Success Rate Curve Research Articles

Evaluation of Deep Isolation Forest (DIF) Algorithm for Mineral Prospectivity Mapping of Polymetallic Deposits

Mineral prospectivity mapping (MPM) is crucial for efficient mineral exploration, where prospective zones are identified in a cost-effective manner. This study focuses on generating prospectivity maps for hydrothermal polymetallic mineralization in the Feizabad area, in northeastern Iran, using unsupervised anomaly detection methods, i.e., isolation forest (IForest) and deep isolation forest (DIF) algorithms. As mineralization events are rare and complex, traditional approaches continue to encounter difficulties, despite advances in MPM. In this respect, unsupervised anomaly detection algorithms, which do not rely on ground truth samples, offer a suitable solution. Here, we compile geospatial datasets on the Feizabad area, which is known for its polymetallic mineralization showings. Fourteen evidence layers were created, based on the geology and mineralization characteristics of the area. Both the IForest and DIF algorithms were employed to identify areas with high mineralization potential. The DIF, which uses neural networks to handle non-linear relationships in high-dimensional data, outperformed the traditional decision tree-based IForest algorithm. The results, evaluated through a success rate curve, demonstrated that the DIF provided more accurate prospectivity maps, effectively capturing complex, non-linear relationships. This highlights the DIF algorithm’s suitability for MPM, offering significant advantages over the IForest algorithm. The present study concludes that the DIF algorithm, and similar unsupervised anomaly detection algorithms, are highly effective for MPM, making them valuable tools for both brownfield and greenfield exploration.

Deep embedded clustering: Delineating multivariate geochemical anomalies in the Feizabad region

The objective of this research was to identify geochemical anomalies associated with mineralization in the Feizabad region using deep embedded clustering (DEC). Two approaches were employed: using all geochemical elements (non-selective) and using factor scores (selective) to construct input data. Results showed that the non-selective approach was highly successful in identifying anomalies, not only confirming known mineral occurrences (KMO) but also introducing new exploration targets. Additionally, model evaluation using the success rate curve (SRC) confirmed its high predictive power. This method is particularly useful in green field. The DEC model, by combining deep neural networks (DNN) and clustering algorithms, reveals complex patterns in geochemical data. This algorithm increases clustering accuracy through automatic data clustering, gradual input of large data, and compatibility with imbalanced data. Furthermore, by simultaneously integrating feature learning and clustering, it provides higher optimization and accuracy. The anomalies identified by the DEC model are consistent with granodiorite units in the region and may indicate the potential for mineralization in the study area. Additionally, the identified targets are associated with granitoids, hydrothermal alteration zones, and intrusive rocks.

A hybrid framework for Detection of Multivariate porphyry Cu Signatures and Anomaly Enhancement: Incorporation of SFA, GMPI, and Grey Wolf Optimization

Geochemical data from stream sediment are commonly employed for regional scale mineral exploration, as they can be utilized to detect geochemical anomalies associated with mineralization processes. However, recognizing efficient multielement geochemical signatures related to target mineralization using stream sediment geochemical data can be challenging due to the intricate nature of mineralization events. To achieve this goal, this article proposes a combination of statistics, logistic functions and machine learning techniques. Initially, the staged factor analysis (SFA) approach was applied to the data to determine the geochemical footprints associated with porphyry copper mineralization in the Varzaghan area. Subsequently, the geochemical mineralization probability index (GMPI) was utilized to facilitate appropriate separation, more precise identification, and the transfer of data into a fuzzy space. By using a combination of SFA and GMPI, it was feasible to enhance the detection of geochemical halos and concealed anomalies. Ultimately, the output of these two approaches was subjected to an improved clustering technique known as GWO-K-means to detect geochemical anomalies. The findings indicated that there is a strong relationship between known mineral occurrences (KMOs) and the identified anomalous areas. The success rate curve demonstrated that the optimized model outperformed the K-means model, proposing that optimization can boost the accuracy in identifying geochemical targets. Overall, the outcomes of this study can be beneficial for proper planning of future exploration activities.

