Seabed geoacoustic parameters are crucial for coupled geoacoustic and underwater acoustic. Although sediment grain-size and compositional data in the Qiongdongnan region are abundant, measurements of seabed geoacoustic parameters remain limited. Based on measured grain-size and compositional data from 221 sediment stations in the Qiongdongnan region, this study uses forward modelling with the Helgerud effective medium theory and fluid–solid Zoeppritz equations to estimate sediment density, wave velocities, attenuation, and reflection coefficients. The results show that sediment density ranges from 1.42 to 1.92 g/cm3, with higher values in reef-island domain and lower values in deep-water domain. P- and S-wave velocities range from 1520 to 1670 m/s and 90 to 180 m/s, respectively, and show spatial patterns consistent with sediment density. Attenuation ranges from 0.63 to 0.78 dB/λ, with higher values in deep-water domain and lower values in reef-island domain, while the normal-incidence reflection coefficients range from 0.17 to 0.37 and show opposite trend. As the incidence angle increases from 0° to 90°, the reflection coefficient shows three regimes: creeping at 0–40°, accelerated at 40–70°, and gradual increase at 70–85°. The analysis suggests that seafloor geoacoustic parameters are mainly controlled by hydrodynamics, geomorphology, and sediment supply. High-energy shelf environments promote sediment coarsening and compaction, leading to higher density and wave velocities but lower attenuation, whereas fine-grained deposition in weakly dynamic deep-sea domain results in the opposite trend. Seafloor reflection behaviour is controlled by sediment properties, with P-wave reflection coefficient showing segmented angle dependence, rapid change near the critical angle, and varying characteristics across domains.
The Brazilian Equatorial Margin (BEM) is an oblique-transform continental margin inherited from the Pangea breakup and the opening of the Equatorial Atlantic Ocean. Located in the central portion of the BEM, the Barreirinhas Basin is a frontier region with only three deep-water wells. Efforts have been made to understand its tectono-depositional evolution and potential for petroleum exploration. This study provides the first integrated characterization of fluid migration systems in the offshore Barreirinhas Basin, identifying and mapping seismic indicators, as well as connecting them with regional settings. We analyzed conventional 2D and 3D seismic reflection data, integrated with available well data. Additionally, seismic attributes (instantaneous amplitude, pseudo-relief, similarity, and chaos) were applied to enhance the interpretation of conventional seismic profiles. Our results show diverse fluid migration features, from deep strata to the seafloor, constrained in four main migration systems. Two systems are correlated to regional gravitational collapse. Fluid conduits, pockmarks, mounds, and faulted zones indicate focused fluid flow. A third system is developed in a region dominated by transform tectonics related to the Romanche Fracture Zone, in which dense fault networks are observed within a laterally extensive interval from the slope to the basin plain. The fourth system is located in a broad region of the basin plain, characterized by numerous seafloor mounds associated with fluid conduits and seismic indicators of hydrocarbon accumulation. These findings improve the knowledge about an underexplored basin and provide a foundation for the understanding of fluid migration systems and their implications, including marine processes and products, emplacement of igneous bodies, and hydrocarbon prospectivity.
Submarine pockmarks are among the most ubiquitous geomorphological expressions of seafloor fluid escape, commonly occurring in high-density fields that may include hundreds to tens of thousands of individual features. Under these conditions, traditional manual morphometric analyses become time-consuming and impractical, limiting the robustness and reproducibility of pockmark-based studies. In this work, we present an open-source Python-based semi-automated workflow designed to systematically analyse the morphometry of pockmark populations using Digital Elevation Models (DEMs) and polygonal vectors as primary inputs. The workflow enables extraction of a comprehensive set of planimetric, elevation-based, geometric, and volumetric measurements, thereby minimizing operator-dependent biases. The approach has been tested on two contrasting geological case studies: the tectonically and volcanically active Graham Bank region in the Sicily Channel, mapped using high-resolution multibeam bathymetry, and the passive north-western Sardinian continental margin, investigated using regional-scale bathymetric grids from EMODnet Bathymetry. These datasets allow evaluation of the workflow performance across different geological settings, spatial resolutions, and pockmark size classes. Results demonstrate that the workflow provides robust and internally consistent morphometric outputs, suitable for large-scale statistical analyses and comparative studies. Although developed and validated on submarine pockmarks, the method is inherently general and can be applied to any negative landform that can be represented as a closed polygon and for which a DEM is available, such as karstic sinkholes or other collapse-related depressions. The proposed workflow therefore represents a flexible and reproducible methodological framework with broad applicability across geomorphology, environmental sciences, and hazard-related studies.
