Journal of Applied Geophysics最新文献_第4页

Interferometric synthetic aperture sonar for high-resolution seafloor mapping and imaging in contrasting geomorphological and benthic settings 干涉式合成孔径声纳在对比地貌和底栖环境中的高分辨率海底测绘和成像

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2026-01-01 DOI: 10.1016/j.jappgeo.2025.106087

Caroline Gini , John W. Jamieson , Craig J. Brown , Brian Carroll , Richard Charron , David Shea , Katleen Robert

Seafloor exploration for geological or biological studies ideally requires high resolution survey data over a large area. However, there is a trade-off between resolution and coverage using conventional acoustic imaging and mapping techniques. The recent development of interferometric synthetic aperture sonar (InSAS), which provides high resolution imagery and bathymetry (3 and 25 cm/pixel, respectively) over large areas of seafloor, and has primarily been used for military and commercial purposes, opens the door for new applications for geological mapping, and seafloor classification and monitoring. For these applications, the processing steps, survey parameters and requirements for InSAS surveying, compared to conventionally used techniques such as multibeam echosounders and side-scan sonars, are not well defined. In this study, we describe and discuss the results of InSAS surveys in two contrasting geological and benthic settings: a relatively flat continental shelf, and a topographically complex mid-ocean ridge. Features of sizes down to 6 cm were identified on the imagery, including lava flow crust lineations, bedrock sedimentary bedding, gravels, and discarded rope. We found that seafloor features <1 m high were better imaged than taller features, such as hydrothermal vents or faults. We test and quantify survey parameters necessary to optimize data quality for effective use for scientific applications. Our results indicate that seafloor bathymetry is the most important consideration to maximize likelihood of data generation success and data quality when planning InSAS surveys.

地质或生物研究的海底勘探理想地需要大面积的高分辨率调查数据。然而，在使用常规声学成像和制图技术的分辨率和覆盖范围之间存在权衡。干涉合成孔径声纳（InSAS）的最新发展，在海底大面积提供高分辨率图像和测深（分别为3和25厘米/像素），主要用于军事和商业目的，为地质测绘和海底分类和监测的新应用打开了大门。对于这些应用，与传统技术（如多波束回声测深仪和侧扫声纳）相比，InSAS测量的处理步骤、测量参数和要求并没有很好地定义。在本研究中，我们描述并讨论了InSAS在两种不同地质和底栖环境下的调查结果：相对平坦的大陆架和地形复杂的洋中脊。在图像上识别出小至6厘米的特征，包括熔岩流地壳线、基岩沉积层理、砾石和丢弃的绳索。我们发现，1米高的海底特征比更高的特征（如热液喷口或断层）成像更好。我们测试和量化必要的调查参数，以优化数据质量，有效地用于科学应用。我们的研究结果表明，在规划InSAS调查时，海底测深是最大限度地提高数据生成成功率和数据质量的最重要考虑因素。

{"title":"Interferometric synthetic aperture sonar for high-resolution seafloor mapping and imaging in contrasting geomorphological and benthic settings","authors":"Caroline Gini , John W. Jamieson , Craig J. Brown , Brian Carroll , Richard Charron , David Shea , Katleen Robert","doi":"10.1016/j.jappgeo.2025.106087","DOIUrl":"10.1016/j.jappgeo.2025.106087","url":null,"abstract":"<div><div>Seafloor exploration for geological or biological studies ideally requires high resolution survey data over a large area. However, there is a trade-off between resolution and coverage using conventional acoustic imaging and mapping techniques. The recent development of interferometric synthetic aperture sonar (InSAS), which provides high resolution imagery and bathymetry (3 and 25 cm/pixel, respectively) over large areas of seafloor, and has primarily been used for military and commercial purposes, opens the door for new applications for geological mapping, and seafloor classification and monitoring. For these applications, the processing steps, survey parameters and requirements for InSAS surveying, compared to conventionally used techniques such as multibeam echosounders and side-scan sonars, are not well defined. In this study, we describe and discuss the results of InSAS surveys in two contrasting geological and benthic settings: a relatively flat continental shelf, and a topographically complex mid-ocean ridge. Features of sizes down to 6 cm were identified on the imagery, including lava flow crust lineations, bedrock sedimentary bedding, gravels, and discarded rope. We found that seafloor features <1 m high were better imaged than taller features, such as hydrothermal vents or faults. We test and quantify survey parameters necessary to optimize data quality for effective use for scientific applications. Our results indicate that seafloor bathymetry is the most important consideration to maximize likelihood of data generation success and data quality when planning InSAS surveys.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"246 ","pages":"Article 106087"},"PeriodicalIF":2.1,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Estimation of S-wave velocity of gas hydrate-bearing sediments using an improved Iso-frame model 基于改进等框架模型的含天然气水合物沉积物横波速度估算

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-31 DOI: 10.1016/j.jappgeo.2025.106086

Ruotong Zhao , Hemin Yuan , Xin Zhang

The micro-scale physical properties of gas hydrate-bearing sediments (GHBS) play a crucial role in elucidating their macro-scale elastic responses, thereby affecting the effectiveness of seismic exploration. Hydrate may have various morphologies in sediments, casting different influences on the elastic properties of GHBS. Various models have been proposed to simulate the hydrates with different morphologies. However, few of them have addressed the generation environment of the different morphologies. In this work, we characterized the elastic properties of GHBS based on the different generation mechanisms using an improved Iso-Frame (IF) model. Based on laboratory observations, we identified different IF values corresponding to P- and S-wave velocities, respectively, reflecting varying influences of hydrate on GHBS elastic properties. Afterwards, we derived the relation between IF_P and IF_S by studying laboratory and well log data statistically for excess-water and excess-gas scenarios, revealing the influences of generation mechanism on GHBS elastic properties. Then these relations were applied on the prediction of S-wave velocity, and the results were compared with the predictions of original IF model and commonly-used hydrate models, which demonstrated that the modified model has improved the Vs prediction. This work highlights the different bulk-shear moduli relations based on the hydrate generation mechanism and provides an alternative route of modeling GHBS, which can facilitate the characterization of GHBS elastic properties.

