Recent studies have revealed the remarkable capabilities of multimodal large models (MLLMs) in general vision-and-language tasks, generating increasing interest in their application to geoscientific domains. Rock image recognition and lithological description constitute fundamental skills for geologists and represent one of the earliest application scenarios in geosciences where artificial intelligence technologies have been actively explored. As rock image recognition and description are inherently multimodal tasks, involving both visual and textual data, models are required to understand the interactions between modalities and how these interactions affect information transfer. However, the simultaneous comprehension of rock images and their corresponding lithological descriptions remains a significant challenge for current machine learning frameworks. The impact of multimodal co-training on downstream geoscientific performance is still unconfirmed. This study presents PetroMind, a domain-adapted multimodal petrographic foundation model built upon Qwen2.5-VL, together with SA-Rock, a novel LLM-based metric designed to assess the semantic accuracy of rock image generation descriptions to evaluate the capabilities of MLLMs in petrographic multimodal tasks. In the rock image classification task, PetroMind achieves performance comparable to ViT-Base-Patch16-224, with accuracy and Macro-F1 scores exceeding or approaching 97%. This demonstrates PetroMind’s strong capability in long-tailed image learning and highlights its effectiveness in substantially improving classification accuracy for few-shot rock image categories. In the lithological description generation task, PetroMind attains BLEU-4 and SA-Rock scores of 0.644 and 7.864, respectively, indicating good few-shot learning performance. The SA-Rock metric shows that the model produces highly accurate descriptions of rock structure, texture, and colour, while leaving considerable scope for improvement in the description of mineral composition. Ablation experiments further indicate that task-specific LoRA adapters are more effective for high-resource tasks such as image classification, whereas a single shared LoRA adapter demonstrates superior multi-task interaction in low-resource captioning scenarios. This study demonstrates the potential of MLLM-based architectures in jointly understanding rock images and their associated geological descriptions.

Constraining the onset of the Lhasa-Qiangtang collision is essential for understanding the assembly and uplift of the central Tibetan Plateau. We compile 202 Paleozoic-Cenozoic paleomagnetic poles from the Lhasa and Qiangtang terranes published between 1980 and 2023 and screen them using the Van der Voo quality criteria complemented by reliability considerations; 22 poles affected by local rotations/outliers are further excluded. Using the screened dataset, we construct apparent polar wander paths (APWPs) for both terranes by calculating 20 Myr running-mean poles and fitting a spherical smoothing spline (smoothness parameter = 50). We also reconstruct paleolatitude drift curves for the northern Lhasa margin and southern Qiangtang margin at a fixed reference site (31.8 degrees N, 87.7 degrees E). The APWPs and paleolatitude trajectories indicate progressive convergence during 150-120 Ma, with Lhasa poles approaching contemporaneous Qiangtang poles. The two paleolatitude curves become indistinguishable within combined 95 % uncertainties at similar to 118 Ma; given the 20 Myr window, this corresponds to an onset time of similar to 118 +/- 10 Ma for collision initiation in the Early Cretaceous.

Seismic horizon tracking is a fundamental and critical step in seismic interpretation. Efficient and robust identification of seismic horizons can significantly enhance the accuracy and efficiency of geological modeling. In light of this, this study investigates seismic horizon identification from an image processing perspective: Edge detection facilitates the extraction of overall reflectors and is suitable for preliminary interpretation, while deep learning-based segmentation is employed to identify target horizons in detailed interpretation stages. This research takes the three-dimensional seismic data from the Condabri area of the Surat Basin, Australia, as a case. It begins with preprocessing the seismic data, including amplitude gain recovery to enhance deep seismic imaging quality and 3D seismic data filtering to suppress random noise, thus facilitating subsequent seismic interpretation and analysis. After preprocessing, the overall seismic horizons of the study area were extracted using an improved Canny edge detection algorithm. The results show that the improved Canny edge detection method can effectively improve the detection ability of seismic events compared with the original operator. The edge detection results provided a foundation for further detailed interpretation by highlighting the distribution of seismic horizons. Based on these results and well-log interpretations, the target horizons were manually interpreted as training labels for the subsequent improved UNet model. Comparative and ablation experiments were also carried out to evaluate the model's performance in horizon identification. The findings demonstrated that the proposed model outperformed the original UNet model and others, such as SegNet and UNet++, showing improvements across various evaluation metrics. The prediction results on the test set further validate the model’s capability to identify continuous seismic horizons across fault zones accurately. Following vectorization, the predicted horizons exhibit high consistency with manually interpreted results, indicating the model's effectiveness in capturing complex stratigraphic features. This study provides a practical and scalable approach for improving the efficiency and accuracy of seismic interpretation workflows.

