Supercritical geothermal can extract more than ten times the energy of conventional Enhanced Geothermal System (EGS), and become the development direction of new energy. Although China has become the country with the largest direct utilization of medium and low temperature geothermal resources, the level of exploration and development of deep underground high temperature geothermal resources needs to be improved. In this paper, we analyze the research progress of high temperature and high pressure physics experiment, numerical simulation, geophysical exploration and monitoring methods for supercritical geothermal, and the rock-fluid-gas geophysical properties of three-phase medium are summarized and analysis. And we give the typical high temperature geothermal area in China for supercritical geothermal resource exploration potential evaluation preliminary discussions, The potential exploration areas of deep supercritical geothermal resources based on geophysical survey results are predicted to provide support for the commercial utilization of supercritical geothermal resources in China.

A cast of east-west trending continental slivers, which were composed of the Cimmerian Supercontinent, including the Lhasa continental block, started to rift apart from the northern margin of Gondwana in the late Paleozoic. According to this comprehensive tectonic scenario, several points of view have been proposed on when and from which continent the Lhasa block rifted away from the northern margin of Gondwana over the past decades, however, the issue still remains controversial. To this end, in this study, we aim to provide quantitative constraints on this topic by employing isotope-based geochronological dating and paleomagnetic techniques. Through carrying out an integrated study including zircon U-Pb geochronology, paleomagnetism, petrography, and rock magnetism on the early Middle Permian Luobadui Formation basalts and andesitic basalts developed in and around the northwest part of the Linzhou County of south Xizang (i.e., Tibet), it reveals that the Luobadui Formation volcanics was formed at ca. 272 Ma (i.e., Guadalupian stage of Middle Permian). Based on rock magnetic results carried out by previous scholars and petrographic investigations performed in this study, it shows that the Luobadui Formation volcanic samples experienced low-grade metamorphism along margins of some rock-forming minerals like plagioclase, quartz. Typical accessory minerals like magnetite particles (is also a kind of iron-bearing oxides) distributed in Luobadui Formation volcanic rock samples was also impacted by this low-grade metamorphism. Rock magnetic data sets also suggest that the dominant remanence carriers in the Luobadui Formation volcanic rocks are pseudo-single domain (PSD) magnetite grains in this study. Statistical analysis on filtered characteristic remanent magnetizations from 21 paleomagnetic specimens of the Luobadui Formation volcanic rocks reveals that the Fisherian mean direction and associated parameters over the sampling unit of the Luobadui Formation volcanic rocks are Ds±ΔD=124.2°±8.8°, Is±ΔI=34.5°±7.2°, k_s=20.3, α_{95 s}=7.2° after bedding correction. Thus, based on already obtained statistical mean direction over the selected 21 samples of the Middle Permian Luobadui Formation volcanics, the paleomagnetic pole can be computed as λ_p=16.8°N, φ_p=325.5°E, K=20.7, A₉₅=7.2°. The paleosecular variation has been adequately averaged out according to paleosecular variation (PSV) evaluation criteria in paleomagnetism for the sampled Luobadui Formation volcanics. Therefore, in combination of positive fold test results for the paleomagnetic data sets with rock magnetic and petrographic results, we believe that our obtained paleomagnetic results are primary origin. It can be calculated (reference site: 30.01°N, 90.99°E) that the Lhasa block was at 19.7°±7.2°S in southern hemisphere at ca. 272 Ma. In other words, the Lhasa block was still in middle-low latitude of southern hemisphere then. Combined with previously published results from the Lhasa block, we intend to think that the Lhasa continental block along with the Cimmerian supercontinent likely rifted away from northern periphery of Gondwana and started its long-lasting northward travel journey in early Permian or even much earlier period. Combined with previously published paleomagnetic and geological results, it shows that the total northward convergence is estimated as 1776±700 km of the Lhasa block travelling from low-middle latitude band of 19.7°±7.2°S at ca. 272 Ma to the equatorial area of 3.7°±3.4°S at ca. 180 Ma. This further indicates that the average motion rate of the Lhasa block was around 1.9 cm/a during the time period of 272~180 Ma. In other words, the Lhasa block likely rifted away from northern margin of Gondwana in Early Permian and immediately started its northward drift and motion journey.

As an important component of Transient Electromagnetic (TEM) prospecting, the performance of a TEM transmitter almost determines the effectiveness of this prospecting method. In order to accurately prospect shallow and even very shallow layers using multi-turn small coils, this paper designs an improved TEM transmitter based on SiC MOSFET. SiC MOSFET is a third-generation semiconductor that offers significant improvements in voltage withstanding, current withstanding, heat dissipation, and response speed compared to widely used Si IGBT. The shut-off time, as a critical parameter of the transmitter, essentially determines the degree of coupling between the primary and secondary field signals. To prevent losing information in even very shallow and relatively shallow layers, the transmitter designed in this paper supports switching transmission between large and small currents. When prospecting in even very shallow or relatively shallow layers, a small current of 1.1 A can be selected, with a shut-off time of only 4 μs. When prospecting in shallow layers, a large current of 16.2 A can be selected, with a shut-off time of 35 μs. To provide transmitting current data use for post-processing, this paper designs a current acquisition system that supports dynamic sampling rate to collect the current waveform in the whole time. When the current waveform is in the rising or falling edge area, the acquisition system automatically selects a high sampling rate of 1.8 MSPS for sampling. When the current waveform is in the steady-state area, the acquisition system automatically selects the lowest sampling rate of 50 KSPS for sampling. Tests show that by sampling with dynamic sampling rate, the acquisition accuracy and data volume can be effectively balanced, thereby ensuring the stability of the storage system. In addition, the transmitter board has a small size of only 255 mm×192 mm, and supports 12 V battery power supply, so it has good portability and can improve field prospecting efficiency to a certain extent.

The spatial and temporal distribution of the palaeomagnetic data of the South China Block is uneven, which limits our understanding of the paleogeography position of the South China Block in the Early Paleozoic. To search for suitable strata for systematic study of paleomagnetism, this study carries out a combined study including petrography, rock magnetism, and magnetic fabric on the mudstone-siltstone of the Cambrian strata from the Zhenba area in the northwestern edge of the South China. The results show that the magnetic minerals of the strata are mainly composed of titanium-bearing hematite or magnetite and magnetic pyrite, while the magnetic fabric is characterized by depositional-weakly deformed magnetic fabric. It is concluded that most of the Cambrian strata in this area have not been tensely modified by tectonic activities, and it is expected to retain the primary magnetic component acquired during deposits, which makes it possible to obtain reliable paleomagnetic data.

