Pre-stack Data Research Articles

Irregularly sampled seismic data interpolation via wavelet-based convolutional block attention deep learning

Seismic data interpolation, especially irregularly sampled data interpolation, is a critical task for seismic processing and subsequent interpretation. Recently, with the development of machine learning and deep learning, convolutional neural networks (CNNs) are applied for interpolating irregularly sampled seismic data. CNN based approaches can address the apparent defects of traditional interpolation methods, such as the low computational efficiency and the difficulty on parameters selection. However, current CNN based methods only consider the temporal and spatial features of irregularly sampled seismic data, which fail to consider the frequency features of seismic data, i.e., the multi-scale features. To overcome these drawbacks, we propose a wavelet-based convolutional block attention deep learning (W-CBADL) network for irregularly sampled seismic data reconstruction. We firstly introduce the discrete wavelet transform (DWT) and the inverse wavelet transform (IWT) to the commonly used U-Net by considering the multi-scale features of irregularly sampled seismic data. Moreover, we propose to adopt the convolutional block attention module (CBAM) to precisely restore sampled seismic traces, which could apply the attention to both channel and spatial dimensions. Finally, we adopt the proposed W-CBADL model to synthetic and pre-stack field data to evaluate its validity and effectiveness. The results demonstrate that the proposed W-CBADL model could reconstruct irregularly sampled seismic data more effectively and more efficiently than the state-of-the-art contrastive CNN based models.

An Automatic Velocity Analysis Method for Seismic Data-Containing Multiples

Normal moveout (NMO)-based velocity analysis can provide macro velocity models for prestack data processing and seismic attribute inversion. Datasets with an increasing size require conventional velocity analysis to be transformed to a more automatic mode. The sensitivity to multiple reflections limits the wide application of automatic velocity analysis. Thus, we propose an automatic velocity analysis method for seismic data-containing multiples to overcome the limit of multiple interference. The core idea of the proposed algorithm is to utilize a multi-attribute analysis system to transform the multiple attenuation problem to a multiple identification problem. To solve the identification problem, we introduce the local similarity to attribute the predicted multiples and build a quantitative attribute called multiple similarity. Considering robustness and accuracy, we select two supplementary attributes based on velocity and amplitude difference, i.e., velocity variation with depth and amplitude level. Then we utilize the technique for order preference by similarity to ideal solution (TOPSIS) to balance weights for different attributes in automatic velocity analysis. An RGB system is adopted for multi-attributes fusion in velocity spectra for visualization and quality control. Using both synthetic and field examples to evaluate the effectiveness of the proposed method for data-containing multiples, the results demonstrate the excellent performance in the accuracy of the extracted velocity model.

An integrated study on fracture prediction using 3D P-wave field seismic data

An integrated study on fracture prediction by analyzing seismic anisotropy within a 3D P-wave seismic data set from an oilfield in Northwest China is performed. The research area is dominated by a regional anticline and the target layer is a fractured gas reservoir at a depth of approximately 3 km under the anticline. Two major aspects are considered. First, two P-wave poststack attributes (reflection strength and frequency content in selected lines) are extracted and their features in the target layer are examined, looking for the possible fracture distribution in two prospective zones. Second, the azimuthal attribute analysis approach is applied to estimate fracture density and orientation in the whole area from the prestack data (amplitude and traveltime). The result from the poststack data indicates the high possibility of fracture distributions in these two zones, and the prediction from azimuthal amplitudes in the prestack data is consistent with the fracture orientation from the well interpretation and the observation of outcrops at the top of the anticline but suffers from a poor offset-depth ratio and acquisition footprint. The prediction from azimuthal traveltime is strongly influenced by the regional anticline structure and, therefore, is not reliable. Nevertheless, the prediction from post- and prestack attributes provides more details of the fracture distribution in the target reservoir, which can expand understanding of the reservoir fracture distribution.

