ALBERT

Description: Expansion of urban areas in Central Asia increases their exposure to seismic hazard, but at present no earthquake early warning (EEW) systems exist in the region. Such systems, successfully implemented in other regions, aim to provide warning of the order of tens of seconds about impending disasters, enabling the first rapid-response steps to be taken. The feasibility of such systems for Bishkek, Kyrgyzstan, has been demonstrated. This study investigates how the use of the spectral content, instead of just ground-motion thresholds, can be used to improve the performance of proposed regional warning systems. We find that using the spectral content of the first few seconds after the P -wave arrival can provide timely warning for events closer to the target city than was possible with the threshold systems. It is further shown that for events less than 60 km from the target, any regional system needs to be complemented with an onsite one to provide a comprehensive EEW system.

Description: In this study we derive a spectral model describing the source, propagation and site characteristics of S waves recorded in central Italy. To this end, we compile and analyse a high-quality data set composed of more than 9000 acceleration and velocity waveforms in the local magnitude ( M l ) range 3.0–5.8 recorded at epicentral distances smaller than 120 km. The data set spans the time period from 2008 January 1 to 2013 May 31, and includes also the 2009 L'Aquila (moment magnitude M w 6.1, M l = 5.8) sequence. This data set is suitable for the application of data-driven approaches to derive the empirical functions for source, attenuation and site terms. Therefore, we apply a non-parametric inversion scheme to the acceleration Fourier spectra of the S waves of 261 earthquakes recorded at 129 stations. In a second step, with the aim of defining spectral models suitable for the implementation in numerical simulation codes, we represent the obtained non-parametric source and propagation terms by fitting standard parametric models. The frequency-dependent attenuation with distance r shows a complex trend that we parametrize in terms of geometrical spreading, anelastic attenuation and high-frequency decay parameter k. The geometrical spreading term is described by a piecewise linear model with crossover distances at 10 and 70 km: in the first segment, the spectral ordinates decay as 〈 tex – mathid = " IM 0001" 〉 r – 1.01 while in the second as 〈 tex – mathid = " IM 0002" 〉 r – 1.68 . Beyond 70 km, the attenuation decreases and the spectral amplitude attenuate as 〈 tex – mathid = " IM 0003" 〉 r – 0.64 . The quality factor Q ( f ) and the high-frequency attenuation parameter k , are 〈 tex – mathid = " IM 0004" 〉 Q ( f ) = 290 f 0.16 and k = 0.012 s, respectively, the latter being applied only for frequencies higher than 10 Hz. The source spectra are well described by 2 models, from which seismic moment and stress drops of 231 earthquakes are estimated. We calibrate a new regional relationship between seismic moment and local magnitude that improves the existing ones and extends the validity range to 3.0–5.8. We find a significant stress drop increase with seismic moment for events with M w larger than 3.75, with so-called scaling parameter  close to 1.5. We also observe that the overall offset of the stress-drop scaling is controlled by earthquake depth. We evaluate the performance of the proposed parametric models through the residual analysis of the Fourier spectra in the frequency range 0.5–25 Hz. The results show that the considered stress-drop scaling with magnitude and depth reduces, on average, the standard deviation by 18 per cent with respect to a constant stress-drop model. The overall quality of fit (standard deviation between 0.20 and 0.27, in the frequency range 1–20 Hz) indicates that the spectral model calibrated in this study can be used to predict ground motion in the L'Aquila region.

Description: In this study we derive a spectral model describing the source, propagation and site characteristics of S waves recorded in central Italy. To this end, we compile and analyse a high-quality data set composed of more than 9000 acceleration and velocity waveforms in the local magnitude ( M l ) range 3.0–5.8 recorded at epicentral distances smaller than 120 km. The data set spans the time period from 2008 January 1 to 2013 May 31, and includes also the 2009 L'Aquila (moment magnitude M w 6.1, M l = 5.8) sequence. This data set is suitable for the application of data-driven approaches to derive the empirical functions for source, attenuation and site terms. Therefore, we apply a non-parametric inversion scheme to the acceleration Fourier spectra of the S waves of 261 earthquakes recorded at 129 stations. In a second step, with the aim of defining spectral models suitable for the implementation in numerical simulation codes, we represent the obtained non-parametric source and propagation terms by fitting standard parametric models. The frequency-dependent attenuation with distance r shows a complex trend that we parametrize in terms of geometrical spreading, anelastic attenuation and high-frequency decay parameter k. The geometrical spreading term is described by a piecewise linear model with crossover distances at 10 and 70 km: in the first segment, the spectral ordinates decay as 〈 tex – mathid = " IM 0001" 〉 r – 1.01 while in the second as 〈 tex – mathid = " IM 0002" 〉 r – 1.68 . Beyond 70 km, the attenuation decreases and the spectral amplitude attenuate as 〈 tex – mathid = " IM 0003" 〉 r – 0.64 . The quality factor Q ( f ) and the high-frequency attenuation parameter k , are 〈 tex – mathid = " IM 0004" 〉 Q ( f ) = 290 f 0.16 and k = 0.012 s, respectively, the latter being applied only for frequencies higher than 10 Hz. The source spectra are well described by 2 models, from which seismic moment and stress drops of 231 earthquakes are estimated. We calibrate a new regional relationship between seismic moment and local magnitude that improves the existing ones and extends the validity range to 3.0–5.8. We find a significant stress drop increase with seismic moment for events with M w larger than 3.75, with so-called scaling parameter  close to 1.5. We also observe that the overall offset of the stress-drop scaling is controlled by earthquake depth. We evaluate the performance of the proposed parametric models through the residual analysis of the Fourier spectra in the frequency range 0.5–25 Hz. The results show that the considered stress-drop scaling with magnitude and depth reduces, on average, the standard deviation by 18 per cent with respect to a constant stress-drop model. The overall quality of fit (standard deviation between 0.20 and 0.27, in the frequency range 1–20 Hz) indicates that the spectral model calibrated in this study can be used to predict ground motion in the L'Aquila region.