含天然气水合物沉积物的微尺度物性对阐明其宏观弹性响应起着至关重要的作用，从而影响地震勘探的有效性。水合物在沉积物中可能具有不同的形态，对GHBS的弹性性能产生不同的影响。人们提出了各种模型来模拟不同形态的水合物。然而，很少有人解决了不同形态的生成环境。在这项工作中，我们使用改进的等框架（IF）模型表征了基于不同生成机制的GHBS的弹性特性。基于实验室观测，我们分别确定了P波和s波速度对应的不同IF值，反映了水合物对GHBS弹性性能的不同影响。随后，通过对过量水和过量气情景下的实验室和测井数据进行统计分析，得出了IFP与IFS之间的关系，揭示了生成机制对GHBS弹性性质的影响。将这些关系应用于横波速度预测，并与原中频模型和常用水合物模型的预测结果进行了比较，结果表明，修正模型提高了横波速度预测的准确性。本研究突出了基于水合物生成机制的不同体积-剪切模量关系，为GHBS的建模提供了另一种途径，有助于表征GHBS的弹性特性。

{"title":"Estimation of S-wave velocity of gas hydrate-bearing sediments using an improved Iso-frame model","authors":"Ruotong Zhao , Hemin Yuan , Xin Zhang","doi":"10.1016/j.jappgeo.2025.106086","DOIUrl":"10.1016/j.jappgeo.2025.106086","url":null,"abstract":"<div><div>The micro-scale physical properties of gas hydrate-bearing sediments (GHBS) play a crucial role in elucidating their macro-scale elastic responses, thereby affecting the effectiveness of seismic exploration. Hydrate may have various morphologies in sediments, casting different influences on the elastic properties of GHBS. Various models have been proposed to simulate the hydrates with different morphologies. However, few of them have addressed the generation environment of the different morphologies. In this work, we characterized the elastic properties of GHBS based on the different generation mechanisms using an improved Iso-Frame (IF) model. Based on laboratory observations, we identified different IF values corresponding to P- and S-wave velocities, respectively, reflecting varying influences of hydrate on GHBS elastic properties. Afterwards, we derived the relation between IF<sub>P</sub> and IF<sub>S</sub> by studying laboratory and well log data statistically for excess-water and excess-gas scenarios, revealing the influences of generation mechanism on GHBS elastic properties. Then these relations were applied on the prediction of S-wave velocity, and the results were compared with the predictions of original IF model and commonly-used hydrate models, which demonstrated that the modified model has improved the Vs prediction. This work highlights the different bulk-shear moduli relations based on the hydrate generation mechanism and provides an alternative route of modeling GHBS, which can facilitate the characterization of GHBS elastic properties.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106086"},"PeriodicalIF":2.1,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A new method for constructing dynamic-static conversion models for rock mechanics based on acoustic wave propagation characteristics 基于声波传播特性的岩石力学动静转换模型的新方法

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-29 DOI: 10.1016/j.jappgeo.2025.106083

Sheng Wang , Kaizhou Xu , Haixue Wang , Xiaofei Fu , Chengxuan Ren

To overcome the limitations of conventional dynamic-static parameter conversion methods characterized by low accuracy and destructive core dependency, this study presents a novel non-destructive approach for high-precision dynamic-static mechanical parameter conversion based on acoustic wave propagation characteristics. Considering that the acoustic wave wavelength is much smaller than the measurement scale of the rock body, a stratified mechanical model for heterogeneous rocks was constructed. The bulk density and reciprocal of wave velocity for each sub-sample layer of the rock sample are set to follow normal distributions, with the variance and expected values of these distributions determined by experimental measurements of wave velocity and bulk density in macroscopic rock samples. Under these constraints, target values of layer parameters are determined through a statistical search algorithm, establishing the dynamic-static mechanical parameter conversion model. Taking Bozhong Oilfield reservoir cores as an example, mechanical and acoustic experiments demonstrate that static parameters obtained with the new method show close agreement with test values, yielding average deviations of 0.048 for Young's modulus and 0.066 for Poisson's ratio. Compared with conventional methods, the dynamic-static conversion relationship developed through this new method proves more effective and accurate in estimating stratum static mechanical parameters when applied to well logging interpretation data. The new method enhances core utilization while maintaining accuracy, offering a cost-efficient solution for reservoir mechanical characterization.

针对传统动-静参数转换方法精度低、依赖破坏性核心的局限性，提出了一种基于声波传播特性的高精度动-静力学参数转换的非破坏性方法。考虑到声波波长远小于岩体的测量尺度，建立了非均质岩石的分层力学模型。设置岩样各子样层的容重和波速倒数服从正态分布，通过实验测量宏观岩样的波速和容重来确定这些分布的方差和期望值。在这些约束条件下，通过统计搜索算法确定层参数目标值，建立动-静态力学参数转换模型。以渤中油田储层岩心为例，力学和声学实验表明，新方法得到的静态参数与试验值吻合较好，杨氏模量和泊松比的平均偏差分别为0.048和0.066。与常规方法相比，将该方法建立的动静转换关系应用于测井解释资料中，对地层静态力学参数的估计更加有效和准确。新方法提高了岩心的利用率，同时保持了精度，为油藏力学表征提供了一种经济高效的解决方案。

