Abstract:In the year 2023,China successfully built the world's most extensive earthquake Early Warning Station Network.This network is characterized by its dense distribution of sta-tions,which serve as a robust data foundation for various critical applications,including earth-quake localization,seismic monitoring,earthquake detection,and numerous other related fields.In recent years,the application of deep learning methods in seismic phase picking has grown significantly,leading to the development of highly effective neural network models.Some of the most representative and widely recognized models in this domain include Phase-Net,LPPN,and CSESnet,all of which are based on convolutional neural network(CNN)archi-tectures.Additionally,there is EQTransformer,which is a more advanced neural network model that combines convolutional neural network,recurrent neural network(RNN),and Trans-former-based architectures to achieve superior performance.These deep learning methods have consistently demonstrated their ability to outperform traditional phase picking approaches,par-ticularly when trained on large-scale labeled datasets containing high-quality earthquake data.Despite the success of these deep learning-based models,they are predominantly trained using datasets that mainly consist of data recorded by velocity seismometer,which poses a limitation. The earthquake Early Warning Station Network records data from three types of seismic instruments;velocity seismometer,accelerator,and seismic intensity meter(MEMS)instru-ments.The differences in data characteristics between these instrument types lead to a decline in the accuracy of current deep learning models when applied to data recorded by accelerometers or MEMS instruments.Furthermore,there is no large-scale,high-quality dataset recorded by early warning stations to train a neural network model for performing phase picking in Early Warning Station Network data.To address the issue of declining accuracy in neural network models when processing data from Early Warning Station Network data,we leveraged transfer learning,a machine learning technique that has proven to be highly effective in scenarios where labeled data is limited.Transfer learning enables pre-trained model to transfer knowledge acquired in one domain to a different but related domain,thereby enhancing model perform-ance with minimal additional data and training costs. In our study,we utilized the Bidirectional Recurrent Neural Network(BRNN)model pub-lished by Yu et al(2023)as the pre-trained model.Then,we perform transfer learning using 26,739 pieces of manually labeled data from Sichuan region.The data included in these samples are from velocity seismometer,accelerator,and MEMS instruments.To explore the effect of training iterations on model performance,we trained three separate transfer learning models with iteration counts of 500,1 000,and 10 000,respectively. The evaluation results,based on standard performance metrics such as precision,recall,Fl score,and the mean and standard deviation of time residuals between model-predicted and manually labeled phase arrivals,demonstrated that transfer learning significantly improved the detection accuracy of both Pg and Sg phases recorded by early warning station.The transfer learning model achieved an Fl score of 0.899 for the Pg phase and 0.862 for the Sg phase,cor-responding to improvements of 0.02 and 0.067,respectively,compared to the pre-trained model.An analysis of different training iteration counts revealed that model performance stabi-lized after approximately 500 iterations,with further increases to 1 000 and 10 000 iterations yielding only marginal performance improvements.To assess the robustness of the transfer learning model,we tested its performance across various settings,including different instru-ment types,signal-to-noise ratio(SNR)conditions,and epicentral distances.The results indi-cated that transfer learning enhanced the model's generalization ability.For instance,under low SNR conditions ranging from 0 to 5 dB,the F1 scores for Pg and Sg phases improved signific-antly,increasing from 0.697 and 0.479 to 0.817 and 0.676,respectively.The transfer learn-ing model achieved an F1 score of 0.891 for detecting Pg in accelerator data,an improvement of 0.019 compared to the pre-trained model.The detection accuracy for Sg showed the greatest improvement,with an F1 score of 0.852,an increase of 0.102 compared to the pre-trained model.Additionally,detection results for individual earthquake events revealed that the pre-trained model exhibited significant failures to detect Sg phase at epicentral distances exceeding 80 km.In contrast,the transfer learning model effectively reduced these misses,achieving marked improvements in Sg phase detection for such cases.Furthermore,we evaluated the per-formance of the transfer learning model on the pre-trained dataset to verify its generalization ability.The results showed that the transfer learning model improved detection accuracy for the pre-trained dataset.Specifically,the F1 scores for Pg and Sg phases increased to 0.842 and 0.866,representing improvements of 0.024 and 0.045 compared to the pre-trained model,respectively. In summary,this study demonstrates that transfer learning significantly reduces the reli-ance on large volumes of manually labeled data,thereby lowering the associated training costs and resource requirements.The transfer learning method in this study included data from differ-ent instrument types,enriching the training samples and enabling the model to learn a broader and more diverse range of waveform characteristics,thereby improving model's performance.The transfer learning model developed in this research can better utilize the data recorded by Early Warning Station Network for related research,such as velocity structure inversion and seismic activity analysis.

Abstract:Some seismic waves excited near the low-velocity subducting oceanic crust are trapped in the low-velocity layer and propagate as guided waves,which are recorded by forearc seismic stations and can be used to image oceanic crustal structures.The subducting oceanic crust plays a critical role in the exchange of material and energy between the Earth's surface and its interior.The subducting oceanic crust transports water into the Earth's interior,contributing to partial melting of the mantle wedge and island arc magmatism,and is therefore characterized by a low-velocity layer in its velocity structure.Studies of the subducting oceanic crustal struc-ture can enhance our understanding of processes of subduction metamorphism and subduction dynamics. When seismic waves are excited by a seismic source near the low-velocity oceanic crust at the top of the subducting slab,a portion of these waves is trapped in the low-velocity layer and propagates as guided waves,while another portion leaks into the high-velocity slab mantle and travels as head waves.The guided wave phases continue to interact with the oceanic crustal structure during propagation,and their arrivals and waveform characteristics carry information about the velocity and geometry changes of the oceanic crust.After propagating a certain dis-tance,the guided waves can leak out to the Earth's surface due to the curvature of the slab and be recorded at forearc seismic stations.In recent years,guided wave information,which is more sensitive to the low-velocity layer structure,has been widely used to image the seismic wave velocity structure.Structural characterization studies of the oceanic crust with guided waves have been conducted in many subduction zones around the world,including Japan sub-duction zone,Alaska-Aleutian subduction zone,Nicaraguan subduction zone,Peru-Chile sub-duction zone,and Hellenic subduction zone. The Alaska subduction zone,with abundant seismic events recorded by permanent seismic stations,is a natural testing laboratory for imaging oceanic crustal structure with guided waves. Waveforms from two intermediate-depth earthquake events,located at the top of the sub-duction seismic zone at the eastern and western sides of the Yakutat terrane,are used to analyze the propagation characteristics of guided wave phases in this study.Clear guided wave arrivals recorded by regional stations for both events show the significant influence of the subducting oceanic crust.When the epicentral distance is greater than 200 km,the amplitudes of the guided wave phases are often much larger than those of other phases in the wave trains,indicating that the low-velocity oceanic crustal structure in this area is very conducive to the generation and propagation of guided waves. The arrival delays of the guided waves increase with propagation distance in the Yakutat area at the eastern end of Alaska subduction zone,while no usual guided wave is observed in the Pacific slab area to the west of the Yakutat terrane.This implies that the Yakutat and Pacific oceanic crusts do not directly connect to each other and that the guided waves cannot effectively propagate from the former to the latter.Thus,the propagation characteristics of the guided waves effectively indicate the boundary between the Yakutat terrane and the Pacific slab. To understand the propagation characteristics of guided wave phases in the subducting Yakutat oceanic crust among ambient mantle,we additionally selected small earthquake events with focal depths＞70 km at the top of the subduction seismic zone on both sides of the Yakutat terrane.In the records of these earthquake events observed at the same station,the arrival time difference between the P-P and the S-S guided wave phases increases with the focal depth,indicating that the Yakutat oceanic crust is continuously distributed and subducts into the Earth's interior. The arrival times of the seismic phases were used to infer the apparent velocity anomalies of the guided wave phases on four profiles in the region.P and S guide wave apparent velocity anomalies of-19％and-15％are commonly inferred in the Yakutat area.Similar apparent velocity anomalies are found in the guided waves from seismic events at depths of 70-100 km on both the eastern and western sides of the Yakutat terrane,with the apparent velocity anom-alies on the western side of the Yakutat terrane are slightly smaller than those on the eastern side.These apparent velocity anomalies are reduced to-6％ in the seismic events recorded at depths of 100-130 km on the eastern side of the Yakutat terrane. The above results indicate that the Yakutat oceanic crust subducts into deep Earth in the region,and the velocity anomalies are reduced due to the dehydration of the Yakutat oceanic crust during its subduction.

Abstract:The seismic design of nuclear power plants not only needs to meet the requirements for structural safety but also ensure the safe operation of nuclear facilities in extreme situations such as earthquakes and prevent accidents like radioactive material leakage.Therefore,study-ing the seismic performance and against progressive collapse of the main plant in the conven-tional island has important practical significance.This article is based on the alternate path method and identifies the key columns in the gable structure using node displacement and bear-ing capacity methods.These key columns play a crucial role in the stability of the entire struc-ture under earthquake action.If a key column fails,it may cause the progressive collapse of the structure,posing a serious threat to the safe operation of the nuclear power plant.After identify-ing the key columns,a detailed analysis of the plastic hinges of the gable structure,as well as the vertical displacement and rotation changes of the failed columns,was conducted using stat-ic and dynamic nonlinear analysis methods.A plastic hinge is a nonlinear deformation form that may occur in structures under earthquake action,and its formation and development have a sig-nificant impact on the seismic performance of structures.By analyzing the vertical displace-ment and angle changes of the failed column,the against progressive collapse performance of the remaining structure can be more accurately evaluated.The results indicate that the failure of the edge columns of the key columns has a greater impact on the against progressive collapse of the remaining structure,which is consistent with the results of the component importance ana-lysis. In addition,the results of static nonlinear and dynamic nonlinear analyses were compared.The amplification effect of the load is considered in static analysis,so the predicted against pro-gressive collapse is generally higher than the results of dynamic analysis.This discovery sug-gests that in practical engineering,we cannot solely rely on the results of static analysis.In-stead,we should comprehensively consider dynamic effects to more accurately evaluate the progressive collapse performance of structures.Finally,three different optimization methods were proposed to enhance the overall safety margin of the structure.Using traditional methods,such as increasing the amount of reinforcement and enlarging the cross sectional size,to en-hance the structure's against progressive collapse can lead to an increase in material usage and construction difficulty.In addition to applying traditional methods,diagonal braces were also added and the support layout of the structure was optimized.It was found that adding diagonal braces can more effectively enhance the against progressive collapse and deformation ability of the frame and bent structure of the main plant in the conventional island.This optimization measure not only enhances the overall against progressive collapse performance of the structure but also offers better economic efficiency and construction convenience.Therefore,in the seis-mic design of nuclear power plants,multiple factors need to be comprehensively considered,including structural form,material selection,construction quality,etc.By optimizing the structural design and reinforcement measures,the seismic performance of the main plant in the conventional island can be significantly enhanced,thereby ensuring the safe operation of nucle-ar power plants under extreme conditions such as earthquakes. In summary,this study employed the alternate path method to identify key columns and adopted static and dynamic nonlinear analysis methods to comprehensively explore the against progressive collapse ability of the gable wall structure of the main plant in the conventional is-land.The impact of key column failure on structural performance was revealed,and effective structural optimization measures were proposed,providing an important theoretical basis and practical guidance for seismic design and reinforcement in nuclear power plants.Research on the seismic performance of future nuclear power plants will pay more attention to the in depth integration of structural engineering,seismology,materials science,and intelligent techno-logy.This is to comprehensively optimize the seismic design of nuclear power plants,further enhance their seismic safety,and provide strong guarantees for the sustainable development of nuclear energy.