Description: In this article, we report on the application of GFZ-Sentry software for decentralized on-site earthquake early warnings to a large data set of recordings collected by the Japanese K-NET and KiK-net seismic networks. The data set is composed of 3,225 three-component recordings from 24 seismic events. The magnitudes of the selected earthquakes fall into a range of M JMA  6.0–7.3 ( M w  5.7–6.9) and cover hypocentral (epicentral) distances between 5 and 107 km (2 and 103 km, respectively). The data have been coordinated in real time with velocity and displacement; in this manner, the peak ground displacement (Pd) within the first few seconds (to a maximum of 3 s) after the P -wave arrival in the vertical component is determined. This value is used to estimate the peak ground velocity (the median, along with the associated 16% and 84% confidence intervals) on the horizontal components using empirical relationships. Based on these values, the traffic light status (green, orange, red) is then determined following the methodology introduced by Parolai et al. (2015) , which uses three matrices to show the relationship between the expected ground motion and the possible damage (in terms of seismic intensity) that may arise. The performance of the software was evaluated without making ad hoc calibrations for the area or the selected thresholds and was found to be quite reliable. For example, in 90% of cases, assignment of the "red" status is followed by shaking that leads to a seismic intensity equal to or greater than V (very light potential damage). Additionally, all of the recordings leading to an intensity greater than VII (moderate to heavy damage) were correctly classified by a red status. Similarly, when considering all the declared green statuses, it is remarkable that only in 10% of cases was there a "missed alarm" (i.e., a green status is determined, but due to the level of observed shaking, it is later seen that the status should have been red).

Description: The accurate determination of stress drop, seismic efficiency and how source parameters scale with earthquake size is an important issue for seismic hazard assessment of induced seismicity. We propose an improved non-parametric, data-driven strategy suitable for monitoring induced seismicity, which combines the generalized inversion technique together with genetic algorithms. In the first step of the analysis the generalized inversion technique allows for an effective correction of waveforms for attenuation and site contributions. Then, the retrieved source spectra are inverted by a non-linear sensitivity-driven inversion scheme that allows accurate estimation of source parameters. We therefore investigate the earthquake source characteristics of 633 induced earthquakes (M w 2-3.8) recorded at The Geysers geothermal field (California) by a dense seismic network (i.e., 32 stations, more than 17.000 velocity records). We find a non-self-similar behavior, empirical source spectra that require an ω γ source model with γ 〉 2 to be well fit and small radiation efficiency η SW . All these findings suggest different dynamic rupture processes for smaller and larger earthquakes, and that the proportion of high frequency energy radiation and the amount of energy required to overcome the friction or for the creation of new fractures surface changes with earthquake size. Furthermore, we observe also two distinct families of events with peculiar source parameters that in one case suggests the reactivation of deep structures linked to the regional tectonics, while in the other supports the idea of an important role of steeply dipping faults in the fluid pressure diffusion.

Description: The Kivu rift is located in the bordering region of the Democratic Republic of Congo and Rwanda, in the western branch of the East African rift. Here, the active volcanoes Nyamulagira (the most active in Africa) and Nyiragongo (host to the largest persistent lava lake on Earth) threaten the city of Goma and neighboring agglomerations, and destructive earthquakes can also affect the region. Despite this high level of hazard, modern seismic monitoring infrastructure was lacking in the area until very recently, leaving many aspects about the volcanic activity and seismicity up to speculation. In order to remedy this unsatisfactory situation, the first dense real-time telemetered broadband seismic network, KivuSNet, was deployed in the region, with the first two stations in 2012/2013 followed by six additional ones in 2014. Since October 2015, a network of 13 stations is running in the Kivu rift, and with currently seven additional stations in the process of installation, this network is under continuous development. KivuSNet opens a new window for the seismological knowledge in this highly active rifting region. It allows for unprecedented insights into tectonic and volcanic seismicity, tremor patterns, and Earth structure as well as for sustainable real-time monitoring of the volcanoes. Together with the often collocated KivuGNet geodetic stations, KivuSNet closes a dramatic observational gap in this region. This article presents the key features of the network, discusses technical aspects, and provides an overview of first results obtained using the thus far acquired data, showing KivuSNet’s wide potential.