{"title":"A new method for constructing dynamic-static conversion models for rock mechanics based on acoustic wave propagation characteristics","authors":"Sheng Wang , Kaizhou Xu , Haixue Wang , Xiaofei Fu , Chengxuan Ren","doi":"10.1016/j.jappgeo.2025.106083","DOIUrl":"10.1016/j.jappgeo.2025.106083","url":null,"abstract":"<div><div>To overcome the limitations of conventional dynamic-static parameter conversion methods characterized by low accuracy and destructive core dependency, this study presents a novel non-destructive approach for high-precision dynamic-static mechanical parameter conversion based on acoustic wave propagation characteristics. Considering that the acoustic wave wavelength is much smaller than the measurement scale of the rock body, a stratified mechanical model for heterogeneous rocks was constructed. The bulk density and reciprocal of wave velocity for each sub-sample layer of the rock sample are set to follow normal distributions, with the variance and expected values of these distributions determined by experimental measurements of wave velocity and bulk density in macroscopic rock samples. Under these constraints, target values of layer parameters are determined through a statistical search algorithm, establishing the dynamic-static mechanical parameter conversion model. Taking Bozhong Oilfield reservoir cores as an example, mechanical and acoustic experiments demonstrate that static parameters obtained with the new method show close agreement with test values, yielding average deviations of 0.048 for Young's modulus and 0.066 for Poisson's ratio. Compared with conventional methods, the dynamic-static conversion relationship developed through this new method proves more effective and accurate in estimating stratum static mechanical parameters when applied to well logging interpretation data. The new method enhances core utilization while maintaining accuracy, offering a cost-efficient solution for reservoir mechanical characterization.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106083"},"PeriodicalIF":2.1,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884267","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Induced polarization effects in transient electromagnetic data: A case study from the Hawiah volcanogenic massive sulfide deposit in Saudi Arabia 瞬变电磁数据中的诱导极化效应：以沙特阿拉伯Hawiah火山块状硫化物矿床为例

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-27 DOI: 10.1016/j.jappgeo.2025.106084

Hesham El-Kaliouby , Juntao Lu , Panagiotis Kirmizakis , Abid Khogali , Tim Eatwell , Tomos Bryan , Pantelis Soupios

In exploring volcanogenic massive sulfide (VMS) deposits, the transient electromagnetic (TEM) method is highly effective due to its sensitivity to metallic conductors such as sulfide minerals. A recent TEM survey over the Hawiah VMS deposit in Saudi Arabia utilized ten stations to investigate the distribution of VMS targets. Several stations exhibited anomalous fast decay curves attributed to induced polarization (IP) effects, a common characteristic of VMS minerals. Spectral IP (SIP) measurements of local VMS core samples indicated large IP effects, aligning with results from a previous time domain IP (TDIP) survey in the area. IP effects can distort TEM measurements, causing as fast decay followed by sign reversals, complicating interpretation when using conventional resistivity-only (RO) workflows. These distortions can result in unreliable interpretations, especially in regions with complex subsurface conditions. In this paper, we present a synthetic model analysis based on geological setting information, demonstrating that in scenarios with moderately resistive backgrounds and deeply buried high-polarizable bodies, TEM curves can exhibit anomalous fast decay in the middle to late time, aligning with the observed field data. Thus, this behavior serves as a key indicator of IP effects in the survey area. Through RO and IP-incorporated inversions, along with an uncertainty analysis of the resulting resistivity models, our findings show that the data is better fitted using an IP-incorporated inversion approach. This highlights the importance of analyzing anomalous decay in TEM data and supports adopting refined methodologies that account for IP effects. Such approaches are crucial for achieving accurate and reliable evaluations in areas with highly conductive and polarizable materials, like VMS deposits.

瞬变电磁（TEM）方法对金属导体（如硫化物矿物）的敏感性使其在火山成因块状硫化物矿床勘探中具有很高的效率。最近对沙特阿拉伯Hawiah VMS矿床进行的TEM调查使用了10个站点来调查VMS目标的分布。一些站点显示出异常的快速衰减曲线，这归因于诱导极化（IP）效应，这是VMS矿物的共同特征。对当地VMS岩心样本的频谱IP （SIP）测量表明，存在较大的IP效应，这与之前在该地区进行的时域IP （TDIP）调查结果一致。IP效应会扭曲TEM测量结果，导致快速衰减，然后是符号反转，这使得使用传统的纯电阻率（RO）工作流程进行解释变得复杂。这些扭曲可能导致不可靠的解释，特别是在地下条件复杂的地区。基于地质背景信息的综合模型分析表明，在中等电阻背景和深埋高极化体的情况下，瞬变电磁法曲线在中后期表现出异常的快速衰减，与现场观测数据一致。因此，这种行为可以作为调查区域IP效应的关键指标。通过RO和ip结合反演，以及对所得电阻率模型的不确定性分析，我们的研究结果表明，使用ip结合反演方法可以更好地拟合数据。这突出了分析TEM数据中异常衰减的重要性，并支持采用精细的方法来解释激电效应。这种方法对于在具有高导电性和极化材料的地区（如VMS沉积物）实现准确可靠的评估至关重要。

