On the selection of GMPEs for Vrancea subcrustal seismic source

The Vrancea subcrustal seismic source is characterized by large magnitude ( $M_{W} \ge 7$ ) intermediate-depth earthquakes that occur two or three times during a century on average. In this study several procedures are used to grade four candidate ground motion prediction equations proposed for Vrancea source in the SHARE project. In the work of Delavaud et al. (J Seismol 16(3):451–473, 2012) four ground motion prediction models developed for subduction zones (Zhao et al. in Bull Seism Soc Am 96(3):898–913, 2006; Atkinson and Boore in Bull Seism Soc Am 93(4):1703–1729, 2003; Youngs et al. in Seism Res Lett 68(1):58–73, 1997; Lin and Lee in Bull Seism Soc Am 98(1):220–240, 2008) are suggested as suitable for Vrancea subcrustal seismic source. The paper presents the appropriateness analysis of the four suggested ground motion prediction equations done using a dataset of 109 triaxial accelerograms recorded during seven Vrancea seismic events with moment magnitude $M_{W}$ between 5.4 and 7.4, occurred in the past 35 years. The strong ground motions were recorded in Romania, as well as in Bulgaria, Republic of Moldova and Serbia. Based on the ground motion dataset several goodness-of-fit measures are used in order to quantify how well the selected models match with the recorded data. The compatibility of the four ground motion prediction models with respect to magnitude scaling and distance scaling implied by strong ground motion dataset is investigated as well. The analyses show that the Youngs et al. (Seism Res Lett 68(1):58–73, 1997) and Zhao et al. (Bull Seism Soc Am 96(3):898–913, 2006) ground motion prediction models have a better fit with the data and can be candidate models for Probabilistic Seismic Hazard Assessment.     

Strong motion uncertainty determined from observed records by dense network in Japan

The variation of ground motions at specific stations from events in six narrow areas was inspected by using K-NET and KiK-net records. A source-area factor for individual observation stations was calculated by averaging ratios between observed values for horizontal peak acceleration and velocity, as well as acceleration response spectra for 5% damping, and predicted values using a ground-motion model (usually known as an attenuation relation) by Kanno et al. (Bull Seismol Soc Am, 96:879–897, 2006). Standard deviations between observed and predicted amplitudes after the correction factor are less than 0.2 on the logarithmic scale and decrease down to around 0.15 in the short-period range. Intra-event standard deviation clearly increases with decreasing distance due to differing paths around near source area. Standard deviations may increase with amplitude or decrease with magnitude; however, both amplitude and magnitude of the data are strongly correlated with distance. The standard deviation calculated in this study is obviously much smaller than that of the original ground-motion model, as epistemic uncertainties are minimized by grouping ground motions at specific stations. This result indicates that the accuracy of strong ground motion prediction could be improved if ground-motion models for specified region are determined individually. For this to be possible, it is necessary to have dense strong-motion networks in high-seismicity regions, such as K-NET and KiK-net.     

The importance of crustal structure in explaining the observed uncertainties in ground motion estimation

In this short article, the possible reduction in the standard deviation of empirical ground motion estimation equations through the modelling of the effect of crustal structure is assessed through the use of ground-motion simulations. Simulations are computed for different source-to-site distances, focal depths, focal mechanisms and for crustal models of the Pyrenees, the western Alps and the upper Rhine Graben. Through the method of equivalent hypocentral distance introduced by Douglas et al. [(2004) Bull Earthquake Eng 2(1): 75–99] to model the effect of crustal structure in empirical equations, the scatter associated with such equations derived using these simulated data could be reduced to zero if real-to-equivalent hypocentral distance mapping functions were derived for every combination of mechanism, depth and crustal structure present in the simulated dataset. This is, obviously, impractical. The relative importance of each parameter in affecting the decay of ground motions is assessed here. It is found that variation in focal depth is generally more important than the effect of crustal structure when deriving the real-to-equivalent hypocentral distance mapping functions. In addition, mechanism and magnitude do not have an important impact on the decay rate.     