Spatial multi-criteria approaches for estimating geogenic radon hazard index

The geogenic radon hazard index (GRHI) map plays a crucial role in evaluating radon exposure risks. The construction of this map requires a comprehensive analysis of radon levels in soil gas and some critical factors, such as uranium content in bedrock, soil permeability, and geological inhomogeneities. In this context, the spatial multi-criteria decision analysis is proposed to integrate the GRHI-based criteria for identifying the high-potential radon areas. In particular, the multivariate integration involves the fuzzy gamma operator and a hybrid multi-criteria decision-making technique, namely AHP-TOPSIS, which represents a novel approach in GRHI mapping. Thus, a comparison is provided through the definition of the GRHI map of an unexplored study area, that is the Apulia region, located in Southern Italy. In order to evaluate the output maps, high radon potential areas are identified based on some available indoor radon measurement data. The success-rate curve, as a valid evaluation metric, is employed for the performance assessment and comparison of these two methods. The results demonstrate that although both generated GRHI maps are closely correlated with high-potential radon zones in Apulia, the hybrid AHP-TOPSIS method is preferable in identifying areas with elevated radon potential.

Experimental study on multi-point ignition of NH3/air by high-frequency nanosecond dielectric barrier discharge

Non-equilibrium plasma ignition technology, with its ability to expand ignition limits and increase ignition volume, is considered an important development direction of advanced ignition systems, and has significant advantages in improving ignition stability and combustion rate. This study investigates the ignition and combustion assistance effect of high-frequency nanosecond surface dielectric barrier discharge (nSDBD) on NH3/air mixture in a constant volume combustion chamber (CVCC). The study reveals the influence of high-frequency nSDBD on the growth process of the flame kernel in NH3/air mixture. As the discharge pulse number (PN) increases, the flame kernel develops from the middle position of the discharge filament to the head of the discharge filament, while there is no obvious flame at the root and middle of the discharge filament. Continuous discharge causes the flame kernel to appear concave and the center of the flame kernel to shift towards the wall of the CVCC. In the NH3/air mixture, the initial flame kernels exhibit dissipation phenomenon, resulting in a discrepancy between the number of initial flame kernels and the number of stable combustion flame kernels. As the discharge frequency or pulse number increases, and the number of stable combustion flame kernel increases. The adjustment of discharge frequency and pulse number can effectively control the flame development time (FDT) and flame rise time (FRT) of NH3/air mixture. As PN increases, both FDT and FRT are shortened. When discharge frequency is 20 kHz, the maximum shortening of FDT and FRT are 30 ms and 14 ms, respectively, about 55 % and 30 %. The effects of discharge frequency on FDT and FRT exhibit a non-monotonic variation pattern, and it is necessary to comprehensively consider various discharge parameters to achieve the best ignition effect. The ignition success rate curve of high-frequency nSDBD in NH3/air mixture is steep. The study results of this paper provide new discoveries and experimental data for advanced ignition systems.

Adopting a cell-based method to identify mineralized geological structures in support of mineral prospectivity mapping at the Kuh–Lakht epithermal gold deposit, Central Iran

Mineral prospectivity mapping (MPM) is a new tool in mineral exploration with the objective of gathering and integrating multi-source evidence to provide decision-making criteria for exploration targeting. Ideally, the MPM approach accounts for the uncertainties associated with different predictive modeling techniques. The cell-based association (CBA) is a computational workflow that addresses the prediction-related bias of the MPM protocols. In this contribution, the CBA technique was employed to prioritize grid cells according to the importance of mineralized geological features and to identify geological settings favorable for ore deposition. The Kuh-Lakht epithermal gold deposit in Central Iran was adopted as a case study to examine the relevance of the proposed workflow and define target areas for brownfield exploration. To assess the significance of the proposed technique in dealing with evidence layers rather than training locations (i.e., mineralized/non-mineralized sites), two geochemical evidence layers, single-element and multi-element signatures, were used to train the CBA technique. An appraisal analysis through prediction-area plots and success-rate curves indicated that multi-element geochemical signatures are more relevant for delineating mineralized geological structures and vectoring for concealed gold mineralization. Notably, the input of single-element geochemical data into the CBA workflow enabled the prediction of additional exploration targets, whereas multi-element geochemical data failed to achieve the same results. The results indicated that the CBA technique can efficiently cope with the uncertainty of insufficient data input and permit the delineation of exploration targets. Additionally, our results suggest that the incorporation of a ranking system of critical geological factors that facilitate the gold mineralization into conventional MPM protocols can improve the robustness of mineral exploration targeting.

