Geochemical Exploration Research Articles

Prediction of Au-Polymetallic Deposits Based on Spatial Multi-Layer Information Fusion by Random Forest Model in the Central Kunlun Area of Xinjiang, China

In recent years, there has been a growing emphasis on combining intelligent prospecting algorithms, such as random forest, with extensive geological and mineral data for the purpose of quantitatively predicting exploration geochemistry. This approach holds significant importance for enhancing the accuracy of target delineation. The central Kunlun area in Xinjiang possesses highly favorable ore-forming geological conditions, offering excellent prospects for mineral exploration. However, the depletion of shallow deposits coupled with a decade-long gap in geological exploration have presented increasing challenges in the quest to discover substantial metal resources. Consequently, there is now a severe shortage of reserve assets in the region, prompting an urgent need for the implementation of new theories, methods, and technologies in mineral resource investigation and evaluation efforts. The researchers used geological and regional geochemical data to construct a random forest metallogenic discriminant model for predicting the mineralization of gold polymetallic minerals in the central Kunlun area of Xinjiang and delineating the metallogenic target area. Two different sampling methods were compared to quantitatively predict gold polymetallic mineral resources. The results indicate that the selected training samples offer higher prediction accuracy and reliability by fully capturing the complex information of the original data. The random forest model using select training samples has valuable applications in metallogenic prospect prediction and potential division due to its ability to consider the actual exploration cost and identify small areas with high potential and a high proportion of ore. This study significantly improves prediction accuracy, reduces exploration risk, and expands the use of machine learning algorithms in mathematical geology in the central Kunlun area of Xinjiang.

Prediction of Au-Associated Minerals in Eastern Thailand Based on Stream Sediment Geochemical Data Analysis by S-A Multifractal Model

Conducted within the scope of geochemical exploration in eastern Thailand, this study aims to detect geochemical anomalies and potential mineral deposits. The objective was to interpret intricate spatial dispersion patterns and concentration levels of deposit pathfinder elements, specifically arsenic (As), copper (Cu), and zinc (Zn), using a comprehensive array of stream sediment geochemistry data. Methodologies involved integrating multifractal properties and traditional statistics, facilitated by the GeoDAS and ArcGIS platforms as instrumental analytical tools. In total, 5376 stream sediment samples were collected and evaluated, leading to the development of an in-depth geochemical map. The results indicated distinct geological units marked by substantially elevated average values of the aforementioned elements. Identification of geochemical anomalies was achieved through the spatial distribution method and the subsequent application of the spectrum-area (S-A) multifractal model. An intriguing link was found between high As concentrations and gold deposits in the area, suggesting As as a viable pathfinder element for gold mineralization. The anomaly maps, generated from the stream sediment data, spotlighted potential zones of interest, offering valuable guidance for future mineral exploration and geological inquiries. Nonetheless, it is vital to recognize that the increased values noted in these maps may be influenced by regional geological factors, emphasizing the necessity for a diverse set of analytical methods for accurate interpretation. This study’s significance lies in its pioneering use of the S-A multifractal model in geochemical data analysis. This innovative approach has deepened our comprehension of geochemical dispersion patterns and improved the precision of mineral exploration.

Geophysical and Geochemical Exploration of the Pockmark Field in the Gulf of Patras: New Insights on Formation, Growth and Activity

The Patras Gulf Pockmark field is located in shallow waters offshore Patras City (Greece) and is considered one of the most spectacular and best-documented fluid seepage activities in the Ionian Sea. The field has been under investigation since 1996, though surveying was partially sparse and fragmentary. This paper provides a complete mapping of the field and generates new knowledge regarding the fluid escape structures, the fluid pathways, their origin and the link with seismic activity. For this, data sets were acquired utilising high-resolution marine remote sensing techniques, including multibeam echosounders, side-scan sonars, sub-bottom profilers and remotely operated vehicles, and laboratory techniques focusing on the chemical composition of the escaping fluids. The examined morphometric parameters and spatial distribution patterns of the pockmarks are directly linked to tectonic structures. Acoustic anomalies related to the presence of gas in sediments and in the water column document the activity of the field at present and in the past. Methane is the main component of the fluids and is of microbial origin. Regional and local tectonism, together with the Holocene sedimentary deposits, appear to be the main contributors to the growth of the field. The field preserves evidence that earthquake activity prompts the activation of the field.

