Journal of Geophysical Research: Solid Earth最新文献

{"title":"Effects of Regional Stress State and Pore Fluid Pressure on the Onset and Style of Caldera Collapse","authors":"Matías A. Villarroel,&nbsp;Martin P. J. Schöpfer,&nbsp;John Browning,&nbsp;Eoghan P. Holohan,&nbsp;Claire E. Harnett,&nbsp;Carlos J. Marquardt,&nbsp;Pamela P. Jara","doi":"10.1029/2024JB031054","DOIUrl":"https://doi.org/10.1029/2024JB031054","url":null,"abstract":"<p>Collapse calderas result from subsidence of a magma reservoir roof during large-volume eruptions. Whilst calderas form in various tectonic settings, it is unclear how regional (“far-field”) forces influence caldera fault nucleation, orientation and architecture. Furthermore, although the presence of a pore fluid is known to reduce the effective stress, it is typically neglected in past caldera collapse models. Utilizing two-dimensional Distinct Element Method (DEM) models, we explore the influences of regional stress and pore fluid pressure on the evolutions of stress, strain and faulting during caldera subsidence. We simulate a shallow magma volume as an inviscid inclusion within a homogeneous crust and decrease the inclusion's pressure to model magma withdrawal. Results reveal that the critical underpressure needed to trigger collapse is reduced in extensional regimes, particularly in fluid-saturated conditions, due to lowered frictional resistance on faults. We observe three progressive deformation stages: initial fracturing at the reservoir roof, collapse onset, and complete roof failure. The geometry of faults depends on the tectonic setting, with extensional conditions favoring steeper fault dips and compressional settings promoting shallower, outward-dipping reverse faults. Models simulating a fluid-saturated crust exhibit similar effects to those models that simulate lower strength materials. This study highlights the need to account for regional stress states and crustal properties in volcanic hazard assessment, especially in caldera systems with complex hydrothermal or tectonic influences. Our findings are compared with recent collapse episodes, underscoring the utility of DEM modeling in understanding crustal responses to magma depletion.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1029/2024JB031054","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144085177","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Co-Evolution of Specific Stiffness and Permeability of Rock Fractures Offset in Shear","authors":"Xinxin He,&nbsp;Pengliang Yu,&nbsp;Agathe Eijsink,&nbsp;Chris Marone,&nbsp;Parisa Shokouhi,&nbsp;Jacques Rivière,&nbsp;Shimin Liu,&nbsp;Derek Elsworth","doi":"10.1029/2024JB030633","DOIUrl":"https://doi.org/10.1029/2024JB030633","url":null,"abstract":"<p>Fractures and faults represent planes of weakness and compliance in rock masses that serve as focal points for both microearthquakes and fluid transport, with seismicity and permeability evolution closely linked. Contact stiffness is highly stress-sensitive and directly influences permeability. We explore the co-evolution of specific stiffness and permeability of rough fractures under normal stress and shear offset using numerical simulations. Individual rough fractures are represented by variable amplitude (Root mean square) and wavelength (<i>λ</i>) using a granular mechanics model. Contacting rough surfaces are mated, offset in shear, and then compacted in displacement mode. The compacting fractures generate stress-dependent changes in contact porosity, which govern both permeability and stiffness evolution. We establish a universal dimensionless relationship linking specific stiffness and permeability that inherently incorporates the effects of surface roughness, shear offset, and microcracking. The observed cracking effect—where local stress redistribution and pressure-driven microcrack propagation dynamically alter the aperture field—introduces a nonlinear permeability response at high stress. Increased roughness amplitude and larger shear offsets reduce stiffness while dampening permeability sensitivity to stress, demonstrating a strong interplay between surface texture and hydro-mechanical behavior. While the model captures this behavior effectively, deviations emerge at very low porosities due to extreme aperture sensitivity in this limit.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1029/2024JB030633","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144085176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Link Between Enhanced Pore Surface Relaxivity and Mineral Alteration in Basalts","authors":"Zuhao Kou,&nbsp;James Howard,&nbsp;Olivia Terry,&nbsp;Tianxiao Shen,&nbsp;Lucas Abernathy,&nbsp;Dragan Veselinovic,&nbsp;Michael Dick,&nbsp;Shaina Kelly","doi":"10.1029/2025JB031227","DOIUrl":"https://doi.org/10.1029/2025JB031227","url":null,"abstract":"<p>Hydrothermal alteration significantly affects the mineralogical and geochemical composition of subsurface rocks. This research utilized a combination of low-field time-domain nuclear magnetic resonance, gas adsorption-desorption isotherms, and scanning electron microscopy (SEM) with energy dispersive spectroscopy to characterize the pore systems of a range of flow top and flow interior basalt samples from Newberry Volcano drill core. A power-law relationship between hydrothermal mineral alteration and magnetic susceptibility of pore-facing minerals is revealed, suggesting a bulk method for quantifying degree of mineral alteration from core or wellbore data. Transverse relaxation time (<i>T</i>₂) distributions, combined with gas adsorption-based and SEM image-based pore size distributions, yield faster <i>T</i>₂ relaxation or enhanced surface relaxivity values within sample micro- and macropores facing or lined with secondary minerals. This relationship can be used to evidence increased paramagnetic metal ion (e.g., Fe, Mn, Cr, Co, V) accessibility for subsurface engineering applications such as in situ carbon mineralization and critical minerals extraction.