Modelling of Strong Ground Motions from 1991 Uttarkashi,India, Earthquake Using a Hybrid Technique

We present a simple and efficient hybrid technique for simulating earthquake strong ground motion. This procedure is the combination of the techniques of envelope function (Midorikawa et al. Tectonophysics 218:287–295, 1993) and composite source model (Zeng et al. Geophys Res Lett 21:725–728, 1994). The first step of the technique is based on the construction of the envelope function of the large earthquake by superposition of envelope functions for smaller earthquakes. The smaller earthquakes (sub-events) of varying sizes are distributed randomly, instead of uniform distribution of same size sub-events, on the fault plane. The accelerogram of large event is then obtained by combining the envelope function with a band-limited white noise. The low-cut frequency of the band-limited white noise is chosen to correspond to the corner frequency for the target earthquake magnitude and the high-cut to the Boore’s f _max or a desired frequency for the simulation. Below the low-cut frequency, the fall-off slope is 2 in accordance with the ω² earthquake source model. The technique requires the parameters such as fault area, orientation of the fault, hypocenter, size of the sub-events, stress drop, rupture velocity, duration, source–site distance and attenuation parameter. The fidelity of the technique has been demonstrated by successful modeling of the 1991 Uttarkashi, Himalaya earthquake (Ms 7). The acceptable locations of the sub-events on the fault plane have been determined using a genetic algorithm. The main characteristics of the simulated accelerograms, comprised of the duration of strong ground shaking, peak ground acceleration and Fourier and response spectra, are, in general, in good agreement with those observed at most of the sites. At some of the sites the simulated accelerograms differ from observed ones by a factor of 2–3. The local site geology and topography may cause such a difference, as these effects have not been considered in the present technique. The advantage of the technique lies in the fact that detailed parameters such as velocity-Q structures and empirical Green’s functions are not required or the records of the actual time history from the past earthquakes are not available. This method may find its application in preparing a wide range of scenarios based on simulation. This provides information that is complementary to the information available in probabilistic hazard maps.     

Empirical models for estimating liquefaction-induced lateral spread displacement

Existing empirical models for estimating liquefaction-induced lateral spread displacement (D_LL) have been derived from a dataset poorly distributed with respect to earthquake magnitude and source distance, and also produced from different tectonic source types and faulting mechanisms. Both the poor distribution and mixed tectonic source types and faulting mechanisms of the data have an adverse impact on the reliability of the empirical models. To overcome these problems in the development of empirical models, we replace the direct use of magnitude and source distance with pseudo-displacement derived from spectral acceleration attenuation models that are well supported by earthquake data, and use a modification factor to account for effects of the non-linear soil response. Attenuation models derived from very large and reasonably well-balanced datasets have been selected, one being a Japanese attenuation model and other being a combination of the Sadigh et al. model and the Youngs et al. model. These models are capable of accounting for the effects of earthquake tectonic source type and faulting mechanism. We determined the model coefficients by selecting the pseudo-displacements calculated for a number of spectral periods to achieve an unbiased distribution of residuals with respect to earthquake magnitude and source distance. Sensitivity analyses have been carried out based on the new and existing models, and show that the new model is more robust than the existing models. Comparison with a limited number of data from the 1997 Kocaeli, Turkey earthquake suggests our model provides comparable liquefaction-induced lateral displacement D_LL estimates.     

Harmonization check of M w within the central, northern, and northwestern European earthquake catalogue (CENEC)