Faults, as one of the main geological structures, are crucial for analyzing subsurface structures and determining oil and gas enrichment areas. Traditional methods face challenges in the efficiency of seismic data feature extraction and the accuracy of fault identification. This article first outlines the background of seismic fault identification and the limitations of traditional methods, and then explores the application of deep learning methods in this field. In deep learning, fault identification is regarded as an image processing task, usually trained in a supervised learning manner. Data, model, and loss function are the three core elements of supervised deep learning. Data is the foundation for training deep learning models, and the quality, diversity, and representativeness of data are crucial for the training and generalization ability of the model; the model can establish a nonlinear expression of the relationship between input and output, used to learn patterns and rules in the data; the loss function is used to quantify the difference between the model's predictions and the true labels, and a good loss function can guide the model towards more accurate optimization continuously. This paper first introduces the training dataset and methods of data fusion and difference optimization, then discusses the effectiveness of different deep learning models such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Transformer models in seismic fault identification, and finally analyzes the impact of different loss functions. This paper summarizes the current performance, advantages, and challenges of deep learning methods in seismic fault identification and provides an outlook on possible future research directions.

Well logging identification and quantitative characterization of source rocks are of great significance to hydrocarbon reserves evaluation. There are three types of source rocks in the Jurassic Yangxia Formation of Kuqa Depression: coal measure, dark mudstone and carboniferous mudstone. However, due to the depth of burial (average greater than 4500 m), the source rocks are too mature and affected by deep buried ground stress, the traditional ΔlogR method is difficult to apply. Based on this comprehensive use of geochemical analysis and geophysical logging data, this paper first reveals the geological characteristics of the source rocks of the Jurassic Yangxia Formation, and realizes the qualitative identification of the source rocks of different lithologies through conventional logging crossplot. The results show that the overall maturity of source rocks of Yangxia Formation is high and the quality of source rocks is medium to good. The qualitative identification of source rocks can be realized by the intersection of curves such as GR, AC, DEN, CNC and AC, GR, RT, CNC, DEN curves that are sensitive to the response of the source rock are selected to establish a quantitative prediction model of TOC content by using multi-regression analysis methods and BP artificial intelligence method, and the quantitative evaluation of the single well source rock in the study area is realized. The results are in good agreement with the measured core data. The research results are of guiding significance for comprehensive logging evaluation of deep source rock quality in Kuqa Depression.

The Transient Electromagnetic Method (TEM) is a geophysical exploration technique with significant potential for widespread application. Its good adaptability to various terrains and capability for non-invasive detection have made it a mainstream technology in urban underground space surveys in recent years. To enable effective TEM exploration in spatially constrained areas such as urban underground spaces and tunnels, this paper addresses the issues associated with small loop devices, such as high mutual inductance and complex structural design, by designing a transceiver integrated small coil and a matching buffer circuit. Firstly, this paper models the transceiver integrated coil and analyzes it from three aspects: equivalent resistance, inductance, and capacitance. Solutions to these issues are then proposed. To address the resonance problem in the coil, a buffer circuit is designed. Finally, the performance of the coil and its matching buffer circuit is verified using both high and low transmission currents. Testing indicates that even with a transmission magnetic moment of 28.75 Am², the coil's effective resistance, inductance, and capacitance remain very low, measuring 494 mΩ, 1.03 mH, and 260 pF, respectively. Moreover, the coil demonstrates impressive turn-off times, reaching 38 μs and 5 μs for transmission currents of 14 A and 1.2 A, respectively. When the coil is used in conjunction with the buffer module, the system effectively reduces the interference from the primary field response, thereby enhancing the secondary field information. Additionally, since the designed coil device integrates transmission and reception functions and is small in size, it is highly suitable for exploration in spatially constrained areas, significantly improving portability and exploration efficiency.

Airgun source is an important tool for marine geophysical exploration, the construction of the gun body will be pre-injected into the release of high-pressure air into the water column, resulting in continuous oscillation until the rupture of the bubble, artificially generated energy-controllable seismic sub-wave, acquisition and analysis of sub-wave signals propagated underwater, high-resolution deep-earth exploration, and then complete the important task of oil and gas deposits prospecting and other important tasks. After decades of development, the international theory of airgun seismic source is becoming more and more mature, and a number of airgun seismic source products with excellent performance have been born, but the domestic start in this field is relatively late, and there is still a gap in technology compared with foreign countries, and there is no available domestic airgun seismic source. To address this situation, this paper firstly introduces the development history of airgun vibration source and underwater bubble motion theory, including the types and working principles of the existing mainstream airgun vibration source, introduces in detail the domestic and foreign geological exploration research based on airgun vibration source in recent years, and combines the current situation of airgun vibration source with its limitations, summarizes the current limitations of the development of China's airgun vibration source problems, and the future of the airgun vibration source of marine exploration. It also puts forward the prospect of the difficulties that may be faced in the process of localization of core equipment such as airgun seismic source in the future.

Strong ground motion duration plays a critical role in evaluating the nonlinear response of structures and holds significant importance for seismic design and disaster assessment. In this paper, two types of duration, including significant duration for a given energy portion (5%~95%) and bracket duration for given thresholds (0.05 g and 0.10 g), are calculated based on selected data from the NGA-West2 strong earthquake database provided by the PEER. The dependence of duration on magnitude and rupture distance are analyzed for both types of durations, from which an inverse variation trend between significant duration and bracket duration is found. Prediction equations of duration in terms of moment magnitude, rupture distance, and site condition are established using nonlinear regression considering the random effect and validated through comparisons with related studies. Finally, the engineering application of two types of duration is discussed pros and cons.

An M_S6.2 earthquake struck Jishishan County, Gansu Province, China, on December 18, 2023, leading to a tremendous sand boil and causing more than 150 deaths. As the trigger of the sand boils, the ground shaking during the earthquake attracts extensive concerns. In this study, ground motions of the Jishishan earthquake were reproduced by using a stochastic finite-fault modeling approach and compared with observations. To maintain the far-field received energy independent of subfault size, two improvements were made to the source spectral model. The comparison shows that: Simulated ground motions agree well with observations in terms of the waveform, peak, and duration of the acceleration time histories at reported stations. The response spectra of the observed time series are in general well reproduced by the simulation except for the remarkably large amplitude at certain periods that may result from a site response. Both the simulated and observed PGA are higher than the predictions of the empirical model but remain close to +1 standard deviation. The contour map of the simulated PGA and PGV shows a similar pattern to the observation except for the slight underestimation near the northeast and northwest corners of the study area. The maximum intensity derived from the simulated PGA is degree Ⅺ, which is also consistent with the reported shaking intensity.

