Using tsunami deposits to validate inundation modeling at Sedanka Island: Revealing clues about great earthquakes in the Unalaska seismic gap.

Abstract:

The Unalaska seismic gap flanks the eastern part of the zone of aftershocks that followed the great Mw 8.6 Andreanof Islands earthquake in 1957, yet the historical and paleo seismicity in this part of the Aleutian subduction zone is poorly resolved. Inversions of seismic and tsunami waveforms indicate little or no megathrust slip near Unalaska in 1957, leading some to designate the region as a seismic gap. In the absence of a long seismic record, inundation modeling validated by geologic evidence for tsunami inundation provide constraints on rupture parameters of tsunamigenic earthquakes. Parameters such as magnitude and slip distribution have broad uncertainties, but are very important in furthering earthquake and tsunami hazard assessments. Field investigations at Stardust Bay on Sedanka Island (near the eastern end of Unalaska) identified a sequence of six paleotsunami deposits over the last 1770 years that inundated up to 18.5 m above sea level, the youngest deposit probably records the tsunami generated by the 1957 earthquake. Tsunami inundation scenarios, modeled in GeoCLAW (simulating tsunami generation, propagation, and inundation) and using published sources of the 1957 tsunami cannot account for the tsunami deposit runup observed at Stardust Bay. Inundation scenarios that used a variety of alternative hypothetical earthquake sources suggest that at least a part of the Unalaska gap must rupture to agree with paleotsunami deposits. Scenarios modeled using a non-uniform slip distribution on the fault consistent with updip strengthening had significantly higher runup compared to scenarios that used a uniform slip distribution. By modeling a variety of non-uniform slip-distributions on a fault representing the subduction zone spanning the Unalaska Gap, we are trying to determine the minimum amount of slip that results in the observed deposit runup. The slip distribution used for the preferred 1957 tsunami scenario [that best fits the 1957 deposit data] provides an estimate of the minimum average slip of the tsunami source. Coupling the inundation model with a forward sediment transport model might reveal relative tsunami magnitudes for each deposit observed in Stardust Bay.