The East Vietnam Sea (EVS) is a complex marginal sea whose tectonic evolution is recorded in its deep crustal structure. Isolating the gravity anomaly of this deep structure requires removing the gravity effect of the overlying seawater and sedimentary layers, a task complicated by the nonlinear variation of density with depth which is rarely captured by simple constant-density assumptions. This study presents a refined 3D frequency-domain approach to calculate the gravity effects of seawater and sedimentary layers using depth- and latitude-dependent density models constrained by local borehole data. For sedimentary layers, we derive an empirical depth-density function from 23,580 bulk density measurements, capturing compaction-driven density variations down to > 3,000 m. For seawater, we apply a latitude- and depth-dependent density model that accounts for compressibility and thermohaline variations. Results demonstrate that constant-density assumptions introduce errors of up to 7.5 mGal for seawater and up to 200 mGal for sediments, particularly in deep-water areas (> 4,500 m) and thick basins ( 12 km sediment). After removing the gravity effect of these layers, the derived basement-mantle gravity anomaly correlates strongly with Moho depth from CRUST1.0 (R = 0.891), significantly outperforming constant-density (R = 0.807) and previous variable-density models (R = 0.861). These findings demonstrate that locally constrained, depth-dependent density functions are essential for reliable gravity-based crustal interpretation, providing a more robust foundation for tectonic analysis in the EVS.
The well-seismic tie process is critical for integrating well and seismic data to achieve accurate subsurface interpretations. While previous studies have investigated the influence of mud filtrate invasion on well-log responses, its impact on seismic calibration remains a persistent challenge in reservoir characterization. This study builds on existing research by employing Gassmann’s fluid-substitution model and sparse-spike deconvolution to systematically evaluate how mud filtrate alters seismic reflections in both synthetic and real datasets from the Norne field in Norway. Our analysis demonstrates that fluid saturation levels, reservoir porosity, and invasion extent significantly affect P-wave, S-wave, and density measurements, leading to measurable discrepancies in well-seismic correlations. The findings corroborate earlier work showing effects of invasion on elastic properties while providing new insights into wavelet distortion patterns under varying saturation conditions. By quantifying these impacts, our results underscore the need to account for mud filtrate effects in seismic-well calibration workflows to improve the reliability of reservoir models for exploration and production applications.
The Middle Miocene Birsa Formation in the Gulf of Hammamet, northeastern Tunisia, records the evolution of a tectonically influenced deltaic system controlled by variations in accommodation space, sediment supply and relative sea-level changes. This study presents an integrated sedimentological, petrophysical and sequence stratigraphic characterization of the Birsa Formation based on well-log analysis from a concession in the Gulf of Hammamet. The investigated interval consists of interbedded sandstones and shales organized into eight reservoir subzones that exhibit significant lateral and vertical heterogeneity. Gamma-Ray, resistivity density and neutron log responses were used to identify depositional facies ranging from distributary channels and mouth-bar deposits to distal delta-front and prodelta shales reflecting successive lowstand, transgressive and highstand depositional phases. Structural analysis indicates that tectonic inversion reactivated pre-existing extensional faults forming elevated fault-bounded blocks that acted as favorable hydrocarbon traps whereas structurally lower areas are predominantly water-bearing. Petrophysical evaluation reveals that reservoir quality is primarily controlled by effective porosity, clay volume and water saturation. Among the identified depositional sequences, the most favorable reservoir intervals occur within the lowstand systems tract (LST) where sand bodies exhibit higher effective porosity, lower clay content and improved hydrocarbon saturation compared to transgressive and highstand deposits. This case study illustrates the value of integrating sequence stratigraphic interpretation with petrophysical and structural analyses to improve the understanding of reservoir distribution and quality variations within inverted graben settings. The results provide new insights into the combined influence of depositional architecture and tectonic evolution on hydrocarbon accumulation in the Birsa Formation and offer a robust framework for improving reservoir characterization in the Gulf of Hammamet.