Abstract:Daocheng County located on the southeastern edge of the Qinghai-Xizang Plateau is characterized by frequent geological tectonic activities,complex terrain,and variable climate.Landslide are one of the major natural hazards in this region.Understanding the mechanisms and patterns of landslide occurrences in this area is an effective way to reduce the losses caused by geological disasters.The ancient landslide database and its spatial distribution are important foundations for studying the mechanisms of landslide occurrences,but there are still data gaps.Using the Google Earth and visual interpretation through human-computer interaction,a detailed identification of ancient landslides in Daocheng County was carried out.A total of 1 324 landslides were identified,with the largest landslide covering an area of 2.36 km2 and the smallest covering about 1 000 m2,with a total landslide area of 306.78 km2.The scale of an-cient landslides in Daocheng County is primarily large-scale landslides(66.39％),while small and medium-sized landslides are fewer(33.61％). The Daocheng landslide is influenced to varying degrees by factors such as slope,aspect,river,fault,lithology,historical earthquakes and PGA.① The study found that there is a sig-nificant correlation between landslide distribution and altitude in Daocheng County through the analysis of spatial distribution.Most of the study area is concentrated between 4 km and 5 km in elevation,but landslides are mainly concentrated within the range of 2-3 km.With increasing elevation,both the landslide area percentage(LAP)and landslide number density(LND)show a trend of first increasing and then gradually decreasing,reaching a peak in the elevation range of 2-2.5 km.In this range,there are 193 landslides accounting for 14.58％of the total land-slides,and the LAP and LND are 54.74％and 2.35/km2,respectively.By analyzing the topo-graphic and geomorphological maps,it can be inferred that high-altitude areas,due to their gentle terrain,lack the necessary geographical conditions for landslide occurrences,while the steep valleys in middle-and low-altitude areas are more prone to landslides.② Slope is another factor that strongly controls the distribution of landslides.The overall slope in the study area is generally gentle,with areas having a slope of less than 30° accounting for 76％of the total study area.A total of 648 landslides are distributed in these low-slope areas,representing 48.94％of the total number of landslides in the region.LAP and LND show a trend of first in-creasing and then decreasing with the increase of slope.LAP peaks in the slope range of 30°-40°,while LND peaks in the slope range of 60°-70°.With the increase of slope,the density and area percentage of landslides rise significantly.This is mainly because the greater the slope,the more pronounced the effect of gravity,making soil and rock masses more prone to collapse.③ Rivers exert a significant control on the distribution of landslides.As the distance from rivers increases,the number and area of landslides show a clear downward trend.From a distance of＜5 km to 5-10 km from rivers,the number of landslides sharply drops from 776 to 257,and the landslide area also decreases by two-thirds,indicating that rivers have a strong control effect on landslides.The area within 5 km from the river is a high-incidence area for landslides,accounting for 58.61％of the total number of landslides and about 63.08％of the area,especially along the valleys of the Dongyi River and Chitu River.This is mainly due to the continuous erosion of the valley bottom by rivers,which weakens the support on both sides and increases the likelihood of landslides.④ Aspect also plays a significant role in the distribu-tion of landslides.The northeastern aspect has the highest number of landslides,with 229 land-slides accounting for 17.30％of the total,and the landslide area also reaches the maximum value of 53.83 km2,accounting for 17.55％,indicating that landslides are more likely to occur in this aspect.This is followed by the west,northwest,southwest,and north aspects,with landslide number of 215,191,190,and 165,accounting for 16.24％,14.43％,14.35％,and 12.46％,respectively.⑤ In terms of lithology,landslides are mainly distributed in Triassic strata,which are dominated by low-strength soft rocks,making them more prone to landslides.⑥ In the history of Daocheng County,only one Ms5.2 earthquake has occurred,located near the Dongyi River basin in southern Daocheng County.By comparing the landslide distribution density with the location of the earthquake,it can be found that the landslide density near the epicenter is higher than that in other areas.This indicates that earthquakes have a certain influ-ence on the development of landslides,especially the disturbance of geological structure caused by the earthquake in the short time after the earthquake.⑦Daocheng County is primarily divided into two PGA(peak ground acceleration)zones:0.15g in the northeast and 0.lg in the southwest.The number of landslides in these zones is 119 and 1 205,respectively,with LAPs of 1.06％and 5.6％and LNDs of 0.06/km2 and 0.24/km2,respectively.Usually,the number of landslides increases with an increase of PGA because higher PGA indicates stronger seismic ground motion,leading to greater damage to slopes.However,the situation in Daocheng County is the opposite,suggesting a low correlation between landslide distribution and PGA.The density of landslides and the location of faults also do not show a high correlation. In conclusion,the distribution of landslides in Daocheng County is affected by multiple factors,with topography,geomorphology,and geological conditions being the main con-trolling factors.Seismic activity also plays a role in certain local areas.The correlation between landslide distribution and PGA,as well as active faults,is not high.Of course,a deeper under-standing of the mechanisms inducing landslides requires further research.Establishing a land-slide database is a key step in scientifically and systematically addressing landslides.By adopt-ing advanced data-collection technologies,establishing a comprehensive database,and imple-menting effective management measures,more accurate and complete landslide information can be provided to society.This helps to reduce the risk of landslide disasters and provides solid data support for subsequent researches.

Abstract:The Tanlu fault zone is the most active fault zone in eastern China,exhibiting multi-phase,complex,and segmented activity characteristics.Based on its geometric structure and activity,the Tanlu fault zone can be divided into four sections from south to north:the Jiashan-Guangji,Weifang-Jiashan,Xialiaohe-Laizhouwan,andHegang-Tielingsections.Amongthese,the Weifang-Jiashan section displays the strongest tectonic activity,within which the F5 fault(Anqiu-Juxian fault)is identified as the most recent Holocene active fault.Studying its shallow tectonic characteristics is critical for understanding fault evolution and assessing seismic hazards. Domestic and international scholars have reached a consensus on the developmental scale,strike,and activity characteristics of the F5 fault.However,rapid and effective methods for detecting fault zone width and internal medium structure remain lacking.While core drilling offers high precision,its high cost and lengthy duration limit widespread application.Reflec-tion wave exploration can locate fault points by analyzing co-phase axis deformation and dif-fracted wave development,but it fails to resolve detailed internal medium structures.Advances in observation technology and data processing now make high-resolution surface wave explora-tion a promising solution. Rayleigh wave exploration,an emerging geophysical method,enables rapid and cost-effective shear wave velocity measurements,making it a research hotspot.Rayleigh wave methods are categorized into active-source and passive-source techniques based on seismic excitation modes.The active-source method yields high-frequency signals but has limited depth resolution,whereas the passive-source method provides low-frequency signals for deeper exploration.Combining both approaches enhances effective exploration depth while maintain-ing shallow-structure accuracy. This study employs a joint detection method for active and passive surface waves,acquir-ing data using identical array configurations for both sources.Dispersion spectra are superim-posed to improve dispersion curve quality and expand the observable frequency range.A genet-ic algorithm inverts the 2D S-wave velocity structure. Results reveal significant lateral heterogeneity in the crustal medium,with a segmented S-wave velocity structure.Fault points are identified at distances 110 m(F5-1-1),160m(F5-1),and 640 m(F5-2).The F5-1 and F5-2 faults form opposing boundary faults(dip angles 70° and 65°,respectively),creating a graben structure with a 330 m-wide fracture zone.The western boundary is not a single fault but a composite zone comprising F5-1 and its secondary fault F5-1-1.The fracture zone exhibits overall low S-wave velocities but contains localized high-velocity anomalies,likely reflecting F5 fault activity. Comparisons with shallow seismic exploration and borehole profiles confirm that the joint method accurately locates main fault surfaces,while also providing clearer delineation of frac-ture zone width and internal material structure.This approach is particularly effective for shal-low faults or high-interference environments like urban areas.

Abstract:Broadband ground motion simulation is a vital research area in earthquake engineer-ing,providing a critical basis for seismic hazard assessment and engineering design.Despite significant advancements in simulation techniques,existing deterministic physical models still face major limitations,particularly in accurately reproducing high-frequency ground motion characteristics.These shortcomings stem primarily from constraints in source model refinement and the substantial computational resources required for broadband,high-accuracy simulations.However,accurately capturing these high-frequency components is crucial for the seismic design of structures and for ensuring infrastructure safety during earthquakes. To address these challenges,this study integrates the GP15.4 kinematic hybrid source model with a regional site model that incorporates surface topography within the spectral ele-ment method(SEM)framework.This integrated approach combines the broadband strengths of the GP15.4 model with the high accuracy and computational efficiency of SEM.The GP15.4 model achieves its broadband capability by superimposing high-frequency stochastic elements onto a low-frequency deterministic slip model.Specifically,the deterministic slip distribution is converted into the wavenumber domain via a two-dimensional Fourier transform.High-fre-quency stochastic components,conforming to a von Karman autocorrelation function spec-trum,are then generated and superimposed.An inverse Fourier transform reconstructs a spa-tially heterogeneous slip distribution.Rupture time and rise time for each subfault are derived from this slip distribution,with consideration given to depth-dependent rupture velocity vari-ations and the introduction of random perturbations to better simulate high-frequency radiation.The final rupture model is implemented in the SPECFEM3D spectral element code,enabling a complete simulation from fault rupture to seismic wave propagation. To validate the proposed methodology,we first conducted a simulation for a region in western China using a flat-surface model.The results,compared against NGA-West2 ground motion prediction equations(GMPEs),predominantly fell within their expected variability ranges.Subsequently,the model was applied to simulate the 2016 Mw6.5 Norcia,Italy earth-quake.Comparisons of acceleration time histories and response spectra at ten observational sta-tions demonstrated the method's effectiveness and accuracy.These validations confirm that the GP15.4 model within the SEM framework significantly enhances the realistic simulation of ground motion characteristics that are challenging for traditional deterministic methods.Further-more,wavefield snapshots provide visualization of seismic wave propagation from the fault to the surface,elucidating key features such as the concentration and rapid attenuation of near-fault motions and the effects of surface topography. This study establishes a framework applicable to other seismically active regions world-wide.By combining an advanced hybrid source model with SEM,this approach ensures a com-prehensive understanding of seismic wave propagation,making it a valuable tool for both theor-etical research and practical engineering applications.With ongoing advancements in computa-tional power,this method holds promise for delivering even more accurate simulations across broader frequency ranges and under more complex site conditions,thereby offering substantial support for future seismic risk assessment and resilient infrastructure design.