{"title":"Induced polarization effects in transient electromagnetic data: A case study from the Hawiah volcanogenic massive sulfide deposit in Saudi Arabia","authors":"Hesham El-Kaliouby , Juntao Lu , Panagiotis Kirmizakis , Abid Khogali , Tim Eatwell , Tomos Bryan , Pantelis Soupios","doi":"10.1016/j.jappgeo.2025.106084","DOIUrl":"10.1016/j.jappgeo.2025.106084","url":null,"abstract":"<div><div>In exploring volcanogenic massive sulfide (VMS) deposits, the transient electromagnetic (TEM) method is highly effective due to its sensitivity to metallic conductors such as sulfide minerals. A recent TEM survey over the Hawiah VMS deposit in Saudi Arabia utilized ten stations to investigate the distribution of VMS targets. Several stations exhibited anomalous fast decay curves attributed to induced polarization (IP) effects, a common characteristic of VMS minerals. Spectral IP (SIP) measurements of local VMS core samples indicated large IP effects, aligning with results from a previous time domain IP (TDIP) survey in the area. IP effects can distort TEM measurements, causing as fast decay followed by sign reversals, complicating interpretation when using conventional resistivity-only (RO) workflows. These distortions can result in unreliable interpretations, especially in regions with complex subsurface conditions. In this paper, we present a synthetic model analysis based on geological setting information, demonstrating that in scenarios with moderately resistive backgrounds and deeply buried high-polarizable bodies, TEM curves can exhibit anomalous fast decay in the middle to late time, aligning with the observed field data. Thus, this behavior serves as a key indicator of IP effects in the survey area. Through RO and IP-incorporated inversions, along with an uncertainty analysis of the resulting resistivity models, our findings show that the data is better fitted using an IP-incorporated inversion approach. This highlights the importance of analyzing anomalous decay in TEM data and supports adopting refined methodologies that account for IP effects. Such approaches are crucial for achieving accurate and reliable evaluations in areas with highly conductive and polarizable materials, like VMS deposits.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106084"},"PeriodicalIF":2.1,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Deep learning seismic impedance inversion with CVAE-generated labels and bidirectional self-supervised learning 基于cvae生成标签和双向自监督学习的深度学习地震阻抗反演

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-27 DOI: 10.1016/j.jappgeo.2025.106074

Heng Su, Junxing Cao, Linsen Zhao, Pengfei Jian

In recent years, deep learning has gained increasing attention in the field of geophysics due to its powerful feature extraction and nonlinear modeling capabilities, providing innovative solutions to complex subsurface characterization problems. However, its widespread application remains constrained by two major challenges: limited generalization ability and a heavy reliance on large-scale, high-quality labeled data. Although seismic and well-log data are relatively abundant, accurately paired labeled data remain scarce, which largely limits the application of deep learning in the field of geophysics. To address this problem, we propose a novel label-generation approach based on Conditional Variational Autoencoders (CVAEs), and introduce a novel training framework that integrates supervised and self-supervised learning within a bidirectional learning model. Experiments conducted on both synthetic and real post-stack seismic datasets for impedance inversion demonstrate that the proposed method significantly outperforms baseline method in terms of accuracy and generalization. These results highlight the practicality and robustness of our approach, offering a promising solution to the long-standing challenge of data scarcity in geophysical applications.

近年来，深度学习以其强大的特征提取和非线性建模能力在地球物理领域受到越来越多的关注，为复杂的地下表征问题提供了创新的解决方案。然而，它的广泛应用仍然受到两个主要挑战的制约：有限的泛化能力和对大规模、高质量标记数据的严重依赖。虽然地震和测井数据相对丰富，但准确配对标记的数据仍然很少，这在很大程度上限制了深度学习在地球物理领域的应用。为了解决这个问题，我们提出了一种新的基于条件变分自编码器（CVAEs）的标签生成方法，并引入了一种新的训练框架，该框架在双向学习模型中集成了监督学习和自监督学习。在合成和真实叠后地震数据集上进行的阻抗反演实验表明，该方法在精度和泛化方面明显优于基线方法。这些结果突出了我们方法的实用性和稳健性，为地球物理应用中数据稀缺的长期挑战提供了一个有希望的解决方案。

{"title":"Deep learning seismic impedance inversion with CVAE-generated labels and bidirectional self-supervised learning","authors":"Heng Su, Junxing Cao, Linsen Zhao, Pengfei Jian","doi":"10.1016/j.jappgeo.2025.106074","DOIUrl":"10.1016/j.jappgeo.2025.106074","url":null,"abstract":"<div><div>In recent years, deep learning has gained increasing attention in the field of geophysics due to its powerful feature extraction and nonlinear modeling capabilities, providing innovative solutions to complex subsurface characterization problems. However, its widespread application remains constrained by two major challenges: limited generalization ability and a heavy reliance on large-scale, high-quality labeled data. Although seismic and well-log data are relatively abundant, accurately paired labeled data remain scarce, which largely limits the application of deep learning in the field of geophysics. To address this problem, we propose a novel label-generation approach based on Conditional Variational Autoencoders (CVAEs), and introduce a novel training framework that integrates supervised and self-supervised learning within a bidirectional learning model. Experiments conducted on both synthetic and real post-stack seismic datasets for impedance inversion demonstrate that the proposed method significantly outperforms baseline method in terms of accuracy and generalization. These results highlight the practicality and robustness of our approach, offering a promising solution to the long-standing challenge of data scarcity in geophysical applications.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106074"},"PeriodicalIF":2.1,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Error calibration for cross-shaped magnetic gradient tensor system based on the improved differential evolution algorithm 基于改进差分进化算法的十字形磁梯度张量系统误差标定

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-26 DOI: 10.1016/j.jappgeo.2025.106080

Chenxu Dong, Zhuoxuan Li, Yuguo Li, Xuezhen Ding

Errors arising from sensor manufacturing and misalignment during installation have been shown to have a significant impact on the measurement accuracy of magnetic gradient tensor systems. This, in turn, can result in a reduction in detection performance in practical applications such as magnetic anomaly detection. This study proposes a high-precision error calibration method for a cross-shaped magnetic gradient tensor system. This method uses the Success-History based Adaptive Differential Evolution with Linear Population Size Reduction (L-SHADE) algorithm, an enhanced variant of the Differential Evolution (DE) algorithm, to estimate these error parameters. A unified optimization model has been developed that incorporates multiple sensor errors into a single calibration framework, thereby enabling one-step error correction. Simulation experiments are conducted under both conditions of absence of noise and conditions of presence of noise. In conditions of absence of noise, the discrepancy between the calibrated and theoretical total magnetic field is of the order of

10^{- 6}

nT, with all magnetic gradient tensor components and invariants tending toward zero. In field experiments, the maximum deviation between the calibrated total magnetic field and the actual geomagnetic field is 1.92 nT, and the maximum improvement ratio of tensor components reaches 979. The findings from both simulation and field trials have been shown to demonstrate that the proposed algorithm successfully attains high calibration accuracy and computational efficiency, thus providing a pragmatic approach for sensor error correction.

