With the advancement of seismic observation techniques,the field of seismology has accumulated vast amounts of observational data.The application of deep learning methods,which have developed rapidly in recent years,to seismic monitoring and prediction has demonstrated significant advantages.This paper systematically reviews the research progress of deep learning in seismic signal recognition,phase picking,microseismic location,and earthquake prediction, analyzes the strengths and limitations of existing methods,and explores future development directions.Through the analysis and integration of current information,it is demonstrated that deep learning holds great potential in earthquake catalog construction,precursor information extraction,and multi-source data fusion for prediction.However,challenges remain,including model generalizability,data heterogeneity,and the interpretability of physical mechanisms.

Based on the analysis of the earthquake swarm sequence, the seismic activity anomalies related to the two earthquake swarms are summarized, and the influence of the 5-magnitude earthquake swarm on the seismic trend in Xinjiang is analyzed. The results show that the earthquake sequence in Yuli-Kuche area in 2024 belongs to the earthquake swarm type, including two earthquake swarms with magnitude greater than 5.0. The whole sequence activity shows an increasing trend and is still in a state of continuous activity. There are six anomalies related to two earthquake swarms with M≥5.0, which are the frequency of earthquakes with M≥3.0 in the middle section of Tianshan Mountains, the high frequency of earthquakes with M≥3.0 in the middle section of Tianshan Mountains, the high frequency of earthquakes with M≥3.0 in Kuche-Shaya area, the 4.5 earthquake window near Baicheng, the significant activity of earthquake swarms near Baicheng, and the stress drop in Kuqa earthquake area.

To address the geological hazard issues in the quartz diorite mining area of Yutian County, Xinjiang, this study proposes a risk assessment method integrating unmanned aerial vehicle (UAV) oblique photography and multi-source data coupling analysis. Based on the SWDC-5 five-lens oblique photography system, a high-precision 3D realistic model was constructed. Multi-source data, including topographic parameters (elevation zoning, slope gradient, slope aspect differentiation), hydrological geomorphology (river lateral erosion, gully headward erosion), and human activity disturbances (road and house construction), were integrated to establish a "topographic parameters-hydrological geomorphology-human activity disturbances" multi-source driven evaluation model. The results show that: ① The intersection areas of SE-facing steep slopes (slope>35°) and NW-SE structural zones are high-risk areas for disasters, and the model-predicted core risk zones for geological hazards have fully covered historical disaster points; ② Five core risk zones were delineated, and a hierarchical prevention and control strategy was proposed, including anchoring engineering, ecological slope protection, and residential relocation, which can effectively reduce the disaster occurrence rate.

Taking the artificial blasting and natural earthquakes recorded at three seismic stations in Hami area as research objects, the main features of these two types of events are summarized by using the three elements of earthquakes, waveform recording, amplitude ratio and spectrum analysis. The results show that the occurrence time, epicenter location, magnitude, and S-P wave arrival time difference of the artificial blasting are relatively fixed, and there are basic features such as vertical P wave initial motion upward, large P wave amplitude, short-period Rayleigh surface wave development, and large period. The P_m/S_m value of artificial blasting is more than 0.75, and the correct identification rate is 94.47%; the P_m/S_m value of natural earthquake is less than 0.75, and the correct identification rate is 98%. The low-frequency component of artificial blasting is more, and the dominant component is mainly distributed in the range of 0-15 Hz, with fast attenuation; the high-frequency component of natural earthquakes is dominated by high-frequency components, with a wide range of dominant frequencies (mainly distributed in the range of 0-35 Hz), and the energy attenuation is slow.The S-transform shows that the main energy of the artificial blasting is concentrated in the low-frequency portion and is relatively single, whereas the natural earthquakes are rich in high-frequency energy with a more discrete distribution, which is “multi-peak” in nature.

Based on the analysis of the basic situation of Karamay, according to the seismic geological data, seismic zoning and active fault detection results, historical seismic monitoring data, the characteristics of seismic activity and seismic structure are analyzed and studied, the risk of seismic disaster in Karamay is evaluated, the weak links of earthquake response in Karamay are analyzed, and the preventive measures are put forward for the key hidden dangers in the region.

