Rock fragmentation is a critical process in mining engineering, tunnel excavation, coalbed-gas exploitation and ultra-deep well construction. Conventional mechanical rock-breaking techniques suffer from severe bit abrasion, high construction cost, and low operational efficiency when applied to high-strength rock under complex geological conditions. As an emerging high-efficiency fragmentation technology, high-voltage electrical pulse (HVEP) rock breaking can induce rock failure via instantaneous energy release. Nevertheless, a comprehensive quantitative evaluation framework for fracture initiation and rock mass damage evolution induced by HVEP has not been well established. This study aims to reveal the intrinsic mechanism of fracture initiation and propagation in rock mass subjected to HVEP loads, clarify the temporal relationship between rock mass damage and fracture evolution driven by shock waves, and identify the effects of key discharge circuit parameters on rock-breaking performance.

Based on fracture mechanics and damage mechanics theories, a coupled phase-field fracture numerical model was constructed by integrating the RLC discharge circuit and the Weizel-Rompe plasma-arc impedance model. A tensile-compressive strain decomposition algorithm was adopted to distinguish tension-dominated fracture from pure compressive elastic deformation. Numerical simulations were carried out under typical working conditions: Discharge voltage 9-13 kV, capacitance 2-5 μF, circuit inductance 5-5000 μH, and plasma-channel length 0.03-0.06 m. The phase-field variable was employed to quantitatively calculate initial fracture propagation length and damage area of the rock mass. The effects of shock wave characteristics, discharge voltage, capacitance, circuit inductance, and plasma-channel length on rock mass fracture behaviors were systematically analyzed.

The simulation results demonstrated that rock mass damage evolution was strongly correlated with shock wave intensity and its rising rate. Damage did not equal fracture propagation, and damage initiation preceded macro-fracture initiation and served as a precursor of fracture formation. Higher shock wave intensity and faster rising rate accelerated fracture propagation, increased the degree of rock mass damage, and expanded the damage extent. Rock-breaking performance improved with the increase of discharge voltage and energy storage capacitance, whereas it degraded as plasma-channel length increased. Minor variations in circuit inductance had little influence on rock-breaking performance, while a substantial increase in inductance suppressed instantaneous energy release and weakened rock fragmentation. The waveform of displacement at the HVEP action point was similar to that of shock wave pressure. Influenced by rock material deformation and internal energy dissipation, the displacement response exhibited an obvious hysteresis relative to shock wave pressure.

The proposed coupled model realizes quantitative characterization of fracture growth and rock mass damage under HVEP loading. The findings can offer theoretical references for parameter selection and commissioning of HVEP rock-breaking equipment in practical rock-engineering applications.

Traditional mineral prospectivity models show limited capacity for fusing multi-source geoscience datasets under complex geological settings. Current ensemble-learning-based mineral prospectivity studies still have two major drawbacks: Insufficient analysis of complementary effects among base learners in identifying geological ore-controlling variables, and inadequate integration of prior geological constraints within data-driven modelling workflows. To address these gaps, this study constructs a two-layer Stacking ensemble learning framework for skarn-type iron-gold polymetallic mineral prospectivity prediction.

The Stacking architecture consisted of a base-learner layer and a meta-learner layer. Random forest (RF), extreme gradient boosting (XGBoost), and categorical boosting (CatBoost) were selected as base learners, while logistic regression was adopted as the meta-learner to synthesize prediction outputs from individual base models. A perturbation-based feature importance scoring evaluator was developed to quantitatively interpret geological predictive variables under nonlinear model conditions. Taking the skarn-type iron-gold polymetallic deposits in the Miaoshan-Xintai area of western Shandong as the research case, this study integrated multi-source geological, geophysical, and geochemical datasets. Redundant variables were eliminated through combined Pearson linear-correlation and Spearman rank-correlation analyses, and 17 core predictive factors were finally retained. The synthetic minority oversampling technique (SMOTE) was applied to mitigate severe sample imbalance between mineralized positive samples and barren negative samples. A five-fold stratified cross-validation strategy was implemented for model training and comprehensive performance assessment.

The Stacking ensemble model outperformed three standalone base-learner models (RF, XGBoost, and CatBoost) in multiple quantitative indicators including accuracy, precision, F1-score, and AUC value. The spatial distribution of predicted mineral prospectivity probabilities showed high consistency with the locations of known skarn-type deposits in the study area. Three base learners exhibited obvious complementary capabilities in identifying ore-controlling geological factors. Model interpretation results further revealed that fault structures and Late Yanshanian intrusive rock bodies represented the dominant ore-controlling factors for skarn-type iron-gold polymetallic mineralization in this study area.

By integrating model-derived mineral prospectivity maps with regional metallogenic geological constraints, three prospecting target areas with exploration potential are delineated for further field investigation. The proposed perturbation-driven feature-importance evaluation tool provides a useful methodological reference for mineral prospectivity modelling. The technical workflow combining multi-source geodata fusion and heterogeneous-model collaborative optimization substantially improves the reliability of mineral prospectivity prediction and can provide technical support for the new-round national mineral exploration breakthrough strategy. Nevertheless, further improvements are still required in intelligent hyper-parameter optimization and multi-scale sample balancing strategies, thereby enhancing the generalizability of the mineral prospectivity approach based on ensemble learning.

The Arqiale Pb-Zn-Cu deposit is situated on the southwestern margin of the Yishijilik metallogenic belt in the western Tianshan Orogen, Xinjiang, northwestern China. The ore bodies occur as stratiform-stratoid bodies within limestone of the Lower Carboniferous Akshak Formation. No ore-related intrusive bodies have been identified in surface exposures or drill cores in the mining area. Two competing genetic models, i.e., sedimentary exhalative (SEDEX) and distal skarn-type origins, have long been proposed for this deposit. The genetic link between mineralization and magmatism remains unclear, restricting exploration targeting at depth and along the periphery of the deposit. This work aims to resolve this genetic controversy by elucidating mineral-forming mechanisms, constraining ore-forming physicochemical conditions, and determining deposit genesis, providing mineralogical evidence for regional prospecting.

Two generations of magnetite (granular and radial aggregates) and abundant ilvaite in the skarn ores were selected for integrated research. Systematic petrographic observations were conducted using transmitted- and reflected-light microscopy, and the major-element compositions of magnetite and ilvaite were determined by electron probe microanalysis (EPMA). Multiple geochemical thermometers, including the Al+Mn-Ti+V discrimination diagram and Mg-in-magnetite geothermometer, were applied to estimate ore-formation temperatures, and major-element ratios in ilvaite were used to evaluate the redox state of ore-forming fluids.

Both granular and radial magnetite showed geochemical signatures typical of hydrothermal skarn magnetite, characterized by low Ti and V contents and elevated Al and Mn concentrations. The Al+Mn-Ti+V diagram yielded mineralization temperatures ranging from 200 to 300 °C, whereas the Mg-in-magnetite geothermometer showed average temperatures of 381-392 °C, collectively indicating a moderate- to low-temperature hydrothermal mineralization environment. Ilvaite in the ores was classified as a manganiferous ilvaite variety characterized by low Fe²⁺ and high Mn²⁺. Its TFeO/(TFeO+MnO) ratios varied from 0.86 to 0.93, indicating relatively oxidizing conditions of ore-forming fluids. Petrographic observations revealed well-developed skarn mineral assemblages of garnet-hedenbergite-ilvaite-actinolite and pervasive metasomatic ore textures. A distinct vertical metal-zoning pattern was recognized. Shallow levels were dominated by Pb-Zn mineralization, whereas Cu-dominated mineralization prevailed at deeper parts.

Integrated mineralogical and geochemical evidence indicates that the Arqiale deposit is a distal skarn-type Pb-Zn-Cu deposit genetically linked to concealed intermediate-acid intrusions at depth. The metasomatic textures and skarn paragenesis contradict the typical features of SEDEX deposits. According to the observed metal zonation, the deep northern part of the mining area possesses favorable prospecting potential for concealed contact-zone Fe-Cu ore bodies related to hidden magmatic intrusions.

The Qimantag area in Qinghai Province is located on the southern margin of the Qaidam Basin and belongs to the East Kunlun metallogenic belt. Extensive Late Paleozoic to Mesozoic magmatic activities are developed in this region, producing numerous Fe-Cu-Pb-Zn polymetallic deposits, such as Hutouya, Kaerqueka, and Niukutou. Previous studies have confirmed that most deposits in this belt record two episodes of magmatism and mineralization. The Yemaquan deposit is a typical skarn-type Fe-polymetallic systems in the Qimantag area. Previous studies have focused only on the Late Triassic granite-skarn metallogenic system of this deposit. Triassic magmatism has generally been considered the sole controlling factor for mineralization. The metallogenic potential of Devonian intrusive rocks lacks chronological and geochemical constraints. Whether two independent magmatic-mineralization episodes occurred in the Yemaquan deposit has long remained controversial. To resolve this scientific issue, this study systematically constrains the timing, magma source, and geodynamic setting of two episodes of intrusions and associated skarn mineralization.

