Frequency dispersion amplifies tsunamis caused by outer-rise normal faults

Although tsunamis are dispersive water waves, hazard maps for earthquake-generated tsunamis neglect dispersive effects because the spatial dimensions of tsunamis are much greater than the water depth, and dispersive effects are generally small. Furthermore, calculations that include non-dispersive effects tend to predict higher tsunamis than ones that include dispersive effects. Although non-dispersive models may overestimate the tsunami height, this conservative approach is acceptable in disaster management, where the goal is to save lives and protect property. However, we demonstrate that offshore frequency dispersion amplifies tsunamis caused by outer-rise earthquakes, which displace the ocean bottom downward in a narrow area, generating a dispersive short-wavelength and pulling-dominant (water withdrawn) tsunami. We compared observational evidence and calculations of tsunami for a 1933 Mw 8.3 outer-rise earthquake along the Japan Trench. Dispersive (Boussinesq) calculations predicted significant frequency dispersion in the 1933 tsunami. The dispersive tsunami deformation offshore produced tsunami inundation heights that were about 10% larger than those predicted by non-dispersive (long-wave) calculations. The dispersive tsunami calculations simulated the observed tsunami inundation heights better than did the non-dispersive tsunami calculations. Contrary to conventional practice, we conclude that dispersive calculations are essential when preparing deterministic hazard maps for outer-rise tsunamis.

Seismicity around the trench axis and outer-rise region of the southern Japan Trench, south of the main rupture area of the 2011 Tohoku-oki earthquake

SUMMARY The 2011 Mw 9.0 Tohoku-oki earthquake ruptured the subduction megathrust fault in the central Japan Trench. We investigated the aftershock activity in the southern Japan Trench to the south of the main rupture area using ocean bottom seismographs deployed both landward and seaward of the trench. In the trench-outer rise region seaward of the trench axis, we identified several ∼100-km-long linear earthquake trends both parallel and oblique to the southern Japan Trench. The earthquake trend oblique to the southern Japan Trench is a southward extension of the trench-parallel linear earthquake trend in the central to northern Japan Trench. The trench-parallel normal-faults in the trench-outer rise region could extend linearly, despite the change of the trench strike from N–S to NNE–SSW to the south of the main rupture area. Normal-faults oblique to the trench should be considered as substantial parts of large intraplate normal-faulting earthquakes. In addition, intraplate seismicity coinciding with the lower velocity oceanic mantle suggest that the structure heterogeneity would be indicative of normal-faults extending into the mantle. In the trench landward area, earthquake activity showed along-trench variations. Earthquakes along the shallow megathrust interface near the trench were observed south of 37°N. These shallow near-trench regular earthquakes, which are located close to the episodic tremors and temporally correlated with the tremor activities, suggest that the afterslip on the plate interface likely extended to the shallow plate interface close to the trench axis. Smaller spatial scale structure heterogeneity, such as the thickness variation in the channel-like low-velocity sedimentary unit, likely relate to the proximity of the regular earthquakes and slow slip which results in the formation of diverse slip behaviours in the shallow subduction zone of the southern Japan Trench.

The Mw7.1 September 19th Puebla-Morelos (Mexico) earthquake triggered by brucite and antigorite dewatering

<p>Subduction controls key geological processes at convergent margins including seismicity and resultant seismic hazard. The September 19th 2017 Mw7.1 Mexican earthquake nucleated ~250 km from the trench within the Cocos plate near its Moho, ~57 km below Earth’s surface. The prevailing hypothesis suggests that this earthquake resulted from bending stresses occurring at the flat-to-steep subduction transition. Here, we present an alternative, but not mutually exclusive, hypothesis: the dehydration reaction brucite + antigorite = olivine + H2O in the slab mantle controls intermediate-depth seismicity along the flat portion of the subducted Cocos plate. This reaction releases a substantial amount of H2O, resulting in a large positive volume change, and thus in an increase in pore fluid pressure at the appropriate depth–temperature conditions to cause the Puebla-Morelos and other intraslab earthquakes in Mexico. The amount of H2O released by this reaction depends on the degree of serpentinization of the oceanic mantle prior to subduction. Only oceanic mantle with > 60% serpentinization—as expected along abundant deep extensional faults at the mid-ocean-ridge or where the plate bends at the outer rise—will stabilize brucite, and thus, will experience this reaction at the same depths where the September 19th 2017 earthquake nucleated.</p>

