Geophysical Applications Research Articles

Carbon Soil Mapping in a Sustainable-Managed Farm in Northeast Italy: Geochemical and Geophysical Applications

Recently, there has been increasing interest in organic carbon (OC) certification of soil as an incentive for farmers to adopt sustainable agricultural practices. In this context, this pilot project combines geochemical and geophysical methods to map the distribution of OC contents in agricultural fields, allowing us to detect variations in time and space. Here we demonstrated a relationship between soil OC contents estimated in the laboratory and the apparent electrical conductivity (ECa) measured in the field. Specifically, geochemical elemental analyses were used to evaluate the OC content and relative isotopic signature in collected soil samples from a hazelnut orchard in the Emilia–Romagna region of Northeastern Italy, while the geophysical Electromagnetic Induction (EMI) method enabled the in situ mapping of the ECa distribution in the same soil field. According to the results, geochemical and geophysical data were found to be reciprocally related, as both the organic matter and soil moisture were mainly incorporated into the fine sediments (i.e., clay) of the soil. Therefore, such a relation was used to create a map of the OC content distribution in the investigated field, which could be used to monitor the soil C sequestration on small-scale farmland and eventually develop precision agricultural services. In the future, this method could be used by farmers and regional and/or national policymakers to periodically certify the farm’s soil conditions and verify the effectiveness of carbon sequestration. These measures would enable farmers to pursue Common Agricultural Policy (CAP) incentives for the reduction of CO2 emissions.

Accuracy assessment of recent global ocean tide models in coastal waters of the European North West Shelf

The accuracy of global ocean tide models is assessed in coastal waters of the European North West Shelf to ascertain where higher resolution local (forecast) models are most needed for geophysical and navigational applications, and which global models are most suitable for providing boundary conditions for regional and local tide models. Five recent global ocean tide models (FES2014b, EOT20, TPXO9-atlas-v5, GOT4.10c, and DTU16) are considered, with the models first compared by interpolating them onto common grids and computing the mean absolute deviation at each grid point. Coastline tide gauge and offshore bottom pressure sensor data were collated from several sources to give a total of 279 observation sites for evaluating model accuracy, including observational values from 137 locations that have not previously been released and have therefore not been assimilated into any of the global models tested. The residual errors between each model’s predicted phasor and the corresponding observed phasor were calculated at each observation location, and quantified using the root mean square (RMS) and median absolute residual (MAR) for the eight tidal constituents M2, S2, N2, O1, K1, K2, P1, and Q1. To avoid RMS values being biased by observation point density, a Voronoi-weighted RMS based on the water area of the Voronoi polygon about each observation location was also developed and used. Four zones were defined based on ocean depth to gauge model performance, and model inaccuracy is again demonstrated in near-shore regions. Seven further zones were defined based on geographical areas, which reveals inhomogeneity among the global models. The smallest overall root sum square (RSS) RMS error across all eight constituents arises with FES2014b, although TPXO9-atlas-v5 has the best performance when using the MAR and Voronoi-weighted RMS metrics. Using only the 137 observation sites that have not been assimilated by any model and therefore provide an independent accuracy assessment, FES2014b exhibits the smallest errors at the coastline, with an RSS RMS of 24.46 cm. All models exhibit larger errors with the 137 independent observation sites than with all 279 observation sites, with an average overall increase in RSS RMS error of 12%, and an increase of 30% for coastline tide gauges, highlighting the need for local model development in these areas.

Energetics of particle-size segregation

We introduce a continuum framework for the energetics of particle-size segregation in bidisperse granular flows. Building on continuum segregation equations and a recent segregation flux model, the proposed framework offers general analytical expressions to study the physics of granular flows from a mechanical energy perspective. We demonstrate the framework's applicability by examining the energetics of shear-driven granular flows. Numerical experiments with varying frictional coefficients and particle-size ratios reveal two distinct phases in the energetics, marked by the separate onset of particle segregation and diffusive remixing. Furthermore, our numerical simulations alongside previous experimental results show that the bulk Richardson number $Ri$ , defined as the potential energy to kinetic energy ratio at steady state, follows the scaling relationship $Ri\equiv \hat {E}^{(s)}_{gp} / \hat {E}^{(s)}_{k} \propto Pe^{-1/2}_{sr}$ for $0.1 \leq Ri\leq ~10^{3}$ and $10^{-4} \leq Pe_{sr} \leq ~300$ , the segregation–rheology Péclet number. Finally, we present a Péclet-number-dependent theoretical expression for the degree of mixing (or segregation), validated by the compiled numerical and experimental dataset. Our findings hint that the bulk segregation–mixing state can be predicted and controlled using the segregation Péclet number $Pe$ and $Pe_{sr}$ , both determined from known system parameters, providing an instrumental tool for engineering and geophysical applications.