Taking Ganhezi fault soil gas observation station as an example, the morphology characteristics of gas components across the fault are measured, and the observation experiment analysis of influence of the thickness of the covering layer, the gas flow rate, the volume of the gas collection, the air temperature and the atmospheric pressure on the air radon is quantitatively out. The results show that the hanging wall of the reverse fault near the fault surface is an ideal location for fault gas observation, and the concentration of fault gas radon tends to be stable after the observation depth is more than 1.5 m. The gas flow rate should be controlled at 0.5 L·min^-1; the smaller the gas collection volume, the higher the gas flow rate, the higher the atmospheric radon concentration, the lower the influence of temperature and pressure on the data, and the more stable the data. Through quantitative analysis of the influencing factors of fault soil gas radon concentration, it provides a reference for the optimization and transformation of some observation stations in the future.

In response to the high incidence of faults in the power supply and communication systems at earthquake early warning stations, an integrated monitoring software has been designed. This software achieves information integration for intelligent power supplies from different manufacturers, rapid data analysis, swift fault diagnosis, advance warnings for power supply faults, and precise information dissemination. The software consists of four main components: information collection, information analysis, information display, and information transmission. Information collection is primarily used to gather power supply data and network communication status from early warning stations; information analysis is focused on extracting analysis of abnormal information; information display is used to visualize power supply data and alert notifications for abnormal information; and information transmission is aimed at sending abnormal information to mobile devices. This monitoring software effectively addresses the challenges faced by early warning duty personnel, significantly reducing the pressure of monitoring duties. It has been applied at the Urumqi Seismic Monitoring Center Station, where it has shortened the fault response time and improved the efficiency of fault handling by the early warning duty staff, demonstrating high practicality.

In order to find out the influence degree of different geological factors on seismic activity in Pamir area of the western section of southern Tianshan Mountains in Xinjiang, three influencing factors of plate movement, fault and rock properties were selected, and the contribution degree of different influencing factors to seismic activity was analyzed by grey correlation method. Taking the seismic activity of the three tectonic units as the reference sequence, and the plate movement, fault and rock properties as the comparison sequence, the grey correlation mathematical model is established. The resolution coefficients are set to 0.1,0.3 and 0.5 respectively, and the correlation coefficient and correlation degree between the reference sequence and the comparison sequence of different tectonic units are solved. The results show that the average correlation degree between plate movement, fault and rock properties and seismic activity is 0.682, 0.646 and 0.510, respectively. There is a high correlation between plate movement and seismic activity, which is the main geological factor controlling seismic activity in the region.

Based on the observation data of the geophysical observation of the Xiaguan central station in the past ten years since 2014, the classification and in-depth analysis are carried out to identify the main sources of interference of each measurement item, and the data variation characteristics and identification methods of various typical disturbances are summarized. This provides a detailed reference and reliable basis for the processing and application of geophysical observation data analytics in the jurisdiction.

The observation data of two different types of gravimeters in the same observation cave are compared and analyzed from the aspects of observation data quality, background noise level, recording earthquake ability and co-seismic response characteristics. The results show that : GS-15 gravity data has better availability and continuity, while gPhone gravity observation accuracy and background noise level are better ; the GS-15 gravity record has more advantages in the number of strong earthquakes, the duration of earthquakes, and the duration of co-seismic response, while the gPhone gravity record has earlier initial motion time and co-seismic response. The initial motion time and duration of the recorded earthquake and coseismic response are related to the filter constant and damping coefficient of the instrument, and the initial motion direction of the response pattern is related to the azimuth of the station and the epicenter.

To explore the coseismic response capacity of the water level in Well Chuan-41 to near-field and far-field earthquakes, this study selected near-field earthquakes with a surface wave magnitude M_S≥6.0 and global earthquakes with an M_S≥6.5 as research objects during the period of 2017-2023. It systematically analyzed the coseismic response characteristics of the water level in Well Chuan-41, and comprehensively evaluated the seismic response capacity of the well from three core dimensions: seismic energy density, relationship between coseismic response amplitude and earthquake magnitude as well as well-earthquake distance, and changes in tidal parameters.The results show that: Well Chuan-41 exhibits good coseismic response capacity to far-field earthquakes with M_S≥7.5, while its coseismic response capacity to near-field earthquakes with M_S<6.0 is relatively weak; the main type of coseismic response of the well's water level is oscillatory change. The specific laws are as follows:The energy density threshold required for far-field earthquakes to trigger the coseismic response of this well is approximately 10^-6 J·m^-3.The surface wave period of small and moderate-magnitude near-field earthquakes is significantly different from the natural vibration period of Well Chuan-41, resulting in the failure to effectively record its coseismic response. The dynamic stress changes caused by large-magnitude near-field earthquakes may lead to changes in the permeability of the aquifer associated with Well Chuan-41.