Urban flash flood prediction modelling using probabilistic and statistical approaches

The development of a detailed strategy to mitigate the negative consequences of any natural calamities depends on accurately identifying sensitive zones where natural hazards frequently happen. In the present investigations, three widely utilized probabilistic approaches viz., frequency ratio (FR), statistical index (SI), and weighting factor (WF) have been utilized for prediction of flsh flood susceptibility zones in the Coochbehar urban and peri-urban area (CUPUA) (area = 26.22 km2). Ten flash flood conditioning factors have been used in this assessment based on previous literatures and experts' opinions. In the FR model, 29.40 % area is observed in the high and very high flood zones, whereas 36.27 % and 31.16 % area is identified in SI and WF model, respectively. The FR model demonstrates that five conditioning factors, viz., topographic position index (TPI), land use and land cover (LULC), normalized difference vegetation index (NDVI), distance to drainage (DtD) and rainfall were highly impacted in flash flood prediction (FFP) analysis; in SI model, LULC is the major influencing parameter, and in WF model LULC, rainfall, NDVI, and distance to road (DtR) are the effective parameters. The success rate curve of the FR, SI and WF models manifest SI model has highest training (AUC=0.903) and prediction (AUC=0.925) accuracy, and FR and WF also have very good accuracy as their AUC values are 0.899 and 0.877 (in success rate curve) and 0.900 and 0.881 (in prediction rate curve). Therefore, the application of probabilistic approaches in this active flash flood-prone region is excellently performed, and the results of this study will help hydrologists, engineers, and water management administrators to control the areas that are extremely susceptible to flash floods and reduce possible damages.

A deep autoencoder network connected to geographical random forest for spatially aware geochemical anomaly detection

Machine learning (ML) and deep learning (DL) techniques have recently shown encouraging performance in recognizing metal-vectoring geochemical anomalies within complex Earth systems. However, the generalization of these techniques to detect subtle anomalies may be precluded due to overlooking non-stationary spatial structures and intra-pattern local dependencies contained in geochemical exploration data. Motivated by this, we conceptualize in this paper an innovative algorithm connecting a DL architecture to a spatial ML processor to account for local neighborhood information and spatial non-stationarities in support of spatially aware anomaly detection. A deep autoencoder network (DAN) is trained to abstract deep feature codings (DFCs) of multi-element input data. The encoded DFCs represent the typical performance of a nonlinear Earth system, i.e., multi-element signatures of geochemical background populations developed by different geo-processes. A local version of the random forest algorithm, geographical random forest (GRF), is then connected to the input and code layers of the DAN processor to establish nonlinear and spatially aware regressions between original geochemical signals (dependent variables) and DFCs (independent variables). After contributions of the latter on the former are determined, residuals of GRF regressions are quantified and interpreted as spatially aware anomaly scores related to mineralization. The proposed algorithm (i.e., DAN‒GRF) is implemented in the R language environment and examined in a case study with stream sediment geochemical data pertaining to the Takht-e-Soleyman district, Iran. The high-scored anomalies mapped by DAN‒GRF, compared to those by the stand-alone DAN technique, indicated a stronger spatial correlation with locations of known metal occurrences, which was statistically confirmed by success-rate curves, Student's t‒statistic method, and prediction-area plots. The findings suggested that the proposed algorithm has an enhanced capability to recognize subtle multi-element geochemical anomalies and extract reliable insights into metal exploration targeting.

