Extreme drought-induced significant reduction of particulate organic carbon (POC) and its impact on apparent oxygen utilization (AOU) in the Changjiang Estuary and adjacent East China Sea.

• Geothermal reservoirs of the Western Bohemian Massif are low-enthalpy (< 100 °C). • Reservoirs occur in crystalline basement at ∼2–4 km depth. • Thermal water circulation is focused along deep permeable fault systems. • Hydraulic connectivity exerts important controls on thermal water composition. The fault-bound basement-hosted geothermal systems of the Western Bohemian Massif (Germany) remain poorly constrained in terms of reservoir temperatures, circulation depths and recharge. Here, we present some of the first systematic constraints on these parameters based on a geochemical survey of regional thermal water, non-thermal groundwater and surface water, combining isotope techniques, a multi-method geothermometric approach and mixing calculations. Stable isotope ratios (δ 18 O and δ 2 H) indicate that all studied thermal waters are derived from meteoric precipitation. The geothermal reservoirs of the region are identified as low-enthalpy “hot water type” systems with temperatures generally < 100 °C. The occurrence of local warm springs is linked to radiogenic heating in primarily granitic reservoirs at depth and migration of thermal water to the surface along long-lived permeable fault systems. Despite having interacted with similar reservoir lithologies at comparable temperatures, thermal waters exhibit pronounced hydrochemical variability in terms of composition, mineralization and extent of water-rock interaction. These differences primarily reflect structural characteristics of the individual geothermal systems, such as hydraulic openness and fault connectivity, which in turn affect recharge rates, residence times, extent of dilution with non-thermal groundwater and the supply of dissolved CO 2 . Together, these results demonstrate that hydraulic connectivity and permeable fault architecture exert a first-order control on geochemical signatures in the geothermal systems of the Western Bohemian Massif and similar basement-hosted settings.

The Western Qinling Orogen (WQO) is a major gold region in China, with gold reserves exceeding 1, 200 tons. Its complex topography, climate, and semi-humid grassland cover provide an ideal setting for evaluating deep-penetrating geochemical (DPG) methods. In this study, three DPG methods-the Fine-Grained Soil Prospecting Method (FGSPM), leaching of mobile forms of metals in overburden (MOMEO), and soil gas geochemical survey (H2S, SO2, and CO2)-were employed in the southern Zhaishang concealed gold deposit, which is covered by semi-humid grassland and various overburden materials. Results indicate that no single method reliably identifies concealed orebodies or alteration zones, yet their combination improves detection, especially for deep and thin ores. Gas geochemistry exhibits higher sensitivity. Differences in response between FGSPM and MOMEO likely stem from variations in sampling and chemical analysis. For FGSPM, fine-grained soil fractions (< 75μm and 75-125μm) exhibit similar capacities for adsorbing mobile Au, predominantly occurs in water-soluble and clay-adsorbed forms, with strong spatial heterogeneity. H2S and SO2 serve as effective gas indicators in semi-humid regions, though uncertainties affected by various factors (e.g., biological activity) require further verification. These findings offer a valuable case study for DPG applications and enhance the understanding of geochemical exploration in covered terrains.

Soil gas geochemical survey exploring for concealed sediment-hosted Au deposits in the semi-humid region: A case study from the Zhaishang Au-deposit, West Qinling, China

Fuzzy logic-based mineral potential mapping was applied to the Tisová–Klingenthal Cu–Co VMS deposit (Erzgebirge) in the Czech–German border region. The study area is characterized by heterogeneous geological and geochemical datasets derived from differing national surveys and historical mining. Using the Exploration Information System (EIS) toolkit, a knowledge-driven fuzzy logic approach integrated key spatial datasets, including copper and zinc soil and stream sediment anomalies and metabasalt lithology, relevant to Besshi-type VMS deposits. Three prospective anomalies were identified: the historic Tisová mine and two additional targets aligned along the same stratigraphic horizon. Artificial Neural Network (ANN) modelling was limited by insufficient training data, resulting in overfitting and reduced predictive reliability. Follow-up soil geochemical surveys conducted over the largest anomaly returned locally elevated copper values but did not conclusively confirm mineralisation. The results demonstrate that fuzzy logic provides a flexible and interpretable framework for mineral potential mapping in complex, data-scarce environments and highlight the need for iterative modelling and targeted exploration.