Abstract:Due to significant interference from the operation of the Baoji-Lanzhou High-Speed Railway(with the nearest electrode only 600 meters away),the reliability of surface geo-resistivity observation data at Tongwei seismic station has decreased.Therefore,a downhole georesistivity observation system was built in 2021 to ensure the continuity of observations.Using existing potential analytical expressions and programs,we analyzed the variation of the influence coefficients of downhole georesistivity observations at Tongwei station with electrode burial depth and current electrode spacing,based on the electrical sounding inversion data of the station.The results indicate that:① With the power electrode spacing fixed at Tongwei sta-tion,the influence coefficients of each layer show complex variation with electrode burial depth;② With the electrode burial depth fixed,the shallow-layer influence coefficients typic-ally first increase and then decrease,whereas the influence coefficients of deep media rises as the current electrode spacing increases.For quantitative analysis,specific observation paramet-ers of Tongwei station(the current electrode spacing AB=550 m,the electrode burial depth h=100 m)were adopted.The results show that the influence coefficient B2 of the second-layer medium is the most significant contributor to georesistivity,indicating that variations of the lay-er's true georesistivity govern the observed values.Further comparative analysis of the down-hole and surface observation data at Tongwei station shows that the dispersion degree of the downhole observed georesistivity is significantly lower than that of the surface,reflecting that the downhole observation has a strong ability to suppress surface and shallow interference,thus verifying the rationality and practicality of the downhole observation electrode layout design. Downhole observations aim to suppress surface interference and emphasize resistivity changes in deep rock masses associated with earthquake preparation.Through comparative ana-lysis of surface and downhole data,it is found that the apparent resistivity of Tongwei station's downhole N20°W and EW channels has shown a synchronous and rapid decline since October 6,2023.This stands in contrast to the normal annual variation pattern——during this period,the data should otherwise be in the rising phase——with the respective anomaly amplitudes of the two channels reaching-0.29％and-0.08％.During the recovery process of the measured values,the Jishishan Ms6.2 earthquake occurred on December 18,2023,approximately 230 km away from Tongwei station.The aforementioned analysis indicates that the influence coefficient of each media layer exhibits a complex variation trend with electrode burial depth and power electrode spacing.When the power electrode spacing is fixed,the influence coeffi-cient does not always change monotonically with the increasing electrode burial depth.Simil-arly,with fixed electrode burial depth,the variaiton of the influence coefficient with power electrode spacing is also relatively complex.However,once the electrode burial depth reaches a certain threshold,it can effectively suppress surface interference and reflect the seismogenic activities information in deep media.This suggests that the downhole observation of Tongwei station can highlight the resistivity variation of deep rock mass induced by earthquake prepara-tion. Based on the theoretical analysis of the influence coefficient for the horizontal layered me-dium model at Tongwei station,this study reveals the mechanism underlying the differences in annual variation amplitudes between surface and downhole observations.It proposes an opti-mized layout scheme for the burial depth and current electrode spacing of downhole observa-tion,and finally evaluates the downhole measuring devices at Tongwei station.The main con-clusions are as follows:① Both the surface and downhole observation data of Tongwei station exhibit a normal annual variation pattern of"lower in summer and higher in winter".The geor-esistivity curves of the two measuring channels in downhole observations show good consist-ency and synchronization,with an annual variation amplitude only about 1/3 of that of surface georesistivity.This indicates that when the electrode burial depth h≥100 m,surface and shal-low-layer interferences can be effectively suppressed,and the annual variation amplitude of observation data can be significantly reduced;② Theoretical analysis of influence coefficients shows that it is reasonable to set the electrode burial depth h of downhole observations at Tong-wei station within the range of 100-150 m and half of the current electrode spacing AB/2 within 200-300 m,which further confirms that the electrode layout design for downhole observations at Tongwei station(h=100 m,AB=550 m)is quite ideal.③ With the increase of electrode urial depth,the ability to suppress surface interferences is enhanced,and the georesistivity changes of deep media is highlighted simultaneously.(4)The influence coefficient of the first-layer medium on surface observations at Tongwei station is 0.64％,while that in downhole ob-servations is only 0.15％.Meanwhile,the influence coefficient of the third-layer medium on surface observations is lower than that in downhole observations.This demonstrates that the electrode layout design for downhole georesistivity observations at Tongwei station is relatively rational,which can provide theoretical basis and practical reference for the design and construc-tion of downhole georesistivity stations in China.

Abstract:This study analyzes earthquake catalogs within the range of 29°N-37°N and 117°E-124°E from 2014 to 2023.Utilizing the lattice searching method,we focusing on the spatial and temporal variations of the b-value before and after the 2021 Ms4.2 Tianning,Jiangsu earthquake.Additionally,apparent stress is calculated and analyzed using seismic waveforms data from the Shanghai Seismic Network for the same period.By integrating the temporal changes in both b-value and apparent stress,this research reveals the evolution of regional stress before and after the Tianning earthquake,thereby providing valuable insights for forecasting future earthquakes of Ms4.0 and above in the study region. The completeness and reliability of the earthquake catalog are essential for accurate b-value estimation.Following the removal aftershocks,we determined the spatial distribution of the minimum complete magnitude Mc.Based on the G-R relationship in the study area,the minimum complete magnitude was set at ML2.0,and the upper limit magnitude was taken as ML4.3.A total of 1 489 seismic events were selected under these criteria. By analyzing the spatial and temporal variations in b-value before and after the Tianning earthquake and comparing them with those associated with the 2021 Ms5.0 Dafeng Sea area,Jiangsu earthquake,which occurred about 200 km away,35 days interval,we found that both earthquakes occurred at or near the edges of the b-value drop regions.Notably,the b-value anomaly zones prior to these two earthquakes did not overlap,suggesting distinct regions of stress accumulation.The temporal evolution of b-value in these anomaly zones also displayed significantly different.Prior to the Dafeng Sea area earthquake,the b-value declined over a rel-atively short period of about 7 months,followed by a rapid decrease after the mainshock.In contrast,before the Tianning earthquake,the b-value exhibited a prolonged decline lasting about 21 months,with a clear and gradual recovery in the post-seismic period.These differing evolutionary patterns may be attributed to factors such as regional tectonic setting and variations in stress release heterogeneity between the two events. Apparent stress values for 134 ML≥2.0 seismic events were determined by calculating their seismic moments and radiated energies.A linear relationship between apparent stress and magnitude was identified in the study area,with a correlation coefficient of approximately 0.66,indicating that larger magnitudes generally correspond to higher apparent stress.To more accurately characterize the dynamic changes in apparent stress surrounding the Tianning earth-quake while minimizing the influence of magnitude,the differential apparent stress method was applied.We analyzed the differential apparent stress and its average value for 56 earthquakes in the magnitude range 2.0≤ML＜3.0 and 36 earthquakes in the range of 3.0≤ML＜4.0 located near the Tianning earthquake,before and after the event.Pronounced temporal variations were observed:the average differential apparent stress increased prior to the earthquake and decreased afterward.The timing of the stress increase varied with magnitude.For events with magnitude range of 2.0≤ML＜3.0,the differential apparent stress began to rise significantly in January 2020,whereas for those with the magnitude range of 3.0≤ML＜4.0,the increase star-ted in May 2020. The spatial and temporal variations in b-value are influenced by factors such as medium in-homogeneity and fault type,while apparent stress is also affected by seismic source mechan-ism.Consequently,accurately determining stress changes based on either parameter alone is challenging.However,the Joint analysis of b-value and apparent stress can effectively reduce uncertainty and provide more reliable constrains on regional structural stress evolution.The observed characteristic decrease in b-value,coupled with the increase in apparent stress prior to the Tianning earthquake,indicates a state of stress accumulation in the pre-seismic period.The subsequent stress release during the earthquake led to an increase in b-value and a simultaneous decrease in apparent stress.These findings confirm that the combined changes in b-value and apparent stress can serve as effective indicators for monitoring the process of stress accumula-tion and release associated with earthquakes. Therefore,the long-term and continuous monitoring of b-value and apparent stress provides a robust approach for structural stress monitoring.Specifically,the timely identifica-tion and tracking of the key phase characterized by a decreasing b-value and a rising apparent stress can effectively reveal ongoing stress accumulation processes.Consequently,this method-ology holds significant potential for improving the prediction of earthquakes above Ms4.0 in the Shanghai monitoring region.