Vulnerability assessment of landslides along the Yunnan section of the Northern Tropic of Cancer based on fuzzy evidence weight model

AbstractFrequent geological disasters pose significant threats to both human lives and property. High-risk landslides, with features such as rapid occurrence, wide distribution, and strong destructiveness, are of particular concern. To investigate the distribution characteristics and threat range of landslides, the Yunnan section of the Northern Tropic of Cancer was selected as the study area. Based on the fuzzy evidence weight model and using a grid size of 1000 m × 1000 m as the study unit, a vulnerability assessment of landslides along the Northern Tropic of Cancer in Yunnan was conducted. This assessment involved constructing a set of 13 evaluation indicator factors from three aspects: geological environment, geographical environment, and human engineering activities. The assessment results were subsequently subjected to precision verification. The results are as follows: (1) The overall vulnerability of landslides in the Yunnan section of the Northern Tropic of Cancer is moderate, covering an area of 25,968.79 km2. This area constitutes the largest proportion, accounting for 46.30% of the entire study area and is primarily distributed in counties such as Yanshan, Funing, and Shuangjiang. The highly vulnerable area covers 15,942.27 km2, representing 28.43% of the total study area. This region is mainly distributed in counties such as Gejiu, Yuanjiang, and Mojiang. (2) The main triggering factors for landslides are twofold: first, large-scale slope instability caused by excavation in human engineering activities, leading to extensive unstable slopes affected by long-term weathering and gravitational effects. Second, soil erosion resulting from land cultivation leads to the collapse and sliding of weathered and fragmented rocks under the influence of heavy rainfall and water erosion. During routine geological disaster patrols, special attention should be given to these factors, with a focus on implementing engineering measures in areas with more fragile geological environments. (3) Precision verification of the assessment results was conducted using both receiver operating characteristic (ROC) curves and Success rate curves. The area under the curve (AUC) values for the success rate curve and ROC curve were 0.85 and 0.74, respectively. The verification results showed high accuracy and feasibility of the evaluation method.

Landslide susceptibility modeling in the Kulfo river catchment, rift valley, Ethiopia: An integrated geospatial and statistical analysis

Landslides occur when debris, rocks, or soil particles move downward. Examining the susceptibility of landslides is essential for safeguarding human well-being and assessing the consequences of landslides on the natural surroundings and ecosystems. This study utilized the frequency ratio technique to evaluate the probability of landslides happening in the Kulfo River watershed, situated in the Rift Valley area of Ethiopia. In order to ensure a comprehensive analysis, many data sources were employed, including satellite images, geological data, and historical records of landslides. This study developed a systematic approach to assess the probability of landslides by considering ten (10) influential factors: land use/land cover, slope, aspect, elevation, curvature, lithology, proximity to lineament, and normalized difference vegetation index (NDVI). The ten influential elements were prioritised based on literature review, expert knowledge, and preliminary study area analysis. The aforementioned causal elements are integrated with a comprehensive landslide inventory map. The sensitivity of the study's area to landslides was mapped using a frequency ratio (FR) model. Subsequently, it was classified as representing various degrees of vulnerability, spanning from extremely minimal to quite significant. The effectiveness of the suggested model was measured by evaluating the accuracy of the generated map of landslide susceptibility by the area under the curve (AUC) technique. Based on the most recent study results, the success rate curve has an area under the curve (AUC) of 79.6%, which indicates a highly satisfactory level of performance. Policymakers may utilize the findings of this study to make educated decisions on how to mitigate the risks of landslides in relation to land use and preparedness for disasters.

Assessing classification system for landslide susceptibility using frequency ratio, analytical hierarchical process and geospatial technology mapping in Aizawl district, NE India

Landslides are a severe geological hazard in northeastern India, particularly in Mizoram’s Aizawl district. To assess this hazard, we conducted a landslide susceptibility mapping study using the frequency ratio (FR) and analytical hierarchy process (AHP) methods. Twelve factors, including slope, elevation, aspect, curvature, topographic wetness index, topographic ruggedness index, rainfall erosivity factor, geology, land use land cover, normalized difference vegetation index (NDVI), and distance to road, were considered. The final susceptibility maps were created by combining these factors, each weighted according to its importance. Since no standard classification guidelines existed, we compared different systems like GIS-based geometric interval (GIC), natural breaks (NBC), and quantile interval (QIC) with mean-standard deviation (MSC). The area was divided into five susceptibility zones: very high, high, moderate, low, and very low. We evaluated the accuracy of these systems using success rate curve (SRC) and quality sum ratio (Qs) techniques. SRC results showed similar performance for GIC, NBC, QIC, and MSC classifications. The AHP method also yielded comparable results. After analyzing the AUC under SRC and Qs, we selected the MSC classification system as the optimal approach for landslide susceptibility mapping. Notably, the distance to roads factor had the highest predictive value, highlighting its importance in identifying slope instability for effective management strategies. As landslides pose significant risks in northeastern India, and our study using FR and AHP methods, along with MSC classification, provides valuable insights for mitigating this geological hazard.