After analyzing the reliability of the tilt observation data of No.2 borehole in the back mountain of Karamay, it is confirmed that the tidal factor is basically stable. By eliminating the interference factors such as human influence, instrument failure and weather change, the credibility of anomaly identification before earthquakes is improved, and the typical interference and anomaly characteristics before earthquakes are classified and quantitatively analyzed. The determination and test of the optimal threshold by R value score show that the missing rate of inclination anomaly of No.2 borehole in Houshan, Karamay is low.

This paper focuses on the broadband mobile seismic array project constructed in the Southwestern Tianshan-West Kunlun region in 2023, detailing its system design and implementation. By increasing the station foundation depth and optimizing base treatment methods, the project effectively reduced environmental noise interference on observational data. Using three-component observational data recorded by the stations, the power spectral density (PSD) and root-mean-square (RMS) values of station noise were calculated, enabling a comprehensive evaluation of the noise levels at the seismic bases. Results indicate that the newly deployed stations exhibit favorable overall noise characteristics, with consistent noise levels across all three components. Additionally, the paper assesses the seismic monitoring capabilities of the array through ambient noise analysis. The findings demonstrate that the seismic monitoring threshold in the region improved from M_L2.0 to M_L0.5 after array deployment, significantly enhancing the area's earthquake detection capacity.

This study employs wavelet analysis, correlation analysis, and regression analysis to quantitatively investigate the impact features and mechanisms of temperature, atmospheric pressure, and rainfall on the tilt of KurkureH2 mountain broad-band. Further, a regression analysis is conducted to remove the influence of temperature on the NS component through regression, and the R-value rating method is used for earthquake capability testing, followed by a summary of their anomalous characteristics. The results indicate that:① atemperature and rainfall are the main influencing factors of annual variation. There is a significant linear negative correlation between temperature and NS component. When the phase of NS component (detrended) lags behind temperature for 17 days, the correlation coefficient between the two is up to -0.751;③ the broad-band tilt is influenced by rainfall from April to September each year, with the two components showing an “M” type of yearly change, during which the rainfall during this period suppressed the linear characteristics of the broad-band tilt influenced by temperature;④ the results of R-value test show that the best forecast window for the NS component is 350 days, with optimal R=0.64, R₀=0.71.

In order to study the application of ABAQUS in structural ground motion time history analysis, the method and key points of structural elastic-plastic time history analysis using this software are summarized by taking wave amplitude modulation, model establishment, material attribute assignment and grid division as steps. By establishing a RC frame model for example analysis, the damage of the model under earthquake action is studied. The results show that ABAQUS can clearly reflect the response of the structure in the whole process of earthquake. The output results can be used to analyze the modal shape of the structure, quantify the damage of the whole and the components, and show the spatial distribution of the damage factor. This method has a good application prospect in disaster loss pre-assessment, earthquake disaster prevention and emergency command.

Based on wavelet analysis methodology, this study investigates the multi-scale time-frequency characteristics of water level response to barometric pressure in the New 32 well, Bole, Xinjiang, using monthly averaged data (2015-2024) and hourly sampled data (2024). The results reveal that: (1) the dominant water-level oscillation period (256 days) exhibited concentrated energy from 2017 to 2019, while high-frequency signals manifested a bimodal pattern: intensive short-period disturbances (0.5～2 h) in summer and diurnal-to-weekly scale fluctuations (32～128 h) in winter; (2) The barometric pressure component in the long-period band (256～512 days) progressively intensified, with summer short-period power (0.5～2 h) reaching its maximum; (3) At the long-period scale (512 days), water level and pressure showed positive coherence during 2018—2021 that transitioned to negative coherence post 2022. High-frequency short-period variations (0.5～2 h) were completely synchronized in summer, while phase lags occurred at the diurnal scale (32 h) during autumn.This research provides time-frequency criteria for identifying fluid precursors in seismically hazardous regions.

