Search Results (337)

Geological CO₂ storage is essential to pathways to carbon neutrality, but its deployment depends on trustworthy estimates of storage capacity, injectivity, trapping, reactive evolution, and containment. Digital rock physics can provide pore-scale inputs to these estimates from X-ray and electron microscopy images, but translation to formation-scale performance requires additional geological, fluid, and operational information. Each image-derived result depends on image segmentation, which converts grayscale data into pore, mineral, fracture, and fluid phases. This review evaluates classical methods, machine learning, deep learning, transformers, and foundation models according to whether they support reliable storage decisions rather than image overlap scores alone. Evidence is synthesized from imaging of dry rocks, CO₂–brine experiments, multiscale studies of carbonates and shales, and analyses of fractured rocks. We introduce a framework with seven dimensions: class accuracy, boundary fidelity, topology, morphology, calibrated uncertainty, sensitivity of physical properties, and consequences for engineering decisions. The evidence shows that visually similar segmentations can yield substantially different predictions of permeability, connected porosity, residual trapping, reactive surface area, and leakage paths when errors occur at critical pore throats, fluid interfaces, or fractures. We therefore recommend selecting methods according to storage task and lithology, validating them on independent samples, propagating ensembles of plausible segmentations, using metrics that account for topology, and comparing predictions with laboratory measurements. The central message is simple: segmentation should be treated as both a measurement process and a form of risk control. Segmentation with auditable and quantified uncertainty has the potential to improve the inputs to site screening, injection design, and monitoring. These project-level benefits are proposed consequences requiring upscaling and project-specific validation, not outcomes demonstrated by this review. Full article

Seasonal rainforests in the southern Yucatán Peninsula invaded by bracken fern (Pteridium aquilinum) are increasingly exposed to recurrent wildfires that may hinder native tree regeneration. Understanding fire severity dynamics and post-fire vegetation recovery is important for improving fire ecology knowledge and informing restoration strategies. This study evaluated wildfire severity and vegetation recovery in communal lands of the Yucatán Peninsula affected by P. aquilinum. Moderate-Resolution Imaging Spectroradiometer (MODIS) active-fire data were used to reconstruct fire history over a 22-year period, while Sentinel-2 imagery was employed to delineate burned areas following the 2020 and 2024 wildfire events. Fire severity was assessed using the differenced Normalized Burn Ratio (dNBR) with locally calibrated United States Geological Survey (USGS) thresholds, while supervised classification was applied to evaluate vegetation recovery one year after the 2020 wildfire. The study area exhibited a recurrent fire regime with an average fire-return interval of ~5 years and consistent fire-severity dynamics across both wildfire events. Areas dominated by P. aquilinum consistently experienced high fire severity and limited vegetation recovery, whereas areas with greater native tree coverage generally exhibited lower severity classes and more favorable recovery. These findings highlight the importance of controlling bracken fern dominance and promoting native tree recovery to enhance ecosystem resilience in tropical communal landscapes. Full article

Cross-well lithology recognition is important for reservoir characterization, but its application to newly drilled wells is constrained by inter-well variations in logging responses and limited lithological labels. This study proposes a geoscience knowledge-guided machine-learning method for lithology recognition under limited target-well labels. Geochemical indicators and conventional logging data are combined to construct geologically interpretable features, while few-shot transfer learning is used to reduce class imbalance and inter-well distribution differences. Geological knowledge is introduced as probabilistic constraints and integrated with model predictions through uncertainty-aware fusion. The method is evaluated using volcanic reservoir well-logging data from the Pearl River Mouth Basin. Three representative wells are used as the source domain and an independent newly drilled well as the target domain, with three labeled samples per lithology class used for adaptation and the remaining samples for independent testing. Repeated experiments with 10 random seeds yield an accuracy of 88.03% ± 6.09%, with improved classification performance and lower variability than the compared baseline methods. The results show that combining geological knowledge with few-shot learning can improve the robustness of cross-well lithology recognition and provide a practical approach for lithology prediction in heterogeneous volcanic reservoirs. Full article

