Aftershock Generation Research Articles

Evaluation of the b Maps on the Faults of the Major (M > 7) South California Earthquakes

AbstractWe use the Godano et al. (2022, https://doi.org/10.1029/2021ea002205) method for evaluating the b maps of the faults associated with the largest earthquakes M ≥ 7.0 that occurred in California. The method allows an independent evaluation of the b parameter, avoiding the overlap of the cells and the omission of some earthquakes, while keeping all the available information in the catalog. We analyzed four large earthquakes: Landers, Hector Mine, Baja California, and Searles Valley. The maps obtained confirm that the b value can be considered as a strain meter and allow us to elucidate the presence of barriers, such as obstacles to the propagation of the fracture, on the fault of the analyzed earthquakes. A further estimated parameter is the time window during which aftershocks occur in the cell, Δt. This quantity is very useful for a better definition of the aftershock generation mechanism. It reveals where the stress is released in a short time interval and how the complexity of the faulting process controls the occurrence of aftershocks on the fault, and also the duration of the entire sequence.

Aftershocks following the 2011 Tohoku-Oki earthquake driven by both stress transfer and afterslip

Aftershocks are a fundamental characteristic of seismicity, and their generation mechanism is mainly characterized by two physical models, stress transfer from large earthquakes and afterslip-induced stress loading. However, the contribution of each mechanism to aftershock generation remains unclear. Here we investigate the spatiotemporal variations in aftershock activity following the 2011 Tohoku-Oki earthquake by applying the Hierarchical Space–Time Epidemic-Type Aftershock Sequence (HIST-ETAS) model to the decade of recorded seismicity since the mainshock. Using the estimated HIST-ETAS model, we categorize the aftershocks into background earthquakes (which are caused by aseismic phenomena) and triggered earthquakes (which are caused by earthquake-to-earthquake interactions). Most of the earthquakes that occurred updip of the large coseismic slip zone along the Japan Trench are triggered earthquakes, consistent with the lack of afterslip in this area. Conversely, background earthquakes are the predominant earthquake type in the long-term downdip of the large coseismic slip zone, and they positively correlate with the afterslip evolution. Our results suggest the importance of combining these two end-member aftershock generation models to explain aftershock activity and thus provide new insights into the relationship between afterslip and spatiotemporal aftershock distribution. Our classifications may also contribute to the monitoring of afterslip in a given region that hosts a large earthquake, particularly where geodetic observation networks are too sparse to evaluate afterslip evolution.

Fault Roughness Promotes Earthquake‐Like Aftershock Clustering in the Lab

AbstractEarthquakes rarely occur in isolation but rather as complex sequences of fore, main and aftershocks. Assessing the associated seismic hazard requires a holistic view of event interactions. We conduct frictional sliding experiments on faulted Westerly Granite samples at mid‐crustal stresses to investigate fault damage and roughness effects on aftershock generation. Abrupt laboratory fault slip is followed by periods of extended stress relaxation and aftershocks. Large roughness promotes less co‐seismic slip and high aftershock activity whereas smooth faults promote high co‐seismic slip with few aftershocks. Conditions close to slip instability generate lab‐quake sequences that exhibit similar statistical distributions to natural earthquakes. Aftershock productivity in the lab is linearly related to the residual strain energy on the fault which, in turn, is controlled by the level of surface heterogeneity. We conclude that roughness and damage govern slip stability and seismic energy partitioning between fore, main and aftershocks in lab and nature.

Relative Afterslip Moment Does Not Correlate With Aftershock Productivity: Implications for the Relationship Between Afterslip and Aftershocks

AbstractAseismic afterslip has been proposed to drive aftershock sequences. Both afterslip moment and aftershock number broadly increase with mainshock size, but can vary beyond this scaling. We examine whether relative afterslip moment (afterslip moment/mainshock moment) correlates with several key aftershock sequence characteristics, including aftershock number and cumulative moment (both absolute and relative to mainshock size), seismicity rate change, b‐value, and Omori decay exponent. We select Mw ≥ 4.5 aftershocks for 41 tectonically varied mainshocks with available afterslip models. Against expectation, relative afterslip moment does not correlate with tested aftershock characteristics or background seismicity rate. Furthermore, adding afterslip moment to mainshock moment does not improve predictions of aftershock number. Our findings place useful empirical constraints on the link between afterslip and potentially damaging Mw ≥ 4.5 aftershocks and raise questions regarding the role afterslip plays in aftershock generation.