Three-dimensional mineral prospectivity modeling (3D MPM) is an emerging tool for targeting deep-seated concealed mineralization. Recently, three-dimensional convolutional neural network (3D CNN) has gained increasing application in 3D MPM due to their ability to integrate multi-source 3D predictive maps. However, existing 3D CNN-based models typically require extensive preprocessing of 3D geological models before they can be integrated with continuous predictive maps. This preprocessing often relies on artificially selected spatial analysis methods to quantify the influence of geological structures on mineralization. While useful, this approach may not be able to fully extract the ore-controlling features from the 3D geological models and may result in a potential risk of acquiring ineffective information by the deep learning model. To address these issues, this paper proposes a hybrid 3D GCN-CNN model that combines the 3D graph convolutional network (3D GCN) module for modeling distance relationships among geological structures with the 3D CNN module for extracting geometric morphology and spatial distribution patterns of geological bodies. This hybrid framework directly extracts information from 3D geological models without spatial analysis. The proposed methodology is applied to a case study in the eastern Chating area in Anhui Province, China. By employing the 3D GCN–CNN hybrid model, this paper demonstrates its superior performance compared to the 3D CNN model. Drilling validation confirmed the effectiveness of the 3D GCN-CNN model, suggesting that the proposed framework offers an efficient method for deep mineral exploration.

Lode gold deposits are among the most important types of gold ore deposits and are typically characterized by medium to low-temperature CO2-rich hydrothermal fluids. Consequently, they are commonly accompanied by the precipitation of abundant carbonates. However, the key role of CO2 in the transport and precipitation of gold remains unclear. Here, we discuss the genetic relationships between carbonate and gold associated with the Rongdu deposit, which is the largest gold deposit in the Wuhe mining district along the southern margin of the North China Craton (NCC). The C-O isotopic compositions of carbonates from different generations suggest that the CO2-rich hydrothermal fluids of the gold deposit were derived from magma. The precipitation of carbonates preceded the formation of high-grade gold ores, possibly as a result of large-scale fluid boiling. After the formation of ankerite (Ank1), the deposit underwent repeated magmatic hydrothermal fluid replenishment; however, external fluids were not introduced during this process. Moreover, the fluctuating anomalous Fe content in the carbonates suggests a competitive relationship for the incorporation of iron between sulfides and carbonates. The formation of carbonates hindered gold precipitation to some extent, but as an indicator of fluid boiling, it can serve as a prospecting indicator for gold in the area.

Evaluating and predicting how carbon storage (CS) is impacted by land use change can enable optimizing of future spatial layouts and coordinate land use and ecosystem services. This paper explores the changes in and driving factors of Zunyi CS from 2000 to 2020, predicts the changes in CS under different development scenarios, and determines the optimal development scenario. Woodland and farmland are the main land use types in Zunyi. Land use change was reflected mainly in the mutual conversion among woodland, farmland, and grassland and by their conversion to construction land and water. In 2000, 2010, and 2020, the CS in Zunyi was 658.77 x 10<^>6 t, 661.44 x 10<^>6 t, and 658.35 x 10<^>6 t, respectively. Woodland, farmland and grassland conversions to construction land and water were primarily responsible for CS loss. The normalized difference vegetation index (NDVI) is the main factor influencing the pattern of CS (q > 10%). Furthermore, the impacts of the human footprint index and population density are increasing. In 2030, the CS of Zunyi is trending downward. Under the ecological-farmland conservation scenario (ECS), the CS is estimated to be 656.67 x 10<^>6 t, with the smallest decrease (- 0.26%) among timepoints. The effective control of woodland and farmland weakens the trend of CS reduction.