The Pearl River Mouth Basin (PRMB), a critical Cenozoic marine sedimentary basin and offshore hydrocarbon production area in China, lacks a precise chronological framework for its Neogene strata, hindering in-depth understanding of climate change drivers and sedimentary evolution during this period. In this study, we present a cyclostratigraphic analysis of gamma-ray (GR) well log data from the Zhuhai and Zhujiang Formations of Well LW-X in the Baiyun Sag, PRMB. Time series analysis of the GR data shows significant peaks with periods at 405 kyr and 100 kyr (long and short eccentricity), 41 kyr (obliquity), and 20 kyr (precession). Based on 405 kyr long eccentricity tuning of GR series, a high-resolution astronomical timescale of approximately 12.5 Myr spanning the Middle Oligocene to the Early Miocene was established. A marked decrease in GR values, coupled with abrupt increases in sedimentation rates and calcareous nannofossil abundance after 23 Ma, was identified and interpreted as the Mi-1 glaciation event. We applied sedimentary noise model to reconstruct relative sea-level changes in our studied section based on the GR data. This reconstruction reveals a 1.2 Myr cyclicity in relative sea-level change over the studied interval, suggesting a close relationship with long obliquity period climate forcing. These cycles exhibit strong consistency with the eleven third-order sequences interpreted by previous researchers (Boulila et al. 2011; Hardenbol et al. 1998). Taken together, we infer that 1.2 Myr long obliquity forcing has played an important role in controlling climate, sea-level changes and third-order sequence development during the Middle Oligocene to Early Miocene.
Seismic facies interpretation in the Arctic continental margin is essential for understanding complex subsurface structures shaped by permafrost, gas hydrates, and glacial-interglacial climatic fluctuations. Conventional seismic facies interpretation requires substantial effort and is highly susceptible to interpreter subjectivity. To overcome these limitations, this study proposes an unsupervised learning framework that integrates multi-attribute analysis with a convolutional autoencoder for automatic seismic facies classification along the western Chukchi Rise, Arctic Ocean. The workflow consists of three components: (1) extraction of seismic attributes, (2) feature learning using an autoencoder, and (3) dimensionality reduction through Uniform Manifold Approximation and Projection (UMAP). The reduced embeddings were subsequently clustered using hierarchical agglomerative clustering (HAC), Gaussian mixture model (GMM), and self-organizing map (SOM), and their quantitative performance was evaluated. Through the numerical experiments, the autoencoder effectively captures latent structural and textural patterns, while UMAP preserves meaningful relationships among attributes during dimensionality reduction. Among the three clustering methods, HAC achieved the highest performance, successfully distinguishing seismic facies based on subtle variations in amplitude and lateral continuity. The proposed approach improves the objectivity and efficiency of seismic facies classification for characterizing Arctic subsurface structures. These findings demonstrate that the proposed framework can reliably delineate seismic facies within the geologically complex Arctic environment.
Nuclear Magnetic Resonance (NMR) logs provide critical insights into pore-scale fluid distribution, including clay-bound water (CBW), free fluid index (FFI), and bulk volume irreducible (BVI), but their high acquisition cost limits widespread availability. This study presents an interpretable machine learning (ML) workflow for predicting these NMR-derived parameters from conventional well logs, including gamma ray (GR), bulk density (RHOB), neutron porosity (APLC), photoelectric factor (PEF), and resistivity (RLA5). Data from three wells (WEB-1, PAPYRUS-1, and BAMBOO-1) in the West El-Burullus gas field, offshore the Nile Delta, Egypt, are used to develop and evaluate two supervised learning models: Extreme Gradient Boosting (XGB) and Multilayer Perceptron (MLP). Model development incorporated rigorous data preprocessing, systematic hyperparameter optimization, and SHAP-based interpretability analysis to ensure physically meaningful predictions. Results demonstrate that the XGB model consistently outperformed the MLP model, achieving testing R² values of 0.885, 0.895, and 0.846 for CBW, FFI, and BVI, respectively. Blind-well validation further confirmed model generalization, with XGB achieving R² values of 0.811 for CBW, 0.825 for FFI, and 0.774 for BVI in the PAPYRUS-1 well. In the BAMBOO-1 well, where NMR measurements were unavailable, the predicted profiles reproduced geologically consistent stratigraphic trends. SHAP analysis revealed that model predictions were controlled by physically plausible relationships between conventional logs and reservoir properties, enhancing confidence in the workflow. The proposed workflow demonstrates that ensemble-based approaches such as XGB can provide reliable, interpretable, and computationally efficient predictions of advanced petrophysical properties from conventional logs. Although the results are encouraging, the limited dataset size constrains broader generalization. Future work should focus on expanding the database, improving BVI prediction through advanced feature engineering, and validating the workflow across a wider range of geological settings.