Large data sets covering large areas and time spans and composed of many different independent sources raise the question of the obtained degree of harmonization. The present study is an analysis of the harmonization with respect to the moment magnitude M _w within the earthquake catalogue for central, northern, and northwestern Europe (CENEC). The CENEC earthquake catalogue (Grünthal et al., J Seismol, 2009) contains parameters for over 8,000 events in the time period 1000–2004 with magnitude M _w ≥ 3.5. Only about 2% of the data used for CENEC have original M _w magnitudes derived directly from digital data. Some of the local catalogues and data files providing data give M _w, but calculated by the respective agency from other magnitude measures or intensity. About 60% of the local data give strength measures other than M _w, and these have to be transformed by us using available formulae or new regressions based on original M _w data. Although all events are thus unified to M _w magnitude, inhomogeneity in the M _w obtained from over 40 local catalogues and data files and 50 special studies is inevitable. Two different approaches have been followed to investigate the compatibility of the different M _w sets throughout CENEC. The first harmonization check is performed using M _w from moment tensor solutions from SMTS and Pondrelli et al. (Phys Earth Planet Inter 130:71–101, 2002; Phys Earth Planet Inter 164:90–112, 2007). The method to derive the SMTS is described, e.g., by Braunmiller et al. (Tectonophysics 356:5–22, 2002) and Bernardi et al. (Geophys J Int 157:703–716, 2004), and the data are available in greater extent since 1997. One check is made against the M _w given in national catalogues and another against the M _w derived by applying different empirical relations developed for CENEC. The second harmonization check concerns the vast majority of data in CENEC related to earthquakes prior to 1997 or where no moment tensor based M _w exists. In this case, an empirical relation for the M _w dependence on epicentral intensity (I ₀) and focal depth (h) was derived for 41 master events, i.e., earthquakes, located all over central Europe, with high-quality data. To include also the data lacking h, the corresponding depth-independent relation for these 41 events was also derived. These equations are compared with the different sets of data from which CENEC has been composed, and the goodness of fit is demonstrated for each set. The vast majority of the events are very well or reasonably consistent with the respective relation so that the data can be said to be harmonized with respect to M _w, but there are exceptions, which are discussed in detail.     

Strong ground motion simulation for the 2013 Lushan M W6.6 earthquake,Sichuan, China,based on the inverted and synthetic slip models

It is well known that quantitative estimation of slip distributions on fault plane is one of the most important issues for earthquake source inversion related to the fault rupture process. The characteristics of slip distribution on the main fault play a fundamental role to control strong ground motion pattern. A large amount of works have also suggested that variable slip models inverted from longer period ground motion recordings are relevant for the prediction of higher frequency ground motions. Zhang et al. (Chin J Geophys 56:1412–1417, 2013) and Wang et al. (Chin J Geophys 56:1408–1411, 2013) published their source inversions for the fault rupturing process soon after the April 20, 2013 Lushan earthquake in Sichuan, China. In this study, first, we synthesize two forward source slip models: the value of maximum slip, fault dimension, size, and dimension of major asperities, and corner wave number obtained from Wang’s model is adopted to constrain the generation of k ^?2 model and crack model. Next, both inverted and synthetic slip models are employed to simulate the ground motions for the Lushan earthquake based on the stochastic finite-fault method. In addition, for a comparison purpose, a stochastic slip model and another k ^?2 model (k ^?2 model II) with 2 times value of corner wave number of the original k ^?2 model (k ^?2 model I) are also employed for simulation for Lushan event. The simulated results characterized by Modified Mercalli Intensity (MMI) show that the source slip models based on the inverted and synthetic slip distributions could capture many basic features associated with the ground motion patterns. Moreover, the simulated MMI distributions reflect the rupture directivity effect and the influence of the shallow velocity structure well. On the other hand, the simulated MMI by stochastic slip model and k ^?2 model II is apparently higher than observed intensity. By contrast, our simulation results show that the higher frequency ground motion is sensitive to the degree of slip roughness; therefore, we suggest that, for realistic ground‐motion simulations due to future earthquake, it is imperative to properly estimate the slip roughness distribution.     

A regional near-surface high frequency spectral attenuation (kappa) model for northwestern Turkey

One approach to model the high-frequency attenuation of spectral amplitudes of S-waves is to express the observed exponential decay in terms of Kappa (κ) factor [1]. Kappa is a significant parameter used for identifying the high-frequency attenuation behavior of ground motions as well as one of the key parameters for stochastic strong ground motion simulation method. As of now, there is not a systematic investigation of the Kappa parameter based on the recently-compiled Turkish ground motions. In this study, we examine a strong ground motion dataset from Northwestern Turkey with varying source properties, site classes and epicentral distances. We manually compute κ from the S-wave portion of each record and study both horizontal and vertical kappa values. We use traditional regression techniques to describe the (potential) relationships between kappa and selected independent variables such as the site class, distance from the source or magnitude of the event. A linear effect of magnitude on kappa is not found statistically significant for the database studied herein. We express the initial findings of a regional κ model for Northwestern Turkey as a function of site class and epicentral distances. Single station analyses at selected sites confirm the regional model. Finally, we present stochastic strong motion simulations of past events in the region using the proposed kappa model. Regardless of the magnitude, source-to-site distance and local site conditions at the stations, the high-frequency spectral decay is simulated effectively at all stations considered.     