A practical approach to develop a proper anisotropic rock physics model for media comprising of multiple fracture sets in the absence of sophisticated laboratory/wireline data

For economical production from a fractured reservoir, a characteristic analysis of the fracture parameters like its density and orientation within the reservoir is essential to improve the fluid flow during extraction. This study deals with the development of a proper anisotropic rock physics model for a media with multiple fracture sets to study the spatial distribution of important fracture parameters i.e., fracture density and orientation in the absence of sophisticated laboratory/wireline and pre-stack seismic data. The crest of hydrocarbon producing fault-bounded Balkassar Anticline in Northern Potwar, Upper Indus Basin, Pakistan is selected as a case study representing a potential zone for development of fractures at reservoir level (Sakesar Limestone). The methodology consists of the interpretation of 3D post-stack seismic and conventional wireline log data to demarcate the reservoir containing fractures. The Ant-tracking discrete fracture network (DFN) attribute is applied on 3D post-stack seismic data to obtain an initial estimate about the presence of fracture corridors and their orientations. Based on this initial estimate, a proper rock physics model has been developed utilizing inverse Gassmann relations, T-matrix approximation, and Brown and Korringa relations. The output from the developed rock physics model has been displayed in the form of 13 effective independent elastic stiffness constants (monoclinic symmetry–representing media comprising of multiple fracture sets) as a function of fracture densities and azimuthal fracture orientations. A clear decreasing trend in effective elastic stiffness constants with increasing fracture densities can be observed. Similarly, a periodic trend of effective elastic stiffness constants with fracture orientations can be observed. These trends are more or less expected, but they would have been difficult to quantify without a proper rock physics model. The use of independent effective elastic constants for the generation of synthetic seismic amplitude versus angle and azimuth (AVAZ) data and its correlation with observed seismic AVAZ data in a geostatistical sense has been discussed.

Accelerating seismic scattered noise attenuation in offset-vector tile domain: Application of deep learning

Recent years have witnessed many practical applications of supervised deep learning in seismic processing. However, a weak generalization behavior prevents widespread implementation on large-scale prestack data sets for coherent noise attenuation. This is particularly true when addressing strong near-surface scattered noise in land seismic data. To alleviate this problem, we have combined deep learning with an offset-vector tile (OVT) partitioning method to suppress strong scattered noise. With the OVT partitioning method, seismic data are spatially uniformly sampled, offering a favorable foundation for network learning. Specifically, the reflection probability distribution is more stationary than the noise distribution, making it easier for the network to learn the reflections. Accordingly, we use the direct signal learning strategy rather than the commonly used residual learning strategy to train the network. To construct high-quality training labels, we adopt the 3D continuous wavelet transform (3D CWT), which can exploit the 3D spatial correlation in OVT gathers. General use of these labels can produce results similar to 3D CWT but is highly efficient. To further improve denoising performance, we propose a training sample construction approach that leverages middle-offset OVT volumes with varying azimuths in light of midoffset relatively high signal-to-noise ratio characteristics. The field data experiment demonstrates that our proposed method also has an excellent generalization ability. Despite only using six middle-offset gathers for training, the trained network is able to effectively process 1260 OVTs in a timely manner.

Robust prestack seismic facies analysis using shearlet transform-based deep learning

Abstract One of the primary purposes of seismic stratigraphy is to evaluate the components of seismic layer relationships within a depositional chronology. Prestack seismic images contain a wealth of information, such as variations in the offset and azimuth of a seismic event, and naturally produce higher-resolution seismic facies analysis results than poststack data. However, prestack data usually suffer from potential unreliability issues due to low signal-to-noise ratios. As this is often overlooked, the present facies analysis methods sometimes fail to extract accurate features from prestack images, which inevitably influences the facies analysis results. To address this issue, this article provides a robust data-driven technique for extracting offset-temporal features via shearlet transform-based deep convolution autoencoders (STCAEs). Unlike the present facies analysis in the time domain, STCAE can optimally represent prestack images at multiple scales and directions through the two-dimensional shearlet transform, which preserves fine edges while suppressing noise in prestack images. Subsequently, robust features are extracted from prestack images in a data-driven manner through a contractive convolutional autoencoder network. We compare our method with other advanced methods and demonstrate the advantages of the proposed approach in classifying seismic layers in prestack data.

Two-stage semi-supervised learning inversion for reservoir physical parameters

Reservoir physical parameters are crucial indexes for reservoir description. Deep learning is a data-driven algorithm, which may avoid some issues in conventional deterministic and stochastic inversion methods, such as cumulative and approximate errors. We develop a two-stage semi-supervised learning inversion method to predict porosity, shale content, and water saturation. Firstly, an inversion network is designed to predict physical parameters from pre-stack seismic data and initial models. The Spatio-temporal characteristics of seismic data and the low-frequency structural information of the initial model can be utilized efficiently by the network. In addition, considering the high production cost of label data in exploration geophysics, we construct a semi-supervised learning inversion workflow to relieve the dependence on the label in training. The rock physics model and seismic convolution forward model are invoked as a forward process in the workflow. Finally, we propose a two-stage inversion strategy to deal with the inversion of physical parameters with different data characteristics. The proposed method is applied to a field survey successfully. A high-resolution and high-accuracy inversion result can be obtained with our method, and potential locations of the gas reservoir can be indicated by prediction results. Meanwhile, the comparisons of the two stages demonstrate that illusions in predicted water saturation can be eliminated effectively.