Accurate ore classification is essential for geological exploration and mineral resource assessment, particularly in geologically complex settings. This study presents an interpretable classification framework that integrates the Light Gradient Boosting Machine (LightGBM) algorithm with post hoc model interpretation using SHapley Additive exPlanations (SHAP). The framework is applied to geochemical and spatial data from the Qiaomaishan Cu-S polymetallic deposit in Anhui Province, China. A dataset comprising 1588 samples-each containing concentrations of 29 geochemical elements along with 3D spatial coordinates-was used to train and evaluate 5 machine learning models: Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Random Forest (RF), LightGBM, and CatBoost. Among these, LightGBM achieved the best performance, with an average F1 score of 0.990 across 10 ore and lithological categories. An ablation experiment further confirmed the critical role of spatial coordinates, as removing them led to a notable drop in classification accuracy, particularly for underrepresented ore types. To interpret the LightGBM model, SHAP analysis was used to quantify feature contributions, identifying key geochemical elements such as W, Sr, Ca, and Fe as significant drivers of classification-consistent with the known mineralogical characteristics of the deposit. Moreover, SHAP beeswarm plots provided insights into the direction and magnitude of each feature's influence, enhancing the model's geological interpretability. SHAP-based feature selection further improved classification performance for underrepresented classes, including copper–sulphur–tungsten ore and copper ore. The proposed framework demonstrates strong potential for facilitating automated ore identification and supporting data-driven decision-making in complex geological environments.

The application of hydrochar in the field of photocatalysis is gradually attracting attention, nevertheless, the impact of coexisting substances on the photocatalytic activity of iron-based hydrochar remains insufficiently explored. The study converted iron-enriched sludge into hydrochar and thoroughly investigated the influence of citric acid (CA) on the photocatalytic activity of hydrochar for the degradation of rhodamine B (RhB). The addition of CA significantly enhanced the photocatalytic degradation rate constant (k) of RhB by hydrochar, increasing it from 0.062 +/- 0.015 h-1 to 0.309 +/- 0.007 h-1 . Furthermore, the removal efficiency of RhB also increased from 21.6 +/- 1.0 % to 93.6 +/- 2.0 %. This enhancement was attributed to the formation of a Fe 3+-CA complex within the hydrochar matrix. The complex activated molecular oxygen produced hydroxyl radicals for the effective degradation of RhB. The Fe-O-Si bonds in the hydrochar diminished the excitation energy and increased the solar absorption of the Fe 3+-CA complex. The structure of dissolved organic matter (DOM) derived from hydrochar transformed from a dispersed to an aggregated state following the addition of CA, and the triplet state DOM (3DOM*) transferred energy to oxygen under light excitation to generate singlet oxygen. This study underscores that coexisting substances in the environment affect the photocatalytic process of hydrochar, deepening the understanding of the environmental behavior and potential application of hydrochar.

The Paodaoling Au deposit is the largest Au deposit in the Middle-Lower Yangtze River Valley Metallogenic Belt (MLYB), with Au reserves of 36.5 tons and an average grade of 1.93 g/t. On the basis of geological field surveys and mineralogical studies, LA-ICP-MS mapping and in situ S-Pb isotope and LA-ICP-MS U-Pb dating of dolomite were carried out to identify the genetic type, metallogenic age and source of ore-forming materials of the deposit. In the Paodaoling Au deposit, hydrothermal fluids filled mainly high-angle tension joints with dip angles of 75-85 degrees, as well as associated structures, and shear joints with dip angles of 50 degrees, suggesting that the fluids filled the host granodioritic porphyry when it was completely consolidated. LA-ICP-MS U-Pb dating of dolomite associated with gold mineralization indicates a metallogenic age of 137.3 +/- 2.3 Ma (n = 58; MSWD = 3.6; 2 sigma), which is obviously later than that of the host granodioritic porphyry (147-141 Ma). The lead isotope data reveal ratios of Pb-206/Pb-204 = 18.190-18.450, Pb-207/Pb-204 = 15.594-15.646, Pb-208/Pb-204 = 38.467-38.625, which plot in the mixed region between the mantle and the upper crust. The S isotopes indicate that the ore-forming fluids of the Paodaoling Au deposit were derived from magma and that sulfur from the strata was added during the process of gold mineralization. Au exists mainly in the form of a solid Au+1 solution in arsenopyrite and arsenian pyrite as the low sulfidation state. The Paodaoling Au deposit formed in an intracontinental orogenic setting without contemporaneous volcanic rocks, and its ores are characterized by enrichment in As and depletion in Te, Se, and Ag, with Ag:Au < 1. The mineral associations and characteristics of the ore-forming elements of the Paodaoling Au deposit are similar to those of typical LS epithermal deposits that are related to subalkalic igneous rocks. Thus, this study confirms for the first time the existence of LS epithermal deposits associated with porphyry and skarn Cu(Au) in the MLYB, which is highly important for the metallogenic theory of porphyry and skarn-epithermal deposits in intracontinental orogeny and for guiding regional geological prospecting.

