Search Results (69)

This research provides an extensive evaluation of liquefaction induced by earthquakes in the Ionian Islands, specifically Lefkada, Cephalonia, Ithaki, and Zakynthos, through the compilation of a liquefaction inventory and GIS-based liquefaction susceptibility index (LiSI) maps. A total of 49 liquefaction sites from 20 causative earthquakes confirm that liquefaction is a recurrent geohazard in the area, primarily affecting coastal and low-lying areas with unconsolidated post-alpine deposits. The relationship between earthquake magnitude and maximum epicentral distance of observed liquefaction is consistent with global empirical datasets, indicating that moderate to strong earthquakes (M_w = 5.9–7.4) can induce liquefaction at considerable distances. The susceptibility model integrates eleven conditioning variables, classified as geological and geomorphological variables, hydrological indices and optical satellite imagery-derived data, within an analytic hierarchy process (AHP) framework. Lithology, age, and geomorphological unit emerged as the dominant conditioning variables. The LiSI maps confirm the zones previously identified in the inventory. Model validation and sensitivity analysis including confusion matrix components, key performance metrics and ROC analysis in coarser grid sizes demonstrate performance ranging from excellent (Zakynthos) to moderate (Lefkada and Cephalonia), while remaining inconclusive for Ithaki due to data limitations. The model exhibits generally conservative behavior, characterized by high precision and specificity but variable sensitivity, while it is largely stable across spatial resolutions in most cases. Full article

This study presents a site-specific calibration of the ratio between S-wave and P-wave peak ground acceleration (PGA) for use in low-cost, on-site earthquake early warning (EEWS) systems in Iași, Romania. A dataset of 25 intermediate-depth Vrancea earthquakes (Mw 4.1–5.7; epicentral distances 150–210 km) was analyzed. PGA values were extracted for the P- and S-wave windows on both horizontal components and combined using geometric means. The resulting S/P amplitude ratios yield a median value of k_S/P = 6.19 and a logarithmic standard deviation of σlog₁₀ = 0.31, corresponding to a multiplicative uncertainty factor of approximately ×2. These results indicate that S-wave amplitudes are typically six times larger than P-wave amplitudes at this site, consistent with soft-soil amplification observed in comparable stations in Japan and Italy. The calibrated ratio can be used as a site-specific input for future MEMS-based on-site EEW implementations to estimate the expected S-wave PGA immediately after P-wave detection, with the observed S–P delays in Iași indicating a typical available warning window of 20–22 s. Full article

The M_JMA 7.6 (Mw 7.5) Noto Peninsula Earthquake of 1 January 2024 in Japan triggered widespread slope failures across northern Noto region, but their spatial controls and susceptibility patterns remain poorly quantified. Most previous studies have focused mainly on fault rupture, ground deformation, and tsunami impacts, leaving a clear gap in machine learning based assessment of earthquake-induced slope failures. This study integrates 2323 mapped landslides with eleven conditioning factors to develop the first data-driven susceptibility framework for the 2024 event. Spatial analysis shows that 75% of the landslides are smaller than 3220 m² and nearly half occurred within about 23 km of the epicenter, reflecting concentrated ground shaking beyond the rupture zone. Terrain variables such as slope (mean 31.8°), southwest-facing aspects, and elevations of 100–300 m influenced the failure patterns, along with peak ground acceleration values of 0.8–1.1 g and proximity to roads and rivers. Six supervised machine learning models were trained, with Random Forest and Gradient Boosting achieving the highest accuracies (AUC = 0.95 and 0.94, respectively). Explainable AI using SHapley Additive exPlanations (SHAP) identified slope, epicentral distance, and peak ground acceleration as the dominant predictors. The resulting susceptibility maps align well with observed failures and provide an interpretable foundation for post-earthquake hazard assessment and regional risk reduction. Further work should integrate post-seismic rainfall, multi-temporal inventories, and InSAR deformation to support dynamic hazard assessment and improved early warning. Full article

Soil liquefaction remains a critical geotechnical hazard during earthquakes, posing significant risks to infrastructure and urban resilience. Traditional empirical methods, while practical, often fall short in capturing complex parameter interactions and providing interpretable outputs. This study presents an interpretable machine learning (IML) framework for soil liquefaction assessment using Rough Set Theory (RST) to generate a transparent, rule-based predictive model. Leveraging a standardized SPT-based case history database, the model induces IF–THEN rules that relate seismic and geotechnical parameters to liquefaction occurrence. The resulting 25-rule set demonstrated an accuracy of 86.2% and strong alignment (93.8%) with the widely used stress-based semi-empirical model. Beyond predictive performance, the model introduces scenario maps and parameter interaction diagrams that elucidate key thresholds and interdependencies, enhancing its utility for engineers, planners, and policymakers. Notably, the model reveals that soils with high fines content can still be susceptible to liquefaction under strong shaking, and that epicentral distance plays a more direct role than previously emphasized. By balancing interpretability and predictive strength, this rule-based approach advances site-adaptive, explainable, and technically grounded liquefaction assessment—bridging the gap between traditional methods and intelligent decision support in geotechnical engineering. Full article