Sheared granular matter and the empirical relations of seismicity.

The frictional instability associated with earthquake initiation and earthquake dynamics is believed to be mainly controlled by the dynamics of fragmented rocks within the fault gauge. Principal features of the emerging seismicity (e.g., intermittent dynamics and broad time and/or energy scales) have been replicated by simple experimental setups, which involve a slowly driven slider on top of granular matter, for example. Yet these setups are often physically limited and might not allow one to determine the underlying nature of specific features and, hence, the universality and generality of the experimental observations. Here, we address this challenge by a numerical study of a spring-slider experiment based on two-dimensional discrete element method simulations, which allows us to control the properties of the granular matter and of the surface of the slider, for example. Upon quasistatic loading, stick-slip-type behavior emerges which is contrasted by a stable sliding regime at finite driving rates, in agreement with experimental observations. Across large parameter ranges for damping, interparticle friction, particle polydispersity, etc., the earthquake-like dynamics associated with the former regime results in several robust scale-free statistical features also observed in experiments. At first sight, these closely resemble the main empirical relations of tectonic seismicity at geological scales. This includes the Gutenberg-Richter distribution of event sizes, the Omori-Utsu-type decay of aftershock rates, as well as the aftershock productivity relation and broad recurrence time distributions. Yet, we show that the correlations associated with tectonic aftershocks are absent such that the origin of the Omori-Utsu relation, the aftershock productivity relation, and Båth's relation in the simulations is fundamentally different from the case of tectonic seismicity. This, we believe, is mainly due to a lack of macroscale relaxation processes that are closely tied to the generation of real aftershocks. We argue that the same is true for previous laboratory experiments.

Bayesian Inference of Centroid Moment Tensors of the 2019 Ambon (Mw 6.5) Aftershock Earthquake Sequence, Indonesia: A Preliminary Result

The Ambon Mw 6.5 earthquake on September 26th, 2019, had contributed to give severe damages and significantly increased seismicity around Ambon Island and surrounding areas. Mainshock was followed by aftershocks with spatial distribution added to the impact of destructions in this region. We investigated aftershocks sequences to reveal the effect of mainshock toward the change in the in-situ stress field, including the possibility of the existing faults reactivation and the generation of aftershocks. We inferred centroid moment tensor (CMT) for significant aftershock events with Mw more than 4.0 using waveform data recorded from October 18th to December 15th, 2019. The aftershock focal mechanism was determined using the Bayesian full-waveform inversion code ISOLA-Obspy. This approach provides the uncertainty of the CMT model parameters. From ten CMT solution we had inferred in three seismic clusters, we found that majority of events have a strike-slip mechanism. Four events located on the south of the N-S trendings have a dextral strike-slip fault type, reflected the rupture of the mainshocks fault plane. Three events in the cluster of Ambon Island are dextral strike-slip, confirming the presence of the fault reactivation. Meanwhile, three CMT solutions in the north show the dextral strike-slip faulting and may belong to the mainshock main fault, connected with the cluster in the south.

Decomposition of elastic potential energy and a rational metric for aftershock generation

