scholarly journals Predicting Imminence of Analog Megathrust Earthquakes With Machine Learning: Implications for Monitoring Subduction Zones

2020 ◽  
Vol 47 (7) ◽  
Author(s):  
F. Corbi ◽  
J. Bedford ◽  
L. Sandri ◽  
F. Funiciello ◽  
A. Gualandi ◽  
...  
2020 ◽  
Vol 36 (3) ◽  
pp. 1271-1297
Author(s):  
Kenneth W. Campbell

In this article, I propose a method for estimating the magnitude [Formula: see text] at which subduction megathrust earthquakes are expected to exhibit a break in magnitude scaling of both seismic source dimensions and earthquake ground motions. The methodology is demonstrated by applying it to 79 global subduction zones defined in the literature, including Cascadia. Breakpoint magnitude is estimated from seismogenic interface widths, empirical source scaling relations, and aspect ratios of physically unbounded earthquake ruptures and their uncertainties. The concept stems from the well-established observation that source-dimension and ground motion scaling decreases for shallow continental (primarily strike-slip) earthquakes when rupture exceeds the seismogenic width of the fault. Although a scaling break for megathrust earthquakes is difficult to observe empirically, all of the instrumentally recorded historical [Formula: see text] mega-earthquakes have occurred on subduction zones with [Formula: see text] (8.1–8.9), consistent with an observed break in source scaling relations derived from these same events. The breakpoint magnitudes derived in this study can be used to constrain the magnitude at which the scaling of ground motion is expected to decrease in subduction ground motion prediction equations.


2021 ◽  
Author(s):  
Susan Bilek ◽  
Emily Morton

<p>Observations from recent great subduction zone earthquakes highlight the influence of spatial geologic heterogeneity on overall rupture characteristics, such as areas of high co-seismic slip, and resulting tsunami generation.  Defining the relevant spatial heterogeneity is thus important to understanding potential hazards associated with the megathrust. The more frequent, smaller magnitude earthquakes that commonly occur in subduction zones are often used to help delineate the spatial heterogeneity.  Here we provide an overview of several subduction zones, including Costa Rica, Mexico, and Cascadia, highlighting connections between the small earthquake source characteristics and rupture behavior of larger earthquakes.  Estimates of small earthquake locations and stress drop are presented in each location, utilizing data from coastal and/or ocean bottom seismic stations.  These seismicity characteristics are then compared with other geologic and geophysical parameters, such as upper and lower plate characteristics, geodetic locking, and asperity locations from past large earthquakes.  For example, in the Cascadia subduction zone, we find clusters of small earthquakes located in regions of previous seamount subduction, with variations in earthquake stress drop reflecting potentially disrupted upper plate material deformed as a seamount passed.  Other variations in earthquake location and stress drop can be correlated with observed geodetic locking variations. </p>


2019 ◽  
Vol 219 (1) ◽  
pp. 645-661 ◽  
Author(s):  
Hiroo Kanamori ◽  
Luis Rivera ◽  
Lingling Ye ◽  
Thorne Lay ◽  
Satoko Murotani ◽  
...  