Characterizing a Wedged Chalk Prospect in the Danish Central Graben Using Direct Probabilistic Inversion

A novel direct probabilistic inversion using seismic pre-stack data as input to characterize a wedged chalk reservoir prospect was demonstrated from the Upper Cretaceous unit, Danish North Sea. The objective was to better resolve the lateral extent and pinch-out of the chalk prospect in a frontier exploration setting and compare the results with a more traditional deterministic inversion and geostatistical reservoir modeling. The direct probabilistic inversion results provided additional reservoir insights that were challenging to obtain from the more traditional workflows and are also more flexible for associated uncertainty assessments. Hence, this study demonstrates the usefulness of such direct probabilistic inversions even with suboptimal data availability.

Estimation of porosity from well logs and seismic using artificial neural network

The Artificial Neural Network (ANN) is widely used to map and estimate reservoir properties. Since ANN has the ability of non-linear computing and self-error correction, it serves as an alternative method to enhance reservoir characterization by improving reservoir properties' prediction. This study employed an integrated approach where seismic, and well log data are used alongside ANN to evaluate petrophysical properties in Field X, South Caspian Basin. The study field is situated in the South Caspian Basin, which developed during Tertiary-Quaternary. The South Caspian Basin covers the southern part of the Caspian Basin, coastal regions of eastern Azerbaijan, northern Iran, and western Turkmenistan. The field is geologically situated on an elongated and multi-crest, anticlinal feature, better known as Apsheron Sill. Available data set consists of log data from four wells: X1, X2, X3, X4, and 3D pre-stack seismic data. Firstly, the reservoir properties were calculated from the well log data. The simultaneous inversion was carried out to obtain elastic rock properties like the P-impedance, S-impedance, density. After that, elastic properties were used to obtain 3D dimensional SQp and SQs attributes. Furthermore, the Radial basis function Neural Network (RBFN) model was created with optional inputs: volume of P-impedance, S-impedance, density, SQp, SQs and property logs and outputs: porosity. Moreover, this research work demonstrates an application of SQp and SQs attributes for properties prediction using neural network. The neural network's primary purpose is to determine a relationship between obtained volumes and reservoir parameters, porosity at well locations. The RBFN model is successfully trained and validated on the field data. The results demonstrated an excellent correlation between actual property logs. They predicted properties, which gives confidence for spatial prediction in areas where well logs are not available.

AVA inversion of pre-stack seismic data with cross-gradient constraints

The ill-posed feature is one basic attribute of geophysical inversion methods. As an example of geophysical inverse problem, AVA (Amplitude variation with incident angle) inversion of pre-stack seismic data is susceptible to noise and uncertainty in the acquisition. To get stable and accurate inversion results, the regularization constraints on model parameter need to be added into the objective function of AVA inversion. In AVA inversion, the most commonly used regularization includes sparsity constraint (e.g. L1-norm regularization, Cauchy regularization) and a priori model parameters constraint, and so forth. However, the existing AVA inversion methods do not consider the structural similarity of different model parameters. All of the different model parameters represent the same underground geological structure, so they should have similar structure. This paper adopts the cross gradient to measure the structural similarity of different model parameters. Next, the cross gradients of different model parameters are added into the objective function of AVA inversion as a regularization term to implement structural similarity constraint. Results of the model numerical tests and real seismic data indicate that the AVA inversion with cross-gradient constraint has higher stability compared to the AVA inversion without structural similarity constraint, especially for the density inversion results.

