Search Results (624)

Porosity prediction from seismic data is of significance in reservoir property assessment, reservoir architecture delineation, and reservoir model building. However, it is still challenging to use traditional model-driven methodology to characterize carbonate reservoirs because of the highly nonlinear mapping relationship between porosity and elastic properties. To address this issue, this study proposes an advanced spatiotemporal deep learning neural network for porosity prediction, which uses the convolutional neural network (CNN) structure to extract spatial characteristics and the bidirectional gated recurrent unit (BiGRU) network to gather temporal characteristics, guaranteeing that the model accurately captures the spatiotemporal features of well logs and seismic data. This method involves selecting sensitive elastic parameters as inputs, standardizing multiple sample sets, training the spatiotemporal network using logging data, and applying the trained model to seismic elastic attributes. In blind well tests, the CNN–BiGRU model achieves a 54% reduction in the root mean square error and a 6% correlation coefficient improvement, outperforming the baseline models and traditional nonlinear fitting (NLF). The application of the proposed method to seismic data indicates that the model yields a reasonable porosity distribution for tight carbonate reservoirs, proving the strong generalization ability of the proposed model. This method compensates for the limitations of individual deep learning models by simultaneously capturing the spatial and temporal components of data and improving the estimation accuracy, showing considerable promise for accurate reservoir parameter estimation. Full article

The Himalayan belt, formed due to the Cenozoic convergence between the Eurasian and Indian craton, acts as a storehouse of large amounts of strain, resulting in large earthquakes from the Western to the Eastern Himalayas. Glaciers also occur over a major portion of the high-altitude Himalayan region. The impact of earthquakes can be easily studied in the plains and plateaus with the help of well-distributed seismogram networks and these regions’ accessibility is helpful for field- and lab-based studies. However, earthquakes triggered close to high-altitude Himalayan glaciers are tough to investigate for the impact over glaciers and glacial deposits. In this study, we attempt to understand the impact of earthquakes on and around Himalayan glaciers in terms of vertical displacement and coherence change using space-borne synthetic aperture radar (SAR). Eight earthquake events of various magnitudes and hypocenter depths occurring in the vicinity of Himalayan glacial bodies were studied using C-band Sentinel1-A/B SAR data. Differential interferometric SAR (DInSAR) analysis is applied to capture deformation of the glacial surface potentially related to earthquake occurrence. Glacial displacement varies from −38.9 mm to −5.4 mm for the 2020 Tibet earthquake (M_w 5.7) and the 2021 Nepal earthquake (M_w 4.1). However, small glacial and ground patches processed separately for vertical displacements reveal that the glacial mass shows much greater seismic displacement than the ground surface. This indicates the possibility of the presence of potential site-specific seismicity amplification properties within glacial bodies. A reduction in co-seismic coherence around the glaciers is observed in some cases, indicative of possible changes in the glacial moraine deposits and/or vegetation cover. The effect of two different seismic events (the 2020 and 2021 Nepal earthquakes) with different hypocenter depths but with the same magnitude at almost equal distances from the glaciers is assessed; a shallow earthquake is observed to result in a larger impact on glacial bodies in terms of vertical displacement. Earthquakes may induce glacial hazards such as glacial surging, ice avalanches, and the failure of moraine-/ice-dammed glacial lakes. This research may be able to play a possible role in identifying areas at risk and provide valuable insights for the planning and implementation of measures for disaster risk reduction. Full article

This study evaluates urban resilience to earthquakes in Daegu Metropolitan City, South Korea, by analyzing outdoor evacuation sites through a dual-axis matrix framework to provide feasible solutions for enhancing urban resilience. Evacuation capacity was assessed by use of resident and de facto population data, while Geographic Information System (GIS) network analysis identified evacuation-feasible and evacuation-infeasible areas. The matrix categorizes areas along two axes: capacity (x-axis) and evacuation-infeasible areas (y-axis), facilitating targeted improvement strategies. Findings reveal that only 54 of 139 census blocks possess sufficient capacity and no evacuation-infeasible areas. For areas with adequate capacity but extensive infeasible areas, redistributing evacuation sites is recommended to improve accessibility. Areas with limited capacity but no infeasible areas require additional outdoor evacuation sites to accommodate the population. In regions constrained by both capacity and accessibility, establishing new evacuation sites within infeasible areas is essential. For critically low-capacity areas without infeasible areas, multi-use spaces, such as disaster prevention parks, are desirable to address evacuation needs. Lastly, areas lacking both capacity and accessibility urgently require new evacuation sites concentrated in infeasible areas. By simplifying complex variables into a capacity–accessibility matrix, this study integrates population dynamics, spatial accessibility, and site capacity, offering implementable solutions for earthquake preparedness in densely populated urban settings. Additionally, this approach supports urban planning efforts to mitigate seismic damage and enhance urban sustainability. Full article

