Van Daele, Maarten ; Vanneste, Kris ; Moernaut, Jasper ; Araya-Cornejo, Cristian ; Pille, Thomas ; Schmidt, Sabine ; Kempf, Philipp ; Meyer, Inka ; Cisternas, Marco

Talk presented at Seismological Society of America Annual Meeting 20019, Seattle, USA on 2019-04-26

Abstract: Subduction zone seismicity arises from megathrust, crustal and intraslab earthquakes, and understanding the recurrence patterns of each type is crucial for hazard assessments. Specifically intraslab earthquakes have recently proven to pose a significant hazard along subduction zones, as evidenced by the destructive September 2017 Chiapas Mw 8.2 (Mexico) earthquake and the recent November 30 2018 Mw 7.0 Anchorage (Alaska) earthquake. Interestingly, intraslab earthquakes have a higher-frequency source spectrum then megathrust earthquakes as a result of the higher stress drop for a given magnitude. Here we compare acceleration response spectra of recent and historical intraslab and megathrust earthquakes along a longitudinal profile in central Chile and show that this difference in source frequency spectrum increases when moving land inward (e.g., towards the Andes in South America). We further illustrate how these differ-ent response spectra provide an opportunity for lacustrine paleoseismology, as higher-frequency accelerations from intraslab earthquakes are hardly attenu-ated in rocks of a lake’s watershed, whereas lower-frequency accelerations from megathrust earthquakes are amplified in soft lake sediments. We indeed observe this in Lo Encañado Lake, located in the central Chilean Andes, where megathrust earthquakes merely trigger subaquatic (soft sediment) slope failures, while intraslab earthquakes also trigger subaerial rock slides/avalanches, resulting in an additional postseismic turbidite in the lake. We conclude that lakes may be the only geological archives that hold intraslab paleoseismic records, and that lakes in other subduction zones (e.g., Cascadia) may hold similar records. Combining this kind of lacustrine records with coastal and deep-marine paleoseismology will permit the construction of recurrence models for the different seismogenic sources, and to evaluate the temporal correlations among them, thereby improving hazard assessments.