GENERALIZED DATASET OF GEOLOGICAL AND GEOPHYSICAL INFORMATION ON THE EASTERN SECTOR OF THE RUSSIAN ARCTIC FOR MACHINE LEARNING – BASED ANALYSIS

The article presents a practical approach to the geological and geophysical spatial data collection and preliminary processing to use in machine learning models for geophysical applications. According to the established principles for estimating efforts in data analysis, which are confirmed by the results of surveys among specialists, this stage is viewed as major time and resource-consuming, amounting up to 80% in total volume of data analysis for a hypothesis testing project. The paper focuses on creating a consistent data set that integrates geological and geophysical information on a given region. We consider problems of different sources in the geodata representation to be related to their format (vector/raster), scale, type of attribute information (quantitative/qualitative) and their availability. The algorithm formalization and synthesis for combining geospatial data and converting them into quantitative vectors is a critical aspect. Combining various data draws on the concept of neighborhood fitting in with the data selection techniques and data consolidation strategy. The paper presents the general architecture of the software and hardware complex which includes a module for data collection and transformation in Python using the Pandas library, a data storage system based on the PostgreSQL DBMS (Database Management System) with the PostGIS extension. It is shown that for the considered class of problems in geophysics, it is sufficient to use a relational DBMS for data storing and processing. If the problem dimension increases, it is proposed to use the Big Data technology based on Apache Hadoop for scaling the system. A practical application of the proposed approach is demonstrated as results of data collection for the Caucasus region and eastern sector of the Russian Arctic. Based on the prepared data, experiments were carried out using machine learning models for recognition of locations of potential strong earthquakes and for sensitivity estimation of several geophysical features of these regions. The article presents the experimental results and evaluation of their efficiency.

Evanescent Acoustic Waves

A theoretical study of “geometric” SP-evanescent (head) waves propagating in an isotropic homogeneous half-space or half-plane with a free boundary shows that these waves can satisfy the condition of absence of effort on the boundary plane if and only if the Lamé parameter λ is vanishingly small, which makes the existence of head waves of this type practically impossible. The analysis is based on the Helmholtz representation for the displacement field in combination with the decomposition of the stress and strain tensor into spherical and deviatoric parts. The obtained result about the non-existence of this type of evanescent waves can find application in theoretical geophysics in the study of seismic wave fields in the vicinity of earthquake epicenters, as well as in non-destructive acoustic diagnostic methods.

Score filtering for contextualized noise suppression of Poisson‐distributed geophysical signals

AbstractGeophysical and remote sensing products that rely on Poisson‐distributed measurement signals, such as cosmic‐ray neutron sensing (CRNS) and gamma spectrometry, often face challenges due to inherent Poisson noise. Common techniques to enhance signal stability include data aggregation or smoothing (e.g., moving averages and interpolation). However, these methods typically reduce the ability to resolve detailed temporal (stationary data) and spatial (mobile data) features. In this study, we introduced a method for contextual noise suppression tailored to Poisson‐distributed data, utilizing a discrete score attribution system. This score filter evaluates each observation against eight different criteria to assess its consistency with surrounding values, assigning a score between 0 (very unlikely) and 8 (very likely) to indicate whether the observation is likely to act as noise. These scores can then be used to flag or remove data points based on user‐defined thresholds. We tested the score filter's effectiveness on both stationary and mobile CRNS data, as well as on gamma‐ray spectrometry and electromagnetic induction (EMI) recordings. In our examples, the score filter consistently outperformed established filters, for example Savitzky–Golay and Kalman, in direct competition when applied to CRNS time series data. Additionally, the score filter substantially reduced Poisson noise in mobile CRNS, gamma‐ray spectrometry and EMI data. The scoring system also provides a context‐sensitive evaluation of individual observations or aggregates, assessing their conformity within the dataset. Given its general applicability, customizable criteria and very low computational demands, the proposed filter is easy to implement and holds promise as a valuable tool for denoising geophysical data and applications in other fields.