Lithium (Li) is a strategic element in the global energy transition, largely driven by its use in lithium-ion batteries for electric vehicles. This study examines the distribution of Li and associated elements (Cs, Rb, K, Be, Sn) in Minas Gerais State, southeastern Brazil, through a low-density stream sediment geochemical survey combined with robust statistical approaches. Baseline values were defined using different calculation methods, multivariate statistics, and compositional data analysis, aiming to delineate Li-rich zones and evaluate the Araçuaí Fold Belt, a key domain of the Eastern Brazilian Pegmatite Province (EBPP). Results show that lithological diversity across the seven geotectonic units strongly controls geochemical baselines. The Araçuaí Fold Belt exhibits the largest concentration range and highest correlation coefficients between Li and related elements, reflecting a common source in pegmatites and parental granites. Among the tested methods, mMAD provided more consistent baselines and thresholds than TIF. Staged factor analysis (clr and ilr transformations) characterized main geochemical associations and supported the generation of the GMPI mineralization map, a predictive layer integrated into a Mineral Prospectivity Map (MPM). The study demonstrates how stream sediment geochemistry, treated with advanced statistical methods, can effectively guide mineral exploration strategies for critical elements such as lithium.

This study incorporates background correction and dilution correction methods for delineating Au geochemical anomalies using data of catchment outlet sediments from two horizons: top outlet sediments (TOS, 0–10 cm) and bottom outlet sediments (BOS, 60–90 cm), collected across the entire Queensland region as part of the National Geochemical Survey of Australia (NGSA) project. The dataset comprises Au concentrations from catchments covering numerous lithological units, which were merged into seven dominant lithologies for analysis. Multiple regression analysis was applied to estimate background Au values owing to lithology, enabling separation of background signals from mineralisation-related anomalies. Results show that dilution-corrected Au residuals improve anomaly detection, with BOS samples outperforming TOS in spatial alignment with 26 of 30 known gold deposit clusters, compared with 14 clusters identified by raw TOS data. The BOS horizon benefits from enrichment in fine sediment fractions (<75 µm) and reduced effects of surface weathering, providing a more stable and representative geochemical signal. Elevated Au levels associated with dominant lithologies may mask true anomalies, underscoring the importance of background correction.

A multiscale convolutional autoencoder for fusing element concentrations and key geological information for geochemical anomaly recognition

Modelling of geochemical survey data via graph self-supervised learning and distance-encoding Transformer

Natural hydrogen (H2) emerges as a clean energy resource generated by different geological processes in multiple geological contexts. Current hydrogen exploration activities rely mainly on surface geochemical surveys focused on diagnostic morphological expressions. Analogous to the early stages of hydrocarbon exploration, these first-pass indicators are used to delineate prospective corridors and rank targets, thereby de-risking and guiding subsequent subsurface investigations. However, these surface expressions remain elusive due to sparse sampling and the absence of robust detection workflows. This study presents a machine learning-based remote sensing technique for pixel-wise hydrogen prospectivity mapping applied on the Atyrau region in the southeastern part of the petroliferous Pricaspian Basin, western Kazakhstan. We utilize Sentinel-2 Level-2A imagery and extract 13 spectral bands (B1-B12, B8A), 5 spectral indices (e.g., NDVI, NDWI), and 4 texture features, forming a 22-dimensional input feature space. Four classifiers, Random Forest (RF, [Formula: see text]), eXtreme Gradient Boosting (XGB, [Formula: see text], max depth = 5), Stochastic Gradient Descent (SGD, loss = log), and LightGBM (LGBM, learning rate = 0.1), were trained on 120,560,400 labeled pixels with 4,621,344 positive labels derived from high-confidence expert polygons. Performance metrics revealed strong recall for XGB (91.8%) and LGBM (93.1%), albeit with lower precision (6.8% and 6.4% respectively), reflecting class imbalance and rare-event detection challenges. The ROC AUC ranged from 0.50 (SGD) to 0.81 (XGB), with Precision-Recall AUC peaking at 0.17 for RF. Explainable AI techniques based on SHAP (SHapley Additive Explanations) were applied to quantify pixel-wise feature contributions for each classifier. SHAP attributions, overlaid on geologic cross-sections, revealed spatial correlations with known faults and surface anomalies. Final predictions, rasterized into labeled GeoTIFFs at 10m resolution, indicate locations of anomalies. These promising screening-type exploration results allow for narrowing down areas for direct surface measurements and focusing on understanding the hydrogen system and play concepts. This demonstrates that workflow provides a scalable, interpretable, and cost-effective methodology for natural hydrogen exploration and can be adapted to other data-sparse basins worldwide.