This study focused on the Late Triassic intrusive rocks in the western Yemaquan deposit, the Devonian granodiorite from deep drill cores in the M13 magnetic anomaly zone in the southeast, and garnet from skarn in the contact zone. Field geological mapping and petrographic microscopic identification were conducted. Zircon and garnet LA-ICP-MS U-Pb dating, whole-rock major- and trace-element analysis, and in situ Lu-Hf isotope analysis of zircon were also conducted to systematically constrain the formation ages, magma sources, and tectonic settings of the two episodes of intrusive rocks.

Zircon ²⁰⁶Pb/²³⁸U U-Pb dating yielded consistent weighted average ages of 225.0-223.4 Ma (MSWD=1.7-2.4) for diorite, monzogranite, and K-feldspar granite from the southwestern part of the deposit. The ages were highly consistent with the mica Ar-Ar mineralization age of (222±1.3) Ma from the deposit, constraining the timing of Late Triassic magmatism and associated skarn mineralization. Deep granodiorite samples from the M13 zone yielded zircon ages of 392.8-392.0 Ma, whereas garnet from the skarn yielded a Tera-Wasserburg isochron age of (407.7±5.1) Ma. The two sets of ages were basically consistent within the error range, demonstrating the synchronous development of magmatic activity and Fe-polymetallic mineralization during the Early-Middle Devonian. Whole-rock geochemistry analyses indicate that both intrusive suites belonged to the high-K calc-alkaline series. The Late Triassic rocks were metaluminous to weakly peraluminous, whereas the Devonian granodiorite underwent intense hydrothermal alteration and severe Na loss, resulting in strongly peraluminous characteristics. Both intrusive had crust-mantle mixed magma sources: The primary magmas originated from partial melting of lithospheric mantle and experienced crustal material assimilation during upward migration. The two magmatic episodes formed during different tectonic stages. The Devonian intrusions formed in a post-collisional extensional setting after the closure of Proto-Tethys Ocean and continental block collision, whereas the Late Triassic intrusions formed in an intraplate extensional setting after the termination of Paleo-Tethys Ocean subduction. Integrated geochronological and geochemical data demonstrated that the Yemaquan Fe-polymetallic deposit recorded two episodes of magmatic intrusion in the Early-Middle Devonian and Late Triassic, respectively, which triggered two independent episodes of skarn-type Fe-Cu-Pb-Zn mineralization. This study first identified the Devonian skarn mineralization event in the Yemaquan ore field and revised the conventional understanding that only Triassic mineralization occurred in the deposit.

The findings improve the metallogenic evolution theory of the Qimantag belt, providing important theoretical foundations and exploration directions for targeting two phases of intrusive rock-related polymetallic deposits in this region.

The Zhenghe area in Fujian Province, southeastern China, is a gold-polymetallic ore concentration region, where magmatic-hydrothermal deposits are well developed and exhibit complex and diverse patterns of mineralization and alteration. However, due to complex geological conditions and limited research, the characteristics of mineralization and alteration remain unclear.

This study focused on the Yanpitou polymetallic exploration area on the northwestern margin of the Dongkeng volcanic basin. Detailed field geological surveys, drillhole mineralization-alteration zoning, and quantitative alteration analyses were conducted to provide valuable information on wall-rock alteration and geochemistry for further deep exploration.

The results indicated that the Yanpitou area was characterized by both Pb-Zn and Cu-Mo mineralization. With mineralization as the center, alteration exhibits a clear zoning pattern, which could be divided into the following zones: Potassic alteration zone, skarn zone, metal mineralization zone, mica schist zone, and hornfels zone. From the potassic to skarnized zone, elements such as Si, Al, TFe (total iron), Ca, and K were depleted. In contrast, from the skarnized and hornfels zones to the mineralized zone, most elements migrated toward the mineralized zone, indicating that the ore-forming fluids were active during mineralization. The results suggested that potassic and skarn alterations were closely associated with mineralization in the Yanpitou polymetallic exploration area, and the mica schist zone was also linked to mineralization.

Based on previous research, a mineralization-alteration zoning model for the Yanpitou polymetallic exploration area was established. The results indicate significant potential for deep exploration. Future exploration should focus on the northwestern direction toward the Huangshegang area in the Yanpitou region, where extensive skarn deposits occur within the Tieshan complex.

Continental interlayer-type shale oil in China has become a critical strategic resource for increasing reserves and production. High-clay siltstone interlayers in the First Member of Qingshankou Formation (Qing-1 Member) of Gulong Depression, Songliao Basin, are important carriers of shale-oil micro-migration and local accumulation. Nevertheless, controlled jointly by clay minerals, fine throats, and heterogeneity in pore-throat connectivity, high-clay siltstone interlayers within shale-oil sequences commonly show distinct oil-bearing characteristics but poor fluid mobility. Existing studies lack quantitative coupling relationships between pore-throat parameters derived from high-pressure mercury intrusion porosimetry (MIP) and fluid-mobility indicators obtained from nuclear magnetic resonance (NMR) for such high-clay interlayers, and sweet spot classification still largely relies on empirical judgment. This study targets high-clay siltstone interlayers of the Qing-1 Member in the Gulong Depression, Songliao Basin. Under the constraints of the same well and identical lithofacies, this study aims to clarify the intrinsic relationship between pore-throat structure and fluid mobility, establish a quantitative sweet spot identification criterion, and mitigate misjudgment risks caused by merely using oil-bearing indicators for reservoir evaluation.

Ten groups of paired core samples from Well PⅡ were collected, and the mineral difference index (MDI)≤15% was used to ensure mineralogical comparability between MIP and NMR subsamples. Integrated laboratory tests, including high-pressure mercury intrusion porosimetry, two-dimensional NMR under original oil-bearing conditions, and centrifugal NMR after oil washing and simulated formation-water saturation, were conducted to obtain key parameters such as median pore-throat radius r₅₀, sorting coefficient σ, mercury withdrawal efficiency, transverse relaxation time T₂ cutoff value, movable-fluid saturation S_MF, and saturated-state geometric mean of T₂. A composite movable-oil index (CMOI=S_o×S_MF/100) was constructed to jointly characterize the contributions of both oil-bearing characteristics and fluid mobility. Based on micro-scale computed tomography (CT) scanning and scanning electron microscopy (SEM) observations, microscopic mechanisms of throat narrowing and pore-throat connectivity impairment induced by clay film coating and bridging infill were revealed to explain the statistical patterns observed in the macroscopic experiments.

For high-clay siltstone interlayers, pore-throat parameters from MIP showed significant correlations with NMR mobility indicators. With increasing clay content, r₅₀ decreased, while S_MF, CMOI, and the saturated-state geometric mean of T₂ decreased simultaneously. Affected by clay-bound water, the T₂ cutoff value was not suitable for cross-sample comparison. Instead, the saturated-state geometric mean of T₂ was more suitable for characterizing differences in sample mobility. Based on the analysis of well-interval samples, a empirical sweet spot identification combination was proposed: r₅₀>0.4 μm, S_MF>40% and CMOI≥20%. Four evaluation zones (Type A-D) were further defined on the S_MF-r₅₀ cross-plot. The samples occurred only in types A, B, and D, with no Type-C samples observed. Type-B samples were characterized by large pore-throat sizes but limited mobility, reflecting the reservoir characteristics of "large pore-throat size does not guarantee high mobility".

Constrained by the sample size and sampling scope, the proposed identification criteria are only applicable to the same-lithofacies interval of the Qing-1 Member within Well PⅡ. Further sample expansion and verification are required for threshold calibration. The joint discrimination scheme combining CMOI and S_MF-r₅₀ can effectively distinguish oil-bearing reservoirs with limited mobility, and it can be applied to relative ranking and sweet spot classification of analogous siltstone interlayers within this well interval.

Deep-to-ultra-deep hydrocarbon reservoirs are characterized by long-term fluid evolution histories and complex hydrocarbon accumulation mechanisms. The Tarim Basin is a typical exploration target for deep-to-ultra-deep hydrocarbons in China. However, only a small number of boreholes have penetrated the Lower Paleozoic Sinian-Cambrian source rocks and reservoirs. At present, direct isotopic chronological constraints on the complete source-to-reservoir fluid evolution and hydrocarbon charging processes remain scarce, which restricts the in-depth understanding of hydrocarbon accumulation mechanisms at depths of 10 000 m within the Tarim Basin.

