Search

The integration of deep learning techniques from the field of image generation into digital rock analysis has resulted in substantial advancements. However, generating high-quality and highly controllable heterogeneous digital rocks still presents challenges and necessities. This paper reports a digital rock reconstruction method using a multi-condition denoising diffusion probabilistic model (MCDDPM). This method has the diffusion and denoising processes inherent in the diffusion model, ensuring high-quality reconstructed samples. Meanwhile, various physical property parameters are embedded as conditions in the network layers of the model at multiple scales, enhancing the conditional control ability of this method. We compared the porosity, pore size distribution, and two-point correlation coefficient under the same conditions, as well as the evaluation indicators of different methods. The comparative results provide convincing evidence for the accuracy of our proposed digital rock reconstruction method. In addition, studies on the control effect of different conditions on reconstructed samples also demonstrate the effectiveness of our strategy. Therefore, MCDDPM can reconstruct reasonable heterogeneous digital rocks based on combinations of physical property conditions, which provides effective technical support for expanding the library of heterogeneous digital rock samples and even achieving zero-shot learning of digital rocks. • The real numerical rock possesses strong heterogeneity, and the method proposed in this paper can effectively reconstruct heterogeneous digital rock. • Compared to other deep learning network models, the reconstruction effect of the method proposed in this paper is at a high level. • The method we propose supports multiple physical property conditions as input. [ABSTRACT FROM AUTHOR]

Frequency–time conversion is a crucial step in grounded electrical-source transient electromagnetic response calculation, and the performance of the algorithm is directly related to the overall accuracy and speed of forward modeling. In mainstream algorithms, algorithms with high accuracy often have slow computation speed while algorithms with high efficiency have unsatisfactory accuracy, especially when facing inversion problems that are difficult to meet requirements. This paper introduces three inverse Laplace transform algorithms for this problem: the Gaver–Stehfest algorithm, the Euler algorithm, and the Talbot algorithm. The performance of each algorithm in forward modeling was analyzed using half-space and layered models, and the optimal selection schemes for algorithm weight coefficients were provided. The numerical calculation results show that the Gaver–Stehfest algorithm has a unique advantage in computational efficiency, while the Talbot algorithm and Euler algorithm meet the accuracy requirements. After considering both accuracy and efficiency, the Talbot algorithm is selected to replace conventional algorithms for calculation of grounded electrical-source transient electromagnetic forward modeling. In addition, this paper combines the characteristics of the Gaver–Stehfest algorithm and the Talbot algorithm to implement an adaptive hybrid algorithm. This algorithm uses the Gaver–Stehfest algorithm for forward modeling in the early times and the Talbot algorithm to compensate for the decrease in accuracy in the later times. Through the comparison of forward modeling calculations, it can be seen that the hybrid algorithm proposed in this paper fully utilizes the advantages of both algorithms. The hybrid algorithm greatly improves computational speed while meeting accuracy requirements, and has significant advantages over conventional algorithms. • Compared and analyzed the characteristics of three inverse Laplace transform algorithms. • The optimal selection for the number of nodes with inverse Laplace transform algorithms coefficients is provided. • The Talbot algorithm has been selected as the most suitable algorithm for forward modeling of grounded electrical-source TEM. • Proposed an adaptive hybrid algorithm and verified its advantages. [ABSTRACT FROM AUTHOR]

3D finite-difference time-domain numerical simulation and reconstruction based on the domain decomposition technique are essential parts of high-performance computation for reverse-time migration and full-waveform inversion. However, the low GPU utilization in computing for small-sized models and the tremendous memory consumption for large-sized models may result in low computational efficiency and high memory costs. This paper proposes a contiguous memory management (CMM) method and a variable-order wavefield reconstruction (VWR) method. The CMM allocates the memory of many small-sized arrays used for MPI communications on a larger-sized contiguous memory block, which aims to reduce the number of MPI communications between subdomains and improve the communication bandwidth, thus reducing the MPI time overhead and improving the GPU utilization. Meanwhile, the VWR can flexibly set the number of layers of boundary wavefield used for source wavefield reconstruction according to the host memory capacity and accuracy requirements. Since one layer of boundary wavefield could be stored using the VWR, the memory consumption of host memory can be significantly alleviated. Numerical experiments show that GPU utilization in computing for the model with a size of 12 1 3 can be improved from 25% to 90% using the CMM method, and the VWR method can reduce memory consumption by about 86% while maintaining good accuracy in wavefield reconstruction. In addition, the issue of how to obtain a domain decomposition scheme with optimal performance is discussed in this paper. • A contiguous memory management strategy is proposed for improving the efficiency of MPI communication. • A variable-order wavefield reconstruction strategy is developed to reduce memory consumption. • The issue of how to obtain a domain decomposition scheme with optimal performance is discussed. [ABSTRACT FROM AUTHOR]

Spike is a type of abnormal waveform observed in strong motion records. It is characterized by a sudden and significant increase in peak acceleration, exceeding the theoretical range, and being greatly uncoordinated with the surrounding record points. The spike severely affects the reliability of subsequent analysis of strong motion records. This paper conducts research based on the strong motion records of the Turkey-Syria earthquakes in 2023. It focuses on the spike represented as Peak Ground Acceleration (PGA) and proposes a spike recognition model for strong motion records using Bayesian-optimized Support Vector Machine (SVM). Firstly, a preprocessing method for strong motion records based on adaptive waveform scaling is proposed. This method enables the adaptive normalization of strong motion records with different scales, thereby eliminating the influence of different scales on the accuracy of manual annotations. Furthermore, a new method for characterizing strong motion records is proposed. It captures shape features (two-dimensional) through one-dimensional data, enabling the spatial distribution of strong motion records to be represented by a shape feature vector. Finally, a Bayesian optimization-based improved SVM model is proposed. It utilizes the feature extraction method proposed in this paper for the identification of spikes in strong motion records. The experimental results show that the proposed model in this paper achieved excellent performance, with an F1-score of 0.868 and an AUC value of 0.93 on the test set. • A normalization method for strong motion records at different scales is proposed. • Proposed a numerical computation-based method for spike feature extraction. • Developed a spike identification model based on Support Vector Machine. • Optimized our model's hyperparameters using bayesian optimization. • Achieved good results on the dataset of Turkey-Syria earthquakes. [ABSTRACT FROM AUTHOR]