A Strong Noise Reduction Network for Seismic Records

Noise reduction is a critical step in seismic data processing. A novel strong noise reduction network is proposed in this study. The network enhances the U-Net architecture with an improved inception module and coordinate attention (CA) mechanism, suppressing noise and enhancing signal clarity. These enhancements improve the network’s capability to distinguish between signal and noise in the time–frequency domain. We trained and tested our model on the STEAD dataset, which eliminated noise across various frequency bands, improved the signal-to-noise ratio (SNR) of seismic records, and reduced the waveform distortion significantly. Comparative analyses against U-Net, DeepDenoiser, and DnRDB models, using signals with SNRs ranging from −14 dB to 0 dB, demonstrated our model’s superior performance. At the same time, we demonstrated that the Inception Conv Block has a significant impact on the denoising ability of the network. Furthermore, validation using the “Di Ting” dataset and real noisy signals confirmed the model’s generalizability. These results show that the proposed model significantly outperforms the comparative methods in terms of the SNR, correlation coefficient (r), and root mean square error (RMSE), delivering higher-quality seismograms. The enhanced phase-picking accuracy underscores the potential of our approach to advance in geophysics applications.

The influence of lamination and fracture on the velocity anisotropy of tectonic coals

The elastic anisotropy of coal formations is important for the prediction of unconventional coalbed methane reservoir performance and hydraulic fracturing. Although essential for several geophysical applications, the elastic anisotropy of coals is still poorly understood. Therefore, three sets of cylindrical coal samples are drilled parallel and perpendicular to the bedding plane, and the ultrasonic velocities under different confining pressures are measured in the laboratory. We combine experimental observations and modeling analysis to evaluate the elastic anisotropy of the coals. The results suggest that the velocity of coals parallel to the bedding plane is less sensitive to confining pressure than those perpendicular to the bedding plane. The fracture densities of coals parallel to the bedding plane are much lower than those perpendicular to the bedding plane, and most of the fractures are closed when the pressure is higher than 15 MPa. The organic matter (OM), clay content, and their preferred orientation play dominant roles in velocity anisotropy, and the fracture effect can significantly enhance it. The velocity anisotropy parameters of coals are positively correlated with the OM and clay content, and the preferred orientation of OM and clay plays a significant role in coal anisotropy. The results in this study are helpful in the characterization of the fractures of coals and can provide an important rock-physics basis for microseismic fracture monitoring and reservoir prediction.

Petrophysical cobalt exploration indicators in Kuusamo, northern Finland

The increasing global demand for cobalt, a critical green transition metal, encourages the development of improved exploration workflows, as Co typically occurs as a by-product of other metals, and the exploration and processing methods are rarely optimised for it. This paper investigates the possibility of extracting geophysical cobalt exploration indicators from petrophysical properties in Kuusamo, one of the most prominent cobalt-enriched areas in the largest European Co producer, Finland. The study region hosts 20 Au–Co deposits and occurrences classified as orogenic gold deposits with atypical metal association. Two separate mineralization stages have been distinguished: a pyrrhotite-dominated Co-stage and a pyrite-dominated Au–Co stage. The assumption in this study was that Co-enriched samples could be distinguished from Au-enriched samples and host rocks in petrophysical data due to the mineral associations. The investigation used an extensive previously acquired petrophysical and geochemical drill hole dataset from six targets complemented with new thermomagnetic data for magnetic mineral phase detection. The samples were divided into ore classes for the petrophysical data analysis. The results indicate that cobalt’s association with monoclinic pyrrhotite is a key factor for geophysical applications: the high remanent magnetisation characteristic of monoclinic pyrrhotite suggests that Co-enriched ores should typically associate with magnetic anomaly sources. Co enrichment is also related to higher electrical conductivities in comparison to Au. The workflow applied in this study illustrates how key petrophysical parameters can be connected with ore-bearing minerals and rocks, and the results indicate how cobalt could be explored more directly and effectively in the study area.