This study focused on fracture- and vug-filling vein minerals in core samples from the Cambrian Yuertusi Formation (source rock) and the Sinian Qigebulake Formation (reservoir) of Well Qitan-1 (QT1). The fluid sources and genesis of multiple generations of veins were identified. Multiple analytical approaches were integrated, including petrographic observation of fluid inclusions, micro-in-situ U-Pb dating of carbonate veins, Re-Os isotopic dating of solid bitumen, REE-Sr isotopic geochemistry, fluid inclusion microthermometry, and basin burial-thermal-history modelling, to reconstruct the dynamic hydrocarbon accumulation sequence of deep source-reservoir systems.

Two generations of fracture-filling calcite veins were developed in the Cambrian Yuertusi Formation source rocks of northern Tarim. REE and Sr-isotopic signatures indicated that both calcite veins were of hydrothermal origin, precipitated from deep Sr-rich fluid systems. Correspondingly, two generations of pore-filling dolomite veins occurred in the Sinian Qigebulake Formation reservoir. Geochemical proxies indicated that these dolomite veins precipitated from formation diagenetic fluids, and the Sr-isotopic compositions of the second-generation dolomite veins recorded fluid mixing with contemporaneous Cambrian seawater. The first-generation calcite vein in the source rocks ((466±5) Ma) and the first-generation dolomite vein in the reservoirs ((460±10) Ma) crystallized during the Middle Ordovician. These early-formed veins were later modified by hydrocarbon charging and trapped oil inclusions, which recorded the primary oil-charge event during the Permian (Hercynian), indicating good source-reservoir matching. Burial-thermal modelling revealed that the Yuertusi Formation source rocks reached the oil-generation threshold in the Carboniferous. Both the second-generation calcite vein in the source rocks ((263±69) Ma) and second-generation dolomite vein in the reservoirs ((55±15) Ma) contained abundant oil inclusions. Constrained by burial-history simulation, the large-scale secondary hydrocarbon charging event took place in the Miocene, showing good spatiotemporal coupling between hydrocarbon generation and accumulation. A large amount of solid residual bitumen occurring after the second-generation dolomite veins in the reservoirs yielded a Re-Os age of (30±14) Ma, documenting the destruction and readjustment of Sinian paleo-hydrocarbon reservoirs in the early Oligocene driven by Himalayan tectonic uplift.

This study conducts systematic isotopic-chronology and fluid evolution analyses for deep source-reservoir intervals in the Tarim Basin. It provides direct experimental evidence for reconstructing the whole process of hydrocarbon generation, expulsion, accumulation, and preservation within the Lower Paleozoic Sinian-Cambrian source-reservoir system.

The Chang 8 Member of the Yanchang Formation in the Ordos Basin is a typical ultra-low-porosity and ultra-low-permeability tight sandstone reservoir characterized by severe microscopic pore-throat heterogeneity and complex distribution patterns of residual oil after water flooding. Existing studies utilizing digital core simulation mostly focus on the independent characterization of pore structure or single displacement condition, and few quantitative investigations have systematically revealed the synergistic effects of intrinsic pore-throat properties and displacement conditions. A comprehensive quantitative understanding of seepage and displacement evolution mechanism under the joint control of microscopic pore-throat structure and displacement parameters can provide a microscopic theoretical basis for efficient water-flooding development of tight oil in Fuxian area, Ordos Basin.

In this study, the tight sandstone reservoir in the Chang 8 Member of the Yanchang Formation in Fuxian area, Ordos Basin was selected as the study object. Two representative tight sandstone core samples (labeled YP1 and YP2) were collected and scanned by micro-CT scan with a spatial resolution of 1 μm. A series of image preprocessing procedures, including grayscale enhancement, non-local mean filtering, and threshold binarization, were performed to eliminate instrument noise, and the maximal ball algorithm was adopted to reconstruct 3D pore-throat network. Representative elementary volume (REV) analysis was carried out to confirm the stable core calculation unit with a side length of 350 μm. Subsequently, grid coarsening and geometric defect repair were conducted to reduce computational cost. A microscopic seepage numerical model coupling Navier-Stokes (N-S) equations with the Cahn–Hilliard (C-H) phase-field method was established to simulate single-phase water seepage and transient oil-water two-phase displacement separately. Two sets of controlled variables—multiple gradient driving pressures (0.25-50 MPa) and eight groups of oil-water viscosity ratios (0.06-3.00)—were designed to quantitatively compare the seepage and displacement response patterns of YP1 and YP2 with distinct pore-throat configurations.

Under single-phase seepage conditions, fluid velocity and differential pressure were concentrated within narrow throats to form high stress concentration zones, while fluid velocity and pressure fluctuated slightly inside large pore cavities. Sample YP1 was characterized by large pores, small throats, and strong microscopic heterogeneity, and its displacement evolution curves showed segmented steep changes. To achieve the same residual oil saturation as YP2, YP1 required remarkably higher driving pressure. For both samples, the calculated absolute permeability declined continuously as driving pressure increased, while absolute permeability remained stable in the low-pressure range. Under oil-water two-phase seepage conditions, stress concentration generated by small throats and intricate pore-throat networks easily caused fluid blockage and displacement stagnation. Throat radius and pore-throat heterogeneity directly controlled the advancing speed and sweep range of water flooding, while the volume fraction of connected pores determined the scale of effectively mobilizable pore space and ultimately controlled final displacement efficiency. Isolated micropores were prone to forming permanent residual oil retention zones. Quantitative simulation results showed that, for cores with a high proportion of throats smaller than 8 μm, increasing driving pressure could greatly improve displacement efficiency, whereas cores dominated by large throats showed limited efficiency improvement after increasing pressure. For samples with abundant pores larger than 18 μm and well-developed preferential seepage channels, reducing the oil-water viscosity ratio could expand the sweep range and improve displacement performance. In contrast, reservoirs dominated by micropores showed higher seepage resistance and lower displacement efficiency when the viscosity ratio decreased. The simulated absolute permeability values (0.32×10⁻³-0.75×10⁻³ μm²) were consistent with laboratory core measurements, which verified the reliability of the established simulation model.

This study distinguishes the respective controlling effects of pores and throats on microscopic oil-water displacement and clarifies their coupling response patterns under different water-flooding parameters. The quantitative relationships among pore-throat configuration, driving pressure, and viscosity ratio can provide a basis for optimizing water-flooding parameters and predicting residual oil for Chang 8 tight oil reservoirs in Fuxian area, Ordos Basin. Unlike previous single-factor simulation studies, this study establishes a complete quantitative evaluation framework considering the joint influence of intrinsic pore-throat properties and artificial displacement conditions, which offers new microscopic insights for the efficient exploitation of analogous tight sandstone reservoirs worldwide.

CO₂-enhanced oil recovery (CO₂-EOR) serves as a vital technology that simultaneously realizes crude oil production increase and geological carbon sequestration. However, existing research lacks systematic comparative investigations on wettability evolution under coupled temperature-pressure conditions and different gas phases (CO₂, N₂) in deep high-temperature high-pressure sandstone reservoirs, and the interfacial evolution mechanism of the CO₂-water-oil-quartz multiphase system remains unclear, which restricts the optimization of field injection-production parameters for CO₂ flooding. This study carries out microscopic experimental research to clarify the coupling mechanism of contact angle and wettability controlled by temperature, pressure, and gas composition.

A high-temperature high-pressure visual quartz capillary experimental setup was constructed, with n-hexadecane adopted to simulate formation crude oil. Continuous temperature gradients ranging from 25℃ to 200℃ and pressure gradients from 5 MPa to 30 MPa were set up for testing. Contact angles of two-phase CO₂-water, water-hexadecane systems, and three-phase systems filled with CO₂ or N₂ were measured separately. Combined with the variation patterns of interfacial tension and fluid viscosity, the internal control mechanism of gas composition on reservoir wettability was analyzed quantitatively.

The test results indicated that the CO₂-water contact angle increased slightly with the growth of temperature and pressure, while the water-hexadecane contact angle decreased remarkably with temperature rise and showed weak response to pressure variation. Gas injection could adjust oil-water interfacial tension to alter reservoir wettability, and CO₂ exerted a stronger viscosity-reducing effect on crude oil than N₂. The critical temperature range for complete wettability transition of quartz mineral was from 120 ℃ to 150 ℃. Above this threshold, the quartz surface completely transformed from oil-wet to water-wet. Heating could effectively reduce crude viscosity and enhance the seepage capacity of oil phase, while pressure had a limited effect on the physical properties of oil and water fluids.