Warped mapping–based blind deconvolution for resolution improvement

ABSTRACTOwing to the influence of absorption and scattering, the components of seismic waves, especially those with higher frequencies, always suffer amplitude attenuation and phase distortion during their propagation through the earth. Nonstationary blind deconvolution takes such attenuations into account for prestack seismic data. This not only compensates for attenuation‐induced energy loss but also simultaneously addresses the energy loss associated with the reflectivity series and wavelet. Additionally, nonstationary blind deconvolution can greatly improve the resolution of seismic data. However, prestack nonstationary blind deconvolution is difficult to implement because it is not easy to estimate the attenuation function of prestack seismic data unless the quality factor of the Earth is assumed to be constant. In this regard, we introduced a nonstationary blind deconvolution method for prestack gathering. First, we used warped mapping to calculate the attenuation function of the prestack data. Then, we incorporated the attenuation function into the nonstationary sparse spike deconvolution based on Toeplitz‐sparse matrix factorization for reflectivity series and wavelet estimation. The proposed method is velocity‐independent and can be adapted to the time‐varying Q model. To validate its stability and effectiveness, numerical and real data examples were adopted. The results show that the proposed method provides a straightforward routine for the estimation of prestack reflectivity coefficients for further processing and inversion.

Porosity prediction from pre-stack seismic data via committee machine with optimized parameters

Prediction of porosity from the seismic data via geophysical methods when limited number of wells are available is a challenging task that has high uncertainties. This study aims to construct a hybrid data-driven predictive model to establish a quantitative correlation between seismic pre-stack (SPS) data and the porosity. First, three intelligent models that are optimized by bat-inspired algorithm (BA): optimized neural network (ONN), optimized fuzzy inference system (OFIS), and optimized support vector regression (OSVR) are constructed for relating porosity to the SPS data. Then, to benefit from all individual optimized models, a final hybrid model was built via committee machine (CM) where single models are combined with a proper weight to predict porosity in the reservoir space. This approach is examined on the SPS data from an oil field in the Persian Gulf with a single exploratory well where input parameters (Vp, Vs, and ρ) to the AI models are derived from a two-parameter inversion method. We found that the coefficient of determination, root mean square error, average absolute relative error, and symmetric mean absolute percentage error for the CM are 0.923615, 0.015793, 0.132280, and 0.061310, respectively. Moreover, based on four statistical indexes that are calculated for each model, CM outperformed its individual elements followed by the OSRV. A comprehensive analysis of the results confirms that CM with the OM elements is a superior approach for computing porosity from the SPS in the well and then throughout the entire reservoir volume. This strategy can aid petroleum engineers to have a better forecast of porosity population in the reservoir static model immediately following the data that is obtained from the first exploratory well. Ultimately, successful implementation of this approach will promptly delineate sweet spots that can replace uncertain and complicated conventional geophysical methods.

A New Stochastic Process of Prestack Inversion for Rock Property Estimation

In order to enrich the current prestack stochastic inversion theory, we propose a prestack stochastic inversion method based on adaptive particle swarm optimization combined with Markov chain Monte Carlo (MCMC). The MCMC could provide a stochastic optimization approach, and, with the APSO, have a better performance in global optimization methods. This method uses logging data to define a preprocessed model space. It also uses Bayesian statistics and Markov chains with a state transition matrix to update and evolve each generation population in the data domain, then adaptive particle swarm optimization is used to find the global optimal value in the finite model space. The method overcomes the problem of over-fitting deterministic inversion and improves the efficiency of stochastic inversion. Meanwhile, the fusion of multiple sources of information can reduce the non-uniqueness of solutions and improve the inversion accuracy. We derive the APSO algorithm in detail, give the specific workflow of prestack stochastic inversion, and verify the validity of the inversion theory through the inversion test of two-dimensional prestack data in real areas.

SH-SH wave inversion for S-wave velocity and density

SS-waves (SV-SV waves and SH-SH waves) are capable of inverting S-wave velocity ([Formula: see text]) and density ([Formula: see text]) because they are sensitive to both parameters. SH-SH waves can be separated from multicomponent data sets more effectively than the SV-SV wave because the former is decoupled from the PP-wave in isotropic media. In addition, the SH-SH wave can be better modeled than the SV-SV wave in the case of strong velocity/impedance contrast because the SV-SV wave has multicritical angles, some of which can be quite small when velocity/impedance contrast is strong. We derive an approximate equation of the SH-SH wave reflection coefficient as a function of [Formula: see text] and [Formula: see text] in natural logarithm variables. The approximation has high accuracy, and it enables the inversion of [Formula: see text] and [Formula: see text] in a direct manner. Both coefficients corresponding to [Formula: see text] and [Formula: see text] are “model-parameter independent” and thus there is no need for prior estimate of any model parameter in inversion. Then, we develop an SH-SH wave inversion method and demonstrate it by using synthetic data sets and a real SH-SH wave prestack data set from the west of China. We find that [Formula: see text] and [Formula: see text] can be reliably estimated from the SH-SH wave of small angles.