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144085178","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effective Inversion of Yellowstone Airborne TEM Data Using Deep Learning","authors":"Siyuan He,&nbsp;Ziang He,&nbsp;Xiangyun Hu,&nbsp;Carol Finn,&nbsp;Lichao Liu,&nbsp;Esben Auken,&nbsp;Hongzhu Cai","doi":"10.1029/2024JB029855","DOIUrl":"10.1029/2024JB029855","url":null,"abstract":"<p>Deep learning methodologies can significantly accelerate the interpretation of airborne transient electromagnetic (ATEM) data. Nevertheless, it remains challenging for deep learning methods to deal with data vectors with missing values. This study introduces innovative processing techniques for transient electromagnetic data, enabling the trained neural network to effectively manage data vectors with missing values. Furthermore, it presents a comprehensive analysis within the Yellowstone National Park study area, comparing the performance of networks trained on real field data sets and synthetic data sets in ATEM data inversion. The results strongly support the superiority of networks trained on field data sets over those trained on synthetic ones. In addition, the research highlights two key factors differentiating these data sets—noise levels and the distribution of resistivity models. It examines the variations in the distribution of resistivity models across data set types and their consequential effects on inversion results. This study underscores the critical importance of utilizing real field data on network training, demonstrating its remarkable effectiveness in deciphering intricate geological structures and achieving detailed imaging of the subsurface conductivity.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144066474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Novel Approach Combining GRACE(-FO) and Altimetry Reveals a Recent East-to-West Shift in the Slowdown of Glacier-Scale Mass Loss in the Greenland Ice Sheet","authors":"Haotian Wu,&nbsp;Jiangjun Ran,&nbsp;Shoaib Ali,&nbsp;Natthachet Tangdamrongsub,&nbsp;Xavier Fettweis","doi":"10.1029/2024JB030862","DOIUrl":"10.1029/2024JB030862","url":null,"abstract":"<p>The mass loss of the Greenland Ice Sheet (GrIS) has profound impacts on sea levels, the water cycle, and global climate variability. The Gravity Recovery and Climate Experiment (GRACE) and its follow-on mission (GRACE-FO) provide accurate but limited spatial resolution observations of GrIS mass changes. Therefore, we developed a novel multi-time scale weighted forward modeling (WFM) approach that combines GRACE(-FO) observations with satellite altimetry data to improve mass change estimations in the GrIS at high-resolution (25 km × 25 km). The WFM solution effectively represents the glacier-scale interannual mass variations, with an average correlation of 0.71 with the Input-Output method, higher than Mascon products. The WFM solution reveals a spatial pattern of glacier mass change from 2020 to 2023, indicating that the slowdown in the GrIS glacier mass loss has shifted from the east to the west compared to 2013–2018; the mass loss rate decreased by 44.9 ± 1.04 Gt/yr in the western GrIS and increased by 42.3 ± 0.98 Gt/yr in the eastern GrIS. The most pronounced mass loss slowdown occurred at Jakobshavn Glacier (7.3 ± 0.07 Gt/yr). In this pattern, the trough of low-pressure west of the GrIS moved westward, and a high-pressure anomaly over the North Atlantic south of the GrIS intensified southwesterly winds over the GrIS. These winds transported warmer, moister air from the Atlantic toward the western GrIS, leading to increased snowfall and rainfall, thereby promoting glacier mass accumulation. If this pattern continues, it could benefit the preservation of the ice in the western GrIS.