Phosphate mining secures the world’s fertilizer supply, yet it leaves behind large volumes of mine waste rock (MWR), most of which is stockpiled as a liability rather than managed as a resource. These stockpiles occupy land, affect local water quality, and represent a secondary resource left unused, at a cost to the sustainability of the sector. The practical obstacle to reuse is rarely a shortage of applications; it is the slow, sample-by-sample expert assessment needed to decide which material qualifies for which pathway. This study tests whether that decision step can be automated with the characterization data mines already collect. Fifty-two waste rock samples from two drillholes at the Benguerir mine (Morocco) were characterized mineralogically (QEMSCAN), chemically (ICP-AES), and geotechnically, then used to train a two-stage machine learning pipeline: geological facies classification (four classes) followed by valorization pathway prediction (brick and ceramic manufacturing, aggregates and concrete, or phosphorus recovery). Each of these pathways substitutes secondary material for primary raw material extraction, which is where the sustainability gain lies. Under repeated stratified cross-validation with fold-level preprocessing, XGBoost outperformed four alternative algorithms, reaching weighted F1-scores of 0.893 ± 0.095 for facies and 0.957 ± 0.079 for valorization. However, the phosphorus-recovery class contained only two independent samples, so its class-specific metrics should be considered descriptive. SHAP analysis confirmed that the predictions rest on the evidence practitioners themselves use: lithology, mechanical strength (UCS, RQD), and carbonate-silica chemistry, consistent with geological and engineering knowledge. Feeding predicted facies into the valorization stage yielded no measurable gain in accuracy (ΔF₁ = +0.005, p = 0.143); the value of the staging lies in its alignment with established geological workflows. Presented openly as a single-site proof of concept, the framework shows that routine characterization data can support consistent, auditable, and near-instant reuse screening, and it maps the path from this demonstration to multi-site operational deployment. Full article

Various versions (local and regional, as well as stationary and non-stationary) of the method of linear integral representations in solving inverse problems of geophysics are considered. Particular attention is paid to the issues of unique solvability of variational problems that arise in the interpretation of geophysical data. This paper emphasizes the application of the approximation approach in constructing analytical models of any physical fields, and highlights a class of problems that can be considered conditionally solvable in closed form. This article proves theorems on the linear independence of function systems, a finite linear combination of which represents the solution to the inverse problem. Theorems of this kind are of fundamental importance both for investigating the well-posedness of the inverse problem formulation and for practical application, as choosing an observation network with prescribed properties will reduce the costs of conducting geological and other surveys. Full article

Groundwater is essential for drinking and irrigation to ensure a stable life and economic development, especially in semi-arid to arid rural areas such as the Lagnadiz district in central Morocco. However, its occurrence is highly variable in discontinuous media because of its uneven circulation along fractures, making it difficult to locate productive drilling sites. This study aimed to identify suitable locations for the establishment of productive wells by combining Sentinel-1 and Sentinel-2 imagery, a DEM, and geological data. Lineaments were automatically extracted and validated using Google Earth imagery, slope and hillshade maps, and field observations. Six thematic layers (lithology, drainage density, distance to faults, slope, lineament density, and lineament intersection density) were weighted using the Analytic Hierarchy Process and integrated through Weighted Linear Combination. The resulting groundwater potential map shows that high potential areas cover 38.53% of the study area, followed by moderate potential areas (34.84%) and low potential areas (26.63%). A ±10% weight sensitivity analysis showed strong map stability, with r values of 0.9949–0.9999 and class agreement rates of 93.31–99.61%. The map was independently validated using exploitation-yield data from 22 boreholes that were not involved in its construction. Using a productivity threshold of 5 m³/h, the ROC analysis produced an AUC of 0.795, with a 95% confidence interval ranging from 0.581 to 0.949, indicating an acceptable ability of the GWPI map to distinguish between the two productivity groups. Full article

Advanced multispectral and hyperspectral remote sensing supports mineral and alteration mapping, but its imagery, specialist expertise and validation requirements may remain inaccessible to artisanal and small-scale mining operators during first-pass evaluation. This study evaluates how much defensible geological reconnaissance information can be extracted from rendered Google Earth imagery without interpreting display color as mineralogical evidence. Google Earth Pro RGB exports displaying Airbus Pléiades imagery were processed using a reproducible workflow combining preprocessing, RGB–HSV transformation, color-class delineation, spatial-pattern assessment and lineament analysis. Image-derived classes were treated as color-defined surface indicators and evaluated against documented geological and structural evidence. Two Chilean IOCG-related reference cases represented contrasting geometries: Farellon displayed narrow, structurally aligned patterns, whereas El Morado displayed a broader, discontinuous corridor-scale distribution. Two underexplored Chilean targets displayed corresponding linear and patchy patterns. A documented skarn occurrence served as a contrasting mineral-system case. Kabul Koh and Ziarat Malik Karkam in Balochistan’s porphyry Cu–Au-prospective Chagai magmatic belt were used to test cross-regional and cross-mineral-system applicability. Across the cases, the workflow delineated contrasting surface-pattern geometries and their spatial relationships with interpreted structures without inferring mineral identity or deposit type from the RGB–HSV classes alone. The method therefore provides a low-cost, constraint-conditioned reconnaissance and target-prioritization procedure for arid copper-prospective terrains, supporting preliminary reassessment of existing or abandoned artisanal workings. Full article