This paper reviews and summarizes the anomalous phenomena preceding the M_S5.0 earthquake in Kuche, Xinjiang, on February 5, 2025. It identifies two seismological anomalies and three point geophysical observation anomalies before the earthquake. The seismological anomalies mainly involve an increase in 3rd-degree earthquakes in the central Tianshan region and an enhanced area for 3rd and 4th-degree earthquakes. The point deformation anomalies before the earthquake include background anomalies and short-term anomalies. The background anomaly is the horizontal swing in Korla, showing a significant trend reversal; the short-term anomalies are: (1) the Kuche borehole tilting EW component saw a significant acceleration of eastward tilt from December 12, 2024, with the anomaly ending on December 16, 2024; (2) the vertical swing NS component in Korla exhibited accelerated southward tilt rate anomaly on January 1, 2025. By comparing the anomaly characteristics before the M_S5.6 earthquake in Kuche on January 16, 2020, it is observed that the point deformation anomalies are mainly located at the Korla seismic station. Both earthquakes had background anomalies before the event, and the spatial evolution of mid-term and foreshock anomalies showed a migration from the near field to the periphery. Further comparison of the variation and evolutionary characteristics of the number of deformation anomalies in the middle section of the Tianshan Mountains during M_S≥5.0 earthquakes reveals that the number of anomalies in the medium and short term shows a clear clustering characteristic. The number of anomalies increases or decreases rapidly before the earthquake, which is similar to the anomaly pattern evolution process before the M_S5.0 earthquake in Kuche.

This paper primarily provides an overview of the construction content of the Yunnan sub - project of the National Earthquake Intensity Rapid Reporting and Early Warning Project. It introduces the construction of general stations, and when exploring the construction quality of general stations, conducts an analysis from multiple aspects such as equipment installation, station inspection, operation rate improvement, and background noise analysis. Through analysis, it is found that during the construction of general stations, there are issues such as non-standard installation and uneven quality of base station buildings, which negatively impact the construction quality. In terms of the operation rate, problems such as shortages of intensity meters and FSU standby equipment and spare parts, overly high operation rate standards, and a single background noise testing method are encountered. Meanwhile, by comparatively analyzing the construction achievements of general stations, basic stations, and benchmark stations in the early warning system, the advantages and disadvantages of the early warning system are further understood. Finally, by discussing aspects such as improving the communication mechanism, protecting the observation environment, increasing standby equipment and spare parts, optimizing the operation rate standard and background noise testing environment, and enhancing the magnitude estimation ability of the early warning system, this paper summarizes the solutions to problems related to the construction of general stations in the earthquake early warning system in Yunnan region, as well as methods to improve the overall effectiveness of the early warning system.

This study applies a four-dimensional disaster-body model to estimate earthquake disaster in Jiashi County, Xinjiang, by integrating factors such as seismic intensity and fortification capacity.The results show that after incorporating more comprehensive disaster information, the model's output agrees well with actual earthquake loss records.The research achieved refined hazard zoning through historical isoseismal lines and weight analysis; the use of relative values to calculate a disaster-body volume index improved the objectivity of the assessment. Validation in Jiashi County demonstrated that the predicted disaster magnitude matches the actual seismic damage observed in recent years. Although weight settings still partly rely on empirical parameters, the proposed method can provide a scientific basis for regional earthquake risk prevention and mitigation planning.

This study investigates the seismic instability mechanisms of colluvial slopes along National Highway G318 in southeastern Tibet using the particle flow code (PFC). A generalized geological model was constructed based on the geological conditions of the study area, incorporating seismic wave data from the M_W6.9 Nyingchi earthquake. The dynamic response characteristics, including unbalanced forces, velocity, displacement, and microstructural evolution, were systematically analyzed. Results demonstrate that during the initial seismic phase, unbalanced forces of surface particles transmitted through deep layers induced acceleration amplification, with peak surface particle velocities reaching 5 m/s. In the peak seismic phase, resonance between ground motion and particle natural frequencies led to the formation of accumulation mounds at the mid-slope and crest, where maximum displacement increments occurred. During the post-seismic phase, gravitational stress redistribution maintained mid-slope particle velocities at 0.5 m/s, while increased porosity weakened interparticle interlocking, resulting in continuous displacement growth at the crest and mid-slope. Accumulation mound particles migrated toward the slope toe, forming gentle colluvial deposits. The instability mechanism of colluvial slopes under seismic action manifests as downslope sliding triggered by contact force imbalance, accumulation mound formation dominated by resonance effects, and displacement progression driven by gravitational potential energy. These findings reveal the dynamic response mechanisms of colluvial slope instability under earthquakes, providing theoretical support for hazard prevention of similar slopes.