SUMMARY We recently found the original Omori seismograms recorded at Hongo, Tokyo, of the 1922 Atacama, Chile, earthquake (MS = 8.3) in the historical seismogram archive of the Earthquake Research Institute (ERI) of the University of Tokyo. These recordings enable a quantitative investigation of long-period seismic radiation from the 1922 earthquake. We document and provide interpretation of these seismograms together with a few other seismograms from Mizusawa, Japan, Uppsala, Sweden, Strasbourg, France, Zi-ka-wei, China and De Bilt, Netherlands. The 1922 event is of significant historical interest concerning the cause of tsunami, discovery of G wave, and study of various seismic phase and first-motion data. Also, because of its spatial proximity to the 1943, 1995 and 2015 great earthquakes in Chile, the 1922 event provides useful information on similarity and variability of great earthquakes on a subduction-zone boundary. The 1922 source region, having previously ruptured in 1796 and 1819, is considered to have significant seismic hazard. The focus of this paper is to document the 1922 seismograms so that they can be used for further seismological studies on global subduction zones. Since the instrument constants of the Omori seismographs were only incompletely documented, we estimate them using the waveforms of the observed records, a calibration pulse recorded on the seismogram and the waveforms of better calibrated Uppsala Wiechert seismograms. Comparison of the Hongo Omori seismograms with those of the 1995 Antofagasta, Chile, earthquake (Mw = 8.0) and the 2015 Illapel, Chile, earthquake (Mw = 8.3) suggests that the 1922 event is similar to the 1995 and 2015 events in mechanism (i.e. on the plate boundary megathrust) and rupture characteristics (i.e. not a tsunami earthquake) with Mw = 8.6 ± 0.25. However, the initial fine scale rupture process varies significantly from event to event. The G1 and G2, and R1 and R2 of the 1922 event are comparable in amplitude, suggesting a bilateral rupture, which is uncommon for large megathrust earthquakes.


2020 ◽  
Author(s):  
Mario D'Acquisto ◽  
Matthew Herman ◽  
Rob Govers

<div> <p>During and after a large megathrust earthquake, the overriding plate above the rupture zone moves oceanward. Enigmatically, the post-seismic motion of the overriding plate after several recent large earthquakes, further along strike from the rupture zone, was faster in the landward direction than before the event. Previous studies interpreted these changes as the result of increased mechanical coupling along the megathrust interface, transient slab acceleration, or bulk postseismic deformation with elastic bending mentioned as a possible underlying mechanism. Before invoking additional mechanisms, it is important to understand the contribution of postseismic deformation processes that are inherent features of megathrust earthquakes. We thus aim to quantify and analyse the deformation that produces landward motion during afterslip and viscous relaxation. </p> </div><div> <p>We use velocity-driven 3D mechanical finite element models, in which large megathrust earthquakes occur periodically on the finite plate interface. The model geometry is similar to most present-day subduction zones, but does not exactly match any specific subduction zone. </p> </div><div> <p>The results show increased post-seismic landward motion at (trench-parallel) distances greater than 450 km from the middle of the ruptured asperity. Similar patterns of landward motion are generated by viscous relaxation in the mantle wedge and by deep afterslip on the shear zone downdip of the brittle megathrust interface. Landward displacement due to postseismic relaxation largely accumulates at exponentially decaying rates until ~6 Maxwell relaxation times after the earthquake. The spatial distribution and magnitude of the velocity changes is broadly consistent with observations related to both the 2010 Maule and the 2011 Tohoku-oki earthquakes.  </p> </div><div> <p>Further model experiments show that patterns of landward motion due to afterslip and to viscous relaxation are insensitive to the locking pattern of the megathrust. However, the locking distribution does affect the magnitudes of the displacements and velocities. Results show that the increased landward displacement due to postseismic deformation scales directly proportionally to seismic moment. </p> </div><div> <p>We conclude that the landward motion results from in-plane horizontal bending of the overriding plate and mantle. This bending is an elastic response to oceanward tractions near the base of the plate around the ruptured asperity, causing extension locally and compression further away along-trench. This elastic in-plate bending consistently contributes to earthquake-associated changes in surface velocities for the biggest megathrust earthquakes, producing landward motion along strike from the rupture zone.</p> </div>


2020 ◽  
Author(s):  
Pousali Mukherjee ◽  
Yoshihiro Ito ◽  
Emmanuel S. Garcia ◽  
Raymundo Plata-Martinez ◽  
Takuo Shibutani

<p>Subduction zones host some of the greatest megathrust earthquakes in the world. Slow earthquakes have been discovered around the subduction zones of the Pacific rim very close to megathrust earthquakes. Investigating the lithosphere of the slow earthquake area versus non slow-earthquake area in subduction zones is crucial in understanding the role of the internal structure to control slow earthquakes. In this study, we investigate the lithospheric structure of stations in the slow earthquake area and non slow-earthquake areas in Chile using receiver function analysis and inversion method using teleseismic earthquakes. Here we focus on, especially the Vp/Vs ratios from both slow and non-slow earthquake areas, because the Vp/Vs ratio is sensitive to the fluid distribution in the lithosphere; the fluid distribution possibly controls the potential occurrence of slow earthquakes. Additionally, the nature of the slab can also play a crucial factor. The Vp/Vs ratio results across depth shows significantly higher value in the deeper oceanic slab region beneath the stations in the slow earthquake areas with higher contrast at the boundary.</p>