SUMMARY The occurrence of earthquakes can be regarded as shear fracture releasing the elastic potential energy stored in the Earth. The potential energy density of linear elastic forces is generally represented in the quadratic form of strain tensor components with the fourth-order coefficient tensor of elastic stiffness. When the material is isotropic, since the stiffness tensor is expressible as a linear combination of two independent symmetric tensors, we can decompose the elastic potential energy density into two independent parts, namely the volumetric part and the shearing part. By definition, the partial derivatives of the elastic potential energy density with respect to volumetric and shearing deformations give the corresponding generalized forces in the sense of Lagrangian mechanics: specifically, one-third of the first invariant of stress tensor (equivalent to the mean stress) to volumetric deformation and the square root of the second invariant of deviatoric stress tensor (equivalent to $\sqrt {3/2} $ times the octahedral shear stress). With these generalized forces instead of the normal and tangential stresses on a specific fault plane, we correctly represented the original concept of Coulomb's failure criterion (shear failure occurs when shearing stress is equal to shearing strength) and defined energetics-based failure stress (EFS). The change in EFS associated with the occurrence of a main fracture (ΔEFS) gives a rational metric for aftershock generation, which can be reduced to previously proposed various metrics in special cases. For example, when the level of background deviatoric stress is much higher than the magnitude of coseismic stress changes, the expression of ΔEFS is reduced to a similar form to the well-known Coulomb failure stress change (ΔCFS). Even in the energetics-based metric, the effects of pore-fluid pressure changes are essential. We theoretically examined the mechanical effects of induced and enforced pore-fluid pressure changes and elucidated that the difference between them is reflected in the focal mechanisms of aftershocks.

Seismic Attenuation Tomography From 2018 Lombok Earthquakes, Indonesia

Local earthquake data was used to determine a three-dimensional (3D) seismic attenuation structure around the aftershock source region of the 2018 Lombok earthquake in Indonesia. The aftershocks were recorded by 13 seismic stations from August 4 to September 9, 2018. The selected data consist of 6,281 P-wave t∗ values from 914 events, which had good t∗ quality in at least four stations. Our results show that the two aftershock clusters northwest and northeast of Lombok Island have different attenuation characteristics. A low P-wave quality factor (low-Qp), low P-wave velocity (Vp), and high ratio of P-wave velocity and S-wave velocity (Vp/Vs), which coincide with a shallower earthquake (<20 km) northwest of Lombok Island, might be associated with a brittle area of basal and imbricated faults influenced by high fluid content. At the same time, the high-Qp, low Vp, and low Vp/Vs, which coincide with a deeper earthquake (>20 km) northeast of Lombok Island, might be associated with an area that lacks fluid content. The difference in fluid content between the northwest and northeast regions might be the cause of the early generation of aftershocks in the northwest area. The significant earthquake that happened on August 5, 2018, took place in a region with moderate Qp, close to the contrast of high and low-Qp and high Vp, which suggests that the earthquake started in a strong material before triggering the shallower aftershocks occurring in an area affected by fluid content. We also identified an old intrusive body on the northeast flank of the Rinjani volcano, which was characterized by a high-Qp, high-velocity, and a high Bouguer anomaly.

Stress and pore fluid pressure control of seismicity rate changes following the 2016 Kumamoto earthquake, Japan

We quantitatively examined the influence of pore fluid pressure and coseismic stress changes on the seismicity rate changes that followed the 2016 Kumamoto earthquake, on the basis of two approaches. One is a numerical calculation of the classic stress metric of ∆CFS, and the other is an inversion analysis of pore fluid pressure fields with earthquake focal mechanism data. The former calculation demonstrated that seismicity rate changes were consistent with the expectation from ∆CFS in 65% of the target region, whereas they were not in the remaining 35% of the region. The latter analysis indicates that seismicity rates increased in the regions where pore fluid pressure before the Kumamoto earthquake sequence was remarkably enhanced above hydrostatic, regardless of values of ΔCFS. This suggests that the increase in pore fluid pressure is one of the important physical mechanisms triggering aftershock generation. We obtained evidence that pore fluid pressure increased around the southern part of the main rupture zone after the mainshock, examining temporal changes in types of focal mechanism data. The average increases in pore fluid pressure were estimated to be 17, 20, and 17 MPa at depths of 5, 10, and 15 km, respectively. These large increases in pore fluid pressure cannot be explained under the undrained condition. The spatial derivative of the pore fluid pressure field in the depth direction implies that fluid supply from greater depths may have controlled increases in seismicity rates that followed the large earthquake.