The portable frequency domain electromagnetic detection instrument has been widely used in the field of urban near-surface exploration because of its high sensitivity, strong anti-interference ability, simple operation and high measurement efficiency. Due to the smaller scale of urban geological exploration targets, three-dimensional inversion is more advantageous than one-dimensional or two-dimensional inversion in terms of data interpretation accuracy. In order to study the effectiveness of the portable frequency domain electromagnetic detection instrument in urban shallow surface exploration and the superiority of 3D inversion, this paper conducts 3D inversion of GEM-2 measured data from a test site in Shanghai, and compares the results with the traditional 1D inversion results. The excavation results show the correctness of the 3D inversion algorithm proposed in this paper and the feasibility and superiority of its application in urban underground space detection.

Ground Penetrating Radar (GPR) is widely used in the fine detection of shallow structures such as urban road disease detection and archaeology, etc. Small-scale underground geological bodies such as cavities and cracks are usually the most concerned detection objects. In general, the energy of hyperbolic diffracted waves generated by small-scale underground geological bodies in the measured GPR profile is weaker than the linear reflected waves generated by the subsurface layered interface, and is easily mixed or masked by the reflected waves with stronger energy, which seriously interferes with the accurate identification and interpretation of small-scale targets. To this end, this paper proposes a multiple singular spectrum analysis method based on k-means clustering algorithm to separate the reflected and diffracted waves in the GPR profile. Then, the k-means clustering algorithm is used to cluster the all singular spectrum, and the singular values in the singular spectrum are divided into k classes according to the similarity, and the several types of singular values representing the reflected wave and the noise are set to zero, and some of the singular values representing the diffracted wave are retained to reconstruct the GPR diffraction wave field. Numerical experiments show that the multiple singular spectrum analysis method based on k-means clustering algorithm can efficiently separate the reflected wave from the diffracted wave. Compared with the reverse time migration of the unseparated GPR data, the imaging resolution of the small-scale geological body is higher and easier to be identified in the separated diffraction wave reverse time migration, which provides a feasible and effective method for the high-precision interpretation of small-scale targets.

Forensic military geophysics is a new branch of military geophysics in the field of military intelligence reconnaissance and identification with high-resolution measurements of the shallow underground layers of geophysical methods on typical military scenes including of the medium and small-scale battlefield spaces of tactics and combat military activities, which is used to directly verify the geological structure information of the reachable area, so as to infer the geophysical characteristics of the detection targets in the unreachable area by a close analogy of similarity in geological environment, and to solve the problem of effective reconnaissance and identification of geological intelligence evidence in the unfamiliar battlefield environment. A series of geophysical methods mainly include electrical methods, electromagnetic methods, seismic exploration and magnetic methods, which are applied to military scenes of the known battlefield environmental survey and the unknown battlefield intelligence reconnaissance. The key of the application to forensic military geophysical methods is similarity, contrast, rapidity, and non-destructive testing, which extremely could be in accordance with reconnaissance and identification to the physical characteristics of hidden objects as evidence constraints of battlefield environmental intelligence. The three types of typical examples consisting of projectile contact explosion damage effect in battlefield reconnaissance, military vehicle maneuverability in trace identification, and hidden target search underwater indicated that the resistivity method, R-wave survey and Ground Penetrating Radar (GPR) have a good identification effect on the remains evidence from military activities in shallow subsurface. This paper proposes that the multi-method application of forensic military geophysics is suitable to investigate and take the evidence for subsurface geological information of battlefield dynamic change conditions. The combination of geological and geophysical methods can effectively enhance the evidence identification effect. A multifaceted approach to the study of forensic geophysics and military geology will promote the theoretical innovation and practical application of military geophysics.

The soil and groundwater pollution caused by gas station leakage is one of the main types of urban underground pollution. The soil and groundwater pollution caused by gas station leaks is one of the main types of urban underground pollution. In order to achieve this goal, we design a new acquisition method of three-dimensional inter-well resistivity tomography in this paper, evaluate and analyze the ability of this technique to finely characterize and quantitatively evaluate the pollution plume through application examples. The study shows that the monitoring technology designed in this paper has good ability to portray the pollution plume formed by small leakage from gas stations, the boundary of the pollution plume formed by secondary repeated leakage, and the diffusion process of the pollution plume caused by the change of the groundwater level; and the inverted three-dimensional spatial resistivity data of the pollution plume and the oil content-resistivity model can be used to estimate the three-dimensional spatial oil content of the plume and the total leakage volume. Based on the time series inversion of the three-dimensional resistivity data, a leakage level criterion is constructed based on the concept of grading. The research paper plays an important role in promoting the three-dimensional fine characterization technology of underground organic matter pollution plumes.