Higher temperature facilitates the hydrophilic transformation of quartz matrix in sandstone reservoirs. Reasonable regulation of flooding temperature can greatly improve the recovery efficiency of CO₂-EOR. This study innovatively conducts comparative micro-capillary experiments with CO₂ and N₂ as contrasting gas phases, and reveals the coupled controlling effect of temperature, pressure, and gas species on multiphase contact angle. The experimental conclusions can provide a reliable basis for optimizing temperature-pressure operation schemes for CO₂ flooding in deep quartz sandstone oil reservoirs.

Rock breakthrough pressure and fracture pressure are two fundamental petrophysical and geomechanical parameters that govern hydrocarbon migration and accumulation, and they also provide important references for evaluating the sealing integrity and safety of caprocks in CO₂ geological storage. At present, most existing studies focus on either breakthrough pressure or fracture pressure separately through laboratory experiments and reservoir evaluation. Few publications systematically compare the discrepancies between these two parameters across diverse lithologies or investigate the underlying mechanisms by which they jointly control hydrocarbon migration and accumulation.

From multidisciplinary perspectives including geology, petrophysics, and rock mechanics, this study integrated laboratory core experimental observations and real-world geological cases from sedimentary basins. It systematically analyzed the available testing techniques for breakthrough pressure and fracture pressure, and compared the distribution characteristics of the two parameters among different rock types. Additionally, this study discussed how the configuration of source-reservoir residual pressure relative to the two critical pressures modulates hydrocarbon transport and accumulation processes.

Experimental datasets showed that ① the relationship between breakthrough pressure and fracture pressure differed significantly among different lithologies. For conventional sandstone reservoirs, breakthrough pressure was far lower than fracture pressure. For tight siltstone reservoirs, no fixed magnitude relationship could be generalized between the two parameters. Controlled by nanoscale pore-throat systems and mechanical anisotropy, mudstone and shale commonly exhibited higher breakthrough pressure than fracture pressure. For instance, measured breakthrough pressure ranges of mudstone and shale were 1.22-30.72, 12.00-68.00 MPa, respectively, while their corresponding fracture pressures fell within 9.83-13.27, 24.63-39.54 MPa. ② The relative magnitudes of source-reservoir residual pressure, breakthrough pressure, and fracture pressure determined hydrocarbon migration patterns. Hydrocarbon migration could hardly occur when residual pressure was below both thresholds. Hydrocarbons migrated through steady pore-dominated seepage if residual pressure was between breakthrough pressure and fracture pressure. Once residual pressure exceeded fracture pressure but remained below breakthrough pressure, hydrocarbons migrated rapidly along newly generated hydraulic fractures. When residual pressure surpassed both values, pores and fractures jointly constituted the hydrocarbon migration pathways. ③ Basin-scale hydrocarbon migration and accumulation could be divided into two modes: steady-state continuous migration and non-steady-state episodic migration. Under the steady-state mode, hydrocarbons underwent persistent percolation via interconnected pore-fracture networks, which was prevalent in conventional reservoirs and the early-charging stage of tight reservoirs. By contrast, episodic migration occurred within overpressured systems, where hydrocarbons were rapidly charged and accumulated through periodically activated hydraulic fractures.

This study provides theoretical support for understanding hydrocarbon migration patterns and accumulation-enrichment mechanisms in conventional and unconventional reservoirs and for evaluating caprock risks in CO₂ geological storage.

In-situ stress is a fundamental geological parameter dominating the design and construction safety of deep mountain tunnels. With the rapid expansion of transportation infrastructure in southwest China, numerous extra-long and deeply buried tunnels inevitably pass through complex tectonic zones with intense horizontal tectonic compression, where severely high in-situ stress easily triggers large deformation of surrounding rock and rockburst hazards. The Qiaojia Tunnel on the Ludian-Qiaojia Expressway, located on the northeastern margin of the Wumeng Mountain tectonic belt, is a separated double-line extra-long tunnel with a maximum burial depth exceeding 1 600 m. Restricted by field construction conditions, only limited borehole measuring points can be arranged for hydraulic fracturing tests, which fail to reflect the overall spatial distribution of the three-dimensional in-situ stress field across the entire tunnel site. To fully characterize the initial in-situ stress distribution and develop a reliable inversion framework applicable to ultra-deep tunnels under strong tectonic extrusion, this paper carries out systematic field measurement and numerical comparative research.

First, six sets of hydraulic fracturing in-situ stress data ranging in depths from 346-637 m were collected from borehole K67+910 in the tunnel exit section and were used to analyze the variation patterns of the three principal stresses with burial depth. The finite element software MIDAS GTS NX was utilized to establish a three-dimensional geomechanical model with a plane size of 2 400 m × 2 400 m. Multiple sets of mechanical parameters for dolomite, limestone, and mudstone are assigned to the model according to official geological investigation documents. Four types of boundary schemes, namely displacement constraint boundary, stress loading boundary, hybrid boundary, and the boundary scheme derived from initial strain energy theory, were separately implemented for comparative inversion tests. The strain-energy-based scheme was selected as the optimal boundary condition due to its superior fitting performance with the measured stress data. After applying the optimal boundary condition to the full-domain model, the inversion results at each measuring point were extracted and compared with field test values, with ±20% defined as the allowable relative error threshold for evaluating inversion accuracy.

Field test results demonstrated that the magnitudes of the three principal stresses increased approximately linearly with burial depth. The magnitude sequence of in-situ stress ranked as follows: maximum horizontal principal stress (S_H) > vertical overburden stress (S_V) > minimum horizontal principal stress (S_h). The dominant azimuth of the measured S_H was NW32°, proving that the regional stress field was dominated by horizontal tectonic stress. The numerical inversion results indicated that the simulated S_H azimuth was concentrated between NW30° and NW35°, which achieved high consistency with field measurement. The relative errors at most measuring points were controlled within 15%, and all errors fell within the permissible range. In addition, the simulated growth gradient of principal stress with depth was consistent with field measurements, verifying the stability and reliability of the strain-energy-based inversion method.

This study systematically clarifies the applicable scope and inherent limitations of four commonly used boundary schemes for in-situ stress inversion. Compared with traditional single-boundary inversion methods, the proposed strain-energy-based approach can simultaneously reconstruct both gravity-induced stress and regional tectonic stress fields without obvious distortion of stress contour distribution. The distribution patterns and quantitative stress data obtained in this research can provide solid theoretical support and practical engineering references for the in-situ stress inversion of analogous ultra-deep extra-long tunnels, and can guide the optimization of surrounding rock support parameters and the early warning of high-stress geological disasters such as rockbursts and large deformation caused by extrusion.

In recent years, extreme rainfall-induced landslide disasters have occurred frequently. Revealing the genetic context of rainfall-triggered landslides and determining rainfall thresholds are critical components of meteorological risk early warning for geological hazards. However, existing regional studies lack quantitative analysis of landslide lag response and comparative evaluation of multiple rainfall threshold models, which restricts the precision of local early warning systems.

This study takes the July 8, 2024, extreme rainfall-induced landslide event in northeastern Chongqing as an example. Using geographic information systems and statistical analysis methods, we conducted quantitative analyses of the spatial coupling relationship between landslide distribution and geo-environmental factors, and the lagging response characteristics of landslides to sequential rainfall processes, based on 71 valid landslide samples after eliminating records with missing rainfall data. I-D, E-D, and E-I rainfall threshold models were established, and the accuracy of these three thresholds was compared using a confusion matrix.

Spatially, landslides were significantly concentrated on north-facing slopes with elevations below 1000 m and gradients from 10° to 30°, predominantly occurring in soft rock strata (rock groups Ⅱ1 and Ⅱ2) and from 500 to 800 m of folds and river networks. Temporally, landslide occurrence was positively correlated with rainfall processes, showing a clear lag effect, peaking on the seventh day of continuous heavy rainfall. The I-D model was found to be more suitable as a rainfall threshold line for the study area compared to the E-D and E-I models. Quantitative evaluation via confusion matrix shows that the I-D model under 50% landslide occurrence probability achieves an accuracy of 0.56 and a false alarm rate of 0.34, presenting the optimal overall performance among the three models.

Landslide disasters in northeastern Chongqing result from the coupling of specific geological conditions and extreme rainfall processes. The established I-D threshold can provide a scientific basis for meteorological risk early warning of regional geological hazards. However, future efforts should focus on increasing sample size and data precision to enhance model accuracy and short-term nowcasting capabilities.

Saline soil widely distributed in Northwest China suffers from unfavorable geotechnical behaviors including collapsibility, salt heaving, and corrosivity, which can easily induce deformation and failure of subgrade and foundation infrastructures. Conventional inorganic stabilization using lime-fly ash tends to cause relatively high environmental burdens. Moreover, the mesoscopic failure mechanisms of saline soil improved by ionic soil stabilizer (ISS) combined with inorganic binders remain poorly understood. This study aims to explore cost-effective and environmentally friendly composite stabilization techniques and clarify the macro-meso mechanical response mechanisms of stabilized saline soil under shear loading. The findings are of great theoretical and practical significance for saline soil engineering in seasonally frozen regions.