In recent years, although a variety of deep learning models have been developed for magnitude estimation, the complex and variable nature of earthquakes limits the generalizability and accuracy of these models. In this study, we selected the waveform data of the Japan earthquake. We applied four deep learning techniques (MagNet combined with bidirectional long- and short-term memory network Bi-LSTM, DCRNN with deepened CNN layers, DCRNNAmp with the introduction of a global scale factor, and Exams with a multilayered CNN architecture) for real-time magnitude estimation. By comparing the estimation errors of each model in the first 3 s after the earthquake, it is found that the DCRNNAmp performs the best, with an MAE of 0.287, an RMSE of 0.397, and an R² of 0.737 in the first 3 s after the arrival of the P-wave, and the inclusion of S-wave seismic-phase information is found to significantly improve the accuracy of the magnitude estimation, which suggests that S-wave seismic-phase waveform features can enrich the model’s understanding of the relationship between the seismic phases. It shows that S-wave phase waveform features can enrich the model’s knowledge of the relationship between seismic fluctuations and magnitude. The epicentral distance positively correlates with the magnitude estimation, and the model can converge faster with the improved signal-to-noise ratio. Despite the shortcomings of model design and opaque internal mechanisms, this study provides important evidence for deep learning in earthquake estimation, demonstrating its potential to improve the accuracy of on-site earthquake early warning (EEW) systems. The estimation capability can be further improved by optimizing the model and exploring new features. Full article

Distributed Acoustic Sensing (DAS) has emerged as a groundbreaking technology in seismology, transforming fiber-optic cables into dense, cost-effective seismic monitoring arrays. DAS makes use of Rayleigh backscattering to detect and measure dynamic strain and vibrations over extended distances. It can operate using both pre-existing telecommunication networks and specially designed fibers. This review explores the principles of DAS, including Coherent Optical Time Domain Reflectometry (COTDR) and Phase-Sensitive OTDR (ϕ-OTDR), and discusses the role of optoelectronic interrogators in data acquisition. It examines recent advancements in fiber design, such as helically wound and engineered fibers, which improve DAS sensitivity, spatial resolution, and the signal-to-noise ratio (SNR). Additionally, innovations in deployment techniques include cemented borehole cables, flexible liners, and weighted surface coupling to further enhance mechanical coupling and data accuracy. This review also demonstrated the applications of DAS across earthquake detection, microseismic monitoring, reservoir characterization and monitoring, carbon storage sites, geothermal reservoirs, marine environments, and urban infrastructure surveillance. The study highlighted several challenges of DAS, including directional sensitivity limitations, vast data volumes, and calibration inconsistencies. It also addressed solutions to these problems, such as advances in signal processing, noise suppression techniques, and machine learning integration, which have improved real-time analysis and data interpretability, enabling DAS to compete with traditional seismic networks. Additionally, modeling approaches such as full waveform inversion and forward simulations provide valuable insights into subsurface dynamics and fracture monitoring. This review highlights DAS’s potential to revolutionize seismic monitoring through its scalability, cost-efficiency, and adaptability to diverse applications while identifying future research directions to address its limitations and expand its capabilities. Full article

