Revisiting earthquake-tsunami models of the 2018 Palu events using near-fault high-resolution imaging and 3D fully-coupled earthquake-tsunami modeling

The 2018 Mw 7.5 Palu earthquake struck the Sulawesi island, Indonesia, in 2018 and was followed by an unexpected tsunami. Using a physics-based, coupled earthquake-tsunami model, Ulrich et al. (2019) showed that direct earthquake-induced uplift could have sourced the tsunami. The 3D dynamic rupture model of the earthquake captures key observations, including the supershear rupture speed and the deformation pattern derived from satellite data. Stress state and fault conditions were tightly constrained by observations combined with simple static analyses based on Mohr-Coulomb theory of frictional failure and a few trial models. The earthquake scenario predicts a combination of up to 6 m of left-lateral slip and 2 m of normal slip on a straight fault segment dipping 65 degrees beneath Palu Bay.

While most studies (e.g. Bai et al., 2018, Ulrich et al., 2019, Oral et al., 2019) suggest a very early supershear transition, the exact timing of the onset of supershear rupture and the driving mechanism of the supershear transition are elusive. Here we revisit the earthquake dynamic rupture modeling based on new high-resolution near-fault deformation maps derived from correlation of optical satellite data. We vary nucleation radius, fault geometry, and off-fault plasticity parametrization to obtain alternative dynamic rupture scenarios. Specific inputs allow delayed transition to supershear. The obtained scenarios are evaluated based on near-fault damage inference.

Additionally, we revisit the tsunami model, adopting advanced strategies for earthquake-tsunami linking and tsunami modeling. In Ulrich et al. (2019), a one-way linking approach with a shallow water equations solver allowed translating the time-dependent seafloor displacements into a tsunami model with wave amplitudes and periods matching those measured at the Pantoloan wave gauge and inundation that is consistent with field survey data. Such modeling workflow yet neglects tsunami generation complexity, acoustic waves, and dispersion, and only approximates horizontal momentum transfer.  We present a 3D fully coupled earthquake-tsunami model (Krenz et al., 2021), that releases these limitations. This allows us to assess how the standard earthquake-tsunami workflow affects our results, and to revisit our conclusions.