In this study, subsulfate saline soil sampled from Yongdeng, Gansu Province was stabilized by the combined addition of ISS, lime, and fly ash. Four ISS contents (0%, 3%, 6%, and 9% by mass of dry soil) were designed for specimen preparation. Laboratory direct shear tests were performed under normal stresses of 50, 100, 150 kPa to analyze the macroscopic shear mechanical properties of stabilized specimens. The three-dimensional discrete element software PFC^3D was adopted to establish numerical specimens. A linear parallel-bonded contact model was utilized, and mesoscopic input parameters were calibrated against laboratory test curves to reproduce real-world shear responses. The evolution patterns of particle displacement and contact force chains during shearing were revealed at the mesoscopic scale, and the progressive breakage of cementation bonds during shearing was quantitatively monitored.

The test results demonstrated that ISS addition remarkably improved the shear resistance of lime-fly ash stabilized saline soil. All specimens exhibited typical brittle failure characteristics. The shear strength of stabilized soil rose first and then declined with increasing ISS content, reaching its peak at an ISS content of 6%. Excessive ISS addition above this threshold induced strength degradation. Cohesion also followed a similar trend of initial increase followed by decrease. Discrete element simulation results indicated that the internal contact force chains of specimens underwent reconstruction during shearing, and cementation bonds near the shear plane broke progressively as shear displacement accumulated. The evolution of the mesoscopic damage factor was divided into three successive phases: Stable development, exponential growth, and gradual stabilization.

ISS affects the strength of stabilized saline soil by altering the thickness of water films on soil particle surfaces and regulating the coverage degree of hydration products. A mesoscopic damage factor formula considering cementation bond breakage is proposed according to the ratio of broken bonds to initial total bonds. This formula can well characterize the shear-induced damage evolution of ISS-lime-fly ash stabilized saline soil. This study deepens the understanding of macro-meso coupling mechanisms for composite-stabilized saline soil and offers reference data for the practical application of ISS-inorganic combined stabilization in saline soil sites.

The Yahuokou landslide in Zhouqu County, Gansu Province, is a typical large creeping landslide distributed along the Bailong River fault zone. Heavy rainfall in 2019 triggered large-scale sliding, causing major economic losses. Similar carbonaceous slate landslides are widely developed in this area. However, previous studies have mainly focused on macroscopic surface deformation monitoring of local landslides, lacking quantitative ring shear tests on strength degradation of slip zone soil under different water contents. In addition, the control mechanism of through-going tensile fractures on rainfall infiltration and slip zone softening remains unclear, and the zoned evolution of landslide instability has not been verified numerically. This restricts the formulation of refined landslide prevention and control schemes. This study aims to clarify the instability mechanism of Yahuokou landslide.

Integrated UAV photogrammetry and field zoned detailed geological surveys were carried out. Argillized carbonaceous slate slip zone soil sampled from the rear edge of the landslide was subjected to multiple long-distance ring shear tests, with four water contents (9%, 12%, 15%, and 18% (saturated)) and three normal stress levels (100, 200, and 400 kPa). After shearing, three-dimensional laser scanning and scanning electron microscopy (SEM) were used to quantitatively calculate the root-mean-square relative roughness Z₂ and observe the micro-morphology of shear surfaces. Based on FLAC^3D, two numerical models with and without fractures were established. Three rainfall scenarios (light, moderate, heavy) were set, and 17 monitoring points were arranged to track the time-series evolution of saturation, pore water pressure, and slope displacement.

Field investigation showed that the landslide was in a long-term creep state, with severe deformation occurring in the rear source zone. The exposed carbonaceous slate slip zone soil had a clay particle content of 26.4% and a hydrophilic clay mineral content of approximately 52%, indicating high water sensitivity. Under long-distance shearing, the soil exhibited significant strain-softening characteristics. When water content increased from 9% to saturated state, the softening ratio rises to 45.49%, and the maximum decrease in shear surface roughness Z₂ was 0.24 at a normal stress of 400 kPa. Clay minerals formed face-to-face aggregative orientation layers on shear surfaces, developing a lubricating structure. The Z₂ value was strongly positively correlated with the residual friction coefficient, with a Pearson correlation coefficient of 0.86755. Increasing water content reduced interface friction through the "polishing effect". Numerical simulation results showed that the landslide remained basically stable under natural conditions. Through-going fractures greatly shortened the saturation duration of slip zones. Slope deformation exhibited a time lag behind the rise of slip zone effective saturation. The overall deformation exhibited characteristics of initiation in the source zone followed by progressive transmission. The Yahuokou landslide showed an overall "front-pull and rear-push" failure mode. The strong water sensitivity of argillized carbonaceous slate slip zone soil was the internal controlling factor, seasonal heavy rainfall was the external triggering condition, and through-going fractures served as preferential infiltration channels to accelerate soil softening.

This study quantitatively elucidates the mesoscale mechanism of water-induced strength degradation of carbonaceous slate slip zone soil and clarifies the mechanism of fracture-controlled seepage-induced instability. Zoned differentiated prevention and control measures are recommended, with priority given to sealing slope surface tensile fractures and combining drainage engineering with anti-slide support. These findings provide a theoretical basis for disaster prevention of similar creeping landslides in the Bailong River basin.

Ground collapse triggered by leakage from fractured underground pipelines frequently occurs in red clay areas worldwide, posing severe threats to urban public safety and causing human casualties and property losses. Physical model tests investigating pipeline-leakage-induced ground collapse have mainly focused on sandy or silty soils in existing studies. Nevertheless, the collapse evolution mechanisms at different pipe flow rates under red clay geological conditions remain insufficiently understood. To fill this research gap, this study carries out targeted physical model experimental research.

Six groups of semi-structured physical model tests under different pipe flow rate conditions were performed using Guizhou red clay as the test material, with full-section pipeline fracture mode and constant overburden thickness kept unchanged throughout all groups. Multiple monitoring instruments, including high-speed cameras, soil pressure sensors, pore water pressure sensors, and laser displacement sensors, were deployed to continuously record multi-field information. The monitored items covered macroscopic soil deformation and failure patterns, wetting front migration behaviors, coupled responses of soil pressure and pore water pressure, and evolutionary characteristics of ground surface displacement.

Under identical pipeline fracture boundary conditions, the increase or decrease of the pipe flow rate exerted prominent control over the erosion and transport capacity of seepage water. As pipe flow rate decreased, soil erosion rate dropped, the migration mechanism of the wetting front transformed, and the dominant development direction of subsurface soil cavities shifted from horizontal toward vertical. Meanwhile, the critical overburden thickness required for collapse occurrence decreased accordingly. When the pipe flow rate fell below a critical threshold, soil cavities still developed inside the stratum, whereas ground collapse did not take place. Distinct soil arching behaviors were observed among different pipe flow rate groups. No significant soil arching effect occurred under high-flow-rate conditions, while well-developed soil arches formed during cavity expansion in low-flow-rate tests. Surface displacement remained extremely weak before collapse occurred, demonstrating the high concealment and suddenness of this geohazard. For those test groups where collapse finally occurred, the failure mode was characterized by overburden breaching driven by continuous seepage water. In addition, the first response location of pore water pressure sensors varied with the changing direction of wetting front migration.

This study reveals the mechanisms by which pipe flow rate controls cavity evolution, hydro-mechanical coupling responses, and final collapse modes in red clay strata. The findings can provide theoretical support for full-life-cycle safety monitoring of urban underground pipelines, ground-collapse risk early warning, and the formulation of corresponding engineering prevention and control measures.

Large-scale colluvial landslides are widely distributed in mountainous areas of western Hubei, and slope deformation evolution is closely coupled with the spatiotemporal distribution of rainfall. Existing related studies mostly focus on single rainfall intensity or short-term monitoring data, lacking systematic comparative research on different rainfall peak patterns and quantitative sensitivity differentiation of multi-geomechanical parameters. This study aims to reveal the response patterns of colluvial slope deformation and evolutionary process under four typical rainfall patterns, and quantitatively identify the sensitivity ranking of various influencing factors to landslide stability.

Taking No.1 colluvial landslide in Machi Village of Shiyan City as the research prototype, a geometrically scaled indoor physical model was established based on similarity theory. Four rainfall scenarios (early-peak, mid-peak, flat-peak and late-peak rainfall) were set up to reproduce the full deformation and failure process of the landslide. On the basis of model test data, orthogonal design combined with analysis of variance (ANOVA) was adopted to distinguish the dominant controlling factors of slope stability.

