Earthquake Dataset Research Articles

Estimating the Probability of Earthquake Occurrence and Return Period Using Generalized Linear Models

In this paper, the frequency of an earthquake occurrence and magnitude relationship has been modeled with generalized linear models for the set of earthquake data of Nepal. A goodness of fit of a statistical model is applied for generalized linear models and considering the model selection information criterion, Akaike information criterion and Bayesian information criterion, generalized Poisson regression model has been selected as a suitable model for the study. The objective of this study is to determine the parameters (a and b values), estimate the probability of an earthquake occurrence and its return period using a Poisson regression model and compared with the Gutenberg-Richter model. The study suggests that the probabilities of earthquake occurrences and return periods estimated by both the models are relatively close to each other. The return periods from the generalized Poisson regression model are comparatively smaller than the Gutenberg-Richter model.

Calibration and Validation of MEEP Method for Location and Magnitude Estimation of Historical Earthquakes from Intra-Carpathian Region of Romania

Intra-Carpathian Region of Romania experienced strong earthquakes as it is mentioned in the national and international databases, e.g. November 26, 1829 (Ms = 6.4), October 10, 1834 (Ms = 6.3), January 26, 1916 (Ms = 6.4), July 12, 1991 (Mw = 5.7). The paper focuses on calibration and validation of MEEP method (Macroseismic Estimation of Earthquake Parameters, Musson & Jimenez, 2008) to provide a homogeneous and repeatable procedure for processing intensity data points (IDPs) which may supply the location and magnitude Mw of historical earthquakes occurred in the Intra-Carpathian Region of Romania and surroundings. The MEEP method is based on 1) grid-search technique for locating the epicentres and 2) the felt area approach to estimate the magnitude. It has the advantage to be used for small and moderate events. We used two high-quality datasets of earthquakes of the 20th century for which macroseismic data points (IDPs), instrumental moment magnitude (Mw) and instrumental epicentral coordinates are available, one for calibration and the other one for validation of the method. The main parameters used through MEEP algorithm were calibrated at regional scale using high-quality instrumental data: 1) crustal attenuation Q3 = 435 (Q at f = 3 Hz, which is understood to be the limit of the human perception of seismic waves), 2) Kövesligethy’s inelastic attenuation (∝ = 0.0065), 3) isoseismic spacing coefficient (k = 3.1) and 4) C = 2015, a constant generally related to tectonic regime. The MEEP method was successfully tested for events occurred within extra-Carpathian regions and the Pannonian Basin.

STanford EArthquake Dataset (STEAD): A Global Data Set of Seismic Signals for AI

Seismology is a data rich and data-driven science. Application of machine learning for gaining new insights from seismic data is a rapidly evolving sub-field of seismology. The availability of a large amount of seismic data and computational resources, together with the development of advanced techniques can foster more robust models and algorithms to process and analyze seismic signals. Known examples or labeled data sets, are the essential requisite for building supervised models. Seismology has labeled data, but the reliability of those labels is highly variable, and the lack of high-quality labeled data sets to serve as ground truth as well as the lack of standard benchmarks are obstacles to more rapid progress. In this paper we present a high-quality, large-scale, and global data set of local earthquake and non-earthquake signals recorded by seismic instruments. The data set in its current state contains two categories: (1) local earthquake waveforms (recorded at “local” distances within 350 km of earthquakes) and (2) seismic noise waveforms that are free of earthquake signals. Together these data comprise ~1.2 million time series or more than 19,000 hours of seismic signal recordings. Constructing such a large-scale database with reliable labels is a challenging task. Here, we present the properties of the data set, describe the data collection, quality control procedures, and processing steps we undertook to insure accurate labeling, and discuss potential applications. We hope that the scale and accuracy of STEAD presents new and unparalleled opportunities to researchers in the seismological community and beyond.