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143933345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep and Clustered Microseismicity at the Edge of Southern New Zealand's Transpressive Plate Boundary","authors":"Jack N. Williams,&nbsp;Donna Eberhart-Phillips,&nbsp;Sandra Bourguignon,&nbsp;Mark W. Stirling,&nbsp;Will Oliver","doi":"10.1029/2024JB030371","DOIUrl":"10.1029/2024JB030371","url":null,"abstract":"<p>Deep (<span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mo>&gt;</mo>\u0000 <mo>∼</mo>\u0000 </mrow>\u0000 <annotation> ${ &gt;} sim $</annotation>\u0000 </semantics></math>20 km) crustal seismicity is typically associated with cold Precambrian lithosphere. However, previous studies have reported that the depth above which 90% of seismicity occurs (D90) in the Southland region of New Zealand is 30–35 km, even though it is situated in relatively young Mesozoic crust at the eastern edge of the transpressive Pacific-Australian plate boundary. To examine this unusual lower crustal seismicity further, we deployed 19 short period seismometers between October 2022 and 2023 through the Southland Otago Seismic Array (SOSA). Supplemented by the permanent GeoNet network and an ongoing array to the west in Fiordland, SOSA provided a station spacing of <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mo>∼</mo>\u0000 </mrow>\u0000 <annotation> ${sim} $</annotation>\u0000 </semantics></math>30 km across this 150 km wide region. In total, SOSA recorded 85 events in and around Southland. These events are not uniformly distributed, with an abrupt decrease in seismicity east of Fiordland. We also identified six spatial-temporal clusters of elevated seismicity across eastern Fiordland and Southland. The distribution of SOSA event depths confirms that seismicity extends into Southland's mid-lower crust and possibly into the upper mantle (D90 = 32 km). This D90 estimate is significantly deeper than suggested by event depths to the north in Otago (D90 = 14 km). Using 1D lithospheric strength profiles, we suggest that the embrittlement of Southland's mid-lower crust reflects the emplacement of mafic terranes in this region during Mesozoic subduction at the Gondwana margin. Our study therefore highlights that a mafic lower crust can lead to deep seismicity within a region undergoing transpression.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1029/2024JB030371","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143933237","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine Learning and Big Data Mining Reveal Earth's Deep Time Crustal Thickness and Tectonic Evolution: A New Chemical Mohometry Approach","authors":"Jianping Zhou,&nbsp;Ehsan Farahbakhsh,&nbsp;Simon Williams,&nbsp;Xiaohui Li,&nbsp;Yongjiang Liu,&nbsp;Sanzhong Li,&nbsp;R. Dietmar Müller","doi":"10.1029/2024JB030404","DOIUrl":"https://doi.org/10.1029/2024JB030404","url":null,"abstract":"<p>Quantitative analysis of crustal thickness evolution across deep time poses critical insights into the planet's geological history. It may help uncover new areas with potential critical mineral deposits and reveal the impacts of crustal thickness and elevation changes on the development of the atmosphere, hydrosphere, and biosphere. However, most existing estimation methods are restricted to arc-related magmas, limiting their broader application. By mining extensive geochemical data from present-day subduction zones, collision orogenic belts, and non-subduction-related intraplate igneous rock samples worldwide, along with their corresponding Moho depths during magmatism, we have developed a machine learning-based mohometry linking geochemical data to Moho depth, which is universally applicable in reconstructing ancient orogenic systems' paleo-crustal evolution and tracking complex tectonic histories in both spatial and temporal dimensions. Our novel mohometry model demonstrates robust performance, achieving an <i>R</i>² of 0.937 and an Root Mean Squared Error of 4.3 km. Feature importance filtering highlights key geochemical proxies, allowing for accurate paleo-crustal thickness estimation even when many elements are missing. Model validation in southern Tibet and the South China Block, regions characterized by well-constrained crustal histories and complex tectonic processes, demonstrates its broad applicability. Reconstructed paleo-crustal thickness records reveal a strong correlation between crustal thickening events and the formation of porphyry ore deposits, offering new insights for mineral exploration in ancient orogens subjected to significant surface erosion. By enabling the reconstruction of crustal thickness across geological timescales, this model enhances our understanding of Earth's internal dynamics and their interactions with surface processes, thereby advancing our comprehension of Earth's geological history.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143930460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The Viscosity of Albitic Melt at High Pressures and Implications for the Mobility of Crust-Forming Magmas","authors":"Aaron Wolfgang Ashley,&nbsp;Mainak Mookherjee,&nbsp;Man Xu,&nbsp;Tony Yu,&nbsp;Geeth Manthilake,&nbsp;Yanbin Wang","doi":"10.1029/2024JB030527","DOIUrl":"10.1029/2024JB030527","url":null,"abstract":"<p>The earliest form of continental crust was produced by tonalite-trondhjemite-granodiorite (TTG) magmas. Molten albite (NaAlSi₃O₈) is representative of TTGs and also a major component of modern crust-forming magma. The viscosity of the melt controls the magma ascent rate and hence influences the production of new continental crust. It is well known that the viscosity (<i>η</i>) of albitic melt exhibits an anomalous pressure (<i>P</i>) dependence. However, prior results on the melt <i>η</i> at high-<i>P</i> differ significantly which limits our ability to predict the movement of crust-forming magma at depth. In this study, we more tightly constrained the <i>P</i>-effect on <i>η</i> in anhydrous albitic melt via high-<i>P</i> and high-temperature (<i>T</i>) falling