由于传感器的制造和安装过程中的不对准引起的误差对磁梯度张量系统的测量精度有很大的影响。这反过来又会导致磁异常检测等实际应用中检测性能的降低。提出了一种十字形磁梯度张量系统的高精度误差标定方法。该方法采用基于成功历史的线性种群大小缩减自适应差分进化（L-SHADE）算法，该算法是差分进化（DE）算法的一种增强变体，用于估计这些误差参数。已经开发了一个统一的优化模型，该模型将多个传感器误差合并到单个校准框架中，从而实现一步误差校正。在无噪声和有噪声两种情况下进行了仿真实验。在无噪声条件下，标定总磁场与理论总磁场之间的差异约为10−6 nT，所有磁梯度张量分量和不变量都趋于零。在现场实验中，标定总磁场与实际地磁场的最大偏差为1.92 nT，张量分量的最大改进比达到979。仿真和现场试验结果表明，该算法成功地实现了较高的校准精度和计算效率，为传感器误差校正提供了一种实用的方法。

{"title":"Error calibration for cross-shaped magnetic gradient tensor system based on the improved differential evolution algorithm","authors":"Chenxu Dong, Zhuoxuan Li, Yuguo Li, Xuezhen Ding","doi":"10.1016/j.jappgeo.2025.106080","DOIUrl":"10.1016/j.jappgeo.2025.106080","url":null,"abstract":"<div><div>Errors arising from sensor manufacturing and misalignment during installation have been shown to have a significant impact on the measurement accuracy of magnetic gradient tensor systems. This, in turn, can result in a reduction in detection performance in practical applications such as magnetic anomaly detection. This study proposes a high-precision error calibration method for a cross-shaped magnetic gradient tensor system. This method uses the Success-History based Adaptive Differential Evolution with Linear Population Size Reduction (L-SHADE) algorithm, an enhanced variant of the Differential Evolution (DE) algorithm, to estimate these error parameters. A unified optimization model has been developed that incorporates multiple sensor errors into a single calibration framework, thereby enabling one-step error correction. Simulation experiments are conducted under both conditions of absence of noise and conditions of presence of noise. In conditions of absence of noise, the discrepancy between the calibrated and theoretical total magnetic field is of the order of <span><math><msup><mn>10</mn><mrow><mo>−</mo><mn>6</mn></mrow></msup></math></span> nT, with all magnetic gradient tensor components and invariants tending toward zero. In field experiments, the maximum deviation between the calibrated total magnetic field and the actual geomagnetic field is 1.92 nT, and the maximum improvement ratio of tensor components reaches 979. The findings from both simulation and field trials have been shown to demonstrate that the proposed algorithm successfully attains high calibration accuracy and computational efficiency, thus providing a pragmatic approach for sensor error correction.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106080"},"PeriodicalIF":2.1,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Microseismic identification and effectiveness assessment of hydraulic-fracturing–induced roof cutting using an STFT–CNN framework 基于STFT-CNN框架的水力压裂顶板切割微震识别与有效性评估

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-26 DOI: 10.1016/j.jappgeo.2025.106078

Haowei Tian , Zhizhong Jiang , Zhenqian Ma , Zhijie Wen , Shaojie Zuo , Yu Liu , Jinhui Li , Wenjian Wang , Mingying Wang , Youchi Jin

During underground mining of solid mineral deposits, hard roofs can induce large-scale strata pressure disasters, posing severe threats to mine safety. Hydraulic-fracturing-based roof-cutting technology offers a means of weakening hard roofs by severing key structural connections, thereby altering the roof structure and reducing the likelihood of roof-related accidents. However, limitations remain in accurately characterizing fracture propagation and evaluating the effectiveness of hydraulic fracturing roof cutting. In this study, a coal mine in southwestern China was selected as the engineering site. By integrating hydraulic-fracturing roof cutting with microseismic monitoring, a hydraulic-fracturing microseismic event recognition model based on the Short-Time Fourier Transform (STFT) and Convolutional Neural Networks (CNN) was developed. Time-frequency analysis revealed that different microseismic signal types exhibit distinct dominant frequency ranges, with hydraulic-fracturing signals concentrated at 130–200 Hz. The proposed STFT–CNN model achieved a recognition accuracy exceeding 92 %. Microseismic source-location and kernel density analyses indicated that hydraulic-fracturing-induced fractures propagated symmetrically along the borehole axis, with an effective influence range of approximately ±5 m horizontally and ± 10 m vertically, and some fractures extending downward into the coal seam. Analysis of microseismic energy evolution showed abrupt energy surges when the working face advanced 1–2 m past the roof-cutting boreholes, with more than three high-magnitude events typically occurring about 3 m behind the working face. Overall, the field results demonstrate that hydraulic fracturing effectively weakened the hard roof and successfully facilitated controlled roof collapse, thereby enhancing the safety of underground coal extraction.