Second‐order analysis of marked inhomogeneous spatiotemporal point processes: Applications to earthquake data

AbstractTo analyze interactions in marked spatiotemporal point processes (MSTPPs), we introduce marked second‐order reduced moment measures andK‐functions for inhomogeneous second‐order intensity‐reweighted stationary MSTPPs. These summary statistics, which allow us to quantify dependence between different mark‐based classifications of points, depend on the specific mark space and mark reference measure chosen. Unbiased and consistent minus‐sampling estimators are derived for all statistics considered, and a test for random labeling is indicated. In addition, we treat Voronoi intensity estimators for MSTPPs. These new statistics are finally employed to analyze an Andaman Sea earthquake data set.

Celebrating the 25th International Conference of The International Environmetrics Society

Celebrating the 25th International Conference of The International Environmetrics Society

S-wave attenuation of the shallow sediments in the North China basin based on borehole seismograms of local earthquakes

S-wave velocity and attenuation structures of shallow sediments play important roles in accurate prediction of strong ground motion. However, it is more difficult to investigate the attenuation than velocity structures. In this study, we developed a new approach for estimating frequency-dependent S-wave attenuation (⁠|$Q_S^{ - 1}$|⁠) structures of shallow sediments based on multiple time window analysis of borehole seismograms from local earthquakes. Multiple time windows for separating direct and surface-reflected S-waves in local earthquake waveforms at borehole stations are selected with a global optimization scheme. With respect to different time windows, the transfer functions between direct and surface-reflected S-waves are achieved with a weighted averaging scheme, based on which frequency dependent |$Q_S^{ - 1}$| values are obtained. Synthetic tests suggest that the proposed method can restore robust and reliable |${\rm{\ }}Q_S^{ - 1}$| values, especially when the data set of local earthquakes is not abundant. We utilize this method for local earthquake waveforms at 14 borehole seismic stations in the North China basin, and obtain |$Q_S^{ - 1}$| values in 2–10 Hz frequency band, as well as average |${V_P}$|⁠, |${V_S}$| and |${V_P}/{\rm{\ }}{V_S}$| ratio for shallow sediments deep to a few hundred metres. Results suggest that |$Q_S^{ - 1}$| values are to 0.01–0.06, and generally decrease with frequency. The average attenuation structure of shallow sediments within the depth of a few hundred metres beneath 14 borehole stations in the North China basin can be modelled as |$Q_S^{ - 1} = {\rm{\ }}0.056{f^{ - 0.61}}$|⁠. It is generally consistent with the attenuation structure of sedimentary basins in other areas, such as Mississippi Embayment sediments in the United States and Sendai basin in Japan.

Regional variations and earthquake frequency–magnitude distribution and fractal dimension in the North of Central-East Iran Blocks (NCEIB)

The Gutenberg–Richter parameters (a and b), fractal dimension (D C ), and relationships between these parameters are calculated for different regions of the North of Central-East Iran Blocks (NCEIB). The whole examined area (between 34°–36° N and 55°–61° E) is divided into 55 equal square grids. Both the a and b values for the frequency–magnitude distribution (FMD) and D C are investigated simultaneously from 55 equal square grids. By using the completeness earthquake dataset for earthquakes of the instrumental period from 1976 to 2015, it is concluded that calculated values of a, b, and D C imply variations of seismotectonic stress. The most vulnerable regions for occurrence of the large earthquakes in the NCEIB are considering the computed lowest b values and the highest D C values. The relationships among D C –b and D C –(a/b) are used to classify the level of earthquake hazards for individual seismic source zones, in which the calibration curves illustrate a positive correlation among the D C and b values (Dc = 0.609 b − 0.008) and also a positive correlation among the D C and a/b ratios (Dc = 0.069 a/b + 0.215) having similar regression coefficients (R2 = 0.80 to 0.87) for both regressions. It is observed that the relationship among a/b and D C may be used for evaluation of seismicity and earthquake hazard assessment because of the high value for correlation coefficients and limited scattering of the calculated parameters.