Groundwater quality degradation in industrial zones presents a critical environmental challenge, as diverse and compositionally complex pollution sources severely constrain precise source identification and the formulation of effective remediation strategies. This study systematically investigates groundwater quality evolution and quantitatively apportions pollution sources in a representative industrial zone in southwestern China, employing an integrated framework of hydrochemical graphical analysis, dual-dimensional hierarchical cluster analysis, and Positive Matrix Factorization (PMF) receptor modeling, based on 189 groundwater samples collected across three hydroperiods (2022–2025) from 94 monitoring wells. (1) The study reveals that overall groundwater quality was unsatisfactory, with Class IV and V waters collectively accounting for 75%, 95%, and 91% across the three campaigns; primary exceedance parameters included ammonia nitrogen, total hardness, Mn, and sulfate. (2) Hydrogeochemical analysis revealed stable HCO₃-Ca type water at background monitoring points, slight contamination influence at diffusion points (occasional HCO₃·SO₄-Ca and Cl·SO₄-Ca types), and pronounced hydrochemical diversification at internal points, evolving from Ca-Cl dominance to the coexistence of Ca-Cl·SO₄, Ca-HCO₃, and other mixed types. (3) The PMF model consistently resolved five pollution sources across all campaigns: agricultural non-point source pollution, geological background, domestic wastewater, industrial emissions, and natural hydrogeochemical evolution. Anthropogenic sources (agricultural, domestic, and industrial) collectively contributed approximately 63.8% of the total contamination load, with individual average contributions of 18.6%, 26.2%, and 19.0%, respectively, indicating that groundwater contamination in the zone is influenced not only by industrial inputs but also substantially by agricultural non-point sources and domestic wastewater. This study reveals a coupled natural–anthropogenic driving mechanism and establishes a replicable integrated framework for pollution source identification and zoned precision management in comparable industrial zone settings. Full article

Landslide susceptibility assessment is essential for maintaining resilient mountain transport corridors where slope failures can disrupt mobility, freight movement, and economic activity. This study evaluated standalone and hybrid machine-learning models for the Kamchik Pass corridor of Uzbekistan, which carries the A-373 Tashkent–Osh highway. Thirteen topographic, hydrological, geological, climatic, vegetation, and land-cover factors were integrated with a landslide inventory. K-nearest neighbours (KNN), Random Forest (RF), XGBoost, and artificial neural network (ANN) models were compared to three RF-based hybrids: RF + KNN, RF + XGBoost, and RF + ANN. Performance was assessed using confusion-matrix metrics and receiver operating characteristic area under the curve (ROC–AUC), together with variable-importance and class-area analyses. All predictors were retained because variance inflation factors remained below 3.3. Slope and elevation were the most consistent predictors across the standalone models. RF + KNN achieved the best performance, with an accuracy of 0.8429, kappa of 0.6857, sensitivity of 0.8857, specificity of 0.8000, and AUC of 0.88. Its map classified 16.94 km², approximately 10.8% of the study area, as high or very high susceptibility. Compared to standalone RF, RF + KNN increased accuracy by 2.86 percentage points and AUC by 0.03, while sensitivity decreased from 0.9429 to 0.8857 and specificity increased from 0.6857 to 0.8000. Because these differences were obtained from a small point-level hold-out set without spatially independent validation, they are interpreted as descriptive rather than evidence of universal model superiority. The maps provide a first-pass susceptibility screening layer for subsequent field verification and asset-exposure analysis; they do not constitute an implemented infrastructure-risk assessment. Full article