The principle restoration of Zhang Heng's Seismoscope and the realization of its seismic detection function are crucial for the seismological community to recognize and accept Zhang Heng's Seismoscope as a scientific instrument. Pillar and Copper-instrument are the two most critical information in historical literature records. The Pillar must support the Copper-instrument, and the Copper-instrument must be placed on the top of the Pillar, otherwise it cannot be called Pillar; Understanding this relationship of support and positioning leads to the emergence of the principle model of the Seismoscope. The "secondary structure excitation model of primary-secondary structure resonance system" is proposed as the principle model of the Seismoscope. By utilizing the resonance amplification effect of the primary-secondary structures (at least 5.0 times), and the lever amplification effect of the trigger mechanism (at least 4.0 times), a relative displacement amplification of at least 20 times for most seismic motions is achieved, with some amplifications exceeding 50 times. Theoretically, this enables the effective excitation of Zhang Heng's Seismoscope under microseisms (imperceptible to humans). Coupled with an automatic locking system, the Seismoscope achieves an automatic seismic detection function. The primary structure (Ground-Motion) can be simplified as a "cantilever structure with a large concentrated mass at the top and supported at the bottom on a horizontal elastic foundation." The secondary structure (Wind-Observation) consists of 8 pendulums corresponding to 8 directions. Each pendulum is suspended using a "pin", and the pendulum rod is made of copper, serving as a tension-compression rod. This design ensures that the pendulum's swing direction is essentially perpendicular to the axis direction of the pin, allowing for the detection of seismic motion direction. The seismic motion direction measured by the Seismoscope is the one caused by microseisms that initially excites the secondary structure to undergo a significant displacement relative to the primary structure. If the direction aligns with the earthquake-source direction, it may be possible to measure the earthquake-source direction. Due to the significant differences in stiffness and mass between the primary and secondary structures, they can be separated and calculated separately as small-damping ideal linear elastic single-degree-of-freedom systems. The relative displacement of the secondary structure given by this simplified method is slightly smaller than the precise calculation results of the ANSYS finite element model, with a deviation of no more than 10%, indicating that the simplified method has high accuracy and credibility. The relative displacement amplification coefficient of the secondary structure is the primary indicator for whether the seismograph is easy to be excited. The relative displacement amplification coefficient spectrum of the secondary structure is proposed as the basic technical diagram for designing the Seismoscope. A statistical analysis was conducted on the calculation results of 18 sets of far-field seismic records. If requiring the relative displacement amplification coefficient of the secondary structure to be no less than 5.0, it is preliminarily believed that the optimal natural vibration period of the primary-secondary structures ranges from 2.1 s to 2.6 s. The on-site installation and adjusting process to control the natural vibration period range of the primary-secondary structures of the Seismoscope for achieving resonance effects is provided. Further field tests are required to verify its seismic detection function. The principle restoration and instrument design of Zhang Heng's Seismoscope have been preliminarily achieved. Using modern seismic observation results and structural dynamic analysis technology, the principle model of the Seismoscope proposed in this paper conforms to historical records and can achieve the (micro-)seismic detection function

With its powerful feature extraction ability, deep learning has shown great potential in various fields and provides new ideas for solving various complex problems. Deep learning models often require a large amount of labeled data for training, but in practice, limited logging data are obtained due to cost, resulting in insufficient training samples. Therefore, this paper proposes a CNN-BiLSTM based semi-supervised learning method for seismic wave impedance inversion. The interpolation resampling technique is used to augment the wave impedance, and then a semi-supervised learning strategy is introduced to train the augmented data, and the unlabeled data information is used to improve the generalization ability and performance of the model. The Marmusi-2 model test shows that it can achieve better inversion results with only a small amount of data augmentation, which verifies the effectiveness of the method in the case of small samples.

The reserve and production of oil and gas exploration and development in the global sea areas have been steadily increasing. The growth rate of ultra-deepwater oil and gas production has exceeded that of deepwater, making it a strategic replacement area for global oil and gas resources. Offshore oil and gas are a crucial part of China's oil and gas energy. With the advancement of offshore exploration towards deepwater and ultra-deepwater, it is extremely urgent to improve the technical service guarantee capabilities for exploration and well logging in these areas. Due to the special characteristics of the offshore exploration environment, such as complex structures, diverse lithologies, high temperatures and high pressures, there are special and complex requirements for the functions and performance of well logging instruments, and the technical difficulty is extremely high. Most offshore drillings are cluster wells or multi-branched wells, featuring large deviation angles, large displacements or horizontal wells. This requires well logging instruments to have higher operational efficiency and measurement accuracy. This paper sorts out and summarizes the technical bottlenecks and challenges faced by China's deepwater well logging technology and equipment. Combining with the development status, it puts forward development trends and suggestions, aiming to provide technical support for the exploration and development of China's deepwater oil and gas resources.

With the rapid development of China's economy, the consumption of shallow mineral resources continues to intensify, and the geological exploration work in China continues to advance to deep and complex areas, the exploration of deep geological resources in our country is also confronted with multiple challenges, such as detection depth, anti-interference ability, and measurement accuracy. The demand for deep exploration is also increasing, and the demand for the domestic distributed magnetotelluric acquisition station is becoming increasingly urgent. In this paper, a domestic ultra-wideband distributed acquisition station DMT-V1 for magnetotelluric detection is designed and developed. Combined with LORA autonomous network and 4G networking communication technology, remote real-time data monitoring and remote data download are realized. The developed acquisition station consists of 2 electric field channels and 3 magnetic field channels, and supports low noise fluxgate sensor, long period induction magnetic sensor, magnetic sensor field calibration; MT (Magnetotelluric) and LMT (ultra-long period Magnetotelluric) methods are supported. The AD converter uses a 32-bit high-precision ADC chip CX1282, and is equipped with a low-power hardware processor. Under ARM intermittent operation mode, the power consumption of the acquisition station can be less than 1W@12VDC. Field detection experiments were carried out on this system. Comparative tests with advanced foreign instruments showed that the performance indicators of the DMT-V1 acquisition station generally reached the international advanced level. Field application tests of the DMT-V1 acquisition station were conducted respectively in different regions, and its detection frequency could reach up to 100000 seconds, supports MT and LMT methods, and has the characteristics of high stability, low power consumption, high precision, light weight and portability. The application scope of this equipment covers shallow mineral exploration, deep and ultra-deep oil and gas exploration, and the investigation of the electrical structure of the crust and upper mantle.

Reservoir pore structure has an important influence on seepage capacity and oil production capacity. Mercury injection experiment is an important method to study reservoir pore structure, but it cannot be carried out in large quantities due to factors such as limited number of cores, high experiment cost and mercury toxicity. Through the analysis of the mercury injection experiment data, it is found that a series of mercury injection pressure is generally a fixed distribution, and the adjacent mercury injection saturation has a good correlation. As long as a mercury injection saturation of a depth point is predicted, and the entire pseudo capillary pressure curve can be predicted for that depth. XGBoost is adopted to predict the mercury saturation, and then the pore throat radius spectrum is obtained. The rock surface relaxation rate is determined by overlapping the pore-throat radius spectrum of the mercury injection experiment and the T₂ spectrum of the NMR experiment, and the two cut-off values for distinguishing small, medium and large pores on the T₂ spectrum are converted into the two cut-off values of the pore-throat radius spectrum. Using two cut-off values to divide the pore throat radius spectrum into three parts, the pore structure index R_{c_index} is proposed. This parameter has a good correlation with the oil layer production measured by the cable formation tester. It is concluded that the pore structure index R_{c_index} predicted by conventional logging curves can continuously predict reservoir production and guide follow-up measures such as test layers selection.