Using the RISP system, a seismic catalog for the Kashi region from January 12, 2020, 00:00 to February 12, 2020, 23:59 was rapidly constructed. Through matching filtering and double-difference relocation methods, the following conclusions were drawn:① The RISP automatic cataloging system demonstrates significant advantages in microseismic detection and catalog production, particularly in faster response times, lower detection thresholds, and positioning results that are comparable to manual methods.② The automatic cataloging results better reveal the distribution of aftershocks from the Kashi M_S6.4 earthquake sequence in terms of depth, aligning with the overall northward dip of the Keping Fault. This suggests that the Kepingtage Fault is the main fault responsible for the east-west directed aftershocks, with activity concentrated near the detachment surface of the Kepingtage thrust structure. It may also have triggered synchronous movement along the Ozgertawu Fault.③ The automatic cataloging identified a significant number of multi-detected small earthquakes; however, verifying the authenticity of these small earthquakes is crucial. From the double-difference relocation results, the positioning of many multi-detected small earthquakes is still not ideal in revealing fault structures. Thus, developing automated verification processes for small earthquakes and improving the accuracy and reliability of automatic seismic catalogs will be key focuses for future research.

Selected observational data from 2014 to 2024 on the two components of vertical swing tilt, cave temperature, and external atmospheric temperature of Kuerle Hohla Mountain were used to study the influence characteristics of temperature on vertical swing tilt and to conduct seismic event analysis on the vertical swing NS component. The results show that:①the annual variation of vertical swing tilt is mainly influenced by the annual changes in air temperature and cave temperature. The effect of air temperature exhibits a more significant linear characteristic with a lag effect, while the effect of cave temperature is real-time, showing a higher correlation with air temperature and a nonlinear relationship.②The seismic event analysis indicates that the NS component has a good correspondence with seismic events. Its response to changes in air temperature and cave temperature is primarily characterized by abnormal annual variation amplitudes and weakened or deviated correlations.

AlphaGUARD radon monitor introduced by the seismic system mainly includes two models: P2000 and PQ2000 Pro. Because the measurement data of this kind of instrument is not connected to the seismic observation network, most of the equipment does not perform regular calibration. If it is directly used for observation, it may lead to significant deviation of the data. Based on the calibration/verification results of the above two types of radon monitors, this paper analyzes the measurement performance and puts forward corresponding work suggestions, in order to provide reference for the use and maintenance of related instruments in the seismic industry. The results show that: 1) the instrument calibration coefficient (K value) fluctuates between 0.91 and 1.14 for a long time, and the annual deviation is less than 5%; 2) The repeatability of the instrument is less than 5% when the measurement value is calculated at 1 h, and the repeatability of the instrument is less than 10 % when the measurement value is calculated at 10 min, and the background accumulation will affect the repeatability of the low radon concentration measurement. 3) In the radon concentration range of (0.4～12) kBq·m^-3, the relative inherent error of the instrument is less than 4%; 4) The cumulative deviation of K value of P2000 model was close to 5% in the third year, while that of PQ2000 Pro model was close to 5% in the fourth year. With the increase of service life, the K value of the instrument showed a significant downward trend. It is suggested that the calibration/verification cycle should be determined according to the specific use of this type of radon measuring instrument. If used as a standard instrument, regular fixed-point traceability should be implemented, and the longest traceability cycle should not exceed 3 years.

This study evaluates the capability of the automatic cataloging system to generate aftershock sequences of strong earthquakes in the Xinjiang region, using the 2016 Hutubi M_S6.2 earthquake in the middle section of the Tianshan Mountains as a case study, with a comparative analysis of manual and automatic cataloging results. The research demonstrates that: (1) Timeliness and detection capability: The automatic cataloging system can rapidly generate aftershock sequence catalogs after an earthquake, with significantly higher efficiency than manual cataloging. The number of detected earthquakes is approximately twice that of manual cataloging, substantially enhancing catalog completeness. (2) Result consistency: The matching rate between the automatic catalog and the manual catalog is 89.5%. Epicenter deviations are mainly distributed within the range of 0~20 km, the lower limit of magnitude processing reaches M_L0.0, magnitude differences are concentrated within ±0.3, and origin time deviations are largely confined to ±2 seconds. The results confirm that the automatic cataloging system is reliable and highly efficient, providing timely and comprehensive data support for scientific research and practical applications such as seismic trend analysis, emergency response, and rupture process inversion.