Is the pathway length of sediments relevant to assess the background value in stream sediment geochemical exploration?

The sample catchment basin approach (SCBA) is one of the methods mainly used to assess the expected chemical composition of a stream sediment sample considering the geolithological variability of the catchment basin upstream of its collection point. In the original SCBA, the catchment surface considered for the purpose is limited to the area between two subsequent collected samples; in the true sample catchment basin approach (TSCBA), the catchment area of each collected sample is limited upstream only by the actual watershed. In both approaches, the dilution is generally assessed in relation to the extension of the single geolithological units falling within the sample catchment area used.However, it makes sense to think that when multiple geolithological units are outcropping upstream of a sampling location, the sediments originating from the ones closer to the collection point (which have traveled a shorter distance) could have a major influence on the geochemical composition of the samples. It is plausible that the degree of dilution occurring during the transport from the source to a given point could also be a function of the distance traveled by the sediments along the fluvial pattern. This also holds in the case of mineralization occurring within the catchment: the farther a mineralization is from a sampling point, the more its geochemical signal should be diluted at the sampling site.In the present study, using the TSCBA, a modification of the existing algorithm is proposed to include both the extension of the geological units and their position relative to the catchment outlet (represented by the sampling point).The territory around Feshark village (Isfahan, central Iran), where stream sediment and rock samples have been collected and analyzed, has been used as the study area. The geochemical background values (corresponding to the expected geochemical values following dilution) for each stream sediment sample were determined using the modified algorithm. The results were compared with those obtained by applying the original equation to the same dataset. The results demonstrated that the new equation could more accurately estimate background (expected) concentrations at the basin outlet, showing a better performance of the new method compared to the original one.

Mineral prospectivity mapping based on Support vector machine and Random Forest algorithm – A case study from Ashele copper–zinc deposit, Xinjiang, NW China

The utilization of machine learning (ML) techniques in conjunction with multi-source geoscience datasets for comprehensive metallogenic prognosis (MP) has emerged as a novel means of geological prospecting. Nevertheless, the representativeness of features and composition of datasets employed in constructing the ML model for MP may substantially impact the prediction model's overall performance and augment the metallogenic prediction's uncertainty. In this study, the Ashele copper-zinc deposit was chosen to conduct the case study to resolve these challenges. In order to improve the representativeness of features employed in ML modeling, the ore-controlling geological and rock geochemical exploration criteria were digitized and subjected to spatial analysis to extract information closely linked to mineralization. To investigate the impact of differences in dataset composition on the performance of predictive models, multiple labeled datasets were generated by randomly sampling from known boreholes. The predictive SVM and RF models were trained using the labeled datasets, validated using 10-fold cross-validation, and assessed using multiple metrics. The evaluation findings indicate that all ML models, subsequent to the enhancement of feature representativeness, exhibited satisfactory performance, as evidenced by a consistent predictive accuracy of approximately 0.88. Additionally, the study revealed that variations in the labeled datasets could adversely affect the predictive models' performance. However, this effect could be alleviated by selecting merit-based predictive models constructed from diverse dataset compositions. A comparative analysis of the ML predictive models revealed that the RBF kernel SVM and RF models demonstrated exceptional performance, with an AUC of 0.972 and 0.968, respectively. As a result, an integrated prediction model was developed by overlaying the RBF kernel SVM and RF models, successfully identifying 98% of the mineralized boreholes in the top 15% of the study area. The spatial distribution of gravity and magnetic anomalies in the study area corresponded well with the prospecting areas identified by the machine learning predictive model, further validating the efficacy of ML methods in MP.