Wide Field Electromagnetic Method(WFEM) is an important means of middle and deep mineral exploration. According to the characteristics of WFEM data, a laterally constrainted pseudo-two-dimensional algorithm (SD-WLCI) based on skin depth weighting is proposed in this paper. Through two two-dimensional theoretical models, it is proved that the algorithm can effectively improve the horizontal continuity of deep inversion results in WFEM and improve the resolution of layer interface. In addition, the influences of three regularization factor search methods (linear search method, cooling method and adaptive regularization factor method) on this algorithm are compared in detail. The results show that the linear search method is superior to the other two methods in terms of work efficiency and stability within the comparison range. Finally, this paper applies the skin-depth weighted lateral constraint algorithm to the inversion of wide-field electromagnetic data measured in a mining area of Dongguashan, Tongling. The inversion results are highly consistent with the actual geological conditions, which provides a reliable geophysical basis for the subsequent prediction of middle-deep mineral resources and borehole location. By comparing with the traditional lateral constraint inversion results, The practicability and necessity of this algorithm in the inversion of the measured data of wide-field electromagnetic method are proved.

In response to the current situation where the acquisition of electrical spectrum parameters of rock and ore mainly relies on indoor observation of samples, and there are few experiments on obtaining electrical spectrum parameters of outcrops in the field, a self-developed measurement device was used to conduct field observation experiments. The device is designed with two switchable signal transmission modes of Stable voltage and current, as well as 8 current output levels ranging from 20 μAto 400 mA. Based on indoor measurement experiments of the standard resistance capacitance network model to ensure the functionality and stability of the device, consistency experiments, spectrum observation comparison experiments, and depth measurement device spectrum observation experiments were conducted using a symmetrical quadrupole device in Daweishan, Liuyang City, Hunan Province to obtain the electrical spectrum parameters of the formation outcrop. The experiment proves that the device can correctly obtain the observed spectrum and obtain the electrical parameters of the underground conductive medium through reasonable analysis of the spectrum. The results indicate that the device has the advantages of simple operation and strong anti-interference ability, providing an independent and practical measurement device for the measurement of the electrical spectrum parameters of geological outcrops. Through reasonable geological interpretation of the observed spectrum, it plays a certain supporting role in related geophysical exploration work and has good application value.

Time-lapse seismic technology plays a crucial role in monitoring changes in the fluid reservoir of offshore oil and gas fields. Conventional time-shifted seismic data collection is required to be repeated, but the actual data collection is difficult to be completely repeated due to different purposes. A lot of 2D and 3D datasets can meet the requirements of time-shift seismic comparison through consistency processing. With the development of multi-component seismic exploration technology, research on time-lapse seismic monitoring by combining multi-component seismic data with early towed streamer data is still limited. In order to further study in time-shift seismic research, This paper conducts a study on non-repetitive time-lapse seismic matching processing based on early towed streamer data and recent Ocean Bottom Cable (OBC) seismic data collected from an oil field in the South China Sea. Due to significant differences in acquisition time, acquisition methods, parameters, and seismic geometry between the two datasets, noise, multiples, and ghost waves exhibit distinct differences. After performing denoising, multiple removal, and ghost wave suppression on the two datasets, the study analyzes their differences, focusing on specific factors such as amplitude, frequency, signal-to-noise ratio (SNR), wavelet phase, and time difference. A mutual equalization method is selected for consistency matching processing of both datasets before pre-stack depth migration. Subsequently, based on the migrated results, it is analyzed whether post-migration consistency matching is required. The results of the study indicate that the raw OBC data has stronger amplitude energy, narrower frequency bandwidth, higher signal-to-noise ratio, and smaller wavelet sidelobes compared to the raw towed streamer data. The consistency processing workflow can effectively match the differences in amplitude energy, frequency bandwidth, signal-to-noise ratio, wavelet phase, and time difference between the two datasets. After consistency matching, both datasets can be used for time-lapse comparison analysis. This research provides a feasible matching processing method for non-repetitive time-lapse seismic studies using multi-component OBC seismic data in conjunction with towed cable data, laying the foundation for subsequent time-lapse seismic studies based on multi-component data.

Carbon Capture, Utilization, and Storage (CCUS) has emerged as a effective strategy for mitigating atmospheric CO₂ emissions. However, the large-scale subsurface injection of CO₂ may induce ground surface deformation, potentially compromising the long-term integrity of CO₂ storage. To investigate surface displacement patterns in CCUS project areas following CO₂ injection, this study employed the Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) technique to analyze 45 ascending Sentinel-1A images acquired between January 2022 and August 2024, enabling comprehensive deformation monitoring over a 2.5-year period in a Chinese CCUS geological storage site. Recognizing the impact of tropospheric delay errors on InSAR measurements, we developed an enhanced atmospheric delay correction approach synergizing Global Navigation Satellite System (GNSS) data and ERA5 reanalysis data, specifically optimized for small-scale regional applications. Analysis of the phase-unwrapped interferograms confirms that our GNSS-constrained atmospheric correction approach significantly reduces tropospheric delay artifacts in localized interferometric observations. The derived InSAR deformation field reveals distinct surface displacement patterns in the CO₂ injection zone, with measured vertical deformation rates ranging from 0 to 10 mm/a. Deformation time series analysis at monitoring points adjacent to injection wells reveals a consistent response: initial surface uplift immediately following CO₂ injection, followed by gradual subsidence after several months. This pattern was confirmed through comparative analysis of InSAR-derived vertical displacements (converted from LOS measurements) and independent GNSS observations. This study demonstrates that the synergistic use of GNSS and InSAR technologies enables precise monitoring of millimeter-scale surface deformation in CCUS operational areas. The GNSS infrastructure serves dual purposes: (1) enhancing InSAR measurement accuracy through tropospheric delay correction, and (2) providing independent validation of InSAR-derived deformation results. These findings underscore the critical importance of integrated geodetic monitoring for ensuring the safety and efficacy of CO₂ geological storage operations.

In recent years, Distributed Acoustic Sensing (DAS) technology has been widely used in earth interior structure research, underground space detection, and microseismic monitoring due to its advantages of high resolution, wide-band measurement, and real-time monitoring. DAS's noise type and complexity are higher than that of conventional geophone data because it is affected by various factors such as optical systems, demodulation algorithms, fiber-optic cable and ground coupling, and transverse and longitudinal waves. The data signal-to-noise ratio is lower than that of the data recorded by conventional geophones under the same noise level. Therefore, higher requirements are put forward for DAS data denoising methods. In recent years, domestic and foreign scholars have researched denoising and weak signal enhancement methods for different types of DAS data. This paper mainly introduces and discusses the latest research progress in denoising DAS data in seismic exploration and natural seismology exploration using traditional physical methods and different denoising methods. The challenges faced by the current DAS data denoising methods are summarized, and the future development trend is forecasted.