Description: A fundamental problem for site-specific ground-motion prediction, commonly required in seismic-hazard assessment, lies in the fact that ground-motion observations over long enough time periods are unavailable at the vast majority of sites. For this reason, most of the ground-motion prediction equations have been derived using observed data from multiple stations and seismic sources, and the standard deviation (sigma) is related to the statistics of the spatial variability of ground motion instead of temporal variability at a single site (ergodic assumption). In this paper, we explore the variability at single sites, decomposing sigma into different parts so that the various contributions to the variability can be identified and the standard deviation for empirical ground-motion prediction models quantified by removing the ergodic assumption. The analysis was conducted using three different data sets. Sigma obtained for Italy using the ergodic assumption is about 0.35log10 units ( Bindi, Pacor, et al. , 2011 ) and decreases to about 0.3 when single stations are considered (15% reduction). The values of single-station sigma obtained in this study for multiple-source data sets are rather stable, in the range 0.18–0.2log10 units, comparable to the findings of previous studies. The reduction of the epistemic uncertainty achieved through the restriction of the analysis to a particular seismic source leads to a sigma of about 0.25log10 units when the ergodic assumption is removed, suggesting that sigma at a particular site, due to a particular earthquake source, may reduce the sigma obtained for the Italian territory by Bindi, Pacor, et al. (2011) by about 30%. Online Material: Tables of ground-motion prediction equation coefficients, site terms, and event-corrected single-station standard deviations.

Description: The wealth of accelerometric recordings collected by the K-NET and KiK-net networks in Japan since 1996 provides a unique opportunity to improve our understanding of many important seismological research questions. Subsets of these data have been used for many case studies, most of them, however, not focusing specifically on the best practices for data selection and giving relatively little attention to the properties and peculiarities directly observable from the data. Yet for many applications, these steps are an important prerequisite for successful and reliable analysis. For this reason, we devote this article to the extraction of a large data set of surface and borehole recordings from the K-NET and KiK-net databases with strong emphasis on data quality and reliability. The final data set available for subsequent work consists of 78,840 records from 2201 earthquakes covering the Japan Meteorological Agency (JMA) magnitude range 2.7-8, observed at 1681 sites throughout Japan. We explain how this data set has been compiled, including automatic phase picking and relocation of events. We also present an overview of the general features of the data set, providing important information for subsequent analysis. Strong amplification effects at high frequencies are immediately visible on the surface recordings. Furthermore, there is a clear presence of downgoing waves in the borehole records, as deconvolution of borehole/surface recording pairs indicates.

Description: The S-wave attenuation characteristics beneath the Vrancea region in Romania are analyzed from the spectra (frequency range 0.5-20 Hz) of more than 850 recordings at 43 accelerometric stations of 55 intermediate-depth earthquakes (M 4.0-7.1) that occurred in the Vrancea seismogenic zone. The method commonly chosen for this type of investigation in the case of crustal earthquakes is the generalized inversion technique (GIT) (e.g., Andrews, 1986; Castro et al., 1990). Yet the Vrancea dataset is entirely different from common crustal datasets. Because of the strong clustering of the hypocenters within a very small focal volume, there are only few crossing ray paths from sources to receivers. As a consequence, inhomogeneities in the attenuation properties are not averaged out, which leads to unphysical results if the standard GIT approach is adopted. The problem is discussed qualitatively by performing tests with synthetic data and solved quantitatively by adapting the GIT technique in view of these peculiarities. With the optimally adapted inversion scheme, it is possible to unravel differences in the attenuation characteristics between two (or more) sets of stations. The results show that the attenuation of seismic waves is roughly comparable in the low frequency range (〈4-5 Hz) but stronger by up to an order of magnitude at higher frequencies within the Carpathian mountain arc as compared with the foreland area. Modeling this strongly frequency-dependent lateral variation of seismic attenuation by a significantly lower Q beneath Vrancea (1) provides a very good fit of observed strong-motion characteristics, (2) sheds new light on the distribution of intensities of the previous strong earthquakes, (3) will have strong implications for future hazard assessment, and (4) is fully compatible with structural models from deep seismic sounding, tomography, and teleseismic attenuation.