Understanding the correlation between natural and accelerated carbonation processes in concrete and their distinct impacts on steel corrosion mechanisms remains a critical challenge for durability prediction. In this study, the steel reinforcement concrete specimens is taken as the main research object, using the methodology of natural and accelerated carbonation, combined with microscopic means, electrochemical parameters, corrosion morphology analysis of steel bars and other evaluation methods, to systematically investigated the coupled effects of CO2 concentration (5-20 %), temperature (20-40 degrees C) and relative humidity (40-70 % RH) on the concrete carbonation kinetics, microstructure evolution and steel corrosion behavior of concrete. The results showed that carbonation depth exhibits a high temperature and CO2 concentration dependence, the relationship between carbonation depth and RH exhibits a quadratic function characteristic, peaking at 55 % RH. Under accelerated carbonation conditions, a low CO2 acceleration (5 %) preserves microstructural similarity to natural carbonation, whereas high CO2 concentrations can cause the substantial reduction of C-S-H but still allow for the observation of Ca(OH)2, high concentrations induce CaCO3 crystallization differences. Compared to temperature and CO2 concentration, the influence of RH on the concrete microstructure is less significant. Under both accelerated and natural carbonation conditions, the corrosion behavior of steel bar in paste is similar but exhibits differences. In most cases, there is a high correlation between accelerated carbonation and natural carbonation, accelerated carbonation tests can effectively reflect the primary characteristics of steel corrosion caused by natural carbonation, offering certain advantages in assessing the corrosion resistance of steel bar. The results presented have a great of importance for advancing concrete carbonation research, enhancing the prediction and improving the reliability of practical applications.

In the field of geosciences, the integration of artificial intelligence is transitioning from perceptual intelligence to cognitive intelligence. The simultaneous utilization of knowledge and data in the geoscience domain is a universally addressed concern. In this paper, based on the interpretability of deep learning models for rock images, rock features such as structure, texture, mineral and macroscopic identification characteristics were selected to extract a rock identification subgraph from the petrographic knowledge graph and carry out rock type similarity reasoning. Comparative experiments were conducted on few-shot learning of rock images under the supervision of rock type similarity knowledge. The results of the few-shot learning comparisons demonstrate that the supervision of rock type similarity knowledge significantly enhances performance. Additionally, rock type similarity knowledge exhibits a marginal effect on improving few-shot learning performance. Given the absence of Chinese word embedding and large-scale Chinese pre-trained language models in the geological domain, graph embedding based on domain-specific knowledge graphs in geosciences can offer computable geoscience knowledge for research dually propelled by data and knowledge.

Nowadays, mineral exploration has increasingly focused on the targeting of deep-seated orebodies. Mineral prospectivity modeling is one of the important approaches facilitating exploration targeting and mitigating risks associated with mineral exploration, particularly under cover. Recent advances in 3D mineral prospectivity modeling, enable the effective extraction of predictive information from three-dimensional geological models, enhancing the accurate identification of the deep-seated orebodies. Notably, these advancements have synergized with deep learning approaches to improve the efficiency of mineral exploration based on their nonlinear and multi-layer sensing attributes, effectively identifying and extracting key relationships between the 3D predictive maps and mineralization. Currently, the main deep learning method used for 3D mineral prospectivity modeling is convolutional neural network (CNN) models. However, the related research has not considered the multiscale features of geological structures, it can be made further improvements on this regard. This paper introduces a multi-scale 3D convolutional neural network model (3D CNN) incorporating a spatial attention mechanism and an Inception module (MSAM-CNN) for 3D mineral prospectivity modeling. By integrating Inception modules and spatial attention mechanisms, the network's capability to identify multi-scale geological features and concentrate on key predictive areas is significantly enhanced. This leads to further improvement in the accuracy and generalization capability of 3D mineral prospectivity modeling. To evaluate the effectiveness of this model, we carry out a case study on 3D mineral prospectivity modeling in the Baixiangshan iron deposit within the Ningwu Basin of the Middle-Lower Yangtze River Metallogenic Belt, China. The results show that the multi-scale 3D convolutional neural network model exhibits remarkable robustness and generalization capabilities. It can effectively delineate targets within the deep and peripheral areas of the deposit, providing Indications for future exploration. The addition, performance indicators, ROC curve, and Capture-Efficiency curve consistently demonstrate that the MSAM-CNN model outperforms the Inception-enhanced CNN (M-CNN), CNN, Random Forest (RF), and Support Vector Machine (SVM) models. All of these indicate that MSAM-CNN can extract the 3D spatial features within 3D predictive maps better during 3D mineral prospectivity modeling, positioning it as a promising tool for yielding more reliable targets for deep-seated mineralization in future exploration.