The high-frequency mass variations of the atmosphere and oceans have significant impacts on the inversion of time-variable Earth gravity fields using GRACE. This paper comprehensively compares and analyzes the atmospheric and oceanic components, as well as their combination, of the AOD1B RL06 and RL07 products released by GFZ using methods including spectral domain analysis, comparison of low-order terms, spatial analysis, principal component analysis, and inversion of time-variable gravity field models. The results show that the differences between the two sets of products in the spectral domain are relatively small, with the main differences being reflected in the oceanic component. However, the comparison results in the spatial domain indicate significant differences in the equivalent water heights of the oceanic component, reaching the decimeter level. The differences between the 60th order time-variable gravity field models inverted based on RL06 and RL07 are relatively small, with differences in the spectral domain mainly concentrated in medium to high orders. The RMS differences of the KBRR residuals after validation for both sets of time-variable models are less than 3.912 nm/s (correlation coefficient: 0.999). However, the KBRR residuals computed based on RL07 products are generally smaller, demonstrating the slight advantage of RL07 in the inversion of time-variable Earth gravity field models.

The fluxgate sensor is a kind of important equipment to measure the vector magnetic fields. Its working principle is based on the non-linear properties of soft magnetic material. It is used to detect weak magnetic fields. In this paper, we introduce the basic working principle, structure, and main applications of fluxgate sensor. We also discuss the research development of fluxgate sensor, focus on the progress of fluxgate sensor in miniaturization, intelligence, digitization, as well as reducing noise. Some further research directions are also provided.

The airborne transient electromagnetic method is an important electromagnetic exploration technology, which obtains the information about the electrical structure of the earth through inversion.But due to the slow progress of the forward, the inversion will consume a significant amount of time.Aiming at the problem of long computation time of traditional forward method, this paper proposes a deep learning-based fast forward method for airborne transient electromagnetic.The method first uses the traditional finite volume method to calculate the induced electromotive force of a large number of different ground resistivity models to form a training dataset; then it designs a ResNet-UNet deep neural network; then it trains the network with the training dataset; finally, it inputs the ground resistivity model into the trained neural network to obtain the forward results.In order to verify the accuracy and efficiency of the method, the forward results of the ResNet-UNet deep neural network and the traditional finite volume method are compared.The experimental results show that the average relative errors of the entire validation set is less than 1.2%, with 87% of the average relative errors falling within the range of 0.1% to 0.3%, and the speed of deep neural network forward is about 2934 times higher than that of the traditional finite volume method, which significantly improves the forward efficiency of the airborne transient electromagnetic.The method is capable of fast forward of airborne transient electromagnetic, which can be put into the existing inversion framework to accelerate the inversion speed of large datasets.

Impedance inversion based on convolution models generally involves inversion of reflection coefficients before impedance inversion, and there are few studies that directly invert impedance using wave equations. Given that wave equation based methods have higher accuracy, this paper proposes a wave impedance inversion method based on wave equations. To ensure the lateral continuity of the impedance inversion model, Fourier series expansion is used in this paper to represent the wave impedance model using Fourier coefficients, which are used as inversion parameters; In the inversion process, an inversion strategy was designed to transition from low wave numbers to high wave numbers and to partition data groups. This not only ensures the inversion from the background wave impedance model to the fine wave impedance model, but also effectively avoids data redundancy, greatly improving computational accuracy and efficiency. The test results of synthetic seismic records and actual seismic data show that the method proposed in this paper has a significant ability to characterize the local edge features of "bead like" reservoirs, which can provide data support for the characterization and early reserve evaluation of fractured reservoirs.

The dispersion and attenuation of seismic waves are critical physical properties of hydrate reservoirs, essential for the precise identification and quantitative characterization of hydrate reservoirs. The coupled effects of hydrate saturation and occurrence modes typically exert significant influences on seismic wave velocity dispersion and attenuation characteristics. However, most existing attenuation characterization theories are established based on assumptions of single or limited hydrate morphologies, making it challenging to accurately describe hydrate reservoirs with complex occurrence states. To address this, this study integrates the generalized effective medium model and two-phase media theory to develop an attenuation theoretical model for hydrate reservoirs that simultaneously considers four occurrence modes: contact-cementing, grain-coating, matrix-supporting, and pore-filling. A comparison of P-wave attenuation characteristics modeled using Biot theory (focusing on global fluid flow in porous media) and BISQ theory (Biot-Squirt theory, incorporating both global fluid flow and local squirt mechanisms) under different hydrate morphologies reveals that the attenuation predicted by Biot theory is significantly lower than that of the BISQ model. This discrepancy arises because the BISQ framework accounts for additional energy loss mechanisms, such as microscopic fluid squirt between hydrate-coated grains, which are critical in heterogeneous hydrate-bearing sediments. Based on the BISQ-derived attenuation model, a morphology-constrained attenuation rock physics template is developed, and a crossplot of P-wave velocity, attenuation, and hydrate saturation is constructed to identify hydrate occurrence modes. Furthermore, the developed morphology-constrained attenuation rock physics template is applied to logging data from Sites 1247B and 1250F of the Ocean Drilling Program (ODP) Expedition 204. Field results demonstrate that the constructed crossplot of P-wave velocity and attenuation versus hydrate saturation accurately identifies hydrate occurrence modes, and the attenuation rock physics template aligns well with the distribution characteristics of logging data. These research outcomes provide a new theoretical foundation and technical methodology for utilizing seismic attenuation attributes to identify and evaluate hydrate reservoirs, confirming the capability of attenuation theory to accurately diagnose occurrence states and estimate reservoir physical parameters.