Numerically Calculated 3D Space-Weighting Functions to Image Crustal Volcanic Structures Using Diffuse Coda Waves

Seismic coda measurements retrieve parameters linked to the physical characteristics of rock volumes illuminated by high frequency scattered waves. Space weighting functions (SWF) and kernels are different tools that model the spatial sensitivity of coda envelopes to scattering and absorption anomalies in these rock matrices, allowing coda-wave attenuation ( Q c o d a ) imaging. This note clarifies the difference between SWF and sensitivity kernels developed for coda wave imaging. It extends the SWF previously developed in 2D to the third dimension by using radiative transfer and the diffusion equation, based on the assumption that variations of Q c o d a depend solely on variations of the extinction length. When applied to active data (Deception Island, Antarctica), 3D SWF images strongly resemble 2D images, making this 3D extension redundant. On the other hand, diffusion does not efficiently model coda waveforms when using earthquake datasets spanning depths between 0 and 20 km, such as at Mount St. Helens volcano. In this setting, scattering attenuation and absorption suffer tradeoffs and cannot be separated by fitting a single seismogram energy envelope for SWF imaging. We propose that an approximate analytical 3D SWF, similar in shape to the common coda kernels used in literature, can still be used in a space weighted back-projection approach. While Q c o d a is not a physical parameter of the propagation medium, its spatially-dependent modeling allows improved reconstruction of crustal-scale tectonic and geological features. It is even more efficient as a velocity independent imaging tool for magma and fluid storage when applied to deep volcanism.

ISC-EHB: reconstruction of a robust earthquake data set

The Engdahl-van der Hilst-Buland (EHB) Bulletin of hypocentres and associated traveltime residuals was originally developed with procedures described by (Engdahl et al. 1998) and ended in 2008. It is a widely used seismological data set, which is now expanded and reconstructed, partly by exploiting updated procedures at the International Seismological Centre (ISC), to produce the ISC-EHB. The reconstruction begins in the modern period (2000–2013) to which new and more rigorous procedures for event selection, data preparation, processing and relocation are applied. The selection criteria minimize the location bias produced by unmodelled 3-D Earth structure, resulting in events that are relatively well-located in any given region. Depths of the selected events are significantly improved by a more comprehensive review of near station and secondary phase traveltime residuals based on ISC data, especially for the depth phases pP, pwP and sP, as well as by a rigorous review of the event depths in subduction zone cross-sections. The resulting cross-sections and associated maps are shown to provide details of seismicity in subduction zones in much greater detail than previously achievable. The new ISC-EHB data set will be especially useful for global seismicity studies and high-frequency regional and global tomographic inversions.

Earthquake behavior of steel cushion-implemented reinforced concrete frames

The earthquake performance of vulnerable structures can be increased by the implementation of supplementary energy-dissipative metallic elements. The main aim of this paper is to describe the earthquake behavior of steel cushion-implemented reinforced concrete frames (SCI-RCFR) in terms of displacement demands and energy components. Several quasi-static experiments were performed on steel cushions (SC) installed in reinforced concrete (RC) frames. The test results served as the basis of the analytical models of SCs and a bare reinforced concrete frame (B-RCFR). These models were integrated in order to obtain the resulting analytical model of the SCI-RCFR. Nonlinear-time history analyses (NTHA) were performed on the SCI-RCFR under the effects of the selected earthquake data set. According to the NTHA, SC application is an effective technique for increasing the seismic performance of RC structures. The main portion of the earthquake input energy was dissipated through SCs. SCs succeeded in decreasing the plastic energy demand on structural elements by almost 50% at distinct drift levels.