AbstractDrilled cores provide first‐hand data for deposit research. The element compositions at different depths of cores are an important basis for evaluating the reserves of mineral deposits. Therefore, a dataset containing a complete drill core element content has great value on practicality. In this study, a pXRF analyser was used to measure the element contents in eight cores of the Xuancheng Qiaomaishan copper–sulphur deposit ZK16 + 02, ZK16 + 03, ZK16 + 04, ZK1804, ZK18 + 02, ZK1403, ZK1602, and ZK1604, and a geochemical dataset was obtained. The dataset includes Cu, S, Fe, W, Mg, Al, Si, As, Mn, Co, Ni, Zn, Se, Ti, V, Cr, Ca, Cl, P, Sr, Y, Zr, Nb, Mo, Pb, Th, U, and other 27 major and trace elements. Based on this dataset, the vertical variation pattern of main and trace element content in drill cores can be examined, the distributions of elements in full hole or exploration line profile can be drawn, and element correlation analysis and principal component analysis can be performed.

Faults serve as oil and gas storage space and transportation channels, so fault identification is significant to oil and gas exploration. Fault extraction methods based on manual identification or seismic body attributes are prone to recognition errors due to human factors or poor data quality. With the development of deep learning, researchers have proposed different network models to extract 3D faults. However, the traditional models still have room for improvement in fine-grained segmentation results and model robustness. Therefore, this study proposes a new multi-scale feature fusion network architecture named MAR-UNet. In order to solve the defect of insufficient fine granularity of traditional model segmentation results, this paper designs a local feature extraction module named Residual Sampling Convolution block (RSC block) and deploys it to MAR-UNet; at the same time, in order to improve the defect that the existing 3D model cannot effectively deal with complex spatial relationship features, this study designs a plug-and-play attention module named Mix Attention Mechanism (MAM) in the model. Finally, this paper proposes a compound loss function named Weight Focal-Dice loss for the model's weak robustness caused by sample imbalance. The results of ablation and cross-experiments show that the loss function proposed in this paper is suitable for accomplishing the fault segmentation task under the influence of sample imbalance, and the model proposed in this paper still shows good reliability and robustness when deploying the model to the actual workspace data.

The latest versions of the Ross-Li model include kernels that represent isotropic reflection of the surface, describe backward reflection of soil and vegetation systems, characterize strong forward reflection of snow, and adequately consider the hotspot effect (i.e., RossThick-LiSparseReciprocalChen-Snow, RTLSRCS), theoretically able to effectively characterize BRDF/Albedo/NBAR features for various land surface types. However, a systematic evaluation of the RTLSRCS model is still lacking for various land cover types. In this paper, we conducted a thorough assessment of the RTLSRCS and RossThick-LiSparseReciprocalChen (RTLSRC) models in characterizing BRDF/Albedo/NBAR characteristics by using the global POLDER BRDF database. The primary highlights of this paper include the following: (1) Both models demonstrate high accuracy in characterizing the BRDF characteristics across 16 IGBP types. However, the accuracy of the RTLSRC model is notably reduced for land cover types with high reflectance and strong forward reflection characteristics, such as Snow and Ice (SI), Deciduous Needleleaf Forests (DNF), and Barren or Sparsely Vegetated (BSV). In contrast, the RTLSRCS model shows a significant improvement in accuracy for these land cover types. (2) These two models exhibit highly consistent albedo inversion across various land cover types (R2 > 0.9), particularly in black-sky and blue-sky albedo, except for SI. However, significant differences in white-sky albedo inversion persist between these two models for Evergreen Needleleaf Forests (ENF), Evergreen Broadleaf Forests (EBF), Urban Areas (UA), and SI (p < 0.05). (3) The NBAR values inverted by these two models are nearly identical across the other 15 land cover types. However, the consistency of NBAR results is relatively poor for SI. The RTLSRC model tends to overestimate compared to the RTLSRCS model, with a noticeable bias of approximately 0.024. This study holds significant importance for understanding different versions of Ross-Li models and improving the accuracy of satellite BRDF/Albedo/NBAR products.