Convolutional Neural Networks (CNNs) have been widely used to model the nonlinear relationship between gravity anomalies and seafloor topography. However, most CNN-based methods operate only in the spatial domain, which limits resolution and hinders the recovery of fine-grained topographic features. To address this, we propose SGL-CNN, a novel framework that extracts multi-input features from both spatial and frequency domains. By integrating multi-component gravity anomalies with long-wavelength bathymetric data, our model simultaneously captures low-, medium-, and high-frequency seafloor components, enabling more detailed topographic reconstruction. We validate SGL-CNN in three representative regions of the Western Pacific–a slope, a seamount, and a trench–against baseline methods (ParkerO, SAS, GGM, LCNN). Accuracy and PSD results show that SGL-CNN outperforms others over seamounts and trenches. Across diverse terrains and depth ranges, its dual-domain three-branch architecture (Spatial, Global and Local Frequency) effectively handles multi-scale wavelength distributions, recovering low-, medium-, and high-frequency components corresponding to slope trends, seamount bodies, and trench fracture zones. Ablation studies confirm the necessity of the proposed architecture and the long-wavelength bathymetric input, and further validation on a high-latitude grid supports its generalizability. In summary, the synergistic fusion of spatial and spectral features in SGL-CNN overcomes spectral truncation issues in single-domain methods, achieving high-resolution bathymetric inversion.
Marine gas hydrates are both an untapped energy source and a drilling hazard, making reliable subsurface assessment essential. We analysed wells in the deepwater Mahanadi Basin, offshore India, to estimate gas hydrate saturation and pore pressure by combining established petrophysical methods with machine learning (ML). Conventional empirical methods, though widely used, often fail to capture complex lithological and geomechanical relationships, especially where limited well information introduces substantial uncertainty. In this study, we present an integrated, data-driven workflow for estimation of gas hydrate saturation and pore pressure using ensemble machine learning (ML) techniques. Archie’s law was first applied to resistivity logs to derive gas hydrate saturations, while Eaton’s and Bower’s methods provided pore pressure profiles from sonic, resistivity, and density data. To overcome empirical limitations, Random Forest, Gradient Boosting, Extremely Randomised Trees, and Bootstrap Aggregating regressors were trained on well NGHP-01-08 and validated on well NGHP-01-19 from the Mahanadi Basin, offshore India. The ML models successfully identified nonlinear patterns in well log responses (resistivity, P-wave velocity, gamma ray, bulk density, and stratigraphic depth bins) and accurately reproduced both smooth background trends and sharp local anomalies, including hydrate-rich intervals near the Bottom Simulating Reflector. Among the tested models, Random Forest achieved the highest predictive accuracy in terms of R^2 across both training and testing phases, followed by Gradient Boosting, Extremely Randomized Trees, and Bootstrap Aggregating. For pore pressure estimation, Random Forest and Gradient Boosting regressors were trained on well NGHP-01-08 and validated on well NGHP-01-19, with both models delivering excellent predictive accuracy across Eaton’s and Bower’s methods, and Gradient Boosting showing improved generalization in heterogeneous intervals. The integrated approach reduces prediction uncertainty, enhances cross-well applicability, and provides a transferable methodology for hydrate and pressure characterization in other offshore basins with similar geological settings.