Mapping the intellectual and conceptual structure of research on CoDa in the ‘Social Sciences’ scientific domain. A bibliometric overview

This article presents a bibliometric analysis and a mapping of the intellectual and conceptual structure of the research conducted on ‘Compositional Data’ (CoDa) in Social Sciences. Since the publication of John Aitchison's seminal documents on compositional data in 1982 and 1986, the scientific production of papers using compositional data methods in Social Sciences has increased exponentially in several fields. The purpose of this paper focuses on multiple categories of Social Sciences and offers a comprehensive approach through bibliometrics methods, of the application of CoDa since 1988. Using various bibliometric techniques of productivity and graphic mapping, this document determines the evolution that research in compositional data has experienced, the most relevant journals and the research topics that have been addressed in different fields of research, among other issues. The findings reveal that the publication of papers applying compositional data has increased exponentially over the last 30 years, highlighting the contributions that have been made by the fields of ‘geosciences’, ‘geography’ and ‘geology’, ‘health’ and ‘environmental science’ to this development. The Journal of Geochemical Exploration, the Journal of Archaeological Science, Archaeometry, the Journal of Archaeological Science-Reports, the International Journal of Environmental Research and Public Health are the main journals that have contributed to this development. Among other contributions of this work, we can highlight that it offers the intellectual, social, and conceptual structure of the research over a long period of time, gathering the growing literature of research progress on CoDa in Social Sciences.

Graph inference algorithms as high-performance tools for detecting geochemical anomalies related to mineralization from geochemical exploration data

Geochemical exploration data usually consists of a small number of labeled data (cells containing known deposits) and a large number of unlabeled data (cells containing none of known deposits). How to make full use of a small number of labeled data and a large number of unlabeled data to build a high-performance anomaly detection model for identifying mineralization anomalies from geochemical exploration data is a challenge faced by geochemical exploration practitioners. The graph inference algorithms including label propagation and label spreading can efficiently build high-performance semi-supervised classification models by learning from sparse labeled and abundant unlabeled data in machine learning. For this reason, the two graph inference algorithms were used to build the semi-supervised classification models to detect Fe-mineralization anomalies from stream sediment survey data of 1: 50,000 scale in the Chengde area (Hebei Province, China). The label propagation model and label spreading model were compared with the Gaussian mixture model and logistic regression model in detecting Fe-mineralization anomalies. The data processing time of the label propagation algorithm and label spreading algorithm is 64.26 s and 28.21 s, respectively; and the data processing time of the Gaussian mixture model and logistic regression algorithm is 37.79 s and 18.98 s, respectively. The area under the receiver operating characteristic curves (AUCs) of the label propagation model and label spreading model are 0.979 and 0.999, respectively; and the AUCs of the Gaussian mixture model and logistic regression model are 0.800 and 0.817, respectively. Therefore, the data modeling efficiency of the label propagation model and label spreading model is comparable to that of the Gaussian mixture model and logistic regression model in detecting Fe-mineralization anomalies. However, the performance of the label propagation model and label spreading model are significantly better than that of the Gaussian mixture model and logistic regression model. It can be concluded that the two graph inference algorithms are potentially useful high-performance tools for detecting mineralization anomalies from geochemical exploration data.

Identifying geochemical anomalies using a new method of Yang Chizhong-spatial scan statistic

Identifying anomalies from geochemical data by modeling of the background and statistical evaluation of anomalies is a major concern in geochemical exploration. This study developed a novel method (namely YangScan) for extracting geochemical anomalies by using Yang Chizhong filtering and a spatial scan statistic. The Yang Chizhong filtering, a progressive filtering method on the principle of weighted moving mean of binomial coefficient, was used to construct the background whose concavity and convexity are consistent with the original data. After removing the background from the original data, a spatial scan statistic based on multidirectional optimization was developed to identify arbitrary shaped and statistically significant anomalies from the residuals. YangScan was tested on both simulated datasets and a stream sediment geochemical dataset collected in the Northwestern Jiaodong Peninsula, Eastern China. The experimental results evaluated by geology consistency, recall, and weights of evidence contrast show that YangScan outperforms the two comparision methods (i.e., the trend surface analysis method and k-nearest neighbor anomaly detector) in identifying weak and irregularly-shaped geochemical anomalies. The distribution of geochemical anomalies linked to Au mineralization identified by YangScan is highly correlated with the trending of known Au deposits and ore-controlling structures. Therefore, YangScan is a powerful tool for identifying geochemical anomalies and can provide a useful reference for mineral exploration.

