Journal of Seismology最新文献

The Isthmus of Panama experiences high seismic activity, having the potential for destructive earthquakes and serious risks to the population. Here, we present a new 1-D P-wave velocity model for Panama, which could be used for routine and accurate determination of earthquake locations, since Panama currently relies on a global velocity model. We used 23,178 P-wave arrival times from 1,672 selected seismic events between 2013 and 2022, recorded by 128 seismic stations across the country. To perform the analysis of P-wave arrival times, we utilized the Particle Swarm Optimization (PSO) method, which propagates multiple particles that explore the solution space to find the best possible velocity model. The new 1-D P-wave velocity model was obtained after multiple PSO runs, using the results of the previous run as a starting model until we find a model that best fits the seismic data. This model consists of 10 layers extending to a depth of 70 km, where the velocities range from 5.76 km/s at depths of 0-5 km to 8.27 km/s in the deepest layer. The station corrections, consistent with the geology of the Isthmus, allowed accurate relocation of earthquakes, achieving an epicentral distance error of ±3 km and a hypocentral distance error of ±6 km. These results are not only relevant for 3-D seismic tomography, but also valuable for seismic hazard and risk assessments in the Isthmus of Panama.

We document a complete seismic hazard study for mainland Norway and the Svalbard archipelago. The study is based on a Probabilistic Seismic Hazard Analysis (PSHA) method, and for the first time a new earthquake catalogue is presented publicly that covers Norway, Svalbard and the adjacent offshore regions. The catalogue is developed from an extensive analysis of historical earthquakes combined with more recent instrumental data with 33,864 reports between 1497 through 2018, and with magnitudes up to Mw 6.7. With this catalogue seismic hazard is computed for 10% exceedance in 475 years through a logic tree computation with 12 branches: two area-zonations, one zonation free branch and four GMPEs. These 12 branches were defined with the aim to reduce the model bias, i.e., to centre the model, and to capture the epistemic uncertainty of the results. While the conventional Vs30 reference velocity is usually around 800 m/s we have targeted a reference velocity of 1200 m/s, based on extensive documentation of Norwegian rock velocities. This has significant bearing on the calculated hazard and provides for results that better reflect the bedrock conditions in Norway. As a result of this, the predicted shaking intensities are lower than the values previously reported in the (1998) national building code. In the Supplementary Information we have provided a brief overview of the seismotectonic setting, some tests that further demonstrate the uncertainty in our hazard estimates, a model for H/V ground-motion response spectra, examples of the sensitivity to the bedrock reference velocity and a comparison between the present study and the ESHM20 results.

Scenario-based disaster prevention, preparedness, and response are developing trends in contingency management. Human walking states in simulated scenarios are studied in this work on the Intelligent Seismic Scenario Experience vibration table. The vibration table adopts a comprehensive method that considers the seismic characteristics, building structure, and dynamic performance of the equipment to ensure the real experience of the human body in earthquake scenarios. The equipment can conduct a full chain of earthquake scenario simulation: based on source physical processes, seismic wave propagation paths, site conditions, and building structures. Seismic scenarios of different sources, site conditions, floors, and response spectra are used to study human perception and reactions during walking. The experimental results are consistent with the description of human action at the current intensity scales of China, Europe, the USA, and Japan. Research shows that the PGV has a significant correlation with the impact of ground motion on human walking and can be used as a key indicator to determine the intensity. However, the correlation between the PGA and the impact on human walking is not strong. The predominant period of the response spectrum is also a key factor affecting human walking states in earthquake scenarios. Vibrations with periods between 0.5 s and 1.9 s have the greatest impact on the walking state, which is usually caused by high-rise buildings and deep soil sites and needs special attention. Moreover, horizontal vibrations dominate impacts on human walking in earthquake scenarios where the PGA ≤ 300 Gal. The results of this experiment can be applied to the study of the relationships between the macroscopic intensity and instrument intensity, the preparation of earthquake intensity scales, guidance on earthquake emergency avoidance actions, and the popularization of earthquake science.