On the aftershock productivity in mining-induced seismicity—insight into seismicity of Rudna copper ore mine, Poland

SUMMARYMining-induced seismic events can be followed by aftershocks that increase the risk associated with the exploitation. The understanding of the aftershock generation process in induced seismicity may improve post-earthquake safety procedures applied in mines. Rudna copper ore mine in southwestern Poland commonly experiences intense and strong seismic activity accompanying the room-and-pillar exploitation of copper ore. Some strong (magnitude >2) mining events are followed by numerous aftershocks and some are not followed by any. In this study, we seek to find whether there is any geological, technological or seismological cause of this diversity. We study 46 strong mining events and focus on their aftershock productivity. We analyse the geological and mining setting of the studied events, their signal similarity, stress drops and the ground motion effect using data from three different seismic networks. Our results show that seismic events producing large aftershock sequences may share similar focal mechanisms and have larger ground effects than events with no aftershocks. The results also indicate the potential differences in stress drops. This interesting observation may help to better evaluate the aftershock hazard in mines. It also indicates the need for a more detailed analysis of the focal mechanisms of strong events and their relationship to the exploitation technique.

Classes of sustained microfracturing produced after formation of hydraulic fractures

Aftershocks were observed to occur after hydraulic fracturing tests on mortar blocks at the laboratory scale. Five aftershocks sequences were recorded and fitted with the original and modified Omori laws. The modified Omori law was found to be a better model for the aftershock sequences studied. Both the estimated decay exponent p and the time shift c fall in the range found from observations of the geological aftershocks. The statistical similarity between the geological and the laboratory aftershocks suggests that there is a common underlying mechanism governing the generation of aftershocks across the scale. In this mechanism the aftershocks correspond to crack production by the mechanism of static fatigue. The power law decay of aftershock frequency (modified Omori law) is explained by the residual stress relaxation caused by the reduction of deformation moduli produced by crack accumulation. Two types of decay exponents p (exponents of the modified Omori's law) were observed: (1) p > 1 caused by static fatigue fracturing of initially non-fractured zones (bridges) within the fracture; (2) p < 1 caused by static fatigue production of parallel cracks outside the fracture surface.

Elastic strain energy and pore-fluid pressure control of aftershocks

Aftershocks are well-known, however their triggering mechanisms remain unclear. The Coulomb failure stress change (ΔCFS) has been widely implemented to understand the spatial distribution of aftershocks. Here we propose a new metric for evaluating aftershock generation based on the energetics of shear faulting in a prestressed state. Unlike ΔCFS, the new metric depends not only on coseismic stress changes but also on background crustal stresses. The energetics-based formulation expands the ΔCFS defined on a specific plane into a generalized failure stress defined in a three-dimensional space. We examined the diagnostic ability of the new metric by applying receiver operating characteristic analysis to 12,175 aftershocks (M≥0) that followed the 1992 Landers earthquake. With a realistic background stress field inferred from thousands of earthquake focal mechanisms, the new metric was able to discriminate the triggering mechanisms of the aftershocks: two-thirds were direct responses to coseismic stress changes and one-fifth resulted from strength decreases owing to pore-fluid pressure increases.

Generation of critical aftershocks using stochastic neural networks and wavelet packet transform

This paper proposes a methodology using wavelet packet transform, principal component analysis, and neural networks in order to generate artificial critical aftershock accelerograms which are compatible with the response spectra. This procedure uses the learning abilities of neural networks, principal component analysis as a dimension reduction technique, and decomposing capabilities of wavelet packet transform on consecutive earthquakes. In fact, the proposed methodology consists of two steps and expands the knowledge of the inverse mapping from mainshock response spectrum to aftershock response spectrum and aftershock response spectrum to wavelet packet transform coefficients of the aftershocks. This procedure results in a stochastic ensemble of response spectra of aftershock (first step) and corresponding wavelet packet transform coefficients (second step) which are then used to generate the aftershocks through applying the inverse wavelet packet transform. Finally, in order to demonstrate the effectiveness of the proposed method, three examples are presented in which recorded critical successive ground motions are used to train and test the neural networks.