Understanding the distribution patterns and genesis of selenium-rich soils, a valuable natural resource, constitutes a fundamental prerequisite for their effective development and utilization. To investigate the geochemical distribution characteristics and influencing factors of selenium-rich soil in Maojie Town, this study conducted a land quality geochemical survey and collected rock and soil samples. Selenium concentrations were determined using atomic fluorescence spectrometry (AFS), and multivariate statistical regression analysis was employed to analyze the data. The results showed that the selenium content in the soils of Maojie Town ranged from 0.13 to 0.85 mg·kg−1, with an average concentration of 0.39 mg·kg−1. The selenium distribution was uneven, with selenium-rich soils mainly concentrated in the central part of the town. The formation of selenium enrichment in Maojie Town is attributed to the geological background of Mesoproterozoic slightly metamorphosed rocks, high-altitude forestlands with abundant organic matter and slow decomposition rates, and the fixation and adsorption effects of clay minerals and acidic soils. These findings provide a scientific basis for the development and utilization of selenium-rich soils and the socioeconomic development of the region.

Abstract This study presents a soil gas geochemical survey in the Kırcaoğlu and Reyhanlı regions of Hatay Province, southeastern Türkiye, following the 2023 Kahramanmaraş earthquake doublet. The aim was to identify concealed faults and assess seismic hazard through analysis of soil CO 2 flux, CO 2 and 222 Rn concentrations, and carbon and helium isotopic compositions. A total of 98 sites were surveyed, and graphical statistical methods were used to establish geochemical anomaly thresholds. In Kırcaoğlu, two prominent gas anomaly zones were delineated with dominant NW‐SE and NE‐SW trends, likely representing buried faults linked to the Yesemek Segment, including one beneath the Reyhanlı Dam. In Reyhanlı, an east‐west gas anomaly suggests a possible westward extension of the Reyhanlı Fault. These findings refine the region's structural framework and highlight seismic risks from buried faults. Isotopic analyses show CO 2 derives from biogenic and deep crustal reservoirs, with 4 He/ 20 Ne and 3 He/ 4 He ratios confirming up to 7.1% crustal helium and &lt;1% mantle helium. Heavier δ 13 C values and elevated crustal helium in Kırcaoğlu support deep gas migration along fault zones. Natural CO 2 emissions are estimated at 66 t/d in Kırcaoğlu and 60 t/d in Reyhanlı. Regionally, emissions from the Amik Basin (∼15,586 t/d) comprise ∼1.5% of Türkiye's daily anthropogenic CO 2 . The overlap between gas anomalies and surface ruptures and liquefaction zones from the 2023 earthquakes confirms the effectiveness of soil gas surveys for buried fault detection. These results highlight the utility of soil gas geochemistry as a non‐invasive tool for fault detection and seismic hazard assessment.

Machine learning (ML) algorithms are widely applied across various fields due to their ability to extract high-level features from large training datasets. However, their use in geochemical prospecting and mineral exploration remains limited because mineralization—a rare geological event—often results in insufficient training samples for supervised ML. Generating adequate training data is thus essential for applying supervised ML in this domain. In this study, we augmented training samples by utilizing adjacent samples centered around known mineral deposits and then employed random forest (RF) modeling to identify multivariate geochemical anomalies associated with mineralization. To evaluate the robustness of data augmentation and gain insights into the geochemical survey data, we applied interpretable ML techniques—feature importance and partial dependence plots (PDPs)—to clarify the data processing within mineral prospectivity mapping. The proposed methodology was tested in the Pangxidong Area, South China. The identified geochemical anomalies show strong spatial correlation with known mineral deposits, while feature importance rankings and PDPs validate the effectiveness of the proposed methodology. This practice enhances the applicability of supervised ML in geochemical prospecting and mineral exploration as well as the application of interpretable techniques for understanding data processing of multi-geoinformation.

