Journal of Seismology最新文献

Crust and uppermost mantle shear-wave velocity in northeast India is elaborated based on the non-linear inversion of the Love wave group velocity data. This is the first time Love wave and SH wave tomography have been carried out in this region. We investigated the vertical and lateral variation of the SH wave velocity and the crustal thickness. We used the data from 26 broadband seismic stations that recorded the earthquakes from 2001 to 2015. Group velocity dispersion curves of 228 moderate and higher magnitude earthquakes are inverted to obtain tomography images of group velocities at periods 6–60 s. Low velocity is confined to the Bengal Basin (BB) and the Indo-Burma Ranges (IBR) at short periods. In contrast, high velocity is present in the Shillong Plateau (SP), Mikir Hills (MH), and Assam syntaxis. Higher velocity indicates an upward buckled crustal structure. The Tibetan plateau has a low velocity, indicating a thick crust and partial melt in the middle to lower crust. The inverted SH wave velocity models also show similar patterns. Based on such analysis, variation in the Moho depth below the study area could be estimated. Moho depth below SP and MH varies from 35 to 45 km. Moho depth below BB varies between ~ 28 to ~ 32 km. The uppermost crust in BB indicates a thick sediment deposit of ~ 15- ~ 20 km. The Moho depth of NE India gradually increases towards the Eastern Himalayas, Lhasa block, and IBR. Steeply dipping Moho below IBR indicates the subduction of the Indian plate below the Burma plate. The Tibet and Lhasa blocks show a crustal thickness of ~ 85 km, which is the maximum compared to other parts of the study area. Highly variable shear wave crustal structures, added with low/high-velocity anomalies at different depths, give new insights into the heterogeneous and complex geotectonics.

The mountainous domain of the Saharan Atlas (northern Algeria) has been long considered as a region of very low seismic activity. However, recent instrumental data reveals significant seismic activity over the last decades. This led to reconsidering this region and re-evaluating its seismic activity. This work investigates the historical seismicity of the western part of the Saharan atlas, from 1867 to 1987. This study consists of compiling a seismic catalogue of the region, based on old and contemporary catalogues, documents, contemporary journalistic sources, and paper seismograms available in the CRAAG archives. A total of 31 seismic events have been gathered, the most important events have been reviewed in detail, including macroseismic intensity maps. In order to unify the Earthquake catalogue for Algeria (ECA) to a common magnitude scale, the surface wave magnitude (Ms) and/or maximum observed intensity (Imax) were determined for each seismic event. The results of this work are intended to be included in the Algerian Macroseismic Database (AMD) and the ECA, to enhance the understanding of the seismic patterns and the tectonic activity in the region.

Northeast India is a tectonically complex region exhibiting the collision between the Indo-Eurasian tectonic plates to the north and subduction between the Indo-Burma plate along its eastern periphery. This study incorporated twenty permanent and three temporary broadband seismic stations’ ambient noise data to decipher the region's crustal structure. We cross-correlate ambient noise records to retrieve group velocity maps for periods ranging from 6 to 25 s. With the help of short-period tomograms (< 10 s), we identify the contact between the undifferentiated granites and Eocene rocks on the Shillong plateau, as well as major geomorphological features of the Bengal basin, Indo-Burma ranges, and the Brahmaputra River valley. For long periods (> 10 s), our results reveal that the basement rocks of the Shillong massif are separated from the sediments of the Bengal basin by a northwardly dipping Dauki thrust fault. The dip of the Dauki fault is constant up to ~ 16 s, and it becomes steeper towards north at higher depths. Our group velocity maps suggest that the tectonic stresses may be responsible for the uplift and expansion of the Shillong massif. Furthermore, the crustal thickness of the Bengal basin increases gradationally towards its eastern boundary, indicating the occurrence of oblique subduction at the Indo-Burmese arc. Again, the 25⁰ latitude differentiates slower and faster velocities in the Indo-Burma region from north to south, which aligns with the thicker crust in the southern region of the Indo-Burma ranges.

Accurate and precise earthquake locations are essential for seismic catalogues and unbiased geophysical research, but their precision is often limited by assumptions about the velocity model used. The construction of a 3D model can improve precision compared to a previously available 1D model, although a quantitative assessment of this impact has been lacking. To address this, we conducted a comparison between earthquake locations in the Alto Tiberina (AT) study area, Northern Apennines of Italy, obtained using three elastic models: (1) a homogeneous half-space model, (2) a 1D elastic model developed for the broader Umbria-Marche region and (3) a 3D elastic model developed ad-hoc for the AT study area, where a dense seismic network operates. Using a Markov chain Monte Carlo algorithm, we evaluated the precision of the hypocentral parameters for each model. Multiple events were analysed at different positions within the study area to test the models performance across varying seismic network configurations. Our results quantify the improvement in earthquake locations, particularly hypocentral depth, achieved with the 3D velocity model, relative to those obtained using the homogeneous and 1D models. For events located at the centre of the seismic network, depth uncertainty decreased to one-third, while events at the network's periphery showed reductions of up to 90% from the homogeneous to the 3D model. Epicentral uncertainties also decreased: by 50% for events located at the border and by up to 90% for events outside the network shifting from the homogeneous to the 3D model. This quantitative analysis underscores the advantages of using a 3D model for earthquake location, particularly in improving hypocentral and epicentral precision across different network positions.