In contrast to a relatively simple whole structure of the African Large Low Velocity Province (LLVP), the Mid-Pacific LLVP appears to be much more complex and likely interacts more with the down-going slab debris from the circum-Pacific subduction zones. Tomographic models show an apparent hole in the Mid-Pacific LLVP, coinciding with observed anomalous SPdKS arrivals. Previous studies have linked these anomalies to a large-scale mega ultra-low velocity zone (ULVZ) exhibiting up to a 45% S-wave velocity reduction. To further investigate this anomaly, we analyzed SKS waveforms from Fiji–Tonga earthquakes recorded by the USArray. Many events display pronounced travel time jumps and waveform distortions near epicentral distances of 100°, consistent with strong multipathing effects. Notably, such complexities are absent in S and SKKS phases, indicating that only the down-going SKS leg is affected. Using waveform modeling, we find that a northeast-dipping high-velocity anomaly approximately 300 km wide, 800 km long, and with a shear velocity increase of ~2% provides a good fit to the observed SKS data. This apparent LLVP hole may represent a localized downwelling within the LLVP or a remnant slab fragment interacting with the deep mantle. Full article

Regional seismic phases are essential for imaging Earth’s internal structure. Although extensive regional seismic networks are publicly available worldwide, only a small fraction of recorded phase arrivals are picked for constraining earthquake source parameters, leaving most data untapped. Recent deep-learning methods offer powerful tools for automatic phase picking, yet their performance often lags behind that of human experts, particularly at relatively large epicentral distances such as the case of the Pn phase (~200–2000 km). Here, we systematically assess the effect of different test-time augmentation strategies on the Pn phase picking performance using PickNet and PhaseNet, along with the Seis-PnSn dataset containing data worldwide to simulate the out-of-distribution situation. We also propose quality control measures to obtain reliable results when ground truths are unknown. Our experiments show that filter-bank augmentation is more effective than the shift augmentation and the rotation augmentation, improving the proportion of picks within ±0.5/1.0 s errors to 53.87%/70.82% compared with the baseline of 48.98%/66.94% for PickNet and ±0.5/1.0 s errors to 48.45%/67.06% compared with the baseline of 46.32%/64.28% for PhaseNet. After the quality control using the standard deviation of different augmentation results, the proportion is further boosted to 67.39%/78.53% for PickNet and 57.99%/74.72% for PhaseNet. Additionally, we provide the workflow in our study as scripts for real-world data processing. Our work enhances both the accuracy and accessibility of regional seismic phase picking, thereby contributing to the studies of Earth’s internal structure and earthquake source characterization. Full article

In this study, a statistical analysis of ground geomagnetic data has been attempted to extract the seismo-electromagnetic (SEM) signatures associated with moderate earthquakes in the region of the seismic gap in the Kumaun Himalaya, Uttarakhand, India. We applied the discrete wavelet transform (DWT) method to the geomagnetic data to identify the ULF energy of the signal. The ULF energy obtained in the central frequency range of 0.01 Hz was further filtered to extract the anomalous ULF energy, which is associated with pre-earthquake processes. We also applied multifractal analysis to the geomagnetic data to classify the complexities in the signal, which are indicative of the seismotectonic environment. We observed enhancements in the ULF energy anomalies associated with large-magnitude earthquakes occurring in the western part of Nepal, even over large epicentral distances (~120 km). The multifractal analysis shows the overlap of anomalies in the Hwp and Hwn signatures in most cases, which suggests that multiple mechanisms generate low- and high-frequency components in the anomalous data. This reflects the complex nature of seismicity in this region of the Main Central Thrust (MCT). Full article