Based on the observational data of the Resistivity Monitoring Station in Shayidong, Korla, Xinjiang, this paper introduces the discovery process and verification of abnormal resistivity data, excluding environmental and meteorological factors,analyzing the reliability of the abnormal change and discussing anomaly characteristics of ground resistivity before earthquake. The results indicate that the observational data of Shayidong ground resistivity are continuously stable with a high degree of abnormal credibility. On May 2023, the data curve showed a trend reversal change, and on April 2024, a yearly variation distortion anomaly appeared again. After the anomaly ended, earthquakes with magnitudes of 5.5 or above occurred within a 300 km radius of the monitoring station. It suggests that the Shayidong ground resistivity monitor can serve as a reliable basis for determining anomalies before earthquake.

Since the late Quaternary, there is still some controversy about the change of the incised rate of the river system in the mainland and its driving factors. In this study, the Yihehusitai River terrace in the southern foothills of the Alataoshan in the northern part of Bole was taken as the research object. Through field geological and geomorphological surveys and high-precision geomorphological measurements by drones, six terraces were identified in the valley. The age sequence of the terraces at all levels was obtained by using the method of OSL dating. The results show that since the late Quaternary, the incised rate of the Yihehusitai River has fluctuated significantly. During the late Pleistocene, the incised rate was 0.64~0.87 mm/a, which was significantly greater than the vertical uplift rate of 0.43 mm/a of the fault, indicating that long-term tectonic movement and climate change played an important role in river incising. The undercutting rate of the T3 terrace is equal to the vertical sliding rate of the fault, indicating that the river undercutting during the T3 terrace is controlled by tectonic movement. Since the late Holocene, the river incision rate has increased to 3~4 mm/a. The rapid increase of river incision rate since the late Holocene is due to the rapid fluctuation and oscillation of climate.

Employing electromagnetic anomaly extraction methods, including DC apparent resistivity, geomagnetic daily ratio, and geomagnetic low-point displacement, this study systematically analyzes the multi-method electromagnetic anomalies observed near the epicenter prior to Kuche M_S5.0 earthquake on February 5, 2025. The results indicate that an annual variation distortion anomaly in DC apparent resistivity was recorded at the Korla station, located approximately 194 km from the epicenter; A geomagnetic daily ratio anomaly emerged 85 days before the earthquake, with the mainshock occurring precisely at the sharp intersection of the daily ratio threshold line within the predicted area; A geomagnetic low-point displacement anomaly, predominantly distributed in a near-NNW direction, appeared across the Xinjiang region 28 days prior to the event. The earthquake was situated near the mutation boundary of this low-point displacement, and the dominant time window for earthquake occurrence was 27±4 days; Spatially, the epicenter is located within the concentrated anomaly prediction zones identified by the Korla DC apparent resistivity, geomagnetic daily ratio, and geomagnetic low-point displacement methods. Temporally, the anomalies exhibit a clear evolutionary progression from medium-term to short-term and imminent-term precursors. These findings provide valuable references for future seismic monitoring and tracking in this region.

To enable earthquake analysis and forecasting personnel to extract anomalous information from observational data quickly and accurately, this study, based on the R language platform, employs methods such as Fast Fourier Transform (FFT), Pertesv’s filtering, and factor decomposition in time series analysis. Specific measures include writing program code to analyze various frequency bands in geophysical observational data, extracting diurnal waves, semi-diurnal waves, seasonal factors, trend variations, and residual random components, and outputting the processed data. The results show that the program can effectively separate components of different frequency bands and identify typical periodic signals and anomalous features. Currently, deformation anomaly analysis in earthquake consultations mainly relies on plotting graphs using EIS2000 and manual interpretation, which suffers from low automation. This study provides data analysts with a convenient and scientific processing tool, thereby improving the efficiency and reliability of earthquake precursor information extraction.

GNSS coordinate time series is the key data for crustal deformation monitoring and seismic research, but its accuracy is seriously disturbed by complex noises such as multipath and atmospheric delay. Noise reduction processing has become a necessary prerequisite for high-precision applications. In recent years, noise reduction methods have developed from traditional empirical mode decomposition (EMD) and wavelet transform to variational mode decomposition (VMD), local mean decomposition (LMD), empirical wavelet transform (EWT) and other adaptive methods, and combined with particle swarm optimization, singular spectrum analysis (SSA), independent component analysis (ICA) and other algorithms to form a variety of hybrid strategies. The principles, advantages and limitations of the above methods are systematically sorted out. Through three typical application cases of post-earthquake deformation monitoring, multipath correction and long-term trend extraction, the performance differences of different methods in actual scenarios are compared and analyzed, and then a comprehensive selection strategy based on accuracy, noise type and efficiency is proposed. Studies have shown that existing methods still face challenges such as different evaluation criteria, insufficient real-time processing capabilities, and poor adaptability to special geographical environments. In the future, we should focus on the development of end-to-end noise reduction driven by deep learning, joint filtering under the constraint of multi-source data fusion, and intelligent automatic processing platform to further improve the availability and reliability of GNSS time series.