Ground-Motion Scaling in the Western Alps

In order to empirically obtain the scaling relationships for the high-frequency ground motion in the Western Alps (NW Italy), regressions are carried out on more than 7500 seismograms from 957 regional earthquakes. The waveforms were selected from the database of 6 three-component stations of the RSNI (Regional Seismic network of Northwestern Italy). The events, M _W ranging between 1.2 and 4.8, were recorded within a hypocentral distance of 200 km during the time period: 1996–2001. The peak ground velocities are measured in selected narrow-frequency bands, between 0.5 and 14 Hz. Results are presented in terms of a regional attenuation function for the vertical ground motion, a set of vertical excitation terms at the reference station STV2 (hard-rock), and a set of site terms (vertical and horizontal), all relative to the vertical component of station STV2.The regional propagation of the ground motion is modeled after quantifying the expected duration of the seismic motion as a function of frequency and hypocentral distance. A simple functional form is used to take into account both the geometrical and the anelastic attenuation: a multi-variable grid search yielded a quality factor Q(f) = 310f ^0.20, together with a quadri-linear geometrical spreading at low frequency. A simpler, bi-linear geometrical spreading seems to be more appropriate at higher frequencies (f > 1.0 Hz). Excitation terms are matched by using a Brune spectral model with variable, magnitude-dependent stress drop: at M _w 4.8, we used Δσ = 50 MPa. A regional distance-independent attenuation parameter is obtained (κ₀ = 0.012 s) by modelling the average spectral decay at high frequency of small earthquakes.In order to predict the absolute levels of ground shaking in the region, the excitation/attenuation model is used through the Random Vibration Theory (RVT) with a stochastic point-source model. The expected peak-ground accelerations (PGA) are compared with the ones derived by Ambraseys et al. (1996) for the Mediterranean region and by Sabetta and Pugliese (1996) for the Italian territory.     

Investigation on the stochastic simulation of strong ground motions for Bucharest area

This short article evaluates the stochastic method of ground motion simulation for Bucharest area using both the single-corner frequency model and recently introduced double-corner frequency models. A dedicated Q model is derived using ground motions obtained during the largest Vrancea earthquakes from the past 30 years. The simulated ground motions are tested against the observed data from the Vrancea earthquakes of August 1986 and May 1990. Moreover, the observed data are also compared against simulations obtained using the Q model derived by Oth et al. (2008). Finally, the results of the simulations show that the derived Q model is better suited for larger magnitude events, while the Q model of Oth et al. (2008) provides better results for smaller earthquakes.     

基于人工神经网络的仪器地震烈度预测模型研究

地震发生后,针对能够快速预测震中附近的烈度分布情况的问题,首先对632次地震触发的台站进行筛选,对2 231个台站触发后20 s内有效的7个地震动参数以及震级和震源距的信息进行提取,并利用人工神经网络对所选数据样本进行训练,建立三种有效的预测模型。研究结果显示模型一所选的输入参数为7个,不利用震源参数,在预测中有着较好的时效性,从第1 s到20 s,预测的平均烈度差值逐渐减小到0.45;模型二所选的输入参数为8个,利用了震源距信息,可以用于烈度级别的预测,预测的平均烈度差值逐渐减小到0.36;模型三所选的输入参数为9个,预测结果较好,可用于震后烈度场的实时预测,平均烈度差值逐渐减小到0.31。利用提出的3种模型对两次地震事件进行烈度预测,预测烈度差值取整后分别有95%和76%以上在1以内,有着较好的结果,可以用于地震预警当中。     

Neo-deterministic seismic hazard scenarios for India—a preventive tool for disaster mitigation