2021 ◽  
Author(s):  
Lise Alalouf ◽  
Yajing Liu

<p>Subduction zones are where the largest earthquakes occur. In the past few decades, scientists have also discovered the presence of episodic aseismic slip, including slow slip events (SSEs), along most of the subduction zones. However, it is still unclear how these SSEs can influence megathrust earthquake ruptures. The Costa Rica subduction zone is a particularly interesting area because a SSE was recorded 6 months before the 2012 Mw7.6 earthquake in the Nicoya Peninsula, suggesting a potential stress transfer from the SSE to the earthquake slip zone. SSEs beneath the Nicoya Peninsula were also recorded both updip and downdip the seismogenic zone, making it a unique area to study the complex interaction between SSEs and earthquakes.</p><p>As most of the shallow SSEs were recorded around the Nicoya Peninsula, we chose to start using a 1D planar fault embedded in a homogeneous elastic half-space, with different dipping angles following several geometric profiles of the subduction fault beneath the Nicoya Peninsula section of the Costa Rica margin. This 1D modelling study allows us to better investigate the interaction between shallow and deep SSEs and megathrust earthquakes with high numerical resolution and relatively short computation time. The model provides information on the long-term seismic history by reproducing the different stages of the seismic cycle (interseismic slip, shallow and deep episodic slow slip, and coseismic slip).</p><p>We study the influence of the variation of numerical parameters and frictional properties on the recurrence interval, maximum slip velocity and cumulative slip of SSEs (both shallow and deep) and earthquakes and their interaction with each other. We then compare our results with GPS and seismic observations (i.e. cumulative slip, characteristic duration, moment rate, depth and size of the rupture, equivalent magnitude) to identify an optimal set of model parameters to understand the interaction between various modes of subduction fault deformation.</p>


Author(s):  
Francisco Acuña ◽  
Gonzalo A. Montalva ◽  
Daniel Melnick

Abstract Time-dependent earthquake forecast depends on the frequency and number of past events and time since the last event. Unfortunately, only a few past events are historically documented along subduction zones where forecasting relies mostly on paleoseismic catalogs. We address the role of dating uncertainty and completeness of paleoseismic catalogs on probabilistic estimates of forthcoming earthquakes using a 3.6-ka-long catalog including 11 paleoseismic and 1 historic (Mw≥8.6) earthquakes that preceded the great 1960 Chile earthquake. We set the clock to 1940 and estimate the conditional probability of a future event using five different recurrence models. We find that the Weibull model predicts the highest forecasting probabilities of 44% and 72% in the next 50 and 100 yr, respectively. Uncertainties in earthquake chronologies due to missing events and dating uncertainties may produce changes in forecast probabilities of up to 50%. Our study provides a framework to use paleoseismic records in seismic hazard assessments including epistemic uncertainties.


2020 ◽  
Author(s):  
Daniel Melnick ◽  
Valentina Maldonado ◽  
Martin Contreras ◽  
Julius Jara-Muñoz ◽  
Joaquín Cortés-Aranda ◽  
...  