The traditional methods for classifying elemental concentrations such as the cumulative frequency method, the logarithmic interval method, and the mean–standard deviation method all have the limitation of depending on a specific dataset. An objective “ruler” that can measure the elemental concentration level regardless of the amount of data (even for a single sample) and enables comparisons among different elements and regions is highly necessary. Recently, the 19-level fixed-value method was proposed as a “ruler” to measure the elemental concentrations of Sn, Li, Mo, and Ni objectively and to facilitate comparisons across elements and regions. However, the method for Au has not been proposed until now. In this paper, we propose the “ruler” for Au, which objectively divides Au concentrations into 19 levels with 18 fixed values from the detection limit to the cut-off grade with easily understood numbers. The “ruler” for Au along with those for Mo and Sn was applied to geochemical survey data at 1:200,000 and 1:50,000 scales, respectively, in the Zhongchuan area of Western Qinling, China, to classify elemental concentrations and draw geochemical maps. The results show that elemental concentrations can be measured using the “ruler” to assess the background, anomaly, and mineralization levels objectively, and the levels can be compared across different elements, regions, and even different scales. Geochemical maps show that in the study area, known gold deposits are all associated with high anomalies or mineralization levels of Au, while the Mo and Sn concentrations are predominantly at background levels. These results are consistent with the known mineral resources in this area. When superimposing geochemical maps of larger scales onto those of smaller scales, the variation in the elemental concentration levels with different survey scales indicates valuable geochemical meanings for mineral exploration.

Geochemical survey of uranium in catchment sediments and prediction of mineralization prospects along the Sino-Mongolian border region

Introduction. The relevance of the study is due to the need to study the aero-technogenic load on new territories adjacent to the production zones of the aluminum industry.Research objective. To assess the content of heavy metals, polycyclic aromatic hydrocarbons, and petroleum hydrocarbons in snow cover using the example of various districts of the city of Tayshet. Materials and methods. Snow geochemical survey, atomic absorption spectrometry; gas chromatography-mass spectrometry, ecological-geochemical and statistical methods.Results. The chemical elements determining the geochemical specificity of emissions identified in the snow samples were divided into the following classes: I – Pb, Hg, Cd; II – Cu, Cr, Fe, Ni; III – Zn, Mn. The median values of heavy metals (HM) were determined in the range of 0.03–111.0; 0.11–102.5 and 0.14–137.0 µg/L for the residential, recreational, and industrial zones, respectively. ΣPAHs ranged from 177 ng/L (T. 5 – recreational zone) to 3526.0 ng/L (T. 9 industrial zone), with a predominance of 4–6-nuclear PAHs (67.0%). The proportion of carcinogenic PAHs in the study increases from the residential (26.1%) to the conditionally industrial (48.9%) zone. The levels of petroleum hydrocarbons (PH) were determined in the concentration range of 21.0–60.0 µg/L.Limitations. One potential limitation of this study is the lack of data on the concentrations of contaminants at environmental sources.Conclusion. The levels of PAH, PH, and HM pollution in the area were low to medium. The characteristic features of the PAH composition under the influence of technogenic emissions were determined. The highest values of the sum of the PAH accumulation factors were noted in areas with multi-story buildings (T. 11 – 67.98) and in the area of aluminum production near checkpoint #7 (T. 9 – 44.51); the indicator of the total HM pollution Zc (46.29 and 65.02) near checkpoint #7 (T. 9) and treatment facilities (T. 5), respectively. The obtained results of the study can be used to predict the technogenic impact on the environment.Compliance with ethical standards. The study does not require submission of the Biomedical ethics committee or other documents.Contribution: Zhurba O.M. – concept and design of the study, search for literature sources, data processing, field work, writing the text; Rukavishnikov V.S. – scientific consulting, relevance, discussion; Shayakhmetov S.F. – concept and design of the study, organization of the study, editing, discussion; Merinov A.V. – collection of literature data, laboratory work, statistical processing, field work, writing; Alekseenko A.N. – collection and processing of material, field work, laboratory research. All authors are responsible for the integrity of all parts of the manuscript and approval of the manuscript final version.Conflict of interest. The authors declare no conflict of interest.Funding. The work was performed within the framework of funds allocated for the implementation of the state task East-Siberian Institute of Medical and Ecological Research.Received: July 16, 2025 / Revised: September 5, 2025 / Accepted: October 15, 2025 / Published: November 14, 2025