Geological reports house a wealth of geological domain knowledge and expert experience knowledge, and the efficient extraction of geological entity relations from these texts holds great significance in constructing a geological knowledge graph. To address challenges such as overlapping entity relations, data imbalance, and the limited incorporation of contextual features in existing deep learning-based entity relation extraction methods, this paper proposes a relational graph convolutional network pointer model based on the pre-trained model of Robustly Optimize Bidirectional Encoder Representation from Transformers Pre-training Approach (RoBERTa). Specifically, our approach tackles the issue of overlapping entity relationships by identifying the primary entity and the subordinate entity associated with a specific relationship. Additionally, by incorporating a relationship graph convolutional network, we can capture multiple relationships among entities. Moreover, we leverage the Biaffine layer to accurately handle entity annotation boundaries, thereby enhancing the precision of entity classification. To address the issue of imbalanced data, we adopt the Focal Loss function during experiments. Furthermore, we construct a geological entity relation dataset comprising 24 types of relationships, which is generated by collecting seven regional geological reports and employing semi-automatic annotation techniques. Finally, the proposed algorithm was tested and comparison with several baseline models, the results demonstrate the superior performance of the proposed model in this paper. • In this paper, a geological entity relationship dataset is constructed. • In this paper, we propose a convolutional pointer network model for relational maps based on the pre-trained RoBERTa model and validate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

The particle-tracking method for transporting radionuclides in multicomponent ocean medium (water and multifractional suspended and deposited sediments) is considered using a probabilistic approach for simulating interaction processes between several states of radioactivity. The state transformations as a result of reactions of the first order were described using the master equation for the probability of the particle being in the given state. Transition probabilities between all possible states can be obtained from the numerical solution of the matrix master equation that is derived in the paper. In the first approximation, the Euler method was used to obtain a solution for the next time step. This approach can be applied to any linear system of equations describing phase transitions with any number of states, but it requires small values of the transition probabilities to ensure only a single-phase change during one time step. The paper also focuses on deriving the Lagrangian interface conditions between the water column and bottom deposition. To apply the probabilistic approach, the boundary conditions were considered as the reaction terms in a thin near-bottom interface layer in which boundary conditions were converted into source terms. For this layer, the corresponding master equation was derived to obtain transitional probabilities for particle states. The developed approaches were tested on numerical and analytical solutions of two test cases. It was found that the optimal thickness of the interface layer must be larger than the maximum vertical displacement of the particle during the one-time step, but it must be small enough to approximate the condition of uniform distribution of concentration in this layer. • Particle-tracking method for the transport of radionuclides in ocean is considered. • Master equation is derived to describe all possible states of contaminant. • Bottom boundary conditions were considered as the reaction terms in interface layer. • Optimal thickness of the interface layer was chosen. • The approach was tested on a numerical and analytical solutions for two test cases. [ABSTRACT FROM AUTHOR]

To address the issues surrounding incomplete coverage of core dictionaries, limited training corpora, and low precision in Chinese geological text segmentation, a knowledge- and data-driven word segmentation method by combining combination probability and machine learning was proposed in this paper. We extracted mathematical feature information from terms in Chinese geological text to construct a Term Combination Probability Model (TCPM) for Chinese word combinations by integrating the combination features of geological terms and the Chinese writing styles under zero-sample conditions. The TCPM was used to extract geological terms with high combination characteristics as a user-defined dictionary, and then a geological corpus was constructed by using a general domain word segmentation method based on this dictionary. After a small amount of manual review and optimization, the geological corpus was trained with a BiLSTM-CRF model to segment Chinese geological text. The proposed method in this paper was tested using a regional geological survey report set in Henan Province, and the precision, recall, and F1-score of the method are 92.65%, 92.53%, and 92.59%, respectively. The experimental results demonstrated that this method, combined with the inherent knowledge features of geological text and the machine learning method, can assist in expanding the core dictionary for Chinese geological text segmentation based on zero-sample, and can improve the segmentation precision of Chinese geological text compared to simple general word segmentation methods or machine learning methods. • TCPM with no sample can reflect the structural features from professional terms in geological field texts. • Chinese word segmentation with custom dictionary from terms extracted by TCPM is better than general segmentation methods. • Combining TCPM with machine learning can improve the accuracy in word segmentation for Chinese geological text. [ABSTRACT FROM AUTHOR]

Seismic data visualization in 3D space is a valuable interpretation tool. Several open-source visualization tools are available. However, little explanation is provided about the inner working of visualization process. The current work discusses a "hello world" equivalent source code for 3D seismic data visualization using Graphical Processing Units (GPUs) with OpenGL and the OpenGL Shading Language (GLSL) programming languages. Rendering is the core process generating 2D image that we see on the screen from the 3D data structures being visualized. Texture mapping-based rendering commonly applied to seismic data starts with creating an OpenGL object called texture. The texture is then mapped over a rectangular object to display a seismic line. The work presented here is performed using OpenGL, GLSL, C++ and Qt toolkit. Here Qt provides the application GUI framework, C++ is used for data I/O, filtering, and sorting, and OpenGL and GLSL are used for 3D rendering. This paper describes the data flow through the application, and two implementations of vertex and fragment GLSL shaders. Visualization is critical for seismic data processing and interpretation. The low-level details of the process presented here will hopefully help readers in obtaining better understanding of the visualization concept and inner working principle and facilitate further development. • Low-level structure of seismic visualization explained in detail with code. • Texture and shader notions are discussed in application to 3D seismic visualization. • "Hello world" equivalent source code for 3D seismic data visualization. [ABSTRACT FROM AUTHOR]