在固体矿床地下开采过程中，硬顶板会诱发大规模的地压灾害，对矿山安全构成严重威胁。基于水力压裂的顶板切割技术提供了一种通过切断关键结构连接来削弱硬顶板的方法，从而改变顶板结构，降低与顶板有关的事故发生的可能性。然而，在准确表征裂缝扩展和评估水力压裂顶板切割效果方面仍然存在局限性。本研究选择西南某煤矿作为工程场地。将水力压裂顶板切割与微地震监测相结合，建立了一种基于短时傅里叶变换（STFT）和卷积神经网络（CNN）的水力压裂微地震事件识别模型。时频分析表明，不同微震信号类型具有不同的主导频率范围，水力压裂信号集中在130 ~ 200 Hz。所提出的STFT-CNN模型的识别准确率超过92%。微震震源定位和核密度分析表明，水力压裂裂缝沿井轴线对称扩展，有效影响范围约为水平方向±5 m，垂直方向±10 m，部分裂缝向下延伸至煤层。微震能量演化分析表明，工作面向切顶钻孔推进1 ~ 2 m时能量突变，工作面后方约3 m处通常发生3次以上的高震级事件。综上所述，现场结果表明，水力压裂有效地削弱了硬顶板，成功地促进了顶板的可控坍塌，从而提高了井下采煤的安全性。

{"title":"Microseismic identification and effectiveness assessment of hydraulic-fracturing–induced roof cutting using an STFT–CNN framework","authors":"Haowei Tian , Zhizhong Jiang , Zhenqian Ma , Zhijie Wen , Shaojie Zuo , Yu Liu , Jinhui Li , Wenjian Wang , Mingying Wang , Youchi Jin","doi":"10.1016/j.jappgeo.2025.106078","DOIUrl":"10.1016/j.jappgeo.2025.106078","url":null,"abstract":"<div><div>During underground mining of solid mineral deposits, hard roofs can induce large-scale strata pressure disasters, posing severe threats to mine safety. Hydraulic-fracturing-based roof-cutting technology offers a means of weakening hard roofs by severing key structural connections, thereby altering the roof structure and reducing the likelihood of roof-related accidents. However, limitations remain in accurately characterizing fracture propagation and evaluating the effectiveness of hydraulic fracturing roof cutting. In this study, a coal mine in southwestern China was selected as the engineering site. By integrating hydraulic-fracturing roof cutting with microseismic monitoring, a hydraulic-fracturing microseismic event recognition model based on the Short-Time Fourier Transform (STFT) and Convolutional Neural Networks (CNN) was developed. Time-frequency analysis revealed that different microseismic signal types exhibit distinct dominant frequency ranges, with hydraulic-fracturing signals concentrated at 130–200 Hz. The proposed STFT–CNN model achieved a recognition accuracy exceeding 92 %. Microseismic source-location and kernel density analyses indicated that hydraulic-fracturing-induced fractures propagated symmetrically along the borehole axis, with an effective influence range of approximately ±5 m horizontally and ± 10 m vertically, and some fractures extending downward into the coal seam. Analysis of microseismic energy evolution showed abrupt energy surges when the working face advanced 1–2 m past the roof-cutting boreholes, with more than three high-magnitude events typically occurring about 3 m behind the working face. Overall, the field results demonstrate that hydraulic fracturing effectively weakened the hard roof and successfully facilitated controlled roof collapse, thereby enhancing the safety of underground coal extraction.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"246 ","pages":"Article 106078"},"PeriodicalIF":2.1,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928775","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Acoustic emission responses and failure characteristics of rocks with varying rockburst tendencies under uniaxial loading 单轴加载下不同岩爆倾向岩石的声发射响应及破坏特征

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-26 DOI: 10.1016/j.jappgeo.2025.106082

Xiaojun Wang , He Zhang , Xiao Feng , Qinglin Chen , Shirong Cao , Haowen Jiang , Jian liu

Understanding damage mechanisms in rocks exhibiting different rockburst tendencies is critical for monitoring rock damage at varied risk levels via acoustic emission (AE) technology. This study conducted rockburst tendency evaluation and uniaxial compression AE tests on limestone, granite, and red sandstone. Metallographic imaging and scanning electron microscopy (SEM) tests were also performed on fracture surfaces. AE responses and failure characteristics of rocks with different rockburst tendencies were analyzed. The parameter r (RA/AF) served as a rock damage index, with its Coefficient of Variation (CV) and AE b-value calculated. The results show that AE ring-down counts have a period of decline during the unstable crack propagation stage. Limestone and granite with rockburst tendencies fail under tension-shear coupling with mutual transitions between failure modes. The CV(r) shows a gradual decrease, stabilization, and significant increase. Red sandstone without rockburst tendencies mainly undergoes tensile action, and its CV(r) remains stable after decreasing. Compared with the traditional b-value, the CV(r) more effectively identifies failure progression, and its abrupt surge serves as a precursor for failure in rocks possessing rockburst tendency.

了解具有不同岩爆倾向的岩石的损伤机制对于利用声发射技术监测不同危险水平的岩石损伤至关重要。研究对石灰岩、花岗岩和红砂岩进行了岩爆倾向评价和单轴压缩声发射试验。对断口表面进行了金相成像和扫描电镜（SEM）测试。分析了不同岩爆倾向岩石的声发射响应和破坏特征。参数r （RA/AF）作为岩石损伤指标，计算其变异系数（CV）和AE b值。结果表明，在不稳定裂纹扩展阶段，声发射衰响次数呈下降趋势。具有岩爆倾向的灰岩和花岗岩在拉剪耦合作用下破坏，破坏模式相互转换。CV(r)呈逐渐下降、稳定和显著上升的趋势。无冲击地压倾向的红砂岩主要受拉伸作用，其CV(r)减小后保持稳定。与传统的b值相比，CV(r)能更有效地识别破坏进程，其突变波动是具有岩爆倾向的岩石破坏的前兆。