Slab narrowing in the Central Mediterranean: the Calabro-Ionian subduction zone as imaged by high resolution seismic tomography

A detailed 3D image of the Calabro-Ionian subduction system in the central Mediterranean was obtained by means of a seismic tomography, exploiting a large dataset of local earthquakes and computing algorithms able to build a dense grid of measure nodes. Results show that the slab is continuous below the southern sector of the Calabro-Peloritan Arc, but the deformation processes developing at its edges are leading to its progressive narrowing, influencing tectonics and magmatism at the surface, and with possible stress concentration in the tip zones. In the southwest, the deformation occurring at a free slab edge lead to propagation of a vertical lithospheric tear in the overriding plate, which extends along a NW-SE fault system (Aeolian-Tindari-Letojanni) up to about 30 km into the Ionian Sea; further southeast, the lithosphere appears only flexed and not broken yet. In the northeast, the slab seems to break progressively, parallel to the trench. Finally, northwest of Mt. Etna, the tomography highlights low VP that can be related to an upwelling of deep mantle material likely flowing laterally through a window opened by the complete slab detachment.

6600 years of earthquake record in the Julian Alps (Lake Bohinj, Slovenia)

AbstractSequences of lake sediments often form long and continuous records that may be sensitive recorders of seismic shaking. A multi‐proxy analysis of Lake Bohinj sediments associated with a well‐constrained chronology was conducted to reconstruct Holocene seismic activity in the Julian Alps (Slovenia). A seismic reflection survey and sedimentological analyses identified 29 homogenite‐type deposits related to mass‐wasting deposits. The most recent homogenites can be linked to historical regional earthquakes (i.e. 1348 ad, 1511 adand 1690 ad) with strong epicentral intensity [greater than ‘damaging’ (VIII) on the Medvedev–Sponheuer–Karnik scale]. The correlation between the historical earthquake data set and the homogenites identified in a core isolated from local stream inputs, allows interpretation of all similar deposits as earthquake related. This work extends the earthquake chronicle of the last 6600 years in this area with a total of 29 events recorded. The early Holocene sedimentary record is disturbed by a seismic event (6617 ± 94 cal yr bp) that reworked previously deposited sediment and led to a thick sediment deposit identified in the seismic survey. The period between 3500 cal yr bpand 2000 cal yr bpis characterized by a major destabilization in the watershed by human activities that led to increases in erosion and sedimentation rates. This change increased the lake's sensitivity to recording an earthquake (earthquake‐sensitivity threshold index) with the occurrence of 72 turbidite‐type deposits over this period. The high turbidite frequency identified could be the consequence of this change in lake earthquake sensitivity and thus these turbidites could be triggered by earthquake shaking, as other origins are discarded. This study illustrates why it is not acceptable to propose a return period for seismic activity recorded in lake sediment if the sedimentation rate varies significantly.

A New Generalization of the Pareto Distribution and Its Application to Insurance Data

The Pareto classical distribution is one of the most attractive in statistics and particularly in the scenario of actuarial statistics and finance. For example, it is widely used when calculating reinsurance premiums. In the last years, many alternative distributions have been proposed to obtain better adjustments especially when the tail of the empirical distribution of the data is very long. In this work, an alternative generalization of the Pareto distribution is proposed and its properties are studied. Finally, application of the proposed model to the earthquake insurance data set is presented.

Aseismic Transform Fault Slip at the Mendocino Triple Junction From Characteristically Repeating Earthquakes

AbstractThe Mendocino Triple Junction (MTJ), at the northern terminus of the San Andreas Fault system, is an actively deforming plate boundary region with poorly constrained estimates of seismic coupling on most offshore fault surfaces. Characteristically repeating earthquakes provide spatial and temporal descriptions of aseismic creep at the MTJ, including on the oceanic transform Mendocino Fault Zone (MFZ) as it subducts beneath North America. Using a dataset of earthquakes from 2008 to 2017, we find that the easternmost segment of the MFZ displays creep during this period at about 65% of the long‐term slip rate. We also find creep at slower rates on the shallower strike‐slip interface between the Pacific plate and the North American accretionary wedge, as well as on a fault that accommodates Gorda subplate internal deformation. After a nearby M5.7 earthquake in 2015, we observe a possible decrease in aseismic slip on the near‐shore MFZ that lasts from 2015 to at least early 2017.