Abstract:The multi-scale structure of the Earth's interior contains important information on plate tectonics and geological processes,making the development of reliable models a critical objective in seismological research.Compared to conventional travel time tomography based on high-frequency ray theory and limited seismic phase arrival time information,waveform inver-sion utilizes the dynamical properties in the waveforms such as the frequency-dependent phases and amplitudes to constrain model parameters,thus representing the most promising approach for constructing high-resolution structural models.Waveform inversion is inherently nonlinear,thus blind use of all records without proper screening during iterative inversion leads to diffi-culties in convergence or getting trapped into local minima.Therefore,one of the key factors for successful waveform inversions is the selection of time windows of signals to constrain the intended structural targets.Proper selection of time windows is crucial for enhancing the utiliza-tion of effective information,reducing noise interference,and improving the accuracy and sta-bility of inversions. To efficiently and conveniently extract effective waveform information,we develop the SeismoClip,an open-source waveform data automatic selection program.SeismoClip is a MATLAB-based interactive tool,designed for waveform inversion.It can be used not only as an independent window selector but also as a post-processing tool for other automatic window selection programs.Various auxiliary functions,such as array analysis,multi-frequency-band filtering,and short-time-average/long-time-average(STA/LTA)analysis,are also employed to significantly improve the accuracy and robustness of window selection performance.Further-more,SeismoClip emphasizes user experience with extensive customizability,allowing re-searchers to make modifications to adapt program functions according to specific research needs.The core function of SeismoClip is to interactively process and select time windows from the input synthetic and observed seismic waveforms,and generate the corresponding window files.The processing flow is mainly divided into four steps:data processing,array analysis,in-teractive window selection,and automated adjustment. In recent years,the widespread deployment of seismic stations has led to the availability of huge volumes of waveform records from densely distributed seismic arrays.The waveform similarity of the same seismic phase between stations and the variation of phase arrival times with epicentral distance can help us determine a reasonable interval for the seismic phase win-dow.Therefore,we have added an array-based module to the SeismoClip to assist in the ana-lysis and extraction of phase windows.Since recorded waveforms are affected by noise and thus have a lower signal-to-noise ratios compared to theoretical seismograms,phase information in records can easily be obscured by noise.Therefore,SeismoClip performs array analysis based on the STA/LTA curve calculated from noise-free synthetic seismograms generated from the same source parameters. During the interactive selection process,SeismoClip displays both observed and synthetic waveforms in array form for reference.The interactive selection process is applied to single-channel waveforms.SeismoClip offers a variety of interactive selection interface options to accommodate different user needs,a including multi-frequency-band filtering single-channel selection interface,an interface highlighting weak-amplitude body-wave phases,and an inter-face based on existing selection results.This allows users to precisely extract seismic phase win-dows based on waveform data.For the interactively selected windows,SeismoClip performs a series of automated adjustments under the constraints of existing window selections to achieve a better fit.Users can choose either directly use the results of the interactive selection or allow the program to make automatic adjustments. In this article,we compare the selection results of SeismoClip with those from the auto-mated time-window selection algorithm FLEXWIN,and discuss the distribution of sensitivity kernels corresponding to different time-windows to highlight the importance of precise time-window selection.Comparative analysis with real data shows that the time windows selected by SeismoClip are generally more accurate and reasonable than those selected by automated selec-tion,particularly in selecting windows from waveforms with low signal-to-noise ratios. Overall,SeismoClip represents as a convenient and effective time-window selection tool for waveform inversion,based on array analysis of specific seismic phase tracking and multi-frequency-band analysis.Moreover,by real-time evaluation of the similarity and cross-correla-tion travel time differences between the observed and synthetic waveforms within the selected time window,it is possible to effectively avoid cycle skipping in waveform inversions.In cases where waveform features are not prominent,SeismoClip can be used to carefully select the time window to obtain appropriate sensitivity kernels and enhance the stability and accuracy of the inversions.For users utilizing an automated time-selection algorithm,we suggest starting with less restrictive parameters to conduct an initial window selection,and then utilizing the"inter-active selection interface based on the eaisting selection results"of SeismoClip to review and refine the initial selection results.

Abstract:Missing seismic traces are inevitable in marine seismic data acquisition due to vari-ous objective factors.Cable towing issues,for instance,can arise from sudden shifts in ocean currents that pull the cables off their intended paths,or from contact with uneven seabed fea-tures that cause temporary disruptions in data transmission.Ocean currents themselves,with their varying velocities and directional changes,exert continuous forces on the acquisition equipment,leading to intermittent loss of signal reception.Weather conditions such as storms or strong winds not only affect the stability of the survey vessel but also create background noise that interferes with seismic wave detection,resulting in missing data points.Additionally,restricted mining areas,where access is limited by environmental regulations,territorial bound-aries,or existing infrastructure,prohibit the placement of sensors in critical locations,leaving gaps in the data coverage.These missing traces can significantly compromise subsequent data analysis:reservoir characterization may misjudge the size and distribution of potential hydrocar-bon reservoirs,and structural interpretation might fail to identify key geological features like faults or folds.Existing reconstruction methods,mostly based on single-algorithm frame-works,struggle to address these issues effectively.Some rely solely on wavelet-based algorithms,which lack the ability to capture directional features of seismic events,while oth-ers use Fourier transform methods that perform poorly with non-stationary data.These single-algorithm approaches often suffer from drawbacks such as slow convergence,where hundreds of iterations may be needed to achieve a stable result,or high computational costs that make them impractical for processing the massive datasets generated by modern high-resolution seis-mic surveys. To address this issue,this study proposes a joint reconstruction method that combines the accelerated linear Bregman method(ALBM)and the iterative shrinkage-thresholding algorithm(ISTA),with Curvelet transform employed as the sparse basis.The choice of Curvelet trans-form is rooted in its unique set of properties that make it ideal for seismic data processing.Its multi-scale nature allows it to decompose seismic signals into different frequency bands,each corresponding to geological features of varying sizes:from large-scale tectonic structures to small-scale reservoir heterogeneities.Directionality enables it to capture seismic events propagating in different directions,such as dipping reflectors that indicate subsurface strati-graphic layers.Sparsity ensures that complex seismic signals can be represented with a relat-ively small number of non-zero coefficients,reducing computational load,while anisotropy makes it highly effective at capturing curve and edge features,such as the boundaries of faults or reservoir edges,with far fewer coefficients than alternative transforms like the wavelet trans-form. The joint algorithm is designed to leverage the complementary strengths of ALBM and ISTA in a structured workflow.In the early iterations,ALBM takes the lead by retaining a lar-ger proportion of unthresholded Curvelet coefficients.These coefficients,which have not been filtered by thresholding operations,contain most of the energy from the main seismic events.By preserving them,ALBM accelerates the convergence process,allowing the algorithm to quickly approximate the general structure of the missing data.As the iterations progress,the algorithm shifts to relying more on ISTA.In the later stages,the remaining signals to be recovered are often weak valid signals,such as reflections from deep reservoirs,which were overshadowed by stronger events in the early iterations.ISTA excels here by applying a series of shrinkage and thresholding operations to the Curvelet coefficients,gradually isolating and enhancing these weak signals,thereby improving the overall reconstruction accuracy. A key innovation of this method is the development of a new acceleration factor that dynamically adjusts from one to two as the iterations proceed.This factor directly controls the proportion of unthresholded Curvelet coefficients retained at each step.In the early iterations,when the factor is closer to one,a higher percentage of unthresholded coefficients is kept,ensuring that the main seismic events are captured quickly.As iterations continue and the factor increases toward two,the proportion of unthresholded coefficients is reduced,allowing ISTA to focus on refining the signal by removing noise and recovering weak features.This dynamic adjustment stands in contrast to traditional linear acceleration factors,which change at a fixed rate and often retain too few coefficients in early stages,and typical acceleration factors that lack flexibility,leading to either slow convergence or loss of weak signals.By enhancing the retention of effective signals in early iterations,the new acceleration factor significantly boosts both convergence speed and reconstruction precision. Theoretical simulations were conducted under three representative scenarios to validate the method's superiority:non-uniform undersampling,uniform undersampling,and noisy condi-tions.In the non-uniform undersampling test,where 40％ of traces were randomly missing,a common scenario in real acquisitions,the proposed method achieved a signal-to-noise ratio(SNR)of 21.5 dB in just 20 iterations.This outperformed traditional linear acceleration factor methods,which required 35 iterations to reach 21.2 dB,and typical acceleration factor meth-ods that took 30 iterations to achieve 21.0 dB.It also showed better alias suppression,with a 30％ reduction in alias energy compared to linear factors,as measured by spectral analysis,and lower reconstruction errors(root mean square error of 0.021 versus 0.035 for linear factors).Under uniform undersampling with 30％ missing traces,simulating systematic equipment fail-ures,the method reached an SNR of 19.8 dB in 18 iterations,preserving the continuity of seis-mic events more effectively than linear(28 iterations,19.5 dB)and typical(24 iterations,19.3 dB)factors.In noisy conditions,where Gaussian noise with a variance of 0.01 was added to 40％undersampling data,the method maintained an SNR of 18.2 dB,demonstrating strong noise resistance compared to linear(16.5 dB)and typical(17.1 dB)factors. Application to real seismic data from deep-sea block A yielded promising results.Block A is located in a geologically complex region with water depth ranging from 1 500 to 3 000 m.The seismic data from this area is characterized by strong multiples,which owns the unwanted reflections from the sea surface and seabed weak reflections from deep reservoirs located over 4000 meters below the seabed,and a network of complex faults that complicate signal interpretation.The dataset used in the application had a sampling rate of 4 milliseconds,a trace interval of 25 meters,and a total of 2 000 traces,covering an area of approximately 50 square kilometers.For 50％randomly undersampling data with 16 consecutive missing traces,which is a challenging scenario that often breaks the continuity of seismic events,the method recon-structed the missing information effectively,restoring the continuity of reflectors across the gaps. Compared with methods using linear and conventional acceleration factors,the proposed method showed clear advantages.After 15 iterations,it achieved an SNR of 13.95 dB,signi-ficantly higher than the 12.30 dB of the linear acceleration factor method and 12.85 dB of the typical acceleration factor method.Energy preservation analysis revealed that it retained over 90％of the energy in the main frequency band(10-60 Hz),which is critical for identifying sub-surface features,compared to 82％for the linear method and 85％for the typical method.It also preserved weak deep reservoir reflections more effectively,with visual inspections showing clearer and more continuous weak signals in the reconstructed sections.In terms of computa-tional efficiency,the method reached an SNR of 13.5 dB in just 73 seconds,saving nearly two-thirds of the time compared to the typical acceleration factor method,which required 220 seconds,and outperforming the 185 seconds of the linear factor method. In summary,the proposed method,with its new acceleration factor,achieves an optimal balance between reconstruction speed and precision.It has demonstrated its ability to adapt to complex scenarios such as uniform missing,non-uniform missing,and noise interference that all common in marine seismic acquisition.By meeting the dual requirements of high efficiency and high precision,it provides a practical solution for processing large-scale marine seismic datasets.Future work will focus on optimizing the acceleration factor's adjustment strategy to handle even more complex geological conditions,such as areas with severe faulting,and to explore its application in 3D seismic data reconstruction,where the data volume is exponen-tially larger and the challenges are more pronounced.