The fault fracture body, consisting of faults, fracture zones, cracks, and the matrix, plays a crucial role in controlling oil and gas accumulation. Understanding its spatial distribution and analyzing the in situ stress field are essential for optimizing well design and fracturing operations. This study integrates geological, logging, and seismic data, and employs advanced techniques such as ant tracking to establish a skeletal model of the fault fracture body. Reverse modeling and optimization reconstruction are used to construct a three-dimensional geomechanical model of the fracture system. Machine learning techniques, specifically a back propagation (BP) neural network, are utilized to invert the boundary conditions of the study area. Finite element numerical simulation software is then applied to model the three-dimensional in situ stress field under coupled flow–solid interaction. The reservoirs in the study area are characterized by a dense structure, low porosity, and low permeability. The results indicate that the maximum horizontal principal stress in the fault fracture reservoir ranges from 68.0 to 72.8 MPa, while the minimum horizontal principal stress ranges from 58.2 to 63.1 MPa. The stress at fractures is lower than that in the surrounding matrix, and stress concentrations occur at both ends of the faults. The in situ stress field exhibits distinct subarea characteristics, with significant stress reductions across fault fractures and directional deflections at faults. These findings provide valuable insights for improving reservoir development efficiency and optimizing well operations in similar geological settings. Full article

Addressing the noise in seismic signals, a prevalent challenge within seismic signal processing, has been the focus of extensive research. Conventional algorithms for seismic signal denoising often fall short due to their reliance on manually determined feature functions and threshold parameters. These limitations hinder effective noise removal, resulting in suboptimal signal-to-noise ratios (SNRs) and post-denoising waveform distortion. To address these shortcomings, this study introduces a novel denoising approach leveraging a DnCNN network. The DnCNN framework, which integrates batch normalization with residual learning, is adept at swiftly identifying and eliminating noise from seismic signals through its residual learning capabilities. To assess the efficacy of this DnCNN-based model, it was rigorously tested against a curated test set and benchmarked against other denoising techniques, including wavelet thresholding, empirical mode decomposition, and convolutional auto-encoders. The findings demonstrate that the DnCNN model not only significantly enhances the SNR and correlation coefficient of the processed seismic signals but also achieves superior noise reduction performance. Full article

Peak Ground Acceleration (PGA) is a measure of the maximum ground shaking intensity during an earthquake. The estimation of PGA in areas affected by earthquakes is a fundamental task in seismic hazard assessment and emergency response. This paper presents an automated service capable of rapidly calculating the PGA’s values in regions impacted by seismic events and publishing its results on an interactive website. The importance of such a service is discussed, focusing on its contribution to timely response efforts and infrastructure resilience. The necessity for automatic and real-time systems in earthquake-prone areas is emphasized, enabling decision-makers to assess damage potential and deploy resources efficiently. Thanks to a collaboration agreement with the Civil Protection Department, we are able to acquire accelerometric data from the Italian National Accelerometric Network (RAN) in real time at the monitoring center of the Osservatorio Vesuviano. These data, in addition to those normally acquired by the INGV network, enable us to utilize all available accelerometric data in the Campi Flegrei area, enhancing our capacity to provide timely and accurate PGA estimates during seismic events in this highly active volcanic region. Full article

The deployment density of surface sensors can significantly impact the accuracy of subsurface shallow seismic field energy inversion. With finite budget constraints, it is often not feasible to deploy a large number of sensors, resulting in limited seismic signal acquisition that hinders accurate inversion of the shallow subsurface explosions. To address the challenge of insufficient sensor signals needed for inversion, we conducted a study on a subsurface shallow wavefield data interpolation method based on transfer learning. This method is designed to increase the overall signal acquisition by interpolating signals at target locations from limited measurement points. Our research employs neural networks to interpolate real seismic data, supplementing the sampled signals. Given the lack of extensive samples from actual data collection, we devised a training approach that combines numerically simulated signals with real collected signals. Initially, we performed conventional interpolation training using a deep interpolation network with complete synthetic gather images obtained from numerical simulations. Subsequently, the feature extraction part was frozen, and the interpolation network was transferred to real datasets, where it was trained using incomplete gather images. Finally, these incomplete gather images were re-input into the trained network to obtain interpolated results at the target locations. Our study demonstrates the superiority of our method by comparing it with two other interpolation networks and validating the effectiveness of transfer learning through four sets of ablation experiments in the actual test. This method can also be applied to other shallow geological structures to generate a large number of seismic signals for energy inversion. Full article