A re-assessment of the macroseismic intensity data was conducted for the second-largest instrumentally recorded event in the Netherlands: the 20 November 1932 Uden earthquake. This event was felt across the Netherlands, Belgium and Germany. Intensity values on the EMS98 scale were assigned based on original data (reports/enquiries/letters), manually for the Netherlands and automatically for Belgium, with existing German data added for completeness. The updated dataset was used to calculate macroseismic location and magnitude using the Bakun and Wentworth (1997, 1999) algorithm. To capture epistemic uncertainty, four newly calibrated intensity attenuation relations were applied and their results averaged. The results using only intensity data from the Netherlands provided stable solutions within the region of maximum observed intensity (I_max = VII). However, including Belgian and German data shifted the source location outside this region and becomes less reliable, likely due to differences in data collection methods and local/regional site effects. Comparison of confidence levels from the BW method with bootstrap modelling showed that almost all bootstrap results fall within the 50% confidence region. A more realistic estimate for the location uncertainty was derived from the bootstrap analysis. The revised source parameters are 51.63°N and 5.61°E ± 7.3 km for the source location and M_L = 5.1 ± 0.3 (M_S 4.9 ± 0.3) for the magnitude.

The significant duration is a crucial intensity measure for earthquake-resistant design and seismic hazard assessment (SHA). The Sichuan-Yunnan region is characterized by a high level of seismic activity and possesses the most concentrated network of seismic stations in China. The ground motion prediction equation (GMPE) is the predominant approach to estimating significant durations. The existing prediction equations for the significant duration are not well-suited for the Sichuan-Yunnan region. This study used data from the National Strong Motion Observation Network System (NSMONS) of China in this region to develop prediction equations for significant durations of D_S5-75 and D_S5-95. The equations took into account variables including moment magnitude (M_w), fault distance (R_rup), average shear wave velocity of 30 m on the soil profile (V_S30), and depth to the top of the rupture (Z_tor). Our database has a singular instance of the Wenchuan earthquake with M_w > 7. The restricted data complicates the calibration of our model for events with M_w > 7. Therefore, we suggest the equations be valid in the Sichuan-Yunnan region for M_w between 4.2 and 7.0, R_rup from 0 to 300 km, and V_S30 values ranging from 139 to 900 m/s.

There are only a few worldwide cases where distant earthquakes have caused damage. One such example is the municipality of Centro located in Tabasco, Southeast Mexico, approximately 360 km from the Mesoamerican trench, where a strong ground shaking was felt during the M_w8.2 earthquake of September 08, 2017. In this study, for 20 sites shear-wave velocity profiles were determined using Multichannel Analysis of Surface Wave and V_P profiles using Seismic Refraction techniques. Also V_S30 (shear-wave velocity up to a depth of 30 m) values were obtained for the same sites. The distribution of the V_S30 values in the study area varied from 120 m/s to 570 m/s and it was observed that sites where damage to buildings were reported lie near areas with V_S30 < 270 m/s. Additionally, the transfer functions of the 20 sites were estimated using the Thomson-Haskell method. The fundamental frequencies (f₀) obtained through transfer functions had values varying from 0.9 ≤ f₀ ≤ 2.0 Hz. These transfer functions were convolved with the signal that represents the record in the bottom of the soil column in the study area to obtain synthetic accelerograms in the municipality of Centro. The only accelerograph station located in the study area (VHSA station) was used as a reference site. The horizontal-to-vertical spectral ratio of the VHSA location was used for site characterization to assess the effects of regional events. The study concludes that several factors contribute to the susceptibility of Centro municipality to distant seismic events. These factors include low shear-wave velocity (V_s), low fundamental frequency (f₀), local site conditions, the presence of buildings on former lake zones, low seismic wave attenuation, and the regions’ overall vulnerability to regional earthquakes.

Considering 1173 recordings of 35 stations from 67 aftershocks of the 12 May 2008 Wenchuan M_w (moment magnitude) 7.9 earthquake, we investigate site amplifications and their variations in the Longmenshan region. Site responses of 35 stations are analyzed using the coda-wave and S-wave methods. For these methods, the site amplifications are computed using a generalized inversion method. Generally, the calculated site amplifications from coda-wave and S-wave inversions are not very large, which is probably resulted from the special geology conditions in the Longmenshan region and most peak ground acceleration values of recordings less than 0.1 g. Because coda-wave amplitudes attenuate slowly along the propagation path, the site amplifications computed through the coda-wave inversion are relatively larger than those through the S-wave inversion. The comparison of the intraevent and interevent residuals of the coda-wave inversion with those of S-wave inversion demonstrates that the coda-wave inversion is more reasonable to calculate the site amplifications of the Longmenshan region. Moreover, for the Longmenshan region, the averaged site amplifications of the young geology sediments are not considerably different from those of the old geology rocks in some frequencies. If the sites of stations are classified by the National Earthquake Hazard Risk Reduction Program (NEHRP) site classifications, the averaged site coefficients of the Longmenshan region are usually smaller than the corresponding NEHRP site coefficients.