Abstract:In the context of near-fault pulse-like ground motions,significant characteristics are often observed,including energy concentration,a prevalence of low-frequency components,and extended periods.These characteristics can have a substantial impact on the seismic response spectra.Therefore,in order to fully consider the influence of near-fault pulse-like ground motions on the design response spectra for seismic isolation structure(hereinafter re-ferred to as the seismic isolation design spectra),near-fault pulse-like ground motions are selec-ted as the research object,and based on the relevant provisions on the seismic isolation design spectra in the Standard for Seismic Isolation Design of Buildings(GB/T 51408-2021),the research on the seismic isolation design spectra of near-fault regions is carried out.For the pur-pose of this study,140 sets of near-fault pulse-like ground motions were meticulously selected from the NGA-West2 database.The selection of these records was based on the criteria,which included a moment magnitude of ≥5.0 and a fault distance ≤40 km,and the velocity pulse identification method based on the wavelet transform was used in the selection process.Based on the seismic motions that have been selected,the seismic response spectra were computed with assurance rates of 50％(average spectrum)and 84％(average spectrum+1 standard devi-ation spectrum),respectively.The actual near-fault seismic response spectra were calculated and then analyzed in comparison with the current seismic isolation design spectra.The results revealed that the current seismic isolation design spectra in China have not fully considered the characteristics of near-fault pulse-like ground motions.This is manifested in the lower spectral values of the current seismic isolation design spectra compared with those of the actual near-fault seismic response spectra in some periods,which may lead to insufficient redundancy for structures in the near-fault regions.In light of the analysis results,we propose a method for adjusting the seismic isolation design spectrum in the near fault regions.The proposed method aims to enhance the seismic safety of seismic isolation structures in this region by optimizing the amplitude enhancement coefficient,characteristic period,and attenuation index of the seismic isolation design spectra.This objective is pursued by retaining the three-stage formula specified in the current seismic isolation design spectra.Firstly,it is proposed to increase the amplitude enhancement coefficient to 1.35 to reflect the energy concentration effect of near-fault pulse-like ground motions,so as to enhance the seismic performance of the structures under seismic actions.Additionally,a recommendation is made to adjust the lower limit of the characteristic period corresponding to earthquakes of different fortification levels.The adjusted lower limits are as follows:0.65 s for fortification earthquakes,0.70 s for rare earthquakes,and 0.75 s for extremely rare earthquakes.These adjustments are intended to comprehensively consider the low-frequency characteristics of near-fault pulse-like ground motions and its substantial impact on structural response.Furthermore,it is recommended that the attenuation index of the des-cending section of the seismic isolation design spectrum curve be adjusted to 0.8.This adjust-ment would serve to more accurately reflect the characteristics of the seismic response in the long period section.Finally,the finding suggests that the values of other relevant parameters continue to adhere to the provisions stipulated in the current Standard for Seismic Isolation Design of Buildings,thereby ensuring the coordination and consistency of the aforementioned parameters.This modification method is useful,providing a reasonable reference for the seis-mic isolation design of structures in near-fault regions.

Abstract:The southeastern Yunnan arcuate structual belt refers to three arc-shaped faults located on the southeastern margin of the Sichuan-Yunnan block,including Qujiang fault,Shiping-Jianshui fault and the central-southern segment of Honghe fault zone.The three faults extend in an arc(in plan view),striking generally in the NW-WNW direction and being arranged from north to south.The eastern side of the region is bounded by the south-central seg-ment of the sinistral nearly north-south Xiaojiang fault zone.Under the combined action of the northeastward extrusion of the Indian Plate,the southeastward escape of the plateau material,and the blocking of the South China block,the tectonic deformation is strong and the seismic activity is frequent in the arcuate structural belt.As the frontier zone for the southeastern slid-ing of the Sichuan-Yunnan block,the special geological structure of the arcuate structural belt may serve as a natural laboratory to constrain and validate the regulatory role of the Honghe fault zone in the evolutionary process of the southeastern margin of Qinghai-Xizang Plateau.GNSS technology enables the quantitative characterization of fault activity,facilitating the un-derstanding of the motion patterns and stress accumulation states of the Honghe fault zone and its surrounding areas.This contributes to elucidating the process of strong earthquake incuba-tion and occurrence in the region,providing reference for future seismicity trend forecasting. This paper builds upon the work of previous researchers,collecting and organizing mul-tiple GNSS observation data sets from the study area since 1999 onwards.Based on the existing GNSS velocity field in the Yunnan region,the regional strain rate fields for the periods 1999-2007,2009-2014 and 2015-2020 are estimated using Kriging interpolation,revealing the current characteristics of regional deformation evolution in the southeastern Yunnan arcuate structual belt.By analyzing the GNSS profiles,the fault slip rate of the central-southern seg-ment of Honghe fault zone,the Shiping-Jianshui fault and the Qujiang fault are derived by fit-ting the elastic dislocation model.Based on the existing GNSS continuous observation data in the study area,five sets of regional blocks crossing the central and southern segments of Honghe fault zone,the Shiping-Jianshui fault,the Qujiang fault,and the central-southern seg-ment of Xiaojiang fault zone were constructed to derive the block strain time series.These ana-lyses yield the following conclusions: 1)Based on the regional strain rate field results,the principal strain rate exhibits near NE-trending extension(tensile)and near NW-trending compression(compressional).This indic-ates that the material from the Qinghai-Xizang Plateau is being extruded and escaping toward the southeastern end of the Sichuan-Yunnan block.Blocked by the South China block,this escape leads to NW-trending compression and NE-trending extension,with tension being the dominant regime.It is inferred that the southeastern Yunnan arcuate structual belt may be a cur-rently forming zone of extension-shear tectonics. 2)The profile analysis results show that the central-southern segment of Xiaojiang fault zone exhibit left-lateral strike-slip with extension,with a slip rate of 9.14 mm/a.The Shiping-Jianshui fault,the central-southern segment of Honghe fault zone all exhibit right-lateral strike-slip characteristics,but overall slip rate is relatively low.The central segment of Honghe fault zone has a slip rate close to zero,indicating a state of slow creep.In the wedge-shaped area of southeastern Yunnan,differential movement between the two boundary faults(the central-south-ern segment of Xiaojiang fault zone and the central segment of Honghe fault zone)is accommodated by secondary faults within the wedge through dextral shear and block rotation.The strike-slip rate of the central-southern segment of Honghe fault zone is smaller than that of other secondary faults in the study area,suggesting that its role as the southwest boundary of the Sichuan-Yunnan active block has been significantly weakened. 3)From the trend of cross-fault block strain,the principal strain of Xiaojiang fault zone is in a distinctly tensile state.The strain of the Qujiang fault is slightly lower than that of the southern segment of Xiaojiang fault zone,indicating weak extensional movement.The south-em segments of Shiping-Jianshui fault and the Honghe fault zone show weakened strain,dom-inated by weak extrusion movement.The central segment of Honghe fault zone represents a low strain area.Within the arcuate structual belt,the maximum shear strain gradually increases from south to north.Analysis of the first shear strain time series reveals that the three NW-trending faults in this area are characterized by right-lateral shear.The shear strain of the Shiping-Jianshui and Qujiang faults is significantly stronger than that of the southern segment of Honghe fault zone,with central segment of Honghe fault zone being the least active.This fur-ther confirms that the central-southern segment of Honghe fault zone are not the main stress-bearing faults in this area,and cannot independently constitute a boundary zone of active tec-tonic units.Deformation of the upper crust on the southeastern margin of the Sichuan-Yunnan block is now distributed among secondary faults within the southeastern Yunnan wedge-shaped tectonic zone,with the Shiping-Jianshui and Qujiang faults being the main stress-bearing faults.This result may provide a preliminary explanation for the difference in seismic potential among the three faults in the southeastern Yunnan arcuate structural belt.

Abstract:There are often differences in the physical and mechanical properties of slope topo-graphy and its constituent rocks in nature.Therefore,under strong seismic conditions,these differences can lead to slopes exhibiting different dynamic response characteristics.In recent years,cross-river large-scale engineering construction has become more and more frequent,and the topography of the river valley has a significant amplification effect on the ground mo-tion,and the appearance of the wave impedance layer increases the sensitivity of the resonance effect of the side slopes,and the joint action of the medium wave impedance ratio n and the to-pography tends to cause a strong amplification effect on slope surface displacements.In order to deeply analyze the dynamic response law under the change in wave impedance of shallow surface-layer medium in combination with the change in slope gradient of river valley bank,a simplified linear model of river valley was established.In this model,the slope gradient of the left bank is set to be fixed at 45°,while that of the right bank is set to be three different choices,which are 30°,45° and 60°,and the bank slope medium is divided into two layers,namely the inner part composed of the micro-neophyte rock and the surface layer made up of weathered rock.On this basis,considering the change in wave impedance ratio of the shallow surface layer of the right bank slope,a total of six values of the wave impedance ratio are set as 1,1.5,2,2.5,3,3.5. The dynamic simulation was carried out using the discrete element software UDEC.The plastic principal model was adopted and the Mohr-Coulomb yield criterion was followed.The results show that: 1)The presence of a shallow wave impedance layer on the river valley slope significantly enhances the dynamic response of the slope,and this enhancement effect is especially obvious on the slope surface where the slope height reaches more than one quarter. 2)The slope gradient of the left bank is fixed at 45°,and of the right bank changes.This causes the left bank slope to exhibit different dynamic response characteristics during seismic activity.This difference lies in distinct force conditions and dynamic responses of the side slopes due to the change in slope gradient of the right bank.Furthermore,the asymmetry of the valley morphology has a significant effect on the ground dynamic response of the bank slopes.Compared with the symmetrical valley morphology this asymmetrical feature can enhance the ground dynamic response of the bank slopes,thus affecting the stability and safety of valley slope. 3)When the seismic wave vertically impinges on the valley slope model,the dynamic response of the right bank slope gradient in the heterogeneous model follows the order of 60°＞45°＞30°,which shows that with the increase of the right bank slope gradient,the dy-namic response of the slope becomes stronger.When the wave impedance ratio n is in the range of 1.5-2.0,the peak acceleration of the slope increases linearly with the wave impedance ratio n,and when n is greater than 2.0,it shows a nonlinear characteristic.The PGA amplification factor and wave impedance ratio n for the monitoring point 5'at the top of each model slope are fitted,and the goodness of fit is greater than 0.8,especially the fitting effect of 30° bank slope gradient is the best,suggesting that the relationship between wave impedance ratio n and PGA amplification factor can be effectively quantified under the specific background of considering influencing factors.At the same time,it is found that when the wave impedance ratio n is 2.5,3,3.5,the relationship between PGA amplification factor and slope gradient at the monitoring point 5' on the top of the slope is a linear increasing function. 4)When the seismic wave propagates into the slope model from the left side of the valley,the PGA amplification factor of the left bank is between 1.03 and 4.02,and that of the right bank is between 0.25 and 1.49,showing a significant backslope effect.The dynamic response of the toe of the left bank of the slope is inhibited by the shallow surface wave impedance n. 5)Stress differentiation exists on both sides of the medium interface,and the tensile stress is distributed on the shallow surface of the slope.The tensile stress makes the surface rock struc-ture tend to be unstable,and the PGA amplification factor of the slope surface is significantly larger than that of the slope's interior.It is not a single condition that leads to the difference in the dynamic response of the valley slope.The combined effects of topographic slope,wave im-pedance ratio of shallow surface layer and propagation direction of seismic wave result in differ-ent in dynamic response of valley slopes.