Seismic Random Noise Attenuation Using DARE U-Net

Seismic data processing plays a pivotal role in extracting valuable subsurface information for various geophysical applications. However, seismic records often suffer from inherent random noise, which obscures meaningful geological features and reduces the reliability of interpretations. In recent years, deep learning methodologies have shown promising results in performing noise attenuation tasks on seismic data. In this research, we propose modifications to the standard U-Net structure by integrating dense and residual connections, which serve as the foundation of our approach named the dense and residual (DARE U-Net) network. Dense connections enhance the receptive field and ensure that information from different scales is considered during the denoising process. Our model implements local residual connections between layers within the encoder, which allows earlier layers to directly connect with deep layers. This promotes the flow of information, allowing the network to utilize filtered and unfiltered input. The combined network mechanisms preserve the spatial information loss during the contraction process so that the decoder can locate the features more accurately by retaining the high-resolution features, enabling precise location in seismic image denoising. We evaluate this adapted architecture by applying synthetic and real data sets and calculating the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM). The effectiveness of this method is well noted.

Resolution enhancement of SMOS brightness temperatures: Application to melt detection on the Antarctic and Greenland ice sheets

A large part of the surface of the Greenland Ice Sheet (GrIS) and the margins of Antarctica are melting every summer, affecting their surface mass balance. Wet/dry snow status has been detected for decades using the peaks of brightness temperature at 19 GHz, and more recently at L-band (1.4 GHz) using both the SMOS and SMAP missions. SMOS owns a longer time series than SMAP with data since 2010, but the 52.5°incidence bin in the Level 3 (L3) product from Centre Aval de Traitement des Données SMOS (CATDS) that was previously used to detect melt suffers from a coarse spatial resolution. For this reason, we developed a new SMOS enhanced resolution brightness temperature (TB) product building on the radiometer version of the Scatterometter Image Reconstruction (rSIR) algorithm. We also exploited the SMOS L1C observations near 40°incidence angle instead of 52.5°as the native spatial resolution of SMOS is better at low incidence. The new product is posted on a 12.5 km polar stereographic grid and covers all the GrIS and Antarctica for 2010–2024 with twice-daily morning and afternoon acquisitions. The effective spatial resolution was evaluated to ∼30 km, a 30% enhancement compared to the SMOS L3TB at 40°and almost a 50% enhancement compared to the SMOS L3TB at 52.5°. Then, we applied a melt detection algorithm to both the enhanced resolution product at 40°and the L3TB product at 52.5°which is used in the literature. The spatial resolution enhancement results not only in the detection of smaller melt regions but also in a widespread increase in the annual number of melt days. This increase is larger than 30 days per year in the GrIS percolation area and on multiple Antarctic ice shelves. This is primarily due to the mix of dry and wet snow regions near the ice shelves grounding line, resulting in lower brightness temperature peaks in the SMOS L3TB product due to a large power spread. These findings highlight the dependence of melt detection in particular, and geophysical applications in general, on the spatial resolution of passive microwave observations. This study provides a new open dataset suitable to monitor melt at the surface and at depth on the two main ice-sheets.

Reflection Measurement of the Scattering Mean Free Path at the Onset of Multiple Scattering.

Multiple scattering of waves presents challenges for imaging complex media but offers potential for their characterization. Its onset is actually governed by the scattering mean free path ℓ_{s} that provides crucial information on the medium microarchitecture. Here, we introduce a reflection matrix method designed to estimate this parameter from the time decay of the single scattering rate. Our method is first validated by an ultrasound experiment on a tissue-mimicking phantom before being applied invivo to a human liver. This Letter opens important perspectives for quantitative imaging of heterogeneous media with waves, whether it be for nondestructive testing, biomedical, or geophysical applications.

The Electrical Properties of Dacite Mixed with Various Pyrite Contents and Its Geophysical Applications for the High-Conductivity Duobaoshan Island Arc