Landslides are frequent and destructive geological disasters. Accurate landslide identification is essential for post-disaster reconstruction and preventing secondary disasters. Deep learning has shown considerable potential for recognizing landslide objects from remote-sensing images; however, existing models still suffer from insufficient detection accuracy in scenarios with complex backgrounds, blurred boundary localization, sample-class imbalance, and difficulty in balancing detection speed and segmentation accuracy. To address these issues, this study investigates landslide identification in the Great Bend of the Yarlung Zangbo River region using improved deep learning models and heterogeneous optical remote-sensing imagery. (1) By introducing the convolutional block attention module (CBAM) into YOLOv8, 82.79% precision was achieved, and the recall improved by 9.56% compared to the original model, reaching a mean average precision of 75.27% while maintaining computational efficiency, outperforming YOLOv5 and the original YOLOv8. (2) Replacing the cross-entropy loss with Focal Loss in DeepLabV3+ improved the landslide edge segmentation by dynamically adjusting the weights of difficult and easy samples. Compared with the original DeepLabV3+, the precision and recall of the DeepLabV3+-FL semantic-segmentation model were improved by 0.26% and 1.93%, respectively, with the mean pixel accuracy and mean intersection over union reaching 81.76% and 62.59%, respectively. Overall, the two improved models enhanced the accuracy of landslide identification and resistance to interference, demonstrating potential for landslide monitoring and emergency response. Full article

Landslide hazard assessment provides an important basis for regional geological disaster prevention and mitigation. Landslide hazard is commonly characterized in terms of three fundamental dimensions: spatial occurrence, temporal occurrence, and potential magnitude. Based on this conceptual framework, Guiyang, a typical mountainous city in western China, was selected as the study area, and slope units were adopted as the basic assessment units. First, eight topographic, geological, and hydrological conditioning factors were incorporated into an information-value model to evaluate the landslide susceptibility of each slope unit. The resulting min–max-normalized susceptibility index was used as the spatial susceptibility component. Second, the historical landslide inventory was combined with kernel density estimation, and a Poisson-based exceedance probability model was used to estimate the model-based exceedance probability of one or more landslides occurring in each slope unit under a 10-year scenario time window. This probability was used as the temporal component. Third, the potential landslide volume of each slope unit was predicted using machine-learning algorithms. The predicted small-, medium-, and large-volume classes were assigned ordinal magnitude weights of 1, 2, and 3, respectively, and were used as the magnitude component. Finally, the three components were integrated using a multiplicative composite index to obtain the relative landslide hazard of each slope unit. The results were as follows. The overall landslide hazard in Guiyang was dominated by medium and low hazard levels. Furthermore, 422 high-hazard slope units were identified, accounting for 15.58% of all slope units. These units were mainly concentrated in the central and northern regions. In terms of administrative divisions, Xifeng County, Kaiyang County, Xiuwen County, Qingzhen city, and Huaxi district ranked among the top five administrative regions with the most high-hazard slope units. Therefore, these regions should be considered key areas for disaster prevention and mitigation. Based on the hazard zonation, 21 selected high-hazard slopes were examined using UAV imagery and on-site inspection to document their representative geomorphological and deformation characteristics. The observations provided qualitative field support for the geological plausibility of selected high-hazard predictions. The findings of this study provide an important decision-making basis for landslide mitigation efforts by the Guiyang Municipal Government. Full article

Landslide damage in mountain forest watersheds may vary not only with local slope conditions but also with hydrological and geomorphological connectivity within watersheds. This study quantitatively analyzed differences in landslide damage characteristics and influencing factors among forest watershed types to support watershed-based landslide damage mitigation. We extracted 1809 landslide damage polygons that overlapped mountain forest watersheds in the Chungcheong region of the Republic of Korea during 2022–2024 and classified them into catchment, slope, and independent zones. Damage area, perimeter, width, length, and elongation ratio were compared among watershed types. Area, perimeter, width, and length were largest in catchment zones and smallest in independent zones. Random forest models were developed and evaluated using repeated stratified 5-fold cross-validation with 13 topographic, soil, forest, geological, and watershed morphometric factors, and SHAP analysis was applied to interpret factor contributions. Watershed area was identified as a common important factor across all watershed types. However, secondary key factors differed by type: relief, DBH class, and slope length were relatively important in catchment zones; slope gradient, DBH class, and relief in slope zones; and relief and slope gradient in independent zones. These findings indicate that landslide damage characteristics and landslide-influencing factors vary by forest watershed type and can support differentiated forest disaster management and landslide mitigation strategies. Full article