In order to study the motion characteristics of the Zhangbo Belt and the constraints affecting its motion, the GNSS data of North China and its adjacent areas are collected and collated. Under the unified space-time datum, the method of calculating the conversion parameters by common points is used to realize the unification of the data reference datum, so that the velocity fields of different datums are converted to the same reference frame, and the GNSS velocity field calculated by the stable Eurasian reference frame is used uniformly. In this study, the velocity field excluding the influence of the 3.11 earthquake in Japan is used to further calculate the strain field in North China and determine the basic characteristics of the strain field in North China. The best block model of North China and its adjacent areas is established, and the theoretical velocity field of North China block and its boundary is analyzed. Combined with the velocity field without ‘rigid’ conversion and the strain field in North China, the comprehensive analysis shows that the theoretical velocity field of Zhangbo zone is 1.5 mm·a^-1.In the eastern section of Zhangbo zone, the velocity field in Tangshan area is quite different from that on both sides, and the strain field gradually decreases from Tangshan to both sides, and the main strain coincides with the NE-trending fault, indicating that the difference of velocity field in the eastern section of Zhangbo zone may be related to the NE-trending fault activity in this area.

As a special product of geological tectonic activity, the eruption process of mud volcano not only reveals the migration characteristics of underground fluid, but also provides important clues for regional tectonic stress changes and seismic activity. In this study, the mud volcanoes in Aiqigou and Baiyanggou were taken as the research objects, and the earthquakes with M_S≥5.0 in the range of 300 km were selected by the combination of macro and micro methods. The correlation between mud volcano eruption activity, gas geochemical characteristics and seismic activity was systematically analyzed, and the R value test was carried out. The results show that when the mud volcano eruption activity is enhanced and the radon content is synchronized with the high value, the corresponding rate of subsequent earthquakes is 55.6%. When the monthly flow observation of CO₂ and CH₄ gas content in the mud volcano showed a continuous upward trend, the R value test was passed. These indicators have certain indicative significance for earthquake prediction and can provide scientific basis for regional earthquake prediction.

In order to solve the problems that the traditional seismic network equipment can not be unified, the structure is complicated, the troubleshooting efficiency is low, and the operation and maintenance is difficult, this study is based on the equipment purchased by the catastrophe prevention project. By deploying SDN and DeepSeek locally, the practical exploration of the integration of SDN network technology and DeepSeek large model is carried out. The research divides the seismic network into control level and forwarding level, builds a centralized management platform, and breaks the bottleneck of traditional seismic network operation and maintenance management. Combined with RAGFlow and Dify, a set of intelligent operation and maintenance management system is formed. The daily warning information is sliced, and the operation and maintenance optimization goal of automatic classification of warning information and intelligent generation of operation and maintenance disposal scheme is realized. It can effectively shorten the time of fault disposal, greatly improve the overall operation and maintenance efficiency and quality of seismic network, and provide practical samples for intelligent operation and maintenance of seismic network.

Taking Tianquan and Mingmuquan springs in Wenquan County, Bozhou, Xinjiang as the research objects, based on the hydrochemical dynamic monitoring data from March 2024 to April 2025, this study comprehensively utilizes methods such as Piper three-line diagram, Gibbs model, and isotopic analysis to explore the hydrochemical characteristics of tectonic fissure springs and their earthquake precursor indicative significance. The results indicate that the hydrochemical types of the two springs are Na-SO₄·HCO₃; the water-rock interaction is in dynamic equilibrium; the hydrochemical components are mainly controlled by the leaching of silicate minerals; and the hydrogen and oxygen isotopic characteristics reveal the mixed recharge process of atmospheric precipitation and deep tectonic water. The study captured the synchronous increase in Na⁺ and Cl^- concentrations before Chabuchaer M_S4.2 earthquake in January 2025, confirming the sensitivity of the hydrochemical components in this area to tectonic stress disturbances. This study provides important hydrogeochemical empirical evidence for identifying earthquake precursors using groundwater chemistry methods in tectonically active areas.