Due to rapid urbanization, rising food demand, and changed precipitation patterns, the waterbodies are contracting their former beds. The continuous shrinking of waterbodies is deteriorating the vital cultural, supporting, provisioning, and regulating services. Thus, understanding and mitigating the impacts of streambed land cover change is crucial for maintaining healthy aquatic ecosystems and improving flood resilience of surrounding population. The existing works use high-resolution aerial imagery focusing on large waterbodies, while ignoring the most vulnerable floodplains of innumerous small water bodies due to high inter-class similarity. This limits the ability to perform a temporal analysis of land cover change along small water bodies. The present work aims to resolve this issue using open-source satellite imagery and taking patched samples along the boundary of small water bodies to identify long-term changes in land cover patterns. Sentinel-2 and Landsat 50 acquired satellite images were used to identify the land cover of this colonized stream bed. The data of Landsat 50 served as historical reference for identifying the changed land use. To capture spatial hierarchies in satellite images effectively, in this paper, a novel dual attention-based vision transformer has been developed for land-cover classification in four categories namely, water, built-up, siltation, and vegetation. The developed model is trained on the data collected from three potential sites in India. The experimental results are validated against seven state-of-the-art deep learning models. The results reveal that the proposed method outperformed all the considered methods by achieving accuracy and precision of 88.4% and 88.9%, respectively, while consuming the least number of parameters. The results reaffirm the concretization and erosion of nature's flood buffers for economic advancement. • A dual-level attention based vision transformer is proposed for processing multi-spectral satellite images. • Open-source satellite imagery based land-cover classification of stream bed is proposed. • A new labeled data set is developed for three study sites in India — Bhalswa lake, Yamuna river, and Anasagar lake. • The proposed model outperforms other considered transformer based and convolution based models. [ABSTRACT FROM AUTHOR]

In this paper, we introduce pySimFrac , an open-source python library for generating 3-D synthetic fracture realizations, integrating with fluid simulators, and performing analysis. pySimFrac allows the user to specify one of three fracture generation techniques (Box, Gaussian, or Spectral) and perform statistical analysis including the autocorrelation, moments, and probability density functions of the fracture surfaces and aperture. This analysis and accessibility of a python library allows the user to create realistic fracture realizations and vary properties of interest. In addition, pySimFrac includes integration examples to two different pore-scale simulators and the discrete fracture network simulator, dfnWorks. The capabilities developed in this work provides opportunity for quick and smooth adoption and implementation by the wider scientific community for accurate characterization of fluid transport in geologic media. We present pySimFrac along with integration examples and discuss the ability to extend pySimFrac from a single complex fracture to complex fracture networks. • A python library to develop synthetic complex fractures. • Fractures can be passed to single- and multiphase simulations for analysis. • Examples of fracture generation and analysis is provided in Jupyter notebooks. • The code-base and user-documentation are made available on GitHub and dfnWorks websites, respectively. [ABSTRACT FROM AUTHOR]

Many geophysical imaging applications, such as full-waveform inversion, often rely on high-performance computing to meet their demanding computational requirements. The failure of a subset of computer nodes during the execution of such applications can have a significant impact, as it may take several days or even weeks to recover the lost computation. To mitigate the consequences of these failures, it is crucial to employ effective fault tolerance techniques that do not introduce substantial overhead or hinder code optimization efforts. This paper addresses the primary research challenge of developing fault tolerance techniques with minimal impact on execution and optimization. To achieve this, we propose DeLIA, a Dependability Library for Iterative Applications designed for parallel programs that require data synchronization among all processes to maintain a globally consistent state after each iteration. DeLIA efficiently performs checkpointing and rollback of both the application's global state and each process's local state. Furthermore, DeLIA incorporates interruption detection mechanisms. One of the key advantages of DeLIA is its flexibility, allowing users to configure various parameters such as checkpointing frequency, selection of data to be saved, and the specific fault tolerance techniques to be applied. To validate the effectiveness of DeLIA, we applied it to a 3D full-waveform inversion code and conducted experiments to measure its overhead under different configurations using two workload schedulers. We also analyzed its behavior in preemptive circumstances. Our experiments revealed a maximum overhead of 8.8%, and DeLIA demonstrated its capability to detect termination signals and save the state of nodes in preemptive scenarios. Overall, the results of our study demonstrate the suitability of DeLIA to provide fault tolerance for iterative parallel applications. • We propose a flexible fault-tolerance library for high-performance computing applications with data synchronization. • The library offers checkpointing rollback and interruption detection with heartbeat monitoring and termination signals. • We successfully applied the proposed fault tolerance library to a full-waveform inversion program. • The proposed library enabled employing the full-waveform inversion application in preemptive cloud instances. [ABSTRACT FROM AUTHOR]