Artificial intelligence technology is one of the most important development directions of detailed reservoir description in the future. Detailed reservoir description provides a high-quality platform and foundation for the development and application of artificial intelligence technology. Artificial intelligence also provides a powerful tool and way for the development and progress of fine reservoir description from digitization to intelligence. The research status, advantages and disadvantages of artificial intelligence technology application in fine reservoir description at home and abroad are compared.The application of artificial intelligence technology almost covers all aspects of detailed reservoir description, mainly including fine stratigraphic division and comparison based on analogy learning, fine interpretation of volcanic reservoir structure based on ant colony algorithm, sedimentary microfacies and reservoir configuration division and identification of expert system, fine logging secondary interpretation based on artificial neural network, fine reservoir evaluation based on grey system theory, training image establishment and multi-point geostatistical modeling based on machine learning, knowledge discovery and data mining reservoir flow unit research, fine reservoir description result management platform based on knowledge system, etc. Finally, 10 problems existing in the application of artificial intelligence technology in fine reservoir description and 10 development directions in the future are pointed out.

Detecting the quality of cement bonding and the corrosion damage condition of the casing is an important measure to maintain the integrity of the wellbore. However, wellbore integrity logging faces challenges such as poor circumferential resolution, low measurement accuracy, and poor evaluation effect of light weight cement. To solve these problems, an ultrasonic vertical incidence and oblique incidence combined measurement design scheme was proposed, and an ultrasonic Lamb-wave scanning imaging logging tool was developed. The tool is composed of a rotary-scanning acoustic assembly, a roller elastic centralizer, a drive mechanism and measuring sensor assembly, and an electronic bin sub. The tool adopts a fast logging operation mode, mainly relying on downhole storage with cable transmission as a subsidiary, and clearly demonstrates the cement bond quality and casing corrosion damage degree with high-resolution images. Field applications indicate that the full 360° scanning measurement of the tool completely covers the entire circumference of the wellbore, and the azimuth imaging map can directly present the circumfluence position of cement loss, channels, and casing corrosion. It holds significant advantages in the evaluation of light cement cementing quality, providing technical support for the safe and stable production of oil and gas wells.

For Bohai metamorphic buried hill, with the characteristics of rapid vertical and horizontal velocity change and strong reservoir heterogeneity, the prediction of fracture reservoir plays a crucial role in the oilfield development. Conventional fracture prediction methods are difficult to accurately predict the buried hill fracture reservoir. A new pre-stack inversion method for high-angle fracture reservoir is proposed in this paper. The method firstly carries out texture analysis based on gray level co-occurrence matrix on the original seismic data, extracts the seismic high-angle reflection information reflecting the response of the fracture reservoir, and integrates it with the low-frequency model of conventional pre-stack inversion to construct a low-frequency model suitable for the strong heterogeneous reservoir in the buried hill of the Archean space, and then carries out fine pre-stack inversion. In this paper, the method is applied to the prediction of fracture reservoir in Bozhong 26-6 oilfield. The results show that the method can improve the inversion effect and the prediction accuracy of fractured reservoir in Bozhong 26-6 oilfield, and provide a reliable basis for the deployment of development wells in Bozhong 26-6 oilfield.

Microseismic signals generated by minor fracturing or deformation in rock masses are often weak and significantly affected by environmental noise, making it challenging to accurately identify effective signals and locate the fracturing source spatially. To eliminate noise superimposed on the fracturing signals and improve the Signal-to-Noise Ratio (SNR) of weak microseismic signals, this paper proposes a denoising method that combines Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and Singular Spectrum Analysis (SSA). First, CEEMDAN is used to obtain the Intrinsic Mode Functions (IMFs) of the signal, and energy entropy is employed to optimize the signal components, removing low-frequency noise. Then, SSA is applied to the reconstructed signal to decompose it into components corresponding to different singular values. Using singular values, the reconstructed components are determined, and the final reconstructed signal achieves secondary filtering. The study is of significant importance for analyzing the location of weak microseismic events induced by fracturing in rock slopes and monitoring the dynamics of landslide hazards.Based on the theoretical and experimental results, the following conclusions can be drawn: (1) The traditional EMD method shows poor frequency separation effect when decomposing weak signals. Due to the strong coupling between microseismic weak signals and random noise, modal aliasing occurs in the components.(2) The simulation results of noisy sinusoidal function waveforms indicate that the SNR of the simulated waveform before denoising was 11.34 dB. After applying this method, the SNR improved to 21.53 dB, the root mean square error was reduced by 74.24%, and the signal energy was maintained at 98%. This method demonstrates a significant denoising effect.(3) Denoising of microseismic signals generated by hydraulic fracturing in the SF-6 well of the Fuling shale gas field in Chongqing shows that the high-frequency band denoising effect is superior to that of EMD and EMD-wavelet threshold methods.(4) Denoising experiments on three microseismic signals effectively removed the background noise, preserving the characteristics of the microseismic weak signals.

Low-permeability tight gas reservoirs are rich in reserves, but their natural gas spatial distribution prediction is extremely challenging due to the complex factors such as reservoir heterogeneity, anisotropy, low porosity, and low permeability. Especially under conditions of limited well data, the lack of core test data, unclear logging and seismic response mechanisms, and insufficient geological understanding restrict the accuracy of gas content identification in low-permeability tight reservoirs. Therefore, this paper proposes a method for gas content tight reservoirs based on the Modern TCN deep learning algorithm under conditions of few wells.First, the sensitive parameters for gas content response are analyzed using well log data, such as sonic time difference (DT), shear sonic time difference (DTS), and density (ρ). Second, the Modern TCN (Modern Temporal Convolutional Network) deep learning network is constructed, with the sensitive parameters as the input for model training and testing. Finally, the decoupled design is used to separate the temporal and feature information of sensitive parameters, fully capturing the gas content characteristics of the reservoir and predicting the spatial distribution characteristics of the reservoir. This method was applied to the gas content identification of tight clastic gas reservoirs in the Huangyan structural belt of the Xihu Sag in a certain sea area, achieving a good well-seismic matching effect. It proves that this method can provide support for exploration and development of low-permeability tight clastic gas reservoirs under few-well conditions.

It is currently challenging to accurately detect overburden cavities ahead of shield tunneling in subway tunnels. Given the shallow depth of subway tunnels, this study investigates the principle of super-ahead detection of overburden cavities in shield tunnels using small coil transient electromagnetic detection technology, as well as the transient electromagnetic anomaly characteristics. Numerical simulation studies show that when the transient electromagnetic coil is located directly above the cavity, an area resembling a "butterfly" is generated in the electric field response. The closer the distance to the cavity, the greater the electric field response influenced by the cavity. Due to the rapid attenuation of the transient electric field in the cavity, differences exist in the apparent resistivity responses, allowing for the identification of underground cavities. Through engineering practice, the transient electromagnetic anomaly characteristics of dry cavities mainly manifest as isolated, discontinuous, and easily closed-loop electromagnetic curves, with high resistivity anomalies in apparent resistivity data. Wet cavities exhibit transient electromagnetic anomaly characteristics with some continuity in the electromagnetic curves, often showing convex or concave features, and low resistivity anomalies in apparent resistivity data. Ground transient electromagnetic methods can effectively identify shallow underground cavities.