This study presents the results of an experimental investigation conducted on a 2:3 scale model of a two-story stone masonry building. We tested the model on the UniKORE L.E.D.A. lab shake table, simulating the Mw 6.3 earthquake ground motion that struck L’Aquila, Italy, on 6 April 2009, with progressively increasing peak acceleration levels. We installed a network of accelerometric sensors on the model to capture its structural behaviour under seismic excitation. Medium-to lower-cost MEMS accelerometers (classes A and B) were compared with traditional piezoelectric sensors commonly used in Structural Health Monitoring (SHM). The experiment assessed the structural performance and damage progression of masonry buildings subjected to realistic earthquake inputs. Additionally, the collected data provided valuable insights into the effectiveness of different sensor types and configurations in detecting key vibrational and failure patterns. All the sensors were able to accurately measure the dynamic response during seismic excitation. However, not all of them were suitable for Operational Modal Analysis (OMA) in noisy environments, where their self-noise represents a crucial factor. This suggests that the self-noise of MEMS accelerometers must be less than 1 µg/√Hz, or preferably below 0.5 µg/√Hz, to obtain good results from the OMA. Therefore, we recommend ultra-low-noise sensors for detecting differences in the structural behaviour before and after seismic events. Our findings provide valuable insights into the seismic vulnerability of masonry structures and the effectiveness of sensors in detecting damage. The management of buildings in earthquake-prone areas can benefit from these specifications. Full article

The use of the Visibility Graph Algorithm (VGA) has proven to be a valuable tool for analyzing both real and synthetic seismicity series. Specifically, VGA transforms time series into a network representation in which structural properties such as node connectivity, clustering, and community structure can be quantitatively measured, thereby revealing underlying correlations and dynamics that may remain hidden in traditional linear or spectral analyses. The time series transformation into complex networks with VGA provides a new approach to analyze seismic dynamics, allowing scientists to extract trends and behaviors that may not be possible by classical time-series analysis. On the other hand, many studies attempt to find viable trends in order to identify preparation mechanisms prior to a strong earthquake or to analyze the aftershocks. In this work, the seismic activity of Southern California Earthquake was analyzed focusing only on the significant earthquakes. For this purpose, seismic series preceding and following each earthquake were constructed using a windowing method with different overlaps and the slope of the connectivity (k) versus magnitude (M) graph (k-M slope) and the average degree were computed from the mapped complex networks. The results revealed a significant decrease in these parameters after the earthquake, due to the contribution of the aftershocks from the main event. Interestingly, the study was extended to synthetic seismicity series and the same behavior was observed for both k-M slope and average degree. This finding suggests that the spring-block model reproduces a relaxation mechanism following a large-magnitude event like those of real seismic aftershocks. However, this conclusion contrasts with conclusions drawn by other researchers. These results highlight the utility of VGA in studying events that precede and follow major earthquakes. This technique may be used to extract some useful trends in seismicity, which could eventually be employed for a deeper understanding and possible forecasting of seismic behavior. Full article

Unreinforced masonry buildings are highly vulnerable to earthquake damage due to their limited ability to withstand lateral loads, compared to other structures. Therefore, a detailed assessment of the seismic response and resultant damage associated with such buildings becomes necessary. The present study employs machine learning models to effectively predict the seismic response and classify the damage level for a benchmark unreinforced masonry building. In this regard, eight regression-based models, namely, Linear Regression (LR), Stepwise Linear Regression (SLR), Ridge Regression (RR), Support Vector Machine (SVM), Gaussian Process Regression (GPR), Decision Tree (DT), Random Forest (RF), and Neural Networks (NN), were used to predict the building’s responses. Additionally, eight classification-based models, namely, Naïve Bayes (NB), Discriminant Analysis (DA), K-Nearest Neighbours (KNN), Adaptive Boosting (AB), DT, RF, SVM, and NN, were explored for the purpose of categorizing the damage states of the building. The material properties of the masonry and the earthquake intensity were considered as the input parameters. The results from the regression models indicate that the GPR model efficiently predicts the seismic response with larger coefficients of determination and smaller root mean square error values than other models. Among the classification-based models, the RF, AB, and NN models effectively classify the damage states with accuracy levels of 92.9%, 91.1%, and 92.6%, respectively. In conclusion, the overall performance of the non-parametric models, such as GPR, NN, and RF, was found to be better than that of the parametric models. Full article