The simulation of geological facies in an unobservable volume is essential in various geoscience applications. Given the complexity of the problem, deep generative learning is a promising approach to overcome the limitations of traditional geostatistical simulation models, in particular their lack of physical realism. This research aims to investigate the application of generative adversarial networks and deep variational inference for conditionally simulating channelized reservoir in underground volumes. In this paper, we review the generative deep learning approaches, in particular the adversarial ones and the stabilization techniques that aim to facilitate their training. We also study the problem of conditioning deep learning models to observations through a variational Bayes approach, comparing a conditional neural network model to a Gaussian mixture model. The proposed approach is tested on 2D and 3D simulations generated by the stochastic process-based model Flumy. Morphological metrics are utilized to compare our proposed method with earlier iterations of generative adversarial networks. The results indicate that by utilizing recent stabilization techniques, generative adversarial networks can efficiently sample complex target data distributions. • Centralizing different theoretical results on GANs in a principled approach. • Presenting a stable architecture combining recent stabilization methods. • Comparing a new method for conditioning geological GANs to a method of the literature. • Demonstrating results on meandering channelized reservoirs simulated with the Flumy model in 3d. [ABSTRACT FROM AUTHOR]

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. [Display omitted] • A hybrid algorithm to recognize metal-vectoring geochemical anomaly patterns. • A deep autoencoder network to learn deep feature codings of multi-element input data. • A geographical random forest regression to quantify spatially aware anomaly scores. • A comparative experiment on a case study from Takht-e-Soleyman district, Iran. • Success-rate curves, Student's t ‒statistic, and prediction-area plots for performance evaluation. [ABSTRACT FROM AUTHOR]

In Naturally Fractured Reservoirs (NFR) diffuse fractures arrangement results from mechanical stratigraphy and tectonic history during failure. Thus, modelling Discrete Fracture Network (DFN) requires to understand and to account for fracture relationships at bed-interface (abutment or crosscutting) in 3D through time (loading path). However, sampling fractures data meaningfully in subsurface has always been a challenge for geologist due to data scarcity. To better understand and forecast fracture networks in stratified rocks, we study outcrops with a focus on geometric relationships between stratigraphic interfaces and fractures. This paper presents an original python toolbox called FracAbut. It is composed of 1 main and 2 auxiliary codes that quantify the geometric relation between fractures and stratigraphic interfaces from 1D (wells, scan-line) and 2D (digital image, photographs data). We calculate the Interface Impedance (II) that accounts for fracture abutment (crossing or not), persistence (single- or multi-bed) and propagation polarity (upward or downward). For each stratigraphic interface FracAbut provides information on fractures (type, number) and interface sensitivity (coupling strength). First, we apply FracAbut on synthetic case studies, then, on naturally fractured and stratified carbonates in Berat, Albania. Using both 1D scan-line and 2D outcrop photograph, we show that i) a mechanical interface can have different coupling above and below based on propagation polarity, ii) FracAbut results can give useful insight on fracture transmissivity, iii) FracAbut is fast and efficient to quantify fracture patterns and classify mechanical interface impact; iv) they are no relation between bed thickness and fracture propagation. • Quantify the impact of mechanical stratigraphy on fracture propagation. • FracAbut is dedicated to process 1D or 2D fracture and bed interface surveys. • Interfaces are characterized based on fractures persistence and abutment. [ABSTRACT FROM AUTHOR]

Oil source correlation can be used to identify the origin of crude oil by linking crude oil to source rocks; however, the manual methods, which are limited by the sample or parameter quantity or imbalanced datasets, are facing uncertainties. Although the existing multivariate statistical techniques can alleviate this problem, they are facing difficulties in processing imbalanced datasets and quantifying source beds. Therefore, a novel oil-source correlation analysis model called SVM-SelectKBest combining a support vector machine (SVM) with a feature selection algorithm to mitigate the common issue of dataset imbalance in oil-source correlations is proposed in this paper. The SVM-SelectKBest offers advantages over normal SVM by dynamically selecting the most relevant features and fine-tuning model parameters to achieve higher accuracy and better generalizability in complex datasets. SVM compensates for class imbalances by heavily penalizing the misclassification of the minority class, and SelectKBest streamlines the feature set to enhance SVM's effectiveness on critical variables. Furthermore, a shallow neural network (SensoryAttentionNet) is introduced into the proposed model to address the issue of quantifying the source bed proportions in crude oil. The result show that SVM-SelectKBest has better performance in identifying key geochemical parameters and discriminating oil source correlation, its accuracy in unbalanced datasets is improved by near 40% compared to SVM. The model obtains 25 key geochemical parameters such as C17 n-heptadecane, Pr pristane, and C18 n-octadecane, it achieves F1 scores of 1.0 on the training, validation, and test sets. SensoryAttentionNet also performs robustly, with a low variance of 0.05 between its predicted and actual values. All the results indicate the effectiveness of the proposed method in dealing with the imbalance problem in oil-source source correlation datasets and in determining the proportional contribution of source beds in crude oil. • A support vector machine model combined with feature selection is presented. • The issue of dataset imbalance in oil-source correlations is mitigated. • A neural network for quantifying the proportion of source beds is presented. • The challenge of quantifying the proportion of source beds is resolved. • The Selection of geochemical parameters are optimized by feature selection. [ABSTRACT FROM AUTHOR]

In this paper, we present a new open-source software "Open-AMA" developed to investigate atmospheric circulation dynamics and environmental research. Open AMA presents an integral package to conduct several air mass analyses. It appears to be a powerful, versatile software package developed to meet the needs of researchers using python and C++ in order to facilitate and speed up working time. This software seamlessly integrates new models for source identification based on air trajectories and ambient air pollution concentration data and enhances certain existing ones. Beyond source identification, it offers a rich array of functionalities for making it automatic, quick and easy to get access many kinds data including gridded meteorological data, trajectory calculations, synoptic parameter extraction from back-trajectories. All this functionalities can be used through a user-friendly graphical interface. Open-AMA can be a significant leap forward in air quality research and analysis, empowering researchers with the tools they need to make informed decisions and address pressing environmental and public health challenges and enhance understanding of pollutant origins in the atmosphere. • Open-AMA is an Open sources software for atmospheric circulation research. • Regional Correlation Analysis model. • Adaptive k-means clustering. • 3D-PSCF and 3D-CWT source–receptor models. • Meteorological data processing. [ABSTRACT FROM AUTHOR]