Current computational resources and physical knowledge of the seismic wave generation and propagation processes allow for reliable numerical and analytical models of waveform generation and propagation. From the simulation of ground motion, it is easy to extract the desired earthquake hazard parameters. Accordingly, a scenario-based approach to seismic hazard assessment has been developed, namely the neo-deterministic seismic hazard assessment (NDSHA), which allows for a wide range of possible seismic sources to be used in the definition of reliable scenarios by means of realistic waveforms modelling. Such reliable and comprehensive characterization of expected earthquake ground motion is essential to improve building codes, particularly for the protection of critical infrastructures and for land use planning. Parvez et al. (Geophys J Int 155:489–508, 2003) published the first ever neo-deterministic seismic hazard map of India by computing synthetic seismograms with input data set consisting of structural models, seismogenic zones, focal mechanisms and earthquake catalogues. As described in Panza et al. (Adv Geophys 53:93–165, 2012), the NDSHA methodology evolved with respect to the original formulation used by Parvez et al. (Geophys J Int 155:489–508, 2003): the computer codes were improved to better fit the need of producing realistic ground shaking maps and ground shaking scenarios, at different scale levels, exploiting the most significant pertinent progresses in data acquisition and modelling. Accordingly, the present study supplies a revised NDSHA map for India. The seismic hazard, expressed in terms of maximum displacement (Dmax), maximum velocity (Vmax) and design ground acceleration (DGA), has been extracted from the synthetic signals and mapped on a regular grid over the studied territory.     

地震动持时的概率危险性分析方法

对各种不同定义的地震动持续时间进行了评述，讨论了地震动持续时间与震级及震中距的各种定量关系，建立了以地震动持续时间为参数的地震危险性分析方法。     

3D crustal structure and long-period ground motions from a M9.0 megathrust earthquake in the Pacific Northwest region

We have developed a community velocity model for the Pacific Northwest region from northern California to southern Canada and carried out the first 3D simulation of a M_w 9.0 megathrust earthquake rupturing along the Cascadia subduction zone using a parallel supercomputer. A long-period (<0.5 Hz) source model was designed by mapping the inversion results for the December 26, 2004 Sumatra–Andaman earthquake (Han et al., Science 313(5787):658–662, 2006) onto the Cascadia subduction zone. Representative peak ground velocities for the metropolitan centers of the region include 42 cm/s in the Seattle area and 8–20 cm/s in the Tacoma, Olympia, Vancouver, and Portland areas. Combined with an extended duration of the shaking up to 5 min, these long-period ground motions may inflict significant damage on the built environment, in particular on the highrises in downtown Seattle.     

阿里亚斯强度衰减关系分析

本文选取美国NGA强震数据库中与我国大陆构造环境相近的美国西部、中亚和我国台湾地区的自由地表强震记录,计算其水平向阿里亚斯强度;采用分步回归法依次求取近场饱和项、距离衰减项、震级项、断层类型项和场地类别项的系数,得出衰减模型,并基于回归结果与实际数据的对比,分析了残差随距离或震级的分布.与前人工作得出的全球活动板块边界、新西兰和我国台湾地区的3种阿里亚斯强度衰减模型进行的比较显示,在M_W6.5或远场区域4种模型的结果较接近,近场区域在高震级和低震级时有一定差异.最后,选取2008年汶川M_W7.9和2013年芦山M_W6.8两次地震的强震记录对模型进行检验,结果显示本文衰减关系均与实际数据拟合得很好,表明本文的模型适用于我国大陆地区的相关研究.      

Spectral shape,epsilon and record selection

Selection of earthquake ground motions is considered with the goal of accurately estimating the response of a structure at a specified ground motion intensity, as measured by spectral acceleration at the first‐mode period of the structure, Sa(T₁). Consideration is given to the magnitude, distance and epsilon (ε) values of ground motions. First, it is seen that selecting records based on their ε values is more effective than selecting records based on magnitude and distance. Second, a method is discussed for finding the conditional response spectrum of a ground motion, given a level of Sa(T₁) and its associated mean (disaggregation‐based) causal magnitude, distance and ε value. Records can then be selected to match the mean of this target spectrum, and the same benefits are achieved as when records are selected based on ε. This mean target spectrum differs from a Uniform Hazard Spectrum, and it is argued that this new spectrum is a more appropriate target for record selection. When properly selecting records based on either spectral shape or ε, the reductions in bias and variance of resulting structural response estimates are comparable to the reductions achieved by using a vector‐valued measure of earthquake intensity. Copyright © 2006 John Wiley & Sons, Ltd.     