Reconstructing seismic data by incorporating deep external and internal learning

Seismic data reconstruction is an inverse problem in the geophysical community. Deep learning-based methods directly learn the projection between undersampled and complete data from large training datasets. However, when the feature difference between the test and training datasets increases, the recovery performance is degraded. Fine-tuning the pretrained network on the undersampled test field data has been introduced to adapt the trained network to new data. However, the pseudo-labels of the fine-tuning process need to be obtained by some traditional method in advance. In this paper, we explore a strategy for interpolating regularly sampled aliased seismic data by incorporating deep external and internal learning, i.e. by combing external pretraining with internal unsupervised fine-tuning. The pretrained network is obtained from an external synthetic dataset. Then, the pretrained network is fine-tuned on the internal dataset to obtain the final trained network adapting the test data. The inputs and labels of the internal dataset are generated solely from the currently regularly undersampled test data. As such, the fine-tuning process is totally unsupervised. Finally, the interpolated result is acquired by feeding the test data into the trained network. The proposed algorithm learns the external priors from large external datasets and it learns the internal features from undersampled test data. Thus, the network has an enhanced capacity to adapt to new data. Poststack and prestack field data are provided to verify the performance of the proposed algorithm.

AVO analysis for high amplitude anomalies using 2D pre-stack seismic data

Amplitude variation with offset (AVO) analysis is an 1 efficient tool for hydrocarbon detection and identification of elastic rock properties and fluid types. It has been applied in the present study using reprocessed pre-stack 2D seismic data (1992, Caulerpa) from north-west of the Bonaparte Basin, Australia. The AVO response along the 2D pre-stack seismic data in the Laminaria High NW shelf of Australia was also investigated. Three hypotheses were suggested to investigate the AVO behaviour of the amplitude anomalies in which three different factors; fluid substitution, porosity and thickness (Wedge model) were tested. The AVO models with the synthetic gathers were analysed using log information to find which of these is the controlling parameter on the AVO analysis. AVO cross plots from the real pre-stack seismic data reveal AVO class IV (showing a negative intercept decreasing with offset). This result matches our modelled result of fluid substitution for the seismic synthetics. It is concluded that fluid substitution is the controlling parameter on the AVO analysis and therefore, the high amplitude anomaly on the seabed and the target horizon 9 is the result of changing the fluid content and the lithology along the target horizons. While changing the porosity has little effect on the amplitude variation with offset within the AVO cross plot. Finally, results from the wedge models show that a small change of thickness causes a change in the amplitude; however, this change in thickness gives a different AVO characteristic and a mismatch with the AVO result of the real 2D pre-stack seismic data. Therefore, a constant thin layer with changing fluids is more likely to be the cause of the high amplitude anomalies.

Low-Frequency Extrapolation of Prestack Viscoacoustic Seismic Data Based on Dense Convolutional Network

Low frequency information in seismic data can improve seismic resolution and imaging accuracy, enhance the quality of inversion, and play an essential role in imaging algorithms such as full-waveform inversion. Sufficiently low frequency data can avoid the cycle skipping phenomenon during full-waveform inversion. During seismic data processing, the protection and reconstruction for low frequency information are therefore of great importance. In this paper, we systematically investigate the extrapolation of pre-stack viscoacoustic seismic low frequency data using a dense convolutional network to effectively establish the nonlinear relationship between high and low frequency data, and realize the extrapolation and reconstruction of viscoacoustic 0-5 Hz low frequency data using 5-30 Hz high-frequency component. And the generalizability of the method for different influencing factors such as wavelets, noise, and models is analyzed using Marmousi2 velocity model forward data. It is demonstrated that the method has high robustness and can be applied to different situations, and the accuracy is higher than that of the traditional convolutional neural networks method. The feasibility of the low frequency extrapolation method based on dense convolutional network is also verified by synthetic data, physical experiment simulation data, and field data testing, and superior to the traditional convolutional neural networks method.