Seismic facies inversion is an important process in the oil and gas industry to estimate subsurface geological facies or rock types based on seismic data. Recently, the Ensemble Smoother with Multi Data Assimilation (ES-MDA) has shown great success in solving complex inverse problems by generating an ensemble of solutions of the model variables for uncertainty quantification. However, there are significant challenges related to the computational cost, Gaussian assumptions, and its application for categorical model variables. Recent developments have proposed to integrate ES-MDA with Variational Autoencoder (VAE) to reparametrize facies into a continuous feature space and reduce computational costs. Despite promising results, it has introduced new limitations concerning network convergence, spatial correlation loss, and manipulation of the latent space. In this paper, a Fully Convolutional Variational Autoencoder (FCVAE) was proposed to outperform the VAE in terms of preserving spatial information, requiring fewer network parameters, presenting higher dimension reduction and yielding better facies estimation. The objective of the application example is to draw sound conclusions based on a 2D synthetic application, where the results of both methods can be compared with reference facies for quality evaluation. • Fully Convolutional VAE to reparametrize facies models. • Latent space with spatial correlation. • Seismic Facies Inversion with ES-MDA and FCVAE. • Pre-salt synthetic dataset with four facies. • Comparison between ES-MDA-FCVAE and ES-MDA-VAE. [ABSTRACT FROM AUTHOR]

Data-driven three-dimensional (3D) mineral prospectivity modeling (MPM) employs diverse 3D exploration indicators to express geological architecture and associated characteristics in ore systems. The integration of 3D geological models with 3D computational simulation data enhances the effectiveness of 3D MPM in representing the geological architecture and its coupled geodynamic processes that govern mineralization. Despite variations in modality (i.e., data source, representation, and information abstraction levels) between geological models and simulation data, the cross-modal gap between these two types of data remains underexplored in 3D MPM. This paper presents a novel 3D MPM approach that robustly fuses multimodal information from geological models and simulation data. Acknowledging the coupled and correlated nature of geological architectures and geodynamic processes, a joint fusion strategy is employed, aligning information from both modalities by enforcing their correlation. A joint fusion neural network is devised to extract maximally correlated features from geological models and simulation data, fusing them in a cross-modality feature space. Specifically, correlation analysis (CCA) regularization is utilized to maximize the correlation between features of the two modalities, guiding the network to learn coordinated and joint fused features associated with mineralization. This results in a more effective 3D mineral prospectivity model that harnesses the strengths from both modalities for mineral exploration targeting. The proposed method is evaluated in a case study of the world-class Jiaojia gold deposit, NE China. Extensive experiments were carried out to compare the proposed method with state-of-the-art methods, methods using unimodal data, and variants without CCA regularization. Results demonstrate the superior performance of the proposed method in terms of prediction accuracy and targeting efficacy, highlighting the importance of CCA regularization in enhancing predictive power in 3D MPM. • Casting 3D mineral prospectivity modeling (MPM) as a deep multimodal fusion problem. • Joint fusion network for integrating features of 3D geological models and simulation data. • Canonical correlation analysis regularization to guide network fusion of the cross-modal gap. • Maximizing the correlation of multimodal information boosts the performance of 3D MPM. [ABSTRACT FROM AUTHOR]

It is of very great significance to achieve accurate solution of equidistant points/equiratio line between two points at a large regional or global scale. At present, no effective method has been found for equidistant/equiratio line between two points on the Earth ellipsoid. In large scale space, the solution plane of equidistant/equiratio points between two points on the map cannot meet the requirements of high-precision space analysis, due to the existence of map projection deformation. The key to solving the above problems is to establish a method generating the equidistant/equiratio line based on solution to geodesic problem, and implement the entire calculation process on the Earth ellipsoid. Based on the Karney's direct and inverse formulas, the paper uses the pendulum method to design a cohesive method for calculating the equidistant/equiratio line between two points on the Earth ellipsoid. Several specific examples are selected for experimental verification, which proves that this algorithm can solve the problem of calculating equidistant/equiratio line of two points of different scale line in short distance, medium distance and long distance, and the precision of solving coordinate space is less than 0.00001 m, which can meet the requirements of long distance, wide range and high-precision spatial analysis. • A unified and efficient algorithm to calculate the equidistant/equiratio line between two points on the Earth ellipsoid. • The calculation is completely processed on the earth ellipsoid, and it can completely get rid of the influence of map projection. • The results can meet the high-precision requirements of spatial analysis on the Earth ellipsoid. [ABSTRACT FROM AUTHOR]