Abstract:Strong motion records have shown that spectral accelerations commonly exhibit nar-row peaks characterized by large amplitudes and small widths,which fail to adequately capture incurrent ground motion attenuation relationships(i.e.,strong motion prediction equations).Leveraging a comprehensive dataset of 190 215 horizontal strong motion records from Japan's KiK-net arrays,this paper systematically parameterized these deviations,identified their gov-eming factors,and developed an empirical model to enhance the accuracy of spectral accelera-tion predictions for seismic hazard analyses and structural engineering applications. The dataset included earthquakes with M≥3 and peak ground acceleration PGA≥5 cm/s2,encompassing various seismic scenarios featuring different magnitude,epicentral distances,and ground motion intensities.Acceleration response spectra(5％damping ratio)were com-puted for all records and normalized by PGA to isolate spectral shape characteristics.Residuals between observed spectra and predictions from a bilinear regression model(dependent on mag-nitude and epicentral distances)revealed distinct single-peaked deviations in 71.6％of records,with 24.9％showing multi-peaked deviations and only 3.5％lacking significant peaks.These deviations,characterized by their narrow bandwidth and elevated amplitudes,were parameter-ized using Gaussian curve fitting into three key features:peak height,peak width,and central period.The analyses demonstrated that such deviations occurred across diverse site conditions including bedrock stations,indicating contributions from both the spectral properties of bed-rock motions and site-specific amplification mechanisms. Statistical analyses were conducted to evaluate correlations between the three key features of single-peaked regular deviations and strong motion characteristics as well as site condition parameters.The results indicated that the peak heights of single-peaked regular deviations ex-hibited a strong positive correlation with the concentration of spectral energy within the devi-ation bandwidth,where higher energy ratios amplified peak amplitudes.Central period and peak width of single-peaked regular deviations showed significant positive relationships with the spectral peak period of borehole motions,earthquake magnitude,and epicentral distance.These relationships reflected the increasing dominance of long-period energy in large-magnitude distant-earthquakes.Site conditions further modulated deviations:softer soils(lower 30-meter average shear wave velocity vS30),thicker overburden(greater depth to bedrock with shear wave velocity vS≥1 km/s),and longer site predominant periods shifted deviations toward longer central periods and broader bandwidths.Weak-motion spectral ratios between surface and borehole motions also demonstrated moderate correlations with peak height,sup-porting the hypothesis that these deviations were derived from resonance between site layers and borehole motions. To address multicollinearity among strong motion characteristics or site condition paramet-ers,Lasso regression was employed to identify dominant factors.For peak height of single-peaked regular deviations,the key predictors included concentration of spectral energy within the deviation bandwidth,30-meter average shear wave velocity vS30,depth to bedrock with shear wave velocity vS≥1 km/s,and weak-motion spectral ratios between surface and bedrock motions.For central period of single-peaked regular deviations,significant predictors were the spectral peak period of borehole motions,vS30,depth to bedrock with vS≥1 km/s,and longer site predominant periods.For peak width of single-peaked regular deviations,several paramet-ers were retained.These include longer site predominant periods,spectral peak period of bed-rock motions,vS30,and depth to bedrock with vS≥ 1 km/s. A multivariate linear regression model was developed to predict the three deviation para-meters.Comparison results showed the model achieved highly consistencies between predicted and observed parameters,and the prediction model considering single-peaked regular devia-tions of spectral accelerations significantly improved the accuracy of spectral shape predictions.The Jaccard similarity coefficients,which quantified the overlap between predicted and ob-served spectra,indicated that both the median and lower-bound of spectral similarity improved to a certain degree.The average Jaccard similarity coefficients increased from 0.728 for the model without considering regular spectral deviations to 0.792 for the model incorporating such deviations,highlighting the model's ability to replicate narrow spectral peaks and improve then overall spectral fidelity. The study underscored the necessity of integrating single-peaked regular deviations into seismic hazard analysis.While the methodology relied on statistical analysis of strong-motion records and did not delve into the physical mechanisms deriving these deviations,the empirical model demonstrated significant improvements in spectral shape predictions.The proposed framework,however,depended on the selected ground motion attenuation relationship;its ap-plicability with alternative prediction equations requires further validation. By bridging the gap between statistical averaging and site-specific spectral features,the findings had potential implications for strong motion prediction,seismic zonation,site-specific hazard assessments,and the design of vibration-sensitive structures.The proposed model en-abled engineers to account for localized spectral amplifications that conventional attenuation models overlook,thereby enhancing the resilience of structures to earthquake-induced forces.Additionally,the methodology provided a framework for future studies to incorporate high-resolution site and ground motion data into probabilistic seismic hazard analyses,fostering a more nuanced understanding of ground motion variability across diverse geological settings.

Abstract:Local site conditions have a significant influence on the seismic ground motions,dir-ectly governing the spatial distribution of earthquake disasters.The site classification is funda-mental in determining seismic parameters for engineering seismic design.The determination of seismic ground motion parameters for different site conditions plays a crucial role in ensuring the seismic safety of engineering structures.The classification index of the site,the determina-tion of the index's limit value,and the site classification scheme are consistently hot topics in earthquake engineering.This paper aims to compare the effects of different site classification methods in China,the United States,and Japan on determining ground motion parameters,particularly on seismic design for engineering structures.By analyzing ground motion attenu-ation relationships for different site classes,it can be revealed how different site conditions af-fect spectral characteristics of ground motions.This provides more accurate ground motion parameters for seismic design,and offers theoretical support and practical guidance for earth-quake disaster risk assessment,engineering seismic design,and relevant policy formulation. Based on the borehole data of 646 observation stations and the acceleration time history data of 25 275 pairs of horizontal strong motions recorded by KiK-net,the site conditions of selected stations were classified according to 12 site classes divided by the site classification methods from China,the United States,and Japan.The strong motion records were then classi-fied and grouped based on the 12 site classes,resulting in 12 strong motion record datasets.Subsequently,the spectral acceleration(including peak ground acceleration)of each record was calculated.The spectral acceleration dataset for each site class was statistically regressed separ-ately,and then the corresponding spectral acceleration attenuation relationships were obtained for the 12 site classes.The difference in spectral accelerations for different site classes calcu-lated using those attenuation relationships is significant as it reflects the site effects on ground motions.A comparison was made regarding differences in the spectral characteristics of ground motions on different class sites.Finally,a comprehensive analysis and comment were presen-ted based on research results. This study conducted a regression analysis of ground motion attenuation relations and obtained the attenuation law of spectral accelerations for 12 class sites.The study also analyzed the influence characteristics of site condition on the spectral accelerations of ground motions.When comparing and analyzing the spectral characteristics of ground motions on different class sites,it was found that Chinese and American site classification methods can clearly distin-guish the influence of different site conditions on the spectral characteristics of ground mo-tions,while Japanese classification method struggles to do so.Specifically,there is almost no overlap between the normalized spectral acceleration curves obtained from Chinese and Americ-an classification methods,indicating significant differences in the effects of different site classes on ground motion.On the other hand,it is showed that Japanese classification method was no significant differences in calculated values of normalized spectral accelerations except for class Ⅰ sites,especially class Ⅲ and Ⅳ sites.Furthermore,it was revealed that overall stand-ard deviation of ground motion attenuation relation calculated by Chinese classification method was lower than that by American and Japanese classification methods.Moreover the clustering and consistency of within-site classifications by Chinese classification method were stronger compared with the other two approaches. After conducting an in-depth study of ground motion attenuation relationships and site clas-sification methods,the following conclusions have been drawn.The site classification methods used in China and the United States are able to effectively reflect the significant differences in spectral characteristics of ground motion among different class sites.They can capture better the influence of site conditions on spectral characteristics of ground motion.However,it is noted that the discriminability of spectral characteristics among different class sites in Japanese classi-fication method is not sufficient. In comparison with the classification methods employed in the United States and Japan,it was observed that the Chinese classification method demonstrates a higher level of clustering and consistency in terms of ground motion at different sites.On the other hand,it was found that Japanese classification method exhibits a lower discriminability and may even fail to ad-equately reflect variations in site effects across different class sites.These findings offer key theoretical support and practical advice for earthquake risk assessment,seismic design,and policy-making.