Taking the network architectures of Bole Earthquake Monitoring Central Station, Xinjiang Uygur Autonomous Region Earthquake Agency before and after the implementation of the Catastrophe Prevention Project as the research object, this paper systematically analyzes the differences before and after the network structure optimization. The original network adopted a bus topology, which carried the risk of single point of failure. Meanwhile, it was plagued by outdated hardware, insufficient security protection, inadequate operation and maintenance audit, failing to meet the requirements for high-speed and stable transmission of earthquake monitoring data. Through a series of optimization measures including equipment upgrading, IRF stacking of core switches, series deployment of firewalls, and introduction of bastion hosts for operation audit, the network achieves high reliability, high security and intelligent operation and maintenance. After the network optimization and reconstruction, data transmission efficiency, management efficiency and system stability have all been greatly improved, which can provide solid technical support for the development of earthquake monitoring work.

In order to solve the problem of insufficient reliability of terminal estimated intensity, based on the point source attenuation model, a comprehensive improvement method of introducing regional correction coefficient k is proposed. The accuracy of estimation is improved by source distance calculation, intensity attenuation model correction and regional parameter optimization. The response data of the special terminal for earthquake early warning in Xinjiang since December 2023 were collected, and the instrument intensity was compared and analyzed. The least square method was used to fit and optimize, and the regional correction coefficient k=0.95 was calculated. The verification results show that the average absolute error of the estimated intensity after correction is reduced from 1.6 degrees to 1.0 degrees, and the root mean square error is reduced from 1.5 degrees to 1.0 degrees, which significantly improves the accuracy and reliability of the estimated intensity of the special terminal for earthquake early warning in Xinjiang. This study provides a technical reference for the optimization of regional earthquake early warning parameters.

Dynamic anomalous changes in subsurface fluid composition are one of the important indicators for identifying seismic precursor signals, and their temporal evolution characteristics can provide observational evidence for capturing short-term and imminent earthquake precursors. Traditional laboratory-based detection methods are inefficient and difficult to meet the requirements of rapid post-earthquake response. In this study, a 30-day comparative test was conducted using the Agilent 990 portable gas chromatograph at Spring 04 and Spring 10 in Urumqi, Xinjiang. The results show that the detection results of the instrument for major components such as CH₄ and CO₂ are in good agreement with those obtained by the benchtop gas chromatograph, with the mean relative error controlled within 5.5%, and the single analysis time shortened to within 10 min. Although limitations remain in trace hydrogen detection, the instrument shows good applicability in major gas component detection, analytical efficiency, and field application, providing experimental evidence for extending seismic emergency fluid monitoring from “sample-and-send laboratory analysis” to “on-site rapid analysis”.

After years of development, GNSS observation technology has been widely recognized and applied in the fields of crustal deformation, geodesy, earthquake prediction and analysis. Based on the GNSS continuous observation station data of the Xinjiang Uygur Autonomous Region Seismological Bureau, this study comprehensively evaluates the data efficiency, signal-to-noise ratio, multipath effect, cycle slip ratio and other indicators, analyzes the factors that may affect the data quality, so as to improve the construction plan, and provide technical support for the data quality improvement and daily operation and maintenance of GNSS observation stations in the future. The results show that the data quality of Xinjiang GNSS continuous observation station is good, and the observation results meet the requirements of the discipline. The quality of single Beidou (BDS) GNSS data is generally better than that of co-site observation (GPS+BDS).

Based on the integrated georesistivity and geomagnetic observation system at the Wenquan New Station in Xinjiang, this study systematically analyzes the interference characteristics of direct-current georesistivity observations on geomagnetic records by designing multi-current gradient experiments. The results indicate that the interference manifests as square-wave pulses in the time domain, synchronized with the power supply cycle, while its frequency-domain energy is concentrated at the fundamental frequency (0.04 Hz) and its harmonics. The interference amplitude (ΔB) exhibits a strict linear positive correlation with the supply current (I) within the range of 0.3～2.5 A (R²>0.999). Furthermore, the interference field demonstrates significant vector characteristics, primarily affecting the vertical component (Z) and total field intensity (F), while having a negligible impact on the horizontal component (H) and declination (D). The linear relationship between interference amplitude and current can be precisely quantified using the current-magnetic field conversion coefficient k, which holds potential for future instrument status monitoring and data consistency correction.