KMKurt J. Marfurt

Kurt J. Marfurt

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Kurt J. Marfurt has a total of 201 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    From Attributes to Geologic Interpretation:  A Critical Thinking Framework for Deepwater Channel Stratigraphic Analysis, Taranaki Basin

    by April Moreno-Ward, Heather Bedle, David Lubo-Robles, Kurt Marfurt

    Effective deepwater channel interpretation and identification of stratigraphic features require more than enhanced seismic visualization. For geoscientists looking to move beyond attribute generation into meaningful geologic interpretation, a systematic and process-based approach is essential. The interpreter must create and follow a systematic approach to translate observations into a defensible depositional model utilizing process-based geologic thinking. It requires the identification of key stratigraphic channel patterns (erosional boundaries and reflector termination characteristics such as onlap and downlap) that reveal depositional setting complexities while assessing confidence levels for each observation. Validation of interpretations across multiple display perspectives is a must, and critical evaluation of alternative scenarios should be considered to avoid blindly following meaningless patterns. We present a process-based decision framework incorporating critical thinking checkpoint questions at each workflow stage, applied to a deepwater channel complex in New Zealand's Taranaki Basin. Building upon the optimization of corendered displays of cosine instantaneous phase and relative acoustic impedance demonstrated in the companion paper, we systematically delineate architectural elements and reconstruct channel system evolution through deliberate observation protocols rather than just pattern matching. The Romney-1 well control provides critical calibration, enabling seismic facies characterization and interpolation from the well-bore into the channel fairway. These checkpoints and systematic framework produce more confidence in the interpretation of channel architecture in exploration settings where well control is limited. This process-based interpretation approach, which integrates depositional mechanics with pattern recognition, captures the full complexity of multi-phase channel evolution through critical thinking. This method requires interpreters to be transparent about what can be confidently interpreted from seismic data (geometry and architectural relationships), what can be inferred with lower certainty (detailed facies predictions), and what remains beyond the resolution of available data.

  • Detection of small-offset faults in seismic data: An application for carbon capture and storage

    by David Lubo-Robles, Heather Bedle, Kurt J. Marfurt

    Accurate seismic fault interpretation is critical to understanding the structural framework of the subsurface and mitigating potential hazards such as induced seismicity or potential fluid migration outside the target reservoir that might affect other formations such as shallow aquifers. In general, faults are identified by mapping discontinuities in the seismic amplitudes or edge-enhancing seismic attributes. However, detecting small-offset faults is challenging due to the lack of clear discontinuities. In this study, we apply volumetric aberrancy and multispectral coherence to (1) delineate faults that exhibit little to no offsets on a seismic volume acquired for carbon capture and storage in the Illinois Basin – Decatur Project and (2) correlate them with the microseismic activity during supercritical CO2 injection. We first apply spectral balancing to improve the vertical resolution of the seismic amplitudes volume. Although spectral balancing improves the resolution, not all improved spectral components aid in delineating subtle faults. Therefore, we only choose the spectral components that best enhance fault offsets to compute the multispectral coherence and quantify discontinuities. We also compute volumetric aberrancy to quantify flexures and their azimuths given by lateral changes in curvature. We find that these seismic attributes are helpful in delineating small-offset faults in the study area. We use a postconditioning workflow called fault enhancement to improve the resolution of the multispectral coherence and quantify the faults’ dip and azimuth. Based on the azimuth estimations, we extract faults aligned with or close to the maximum horizontal stress field. We observe that several of the extracted faults correlate well with the microseismic events, whereas other faults do not match any event, possibly due to the lack of pore pressure variations during injection, stratigraphic and structural variabilities, or because multispectral coherence and volumetric aberrancy do not directly measure fractures in seismic data.

