Coupled, physics-based modeling reveals earthquake displacements are critical in generating the 2018 Palu, Sulawesi tsunami

The 2018 Mw 7.5 Palu, Sulawesi strike-slip earthquake in Indonesia was followed by an unexpected tsunami. Using a physics-based, coupled earthquake-tsunami model, we show that direct earthquake-induced uplift could have sourced the tsunami. The model combines 3D rupture dynamics, wave propagation, tsunami propagation and inundation. The earthquake scenario matches key observations, including the supershear rupture speed and satellite data. This dynamic forward-modeling approach to building a tsunami source is distinct from kinematic methods. A kinematic earthquake source is produced by data inversion and generates static ground displacements to source the tsunami. This solution is non-unique. A dynamic source is also non-unique, but permits constraints from time-dependent data and model input and output are mechanically consistent. Here, the stress state and fault conditions are tightly constrained by observations using a pre-established, efficient algorithm by Ulrich et al. (2019). This consists of simple static analyses based on Mohr-Coulomb theory of frictional failure combined with a few trial models. The remote stress regime applied in this earthquake scenario, reflecting regional transtensional strain, produces a combination of up to 6 m of left-lateral slip and 2 m of normal slip on the straight fault segment dipping 65 degrees beneath Palu Bay. The time-dependent seafloor displacements are translated into bathymetry perturbations with a mean vertical offset of 1.5 m across the submarine fault segment. This sources a model tsunami with wave amplitudes and periods that match those measured at the Pantoloan wave gauge and inundation that is consistent with field survey data. We conclude that a source related to earthquake displacements is probable and that landsliding may not have been the primary source of the tsunami. Current numerical and computational capabilities now allow physics-based modeling to complement traditional methods for use in rapid response.