Pixel-Based Spatio-Statistical Analysis of Landslide Probability in Humid and Seismically Active Areas of Himalaya and Hindukush

The Hindukush and Himalaya regions of Pakistan are chronically prone to several geological hazards such as landslides. Studying landslides in these regions is crucial for risk assessment and disaster management, as well as for determining the effects of adverse climatic conditions, infrastructure management, and increasing anthropogenic activities. High-relief mountains in these regions face severe challenges because of frequently occurring landslides and other natural hazards, especially during intensive rainfall seasons and seismic activity, which destroy infrastructure and cause injuries and deaths. Landslides in the Alpuri Valley (Hindukush) and the Neelum Valley (Himalaya) have been activated through high magnitude earthquakes, intensive rainfalls, snowfall, floods, and man-made activities. Landslide susceptibility mapping in these areas is essential for sustainable development as it enables proactive risk management, up-to-date decision-making, and effective responses to landslide hazards, ultimately safeguarding human lives, property, and the environment. In this study, the relative effect method was applied for landslide susceptibility modeling in both study areas to determine the capability to reduce the effects of landslides, and to improve the prediction accuracy of the method. The relative effect is a statistical model that has only been used for very limited time for landslide susceptibility with effective results. A total of 368 (Neelum Valley) and 89 (Alpuri Valley) landslide locations were identified, which were utilized to prepare the reliable landslide inventory using GIS. In order to evaluate the areas at risk for future landslides activities and determine their spatial relationship with landslide occurrences, the landslide inventory was developed with 17 landslide causative factors. These factors include slope gradient, slope aspect, geology, plan curvature, general curvature, profile curvature, elevation, stream power index, drainage density, terrain roughness index, distance from the roads, distance from the streams, distance from fault lines, normalized difference wetness index, land-use/land-cover, rainfall, and normalized difference vegetation index. Finally, the performance of the relative effect method was validated using the success and prediction curve rate. The AUC-validated result of the success rate curve in the Alpuri Valley is 74.75%, and 82.15% in the Neelum Valley, whereas, the AUC-validated result of the prediction rate curve of the model is 87.87% in the Alpuri Valley and 82.73% in the Neelum Valley. These results indicate the reliability of the model to produce a landslide susceptibility map, and apply it to other landslide areas. The model demonstrated a more effective result in the Alpuri Valley, having a smaller area. However, the results are also desirable and favorable in Neelum Valley, with it being a large area. It will assist in general landslide hazard management and mitigation, and further research studies related to future landslide susceptibility assessments in other parts of the region.

Assessing landslide susceptibility in Lake Abya catchment, Rift Valley, Ethiopia: A GIS-based frequency ratio analysis

Ethiopia's varied landscape, significant rainfall, and diverse geological characteristics pose risks of landslides. The specific research area spans 40 km2 within the Lake Abaya catchment area in the Rift Valley of Ethiopia. This investigation aimed to map landslide susceptibility using remote sensing information, GIS technology, and frequency ratio analysis. It evaluated multiple factors influencing landslide susceptibility. The process involved meticulous mapping of thematic layers, utilizing GIS techniques and diverse data sources, including primary data, satellite imagery, and secondary sources. A combination of Google Earth image analysis and field surveys was used to map landslide susceptibility in inaccessible areas. It was determined that 138 landslide sites existed. Of these, 30% (41 points) were assigned to the test of the model and another 30% to the training of the model, for a total of 97 points. The landslide susceptibility was classified into five categories based on frequency ratio analysis of the landslide susceptibility index (LSI): very low, low, moderate, high, and very high. The northeastern sector of the study area demonstrated a comparatively diminished susceptibility to landslides, ranging from low to moderate, whereas the central and southern regions showcased markedly elevated vulnerability. An evaluation of the model's accuracy using the area under the curve (AUC) method based on test inventory landslide data produced encouraging results: 84.8% accuracy on the success rate curve and 78.8% accuracy on the prediction rate curve. Based on the frequency ratio model, a susceptibility map is derived to represent susceptibility levels accurately.