The twin earthquakes that struck Kahramanmaraş in 2023 (Mw 7.7 and Mw 7.6) caused widespread structural destruction across southeastern Türkiye, underscoring the need for more refined approaches to seismic damage assessment. In this study, a large-scale machine learning (ML) analysis is conducted to identify and classify damage patterns among 304,299 buildings across 11 cities. Ten ML algorithms are implemented, and their performance in the multiclass classification of damage severity is comparatively evaluated (collapsed, urgent demolition, moderately damaged, and severely damaged). Unlike conventional methods that rely on single-station data, the proposed approach integrates ground motion parameters from the six seismic stations closest to each building. These parameters include peak ground acceleration, several distance measures (Joyner–Boore, rupture, and epicentral distances), and site condition indicators such as mean shear wave velocity in the upper 30 m and soil classification, yielding 60 engineered features per building. The analysis reveals that ensemble learning models, particularly the random forest and a voting ensemble, achieve the highest classification accuracies (79.65% and 79.62%, respectively). Moreover, classification performance varies across damage categories: severely damaged structures exhibit the highest F1-score (0.891), whereas collapsed buildings exhibit lower accuracy (F1-score: 0.408). These findings offer practical value for post-earthquake emergency operations. Furthermore, the methodology establishes a precedent for future seismic risk assessments and supports data-driven decision-making. Full article

Coseismic landslides are geological events in which slopes, either on the verge of instability or already in a fragile state, experience premature failure due to seismic shaking. On 5 September 2022, an Ms 6.8 earthquake struck Luding County, Sichuan Province, China, triggering numerous landslides that caused severe casualties and property damage. This study systematically interprets 13,717 coseismic landslides in the Luding earthquake’s epicentral area, analyzing their spatial distribution concerning various factors, including elevation, slope gradient, slope aspect, plan curvature, profile curvature, surface cutting degree, topographic relief, elevation coefficient variation, lithology, distance to faults, epicentral distance, peak ground acceleration (PGA), distance to rivers, fractional vegetation cover (FVC), and distance to roads. The analytic hierarchy process (AHP) was improved by incorporating frequency ratio (FR) to address the subjectivity inherent in expert scoring for factor weighting. The improved AHP, combined with the Pearson correlation analysis, was used to identify the dominant controlling factor and assess the landslide susceptibility. The accuracy of the model was verified using the area under the receiver operating characteristic (ROC) curve (AUC). The results reveal that 34% of the study area falls into very-high- and high-susceptibility zones, primarily along the Moxi segment of the Xianshuihe fault and both sides of the Dadu river valley. Tianwan, Caoke, Detuo, and Moxi are at particularly high risk of coseismic landslides. The elevation coefficient variation, slope aspect, and slope gradient are identified as the dominant controlling factors for landslide development. The reliability of the proposed model was evaluated by calculating the AUC, yielding a value of 0.8445, demonstrating high reliability. This study advances coseismic landslide susceptibility assessment and provides scientific support for post-earthquake reconstruction in Luding. Beyond academic insight, the findings offer practical guidance for delineating priority zones for risk mitigation, planning targeted engineering interventions, and establishing early warning and monitoring strategies to reduce the potential impacts of future seismic events. Full article

This study focuses on determining the high-frequency decay parameter kappa (k) and its site component (k₀) for sixteen accelerometric stations installed in suitable locations in North Macedonia. Kappa characterizes the attenuation of ground motion at high frequencies, describing the decrease in the acceleration amplitude spectrum. It is defined using a regression line in log-linear space, starting from the point where the S-wave amplitude spectrum begins to decay rapidly. The site characteristics of the stations are determined through geophysical and borehole investigations, as well as HVSR mean curves derived from earthquake data. The strong-motion data used in this analysis originate from earthquake events with a moment magnitude greater than 3 (M_W > 3), an epicentral distance less than 120 km (R_epi < 120 km), and a focal depth lower than 30 km (h < 30 km). The records undergo visual inspection and filtering, with those having a signal-to-noise ratio (SNR) below 3 excluded from further analysis. The study examines the correlation between kappa values and various parameters, including magnitude, epicentral distance, average shear-wave velocity in the top 30 m depth (V_S30), and fundamental site frequency (f₀). The importance of this study is the application in the future evaluation/update of seismic hazard analysis of the region. Full article

Prediction of the intensity of earthquake-induced motions at the ground surface attracts extensive attention from the geoscience community due to the significant threat it poses to humans and the built environment. Several factors are involved, including earthquake magnitude, epicentral distance, and local soil conditions. The local site effects, such as resonance amplification, topographic focusing, and basin-edge interactions, can significantly influence the amplitude–frequency content and duration of the incoming seismic waves. They are commonly predicted using site effect proxies or applying more sophisticated analytical and numerical models with advanced constitutive stress–strain relationships. The seismic excitation in numerical simulations consists of a set of input ground motions compatible with the seismo-tectonic settings at the studied location and the probability of exceedance of a specific level of ground shaking over a given period. These motions are applied at the base of the considered soil profiles, and their vertical propagation is simulated using linear and nonlinear approaches in time or frequency domains. This paper provides a comprehensive literature review of the major input parameters for site response analyses, evaluates the efficiency of site response proxies, and discusses the significance of accurate modeling approaches for predicting bedrock motion amplification. The important dynamic soil parameters include shear-wave velocity, shear modulus reduction, and damping ratio curves, along with the selection and scaling of earthquake ground motions, the evaluation of site effects through site response proxies, and experimental and numerical analysis, all of which are described in this article. 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