The results showed that: ① The influence of the four rainfall patterns on pore water pressure in the landslide soil was mainly reflected in the timing of the peak pore water pressure, which varied with the position of the rainfall peak. The earlier the rainfall peak occurred, the earlier the peak pore water pressure appeared. Moreover, the late-peak rainfall pattern induced a relatively larger failure zone and exhibited a stronger disaster-causing effect. ② Under different rainfall patterns, the slope model began to deform primarily from the middle section, first undergoing creep deformation, followed by step-like deformation, and eventually complete failure. ③ ANOVA based on orthogonal tests indicated that the sensitivity of factors affecting the overall stability of the No.1 landslide in Machi Village decreased in the following order: Internal friction angle (φ)> cohesion (c) > cumulative rainfall (T) > permeability coefficient (K_s) > unit weight of slip zone soil (γ) > rainfall pattern (Q). The cohesion (c) and internal friction angle (φ) of the slip zone soil are the key shear strength parameters for evaluating landslide stability. ④ For the local front edge of the landslide, the permeability coefficient (K_s) was an important factor affecting stability, whereas for the overall landslide, cumulative rainfall (T) played a more critical role.

The No.1 landslide in Machi Village is a typical rainfall-induced retrogressive landslide, and late-peak rainfall has the strongest disaster-causing effect. The integrated research framework combining physical model observation, deformation evolution analysis, and multi-factor ANOVA sensitivity quantification proposed in this study has high accuracy and practicability. The research findings can provide theoretical support for deformation monitoring and emergency response of similar colluvial landslides in the mountainous areas of western Hubei.

In strong deformation tectonic zones, the structure of multi-layer karst aquifer systems exhibits significant spatial variability, leading to complex and variable characteristics of groundwater occurrence and migration. This presents considerable challenges for predicting water inrush risks in deep-buried water diversion tunnels. The water conveyance tunnel of the under-construction Yangtze-to-Han River Water Diversion Project traverses the Jindou-Anzizhai strong deformation zone, where multiple exploration boreholes have exposed confined water with high pressure and large flow rates, resulting in an extremely high risk of water inrush during tunnel construction.

A comprehensive combination of hydrogeological drilling, borehole television imaging, hydrochemical analysis, isotopic tracing, and hydrodynamic monitoring techniques was employed to systematically identify the recharge, occurrence, and migration patterns of confined water in the multi-layer aquifer system.

The Jindou-Anzizhai composite anticline is characterized by an interbedded structure of alternating karst aquifers and aquitards, with weak karst development and fracture-dominated water-bearing media. The confined water is characterized by long runoff paths, slow circulation renewal, and the inability to receive direct and rapid recharge from modern precipitation. Groundwater migration is jointly controlled by topographic and tectonic factors, with an overall runoff direction from Southwest to Northeast, and the confined water is discharged upward through faults.

High recharge elevation, long runoff paths, and narrow discharge channels are the fundamental factors governing the formation of high-head confined water. The tunnel is prone to high-pressure water inrush risks when crossing fault fracture zones, fold cores, and the interfaces between aquifers and aquitards in strong deformation tectonic zones, and water inrush is mainly attributed to the release of static groundwater reserves. The research can provide a valuable reference for water hazard identification in deep-buried tunnels located in similar strong deformation tectonic zones.

Groundwater age contains critical information about groundwater circulation and evolutionary processes, which is an important hydrogeological parameter. A systematic bibliometric analysis of 1 804 papers published from 1975 to 2024 related to groundwater age with tracers is conducted to summarize the current research status and analyze development trends.

The results reveal that the number of publications in this field has exhibited an overall exponential upward trend, with the United States, China, and Germany contributing the most publications. Keyword co-occurrence network analysis identifies three key research themes: groundwater quantity, paleoclimate, and groundwater quality. Temporal trends show relatively balanced development across these three themes from 1990 to 2001. During the period from 2002 to 2013, driven by social demand, research placed greater emphasis on groundwater quantity for groundwater resource management. During 2014 to 2024, driven by technological breakthroughs, research focus shifted toward paleoclimate studies to address climate change challenges.

Over the past two decades, breakthroughs in atom trap trace analysis (ATTA) technology have enabled the application of long-lived radioactive noble gas isotopes, creating new opportunities for constructing continuous groundwater chronology sequences spanning 1.3 million years. Important development trends in groundwater chronology include, but are not limited to, improvement of single-tracer age correction models, multi-tracer combination approaches, and coupling with groundwater numerical models. These advancements contribute to enhancing the precision and accuracy of groundwater age dating, providing theoretical support and technical guidance for groundwater resource management and climate change adaptation.

Aniline is a highly toxic, recalcitrant, and stable semi-volatile organic compound (SVOC) of great environmental concern. It is primarily discharged via industrial manufacturing of dyes, pharmaceuticals, pesticides, and rubber additives, and agricultural practices involving aniline-based herbicide application also contribute to environmental aniline accumulation. Classified as a priority pollutant in the United States and the European Union, aniline has also been included in China's List of Toxic and Hazardous Water Pollutants (Batch 2), owing to its carcinogenic, teratogenic, and mutagenic risks to human health and ecosystems. Due to its environmental persistence, complicated multi-media migration and transformation processes and heterogeneous spatial distribution in subsurface environments, effective remediation of aniline-contaminated soil and groundwater remains a prominent bottleneck in modern environmental restoration practices.

Based on bibliometric analysis of publications retrieved from the Web of Science Core Collection spanning from 2005 to 2024, this review systematically synthesizes the sources of aniline pollution, environmental characteristics of contaminated sites, and the development status of detection and analysis technologies, and summarizes the applicable boundaries of various remediation technologies.

Building on these fundamental analyses, a tiered concentration-driven remediation framework is established to support targeted technology selection for different pollution levels. For high-concentration zones (soil aniline mass fraction >1500 mg/kg; groundwater aniline mass concentration >1000 mg/L), source reduction and groundwater plume containment should be prioritized by adopting integrated physicochemical remediation techniques including thermal desorption, chemical oxidation/reduction, soil washing, soil vapor extraction, and groundwater pump-and-treat systems to rapidly reduce contaminant loads and mitigate persistent secondary release risks. For medium-concentration zones (soil aniline mass fraction: (260, 1500] mg/kg; groundwater aniline mass concentration: (100, 1000] mg/L), hydraulic and reactive barrier facilities such as cutoff walls and permeable reactive barriers (PRBs) are recommended to block plume migration and promote natural contaminant attenuation. For low-concentration zones (soil aniline mass fraction ≤260 mg/kg; groundwater aniline mass concentration ≤100 mg/L), low-intervention strategies including microbial remediation, phytoremediation, and monitored natural attenuation (MNA) are preferred to achieve long-term risk control. Critical bottlenecks restricting real-world remediation performance are further elaborated. Key challenges include excessive project costs, unsatisfactory removal efficiency under complex geologic conditions, and prominent gaps between laboratory-scale experimental observations and field-scale engineering applications. Many bioremediation approaches show promising laboratory performance yet fail under high-aniline-load field conditions due to microbial inhibition. Finally, prospective research directions are proposed, covering the development of high-performance remediation materials, performance optimization of multi-technology coupled remediation frameworks, and construction of green-oriented sustainable remediation workflows for aniline-affected sites. This study can serve as a theoretical basis and practical technical reference for risk evaluation, technical screening, and precision remediation of aniline-contaminated sites across diverse hydrogeological settings.

As an emerging contaminant, microplastics have been widely detected in soil and groundwater. However, their migration mechanisms in heterogeneous porous media remain unclear, which limits the prediction of their environmental behaviors and risk assessment.

In this study, laboratory sand-column experiments were conducted to systematically investigate the transport characteristics of fluorescent polystyrene microplastics (200 and 800 nm) in three typical porous media: homogeneous coarse sand (particle size range: 1.18 to 1.40 mm), homogeneous fine sand (particle size range: from 0.12 to 0.13 mm), and a concentric heterogeneous structure (coarse sand in the center and fine sand at the periphery). Different conditions were set, including pH values (5, 7, 9), flow rates (0.5, 1.0 mL/min), and microplastic particle sizes (200, 800 nm), to observe their breakthrough behaviors. After injecting 3 pore volumes (PVs) of the microplastic suspension into the columns pre-equilibrated with 0.01 mol/L NaCl background solution, the columns were flushed with 5 PVs of the background solution. The effluent was collected and its fluorescence intensity was measured to construct breakthrough curves, and the columns were subsequently divided into 10 segments to quantify the retained microplastics and establish retention profiles. Batch adsorption experiments and zeta-potential measurements were also performed to characterize the surface properties of the sands and microplastics.