Identification of geochemical anomalies in multibackground areas using the combined k-means clustering and residual contrast value method: A case study in a district in Hunan, China

Due to the complex characteristics of geological structures, identifying geochemical anomalies in valuable deposits using regional geochemical datasets of stream sediments is challenging. In this study, an effective combined method was proposed to solve the problem of anomaly identification in multibackground areas. First, samples were classified into different clusters through the k-means clustering method using major elements/minerals (such as SiO2 and Na2O) that can reflect the lithology and were chosen as classification indicators. Considering the double restriction of the contour coefficient and lithological background, each sample within the same cluster was considered to have the same background. Then, the residual value between each sample and the mean data of adjacent samples within the same cluster was calculated, and the original data of each sample were replaced with the ratio (a new parameter defined as the residual contrast value) of the residual value and the anomaly threshold obtained for the corresponding cluster. Finally, geochemical maps and anomaly maps were generated using the residual contrast values. A practical example involving a regional geochemical dataset of stream sediments in Hunan, China, was examined in detail to clarify the procedure. Moreover, a comparative analysis through success rate curves of the percentage of deposits correctly determined was performed between the traditional method, singularity method and new combined method. The results showed that the anomalies identified by the proposed method were closely associated with known deposits, and the residual contrast value, which considers the effects of the lithological background, random error, and structural anomalies, could eliminate the influence of lithology and enhance weak anomalies. Moreover, the geological significance of this method is clear, and the calculation procedure is simple. Thus, this method could be applied for identifying regional geochemical anomalies in multibackground areas and could be used as a guide for new exploration targets.

Geologically-constrained GANomaly network for mineral prospectivity mapping through frequency domain training data

Generative adversarial networks (GAN) and various deep autoencoders have been frequently executed to recognize multi-element geochemical anomalies linked to different ore resources in recent decade. Efficient recognition of multi-element geochemical anomaly patterns is a significant issue in mineral exploration targeting. Traditional procedures have not sufficient capability to perform efficient pattern recognition. While, deep learning algorithms as influential subset of machine learning algorithms can present magnificent conclusions in classification and pattern recognition. Because those have robust ability in extracting high-level features of complex inputs. Although, many deep learning algorithms were used to recognize geochemical anomalies but the GANs have demonstrated specific dignity in recognizing multi-element geochemical anomaly patterns. But, these frameworks should be constrained to learn geological knowledge and yield reasonable potential maps. In this regard, a novel geologically-constrained GANomaly was trained with frequency domain training data to recognize multi-element geochemical anomalies. Application of the geologically-constrained GANomaly network with considering mineral system parameters of the Au–Cu mineralization in the Feyzabad district, NE Iran was eventuated to suitable results. The success-rate curves demonstrated that produced map of frequency domain geochemical data has traced 86.68% Au–Cu occurrences via 30% corresponded area while produced map of spatial domain geochemical data has traced 80.13% Au–Cu occurrences via 30% corresponded area.

Lithium exploration targeting through robust variable selection and deep anomaly detection: An integrated application of sparse principal component analysis and stacked autoencoders