Surface geochemical anomaly pattern of medium-to-high temperature geothermal systems in South China and its application

The existence of thermal storage will correspondingly increase the temperature of surrounding strata and promote the continuous expansion, volatilization, upward migration, and loss of gas in the strata. As a result, a low-concentration gas field will be formed in the strata above geothermal reservoirs. Geothermal reservoirs could in turn heat formation water and increase the solubility of soluble inorganic salts in the surrounding rocks and the total dissolved solids (TDS) content in the formation water. Since water can strongly wet and permeate strata, the dissolved inorganic salts migrate into upper strata along with water, giving rise to the formation of a high-concentration inorganic salt field in the strata above geothermal reservoirs. A higher geothermal reservoir temperature corresponds to more significant characteristics mentioned above. Therefore, a medium-to-high temperature geothermal system has a surface geochemical anomaly pattern of high inorganic salt concentrations and low gas concentrations (also referred to as the high-salt and low-gas pattern). This pattern is applied to the surface geochemical exploration of the two geothermal fields in Guangdong Province, i.e., the Huangshadong geothermal field in Huizhou City and the Xinzhou geothermal field in Yangjiang City, revealing low-concentration gas fields above both. The application results also show that the exposed thermal spring water in both geological fields has higher concentration of dissolved inorganic salt than the surface water and nearby seawater, forming high-amplitude anomalies on the surface above geothermal reservoirs. These characteristics, as well as the measured temperature at known geothermal wells, verify the validity of the high-salt and low-gas pattern of medium-to-high temperature geothermal systems proposed in this study. Moreover, the high-salt and low-gas pattern proposed predicts three favorable medium-to-high temperature geothermal zones in the surface geochemical exploration of the Shiba Basin near the Huangshadong geothermal field.

Application and Significance of the Wavelet–Fractal Method on the Data of the Induced Polarization Method in the Graphite Deposits of Datong, China

Wavelet transformation has been widely used in geophysical and geochemical exploration, and the fractal feature of wavelet coefficients has recently stood out from many wavelet threshold methods. We introduced the wavelet–fractal method to analyze the polarizability and resistivity of graphite deposits. Due to the unique nature of graphite-bearing gneiss, characterized by high polarizability and low resistivity, we concluded that the polarizability background mode is better suited to depict the morphology of the graphite-bearing formation, with the resistivity background mode serving as complementary information for verification. Symlets5 is regarded as the optimal mother wavelet to indicate the characteristics of graphite ore by means of comparison and analysis. The polarizability anomalies showed two different linear forms: the direction of the measuring line and the strike of the ore bodies. According to the data of drill holes on the profile, we inferred that the high values of the anomaly mode can be used to delineate the target area where the graphite is enriched. Combining the application of both modes, we used the wavelet–fractal method for the quantitative prediction and effective selection of a potential area with a high grade. The approach used in this current study can be extended to the prospecting of other graphite deposits or sedimentary–metamorphic deposits containing conductive minerals, where geochemical and geophysical data are available.