  • Seismic attributes and fault-fold systems: A case study of quantitative analysis of coal seams in the Southern Qinshui Basin, China

    by Shu Yuan, Rulang Wang, Xiaojuan Huang, Brett M. Carpenter, Kurt J. Marfurt, Zonghu Liao

    The formation of folds often leads to the emergence of subsidiary faults, creating intricate fault-fold systems with fractures. Although existing literature on fault-fold systems predominantly relies on field observations, there has been scant research focusing on their subsurface structures. Using 3D seismic data, this study endeavors to identify and quantify the principal fault-fold systems within the no. 3 coal seam of the Shanxi Formation and the no. 15 coal seam of the Taiyuan Formation, in the southern Qinshui Basin, China. The findings reveal: (1) the presence of five principal fault-fold systems, with four exhibiting a north–south orientation and one aligning east–west; (2) the establishment of an asymmetric model for the fault-fold system through a quantitative analysis of seismic attributes, particularly focusing on the high-angle normal faults along the folds’ limbs; and (3) the asymmetrical fault damage zone width (no. 3 fault-fold system) is quantitatively measured at approximately 200 m on the west side, in contrast to the east side where widths range from 50 to 500 m. These initial assessments indicate the great potential of an attribute-enhanced geologic model, which may provide insights for analyses of the continuity of coal seams, the propagation of fractures, and the migration of gas within fractured reservoirs.

  • Postmigration seismic data conditioning workflow on a merged data set

    by Pamela Blanco Dufau, Heather Bedle, Kurt Marfurt, Brett M. Carpenter

    Analyzing amplitude anomalies in seismic data requires a comprehensive understanding of the geologic context and the accuracy of the seismic image. Over the past four decades, numerous surveys in mature basins such as the U.S. Gulf of Mexico have undergone reprocessing and merging to enhance imaging quality. The merging of seismic data volumes demands careful attention during processing, as the different volumes are often acquired at different times with different hardware, acquisition geometries, and exploration objectives. If insufficient care is taken, significant differences in the amplitude and spectra of the merged survey components can pose challenges when used as input for machine-learning techniques or seismic attribute interpretation and studies. We implement spectral balancing followed by structure-oriented filtering (SOF) to address the previously mentioned discrepancies in a merged survey. Spectral balancing equalizes high and low frequencies, creating a more uniform frequency spectrum. SOF eliminates random and cross-cutting coherent noise while preserving structural and stratigraphic features. This workflow ameliorates the discrepancies between the areas covered by the individual surveys, resulting in a more consistent interpretation across the seam between the two surveys. This study finds that these two processes improved the vertical resolution and lateral delineation of fault and stratigraphic edges. However, we also observe in our data that increasing the spectrum runs the risk of increasing the effect of low- and high-frequency noise, including acquisition footprint. Surprisingly, we find that spectral balancing can diminish the appearance of stratigraphic edges that are fortuitously imaged by the original narrowband data volume. Finally, because spectral balancing and SOF are set to not damage relative amplitudes, we need to apply amplitude gain control to balance them on the shallower parts of the merged survey.

  • Understanding uncertainty in deepwater channel seismic facies classification applying random forest on outcrop-constrained 3D models and synthetic seismic data

    by Karelia La Marca, Heather Bedle, Lisa Stright, Mario Ballinas, Kurt Marfurt

    To understand the uncertainties in seismic interpretation, especially for deepwater seismic facies, we apply a novel approach using outcrop-based synthetic data as ground truth. By applying the random forest (RF) supervised machine learning method we can better understand the influence of classifier hyperparameters on facies prediction accuracy. Based on previous analysis, we chose six seismic attributes that are able to differentiate five deepwater architectural facies: shale (thin-bedded turbidites), channel axis, off-axis, margin, and mass transport deposit. Hyperparameter testing indicated that optimization of the RF classifier is sensitive to the (1) choice of training attributes, (2) original facies proportions, (3) similarity in the seismic expression of different facies, and (4) seismic data resolution and seismic data signal-to-noise ratio. A simple classifier using common RF hyperparameters developed for fluid saturation predicted the facies with 74% accuracy. Although computationally more expensive, optimizing the RF hyperparameters provided approximately 89% accuracy, demonstrating the importance of hyperparameter tuning. When applying the best model to an unseen portion of the model, the position of the channel complex set was predicted accurately, but the accuracy was only 58%. This highlights the limitations of universal models and the persistent uncertainties faced when using ML to enhance our interpretations. Although we use a correlation matrix approach, we acknowledge that uncertainties can only be assessed and not quantified. Optimization of attribute choice and hyperparameter tuning reduce the epistemic uncertainty. We encourage future research to explore this in more detail.