sphere experiments. We limited undesirable drag effects by using small sphere-to-capsule diameter ratios (<i>d</i>/<i>D</i>) such that <i>d</i>/<i>D</i> ≤ 0.12, and evaluated uncertainties due to such drag using a Monte Carlo approach. Our results show that melt <i>η</i> first decreases with <i>P</i> (i.e., ∂<i>η</i>/∂<i>P</i> &lt; 0) and then increases with continued compression (∂<i>η</i>/∂<i>P</i> &gt; 0) with a well-defined <i>η</i> minimum (<i>η</i>_min) at ∼6 GPa along a ∼2,000 K isotherm. We find that the viscosity of the melt can be described by an Arrhenius formalism with an activation volume that varies with <i>P</i> and <i>T</i>. The results indicate that <i>η</i> of aluminosilicate magmas decrease with depth and temperature in the crust, thereby mobilizing the magmas to promote rapid volcanic eruptions. The results also suggest that TTG magmas relevant for the early Earth could pond during ascent due to the anomalous <i>P</i>-effect on <i>η</i>.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143932671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Physics-Informed Deep Learning for Estimating the Spatial Distribution of Frictional Parameters in Slow Slip Regions","authors":"Rikuto Fukushima,&nbsp;Masayuki Kano,&nbsp;Kazuro Hirahara,&nbsp;Makiko Ohtani,&nbsp;Kyungjae Im,&nbsp;Jean-Philippe Avouac","doi":"10.1029/2024JB030256","DOIUrl":"https://doi.org/10.1029/2024JB030256","url":null,"abstract":"<p>Slow slip events (SSEs) have been observed in many subduction zones and are understood to result from frictional unstable slip on the plate interface. The diversity of their characteristics and the fact that interplate slip can also be seismic suggest that frictional properties are heterogeneous. We are however lacking methods to determine spatial variations of frictional properties. In this paper, we employ a Physics-Informed Neural Network (PINN) to achieve this goal using a synthetic model inspired by the long-term SSEs observed in the Bungo channel. PINN is a deep learning technique that can be used to solve the differential equations representing the physics of the problem and determine the model parameters from observations. We start with an idealized case where it is assumed that fault slip is directly observed. We next move to a more realistic case where the observations consist of synthetic surface displacement velocity data measured by virtual GNSS stations. We find that the geometry and friction properties of the velocity weakening region, where the slip instability develops, are well estimated, especially if surface displacement velocity above the velocity weakening region is observed. Our PINN-based method can be seen as an inversion technique with the regularization constraint that fault slip obeys a particular friction law. This approach remediates the issue that standard regularization techniques are based on non-physical constraints. Our results show that the PINN-based method is a promising approach for estimating the spatial distribution of friction parameters from GNSS observations.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1029/2024JB030256","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143930229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RuptureNet2D, a Deep Neural Network Based Surrogate for Dynamic Earthquake Rupture Simulation in Two Dimensions","authors":"Zekang Gong,&nbsp;Ziyi Wang,&nbsp;Chao Liang,&nbsp;Andreas Nienkötter,&nbsp;Jianze Wang,&nbsp;Chunmei Ren,&nbsp;Xinyan Peng","doi":"10.1029/2024JB030069","DOIUrl":"https://doi.org/10.1029/2024JB030069","url":null,"abstract":"<p>Earthquake dynamic rupture simulations are crucial for physics-based seismic hazard assessment. However, due to its intricate dynamics spanning vast spatial and temporal scales, these simulations pose a complex and computationally challenging problem. Here, we propose an end-to-end deep learning model (RuptureNet2D) as a cost-effective alternative (at evaluation time) to expensive numerical simulations. This model is trained on a data set of <span></span><math>\u0000 <semantics>\u0000 <mrow>\u0000 <mo>∼</mo>\u0000 </mrow>\u0000 <annotation> ${sim} $</annotation>\u0000 </semantics></math>300k simulations in two dimensions and is capable of simultaneously predicting two key earthquake source parameters: rupture time and final slip. Testing reveals exceptional model performance on faults with homogeneous and heterogeneous (with one or two asperities) frictional parameters but only requires a fraction (1/100,000 to 1/1,000) of the prediction time compared to numerical simulations. However, the accuracy of the neural network decreases as the fault length and number of asperities increases beyond the range of the training data. Nevertheless, our work demonstrates the applicability and efficiency of neural networks as a surrogate for earthquake dynamic rupture simulations which has a potential of significantly accelerating physics-based earthquake source inversion and advancing our understanding of earthquake physics.</p>","PeriodicalId":15864,"journal":{"name":"Journal of Geophysical Research: Solid Earth","volume":"130 5","pages":""},"PeriodicalIF":3.9,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143930160","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}