Ground motion prediction equations for horizontal and vertical components of acceleration in Northern Iran

Recent studies have shown that the vertical component of ground motion can be quite destructive on a variety of structural systems. Development of response spectrum for design of buildings subjected to vertical component of earthquake needs ground motion prediction equations (GMPEs). The existing GMPEs for northern Iranian plateau are proposed for the horizontal component of earthquake, and there is not any specified GMPE for the vertical component of earthquake in this region. Determination of GMPEs is mostly based on regression analyses on earthquake parameters such as magnitude, site class, distance, and spectral amplitudes. In this study, 325 three-component records of 55 earthquakes with magnitude ranging from M _w 4.1 to M _w 7.3 are used for estimation on the regression coefficients. Records with distances less than 300 km are selected for analyses in the database. The regression analyses on earthquake parameters results in determination of GMPEs for peak ground acceleration and spectral acceleration for both horizontal and vertical components of the ground motion. The correlation between the models for vertical and horizontal GMPEs is studied in details. These models are later compared with some other available GMPEs. According to the result of this investigation, the proposed GMPEs are in agreement with the other relationships that were developed based on the local and regional data.     

Spectral parameters of ground motion in different regions: comparison of empirical models

Fourier-amplitude spectrum is one of the most important parameters describing earthquake ground motion, and it is widely used for strong ground motion prediction and seismic hazard estimation. The relationships between Fourier-acceleration spectra, earthquake magnitude and distance were analysed for different seismic regions (the Caucasus and Taiwan island) on the basis of ground motion recordings of small to moderate (3.5≤M_L≤6.5) earthquakes. It has been found that the acceleration spectra of the most significant part of the records, starting from S-wave arrival, can be modelled accurately by the Brune's “ω-squared” point-source model. Parameters of the model are found to be region-dependent. Peak ground accelerations and response spectra for condition of rock sites were calculated using stochastic simulation technique and obtained models of source spectra. The modelled ground-motion parameters are compared with those predicted by recent empirical attenuation relationship for California.     

Use of spectral acceleration data for determination of three-dimensional attenuation structure in the Pithoragarh region of Kumaon Himalaya

Three-dimensional attenuation structures are related to the subsurface heterogeneities present in the earth crust. An algorithm for estimation of three-dimensional attenuation structure in the part of Garhwal Himalaya, India has been presented by Joshi (Curr Sci 90:581–585, 2006b; Nat Hazards 43:129–146, 2007). In continuation of our earlier approach, we have presented a method in which strong motion data have been used to estimate frequency-dependent three-dimensional attenuation structure of the region. The border district of Pithoragarh in the Higher Himalaya, India, lies in the central seismic gap region of Himalaya. This region falls in the seismic zones IV and V of the seismic zoning map of India. A dense network consisting of eight accelerographs has been installed in this region. This network has recorded several local events. An algorithm based on inversion of strong motion digital data is developed in this paper to estimate attenuation structure at different frequencies using the data recorded by this network. Twenty strong motion records observed at five stations have been used to estimate the site amplification factors using inversion algorithm defined in this paper. Site effects obtained from inversion has been compared with that obtained using Nakamura (1988) and Lermo et al. (Bull Seis Soc Am 83:1574–1594, 1993) approach. The obtained site amplification term has been used for correcting spectral acceleration data at different stations. The corrected spectral acceleration data have been used as an input to the developed algorithm to avoid effect of near-site soil amplification term. The attenuation structure is estimated by dividing the entire area in several three-dimensional block of different frequency-dependent shear wave quality factor Q _β (f). The input to this algorithm is the spectral acceleration of S phase of the corrected accelerogram. The outcome of the algorithm is given in terms of attenuation coefficient and source acceleration spectra. In the present study, this region has been divided into 25 rectangular blocks with thickness of 10 km and surface dimension of 12.5 × 12.1 km, respectively. Present study gives three-dimensional attenuation model of the region which can be used for both hazard estimation and simulation of strong ground motion.