Two-dimensional joint inversion of Magnetotelluric and local earthquake data: Discussion on the contribution to the solution of deep subsurface structures

Joint inversion of data sets collected by using several geophysical exploration methods has gained importance and associated algorithms have been developed. To explore the deep subsurface structures, Magnetotelluric and local earthquake tomography algorithms are generally used individually. Due to the usage of natural resources in both methods, it is not possible to increase data quality and resolution of model parameters. For this reason, the solution of the deep structures with the individual usage of the methods cannot be fully attained. In this paper, we firstly focused on the effects of both Magnetotelluric and local earthquake data sets on the solution of deep structures and discussed the results on the basis of the resolving power of the methods. The presence of deep-focus seismic sources increase the resolution of deep structures. Moreover, conductivity distribution of relatively shallow structures can be solved with high resolution by using MT algorithm. Therefore, we developed a new joint inversion algorithm based on the cross gradient function in order to jointly invert Magnetotelluric and local earthquake data sets. In the study, we added a new regularization parameter into the second term of the parameter correction vector of Gallardo and Meju (2003). The new regularization parameter is enhancing the stability of the algorithm and controls the contribution of the cross gradient term in the solution. The results show that even in cases where resistivity and velocity boundaries are different, both methods influence each other positively. In addition, the region of common structural boundaries of the models are clearly mapped compared with original models. Furthermore, deep structures are identified satisfactorily even with using the minimum number of seismic sources. In this paper, in order to understand the future studies, we discussed joint inversion of Magnetotelluric and local earthquake data sets only in two-dimensional space. In the light of these results and by means of the acceleration on the three-dimensional modelling and inversion algorithms, it is thought that it may be easier to identify underground structures with high resolution.

Functional Link Neural Network Learning for Response Prediction of Tall Shear Buildings With Respect to Earthquake Data

This paper proposes the application of functional link neural networks (FLNNs) for structural response prediction of tall buildings due to seismic loads. The ground acceleration data are taken as input, and structural responses of different floors of multistorey shear buildings are considered as output. It is worth mentioning that handling of large earthquake data has become a great challenge in the design of tall structures viz., that of shear buildings. As such, here, a functional expansion block in FLNN has been used along with efficient Chebyshev and Legendre polynomials. Training is done with one earthquake data set, and testing is done with different intensities of other earthquake data sets; and it is seen that FLNN can very well predict the structural response of different floors of multistorey shear buildings subject to earthquake data. Results of the FLNN are compared with a multilayer neural network (MNN), and it is found that the FLNN gives better accuracy and takes less computation time compared to MNN, which shows the computational efficiency of FLNN over MNN. Numerical examples of two-, five-, and ten-storey buildings are considered, and corresponding results are presented in the form of tables and plots.

An adaptive method for clustering spatio‐temporal events

AbstractThe clustering of spatio‐temporal events has become one of the most important research branches of spatio‐temporal data mining. However, the discovery of clusters of spatio‐temporal events with different shapes and densities remains a challenging problem because of the subjectivity in the choice of two critical parameters: the spatio‐temporal window for estimating the density around each event, and the density threshold for evaluating the significance of clusters. To make the clustering of spatio‐temporal events objective, in this study these two parameters were adaptively generated from statistical information about the dataset. More precisely, the density threshold was statistically modeled as an adjusted significance level controlled by the cardinality and support domain of the dataset, and the appropriate sizes of spatio‐temporal windows for clustering were determined by the spatio‐temporal classification entropy and stability analysis. Experiments on both simulated and earthquake datasets were conducted, and the results show that the proposed method can identify clusters of different shapes and densities.

Site amplification and resonance frequency in the urban environment

The effects of the built environment on site amplification and resonance frequency are studied utilizing 34 earthquake recordings made at three seismic downhole arrays (ATK, FTH and ZYT) and two outcrop stations at Istanbul. The earthquake data set included minor to moderate earthquakes having hypocenters at various distances and azimuths with respect to the arrays. The buildings considered in the study are at distances less than 110m from the arrays and have different dynamic characteristics. The results indicate the impact of the buildings on the site amplification and resonance frequency. At ATK and ZYT arrays, the buildings have a lowering effect on the site amplification, while at FTH array a shift in the resonance frequency is observed. Results from ATK and FTH sites also support previous findings that even in case of similarity between the shear wave velocity profiles amplification levels can differ significantly.