Macroseismic intensity data have been gathered continuously in Belgium since the start of the twentieth century. In this study, we review the applied survey practices used over the previous century: from small-scale ad-hoc improvised surveys to the mass distribution of collective questionnaire versions sent to local authorities. The variety of survey methodologies causes a high degree of heterogeneity within the resulting macroseismic data, increasing the uncertainty in macroseismic applications that rely on homogeneous data. We thoroughly re-evaluated the original source material and intensity scale conversions in order to create the Belgian Traditional Macroseismic (BTM) database, a comprehensive compilation of 20th-century macroseismic data in which all sources are properly referenced. The BTM database currently consists of 23,950 intensity data points (IDPs) on the European Macroseismic Scale for 80 felt earthquakes, ranging from 2.4 to 5.8 on the local magnitude (M_L) scale. Each IDP is provided with a source type and each earthquake is attributed a data quality parameter that indicates the level of uncertainty associated with its IDP source quality. The publication of the BTM database facilitates the use of Belgian macroseismic data for a variety of seismological purposes and allows us to summarize the overall seismic impact on Belgium for the duration of a century. Like in many other parts of the world, traditional procedures have practically been discontinued in Belgium in favour of an online enquiry. However, the potential for high-quality data following a traditional survey methodology is still large and we strongly recommend its continuation.

Peak ground velocity (PGV) is a crucial ground motion parameter correlating with earthquake damage. How to quickly predict PGV at a target site is a core issue of earthquake early warning (EEW) system. By using the embedded characteristics in ground motion sequence, a Long Short-Term Memory (LSTM) networks-based onsite PGV prediction model (LSTM-PGV) is proposed in this paper. The LSTM-PGV model consists of three layer of LSTM and one fully connected layer, and the inputs are sequence features of energy-related, amplitude-related, period-related and distance-related P-wave parameters. The performance of the LSTM model on training, validation and test datasets indicates that the model has good generalization capability, and the predicted PGV and observed PGV can meet the 1:1 relationship in general. Compared with Pd-PGV model, a logarithmic linear regression model where Pd is the peak vertical displacement of the first 3 s P-waves, and LSTM-Pd-PGV model, a LSTM-based model with Pd as the sole input sequency feature where Pd is the maximum vertical displacement continuously changing over time, the proposed model predicts PGV more accurately and stably. Furthermore, the issue of underestimation of PGV for larger earthquakes is alleviated in LSTM-PGV model by using longer length of sequence input. The LSTM model is tested with one off-shore earthquake and one inland earthquake in Japan. The results show that the standard deviation of prediction residual goes from 0.417 at sequence length of 3 s to 0.309 at sequence length of 10 s for the off-shore event, and for the inland event the standard deviation decreases from 0.357 to 0.267 at corresponding sequence length. The prediction timeliness measured by lead time, defined as the time interval between the moment when the observed PGV reaches 17.3 cm/s and the moment when the predicted PGV reaches the same threshold, is also discussed for different magnitudes and hypocentral distances. We believe the proposed LSTM model has promising potential in onsite EEW system for providing accurate and timely PGV prediction.

Seismic recorders register vibrations from all possible sources. Even though the purpose of the seismic instrument is, usually, to record ground motions coming from tectonic sources, other sources such as vehicles can be recorded. In this study, a machine learning model is developed by using a convolutional neural network (CNN) to separate three different classes which are earthquakes, vehicles, and other noises. To do that vehicle signals from various accelerometric stations from Italy are visually detected. Together with the vehicle signals noise and earthquake information coming from Italy are used. Inputs of the database are 10 s long seismic traces along with their frequency content from three channels of the seismic recorder. CNN model has an accuracy rate of more than 99 % for all classes. To understand the capabilities of the model, seismic traces with vehicles and earthquakes are given as input to the model which the model successfully separates different classes. In the case of the superposition of an earthquake and a vehicle, the model prediction is in favor of the earthquake. Moreover, earthquake signals from various databases are predicted with more than 90 % accuracy.