The survey setup of 3D magnetotelluric inversion is typically chosen to best reflect the underlying geological structure. Typically, the individual observation azimuths are distinct from this preferred survey layout orientation; in fact, measured electric and magnetic fields are aligned with local geomagnetic North. When that happens, either data preprocessing (rotation) is required before the inversion may be performed, or the orientation of the modeling grid is aligned with the data orientations. None of these approaches is optimal, possibly resulting in a reduction of data quality, the loss of statistical accuracy of the error estimates, or in unnecessary grid cells, resulting in increased computation times. In situations when the data quality may be improved by allowing distinct individual site orientations, such as when cultural noise or faulty instrumentation cause a degradation of one of the modes, data loss due to rotation is especially detrimental. In this paper, we present a modified 3D magnetotelluric inversion algorithm, based on the ModEM software, that allows for direct interpretation of observations with arbitrary data orientation angles (ModEM ADORA). In ModEM ADORA, the modeling responses are rotated to the observation orientation angles after each forward calculation and before the data misfit is evaluated. This allows for direct interpretation of observation data with distinct orientations at each site and (as needed for some historical data) at each frequency. The derivative of the penalty functional is computed by pre-multiplying the derivatives of the data functionals (or the complete Jacobian matrix) by the relevant sparse block-diagonal rotation operators. We describe the principle and implementation of ModEM ADORA, and provide a collection of synthetic and real data tests for validation and an analysis of effectiveness of the ADORA option of ModEM. We show that by decoupling modeling grid and sensor orientations, ModEM ADORA supports both computational efficiency and ease of use, as well as allows for the maximum quality and quantity of inverted data. • Modeling and observation coordinate systems can now be set arbitrarily for 3D magnetotelluric inversion. • More high-quality data can be incorporated into inversion by choosing optimal data orientation. • Solve for fewer inversion model parameters by rotating the modeling coordinate system to site distribution. • Free for academic use. [ABSTRACT FROM AUTHOR]

Seismic data denoising (SDD) plays an important role in obtaining high-quality data for subsequent seismic imaging and inversion. However, traditional SDD methods still present several disadvantages such as parameter selection difficulties, high computational costs, and a strong dependence on the experience of processing personnel. Deep learning (DL)-based SDD methods can overcome these issues to a certain extent, having the advantages of high computational efficiency and negligible dependency on the experience of processing personnel. However, existing DL-based SDD methods typically adopt a specific single-network architecture, which can be further improved in terms of training efficiency and denoising quality. In this paper, we propose a two-step prediction denoising method based on the combination of two identical or different network architectures for improving the accuracy and efficiency of DL-based SDD methods. Compared with existing SDD methods, the two-step prediction denoising method is more capable of extracting feature information at multiple scales because multiple network architectures are combined in a single denoising task, thus improving the training efficiency and denoising accuracy. In this study, we implemented the two-step prediction denoising method using an improved denoising convolutional neural network (DnCNN) and a nested U-shaped fully convolutional network (U-Net + +). Several synthetic numerical examples based on three different types of noise were denoised, with the results indicating that the proposed method effectively improves the accuracy and efficiency of the DL-based SDD method. Finally, a test using field data was performed to demonstrate the real-world applicability of the novel two-step prediction denoising method. • A modified DnCNN using a different architecture is constructed. • A network-based two-step prediction method is proposed to denoise seismic data. • Examples illustrate the accuracy and feasibility of the method. [ABSTRACT FROM AUTHOR]

This paper presents a model framework based on a deep learning algorithm using Faster R-CNN to improve the accuracy and efficiency of detecting long-distance levee hazards. The proposed method aims to assist technicians in identifying hazards in resistivity inversion profiles, thereby reducing the manual workload and extracting valuable information from a large volume of data using deep learning techniques. To this end, we constructed a dataset of levee hazards using inversion images of electrical resistivity imaging data obtained from numerical simulations, indoor experiments, and actual levee hazards. Subsequently, we utilized image augmentation techniques to expand the dataset and enhance the model's performance, resulting in a total of 2000 labeled hazard images. The network was then trained from scratch by setting the hyperparameters. the trained weight parameters and network architecture were saved for future use. The model's performance was evaluated using standard evaluation metrics, including precision, recall, F1 score, and average precision (AP). The proposed model demonstrated robustness and adaptability, achieving high precision and accuracy with an AP value of up to 100%. Furthermore, we validated the method using three cases, and the results indicated its superiority over traditional manual identification-based or outlier analysis detection of levee hazards in practical scenarios. This study emphasizes the significant advantages of deep learning techniques over the traditional methods, including improved accuracy, automation, speed, and scalability. It suggests that intelligent detection based on deep learning can accurately identify and locate hazards in levees. • A levee hazard detection model is built on the Faster R–CNN for position and prediction. • The model replaces manual identification, enhancing efficiency and accuracy in identifying hazards. • The self-made dataset, constructed from resistivity images, is augmented to improve model performance. • The model achieves excellent precision and accuracy, with an AP value of up to 100%. • Case study is carried out to verify the intelligent hazard identification method. [ABSTRACT FROM AUTHOR]

Pore-scale analysis and characterization of reservoir rocks provide valuable information for the definition and management of underground hydrogen storage and CO 2 storage or sequestration. This article presents an optimized implementation of the A* algorithm, the most popular pathfinding method in the presence of obstacles. In this context, the algorithm is applied to recognize the minimum length connected paths through each flow direction of 3D images of a porous medium representative of a reservoir rock. The identification of the main paths allows the characterization of the pore space and the calculation of fundamental petrophysical properties such as tortuosity and effective porosity, which can be used for permeability estimation. Compared to other algorithms available for pore-scale characterization, such as the pore centroid, A* provides a better approximation of the pore space available for the flow and, therefore, a reliable characterization of the petrophysical properties. On the other hand, path identification is significantly consuming in terms of time and memory. In this paper, an efficient and optimized implementation based on C++/OpenMP programming language is presented. The proposed implementation aims to the analysis of large-scale models profiting from parallelization, memory optimization, and enhanced managing of dead paths. Three test cases of increasing sizes are presented, to analyze the advantages and the disadvantages of the algorithm as the number of explored points increases. The 3D binary images analyzed are related to a synthetic domain (1 million voxels) and two actual sandstone samples (about 4 and 64 million voxels respectively). The code is validated against a Matlab serial implementation, showing a significant improvement in efficiency. Remarkable test cases of several millions of voxels were afforded, overcoming the memory and execution slowness issues. Moreover, the proposed implementation is suitable for large pore-scale models run in HPC environments. [Display omitted] • Computer-aided geometric characterization of porous media images. • Pathfinding algorithm implementation applied to Digital Rocks Physics. • Parallel computing optimization for native serial algorithms. • Validation on real sample binary representations. • Geometrical properties computation for underground storage. [ABSTRACT FROM AUTHOR]