Building geochemical vectors with trace element compositions of sulfides in orogenic gold mineral systems in northern Finland

Exploring and discovering new ore deposits that supply the needed metals for our current societal transition to a green-energy and technologically advanced low-carbon future is becoming increasingly difficult. To fully convert the current system however, a significant increase of mining of minerals is required, particularly since current global mineral reserves are insufficient to cover the growing demand for a fossil fuel-free infrastructure. It is therefore even more important to develop and introduce new time-and cost-efficient exploration tools, which can be utilized in the discovery of new mineral deposits and applied in the early, or through advanced stages of mineral exploration. Given that Finland is one of the major EU countries that covers the supply chain for the EU battery industry, its metallogenic provinces are extensively tested and investigated by new exploration technologies. For this purpose, a comprehensive dataset of trace element contents in single grains of pyrite, pyrrhotite and chalcopyrite was compiled on the basis of results of laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) analyses. Samples were systematically collected from drill cores from orogenic Au-only and Au-base-metal deposits in the Paleoproterozoic Svecofennian orogenic belts in northern Finland. The objective of this dataset is to serve as a reference tool in heavy indicator minerals-based exploration. By performing correlation analysis with the symmetric pivot coordinates approach for compositional data, as well as principal component analysis, it is demonstrated how all analyzed sulfide species have the capacity to discriminate between Au-only and Au-base-metal deposit types, as well as occurrences of free Au versus refractory Au. Element pairs of symmetric coordinates most suitable for differentiating Au-only systems from Au-base-metal settings are Co/Au and Au/Se in pyrite; Ni/Bi and Se/As in pyrrhotite; and Ni/Ag, Ag/In, Se/Ag, Cr/Ag, Cr/In, Cr/Zn, Zn/Se and Sn/Se in chalcopyrite. Results from principal component analysis reveal that pyrites from Au-base-metal deposits are associated with the positive loadings of Se-As-Co onto PC1 and those from Au-only deposits with negative loadings of Bi-Sb-Te-Au and Ni onto PC1 and PC2, respectively. In the case of pyrrhotite, samples from Au-Co and Au-Cu deposits are clearly distinguished between the positive loadings of Hg-Se-Co onto PC1, and the negative loadings of Sb-Au-Bi-Ni onto PC1. Positive loadings of Ag-Se and Co along PC1 are indicative of chalcopyrite from Au-Co deposits, while negative loadings of Ni and Zn onto PC1 are pointing towards an origin from Au-Cu, as well as Au-only deposits. Additionally, when testing the indicator mineral method with recovered pyrites from till samples from the Peräpohja belt, a potential relationship with bedrock pyrites is successfully demonstrated. This study aims to present the versatility of sulfide trace elements in geochemical exploration, as it could complement the mineral exploration industry with a more efficient and less expensive practice for profitable deposit discovery.

CNN2D-SENet-Based Prospecting Prediction Method: A Case Study from the Cu Deposits in the Zhunuo Mineral Concentrate Area in Tibet

Intelligent prospecting and prediction are important research foci in the field of mineral resource exploration. To solve the problem of the performance degradation of deep convolutional neural networks, enhancing the attention to target information and suppressing unnecessary feature information, this paper proposes a new prospecting prediction method based on a two-dimensional convolutional neural network (CNN2D). This method mainly uses known Cu deposits as the positive sample labels, adopts the sliding window method for data enhancement, and uses the window area as a unit to extract spatial variation features. It is important to supplement squeeze-and-excitation networks (SENets) to add an attention mechanism to the channel dimension, assign a weight value to each feature layer, and finally make prospecting predictions by matching the features of the known deposit window area and the features of the unknown window area. This method allows the neural network to focus on certain characteristic channels and realizes prospecting prediction in the case where there are few known deposits so that the deep learning method can be more effectively used for the prospecting prediction of mineralization. Based on geological data, geochemical exploration data of water system sediments, and aeromagnetic data, and via this method, this study carried out prospecting prediction of Cu deposits in the Zhunuo area of Tibet and predicted 12 favorable Cu prospecting prediction areas. Combined with previous research results and field exploration, the predicted result is consistent with the established mineralization and prospecting pattern and has good prospects for Cu deposit prospecting.