  • Characterization of seismic-scale petrofacies variability in the Arbuckle Group using supervised machine learning: Wellington Field, Kansas

    by A. B. Caf, D. Lubo-Robles, K. J. Marfurt, H. Bedle, M. J. Pranter

    The Arbuckle Group in southern Kansas has been investigated for carbon geosequestration-related studies. In this study, we evaluated the seismic-scale petrophysically defined facies variability of the Arbuckle Group at the Wellington Field, Kansas, using quantitative seismic interpretation and a supervised Random Forest classification approach. We first defined three petrophysics-based rock types (petrofacies) from core-derived porosity and permeability measurements using the flow-zone indicator approach. Then, using the artificial neural network, we classified these petrofacies in noncored intervals. We observed that petrofacies 1 corresponds to medium and coarse-grained dolomitic packstone, wackestone, and dolomitic breccia with up to 8% porosity and D-scale permeability values. Whereas petrofacies 2 and 3 correspond to argillaceous and fine-grained micritic dolomites and dolomitic mudstones with lower permeability values for a given porosity, with respect to petrofacies 1. Using the common reflection-point gathers, we performed prestack seismic inversion and calculated various amplitude-variation-with-offset (AVO) attribute volumes. We used these elastic properties and AVO attribute volumes as input for estimating supervised seismic-scale 3D petrofacies and petrofacies probability volumes using the Random Forest algorithm. Results reveal the complex distribution of the petrofacies in the candidate injection and baffle zones in the study area, where petrofacies 1 is mainly prevalent within the lower and upper portions of the Arbuckle group, whereas petrofacies 2 and 3 are mainly present in the middle Arbuckle interval. The workflow we present through this study provides the spatial variability of facies distribution that is reflective of the actual lithology and petrophysical properties of the Arbuckle group in the study area with limited well control.


  • Uncertainty assessment in unsupervised machine-learning methods for deepwater channel seismic facies using outcrop-derived 3D models and synthetic seismic data

    by Karelia La Marca, Heather Bedle, Lisa Stright, Kurt Marfurt

    Unsupervised machine-learning (ML) techniques have been widely applied to analyze seismic reflection data, including the identification of seismic facies and structural features. However, interpreting the resulting clusters often relies on geoscientists’ expertise, necessitating a robustness assessment of these methods. To evaluate their reliability, synthetic data generated from an actual outcrop model are used to demonstrate how two unsupervised methods, self-organizing maps (SOMs) and generative topographic maps (GTMs), cluster deepwater channel-related seismic facies and then measure the associated error. Six seismic attributes, comprising root-mean-square amplitude, instantaneous envelope, peak magnitude, and spectral decomposition frequencies at 20, 40, and 55 Hz, served as input variables. Geobodies are assigned to each cluster formed, and the error in facies clustering is quantified by comparing the actual 3D model with the facies grouped by ML methods on a voxel-by-voxel basis. This allowed for error quantification and the computation of metrics such as F1 score and accuracy through correlation matrices. Key findings reveal that (1) GTM and SOM exhibit similar performance, with a clustering configuration of 81 for GTM slightly outperforming others. (2) Error rates are approximately 2% for the predominant facies (background shale) but significantly higher for individual channel-related facies, suggesting that channel clusters might represent multiple facies. (3) Resolution and imbalanced data distribution impact seismic facies predictability, resulting in nonuniqueness in cluster generation. (4) Using synthetic seismic data proved valuable for experimenting with different unsupervised MLs, highlighting the need for assessing uncertainty in these methods, given their implications for crucial economic decisions reliant on reservoir interpretation, modeling, and volumetric estimations.


  • Seismic reservoir characterization of the Gassum Formation in the Stenlille aquifer gas storage, Denmark — Part 1