Heterogeneous media significantly altered the transport paths and retention distribution of microplastics. The breakthrough curves exhibited a typical bimodal pattern, reflecting the coexistence of preferential flow (in coarse sand regions) and retention zones (in fine-sand regions). The mobility of microplastics increased with increasing pH and flow rate. For example, in fine sand, the breakthrough rates were 103.17%, 87.06%, and 19.98% at pH 9, 7, and 5, respectively. In coarse sand, they increased from 69.30% to 95.87% as the flow rate increased from 0.5 to 1.0 mL/min. Meanwhile, the transport capacity of 200 nm particles was significantly greater than that of 800 nm particles, with breakthrough rates of 96.20% versus 69.30% in coarse sand and 75.61% versus 67.22% in the heterogeneous medium. The underlying mechanisms are as follows. Increased pH and decreased particle size promote transport by enhancing the electrostatic repulsion between microplastics and the medium surface and by weakening physical clogging, respectively, whereas higher flow rates increase the hydrodynamic force on microplastics, thereby reducing their adsorption capacity onto the medium surface.

This study reveals that the bimodal transport pattern of microplastics in heterogeneous media is dominated by the "fast-slow" dual-domain flow structure. It also clarifies the key regulatory roles of environmental factors such as pH, hydrodynamic conditions, and particle size. These findings provide experimental evidence for the behavior simulation and risk control of microplastic pollution in subsurface environments.

As a typical cratonic "cold basin" in China, the Tarim Basin contains abundant ultra-deep hydrocarbon resources. Its low geothermal field exerts fundamental constraints on source-rock thermal maturation, hydrocarbon phase preservation, and hydrocarbon accumulation processes. Clarifying the formation mechanisms of the basin's low geothermal field and the main controlling mechanisms of geothermal differences across tectonic zones is of great theoretical and practical significance for ultra-deep hydrocarbon exploration.

This study systematically compiles and synthesizes published borehole temperature measurements, rock thermophysical parameters, tectono-thermal evolution results, and relevant geothermal literature. By means of comparative analysis and inductive analysis, it reviews the spatial distribution characteristics of present-day geothermal fields, reconstructs regional thermal evolutionary history, and synthesizes the genetic mechanisms for the low-temperature background. Additionally, it distinguishes the main controlling factors of geothermal differences across various tectonic units and further summarizes the constraining effects of low geothermal conditions on deep hydrocarbon accumulation. The results show that the present-day geothermal field exhibits a planar distribution pattern characterized by higher values in uplift zones and lower values in depression zones, with an average geothermal gradient of 18-21 °C/km and terrestrial heat flow values of 35-45 mW/m². Vertically, geothermal gradients gradually decrease with increasing burial depth. Deep carbonate intervals have gradients of approximately 14 °C/km, distinctly lower than those of the shallow clastic sequences (approximately 22 °C/km). Since the Sinian, the basin has experienced long-term regional thermal decay, interrupted by a short-lived geothermal pulse triggered by Permian magmatic events. Thereafter, geothermal gradients gradually declined and stabilized at roughly 20 °C/km from the Mesozoic onward. The low-temperature geothermal background is controlled by the coupled effects of lithospheric thermal architecture, deep geodynamic processes, and sedimentary cover properties. The “cold-mantle and cold-crust” cratonic lithosphere is the fundamental internal cause. Long-term lithospheric cooling since the Permian and the suppression of heat transport induced by Cenozoic intracontinental compression collectively reinforced the cold thermal setting, and thick sedimentary sequences provided additional thermal-blanketing effects. Geothermal differences are controlled by distinct dominant mechanisms in different tectonic units. Geothermal signatures in the Kuqa Depression are mainly governed by Cenozoic orogenic tectonic activities, while the Tabei Uplift is dominated by basement-topography variations. The coupling of low geothermal gradient and overpressure can broaden the effective hydrocarbon-generation window for source rocks. Liquid hydrocarbons can be stably preserved at burial depths of up to 9000 m, whereas further increases in burial depth will trigger extensive thermal cracking of crude oil into gas.

This study systematically clarifies the origin and zonal differential patterns of the low geothermal field, advances the understanding of source-rock thermal maturation and hydrocarbon phase evolution within cratonic ultra-deep settings, and provides critical geothermal references for future deep-to-ultra-deep hydrocarbon exploration in the Tarim Basin.

The Jurassic Shaximiao Formation in the Sichuan Basin represents a critical tight sandstone target for both tight gas production and enhanced geothermal system (EGS) exploitation. Strong reservoir heterogeneity complicates the characterization of hydraulic fracture evolution, restricting the optimization of fracturing treatments and the joint development of tight gas and geothermal energy. This study aims to characterize multi-scale fracture behaviors and reveal the underlying rock-rupture mechanisms of tight sandstone during hydraulic stimulation, providing theoretical support for reservoir stimulation and production enhancement in heterogeneous tight reservoirs.

Large-scale true-triaxial hydraulic fracturing experiments were performed on two groups of tight sandstone specimens (samples α and β). Acoustic emission (AE) signals were recorded to capture microcrack activity at laboratory scale. Field monitoring data collected from six drilling platforms and 12 wells within the study area were analyzed. Laboratory AE responses were integrated with the spatiotemporal distributions of field multi-scale microseismic events to characterize fracture propagation under coupled geological and mechanical controls.

① The complexity of pre-existing natural fractures exerted a dominant control on rock-rupture behaviors. Under true-triaxial confining stress conditions, natural fractures acted as mechanical weak planes and produced localized stress concentrations at fracture tips. Pore pressure generated by hydraulic fracturing reduced the normal stress on fracture planes and activated natural fractures. Pore pressure was transmitted along the fracture planes and continuously accumulated at fracture tips. Once the local pressure exceeded the critical rupture threshold, the fractures propagated forward and generated detectable AE signals. ② Final fracture geometry was jointly governed by the interactions between pre-existing natural fractures and artificially induced hydraulic fracture networks. Where natural fractures were poorly developed, induced hydraulic fractures prevailed and formed complex multi-scale coupled fracture systems. The strike and distribution of subaqueous distributary-channel sand bodies determined the initial orientation of fracture propagation. Fractures preferentially extended along the orientation of sand-body deposition. Thick sand bodies with favorable reservoir properties promoted fracture initiation and propagation, yielding planar fracture patterns closely associated with sedimentary facies distributions. ③ Porosity was a vital parameter controlling fracturing responses. The Shaximiao Formation comprised typical low-porosity and low-permeability reservoirs. Its low matrix permeability increased the difficulty of triggering microseismic events. Rock anisotropy significantly influenced the initiation pressure, AE responses, and the geometric configuration of hydraulic microfracture networks. Intervals with low brittleness index produced sparse microseismic events and poor stimulation performance. By contrast, intervals with favorable logging interpretations showed dense microseismic clusters and effective reservoir stimulation. Reservoirs with higher porosity required larger volumes of fracturing fluid. Porosity, brittleness index, and in-situ stress-difference coefficient were mutually coupled and jointly controlled fracture complexity and stimulated reservoir volume (SRV).

Fracture development within the Shaximiao Formation tight sandstone is governed by a multi-factor coupling mechanism. The interactions among natural and hydraulic fractures, spatial restrictions imposed by sedimentary facies (sand-body distribution), geostress conditions, and intrinsic reservoir properties (low porosity and permeability, brittleness index, rock anisotropy) collectively generate heterogeneous multi-scale fracture networks. Fracture generation follows a mixed tension-shear rupture mode, jointly controlled by geological factors (brittle-mineral fraction, bedding-induced anisotropy) and mechanical factors (in-situ stress contrast and pore-pressure accumulation). The integration of laboratory physical simulation and field microseismic monitoring provides an effective approach for elucidating fracture evolution in highly heterogeneous tight sandstone reservoirs.

Deepwater oil and gas fields are typically characterized by sparse wells, large well spacing, and limited seismic data resolution. Deepwater gravity flow reservoirs develop multilevel architectural units, with complex spatial distribution and stacking relationships of sand bodies. Conventional modeling methods struggle to simultaneously ensure the rationality of geological patterns and the fine characterization of reservoir heterogeneity, posing challenges for geological reservoir modeling under sparse well conditions.

Taking the gas field A in the Baiyun Sag, Pearl River Mouth Basin as a case study, this study proposed a geological modeling method for deepwater gravity flow reservoirs suitable for sparse well conditions. Under joint well-seismic constraints, this method adopted the core concept of "hierarchical constraint and level-by-level nesting". In the first-level modeling, the interpreted channel complex profiles were transformed into a 3D grid through a spatial mapping mechanism, and a combined deterministic and stochastic approach was employed to construct the channel complex model. In the second-level modeling, for the channel complex model, 3D training images reflecting the morphology and stacking relationships of single channels were constructed. Additionally, a multiple-point geostatistics algorithm was utilized to finely characterize single channels and mudstones. For the lobe complex, an object-based modeling method was adopted to establish the mudstone model. Finally, sedimentary units at all levels were nested and integrated according to hierarchical priority to establish a 3D sedimentary facies model, based on which a reservoir petrophysical model was constructed.