When probabilistic seismic hazard climbs volcanoes: the Mt. Etna case, Italy – Part 1: Model components for sources parameterization

Abstract. The volcanic region of Mt. Etna (Sicily, Italy) represents a perfect lab for testing innovative approaches to seismic hazard assessment. This is largely due to the long record of historical and recent observations of seismic and tectonic phenomena, the high quality of various geophysical monitoring and particularly the rapid geodynamics clearly demonstrate some seismotectonic processes. We present here the model components and the procedures adopted for defining seismic sources to be used in a new generation of probabilistic seismic hazard assessment (PSHA), the first results and maps of which are presented in a companion paper, Peruzza et al. (2017). The sources include, with increasing complexity, seismic zones, individual faults and gridded point sources that are obtained by integrating geological field data with long and short earthquake datasets (the historical macroseismic catalogue, which covers about 3 centuries, and a high-quality instrumental location database for the last decades). The analysis of the frequency–magnitude distribution identifies two main fault systems within the volcanic complex featuring different seismic rates that are controlled essentially by volcano-tectonic processes. We discuss the variability of the mean occurrence times of major earthquakes along the main Etnean faults by using an historical approach and a purely geologic method. We derive a magnitude–size scaling relationship specifically for this volcanic area, which has been implemented into a recently developed software tool – FiSH (Pace et al., 2016) – that we use to calculate the characteristic magnitudes and the related mean recurrence times expected for each fault. Results suggest that for the Mt. Etna area, the traditional assumptions of uniform and Poissonian seismicity can be relaxed; a time-dependent fault-based modeling, joined with a 3-D imaging of volcano-tectonic sources depicted by the recent instrumental seismicity, can therefore be implemented in PSHA maps. They can be relevant for the retrofitting of the existing building stock and for driving risk reduction interventions. These analyses do not account for regional M > 6 seismogenic sources which dominate the hazard over long return times (≥ 500 years).

When probabilistic seismic hazard climbs volcanoes: the Mt. Etna case, Italy – Part 2: Computational implementation and first results

Abstract. This paper describes the model implementation and presents results of a probabilistic seismic hazard assessment (PSHA) for the Mt. Etna volcanic region in Sicily, Italy, considering local volcano-tectonic earthquakes. Working in a volcanic region presents new challenges not typically faced in standard PSHA, which are broadly due to the nature of the local volcano-tectonic earthquakes, the cone shape of the volcano and the attenuation properties of seismic waves in the volcanic region. These have been accounted for through the development of a seismic source model that integrates data from different disciplines (historical and instrumental earthquake datasets, tectonic data, etc.; presented in Part 1, by Azzaro et al., 2017) and through the development and software implementation of original tools for the computation, such as a new ground-motion prediction equation and magnitude–scaling relationship specifically derived for this volcanic area, and the capability to account for the surficial topography in the hazard calculation, which influences source-to-site distances. Hazard calculations have been carried out after updating the most recent releases of two widely used PSHA software packages (CRISIS, as in Ordaz et al., 2013; the OpenQuake engine, as in Pagani et al., 2014). Results are computed for short- to mid-term exposure times (10 % probability of exceedance in 5 and 30 years, Poisson and time dependent) and spectral amplitudes of engineering interest. A preliminary exploration of the impact of site-specific response is also presented for the densely inhabited Etna's eastern flank, and the change in expected ground motion is finally commented on. These results do not account for M > 6 regional seismogenic sources which control the hazard at long return periods. However, by focusing on the impact of M < 6 local volcano-tectonic earthquakes, which dominate the hazard at the short- to mid-term exposure times considered in this study, we present a different viewpoint that, in our opinion, is relevant for retrofitting the existing buildings and for driving impending interventions of risk reduction.