Lithium is a strategic metal for high-technology industries that plays a vital role in realizing electromobility and effective energy storage for smartphones and electric/hybrid vehicles, in addition to being important in accumulating energy from renewable sources. Motivated by this, the development of state-of-the-art workflows is imperative to make Li exploration more efficient and guarantee a sustainable supply to the demanding markets in the future. In this paper, a computational protocol was described to address two critical challenges raised in the Li exploration: (i) selection of mineralization-related pathfinder variables and (ii) recognition of multi-element geochemical anomalies related to ore-forming processes. Robust sparse principal component analysis (SPCA) was employed to reduce active (non-zero) loading coefficients and straightforward selection of Li pathfinder variables. The selected variables were then fed into the training network of stacked autoencoders (SAE) to learn deep representations of multivariate input signals and quantify reconstruction errors linked to Li-vectoring geochemical anomalies. As a representative example, the proposed workflow, i.e., SPCA + SAE, was implemented in the R programming language and applied to a real-case experiment with stream sediment geochemical data pertaining to the Moalleman district, NE Iran. Moreover, compositional robust principal component analysis (RPCA) and Mahalanobis distance (MD) technique were adopted to constitute two comparative models, RPCA + SAE and SPCA + MD, as benchmarks to judge the competence of the SPCA + SAE model in variable selection and anomaly detection, respectively. A performance appraisal by success-rate curves and relevant area under the curves (AUCs) indicated that the SPCA + SAE model brings, by calculating the highest AUC, spatial patterns that are more promising for vectoring towards Li-mineralized grounds. Moreover, AUCSPCA+MD was measured to be greater than AUCRPCA+SAE, suggesting robust variable selection is a more important paradigm than deep anomaly detection for Li exploration targeting. Student's t–statistic method was eventually applied to the anomaly map from the SPCA + SAE model to define relevant thresholds for narrowing down prospective areas for the next round of Li exploration within the study area.

A comprehensive evaluation of OPTICS, GMM and K-means clustering methodologies for geochemical anomaly detection connected with sample catchment basins

The process of data-driven clustering to uncover geochemical anomalies linked to sample catchment basins (SCBs) includes a comprehensive framework to discern areas exhibiting unique geochemical attributes within a specified study area. The Ordering Points to Identify the Clustering Structure (OPTICS) method can serve as a robust methodology for detecting geochemical anomalies in SCBs. This is attributed to its capacity to effectively manage varying cluster densities, adaptively identify cluster numbers, exhibit resilience to noise, and display minimum sensitivity to parameters. A comparison was conducted in this research between the outcomes of the OPTICS clustering algorithm and two traditional clustering techniques, namely the Gaussian Mixture Model (GMM) and K-means clustering. In the following, the Expectation-Maximization (EM) technique is employed to train the GMM for clustering. Moreover, the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) as two validate statistical metrics implemented to ascertain the optimal number of components (clusters) belong to the GMM. It should be noted that the effectiveness of the clustering algorithms was further assessed using the Calinski-Harabasz (CH) index and the success-rate curves. OPTICS, a density-based clustering approach, was confirmed to be more effective than K-means and GMM for identifying MVT PbZn anomalies in Varcheh district, western Iran. Furthermore, the specified anomalies show a geo-spatial correspondence with the geological facts, and it has been observed that strong anomalies are more discoverable in close proximity to MVT PbZn occurrences. This work suggests a novel anomaly detection approach based on OPTICS, which exhibits superior performance and data-modeling efficiency. The main emphasis is on effectively distinguishing geochemical anomalies from sample data originating from populations with uncertain distributions.

Aquifers Recharge Potentiality Evaluation using Bivariate and Multivariate Statistical Analysis: A Case Study in Malwathu Oya River Basin

An aquifer is a body of porous rock or sediment saturated with groundwater that plays a vital role in the hydrological cycle. Due to rapid population growth and urbanization, these valuable resources face threats like climate changes, pollution, and over-exploitation. Overdraft, where extraction exceeds recharge, can lead to groundwater depletion. Although Sri Lanka is not considered a water- scarce country, it faces rising water demands due to rapid population growth. Especially in dry zones, groundwater is crucial for drinking and agriculture; unauthorized agro-wells exacerbate groundwater issues. Therefore, this research aims to assess groundwater recharge potential zones in Malwathu Oya river basin using Bivariate and Multivariate statistical analysis. Comparing these techniques will lead to the precise identification of potential recharge areas. And this study primarily focuses on the implementation of GIS to find solutions for this kind of challenge without performing field investigations. Therefore, eight conditioning factors; water level, electrical conductivity, lithology, soil, slope, lineament density, drainage density and land use were selected as groundwater recharge influencing factors. The well inventory database consisted of 380 well locations and, out of these, 70% was considered as the training dataset, and the remaining 30% was taken as the testing data set. Groundwater recharge potential maps (GRP) were developed in ArcGIS 10.8, based on the results obtained from Frequency Ratio (FR), Logistic Regression (LR) and their ensemble method. For testing the accuracy of the FR model, test well locations were cross tabulated with the GRP map, and the highest well percentage was achieved in high groundwater recharge potential zone (37.39%). In the LR model, cross-tabulation results showed that 62.61% of testing wells were in the very high potential zones. The ensemble method of FR and LR revealed that 53.04% of test wells fall into the very high potential class. Hence, it can be assumed that LR predictions are more reliable than the FR model. To further validate the LR model and ensemble model, Receiver Operating Characteristics curve method was used. The Area Under the Curve (AUC) values for Success rate curve (SRC) and prediction rate curve (PRC) in LR method were 63% and 57% respectively. In ensemble method, the AUC values for SRC and PRC were 89% and 75%. These results indicate that the LR model is relatively poor estimator, whereas FR and LR ensemble method is relatively good estimator to evaluate groundwater recharge potential in Mulwathu Oya river basin.
 