Long-term observations in wells make it possible to study changes in groundwater pressure/level during individual earthquakes (seismo-hydrogeodynamic effects—SHGEs) over a wide range of periods of their manifestation. Information on the morphological features and durations of the SHGEs together with data on earthquake parameters form the basis for creating the unique typifications of SHGEs for individual observation wells. With reliable verification, such SHGE typifications provide the practical use of well observation data to predict strong earthquakes and assess their impact on groundwater. During long-term (1996–2022) precision observations of pressure/water level variations in wells of the Petropavlovsk–Kamchatsky test site (Kamchatka Peninsula, northwest Pacific seismic belt), SHGE typifications describing the manifestations of various types of SHGEs at the earthquakes in ranges of magnitudes M_w = 5.0–9.1 and epicentral distances d_e = 80–14,600 km were developed. At the same time, the issue of verifying created SHGE typifications for individual wells in relation to the strongest and closest earthquakes, accompanied by noticeable tremors in the observation area, is relevant. On 3 April 2023, an earthquake, M_w = 6.6 (EQ), occurred at an epicentral distance d_e = 67–77 km from observation wells. Various changes in the groundwater pressure/level were recorded in the wells: oscillations and other short-term and long-term effects of seismic waves, coseismic jumps in water pressure caused by a change in the static stress state of water-bearing rocks during the formation of rupture in the earthquake source, and supposed hydrogeodynamic precursors. The EQ was used to verify the SHGE typifications for wells YuZ-5 and E-1 with the longest observation series of more than 25 years. In these wells, the seismo-hydrogeodynamic effects recorded during the EQ corresponded to the previously observed SHGE during the two strongest earthquakes with M_w = 7.2, d_e = 80 km and M_w = 7.8, d_e = 200 km. This correspondence is considered an example of the experimental verification of previously created SHGE typifications in individual wells in relation to the most powerful earthquakes in the wells’ area. Updated SHGE typifications for wells E-1 and YuZ-5 are presented, showing the patterns of water level/pressure changes in these wells depending on earthquake parameters and thereby increasing the practical significance of well observations for assessing earthquake consequences for groundwater, searching for hydrogeodynamic precursors and forecasting strong earthquakes. The features of the hydrogeodynamic precursor manifesting in the water level/pressure lowering with increased rates in well E-1 before earthquakes with M_w ≥ 5.0 at epicentral distances of up to 360 km are considered. A retrospective statistical analysis of the prognostic significance of this precursor showed that its use for earthquake forecasting increases the efficiency of predicting earthquakes with M_w ≥ 5.0 by 1.55 times and efficiency of predicting earthquakes with M_w ≥ 5.8 by 2.34 times compared to random guessing. This precursor was recorded during the 92 days before the EQ and was identified in real time with the issuance of an early prognostic conclusion on the possibility of a strong earthquake to the Kamchatka branch of the Russian Expert Council for Earthquake Forecasting. Full article

This study aimed to identify the critical ground motion parameters that lead to structural damage and assess their impact on the nonlinear responses of buildings. The analyses are carried out using a calibrated numerical model that was acquired within the ICONS experimental framework that represents reinforced concrete (RC) structures constructed before seismic design regulations were enforced. For the analysis, 30 seismic records were chosen based on magnitude (M), epicentral distance (R), and peak ground acceleration (PGA) for two high seismic activity areas that were observed. Eleven parameters are categorized, traditional metrics, energy-based, spectrum-based, duration-based, and fundamental metrics, and examined based on their main attributes. The results showed a strong relationship between certain seismic properties and the maximum interstory drifts of building as a damage prediction parameter. Peak ground velocity (PGV), specific energy density (SED), and Housner Intensity (HI) were found to be the most important variables in assessing the correlation with possible structural damage. Therefore, the assessment of structural damage based on nonlinear dynamic analysis should primarily incorporate PGV with the possible addition of energy- and spectrum-based metrics as the most reliable ground motion parameters for the selection of earthquake records for time history analysis. 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