{"title":"Acoustic emission responses and failure characteristics of rocks with varying rockburst tendencies under uniaxial loading","authors":"Xiaojun Wang , He Zhang , Xiao Feng , Qinglin Chen , Shirong Cao , Haowen Jiang , Jian liu","doi":"10.1016/j.jappgeo.2025.106082","DOIUrl":"10.1016/j.jappgeo.2025.106082","url":null,"abstract":"<div><div>Understanding damage mechanisms in rocks exhibiting different rockburst tendencies is critical for monitoring rock damage at varied risk levels via acoustic emission (AE) technology. This study conducted rockburst tendency evaluation and uniaxial compression AE tests on limestone, granite, and red sandstone. Metallographic imaging and scanning electron microscopy (SEM) tests were also performed on fracture surfaces. AE responses and failure characteristics of rocks with different rockburst tendencies were analyzed. The parameter r (RA/AF) served as a rock damage index, with its Coefficient of Variation (CV) and AE b-value calculated. The results show that AE ring-down counts have a period of decline during the unstable crack propagation stage. Limestone and granite with rockburst tendencies fail under tension-shear coupling with mutual transitions between failure modes. The CV(r) shows a gradual decrease, stabilization, and significant increase. Red sandstone without rockburst tendencies mainly undergoes tensile action, and its CV(r) remains stable after decreasing. Compared with the traditional b-value, the CV(r) more effectively identifies failure progression, and its abrupt surge serves as a precursor for failure in rocks possessing rockburst tendency.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106082"},"PeriodicalIF":2.1,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145924829","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Geophysical survey methods (GPR and ERT) to find architectural remains from the 17th century at the Fort of San Diego in Acapulco, Mexico. A case study. 地球物理测量方法（GPR和ERT）在墨西哥阿卡普尔科的圣地亚哥堡发现了17世纪的建筑遗迹。案例研究。

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-25 DOI: 10.1016/j.jappgeo.2025.106079

J. Ortega-Ramirez , M. Bano , J.L. Salas-Corrales , R. Junco Sánchez , L.A. Villa-Alvarado

Fort San Diego in Acapulco, Mexico, is an iconic monument, deeply linked to the history of the continent and a vital source of cultural identity for its community and future generations. Given its immense value as a cultural asset, it is essential to understand its architectural evolution, especially as historical records indicate significant alterations due to seismic activity and changes of use over time.

The article presents a geophysical study with the objective of locating and documenting the hidden architectural remains of the fort constructed in the 17th century. Given the paucity of documentation on the fort's modifications, we used non-destructive methods such as georadar (GPR) and electrical resistivity tomography (ERT). Both techniques identified a large anomaly measuring 3 by 6 m beneath the surface of the fort. This anomaly, characterized by multiple GPR diffractions and high electrical resistivity values, was then validated by a small archaeological excavation. The excavation confirmed that the anomaly corresponded to an ancient architectural foundation, visible from a depth of 30 cm down to at least 2.0 m. We hypothesize that this structure represents the remains of a drawbridge that served as the main entrance to the fort before the devastating earthquake of 1776, supporting the theory that the main gate was located on the opposite side to the current one. The study highlights the effectiveness and versatility of geophysical methods as essential tools for the investigation and conservation of cultural heritage, revealing crucial details about the hidden history of the fort.

位于墨西哥阿卡普尔科的圣地亚哥堡是一座标志性的纪念碑，与美洲大陆的历史紧密相连，是其社区和子孙后代文化认同的重要来源。鉴于其作为文化资产的巨大价值，了解其建筑演变是至关重要的，特别是历史记录表明，由于地震活动和使用的变化，随着时间的推移发生了重大变化。本文介绍了一项地球物理研究，目的是定位和记录17世纪建造的堡垒的隐藏建筑遗迹。鉴于缺乏关于堡垒改造的文件，我们使用了非破坏性方法，如地质雷达（GPR）和电阻率层析成像（ERT）。两种技术都发现了一个巨大的异常，在堡垒表面下3米乘6米。该异常具有多个GPR衍射和高电阻率值的特征，随后通过小型考古挖掘进行了验证。挖掘证实，这个异常对应于一个古老的建筑基础，从30厘米到至少2.0米的深度都可以看到。我们假设这个结构代表了1776年毁灭性地震之前作为堡垒主要入口的吊桥的遗迹，支持了主要大门位于当前大门对面的理论。这项研究突出了地球物理方法作为调查和保护文化遗产的重要工具的有效性和多功能性，揭示了有关堡垒隐藏历史的重要细节。

{"title":"Geophysical survey methods (GPR and ERT) to find architectural remains from the 17th century at the Fort of San Diego in Acapulco, Mexico. A case study.","authors":"J. Ortega-Ramirez , M. Bano , J.L. Salas-Corrales , R. Junco Sánchez , L.A. Villa-Alvarado","doi":"10.1016/j.jappgeo.2025.106079","DOIUrl":"10.1016/j.jappgeo.2025.106079","url":null,"abstract":"<div><div>Fort San Diego in Acapulco, Mexico, is an iconic monument, deeply linked to the history of the continent and a vital source of cultural identity for its community and future generations. Given its immense value as a cultural asset, it is essential to understand its architectural evolution, especially as historical records indicate significant alterations due to seismic activity and changes of use over time.</div><div>The article presents a geophysical study with the objective of locating and documenting the hidden architectural remains of the fort constructed in the 17th century. Given the paucity of documentation on the fort's modifications, we used non-destructive methods such as georadar (GPR) and electrical resistivity tomography (ERT). Both techniques identified a large anomaly measuring 3 by 6 m beneath the surface of the fort. This anomaly, characterized by multiple GPR diffractions and high electrical resistivity values, was then validated by a small archaeological excavation. The excavation confirmed that the anomaly corresponded to an ancient architectural foundation, visible from a depth of 30 cm down to at least 2.0 m. We hypothesize that this structure represents the remains of a drawbridge that served as the main entrance to the fort before the devastating earthquake of 1776, supporting the theory that the main gate was located on the opposite side to the current one. The study highlights the effectiveness and versatility of geophysical methods as essential tools for the investigation and conservation of cultural heritage, revealing crucial details about the hidden history of the fort.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"245 ","pages":"Article 106079"},"PeriodicalIF":2.1,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884270","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A data-driven approach for identification of coal-related lithofacies using single and meta-learner ensemble classifiers from well-log data: A case study from Sohagpur coal field, India 利用测井数据中的单学习器和元学习器集成分类器识别煤相关岩相的数据驱动方法：以印度Sohagpur煤田为例