The results demonstrated that, compared with deterministic modeling methods, the proposed method more accurately characterized the heterogeneity and spatial distribution of sand bodies in deepwater gravity flow reservoirs, significantly improving model accuracy.

The research findings provide a reliable theoretical basis and technical support for the exploration, development, and decision-making of deepwater gravity flow reservoirs.

To address the challenges of scarce real-world rockfall samples, highly variable target scales, and complex background interference in mountain road rockfall detection, existing models suffer from weak generalization ability and high missed detection rates under small-sample conditions. Consequently, they fail to meet intelligent monitoring requirements. This study aims to construct a rockfall object detection model capable of adapting to small samples, multiple scales, and complex environments, thereby improving detection accuracy and providing technical support for intelligent rockfall monitoring and early warning on mountain roads.

This study proposed a YOLOv8 rockfall detection model integrating transfer learning and multi-strategy improvements. First, a dataset containing over 3 000 rockfall images was constructed, covering various rock lithologies, different lighting conditions, and varying backgrounds. Data sources included on-site surveillance camera captures, mobile phone photography, and data augmentation operations, thereby expanding data diversity from a small-sample perspective. In terms of model design, transfer learning based on a large-scale dataset was introduced to alleviate insufficient feature learning caused by limited samples. Additionally, the coordinate attention mechanism (CA) was embedded into the feature network to enhance the model's response to the spatial positions of rockfalls while suppressing complex background interference. Meanwhile, a weighted bidirectional feature pyramid network (BiFPN) was adopted to replace the original feature fusion path, improving multi-scale feature flow and enhancing the perception of rockfalls at different scales. Finally, the EIoU loss function was employed to optimize bounding box regression, simultaneously considering center point distance, aspect ratio, and overlap area to improve localization accuracy.

Experimental results showed that compared with the baseline YOLOv8 model, the precision of the proposed model increased by 17.1%, recall by 24.7%, and mAP50 by 17.4%, with significant improvements in all three indicators. In comparison with other lightweight models, the proposed model achieved an mAP50 of 74.6%, significantly outperforming YOLOv5n (51.3%), YOLOv6n (56.7%), and YOLOv8n (57.2%). Under small-sample conditions, the missed detection rate and false alarm rate were substantially reduced, and a good balance between accuracy and efficiency was maintained. Validation on the test set demonstrated that the model exhibited good generalization ability and robustness across various complex backgrounds and different rock lithology scenarios.

Through transfer learning and multi-strategy improvements, the proposed YOLOv8 model effectively overcomes the problems of sample scarcity, scale variability, and background complexity in mountain road rockfall detection, achieving outstanding detection accuracy and robustness. It provides a feasible technical solution for the development of intelligent rockfall monitoring and early warning systems for mountain roads.

Rainfall-triggered clustered landslides frequently occur in the hilly and mountainous regions of southern China, posing severe threats to local lives and property. Such geohazards feature short-term abrupt occurrence, multi-point concurrent failures, and strong spatial aggregation. Critically, landslide sites share highly similar geo-environmental conditions with adjacent stable slopes, which brings great challenges to landslide susceptibility modeling. Conventional negative sample sampling strategies often generate oversimplified non-landslide samples from environmentally distinct stable areas. These easy negative samples make machine learning models achieve seemingly high statistical scores while failing to capture subtle differences between unstable and stable slopes within landslide-prone zones, consequently producing widespread spatial misjudgments. Taking the clustered landslide event triggered by extreme rainfall in October 2023 across the Xinyi-Luoding border area, western Guangdong Province, China as the study object, this study aims to explore how coupling different negative sample sampling strategies with machine learning models affects the performance of landslide susceptibility assessment.

Based on pre- and post-event Planet satellite images, visual image interpretation was implemented to acquire 3792 positive landslide samples. Twelve evaluation factors were selected from five dimensions: meteorology and hydrology, topography and geomorphology, geology and soil, vegetation cover, and human activities. Spearman correlation analysis was applied to detect multicollinearity, and the elevation factor was eliminated due to strong collinearity. Under a 1∶1 positive-to-negative sample ratio and a unified 100-m buffer constraint around landslide boundaries, three types of negative sample datasets were established via buffer-based random sampling, low-slope constrained sampling and K-means unsupervised clustering sampling. The whole dataset was partitioned into a 70 % training subset and a 30 % testing subset. Grid search combined with five-fold cross-validation was adopted for hyperparameter optimization of random forest (RF), extreme gradient boosting (XGBoost), and Stacking ensemble learning models (using RF and XGBoost as base learners and logistic regression as the meta-learner). Model performance was quantified using AUC-ROC, accuracy (ACC), and F1-score, and the natural-breaks method was adopted to generate multi-grade landslide susceptibility maps.

The results demonstrated that Stacking ensemble learning algorithms could achieve relatively high baseline performance, yet negative sample sampling strategies had a prominent influence on model outputs. Noticeably, ACC and F1-score values only showed slight discrepancies across different sampling schemes. Nevertheless, obvious differences were observed in AUC values and spatial pattern rationality of susceptibility maps. Benefiting from its capacity to characterize environmental heterogeneity within non-landslide domains, the K-means clustering sampling combined with the Stacking ensemble model obtained the optimal overall performance with an AUC of 0.93, accompanied by the highest ACC and F1-score. The derived high- and very-high-susceptibility zones were mainly distributed in the northwestern transition zone between low-to-middle mountains and hills, showing good agreement with the spatial pattern of actual landslide clusters. By comparison, buffer-based random sampling yielded moderate performance, while low-slope constrained sampling performed worst and tended to produce overestimated susceptibility and scattered mapping noise.

This study shows that the K-means-based unsupervised clustering sampling can sufficiently capture the environmental heterogeneity of non-landslide areas in clustered landslide areas. When integrated with the Stacking ensemble learning, it improves both statistical accuracy and spatial-geological rationality of susceptibility outputs. The proposed workflow provides a reliable methodological reference for refined susceptibility evaluation of rainfall-induced clustered landslides in hilly and mountainous regions across southern China.

The Upper Cretaceous Qingshankou Formation of the Songliao Basin represents a critical exploration target for continental shale oil in China. Nevertheless, this shale reservoir is characterized by strong reservoir heterogeneity and complex lithological assemblages. Practical exploration is further constrained by limited core test samples and scarce nuclear magnetic resonance (NMR) logging data in the study area. Conventional reservoir evaluation approaches suffer from limited prediction accuracy and poor regional adaptability. Meanwhile, purely data-driven machine learning models frequently suffer from deteriorated generalization ability when training datasets are insufficient. To resolve these bottlenecks, this study develops a multi-parameter coordinated prediction method integrating geological prior knowledge and machine learning algorithms for quantitative well-log evaluation of key shale reservoir parameters.

An integrated prediction framework was constructed by combining an improved ΔlogR method, a lightweight fully connected neural network (FCNN), and core-calibrated optimized empirical formulas to achieve high-precision coordinated inversion of total organic carbon mass fraction w (TOC), three major mineral components, and total porosity. The improved ΔlogR algorithm adopted segmented stratigraphic baseline calibration and dynamically adjustable weighting coefficient D, which mitigated the limitations of fixed-baseline linear assumptions and adapted to nonlinear logging responses within high-maturity shale intervals. The lightweight FCNN took six conventional well-logging curves as input variables: Acoustic travel time (AC), gamma ray (GR), caliper (CAL), compensated neutron log (CNL), logarithmic deep lateral resistivity (logLLD), compensated density (DEN). It established a nonlinear inversion model for felsic, clay, and carbonate mineral contents. For porosity estimation, an acoustic-density-neutron synergistic calculation formula was optimized using core-measured porosity data. It substantially enhanced the prediction accuracy for ultra-low-porosity shale reservoirs.

Blind-well validation demonstrated that the improved ΔlogR method increased the coefficient of determination R² of w (TOC) prediction from 0.613 to 0.786. The lightweight FCNN mineral prediction model yielded an average R² of 0.867 and maintained favorable generalization ability under small-sample conditions. After formula optimization, the R² of porosity prediction increased from 0.68 to 0.91.

The proposed prediction workflow balances geological rationality and physical interpretability and effectively alleviates the problem of insufficient model generalization caused by limited training samples. It remarkably improves prediction accuracy and stability for multi-parameter inversion in complex shale reservoirs and forms an integrated well-log interpretation system applicable to small-sample unconventional reservoirs. This study can provide reliable technical support for "sweet spot" identification, fracability assessment, and efficient development of shale oil reservoirs in the Songliao Basin and other similar continental basins worldwide.