In actual logging operations, various factors such ascomplex wellbore environment, complex geological structures, and wellbore collapses can lead to the issue of missing logging curves, and the cost of remeasuring data is high. To address the issue of missing logging curves, this paper first employs Singular Spectrum Analysis (SSA) to decompose the original logging curves, utilizing the more correlated components for more efficient curve completion. Furthermore, a logging curves completion model based on graph attention network incorporating Multi-Head Attention Mechanism and Bidirectional Gated Recurrent Units (GAT-MABiGRU) is proposed. In the completion experiments for the RHOB and DT logging curves, results show that the GAT-MABiGRU model based on SSA outperforms Support Vector Regression (SVR), Multi-Layer Perceptron (MLP), Long Short-Term Memory Network (LSTM), and Temporal Convolutional Network (TCN) in terms of Root Mean Squard Error(RMSE), Mean Absolute Error(MAE), and coefficient of determination(R²). Ablation experiments and blind well experiments further verify the effectiveness of incorporating SSA and GAT modules in improving the model's prediction accuracy, providing a new method for logging data completion.

Due to the limitation of the resolution of traditional time-frequency analysis algorithm, the energy of time-frequency spectrum diverges when seismic data are converted from time domain to time-frequency domain. The generalized s-transform of multi-channel synchronous extrusion has high time-frequency focusing ability, which can keep fracture information as much as possible and reduce time-frequency error when seismic data are transformed from time domain to time-frequency domain, then the fracture identification in time-frequency domain is carried out. The low frequency band of seismic data is advantageous to Kingdom of Dali-scale fault identification, while the high frequency band of seismic data is advantageous to the identification of small faults. Therefore, the fault identification algorithm based on time-frequency domain can characterize the fault information of different scales. In this paper, the technique of clutter detection and multi-channel synchronous squeezing generalized s-transform are combined to carry out fault identification of seismic data. The theoretical model and the single-channel signal show that the generalized s-transform algorithm of multi-channel synchronous extrusion has high time-frequency focusing ability, the micro-fractures can also be well characterized.

The assessment and analysis of ecosystem service functions are of great significance for the rational allocation and optimization of regional resources. Taking Cangzhou City as the research object, the ecosystem service functions in 2005, 2015, and 2020 were evaluated based on the InVEST model. The Spearman correlation coefficient was used to analyze the trade-off and synergy between ecosystem services, and the self-organizing mapping method was used to identify ecosystem service bundles. The research results indicate that the water yield in Cangzhou City shows a trend of first increasing and then decreasing, while the N/P output first decreases and then increases, while carbon storage and habitat quality remain stable with some improvement. There is a significant synergistic relationship between water yield and N output, water yield and P output, N output and P output, carbon storage and habitat quality. There is a significant trade-off between water yield and carbon storage, water yield and habitat quality, N output and carbon storage, N output and habitat quality, P output and carbon storage, and P output and habitat quality. There are three ecosystem service bundles in Cangzhou City, namely ESB1 focused on water yield and N/P output, ESB2 focused on habitat quality and carbon storage, and ESB3 focused on cultural tourism.The research results provide decision-making basis for improving the overall efficiency of ecosystem services and differentiated management of ecological functional areas in Cangzhou City.

Research has shown that seismic waves usually undergo different degrees of velocity dispersion and attenuation when they encounter hydrocarbon-bearing reservoirs during propagation, which also leads to a close correlation between the reflection coefficient and frequency. Therefore, we can utilize the velocity dispersion property extracted by hydrocarbon-bearing reservoirs AVO inversion for fluid identification. Frequency-dependent AVO inversion is performed based on the amplitude spectrum obtained from the time-frequency analysis of seismic data. The resolution and accuracy of the time-frequency analysis are critical factors influencing the results of dispersion attribute inversion. In recent years, time-frequency analysis methods based on sparse representation have gained attention due to their high time-frequency resolution. This paper proposes a more flexible sparse time-frequency analysis method based on compressed sensing theory and constrained by the L_P quasi-norm. Numerical models demonstrate that this method achieves higher resolution time-frequency distributions, making it suitable for seismic signal analysis. By integrating this L_P quasi-norm sparse time-frequency analysis method with frequency-dependent AVO inversion, it is possible to accurately extract P-wave dispersion attributes, thereby identifying fluids in reservoirs. Field data validation shows that the frequency-dependent AVO inversion method based on sparse time-frequency analysis not only provides high resolution but also offers reliable fluid indicators for hydrocarbon reservoirs, offering strong technical support for the identification of complex reservoirs.

High-precision Earth Orientation Parameter (EOP) forecasting, which encompasses parameters such as Polar Motion (PM), universal time, and length of day, is crucial for facilitating the transformation between terrestrial and celestial reference frames in various applications (e.g., satellite autonomous navigation, deep space exploration, and geodynamic research). However, the mainstream predictive methods (including Least squares, numerical decompositions) have serious limitations, such as tail effect, prior periods and poor estimations of model parameters. To improve the accuracy of EOP forecasting (1~360 days), this paper introduces an integrated model combining Singular Spectrum Analysis (SSA), Prony's method, and Autoregressive (AR) models, demonstrated through the case of PM parameter prediction: Firstly, the SSA is used to separate the principal components (e.g., trend, annual, and Chandler terms) and residual components from the original PM observations. Secondly, combined with the Prony method to model and extrapolate these principal components based on complex exponentials functions; and we combine the widely used AR method to predict their residuals. To verify this combined approach, we conduct multiple prediction experiments based on the IERS EOP 20 C04 data. The experimental results showed that the SSA+Prony+AR model effectively captures the time-varying characteristics and significantly mitigates the tail effects of the PM components. Compared with traditional LS+AR models and IERS Bulletin A forecast products, our proposed model exhibits superior performance in medium to long-term polar motion forecasting, particularly reducing forecast errors by nearly 40% in the X direction. These findings can also provide valuable insights into the forecasting of other EOP parameters.