The March 25, 2020 MW 7.5 Paramushir earthquake

The strong earthquake with moment magnitude Mw = 7.5 occurred on March 25, 2020, in the North Kurils to the southeast of the Paramushir Island. The hypocenter of the earthquake was located under the oceanic rise of deep-sea trench in the subducting Pacific lithospheric plate. This earthquake has been the strongest seismic event since 1900 for an area about 800 km long of the outer rise of the trench. It also was the strongest earthquake for the 300-kilometer long area of the Kuril-Kamchatka subduction zone adjacent to the epicenter. The article summarizes the data on the Paramushir earthquake. Tectonic position of the earthquake, source parameters, features of the aftershock process development, as well as coseismic displacement of the nearest continuous GNSS station are considered. The performed analysis did not allow us to clearly determine the rupture plane in the source. Nevertheless, the study of the features of the outer-rise earthquake is a matter of scientific interest, since the stress state of the bending area of the subducting Pacific lithospheric plate reflects the interplate interaction in the subduction zone.

Dynamic slab segmentation due to brittle-ductile damage interactions in the outer rise

The recycling of oceanic plates by means of subduction represents the major plate driving force and subducting plate strength controls many aspects of the thermo-chemical evolution of Earth. Regardless of its prior history, each subducting plate experiences intense normal faulting1-11 during bending that accommodates the transition from horizontal to downward motion at the outer rise at subduction trenches. Here, we investigate the consequences of this bending-induced plate damage using new numerical, thermomechanical subduction models in which both brittle and ductile deformation, as well as grain size evolution, are tracked and coupled self-consistently. Pervasive slab weakening and pronounced segmentation can occur at the outer rise region due to the strong feedback between brittle and ductile damage localization. The “memory” of bending varies from segmentation to broadly-distributed damage depending on the age of the subducting plate, mantle potential temperature, and the magnitude of strain-induced weakening of outer rise normal faults. This new slab damage phenomenon explains the development of large-offset normal faults8,9, the occurrence of deep compressional thrust-faulting earthquakes12, and the appearance of localized areas of reduced effective viscosity13 observed at subduction trenches. Furthermore, brittle-viscously damaged slabs show a strong tendency for slab breakoff at elevated mantle temperatures. Given Earth’s planetary cooling history14, this implies that intermittent subduction with frequent slab breakoff episodes15,16 may have been characteristic for terrestrial plate tectonics until more recent times than expected from memory-free rheologies17.

Scaling relations of seismic moment and rupture area for outer-rise earthquakes

A new piece-wise scaling relation for rupture area ( S ) and seismic moment ( M 0 ) was obtained for outer-rise earthquakes using data compiled from previous seismological studies. The compiled source parameters were examined carefully to ensure that rupture area dimensions represent the "actual" rupture area. Twenty sets of S and M 0 pairs for ten earthquakes from independent studies were used for regression analysis. To estimate rupture areas of outer-rise earthquakes whose fault down dip tips, respectively, do and do not reach the oceanic lithosphere base, regression parameter b in S=(M 0 /a) 1/b is 2 with the corresponding regression parameter log 10 a 2 being 1.402 (±0.250), and b is 1.5 with the corresponding regression parameter log 10 a 1 being 6.401 (±0.187). The piece-wise scaling relation developed here can fit outer-rise earthquake data from Mw 7.2 to Mw 8.4 and can be used to estimate rupture areas of earthquakes greater than Mw 8.4.