In this work, a series of electrical conductivities of pyrite-bearing dacite were measured under 10−1–106 Hz, 573–973 K, 1.0–3.0 GPa, and different pyrite contents ranging from 0 vol.% to 20 vol.%) using a Solartron–1260 A impedance analyzer. For the dacite sample with 5 vol.% pyrite, the electrical conductivity of the dacite increased with temperature but slightly decreased when the pressure was increased from 1.0 GPa to 3.0 GPa. In the temperature range of 573–973 K, the bulk electrical conductivity of the pyrite-bearing dacite gradually increased with increasing pyrite percentage from 0 vol.% to 20 vol.% at 1.0 GPa. Thus, a positive correlation between the electrical conductivity of the sample and the pyrite content was typically observed. In light of the significant enhancement in the electrical conductivity of the interconnected pyrite in the dacite, the value of the percolation threshold was determined as 7 vol.%. Furthermore, the dominant conduction mechanism of the small polaron for pyrite-bearing dacite was proposed from our obtained results on the chemical compositions and activation enthalpies under high-temperature and high-pressure conditions. A comprehensive consideration of our constructed electrical conductivity–depth profile based on the electrical conductivity of the pyrite-bearing dacite, can provide a good constraint on the volume of pyrite in dacite for high-conductivity Duobaoshan island arc. In conclusion, the presence of pyrite in dacite can provide a reasonable explanation for the high-conductivity anomaly observed in the region of Duobaoshan island arc.

Extending the blended generalized extreme value distribution

AbstractThe generalized extreme value (GEV) distribution is commonly employed to help estimate the likelihood of extreme events in many geophysical and other application areas. The recently proposed blended generalized extreme value (bGEV) distribution modifies the GEV with positive shape parameter to avoid a hard lower bound that complicates fitting and inference. Here, the bGEV is extended to the GEV with negative shape parameter, avoiding a hard upper bound that is unrealistic in many applications. This extended bGEV is shown to improve on the GEV for forecasting heat and sea level extremes based on past data. Software implementing this bGEV and applying it to the example temperature and sea level data is provided.

Introduction to this special section: Geophysical methods in archaeology

Archaeogeophysics, i.e., the application and integration of geophysics into archaeological investigations, is an exciting and growing field of study and an international collaboration at the intersection of the physical and social sciences. Through the incorporation of ground- and drone-based geophysical and remote sensing tools, these collaborative teams continue to advance the field of archaeology by rapidly and accurately revealing hidden features of the past and helping to focus active archaeological digs. Equally relevant, yet often unnoticed, the inclusion of noninvasive geophysical tools likewise provides a means to understand and preserve historically and culturally sensitive sites without the need for extensive and potentially destructive excavations.

Measurement of the Static Nonlinear Third‐Order Elastic Moduli of Rocks: Problems and Applicability

AbstractThe third‐order elastic (TOE) model has been used to describe the widely observed nonlinear mechanical behaviors of earth materials. In addition to linear elastic constants (λ, μ), three nonlinear elastic moduli (A, B, C) are required for isotropic rocks. Contrary to previous research on dynamic TOE moduli, this study followed the protocol to measure strain and stress under uniaxial and hydrostatic compressive tests statically, which were later used to invert for the full set of TOE moduli for four standard rock types with differing pore structures of a porous oolitic limestone, quartz‐rich sandstones, and a dense crystalline basalt. The applicability of the TOE model to characterize nonlinearity depends on the fulfillment of path‐independence and small‐strain assumptions. Using the measured static TOE moduli, the finite element model demonstrates that the stress in the vicinity of the wellbore is more amplified than the stress in the linear elastic case, which leads to a wider zone of rock failure around the wellbore. Due to the long‐standing discrepancy between static and dynamic moduli, the rarely reported full set of static TOE moduli is necessary and will benefit future research in understanding the effect of rock nonlinearity on geophysical and geomechanical applications, such as long‐term safe storage of CO2 and generating process of geohazards.

Analytical prediction of the formation factor for anisotropic mono-sized unconsolidated porous media

The rectangular Representative Unit Cell (RUC) models for mono-sized isotropic unconsolidated porous media have been extended to anisotropic media. Volume averaging was applied to Ohm's law in order to propose analytical expressions for the formation factor for electrical conduction in terms of the pore-scale linear dimensions. The formation factor, crucial for understanding electrical conduction in porous media, has significant applications in subsurface geophysics, petroleum reservoir characterization, and the durability assessment of construction materials. These analytical models provide valuable insights for optimizing the electrical properties of porous structures in various engineering fields, for example in mapping groundwater resources, differentiating oil-bearing zones in reservoirs, and predicting ion migration in concrete. Aspect ratios of the cell and solid dimensions have been introduced, and the formation factor expressed in terms thereof for five different arrays, i.e. a regular, a non-overlapping streamwisely fully staggered, an overlapping streamwisely fully staggered, a non-overlapping transversally fully staggered, and an overlapping transversally fully staggered array. The Laplace equation has furthermore been solved numerically for two-dimensional versions of the different arrays and the formation factor computed for several values of the aspect ratios. The proposed analytical models have been validated against the generated numerical data and compared to experimental and numerical data obtained from the literature. The proposed models have also been adapted to include a percolation threshold porosity in order to improve the model predictions for very low porosity media.