Density stability estimation method from pre-stack seismic data

As one of the most important elastic parameters in seismic exploration, density plays an important role in lithology identification and hydrocarbon indicator. Based on the approximate Zoeppritz equation, the density term is usually estimated using AVO/AVA (Amplitude variation with offset/incident angle) inversion. However, the small contribution of density to the reflection coefficient or the strong correlation with other parameters in the approximate equation renders the density estimation tough. By analyzing the contribution of density to reflection coefficient in different approximate equations, Aki-Richards is finally selected as the inversion equation, in which the density term has a higher contribution to the reflectivity. The conventional inversion approach is to invert several parameters simultaneously on the basis of approximate equation. The same convergence criterion is used for the estimation of the three parameters, but the density estimation error fails to converge to the minimum due to its small sensitivity, which makes density inversion unstable. Furthermore, the conventional inversion methods involve the inverse of large matrix, which also reduces the stability of density estimation. Therefore, a new form of Aki-Richards approximate reflection equation is derived on the assumption that the inverse of coefficient matrix exists. The reflectivity of elastic parameter is independently expressed as the weighted superposition of seismic reflection coefficients at different incident angles. Then utilizing well logging data and seismic reflection coefficient estimated by elastic impedance inversion as constraints, the inverse of coefficient matrix in inversion equation is estimated directly, which avoids inverse of large matrix and improves the precision of density estimation. In physical, the weighted coefficients directly reflect the contribution of elastic parameters reflectivity to seismic data. Finally, different experimental examples show that proposed method can stably estimate density term.

Evaluation of the Influence of Shale on the Petrophysical Properties of Hydrocarbon-Bearing Reservoir Sand in “CAC” Field in the Niger Delta, Nigeria

This study aims at evaluating the influence of the presence of shale on the quality of reservoir sand in “CAC-Field”, Coastal swamp Niger Delta by integrating suites of well logs and 3D pre-stack seismic data. Shales in the reservoir pose interpretation challenges as they form baffles to fluid flow and reduce effective porosity. The data used included well logs (density, gamma ray, neutron, resistivity) and 3D seismic data. Petrel and Interactive Petrophysics software were adopted for the analyses. The Vclay/effective porosity cross-plots were used to determine the clay distribution patterns hence the influence of shale on the petrophysical properties of the hydrocarbon reservoir. Result of the well correlation yielded 12 reservoirs with 4 (RES 4 - RES 7) being hydrocarbon bearing and laterally continuous across the 4 wells, (CAC-1 - CAC-4) forming the focus of the study. Evidence of an NW-SE trending delta progradation in the CAC field is represented by the increasing sandiness downdip, at both intermediate and the shallow horizons. Thickening of the reservoir in some instances may be structurally controlled due to faulting. The results from the petrophysical evaluation show Vclay ranges of 13% - 21% and good to very good porosity values that vary from 15% - 25%. The permeability range from 240.49 - 2406.49 mD except for the sands in RES 7, CAC-3 well where the permeability was low (91 mD). Additionally, the Vclay/Effective Porosity cross-plots indicate essentially laminated and structural clay types with few dispersed clay in RES 7, CAC-3 well. The existence of these 3 clay types did not significantly influence the quality of the sands containing the hydrocarbon in the area, except in RES 7, CAC-3. The compartmentalizing effect of the laminated clay/shale could only possibly affect the vertical flow due to possible baffles to the vertical flow, but the horizontal flow may not have been impeded significantly. The study of the type and pattern of clay has helped to better evaluate the quality and mobility trend of the hydrocarbon in the CAC field.

Evaluating strategies for estimation of local kinematic parameters in noisy land data: quality versus performance trade-offs

Abstract Modern land seismic data are typically acquired using high spatial trace density with small source and receiver arrays or point sources and sensors. These datasets are challenging to process due to their massive size and relatively low signal-to-noise ratio caused by scattered near-surface noise. Therefore, prestack data enhancement becomes a critical step in the processing flow. Nonlinear beamforming had proved very powerful for 3D land data. However, it requires computationally intensive estimations of local coherency on dense spatial/temporal grids in 3D prestack data cubes. We present an analysis of various estimation methods focusing on a trade-off between computational efficiency and enhanced data quality. We demonstrate that the popular sequential «2 + 2 + 1» scheme is highly efficient but may lead to unreliable estimation and poor enhancement for data with a low signal-to-noise ratio. We propose an alternative algorithm called «dip + curvatures» that remains stable for such challenging data. We supplement the new strategy with an additional interpolation procedure in spatial and time dimensions to reduce the computational cost. We demonstrate that the «dip + curvatures» strategy coupled with an interpolation scheme approaches the «2 + 2 + 1» method's efficiency while it significantly outperforms it in enhanced data quality. We conclude that the new algorithm strikes a practical trade-off between the performance of the algorithm and the quality of the enhanced data. These conclusions are supported by synthetic and real 3D land seismic data from challenging desert environments with complex near surface.