Multibeam sonar data are a valuable tool for seafloor mapping and geological studies. However, the presence of outliers in multibeam data can distort the results of analyses and reduce the accuracy of seafloor maps. In this paper, we define a weighting function based on the performance of various outlier detection techniques (ODTs) for detecting outliers in multibeam data, which calculates an outlier probability score for each sounding. Our results show that each ODT has its own strengths and weaknesses, and that a combination of outlier detection techniques is promising to improve reproducibility, explainability and the accuracy of the detection process. To address the challenge of detecting outliers in multibeam data, we propose a weighted outlier detection function that outperforms individual outlier detection techniques in terms of precision, recall and F1 scores by considering their strengths and combining them in a way that accounts for variations in the data. The function detects various types of outliers with high precision and recall values, resulting in valuable improvements in outlier detection performance for multibeam data. Overall, our proposed workflow has the potential to significantly improve the way multibeam data cleaning is performed, with the weighted outlier detection function being applied first, detecting most of the outlier automatically, followed by a domain-expert review of a small group of soundings whose automatic outlier labeling is not unequivocal. • Developed innovative outlier detection for precision in seabed analysis. • Weighted function outperforms in precision, recall, F1 for seabed mapping. • Enhanced geoscientific data quality, preserving sonar data integrity. • Supports sustainable management with accurate seabed mapping accuracy. • Discussed adaptability and limitations, advocating further research. [ABSTRACT FROM AUTHOR]

The parameters of fractures, such as fracture geometries, placement, and conductivity, are critical for the production of horizontal wells in tight gas reservoirs. Current studies on fracture parameter optimization, especially those using the optimization algorithms, are often based on the assumptions that the fractures are uniformly placed with equal length and conductivity. However, non-uniform designs of fracture spacing have been proved to be effective in reducing fracture interference as well as increasing the well production, especially with strong rock heterogeneity. In addition, most of the current literature focuses on identifying optimal hydraulic fracture geometries without considering the fracture propagation process, which may oversimplify the evaluation of completion cost and result in fracture geometries that cannot be realized in practice. In this paper, an automatic optimization process, which could efficiently couple the fracture propagation, is proposed to optimize the fracture parameters. The genetic algorithm (GA) with penalty function method is used to optimize the constraint and non-linear problem. A modified PKN fracture propagation model is established to quickly predict the hydraulic fracture geometries. Fracture locations and fluid injection volume are sequentially optimized with consideration of heterogeneous rock properties. Model parameters are taken from typical tight gas horizontal wells in Jinqiu gas field in Sichuan Province, China. The results show the importance of permeability on both fracture length and production in tight gas well. The lower the permeability, the more significant the production increase after fracture location optimization. More fractures tend to be placed in area with small permeability and fewer fractures are required in highly permeable regions. With an increase in rock heterogeneity, the distribution of fractures becomes more uneven. Conversely, increasing the number of fractures can lead to a more even distribution of fractures. More fluid should be injected in low-permeability regions according to the optimization results. The optimization method is applied to a practical tight gas well in Sichuan Basin and the cumulative gas production can be increased by 6.5% to 8% when compared with the initial uniform designs. • A modified PKN model is proposed for quick prediction of fracture propagation, validated against numerical simulations. • GA is used to optimize fracture locations by considering rock heterogeneity's impact on fracture propagation and production. • Sequential optimization of fracture placement and injection volume is conducted with its reliability verified. [ABSTRACT FROM AUTHOR]

The acquisition of rock property information is at the core of regional geological survey and mineral exploration, but hand-crafted feature-based methods are heavily influenced by human prior knowledge and have limited transferability. End-to-end deep learning techniques, exemplified by convolutional neural networks (CNNs), have attained significant accomplishments in the domain of image classification. However, previous end-to-end CNN-based methods are hard to focus on image critical areas, and they also cannot make full use of global alignment dependency of the rocks. In this paper, rock image classification via a RockS2Net is proposed. The RockS2Net framework adopts a Siamese architecture, characterized by two branches that share parameters, enabling the efficient extraction of both global and local features. This approach facilitates the extraction of global features from entire images and focuses on critical areas to extract local features. The architecture of spatial transformer network (STN) is introduced to transform microscopic images of rock sections to their critical areas. By fusing local features and global features, the properties obtained from microscopic images of rock sections can be more accurately predicted. To test the robust generalizability of the proposed method, the constructed CHN-Rock images dataset is used for experiments and evaluation. Experimental results show that the accuracy of the proposed RockS2Net on the CHN-Rock image dataset is 2–3% higher than that of other rock image classification networks. • The proposed RockS2Net can effectively extract local features of the critical areas and global features from microscopic images of rock sections. • The proposed SGD block, which can strengthen the transfer of features and reduces the loss of features, fuses the two-stream global-local features, and discriminative features of the rock sections can be obtained. • The proposed STN blocks capture key regions by transforming microscopic images of rock sections to reduce the influence of irrelevant features on rock image classification. • The introduced idea of extended convolution is applied to the framework for solving the information loss caused by rock image resolution reduction and down sampling in rock image classification problems. [ABSTRACT FROM AUTHOR]