Global statistics of aftershocks following large earthquakes: independence of times and magnitudes

This paper considers the global statistics of times of largest aftershocks relative to the times of the corresponding main shocks. A large data set was used to show that the time-dependent distribution of largest aftershocks obeys a power law distribution. This is analogous to the Omori law for the sequence of all after- shocks. It is also shown that the times of the second, etc., largest aftershocks obey the same distribution. Thereby, we have confirmed the hypothesis that the times and magnitudes in an aftershock sequence are independent and make a good case for the Reasenberg-Jones representation of the aftershock process as a superposition of the Omori-Utsu law and the Gutenberg–Richter relation. Events that are smaller than the largest in an aftershock sequence show no delay relative to the largest event; this rejects the idea of the after- shock process as a direct failure cascade involving gradual transitions from larger to lesser scales, which imposes certain restrictions on the widely popular stochastic models of aftershock generation as branching processes. The above result is important in practice for prediction of aftershock activity and for assessing the hazard of large aftershocks.

Global statistics of aftershocks following large earthquakes: independence of times and magnitudes

This paper considers the global statistics of times of largest aftershocks relative to the times of the corresponding main shocks. A large data set was used to show that the time-dependent distribution of largest aftershocks obeys a power law distribution. This is analogous to the Omori law for the sequence of all after- shocks. It is also shown that the times of the second, etc., largest aftershocks obey the same distribution. Thereby, we have confirmed the hypothesis that the times and magnitudes in an aftershock sequence are independent and make a good case for the Reasenberg-Jones representation of the aftershock process as a superposition of the Omori-Utsu law and the Gutenberg–Richter relation. Events that are smaller than the largest in an aftershock sequence show no delay relative to the largest event; this rejects the idea of the after- shock process as a direct failure cascade involving gradual transitions from larger to lesser scales, which imposes certain restrictions on the widely popular stochastic models of aftershock generation as branching processes. The above result is important in practice for prediction of aftershock activity and for assessing the hazard of large aftershocks.

Global Statistics of Aftershocks Following Large Earthquakes: Independence of Times and Magnitudes

This paper considers the global statistics of times of largest aftershocks relative to the times of the corresponding main shocks. A large data set was used to show that the time-dependent distribution of largest aftershocks obeys a power law distribution. This is analogous to the Omori law for the sequence of all aftershocks. It is also shown that the times of the second, etc., largest aftershocks obey the same distribution. Thereby, we have confirmed the hypothesis that the times and magnitudes in an aftershock sequence are independent and make a good case for the Reasenberg–Jones representation of the aftershock process as a superposition of the Omori–Utsu law and the Gutenberg–Richter relation. Events that are smaller than the largest in an aftershock sequence show no delay relative to the largest event; this rejects the idea of the aftershock process as a direct failure cascade involving gradual transitions from larger to lesser scales, which imposes certain restrictions on the widely popular stochastic models of aftershock generation as branching processes. The above result is important in practice for prediction of aftershock activity and for assessing the hazard of large aftershocks.

Application of the Artificial Neural Network for Predicting Mainshock-Aftershock Sequences in Seismic Assessment of Reinforced Concrete Structures

ABSTRACT Studies from past earthquakes have shown that aftershocks can increase the vulnerability of structures damaged under mainshocks. However, the main challenge is to take these threats into account during the process of design and retrofit. Using the accelerograms that are compatible with the mainshock and the corresponding aftershocks can lead to a proper evaluation of seismic performance of a structure during the mainshock and aftershock, based on time history analyses. Therefore, it is necessary to simulate the aftershock motions consistent with the predicted mainshock spectrum. This study first investigates the relationship between the frequency content of an aftershock and its corresponding mainshock. It then discusses the development of a method for the generation of artificial aftershocks based on the acceleration response spectrum obtained from the artificial neural networks. To evaluate the effectiveness of the proposed method for the generation of artificial seismic sequences, a large number of time-history analyses have been conducted using both the current (the conventional back-to-back approach) and the proposed methods for three different structures including a column (representing the pier of a bridge), a reinforced concrete (RC) building, and an RC bridge. In these analyses, a wide range of different elements and materials have been employed and different responses, including maximum drift, residual drift, inter-story drift, floor acceleration, column curvature, and base shear, have also been studied. The results revealed that the proposed method produces the appropriate estimation of structural responses, and has a considerably higher efficiency compared to the conventional BTB approach that often introduces considerably large errors in the estimation of responses.