The study of channel behaviour is significant in the design of an underwater acoustic (UWA) communication system. In a time-varying underwater environment, transmitted acoustic signal characteristics might be altered due to the impact of scatterers. To examine the impact, this work presents geometry-based modelling of UWA channels for a mobile receiver under different scattering conditions. The model considers scattering from the sea-surface, seabed, and both sea-boundaries, incorporating uniform and random distributions of scatterers to represent underwater propagation conditions. A total of 21 scenarios were analysed to evaluate the combined influence of transmitter–receiver geometry, scatterer distribution, and receiver mobility on UWA channel characteristics. The analysis was carried out using channel impulse response, temporal correlation, received power, and RMS delay spread metrics. The simulation results show that the configuration with the transmitter located below the receiver generally provides more favourable propagation characteristics, with comparatively higher received power and lower multipath dispersion, followed by the configurations with the transmitter located inline and above the receiver. Seabed scattering environments exhibit stronger attenuation and larger delay spread, while surface scattering has the least impact. The findings demonstrate that transmitter–receiver geometry and scatterer boundary conditions significantly influence channel stability and multipath behaviour in UWA communication environments.
Makran subduction zone (MSZ) along southern Iran and Pakistan has been recognized as a well-known source of earthquake Tsunamis. However, recent geological studies suggest that submarine landslides may also contribute to tsunami generation in this region. In this paper, we identify past submarine landslides offshore of Chabahar using reflection seismic data and investigate their tsunami potential. These landslides represent past mass failure events preserved in the seismic record; their precise age cannot be determined from 2D seismic data alone and would require borehole or sediment core data. No direct historical tsunami records have been associated with these specific events to date. Two landslide sources located 40 km and 120 km offshore are simulated using seismic profiles from the western Makran, with scenarios representing slide volumes ranging from 1 km³ to 13 km³. Time-series analyses reveal maximum station-recorded amplitudes of 1.5 m at Jask, 1 m at Chabahar, < 0.5 m at Pakistani stations, and 2.5–3 m negative waves in Oman. Wave arrivals range from 10 to 61 min across stations, with peak amplitudes not necessarily coinciding with first arrivals, and prolonged wave activity observed at several sites. Periods vary spatially, reflecting local shelf width and coastal geometry: Iranian stations ( 6–10 min), Pakistani stations ( 10–18 min), and Omani stations ( 8–14 min). Longer landslide paths generate stronger waves, with the worst-case scenario producing up to 6 m at Chabahar and 8 m along Oman (Muscat–Sur), flooding inland areas by 300 m. Most events along the Iran–Pakistan coast remain small (< 0.2 m), though localized amplification occurs near 60°–62°E. The exceedance probabilities at Chabahar show a heavy-tailed distribution, with the likelihood of exceeding a 1 m wave height reaching 0.8, highlighting a high potential for extreme events. These findings underscore the importance of incorporating landslide dynamics and high-frequency wave generation into tsunami hazard assessments. Complex wave interactions in the Gulf of Oman further modulate risk, emphasizing the need for high-resolution seafloor mapping, additional seismic profiling, and three-dimensional non-hydrostatic modeling for improved hazard evaluation.
The Northwest (NW) African continental margin is a typical passive margin hosting thick sedimentary sequences. Previous studies on the sedimentary processes shaping the margin were often based on research data from single cruise. Numerous vessel-based multibeam bathymetric datasets are, for example, available for the region, including data collected during transits and cruises where seafloor mapping was not a primary objective in the past 30 years. These data have, however, never been combined in a systematic manner, hindering margin-wide analysis of its sedimentation patterns. Here we compiled raw datasets from 99 cruises and different multibeam systems. The workflow is implemented in a cloud-based LINUX environment using open-source MB-System software. It achieves automation through recursive shell-scripting that handles multi-format conversion, navigation cleaning, and adaptive filtering constrained by a ± 15
The Gunsan Basin in the South Yellow Sea has undergone multi-phase tectonic evolution in the East Asian margin, and its basin structures are important for understanding the regional tectonic processes. Previous studies using ocean-bottom seismometer (OBS) and streamer data have revealed a complex crustal structure in this region, including the crustal thinning due to strike-slip faulting and the development of buried hills. However, the large station spacing in OBS surveys results in low resolution of tomographic inversion, leading to an ambiguity in the interpretation of basin geometry and buried hill structures. A new tomographic workflow involving multiscale parameterization, grid size optimization, and initial model scanning is proposed to improve the fidelity of tomographic inversion. Numerical tests show that optimal grid selection effectively enhances inversion accuracy under large station spacing. The use of initial model scanning combined with inversion parameters optimization significantly improves the stability of the solution. The final inverted model reveals three isolated high-velocity anomalies, located at approximately 40 km, 100 km, and 120 km along the profile, which are consistent with characteristics of buried hills. Those buried hills produce a basement relief of about 2 km, which is further supported by refraction arrival-time interpretation, and distort the sedimentary layers, indicating their significance in controlling the eastern sub-basin structure. The proposed tomographic strategy offers a robust approach for constructing reliable upper crustal models from sparsely spaced OBS data.