IF 2.1 3区地球科学 Q2 GEOSCIENCES, MULTIDISCIPLINARY

Journal of Applied Geophysics

Pub Date : 2025-12-25 DOI: 10.1016/j.jappgeo.2025.106073

Rupam Roy , Dip Kumar Singha , Sayan Ghosh , Laraib Abbas , Debjeet Mondal

This study focuses on building a set of classification models based on single Machine Learning (ML) classifiers, followed by utilization of the trained single classifiers to construct optimized homogeneous, heterogeneous, and stacked ensemble learners to predict coal, carbonaceous shale, and non-coal lithofacies solely based on a set of high resolution conventional well-log data. A total of 6 lithofacies were considered as classes set as the target variable, along with 5 variables as input features obtained from well-log data from 3 wells (well-1, well-2, and well-3) to construct the training data. A couple of wells (well-4, 5) were utilized as blind testing wells to evaluate all 12 classification models, one of them (well-4) having a true litholog. A high inherent imbalance was observed in the class distribution for both the training as well as blind testing datasets. The imbalance issue was resolved by utilizing the class weight parameter assigned to certain classifiers and randomly removing the dominant (the sandstone) class through a random under-sampling operation in Python. The trained single classifiers, despite having overall good performance on blind datasets, were poor at identifying the coal, sandy shale, shaly coal, and carbshale. However, the trained heterogeneous ensemble is proven to be the best classifier among the 12, both class-wise and in terms of overall accuracy (more than 90 %), and the stacked ensemble learner is the second best, as far as coal, carbshale, and shaly coal are concerned. In the case of the homogeneous ensemble learners, the class-wise prediction performance has improved considerably, and all the homogeneous ensemble learners have performed better than their respective single classifier counterparts. This study successfully verified the usefulness of stacked, homogeneous, and heterogeneous ensemble meta learners over the single classifier models, for coal, non-coal, and carbonaceous lithofacies identification, both class-wise and in an overall manner.

本研究的重点是建立一套基于单个机器学习（ML）分类器的分类模型，然后利用训练好的单个分类器构建优化的同质、异质和堆叠集成学习器，仅基于一组高分辨率常规测井数据预测煤、碳质页岩和非煤岩相。将6个岩相作为类集作为目标变量，将5个变量作为从3口井（井1、井2、井3）测井数据中获得的输入特征，构建训练数据。几口井（井4和井5）被用作盲测井，以评估所有12种分类模型，其中一口井（井4）具有真实的岩性。在训练和盲测数据集的类分布中观察到高度固有的不平衡。通过利用分配给某些分类器的类权重参数，并通过Python中的随机欠采样操作随机删除主导（砂岩）类，解决了不平衡问题。尽管训练的单一分类器在盲数据集上具有良好的总体性能，但在识别煤、砂质页岩、页岩煤和碳页岩方面表现不佳。然而，经过训练的异构集成被证明是12个分类器中最好的分类器，无论是在类别方面还是在总体准确率方面（超过90%），而堆叠集成学习器在煤、碳页岩和页岩煤方面排名第二。在同质集成学习器的情况下，类预测性能有了很大的提高，并且所有同质集成学习器都比它们各自的单分类器表现得更好。该研究成功地验证了堆叠、均匀和异构集成元学习器在单一分类器模型上对煤、非煤和碳质岩相识别的有用性，无论是分类还是整体方式。

{"title":"A data-driven approach for identification of coal-related lithofacies using single and meta-learner ensemble classifiers from well-log data: A case study from Sohagpur coal field, India","authors":"Rupam Roy , Dip Kumar Singha , Sayan Ghosh , Laraib Abbas , Debjeet Mondal","doi":"10.1016/j.jappgeo.2025.106073","DOIUrl":"10.1016/j.jappgeo.2025.106073","url":null,"abstract":"<div><div>This study focuses on building a set of classification models based on single Machine Learning (ML) classifiers, followed by utilization of the trained single classifiers to construct optimized homogeneous, heterogeneous, and stacked ensemble learners to predict coal, carbonaceous shale, and non-coal lithofacies solely based on a set of high resolution conventional well-log data. A total of 6 lithofacies were considered as classes set as the target variable, along with 5 variables as input features obtained from well-log data from 3 wells (well-1, well-2, and well-3) to construct the training data. A couple of wells (well-4, 5) were utilized as blind testing wells to evaluate all 12 classification models, one of them (well-4) having a true litholog. A high inherent imbalance was observed in the class distribution for both the training as well as blind testing datasets. The imbalance issue was resolved by utilizing the class weight parameter assigned to certain classifiers and randomly removing the dominant (the sandstone) class through a random under-sampling operation in Python. The trained single classifiers, despite having overall good performance on blind datasets, were poor at identifying the coal, sandy shale, shaly coal, and carbshale. However, the trained heterogeneous ensemble is proven to be the best classifier among the 12, both class-wise and in terms of overall accuracy (more than 90 %), and the stacked ensemble learner is the second best, as far as coal, carbshale, and shaly coal are concerned. In the case of the homogeneous ensemble learners, the class-wise prediction performance has improved considerably, and all the homogeneous ensemble learners have performed better than their respective single classifier counterparts. This study successfully verified the usefulness of stacked, homogeneous, and heterogeneous ensemble meta learners over the single classifier models, for coal, non-coal, and carbonaceous lithofacies identification, both class-wise and in an overall manner.</div></div>","PeriodicalId":54882,"journal":{"name":"Journal of Applied Geophysics","volume":"246 ","pages":"Article 106073"},"PeriodicalIF":2.1,"publicationDate":"2025-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0