Soil gases in mineral exploration: a review and the potential for future developments

To meet the increasing demand for metals to assist in a successful and rapid energy transition, it is crucial to discover more first-class mineral deposits. With most of the world's major deposits occurring near the surface, improved methods for detection at deeper levels are required. This paper summarizes the soil gas studies that have been published in English discussing the use of soil gas as a sample medium for mineral exploration. The potential and reliability of various methods and gas species (O 2 –CO 2 , sulfur gases, polymetallic studies, gaseous hydrocarbons, radiogenic daughters (He, Rn), hydrogen and other gases) are reviewed and the challenges for the broad-scale adoption of soil gas measurement as an exploration tool are discussed. Soil gas composition has promising potential for mineral exploration, but much remains to be understood about the origins and processes affecting it. There has been a great deal of variation among the studies in sampling and analytical techniques, targeted gas(es), targeted commodities and mineralization type, climatic conditions and environmental settings. Improvement is needed in technical consistency, systematic monitoring of the environmental factors shortly before and after sampling, and the impact of microbes on the composition of the gases. In addition, further study is needed into the impact of climate, the cover composition and structure as well as the biological impact of microbes and plant roots before soil gas composition is a reliable exploration method. Supplementary material: Spatial data kmz file of gas study site locations and references presented in Figure 8 are available at https://doi.org/10.25919/5eww-8f16 Thematic collection: This article is part of the Reviews in exploration geochemistry collection available at: https://www.lyellcollection.org/topic/collections/reviews-in-exploration-geochemistry

Detection of potential hydrocarbon accumulations at Qaret El-Soda area, Western Desert, Egypt, based on airborne geophysical survey data

Airborne spectral gamma-ray survey data were processed using Th-normalization technique for oil and gas exploration in the Qaret El-Soda area, Western Desert of Egypt. This technique was applied to suppress the effects of surface lithology, which are the main factors influencing the variation of radioelement content in rocks. Normalization of K and U by thorium yielded residual potassium and residual uranium estimates. Possible occurrences of new hydrocarbon microseepages were determined by mapping low values of residual potassium and high values of residual uranium relative to potassium, which are indicated as DRAD values, which were obtained by subtracting residual potassium from residual uranium values (eUresid – Kresid). Lower residual values of K, which were associated with higher DRAD anomaly values, highlight areas of prospective hydrocarbon accumulations. The obtained results from quantitative analysis and interpretation of aeromagnetic data show sufficiently thick sediments, probably suitable for the accumulation of hydrocarbons. This means that the study area may possess a potential for hydrocarbon exploration if supported by other detailed geophysical and geochemical exploration techniques.

Single particle inductively coupled plasma mass spectrometry: A new method to detect geochemical anomalies in stream sediments

Single particle inductively coupled plasma mass spectrometry (spICP-MS) offers a new tool to explore for concealed mineral deposits. Through detection of metal nanoparticles (NPs) present in stream sediments, subtle geochemical anomalies may be identified farther from their source. Geochemical exploration with spICP-MS lowers the detection limit substantially from conventional ICP-MS for insoluble elements of interest. Analyzing only 197Au signal generated by individual NPs rather than an integrated signal, which can have a significant contribution from detector noise, resulted in a 53-fold reduction in detection limit compared to a digestion/quadrupole ICP-MS determination (21 ng∗L−1 to 0.39 ng∗L−1). Method development included analysis of dilute suspensions of manufactured 50 nm Au particles and three stream sediment samples (5–15 g sonicated in 20 mL 25 μM 11-mercaptoundeconoic acid). Au and Ag NPs were identified in the two sites downstream from a known area of mineralization, whereas none were detected in a background stream sediment site. Average particle numbers and masses of Au and Ag were highest in the sample closest to known mineralization. Additional results indicated that use of surfactant is an important contribution to the procedure. This pilot study of three sites provides a proof of concept in the analysis of spICP-MS for determination of Au and Ag in stream sediment samples, which could augment or improve current cyanide/acid dissolution/ICPMS methods for the ultra-trace concentration range.