As the forefront of inland extension on the Indian plate, the northeastern Tibetan Plateau, marked by low strain rates and high stress levels, is one of the regions with the highest seismic risk. Analyzing seismicity through statistical methods holds significant scientific value for understanding tectonic conditions and assessing earthquake risk. However, seismic monitoring capacity in this region remains limited, and earthquake frequency is low, complicating efforts to improve earthquake catalogs through enhanced identification and localization techniques. Bi-scale empirical probability integral transformation (BEPIT), a statistical method, can address these data gaps by supplementing missing events shortly after moderate to large earthquakes, resulting in a more reliable statistical data set. In this study, we analyzed six earthquake sequences with magnitudes of M_S ≥ 6.0 that occurred in northeastern Tibet since 2009, following the upgrade of the regional seismic network. Using BEPIT, we supplemented short-term missing aftershocks in these sequences, creating a more complete earthquake catalog. ETAS model parameters and b values for these sequences were then estimated using maximum likelihood methods to analyze parameter variability across sequences. The findings indicate that the b value is low, reflecting relatively high regional stress. The background seismicity rate is very low, with most mainshocks in these sequences being background events rather than foreshock-driven events. The p-parameter of the ETAS model is high, indicating that aftershocks decay relatively quickly, while the α-parameter is also elevated, suggesting that aftershocks are predominantly induced by the mainshock. These conditions suggest that earthquake prediction in this region is challenging through seismicity analysis alone, and alternative approaches integrating non-seismic data, such as electromagnetic and fluid monitoring, may offer more viable solutions. This study provides valuable insights into earthquake forecasting in the northeastern Tibetan Plateau. Full article

Hydrophone is a key component of marine seismic exploration systems, divided into a scalar hydrophone and vector hydrophone. The electrochemical vector hydrophone has attracted much attention due to its high sensitivity and low-frequency detection capability. With the development of noise reduction technology, high-frequency noise has been effectively suppressed, while low-frequency noise is still difficult to control, which has become a key issue in the monitoring of underwater target radiation noise. The traditional electrochemical vector hydrophone based on the molecular electron transfer (MET) principle is limited in the working bandwidth in the low-frequency band, which affects the detection capability of low-frequency radiation signals from underwater targets. In order to solve this problem, a frequency response extension method of a MET electrochemical vector hydrophone based on dynamic feedback network is proposed. By introducing a dynamic force balance negative feedback system based on a digital signal processor (DSP), the working bandwidth of the hydrophone is extended, and the detection capability of low-frequency signals is enhanced. At the same time, the system has field adjustability and can resist the long-term system frequency characteristic drift. Experimental results show that the proposed method effectively improves the frequency response performance of the electrochemical vector hydrophone, providing a new technical solution for its application in the monitoring of low-frequency radiation noise from underwater targets. Full article

Radon gas anomalies have been investigated as potential earthquake precursors for many years. In this work, we have studied the possible correlations between radon emissions and the seismic activity rate for a given region to test if the existing correlation may be later used to forecast the occurrence of earthquakes larger than a given magnitude. The Vrancea region (Romania) was chosen as a study area since it is being surveilled by a multidisciplinary real-time monitoring network, and at least seven earthquakes with magnitudes greater than 4.5 Mw have occurred in this area in the period from 2016 to 2020. Our research followed several steps: First, the recorded radon signals were preprocessed (detrended, deseasoned and smoothed). Then, the station’s signals were correlated in order to check which stations are recording radon anomalies due to the same regional tectonic process. On the other hand, the seismic activity rate was computed using the earthquakes in the main catalogue of the region that are able to generate radon emissions and can be registered at several stations. The obtained results indicate a significant correlation between the seismic activity rate and the temporal series of radon anomalies. A temporal lag between the seismic activity rate and the radon anomalies was found, which can be related to the proximity to the epicentre of the main earthquake in each of the studied subperiods. Changes in the regional tectonic stress field could explain why the seismic activity rate and radon anomalies are correlated over time. Further research could focus on obtaining a function to forecast the seismic activity rate using the following as dependent variables: the radon anomalies recorded at several stations, the distance from the stations, and tectonic factors such as the fault system, azimuth, type of soil, etc. Full article