    by Satinder Chopra, Ritesh Kumar Sharma, Kenneth Bredesen, Kurt J. Marfurt

    Seismic reservoir characterization plays an important role in carbon capture and storage analysis. The Havnsø anticlinal structure in Denmark is a prospective CO2 storage site due to its proximity to two large emission sources—a coal-fired power station and a nearby refinery. Although legacy 2D seismic lines over the area outline the anticlinal structure, their quality is insufficient for quantitative interpretation. Earlier studies have shown that the natural gas stored in the Stenlille aquifer exhibits a seismic response similar to the modeled CO2 fluid in the Havnsø structure. Thus, seismic reservoir characterization carried out on the Stenlille aquifer gas storage in terms of identifying spatial distribution of gas and outlining faults would provide insight regarding value addition that seismic data can bring into the proposed CO2 storage at Havnsø. Using the available poststack seismic data, we apply an integrated reservoir characterization analysis. After performing the adequate data conditioning, the impedance of the target Stenlille Formation is estimated through generation of an accurate low-frequency model. Thereafter, multiattribute analysis was used to generate volumetric estimates of porosity, gamma ray, and water saturation within the target formation so that the spatial distribution of gas can be mapped. The resulting porosity and gamma-ray volumes indicate encouraging results and were used for Bayesian classification to predict the probability of the more important lithofacies, namely, sand, shale, moderate-porosity sand, and moderate-porosity shaly sand, which enabled the mapping of high-porosity/facies zones in the two aquifer storage levels. Independently, we make use of unsupervised machine learning applications for seismic facies prediction and compare them at the two storage levels, which will be presented in the part 2 of this paper.


  • Seismic reservoir characterization of the Gassum Formation in the Stenlille aquifer gas storage, Denmark — Part 2: Unsupervised classification

    by Satinder Chopra, Ritesh Kumar Sharma, Kenneth Bredesen, Thang Ha, Kurt J. Marfurt

    Ideally, a good static reservoir model provides an accurate estimate of the extent, porosity, permeability, and lithology of the container as well as the properties of the seal and any faults or fractures that may allow the reservoir to leak. The major pitfall of deterministic, statistical, or supervised learning workflows is that they estimate only the properties sampled by the wells or provided by empirical relations and may miss mapping heterogeneities in the reservoir and seal that can give rise to flow baffles and reservoir leakage. This shortcoming is exacerbated when the number of wells is small, the types of logs recorded are limited, and the migrated seismic gathers are absent or of limited quality. In contrast, unsupervised learning looks for patterns in the seismic amplitude and attribute volumes themselves. In this paper, we apply unsupervised learning algorithms to evaluate the natural gas storage Stenlille aquifer in Denmark and compare the results with a supervised multiattribute regression reservoir characterization described in a companion paper. Specifically, we apply principal component analysis, self-organizing mapping, and generative topographic mapping workflows to extract patterns across eight attribute volumes: relative acoustic impedance, energy, sweetness, gray level co-occurrence matrices (GLCM) entropy, curvedness, and three spectral magnitude volumes. We find that the large-scale patterns are similar, but that the unsupervised learning algorithms provide greater detail. Because our deterministic model was built on poststack data using the limited well-log data available, we believe that the heterogeneity mapped by the unsupervised learning workflows provides a relatively unbiased means of estimating risk in our reservoir model. Quantifying the importance of these anomalies will need to be reconciled with a dynamic reservoir model.


  • Fault enhancement comparison among coherence enhancement, probabilistic neural networks, and convolutional neural networks in the Taranaki Basin area, New Zealand

    by José P. Mora, Heather Bedle, Kurt J. Marfurt

    Fault identification is a critical component of seismic interpretation. During the past 25 years, coherence, curvature, and other seismic attributes sensitive to faults improved seismic interpretation, but human interaction is still required to generate a complete fault interpretation. Today, image enhancement of fault-sensitive attributes, multiattribute fault analysis using shallow learning, and deep-learning algorithms based on extensive training and convolutional neural networks (CNNs) are promising fault interpretation workflows. We have compared three workflows to test fault-detection capabilities; these include image enhancement, probabilistic neural networks (PNNs), and CNNs. We compared results to human-interpreted faults as our ground truth for a merged 3D seismic survey acquired in the Taranaki Basin, New Zealand. We extracted fault surfaces from the results of the workflows using them as seed points for an active contour method. Extracted faults are then compared to the human-interpreted surface using the Hausdorff distance. Data conditioning, including spectral balancing and structure-oriented filtering, improved the performance of all three workflows. Although all three approaches produce enhanced fault volumes, we find differences in fault location and different artifacts (mispredicted faults). Because all three methods exhibit “false positive” predictions, the enhanced multispectral coherence method produces faults and stratigraphic edges in the final image, including residual stair-step artifacts. In our implementation, PNN produces many salt-and-pepper artifacts through the resulting image, suggesting that we might need to include better training data or reduce the volume size to reduce the number of relevant classes to obtain an improved classification. The CNN algorithm is trained with synthetic data that provide rapid results, correctly identifying larger faults, but missing smaller faults and, in some cases, misclassifying mass-transport deposits and angular unconformities as being faults.