<p>Most of the seismic hazard along subduction zones is posed by great tsunamigenic earthquakes associated with the interplate megathrust fault. However, crustal faults are ubiquitous along overriding continental plates, some of which have been triggered during recent megathrust earthquakes. In Chile, the 2010 Maule earthquake (M8.8) triggered a shallow M7 earthquake on the Pichilemu fault, which had not been mapped and was unknown. In fact, M~7 earthquakes have recently occurred along unknown faults in California and New Zealand, emphasizing the need for better and more detailed mapping initiatives. A first step towards a synoptic assessment of seismic hazards posed by continental faults at the national level is mapping at a homogeneous scale to allow for a systematic comparison of faults and fault systems. Here, we present the first map of active and potentially-active faults in Chile at 1:25,000 scale, which includes published studies and newly-identified faults. All the published faults have been re-mapped using LiDAR and TanDEM-X topography, where available. Using different scaling relations, we estimate the seismic potential of all crustal faults in Chile. For specific faults where we have conducted paleoseismic and tectonic geomorphic field studies (e.g., Liquiñe-Ofqui, El Yolki, Mesamavida, and Pichilemu faults) we provide new estimates of slip rate, recurrence interval, and deformation style. We propose a segmentation model of continental faults systems in Chile, which are associated with distinct morphotectonic units and have predominant kinematics and relatively uniform slip rates. Using stress transfer models, we explore the potential feedbacks between upper-plate deformation and the megathrust seismic cycle.</p>


2020 ◽  
Author(s):  
Claudio Petrini ◽  
Luca Dal Zilio ◽  
Taras Gerya

<p>Slow slip events (SSEs) are part of a spectrum of aseismic processes that relieve tectonic stress on faults. Their occurrence in subduction zones have been suggested to trigger megathrust earthquakes due to perturbations in fluid pressure. However, examples to date have been poorly recorded and physical observations of temporal fluid pressure fluctuations through slow slip cycles remain elusive. Here, we use a newly developed two-phase flow numerical model — which couples solid rock deformation and pervasive fluid flow — to show how crustal stresses and fluid pressures within subducting megathrust evolve before and during slow slip and regular events. This unified 2D numerical framework couples inertial mechanical deformation and fluid flow by using finite difference methods, marker-in-cell technique, and poro-visco-elasto-plastic rheologies. Furthermore, an adaptive time stepping allows the correct resolution of both long- and short-time scales, ranging from years to milliseconds during the dynamic propagation of earthquake rupture.</p><p>Here we show how permeability and its spatial distribution control the degree of locking along the megathrust interface and the interplay between seismic and aseismic slip. While a constant permeability leads to more regular seismic cycles, a depth dependent permeability contributes substantially to the development of two distinct megathrust zones: a shallow, locked seismogenic zone and a deep, narrow aseismic segment characterized by SSEs. Furthermore, we show that without requiring any specific friction law, our model shows that permeability, episodic stress transfer and fluid pressure cycling control the predominant slip mode along the subduction megathrust. Specifically, we find that the up-dip propagation of episodic SSEs systematically decreases the fault strength due to a continuous accumulation and release of fluid pressure within overpressured subducting interface, thus affecting the timing of large megathrust earthquakes. These results contribute to improve our understanding of the physical driving forces underlying the interplay between seismic and aseismic slip, and demonstrate that slow slip events may prove useful for short-term earthquake forecasts.</p>


2020 ◽  
Author(s):  
Mohammad Yuzariyadi ◽  
Kosuke Heki

<p>Enhanced interplate coupling has been found for segments adjacent along-strike to megathrust faults after the 2003 Tokachi-Oki and the 2011 Tohoku-Oki earthquakes, NE Japan, and was interpreted as acceleration of the subducting Pacific Plate slab. A similar enhanced coupling was also reported for the segments to the north of the rupture area of the 2010 Maule earthquake, central Chile. We utilize available GNSS data to find such enhanced coupling in worldwide subduction zones including NE Japan, central and northern Chile, Sumatra, and Mexico to investigate their common features. Our study revealed that the accelerations of landward movement of 2.1-9.0 mm per year appeared in adjacent segments following the 2014 Iquique (Chile), the 2007 Bengkulu (Sumatra), and the 2012 Oaxaca (Mexico) earthquakes. We also confirmed that the enhanced coupling is associated with the increase of seismicity for all these six cases. We found that the degree of enhancement depends on the length of the slab and the magnitude of the earthquake, which is consistent with the simple 2-dimensional model proposed earlier.</p>


Sign in / Sign up

Export Citation Format

Share Document