In recent years, development projects targeting coastal waters have increasingly emphasized clean energy production and long-term sustainability. Among various geophysical investigation techniques, seismic refraction surveys provide subsurface velocity structure information beneath the seabed and therefore support marine engineering applications. However, economical and miniaturized refraction survey technologies suitable for shallow coastal waters remain insufficiently developed and rarely validated under real field conditions. In this study, we propose a refraction survey method that economically generates a long-wavelength velocity model in shallow coastal waters characterized by numerous obstacles, and we evaluate its field applicability. We design a multi-vessel-based refraction data acquisition strategy that employs only conventional survey equipment, including a small air gun source and a short-offset seismic streamer with a total active length of approximately 25 m, operated from two 10-ton vessels. In addition, we develop a Global Navigation Satellite System (GNSS)-based triggering system that enables precise source-receiver synchronization without offset constraints, and we verify its operational performance and data reliability through a field application in real coastal waters. Despite imperfect field conditions that result in limited source-receiver geometry, first-arrival travel-time tomography yields a long-wavelength velocity model with an imaging depth of approximately 250 m, which is sufficient for engineering-scale geological interpretation. These results demonstrate that seismic refraction surveys can be applied economically to construct long-wavelength velocity models beneath the seabed in coastal areas where diverse marine engineering activities are conducted.
High-resolution bathymetric models are essential for oceanographic research and marine applications. Deep learning methods are increasingly used for bathymetric inversion, typically relying on gridded shipborne soundings as training labels. However, the confidence of these gridded depth values is highly heterogeneous, as they are derived from interpolating sparse and unevenly distributed point measurements. To address this inherent uncertainty in the training data, we propose the Abundance-Constrained Convolutional Neural Network (ABCNN). Our approach introduces an "abundance field" that quantifies ship sounding density, providing a direct measure of confidence for each grid cell. This field is integrated into an adaptive loss function, which weights the learning process to reliable grid cells and reduce the influence of uncertain ones. Experiments show that ABCNN produces more accurate and robust seafloor topography, offering a principled framework that effectively accounts for the varying quality of gridded bathymetric data.
Advances in marine geophysical imaging, including high-resolution side-scan sonar and multibeam bathymetry integrated with ground-truthing methods (e.g., sediment sampling and underwater video observations), have substantially enhanced our ability to map submerged glacial landforms and reconstruct past ice-sheet dynamics. These techniques provide critical insights into ice-flow patterns, subglacial erosion, and sediment deposition—key processes for understanding both paleo-glaciation and the response of modern ice sheets to climate change. Despite these advances, nearshore environments, which preserve key evidence of past ice-sheet and ice-marginal processes, remain understudied in many regions of Antarctica. To address this gap, we present the first high-resolution geophysical and ground-truthing dataset from the western coast of Horseshoe Island (Western Antarctic Peninsula). This dataset integrates side-scan sonar imagery with underwater video observations to map nearshore erosional and depositional landforms at unprecedented resolution. The identified landforms provide new insights into past ice-flow dynamics and subglacial processes, thereby helping to constrain ice-sheet reconstructions and improve predictions of future change. Distinct morphological boundaries and erosion patterns indicate that meltwater streams have significantly influenced the seafloor, particularly at depths greater than 10 m, reflecting glacial–marine and ice-marginal environmental conditions in the recent past. Within a limited area, geophysical evidence reveals subglacially formed seafloor features, including remnants of a paleo-lake and its associated paleoshoreline, pointing to former subglacial hydrological activity. The orientation of meltwater channels and streamlined glacial landforms indicates a dominant southward paleo–ice flow. This interpretation is supported by greater seafloor incision and deeper bathymetry in the southern sector of the island compared to its western margin. Collectively, these observations advance reconstructions of ice-sheet dynamics by elucidating the role of subglacial and ice-marginal processes in shaping nearshore submarine geomorphology. The integration of high-resolution marine geophysical data with ground-truthing observations provides critical empirical constraints for numerical ice-sheet models and contributes to refining projections of Antarctica’s future contribution to global sea-level rise.