 Keywords: Area under curve, Frequency ratio, GIS, Groundwater recharge, Logistic regression

A data-driven VIKOR procedure for predictive modeling of porphyry copper prospectivity in SE Iran

This research is aimed at comparing the performance of a proposed data-driven VIKOR method with its knowledge-driven version for porphyry copper prospectivity modeling in the Sardouyeh district, Kerman province, SE Iran. Sixteen exploratory layers from various sources were first nominated for this purpose. These layers were provided from the following sources: (1) multi-element geochemical signatures derived from principal component analysis (PCA); (2) distribution of geophysical footprints derived from aeromagnetic and potassium radioelement surveys; (3) proximity to ore host rocks and tectonic controls; and (4) proximity to hydrothermal alterations. For better comprehension of exploration layers, their values were transferred to the same range [0–1] using fuzzy logic. For predictive modeling of porphyry-type copper mineralization, an improved data-driven VIKOR model was introduced in this paper that its main foundation is based on the calculation of criteria and their relevant classes weights in a data-driven way using prediction-area (P-A) plots and normalized density index (Nd), respectively. This methodology was applied twice: once on all exploratory layers and again only on efficient evidence layers. In order to compare and evaluate the performance of the novel methodology introduced for predicting high-favorable areas associated with Cu-porphyry deposits, on the basis of expert opinion and an Analytic hierarchy process (AHP) approach, the AHP-VIKOR procedure was implemented as a knowledge-driven method. A success-rate curve was employed to accurately compare the performance of three models, quantitatively. The findings reveal that the improved data-driven VIKOR approach using efficient evidence layers covered 80 % of Cu occurrences in only 6 % of the total area of the region, which led to much more precise and high-quality outcomes. Moreover, it provides more effective performance to be employed for more detailed exploration programs in this study area and similar studies in other regions.

Groundwater potential zone mapping using GIS and Remote Sensing based models for sustainable groundwater management

The present research is conducted in the southern region of Khyber Pakhtunkhwa, Pakistan, to identify groundwater potential zones (GWPZ). We used three models including Weight of Evidence (WOE), Frequency Ratio (FR), and Information Value (IV) with twelve parameters (elevation, slope, aspect, curvature, drainage network, LULC, precipitation, geology, Lineament, NDVI, road, and soil texture, that have been prepared and integrated into ArcGIS 10.8. The reliability of the applied models’ results was validated using Area Under the Receiver Operating Characteristics (AUROC). The GWPZ were reclassified into five classes, i.e. very low, low, medium, high, and very high zone. The area occupied by mentioned classes using WOE are very low (10.14%), low (19.58%), medium (26.75%), high (27.10%), very high (16.40%), while using FR are very low (20.93%), low (32.38%), medium (18.92%), high (13.13%), very high (14.61%) and using IV are very low (14.41%), low (17.17%), medium (29.01%), high (25.85%), and very High (13.53%). The Success Rate Curve of WOE, FR, and IV were 0.86, 0.91, and 0.87, while the Predicted Rate Curve values were 0.89, 0.93, and 0.90, respectively. The results revealed that all applied statistical models performed very well to delineate GWPZ. However, use of the FR technique is strongly encouraged to evaluate the GWPZ, and its findings are especially useful for managing groundwater resources in urban planning. Our approaches for assessing the GWPZ mapping can be applied in any region with similar scenarios and are recommended as a helpful tool for policymakers to manage groundwater.