Research on Scale Improvement of Geochemical Exploration Based on Remote Sensing Image Fusion

Both remote sensing and geochemical exploration technologies are effective tools for detecting target objects. Although information on anomalous geochemical elemental abundances differs in terms of professional attributes from remote sensing data, both are based on geological bodies or phenomena on the Earth’s surface. Therefore, exploring the use of remote sensing data with high spatial resolution to improve the accuracy of small-scale geochemical data, and fusing them to obtain large-scale geochemical layers could provide new data for geological and mineral exploration through inversion. This study provides a method of fusing remote sensing images with small-scale geochemical data based on a linear regression model that improves the resolution of geochemical elemental layers and provides reference data for mineral exploration in areas lacking large-scale geochemical data. In the Xianshuigou area of Northwest China, a fusion study was conducted using 200,000 geochemical and remote sensing data. The method provides fused large-scale regional chemical data in well-exposed areas where large-scale geochemical data are lacking and could provide potential data sources for regional mineral exploration.

Halogen characteristics and their implications for geochemical exploration in the Qujia gold deposit, Laizhou, Jiaodong

The ore-forming fluids in the Jiaodong gold deposits are characterized by low to medium salinities and Au-HS complexes are the predominant Au species in the fluid. The role of halogen elements in geochemical exploration for the gold deposits is still unclear. This study takes the Qujia gold deposit as an example to explore the issue. Scatter plots and 3D modeling were conducted based on the geochemical data of 1130 borehole samples. The halogen characteristics are governed by fluid-rock interactions, and rock types and heterogeneity. Nevertheless, the halogen characteristics in the alteration zone along the Jiaojia fault zone are still interesting and appealing. In the mineralized zone along the Jiaojia fault, the F and Cl concentrations generally reduce with a decrease in depth; F anomalies evolved from strong positive anomalies at deep depths to negative anomalies at shallow depths; positive Cl anomalies appeared in the upper part or above the mineralized zone while negative Cl anomalies appear in the middle and lower parts of the mineralized zone. The zonation pattern may be attributed to the preferential replacement of OH− by F− due to the acidic pH and the intense beresitization. The consumption of halogen elements may cause the destabilization of metal-Cl complexes and increase the pH of ore-forming fluid, and the deposition of metal sulfides further results in the destabilization of Au-HS complexes and Au deposition. Therefore, F, Cl, and F/Cl ratio can be jointly used as exploration indicators for the geochemical exploration only if rock types, and characteristics of hydrothermal fluid and alteration are examined in advance.

Self-paced ensemble for constructing an efficient robust high-performance classification model for detecting mineralization anomalies from geochemical exploration data

Given a base classifier such as the support vector classifier, the self-training algorithm can be used to build a high-performance classification model to detect mineralization anomalies from geochemical exploration data. However, the established classification model has poor robustness. To solve this problem, the self-paced ensemble algorithm was adopted to establish an efficient robust high-performance model for detecting mineralization geochemical anomalies. The self-paced ensemble algorithm can efficiently build a robust classification model based on a base classifier such as the support vector classifier, decision tree classifier, k-nearest neighbor classifier, gradient boosting classifier and multilayer perceptron. To illustrate the superiority of the self-paced ensemble algorithm, a case study for molybdenum mineralization anomaly detection was carried out in the Molidawa area, Inner Mongolia, China. The self-paced ensemble algorithm and self-training algorithm were used to build classification models based on decision tree classifier to detect molybdenum mineralization anomalies from stream sediment survey data. Each algorithm was repeated five times, each time using the same set of parameters to initialize the algorithm. Thus, five classification models were established for each algorithm. The precision-recall curve (PRC) and area under the precision-recall curve (AUPRC) were used to evaluate the performance of the classification models in geochemical exploration. The results show that compared with the PRCs of the five classification models established by the self-training algorithm, the PRCs of those established by the self-paced ensemble algorithm coincide with each other and are closer to the upper right corner of the precision-recall space. The AUPRCs of the five classification models established by the self-paced ensemble algorithm are all 0.3538, and the AUPRCs of those established by the self-training algorithm are between 0.006624 and 0.02941, much lower than the value 0.3538. In addition, the time of the self-paced ensemble algorithm for geochemical data modeling ranges from 44.84 to 47.05 s, and the time of the self-training algorithm for geochemical data modeling ranges from 51.97 to 59.46 s. Therefore, the data-modeling efficiency, robustness and classification performance of the model established by the self-paced ensemble algorithm are better than those of the model established by the self-training algorithm in detecting molybdenum mineralization anomalies. It can be concluded that the self-paced ensemble algorithm is one of effective tools to establish an efficient robust high-performance classification model for mineralization anomaly detection.