Abstract:On December 18,2023,a MS6.2 earthquake occurred in Jishishan county,Gansu Province,causing significant casualties and serious economic losses.This earthquake occurred in the northeastern margin of the Qinghai-Xizang Plateau,which is at the junction of the Loess Plateau and the Qinghai-Xizang Plateau.A series of studies have been conducted on this earth-quake,primarily focusing on damage assessment or focal mechanism solution.However,from a broader perspective of earthquake precursor research,it has been documented that thermal ra-diation anomalies have been detected prior to many moderate or strong earthquakes.And,the detection of pre-earthquake anomalies for the 2023 Jishishan earthquake is still scarce.Whether there were thermal radiation anomalies prior to Jishishan earthquake is a scientific question worthy of exploring. Currently,there are mainly two types of satellite remote sensing data sources utilized to detect pre-earthquake thermal radiation anomalies,one is the thermal infrared remote sensing data,the other is the passive microwave remote sensing data.However,due to the limited atmosphere penetration capability of thermal infrared band,data-continuity of satellite thermal infrared remote sensing is strongly affected by the variant meteorological conditions.In con-trast,passive micro wave remote sensing data offer unique advantages in terms of data continu-ity for the detection of pre-earthquake thermal radiation anomalies,thanks to its longer wave-lengths and stronger atmospheric penetration capability.To answer the abovementioned scientific question,the microwave brightness temperature(MBT)data covering(34°N-42°N,99°E-110°E)acquired by the Advanced Microwave Scanning Radiometer 2(AMSR2)sensor is collected.This dataset is analyzed using a novel wavelet-based two-step difference(WTSD)method to detect the thermal radiation anomalies preceding the Jishishan earthquake. In the WTSD method,the radiation received by microwave sensors is considered to com-prise three different components,i.e.,the low-frequency background field,high-frequency meteorological signals,and the seismic MBT anomalies.The low-frequency background field is considered as a stable contribution to the microwave radiation,which is mainly caused by topography,surface coverage and seasonal variations with strong regularity.The high-frequency meteorological signals are mainly caused by frequent variation of the meteorological conditions(e.g.,precipitation and temperature change,etc),which are considered as a ran-dom contribution.The seismic MBT anomalies can be extracted by sequentially subtracting the low-frequency background field and the high-frequency meteorological signals from the origin-al MBT data.Specifically,the hierarchical clustering is first conducted to classify the ground covers into different types.Then,MBT background fields for different types of ground covers are established and removed separately by adopting wavelet decomposition on the multi-year MBT data(from 2003 to 2022).After that,a second wavelet decomposition is performed to eliminate the high-frequency meteorological signals,thereby extracting the MBT anomalies as-sociated with the Jishishan earthquake.The temporal,spatial,and intensity evolution charac-teristics of the pre-earthquake MBT anomalies for the Jishishan earthquake are summarized as follows: 1)Temporally,the MBT anomalies lasted for approximately 133 days(from August 7 to December 18,2023)and experienced three different periods,i.e.,the initial active period,the subsequent calm period and the final pre-earthquake rebound period. 2)Spatially,a"//"-shaped distribution was found preceding the earthquake,consisting of MBT increase stripes interspersed with a MBT decrease region.During the active period,two MBT increase stripes appeared,and then gradually extended in the northeastern margin of the Qinghai-Xizang Plateau,forming a"//"-shaped anomalies distribution pattern.However,there was a MBT decrease region between the two MBT increase stripes. 3)In terms of intensity,the maximum intensity within both MBT increase stripes ex-ceeded 10 K,and the daily average intensity within them reached 3 K.It was worth noting that on 19 November(one month preceding the earthquake),both MBT increase stripes showed sig-nificant MBT anomalies decrease of approximately 7 K.Meanwhile,the MBT in the decrease region had been consistently at a low level,with a maximum intensity of less than 2 K.On October 10(approximately two months prior to the earthquake),the MBT anomalies in the de-creasing region also exhibited a significant decrease of around 12 K.Generally,the evolution trend of the MBT decrease region was similar to that of the two MBT increase areas,but in the MBT decrease region,the amplitude of the anomalies was smaller,the calm period was earlier and its decrease amplitude was larger. Interestingly,the spatial distribution of the MBT anomalies preceding Jishishan earth-quake is similar to the infrared anomalies observed during our rock-loading experiment.This indicates a large-scale development of microfractures within the crust preceding the earthquake.According to the remote sensing rock mechanics experiments and the meta-instability model,it can be concluded that the pre-earthquake MBT anomalies are mainly caused by variation of the crustal stress.The calm period shortly preceding the earthquake may imply that the crustal stress has entered the meta-instability stage,which can be regarded as a precursor to the Jishishan earthquake.

Abstract:Focal depth is an important parameter for the study of regional seismicity and seis-mic hazard.Accurate focal depths can provide valuable references for seismic hazard assess-ment and seismogenic mechanism research.However,it is a challenge to determine an accu-rate focal depth for earthquakes that occur in regions with sparse seismic networks.Traditional methods relying on seismic wave(e.g.the P and S phases)arrival times are severely limited by network density,resulting in low measurement accuracy.Nonetheless,utilizing information such as seismic wave amplitudes,spectra,and depth phases,even in sparse seismic networks,can facilitate the accurate determination of focal depths. In recent years,the sPL depth phase method has been widely utilized for determining the focal depths of local small and moderate earthquakes.The travel time of the sPL depth phase is primarily related to the focal depth and almost independent of the epicentral distance.There-fore,utilizing the sPL depth phase method not only avoids the compromise between the origin time and focal depth of earthquakes,but also effectively reduces measurement errors induced by velocity models.Moreover,the CAP(cut-and-paste)method is a full waveform inversion method with significant advantages in determining focal depths for moderate earthquakes. On 2 December,2023,at 01∶36∶33 Beijing time,an earthquake with MS5.0(local mag-nitude is ML5.3)occurred in Mangshi,Yunnan Province,followed by four ML≥3.5 after-shocks.Different institutions have reported significant disparities in the determined focal depth of the Mangshi MS5.0 mainshock.Hence,it is necessary to reassess the focal depth of the MS5.0 mainshock by using more regional seismic waveforms and different methods.Based on the broadband waveform data from the Yunnan Seismic Network and two regional velocity models,this study employed the CAP method to invert the focal mechanisms and focal depths ofthe mainshock(MS5.0)and four aftershocks(ML≥3.5)in the Mangshi earthquake se-quence.Additionally,we employed the sPL depth phase to further determine the focal depths.Our research results indicate that:the Mangshi MS5.0 mainshock is characterized by a strike-slip fault with a significant normal fault component.The optimal double-couple focal mechan-ism solution of the mainshock is as follows:strike 89°,dip 78°,rake-20° for fault plane Ⅰ;strike 183°,dip 70°,rake-167° for fault plane Ⅱ.The four ML≥3.5 aftershocks exhibit strike-slip with thrust or pure thrust mechanisms,the optimal double-couple focal mechanism solu-tions for these aftershocks all feature fault planes trending northeast(NE),and the strike,dip,and rake of the average plane for the four ML≥3.5 aftershocks are approximately 247°,65°,26°,respectively.This orientation of the average plane is consistent with the distribution of the double-difference relocated aftershocks and the orientation of the major axis of seismic intensity for the Mangshi earthquake sequence.Furthermore the focal mechanisms of both the main-shock and the four aftershocks of ML≥3.5 have complex orientations of the P and T axes,in-dicating the presence of a complex stress regime within the source region.It is plausible to spec-ulate that the rupture of the MS5.0 mainshock may have triggered nearby faults(with different fault planes)due to regional stress adjustments,resulting in significant differences in the focal mechanism types between the MS5.0 mainshock and the subsequent four ML≥3.5 aftershocks.Additionally,by using the CAP method,the optimal focal depth of the MS5.0 mainshock in Mangshi is determined to be 7 km,while the focal depths of the four ML≥3.5 aftershocks range from 5 to 7 km.On the other hand,the focal depth of the MS5.0 mainshock in Mangshi is estimated to be 7 km,and the focal depths of the four aftershocks are all approximately 5 km by utilizing the sPL depth phase method.The consistency between the focal depths determined by both two methods(with a difference of less than 2 km)indicates that the Mangshi earthquake sequence mainly occurred in the shallow part of the upper crust. Considering that the epicenter of the Mangshi earthquake is located within the Longjiang Reservoir area,we statistically analyze the relationship between the ML≥1.0 seismic events and water level changes in the Longjiang Reservoir area from 2010 to 2024 to depict the charac-teristics of seismic activities in the reservoir area after the reservoir impoundment.Our study results indicate that seismic events with ML≥3.0 in the Longjiang Reservoir area are closely related to reservoir water levels,except for one ML3.1 earthquake,which occurred during a period of low annual water level on June 8,2013,all other ML≥3.0 earthquakes in the reser-voir area occurred during periods of high annual water levels.Among them,the largest earth-quake,with a magnitude of ML4.2,occurred on September 24,2011,during a period of high water levels after the impoundment of the Longjiang Reservoir. The seismic activities in the Longjiang Reservoir area can be divided into four stages:the first stage is from January 2010 to February 2016,during which seismic activities of ML≥1.0 earthquakes were relatively calm(monthly frequency less than 10 times),with a calm period of about six years;the second stage is from March 2016 to December 2017,during which the wa-ter level of the reservoir changed drastically,and the seismic activities were relatively active,with monthly frequencies of ML≥1.0 seismic activity ranging from a dozen to several dozen times;the third stage is from January 2018 to September 2023,during which the seismicity with magnitude above ML1.0 was relatively calm(monthly frequency less than 10 times)over a period of almost six years;the fourth stage is from October 2023 to December 2023,during which the reservoir was at a high water level and its water level changed drastically followed the Mangshi MS5.0 earthquake on December 2,suggesting that there may be a certain correlation between the water level changes in the Longjiang Reservoir and the occurrence of this MS5.0 earthquake. Given that this earthquake occurred within the Longjiang Reservoir area,as well as factors such as high water levels,shallow hypocentral distribution,and conspicuous discrepancies in the focal mechanism solutions of the mainshock and aftershocks,this study tentatively hypo-thesizes that the infiltration of fluids into pre-existing fault fractures with potential for generat-ing moderate to strong earthquakes within the reservoir area may have facilitated the occurrence of this MS5.0 mainshock.Additionally,the rupture of the mainshock was likely to induce ad-justments of the local stress field,triggering slip along nearby NE-trending faults,and result-ing in the predominant NE-oriented distribution of aftershocks in this seismic source area.

