-mode ocean acoustic waves, and signals consistent with dispersive Scholte waves. We track crests and troughs over time as the respective means of the most extreme 0.1 percent of positive and negative sea surface height anomalies. For each time-dependent source, following transient seismo-acoustic waves, the one-way linked and fully coupled models agree within 0.10 m on both metrics from 50 to 360 s in water deeper than 100 m. At 6 min, all models share a leading drawdown with trough metrics near -0.5 m, whereas the instantaneous Okada source yields a crest metric of 0.35 m, compared with 0.5–0.6 m for time-dependent sources. Instantaneous sourcing, common in hazard assessment, may therefore underestimate tsunami crests from moderate normal-faulting earthquakes. The preferred dynamic rupture source captures the main tsunami periods at Syros, with larger tide gauge mismatches elsewhere. At Samos, full coupling resolves the early near-source seismo-acoustic wavefield, whereas time-dependent one-way linking provides an efficient estimate of the large-scale tsunami.