Regional geochemistry of mercury in the Sino-Mongolian border region

The Sino-Mongolian border region is a hotspot for international geoscience research and exploration. Mercury (Hg) can be used as a reference element for geochemical exploration and ore prospecting. This is because the change in Hg concentration and its spatial distribution can indirectly reflect the regional mineral distribution and regional environmental mercury pollution. Hg pollution caused by anthropogenic activities has received extensive attention because of its hazardous nature and Hg is generally regarded as a global pollutant. Although geochemical explorations have been conducted in the Sino-Mongolian border region, there have been relatively few studies with detailed and consolidated analyses of single elements such as Hg. To address this research gap we have examined the concentration and spatial distribution of Hg in this region. Influencing factors including the geological background, geomorphological landscapes, mineral distribution, development and use, and human activities, were also explored. The Hg concentration in the catchment sediment samples had a median value of 16.5 ng/g considered the geochemical background value, mean value of 18.9 ng/g, and a range of 0.5–712.5 ng/g. Low Hg concentrations are distributed in the Gobi Desert areas of Xinjiang, Inner Mongolia in China, and in the southern Gobi region of Mongolia because of the poor adsorption capacity for Hg by quartz-rich windblown sand. Areas with high Hg concentrations are distributed in the eastern and northwestern parts of the study region. Fifteen anomalous Hg enriched areas were delineated in the study region, among which 24 high Hg concentration centers (>25.4 ng/g, > Q85%) reached the scale of geochemical provinces. The high concentration of Hg in these centers is attributed to the widespread distribution of minerals containing elevated concentrations of Hg, minerals that produce substantial quantities of Hg as a by-product, and anthropogenic contributions such as mining and smelting activities. The enrichment of Hg in certain areas is an effective indicator of the occurrence of ore deposits. The Hg geochemical data obtained from the Sino-Mongolian border region is an important reference for further studies on geology, mineral exploration, and environmental protection.

Detection of geochemical anomalies related to mineralization using the Random Forest model optimized by the Competitive Mechanism and Beetle Antennae Search

This study aimed to detect the geochemical anomalies related to Fe-mineralization in the Hunjiang area in Jilin Province, China. For this purpose, a Random Forest model was optimized from two aspects of parameters' search and seeking the optimal solution by integrating the “Beetle Antennae Search” and “Competitive Mechanism”. The optimized Random Forest model was compared with the original Random Forest model in the detection of geochemical anomalies related to Fe-mineralization. The results showed that the optimized Random Forest model avoids some limitations of the original Random Forest model, with significant advantages in the detection of mineralization-related anomaly information from exploration geochemical data. It allowed automatic optimization of parameters by integrating two heuristic searching processes. Simultaneously, it focussed more on the quest for the best global optimal solution. Through the case study, it is found that the optimized Random Forest algorithm significantly improved the overall performance for the detection of geochemical anomalies related to Fe-mineralization. The geochemical anomalies detected by the optimized Random Forest model showed close spatial correlation with known iron polymetallic deposits. Therefore, using the optimized Random Forest model to detect geochemical anomalies related to mineralization is of great importance. It also provides a better supporting foundation for geochemical exploration and mineral resource exploration in the Hunjiang area.