While geological data gathering expands and advances along with mineral exploration, there is still a need for more innovation and enrichment because there are few complete analytical tools for these data. The light gradient boosting machine (LightGBM) technique is used in this study to estimate the prospectivity for minerals in three dimensions. However, because of the LightGBM model's large number of hyper-parameters, it is difficult to manually configure and alter model hyper-parameters, which has a substantial impact on the trained model's accuracy and dependability. Therefore, to optimize the LightGBM hyper-parameters, we use genetic algorithm (GA), which has the resilience and global optimization searchability in addressing difficult optimization problems. The GA–LightGBM algorithm for 3D mineral prospectivity modeling is the name we give to this combined approach. This study compares the GA–LightGBM algorithm against the GA-optimized support vector machine (SVM) and random forest (RF) algorithms in order to assess its applicability and superiority. The training set accuracy, test set accuracy, and Kappa coefficient values for the GA–LightGBM algorithm model were 0.9763, 0.9651, and 0.9453, respectively. The GA–LightGBM model's receiver operating characteristic curve was quite close to the upper left corner of the graph. The findings show that in terms of application and predictability, the GA–LightGBM ensemble learning approach performs better than the GA-optimized SVM and GA-optimized RF models. The outcomes of this study offer fresh perspectives for 3D mineral prospectivity modeling.

The traditional Fenton system is subject to the low efficiency of the Fe(III)/Fe(II) conversion cycle, with significant attempts made to improve the oxidation efficiency by overcoming this hurdle. In support of this goal, iron-enriched sludge-derived hydrochar was prepared as a high-efficiency catalyst by one-step hydrothermal carbonization and its performance and mechanisms in mediating the oxidation of triclosan were explored in the present study. The hydrochar prepared at 240 °C for 4 h (HC240-4) had the highest removal of triclosan (97.0%). The removal of triclosan in the HC240-4/H2O2 system was greater than 90% in both acidic and near-neutral environments and remained as high as 83.5% after three cycles, indicating the broad pH applicability and great recycling stability of sludge-derived hydrochar in Fenton-like systems. H2O2 was activated by both persistent free radicals (PFRs; 19.7%) and iron (80.3%). The binding of Fe(III) to carboxyl decreased the electron transfer energy from H2O2 to Fe(III), making its degradation efficiency 2.6 times greater than that of the conventional Fenton reaction. The study provides a way for iron-enriched sludge utilization and reveals a role for hydrochar in promoting iron cycling and electron transfer in the Fenton reaction.

三维地质模型是三维成矿预测最重要的数据基础之一,其准确性对于深部预测靶区的圈定具有十分重要的意义.然而,三维地质模型存在一定程度的不确定性,其不确定性主要是受到地质体与地质结构自身的复杂性,数据密度、误差和建模方法等多方面影响.本文以长江中下游成矿带钟姑矿田为例,基于隐式三维地质建模及蒙特卡洛模拟方法,对控制剖面中地质界线产状引起的三维地质模型的不确定性开展定量分析,并度量和评价其对三维成矿预测结果的影响.研究结果显示,深部地质界线的产状倾角及倾向变化会导致三维地质模型产生不同程度的不确定性.通过扰动策略,能够有效开展相关的不确定性分析和度量工作;三维地质模型的不确定性会在一定程度上影响三维成矿预测结果,但三维成矿预测方法具有一定的鲁棒性,基于预测结果圈定的找矿靶区位置相对稳定.对于具有较高不确定性的深部预测靶区,进一步工作可通过降低三维地质模型的不确定性或增加其他预测信息等方式,提高预测结果的确定性,以降低找矿勘探风险.