The French Geological Survey (BRGM) as consultant entity, and the National Mining and Geology Directorate of Uruguay (DINAMIGE) as executant, developed in the 1980s, a geochemical survey in the Precambrian terrains of Uruguay. Some thirty thousand soil and stream sediment samples were collected with direct current plasma spectrometry analysis for 22 elements (Ag, As, B, Ba, Be, Cd, Co, Cr, Cu, Fe, Mn, Mo, Nb, Ni, P, Pb, Sb, Sn, V, W, Y and Zn). Each geographical sheet, including more or less 700 samples, produced a technical report in 63% of total cases. Factor Analysis (FA) techniques applied systematically on that reports, focusing on multivariate geochemical correlations and information hidden in the data structure. Therefore, in this article FA considered all data together, including the 38 geographical sheets.

Geochemical survey of thermal springs from the Atacama region, northern Chile

Abstract This study presents the first comprehensive geochemical and mineralogical investigation of Cretaceous-age S- and I-type granitoids from the Meratus Mountains, South Kalimantan, Indonesia, aimed at evaluating their potential for hard-rock lithium mineralization. Whole-rock geochemical analysis of 17 samples using ICP-MS revealed low bulk lithium concentrations (0.05–3.00 ppm). Despite these low concentrations, X-ray diffraction (XRD) analyses identified discrete lithium-bearing minerals. The granites predominantly host α-spodumene, whereas the associated granodiorites contain a more diverse assemblage comprising γ-eucryptite, α-spodumene, and petalite. Harker diagrams delineate a clear magmatic differentiation trend governed by fractional crystallization, while Principal Component Analysis (PCA) distinguishes between magmatic evolution processes and identifies a distinct elemental association for lithium, indicative of late-stage magmatic-hydrothermal enrichment. The presence of lithium minerals within rocks of low bulk lithium content highlights a critical exploration insight: significant localized mineralization may occur within parental granites and remain undetected by regional geochemical surveys alone. These findings provide the initial mineralogical evidence of lithium prospectivity in the Meratus Mountains and establish a foundation for future exploration targeting discrete, high-grade pegmatitic bodies in Indonesia.

The Xintianling deposit, a skarn-type tungsten mineralization system in the Nanling Range of South China, presents significant challenges in identifying new high-grade orebodies. This study employs an integrated approach, combining the opposing-coil transient electromagnetic (OCTEM) method with geochemical exploration, to delineate and evaluate concealed mineralization within and beyond the known mining area. High-precision geophysical surveys revealed low-resistivity anomalies along the contact zone between Jurassic granite and the Carboniferous Shidengzi Formation limestone. Integration of these anomalies with geochemical element associations (W-Sn-Fe-Bi and Cu-Mo-As) highlights signatures indicative of tungsten mineralization. The results demonstrate that skarn-type orebodies in the mining area are primarily controlled by the axial planes of N–S-striking anticlines and associated secondary folds, with thick, large orebodies preferentially forming in depressions along the granite roof. Comprehensive analysis of the geophysical and geochemical data identified 15 low-resistivity anomalies in the Shanglongshan–Huanggualing target area, of which 14 are interpreted as potential skarn-type mineralized bodies, thereby delineating three potential exploration targets. This integrated methodology establishes a robust scientific foundation for deep and peripheral prospecting in the mining area and provides methodological guidance for exploring similar skarn-type tungsten deposits.