Shear‑wave velocity (Vs) is vital for seismic reservoir characterization, AVO/AVA analysis, rock‑physics modeling and geomechanics, yet complete Vs logs are often unavailable. In this study, we present a comprehensive comparison of 36 methods to estimate Vs. The methods used for this comparison include empirical relations, linear and regularized regressors, power‑ and polynomial‑based transforms, support‑vector, tree and gradient‑boosting algorithms, as well as ensemble frameworks that we applied to data from a clastic-dominated well in Australia’s Cooper–Eromanga Basin. This workflow assumes minimal measured Vs coverage (ultra-sparse to minimally sparse), aiming to reconstruct missing intervals rather than generate complete logs without supervision. To simulate real life operational conditions, Vs values were deliberately removed from randomly selected intervals (training subset) while a depth‑balanced quality‑control (QC) subset retaining measured Vs was reserved for validation. All models were run with their default hyper‑parameters to reflect “out‑of‑the‑box” performance and were ranked with multiple error metrics: coefficient of determination (R2), mean absolute error (MAE), root mean square error (RMSE), and mean square error (MSE), along with depth-dependent residual diagnostics. Extra Trees — an ensemble method using many decision trees where split thresholds are chosen randomly rather than optimally — yielded the highest accuracy in this well (R2 = 0.921; MAE = 82 m/s) for Dandy-001, but applying the same top‑ten models to two different wells (Casimir‑001 and Bagheera East‑001) revealed markedly different model rankings among the three clasticdominated wells (in the same basin). These differences highlight variability in method workability across lithological and data‑quality scenarios. Castagna (Geophysics 50(4):571-581, 1985) and simple interpolation anchored the lower‑bound performance (R2 ≈ 0.87 and 0.63, respectively), while several default machine‑learning configurations (e.g., Gaussian‑Process, SGD) were unstable. Overall, although no single technique dominates in every case, a best-fitting method that most closely reflects the measured data can be identified for each dataset. However, this requires performing a thorough, scenario-specific QC to evaluate potential bias, variance, and situational reliability.
Shear-wave velocity is an important parameter characterizing the physical properties of marine sediments. Understanding the behavior of shear-wave velocity in seafloor surface sediments is critical for interpreting underwater sound propagation, predicting shallow-sea sound fields, and evaluating geomechanical characteristics in marine engineering. Well-logging data obtained from 78 sites across 36 legs of the Ocean Drilling Program/International Ocean Discovery Program (ODP/IODP) were used to analyze shear-wave velocity and Vp/Vs ratios, as well as their controlling factors. Results indicate that the depositional environment plays a significant role in influencing the shear-wave velocity and Vp/Vs ratios of marine sediments. Specifically, sediments from carbonate platforms, continental slopes, and continental rises generally exhibit higher shear-wave velocities and lower Vp/Vs ratios compared to those from abyssal plains, deep-water turbidites, and other deep-water environments. Lithofacies and physical compaction also strongly influence shear-wave velocity and Vp/Vs ratios. Diatomaceous sediments tend to have lower shear-wave velocities and higher Vp/Vs ratios compared to calcareous and terrigenous sediments. The Vp/Vs ratio sharply drops from 8 to less than 2 with increasing shear-wave velocity, indicating sensitivity to physical compaction. We propose empirical models that capture the relationship between the Vp/Vs ratio and shear-wave velocity for different lithofacies, with R² values exceeding 0.90.