Why Do Aftershocks Occur Within the Rupture Area of a Large Earthquake?

AbstractThe mainshock‐aftershock sequence is one of the fundamental characteristics of seismogenesis, yet the physical mechanism of aftershock generation remains poorly understood. The simple explanation that aftershocks are caused by the mainshock's redistribution of strain energy is not always applicable, especially for aftershocks within the mainshock slip area, where strain energy is released. Here we show that the genesis of aftershocks can be modeled using a frictionally heterogeneous fault system. We conducted quasi‐dynamic numerical simulations of fault rupture cycles on a finite fault governed by a rate‐ and state‐dependent friction law. Aftershocks are observed around and within the mainshock rupture area when the frictional heterogeneity varies significantly along the fault. On the other hand, aftershocks are not produced when along‐fault variations in the frictional heterogeneity are small, which mimics the observed lack of aftershocks for repeating earthquakes.

Two complementary stress release processes based on departures from Omori’s law

Three aftershock series in Southern California, associated with the mainshocks of Landers (1992), Northridge (1994) and Hector Mine (1999) are interpreted as the superposition of a long stress relaxation process, in agreement with the Omori’s law, and a number of short episodes with sudden stress release. These short episodes are detected as departures from the expected Omori’s law. Aftershocks following a long stress relaxation process are designed as leading aftershocks (LA) and their generation rate fits well to the modified Omori’s law (MOL). The rest of aftershocks correspond to the different episodes of sudden stress release, every one designed as a cascade (CA). Cascades are found to be characterized by four basic properties. First, although the number of aftershocks NC belonging to every cascade fluctuates greatly along the stress relaxation process, a clear positive trend of NC with the elapsed time since the mainshock is observed in the initial phase of the aftershock sequence. After a critical elapsed time, very short in comparison with the aftershock sequence length, this positive trend diminishes significantly or even becomes negative. Second, the aftershock generation rate, GR, for every cascade, is almost constant. Third, GR for the successive CAs, represented as a function of its starting time measured since the mainshock, decreases according to a power-law. Four, the validity of the Gutenberg-Richter law is preserved for the sequence of all aftershocks belonging to cascades, with values of the b-parameter quite similar to those deduced for the complete aftershock series. Additionally, some statistical and fractal (self-affine) properties of CAs are analyzed. Given that the number of aftershocks with minimum magnitudes assuring catalogue completeness ranges, for the three seismic crises, from 6,000 up to 20,000, properties concerning LA and CA can be established with a high degree of confidence. As a conclusion, a single stress relaxation process following the Omori’s law should be discarded. A physical explanation, based on the complex spatial patterns of the stress field and tectonic fractures, as well as on the two proposed relaxation processes, is qualitatively discussed.

Variation of b and p values from aftershocks sequences along the Mexican subduction zone and their relation to plate characteristics

Aftershock sequences along the Mexican subduction margin (between coordinates 110ºW and 91ºW) were analyzed by means of the p value from the Omori–Utsu relation and the b value from the Gutenberg–Richter relation. We focused on recent medium to large (Mw > 5.6) events considered susceptible of generating aftershock sequences suitable for analysis. The main goal was to try to find a possible correlation between aftershock parameters and plate characteristics, such as displacement rate, age and segmentation. The subduction regime of Mexico is one of the most active regions of the world with a high frequency of occurrence of medium to large events and plate characteristics change along the subduction margin. Previous studies have observed differences in seismic source characteristics at the subduction regime, which may indicate a difference in rheology and possible segmentation. The results of the analysis of the aftershock sequences indicate a slight tendency for p values to decrease from west to east with increasing of plate age although a statistical significance is undermined by the small number of aftershocks in the sequences, a particular feature distinctive of the region as compared to other world subduction regimes. The b values show an opposite, increasing trend towards the east even though the statistical significance is not enough to warrant the validation of such a trend. A linear regression between both parameters provides additional support for the inverse relation. Moreover, we calculated the seismic coupling coefficient, showing a direct relation with the p and b values. While we cannot undoubtedly confirm the hypothesis that aftershock generation depends on certain tectonic characteristics (age, thickness, temperature), our results do not reject it thus encouraging further study into this question.