Three-phase-lag analysis of transversely isotropic double porous thermoelastic waves with liquid medium

This manuscript presents a mathematical framework for a double porous, thermoelastic solid half-space with transverse isotropy in contact with an inviscid liquid half-space. The analysis takes into account the three-phase-lag model within thermoelasticity theory. To gain deeper insights, we formulate the governing equations in a two-dimensional representation and subsequently employ normal mode analysis for their rigorous solution. The governing equations are influenced by both types of voids, anisotropy, thermal effect, and inviscid liquid. It has been discovered that thermoelastic half-space exhibits five coupled longitudinal waves, and in the liquid half-space, there is one mechanical wave. Secular equations are obtained by using thermal and mechanical conditions, resulting in a compact representation of phase velocity, specific loss, penetrating depth, and attenuation coefficient. Analytical presentations have been made to showcase the effects of anisotropy, voids, and liquid on various wave parameters. These effects are compared by illustrating the results graphically using the MATLAB software, allowing for a visual understanding of the comparative outcomes. This study has practical applications in engineering, materials science, geophysics, and environmental studies, contributing to advancements in various industries and scientific fields.

SmoGSI: smoothed multiscale iterative geostatistical seismic inversion

Iterative geostatistical seismic inversion is a vital technique for estimating subsurface properties. However, a conventional single-scale strategy faces challenges in preserving large-scale geological features due to the limited restoration of the type of data template, and conventional multiscale strategies face the challenge in that it is easy to lose the large-scale structure that was previously preserved. This paper introduces a novel smoothed multiscale strategy aimed at overcoming these limitations, which comprises two components: Simulated annealing at the coarsest scale and smooth conversation between two scale grids. This approach offers a smoother way to simultaneously retain large-scale and small-scale structures, improving the overall accuracy of the subsurface property estimations. To validate the effectiveness of our approach, we apply it to both synthetic and real examples. The results show that the simulated annealing strategy at the coarsest scale grid explores the prior space and finds the best large structures to avoid the generated models trapping in the local minimal. Meanwhile, the smooth conversation strategy between two scale grids helps us avoid the damage of the coarser structure. It can be explained that a large weight is assigned to the coarse structure at the beginning of the conversation of two grid scale, reducing the likelihood of the replacing proposed local small-scale geological patterns, which prone to be accepted in the conventional multiscale strategies. The combination of the two strategies used in the proposed smoothed multiscale strategy displays a significant improvement in subsurface property estimation accuracy compared to traditional multiscale strategies. This innovation can have far-reaching implications, benefiting a wide range of geophysical applications and contributing to more accurate and informed decision-making in geological and hydrogeological assessments.

Controllable Seismic Velocity Synthesis Using Generative Diffusion Models

AbstractAccurate seismic velocity estimations are vital to understanding Earth's subsurface structures, assessing natural resources, and evaluating seismic hazards. Machine learning‐based inversion algorithms have shown promising performance in regional (i.e., for exploration) and global velocity estimations, while their effectiveness hinges on access to large and diverse training data sets whose distributions generally cover the target solutions. Additionally, enhancing the precision and reliability of velocity estimation also requires incorporating prior information, for example, geological classes, well logs, and subsurface structures, but current statistical or neural network‐based methods are not flexible enough to handle such multimodal information. To address both challenges, we propose to use conditional generative diffusion models for seismic velocity synthesis in which we readily incorporate those priors. This approach enables the generation of seismic velocities that closely match the expected target distribution, offering data sets informed by both expert knowledge and measured data to support training for data‐driven geophysical methods. We demonstrate the flexibility and effectiveness of our method through training diffusion models on the OpenFWI data set under various conditions, including class labels, well logs, reflectivity images, and the combination of these priors. The performance of the approach under out‐of‐distribution conditions further underscores its generalization ability, showcasing its potential to provide tailored priors for velocity inverse problems and create specific training data sets for machine learning‐based geophysical applications.