Abstract:On June 17,2013,the Tianshui central seismic station submitted an earthquake pre-diction report based on observed anomalies in downhole georesistivity,groundwater radon con-centrations from Wushan Well No.22 and Wushan Spring No.1,and borehole tilt measure-ments from the Wushan vertical pendulum.On July 22 of the same year,an earthquake with MS6.6 struck the border region between Minxian and Zhangxian counties in Gansu Province.This event confirmed that the four sets of geophysical field observations from the Tianshui sta-tion had effectively captured pre-seismic changes within the regional subsurface medium.In this study,we analyze anomalies from these four datasets using the normalized rate method,subor-dinate function method,and tidal factor method,respectively.We systematically summarize the anomalous characteristics of each observation item and examine their correlation with the Minxian-Zhangxian earthquake.Finally,incorporating the focal mechanism solution,we in-vestigate the spatio-temporal characteristics of the geophysical anomalies based on a fault virtu-al dislocation model. The Tianshui georesistivity observation station is located in Yawan Village,Mapaoquan Town,Maiji District.The current downhole observation system,implemented as part of a post-disaster reconstruction project,commenced formal operation in January 2012,with elec-trodes installed at a depth of 100 m.From April 9 to May 9,2013,the hourly georesistivity values along the NS,EW,and N45°W directions at the Tianshui station exhibited synchronous high-frequency disturbances,with maximum variation amplitudes of 1.83％,1.00％,and 4.28％,respectively.The Lushan MS7.0 earthquake in Sichuan Province on April 20,2013 oc-curred during this anomalous interval.A second episode of synchronous high-frequency disturb-ances was recorded between June 12 and August 8,2013,with peak amplitudes of 1.41％,0.85％,and 0.94％,respectively,during which the Minxian-Zhangxian MS6.6 earthquake on July 22,2013,took place.Notably,the disturbance amplitudes were more pronounced prior to the Lushan event,which may be attributed to differences in deep geological structure and seis-mogenic environment between the two earthquake sequences.Furthermore,the daily variation pattern of georesistivity at Tianshui aligns broadly with the source region resistivity changes de-scribed by the DD(dilatancy-diffusion)mode,in which pore fluids play a critical role. Wushan Spring No.1,Well No.22,and the vertical pendulum borehole tiltmeter are all situated at the Wushan seismic station in Wenquan Town,Wushan County.The spring is natur-ally exposed,while the well is fully confined;the two are spaced approximately 100 meters apart.Their background radon concentrations are 240 Bq/L and 470 Bq/L,respectively.Radon variations in both the well and the spring were largely synchronous.Starting in April 2012,concentrations generally followed a"low-high-low"pattern,with the elevated phase lasting about one year.A short-term pre-seismic anomaly began on June 25,2013,exhibiting a"rise-earthquake-decline"sequence.The maximum pre-seismic variations reached 6.36％in Well No.22 and 9.59％in Spring No.1.Following the earthquake,radon levels decreased but did not fully return to pre-anomaly background values.In contrast,the NS and EW compon-ents of the borehole tiltmeter recorded opposing trends:the NS component decreased while the EW component increased.The tilt rate accelerated from 3.11 × 10-3/d during July 2012-Febru-ary 2013 to 5.66 × 10-3/d from February to June 2013.Post-seismic deformation was marked by a clear co-seismic strain step. Prior to the Minxian-Zhangxian MS6.6 earthquake,the normalized rate values for all three georesistivity components exhibited synchronous increasing anomalies,with a predefined threshold of 0.8.The monthly average radon concentration at Wushan Well No.22 increased from March to August 2012,reaching a maximum variation of 13.6％,while its 13-point mov-ing average rose from February to September 2012,with a peak variation of 8.7％.The subor-dinate function μ for this well exceeded the threshold three times before the earthquake,starting 14 months prior to the event.Similarly,the monthly average radon concentration at Wushan Spring No.1 began to increase from March 2012,with a maximum variation of 15.7％,and its 13-point moving average rose from February to November 2012.The μ value for the spring also surpassed the threshold three times,commencing 14 months before the earthquake.Following both the Lushan MS7.0 and Minxian-Zhangxian MS6.6 earthquakes,the μ values began to re-cover in August 2013.Significant changes in the tidal factor were also observed before the Minxian-Zhangxian earthquake.The tidal factor γ values for the NS and EW components showed a marked step-like decrease from April 2012 to January 2013,indicating a medium-to long-term trend change.The earthquake occurred approximately six months after the γ values began to rebound. The following conclusions can be drawn:① All four observation items are situated within a region characterized by compressive stress enhancement;② The geophysical field anomalies recorded at the Tianshui station were mainly distributed within 200 km of the epicenter.Their evolution and amplitude exhibited clear"long-term-intermediate-term-short-term-imminent-term"staging characteristics relative to the epicentral distance,demonstrating notable spatio-temporal correspondence with the Minxian-Zhangxian earthquake;③ The seismogenic process is reflected in the geophysical field through long-term background trends,intermediate-term anomalies,and short-term/imminent changes.After the occurrence of regional earthquakes,it is necessary to systematically analyze observation items with anomalous data,accurately identi-fy precursory signals based on their spatio-temporal distribution,intensity,and waveform mor-phology,and build a representative database of typical regional earthquake cases.

Abstract:Geological CO2 sequestration(GCS)is an important means for human society to alleviate greenhouse gas effects and achieve the so-called"dual carbon goals",namely the"carbon peeking"and"carbon neutrality".However,injecting CO2 into geologic formations on a large scale or at excessive rates may cause overpressure in the injected formation,which may in turn alter the shear strength of the intersected faults or preexisting fractures,posing en-vironmental risks to GCS engineering by triggering ground surface deformation and/or induced seismicity.The pore pressure buildup caused by fluid injection and production is an important cause for ground deformation and induced seismicity.Evaluation of the geomechanical effects caused by CO2 injection is an indispensable part of GCS risk assessment.Based on the inclu-sion theory and Green's function method,the existing analytical solutions for evaluating the geomechanics of a single fault has been expanded,to enables it to evaluate geomechanical ef-fects associated with large-scale CO2 injection into reservoirs penetrated by three faults.A mod-ular PYTHON-scripted utility tool composed of tens of functions was developed based on the analytical solution proposed in this study,which enables rapid assessment of the distribution of the fault slip patches and the maximum size of the fault slip patch.The maximum moment mag-nitude of induced earthquakes can be estimated for various injection scenarios and site settings.The major controlling factors of the seismic risks can be identified as well.Based on the analys-is of case studies,the following conclusions are drawn:① Injection causes horizontal displace-ment to concentrate on the burial depths of the fault that simultaneously contacts the reservoir and the surrounding rock(i.e.,the overlying and underlying formation units),while vertical displacement varies uniformly with depth.Due to lateral confinement,injection causes the reservoir to expand in both the caprock and baserock directions simultaneously.② Injection results in a negative horizontal strain increment and a positive vertical strain increment within the reservoir.The horizontal normal strain is significantly concentrated in the surrounding rock below the hanging wall of the fault and the overlying surrounding rock above the footing wall of the fault.While the vertical normal strain follows a similar pattern to the horizontal normal strain,it has opposite signs.③ The expansion of the reservoir causes the rocks inside the reser-voir to be subject to compressive normal stress.The horizontal normal stress only has a positive range in the surrounding rock below the reservoir on the hanging wall of the fault and the sur-rounding rock above the reservoir on the footing wall of the fault,while the vertical normal stress shows stress concentration in the surrounding rock area near the intersection of the fault and the reservoir,indicating that the surrounding rock in these areas is subject to tensile stress.CO2 injection leads to the concentration of shear stress increment near the four singularities of the fault.④ The risks of on-fault seismicity(i.e.,the maximum fault slip size and the maximal moment magnitude)are relevant to the distance of the fault to the injection center.The major controlling geomechanical parameters of the induced seismicity include the Biot coefficient,Poission's ratio,and the initial stresses and the initial pore pressure.

Abstract:Accurate estimation of input bedrock motion is essential for the seismic design of un-derground structures.However,most seismic observations are recorded at the top surface,making direct measurement of bedrock motion exceedingly challenging.Consequently,deriv-ing actual bedrock seismic motion holds substantial engineering significance.In recent years,the usage of inversion methods to calculate bedrock input motion has garnered significant atten-tion.Inversion involves evaluating bedrock input motion based on surface observations,primarily to address input excitation for site seismic response or soil-structure interaction ana-lyses.However,the inversion analysis of bedrock horizontal seismic motion primarily relies on surface horizontal seismic records in engineering.It is usually assumed that seismic waves are vertically incident,disregarding the incident angle of bedrock motion and treating it as a one-dimensional(1D)inversion problem.In reality,seismic waves at bedrock often have a certain angle of incidence,thus studying the inversion of bedrock motion under oblique incidence has important significance for engineering applications. It is worth noting that current inversion research is limited to the assumption of isotropic media.Nevertheless,natural soils generally exhibit obvious anisotropy due to processes such as weathering and sedimentation,and the horizontal modulus of natural soil is generally different from its vertical modulus.Transversely isotropic(hereinafter referred to as TI)media are com-monly employed to describe these anisotropic geomaterials.These models primarily account for differences in mechanical properties between vertical and horizontal directions.Therefore,it is more in line with the actual conditions to describe soil anisotropy by using TI medium mechan-ical model.Moreover,the existing studies on the layered TI half-space primarily focus on for-ward modeling of site surface responses,with limited attention to the inversion analysis of bed-rock seismic motion in the layered TI sites. Building on the authors' previous researches on bedrock motion inversion in the layered isotropic half-space,this study extends the approach to more realistic transversely isotropic(TI)site conditions,proposing a two-dimensional(2D)inversion method for bedrock motion in the layered TI half-space.The proposed method is based on the exact dynamic stiffness matrix of TI soil layers and the bedrock half-space.Utilizing seismic horizontal and vertical acceleration records of a certain point on the surface,an inversion objective function is formulated,an inversion objective function is established,and the control parameters in the inversion process are updated by employing optimization methods to achieve the inversion of bedrock motion and incident angle.Furthermore the method was validated using three different TI parameter site models as case studies. The research results demonstrated that bedrock motion and incident angle of the TI half-space can be accurately inverted based on the surface acceleration records and TI site character-istics.The Pearson correlation coefficient was selected to evaluate the overall accuracy of the two-dimensional inversion method for the layered TI half-space.An average Pearson correla-tion coefficient of 0.999 was obtained between the inverted bedrock motion and the actual val-ues for qP and qSV wave incidence.The relative error of peak acceleration was used to evalu-ate the local accuracy of the inversion results,yielding a total average relative error of 0.998％for peak ground acceleration between the inverted bedrock motions and the actual input waves under qP and qSV wave incidence.These findings quantitatively confirm the accuracy and precision of the proposed inversion method for calculating bedrock motion of the layered TI half-space. Moreover,the traditional isotropic(ISO)bedrock motion inversion model assumes the vertical material parameters of the site equivalent to the horizontal ones.In this study,the ISO model is applied to evaluate the error of the two-dimensional inversion method for the TI half-space.When seismic waves are incident at a small angle,the inverted acceleration time history and response spectrum of the bedrock motion for the layered TI half-space,used in traditional ISO inversion model,closely approximate the actual input waves.However,as the incident angle increases,errors in inverted bedrock motion for the layered TI site using the ISO model become increasingly pronounced.The inverted response spectrum of bedrock motion exhibits significant differences from the actual values in terms of curve shape,peak amplitude,and peak period,with the inversion accuracy decreasing significantly.This is because the traditional ISO model ignores the influence of the differences in the horizontal and vertical properties of the TI medium on the wave velocity,resulting in certain errors in the model and significant errors in the inversion calculation results.This indicates the necessity of developing a two-dimensional inversion method for the layered TI half-space.