Earthquake catalogue is crucial for seismicity modeling and seismic hazard assessment. That’s why it should be updated, homogeneous and complete. This study presents a comprehensive compilation of the Tunisian seismicity catalogue. Our approach involved integrating instrumental and historical data, ensuring comprehensive coverage. Seismological data, including epicenters, origin times, location (longitude & latitude), and amplitudes, were collected from the Tunisian Seismic Network database which is installed by the National Institute of Meteorology for accurate earthquakes. Additionally, data from three other earthquake catalogues were incorporated which are data from the International Seismological Centre, the Euro-Mediterranean Seismological Centre, and the Algerian Centre for Research in Astronomy, Astrophysics and Geophysics. The updated earthquake catalogue spans from 412 to 2022, encompassing 5,932 events, with the moment magnitude scale established through a new magnitude homogenization equation specifically developed for the Tunisian region. This catalogue is the most updated one for Tunisian seismicity due to the exploitation of recent data detected by modern installed stations by the INM. In fact, it is identified by a notable increase in the total number of events nearly doubled compared to the previous elaborated catalogues. The generated b-value attends 0.83 which is a suitable one for a region with low to moderate seismicity such as Tunisia. Through a parametric study, we compared the computed b-value with those calculated previously and identified the factors and parameters that have the greatest influence on it which are the magnitude conversion and the amount of data treated.

The 2023 Ms 6.2 Jishishan earthquake generated intense near-field ground motions with a peak acceleration of 1.1 g, far exceeding regional model predictions. The physical mechanism behind its high-energy characteristics, however, remained unclear. Using dense strong ground motion records, this study systematically reveals high-frequency ground motion anomalies and a pronounced hanging-wall effect. Based on the relationship between ground motion and source parameters under the elliptical model, we attribute this to a high stress drop (13.36 MPa) and identify significant source radiation modulation. We develop a physics‐based ground motion prediction model that integrates seismic moment and stress drop derived from P‐wave records. This method overcomes the conventional reliance on magnitude alone, capturing high‐frequency dynamic source features, and achieves significantly improved prediction accuracy. Our results highlight the critical need to incorporate dynamic source parameters into seismic hazard assessment, especially for moderate‐to‐strong earthquakes in high‐stress tectonic settings.

In this study, Machine Learning (ML)-based models were developed for the estimation of Pseudo Spectral Acceleration (PSA), a key parameter for identifying strong ground motion in seismic hazard assessments. A large dataset was generated using 1975 earthquake records ( 3.0 ≤ M ≤ 6.8), sourced from 63 strong ground motion stations in different regions of Türkiye. This dataset includes a total of 150,552 samples, which we divided into three groups based on their period (T) values. These 3 datasets were modeled by using 19 different ML algorithms and validated by using tenfold cross-validation. The performances of the models were compared by using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared (R²) criteria. The models trained with the Ensemble Bagged Trees and Fine Tree algorithms were determined to be the best models according to test criteria. The Ensemble Bagged Trees model performed best in the T = 0.05–1.0 s range (R² = 0.90), while the Fine Tree model performed best in T = 1.0–1.9 s (R² = 0.94) and T = 2.0–4.0 s (R² = 0.89). Residual analyses showed that the prediction errors were randomly distributed around the zero line. Moreover, the PSA results of trained ML Models yielded results closer to the observed PSA curve. These results indicate that, within the compiled dataset, the evaluated ML models can predict PSA with lower errors and may serve as a complementary approach to GMPEs in future applications.

In Mexico, the Michoacán region and its adjacent areas are characterized by complex geodynamic interactions that result in significant seismic activity. Historical earthquakes, such as the Mw 8.1 event on September 19, 1985, and the Mw 7.7 event on September 19, 2022, as well as its tectonic activity, indicate an important seismic hazard in this area. The region is divided into three seismotectonic zones: one of subduction-related seismicity, another with intraplate seismicity of the subducted Cocos plate, and the third having shallow seismicity within the Transmexican Volcanic Belt, associated with the North American plate. This study focuses on the statistical analysis of seismicity and earthquake recurrence rates to assess seismic hazard, addressing spatial variations in seismic productivity and stress regimes. In this work, seismic hazard is understood as the probabilistic assessment of earthquake occurrence—severity (e.g., magnitude or intensity) and its temporal and spatial measures—derived from observational data (with no ground-motion modeling). A robust statistical-numerical framework was implemented, incorporating recent approaches such as the b-absolute and a-positive estimators. The seismic hazard assessment in the seismotectonic zones provides improved constraints for assessing the potential occurrence of significant earthquakes (Mw ≥ 4) and offers valuable input for the development of risk mitigation strategies.

The inherent unpredictability of seismic activity has long impeded the development of precise earthquake forecasting models, rendering traditional statistical methodologies insufficient for mitigating disaster impact. This research presents a novel Deep Learning based Generative framework that synthesizes cloud computing alongside IoT for spatial feature extraction, temporal pattern recognition and augmenting training datasets through synthetic seismic event data generation for low-resource seismic regions. Through the use of empirical seismic multi-modal data and computational models within a cloud hosted infrastructure, the proposed research enabled real-time and scalable earthquake prediction features with valuable tests revealing a crucial 15–25% improved prediction accuracy over the traditional methods, further reducing the significant false positives and improved alert response times. This research redefines the methods of earthquake forecasting creating a stage for a versatile GenAI-based predictive opportunities that can be generalized to broader disaster resilience events, such as tsunami, wildfire and landslide based early warning systems.