Recent geophysical and geodetic observations suggest that megathurst faults are frictionally segmented at depth, a condition that seems to allow the occurrence of supercycles of giant tsunamigenic earthquakes. Ignoring regions of previous seismic energy release, giant earthquakes seem capable of propagating from the deeper plate interface all the way to trench, producing both strong shaking and devastating tsunamis. A remarkable example was dramatically revealed by the 2011 Mw 9.0 Tohoku earthquake, which besides rupturing the lower part of the seismogenic zone as anticipated by recurrent Mw 7-8 events, propagated onto the previously thought aseismic shallow zone, even reaching the trench with up to 50 m of highly tsunamigenic slip. Indeed, the downdip frictional segmentation of Tohoku allowed the 2011 Japan event to grow to such a giant tsunamigenic earthquake. This story highlights the importance of depth segmentation in seismic and tsunami hazards assessments. Here we propose a similar behavior for the metropolitan Chilean megathrust, enclosed between the rupture zones of the 2010 and 2015 earthquakes. In this region, large earthquakes (Mw > 8) have historically occurred with a remarkable temporal regularity, in 1730, 1822, 1906 and 1985; however, no doubt remains that the 1730 event was by far the largest of the sequence and the only one capable of producing a large tsunami. The rest of the earthquakes have instead been better characterized by coastal uplift than by their tsunamis. Uplift of the coastal region around Valparaíso was systematically observed in the earthquakes of 1822, 1906 and 1985, but surprisingly not in the largest earthquake of 1730. But how could such a giant earthquake not produced notable coastal uplift in regions where the following much smaller earthquakes did? By combining abundant, recently found historical records with rupture and tsunami modeling we show that, as the Tohoku story, these observations can be explained by a dual seismogenic behavior, in which the 1822, 1906 and 1985 events ruptured downdip segments of the large 1730 earthquake rupture zone. Because little or no shallow slip has been released in over 288 years, surpassing the minimum recurrence for large tsunamis in the region as revealed by paleotsunami studies, our results imply increased tsunami hazards for the most populated coast of Chile. This study is supported by Millennium Nucleus CYCLO: The Seismic Cycle along Subduction Zones and Fondecyt N°1150321.
Fault segmentation across the metropolitan Chilean megathrust: Large tsunamigenic potential for the next great earthquake