This study evaluates Cu–Pb–Zn soil-geochemical anomalies in the Örencik prospect, NW Türkiye, using robust thresholding, multivariate statistics and indicator-kriging probability mapping. The dataset comprises 1100 primary B-horizon soil samples collected on a nested grid locally densified from 50 × 100 m to 50 × 50 m and 25 × 25 m. Background–anomaly boundaries for Cu, Pb, Zn and As were defined using the median and normalised median absolute deviation (nMAD) of log10-transformed concentrations. The element-specific M + nMAD thresholds were 54.68 ppm Cu, 19.18 ppm Pb, 81.23 ppm Zn and 13.68 ppm As. At or above these thresholds, 15.64% of Cu, 16.64% of Pb and 12.18% of Zn observations were classified as anomalous; notably, 133 of the 134 anomalous Zn samples remained within the weak anomaly class. Correlation analysis and PCA distinguish a Cu–As association from a Zn–Co–Ni–V association; PC1 and PC2 explain 69.51% of the total variance, while a Pb-dominated PC3 is treated only as an exploratory secondary component. Indicator kriging delineates compact and discontinuous Cu probability zones, a more continuous Pb belt and a broad Zn pattern. Integration of the robust thresholds, multivariate relationships and spatial probability patterns indicates that Zn variability has a stronger lithological contribution, whereas spatially coherent Cu and Pb zones provide the principal exploration targets. These targets require geological verification and are not interpreted as confirmed economic mineralisation. Full article

Granite residual and slope-wash soils are important disaster-prone geological bodies in the hilly and mountainous areas of Hunan Province, China. Their engineering classification has long relied on manual visual inspection and laboratory testing, which is inefficient and subjective. In this study, an automatic classification method based on deep learning image recognition is proposed for granite residual and slope-wash soils in the mountainous areas of Hunan Province. First, a three-class primary classification scheme was established, comprising residual clay (RNC), residual sandy clay (RNSC), and residual gravelly clay (RNGC), based primarily on the gravel content of particles larger than 2 mm (RNC < 5%, RNSC 5–20%, RNGC > 20%). Second, 7678 geotechnical test records from 21 counties in Hunan Province were collected, and classification labels were assigned through a strategy combining manual verification and automatic inference using Random Forest (5-fold cross-validation macro F1 = 0.913). From approximately 10,096 original field images, 3380 pure soil image patches were retained after segmentation and screening. A dataset of 43,940 samples was then generated through two-stage preprocessing (including denoising and illumination correction) and 13-fold data augmentation. A CNN image classification model was constructed based on a ResNet18 backbone network pre-trained on ImageNet. On the independent test set (6591 images), the primary classification accuracy reached 91.46%, with a macro F1-score of 0.9128; the per-class F1-scores for RNC, RNSC, and RNGC were 0.921, 0.885, and 0.933, respectively. Grad-CAM visualization analysis demonstrated that the model’s attention was primarily focused on soil particle distribution regions rather than non-soil background areas, confirming the effective learning of mixed-grain features. The study shows that the combined application of transfer learning and 13-fold data augmentation can significantly improve classification performance under limited sample size conditions (an improvement of 15.33 percentage points compared to the baseline of 76.13%), demonstrating promising potential for engineering applications. Full article

Geological hazards pose persistent challenges to sustainable land use planning, infrastructure safety, and community resilience in alpine valley regions. Taking Tongren City on the northeastern margin of the Qinghai–Tibet Plateau as the study area, this study assessed composite susceptibility to landslides, rockfalls, and debris flows using 18,136 slope units, 189 historical hazard points, and 92 independent verification points, with the three hazard types combined into a single positive class for regional screening. Twelve conditioning factors were retained after collinearity testing, and negative-sample buffer distances of 200, 500, 1000, and 1500 m were compared before applying Bayesian-optimized random forest (Bo-RF), Bayesian-optimized CatBoost (Bo-CatBoost), and TabPFN models. Model performance was evaluated through repeated buffered spatial block cross-validation, probability calibration, spatial error analysis, independent verification, and SHAP interpretation. The 1000 m buffer produced the best negative-sample performance, with accuracy of 0.905 and an ROC-AUC of 0.972. Under spatial validation, the three models showed comparable discrimination, with ROC-AUC values of 0.9152–0.9161 and accuracy values of 0.8509–0.8549; Bo-CatBoost performed better in probability calibration, whereas TabPFN exhibited weaker spatial clustering of residuals. The TabPFN very-high-susceptibility zone covered 17.03% of the study area and contained 84.66% of historical hazard points, while the corresponding Bo-CatBoost zone captured 72.83% of independent verification points. SHAP analysis identified the mean annual rainfall, distance to water systems, distance to roads, NDVI, and lithology as the principal factors shaping the composite susceptibility pattern. Overall, the proposed framework provides spatial decision support for sustainable land use planning, resilient infrastructure management, targeted hazard investigation, and the efficient allocation of disaster prevention resources in alpine valley regions. Full article