For layered models with a low-velocity half-space, Rayleigh waves can exhibit inversely dispersive patterns and thus are advantageous for providing useful information on the seismic properties of such structures. However, the calculation of dispersion curves usually encounters some difficulties such as root skipping or numerical instability due to the presence of leaky modes. The full-wavefield method and spectral element method can offer better solutions but at the cost of a significant computational effort. For practical applications, most researchers treat the zeros of the real part of the secular function as an approximation to the leaky modes, which may lead to relatively large errors that cannot be ignored. Instead, in this paper, we propose a simple and efficient method to approximate the leaky mode based on the local minima of the absolute value of the secular function. Compared with the dispersion curve generated from the real part of the secular function, the proposed method can provide a more accurate approximation. We further apply a new efficient Monte Carlo search technique to two synthetic datasets and a permafrost example. Results show that our forward and inverse methods are effective, efficient, and robust. Moreover, our numerical experiments indicate that the modified Thomson-Haskell method is more suitable for calculating leaky modes compared with the fast delta method and the generalized reflection/transmission method, and that the inversion results are less sensitive to the effect of the limited spread length on the measured dispersion curves. The present study emphasizes the high efficiency of the proposed method in imaging permafrost structures and foundation features of buildings or roads. • A fast solver for calculating the apparent modes due to a low-velocity half-space. • The modified Thomson-Haskell method is more suitable for calculating the leaky modes. • Application for imaging permafrost structures. [ABSTRACT FROM AUTHOR]

Seismic data are often characterized by low quality due to noise contamination or missing traces. Convolutional neural networks are popular in dealing with denoising and interpolation. However, fixed-size convolution kernels have limited feature extraction range, and popular networks aim at either denoising or interpolating. In this paper, we propose a Swin Transformer convolutional residual network (SCRN) for simultaneous denoising and interpolation. We step-by-step show on synthetic datasets that SCRN can effectively denoise, interpolate, or denoise and interpolate simultaneously. Following this, the trained simultaneous denoising and interpolation model is directly used to handle multi-tasks on the field datasets. We assess the reconstruction effect of SCRN based on synthetic and field datasets. Comparison methods include both supervised ones (DnCNN, Unet and RIDNet) and unsupervised ones (DDUL, MultiResUNet and DL-POCS). Experimental results show that SCRN outperforms the counterparts in terms of (1) quantitative evaluation indices, (2) event continuity and weak signal retention, (3) generalization seismic data ability to suppress various noise levels and interpolate different sampling rates, and (4) preservation of first-arriving edge signals. These results indicate that SCRN can effectively reconstruct seismic data. • Global and local feature fusion helps fully utilize seismic data's local similarity. • Swin Transformer shares temporal information, maintaining the continuity of the events. • Convolutional Neural Networks focus on high-correlation information, preserving weak signals. • Simultaneous denoising and interpolation networks complete the multi-task. • The diversity of the training set enables better generalization of the model. [ABSTRACT FROM AUTHOR]

The interest in the study of electromagnetic propagation through planetary atmospheres is briefly discussed. Special attention is devoted to extremely-low-frequency fields in the Earth's atmosphere for its global nature and possible applications to climate monitoring studies among others. In the Earth's case, the system can be considered as a spherical electromagnetic shell resonator in which two concentric and large conducting spheres with a radius around 6300 km are separated by a very small distance of around 100 km, the atmosphere height. A numerical solution using the Transmission Line Method is proposed. The classical spherical-coordinate description is easy to use, however, the important difference in the dimensions along the three coordinate directions causes high numerical dispersion in the results. A Cartesian scheme with equal node size for all directions is used to reduce this undesired dispersion. A pre-processing stage is the key point introduced to lessen the resulting high demand of memory and time calculation and make the solution feasible. A parallelized Fortran code together with pre- and post-processing Python programs to ease the user interface are provided with this work. Details on the Fortran code and the Python modules are included both in the paper and the source codes to allow the use and modifications by other researchers interested in electromagnetic propagation through planetary atmospheres. The program allows calculation of the time evolution of the electromagnetic field at any point in the atmosphere. It includes the possibility of considering multiple time-dependent sources and different homogeneous and inhomogeneous conductivity profiles to model different situations. Profiles to study day-night asymmetries or locally perturbed profiles which have been attributed to earthquakes in the literature are implemented, for instance. • A parallel algorithm is used to study the electromagnetic propagation in the Earth's atmosphere. • A TLM scheme is used for the numerical modeling of electromagnetic propagation. • Schumann resonances can be obtained in a full 3D electromagnetic model. • The implementation uses Cartesian coordinates to reduce numerical dispersion. [ABSTRACT FROM AUTHOR]

The Bay of Bengal (BoB) has maintained its salinity distribution over the years despite a continuous flow of fresh water entering it through rivers on the northern coast, which is capable of diluting the salinity. This can be attributed to the cyclic flow of high salinity water (≥ 35 psu), coming from Arabian sea and entering BoB from the south, which moves northward and mixes with this fresh water. The movement of this high salinity water has been studied and analyzed in previous work (Singh et al., 2022). This paper extends the computational methods and analysis of salinity movement. Specifically, we introduce an advection based feature definition that represents the movement of high salinity water, and describe algorithms to track their evolution over time. This method allows us to trace the movement of high salinity water caused due to ocean currents. The method is validated via comparison with established observations on the flow of high salinity water in the BoB, including its entry from the Arabian Sea and its movement near Sri Lanka. Further, the visual analysis and tracking framework enables us to compare it with previous work and analyze the contribution of advection to salinity transport. • Novel feature definitions for salinity transport based on advection. • Algorithms for computing and tracking this feature in the ocean. • Visual analysis of high salinity transport due to advection. • Comparison of salinity movement in the Bay of Bengal against existing methods. [ABSTRACT FROM AUTHOR]