After multistage tectonic movement and evolution, large superimposed oil and gas basins generally developed many igneous rocks in the early rifting stages. The lithology and lithofacies of igneous rocks are complex, which is easy to lead to the distortion of the underlying migration velocity field and thus the response of seismic pseudofaults. Also, because of the obvious shielding and absorption effect of igneous rocks on seismic waves, the waveform quality of underlying strata is poor and the seismic response characteristics of faults are fuzzy. Currently, relevant studies have shown that the influence of igneous rock can be eliminated by the prestack depth migration with an accurate igneous rock velocity model. However, improving the accuracy of the velocity model needs to be corrected by well-logging data, resulting in poor applicability of the existing velocity modeling technology underlying igneous rocks without well, which is an obvious technical bottleneck. In this paper, the secondary strike-slip fault in Shuntuoguole low uplift of Tarim Basin, which has great oil and gas exploration potential but a very low degree of drilling, is selected as the research object. Aiming at difficult fault detection underlying igneous rocks caused by lack of drilling, the accuracy of fault seismic identification is improved by “interpretative fault preprocessing” and “fault sensitive attribute optimization.” In addition, through the “extreme hypothesis method” to maximize the complex migration velocity and simulate the underlying target layer distortion maximization, we realize the quantitative elimination of seismic pseudofaults. The practical application shows that this technology can determine the true and fake underlying faults quantitatively without establishing an accurate igneous rock velocity model. It is crucial not only for exploring oil and gas in the Tarim Basin’s secondary strike-slip faults but also for offering a method and technical guide for identifying faults in other basins affected by igneous rocks.

Large superimposed oil and gas basins have undergone multiple periods of tectonic movement and evolution and generally experienced multiple periods of strong magmatic activity in the early stages of rifts or rifts, preserving numerous igneous rocks. As a high-velocity rock mass, igneous rock has two major impacts on the precise structural imaging of its underlying strata [1]. First, igneous rocks strongly shield and absorb seismic waves, leading to the blurring of seismic response characteristics of small structures and faults in the underlying strata. Second, the uneven distribution of thickness, the lateral and vertical variability of lithology, and the significant velocity differences between different lithologies of igneous rock bodies make it difficult to accurately describe the areal distribution, thickness, and velocity of high-velocity igneous rocks before migration imaging. This reduces the accuracy of the migration velocity field and commonly results in false structures and pseudofaults in the underlying seismic data [2-7]. These factors limit the accurate recognition of oil and gas reservoirs underlying igneous rocks. Therefore, in areas with igneous rock development, improving the precision of structural seismic identification and eliminating seismic structural artifacts have become core issues in oil and gas exploration and evaluation.

The Tarim Basin is the largest inland petroliferous basin in China, with prospective resources of 258.89 × 108 t, of which 38% is mainly distributed in the Ordovician carbonate rocks [8]. In the early stage, guided by the theory of “oil controlled by paleo-uplift and enriched in slope” [9], the Lungu-Tahe oilfield, the largest weathering crust oilfield in China, and the Tazhong I gas field, the largest condensate gas field in China, were discovered in the Tabei uplift, Tazhong uplift, and their slopes [10]. With the deepening of the geological theories of hydrocarbon accumulation in carbonate rocks and the advancement of seismic and engineering technologies, the exploration scope has expanded from paleo-uplifts and paleo-slopes to depressions, and the exploration targets have transited from carbonate buried-hill karst and reef-bank karst reservoirs to interlayer karst reservoirs and then to fault-controlled karst reservoirs [11]. CNPC and Sinopec have obtained high-yield oil and gas flows through drilling in several NE-trending main strike-slip faults in the Shuntuoguole uplift, forming the Shunbei and Fuman oil and gas fields with the capacity at billion-ton level each [12-14] (Figure 1(a)). Based on the comprehensive analysis of structural evolution, reservoirs, accumulation and hydrocarbon enrichment of the Shunbei and Fuman oil and gas fields, it is considered that multiphase activities of strike-slip faults and fluid dissolution process formed good fractured-cavity reservoirs in the Yijianfang Formation(O2yj), Yingshan Formation(O1-2y), and Penglaiba Formation(O1p) of Middle–Lower Ordovician carbonate rocks, and the tight carbonate rocks around the fault zone act as lateral seals, forming the physical trap. The overlying very thick Sangtamu Formation(O3s) mudstone is the regional seal, and hydrocarbons from the Yuertusi Formation(Є1y) of Lower Cambrian migrated vertically along the strike-slip faults and accumulated finally (Figure 1(b)). The strike-slip faults have obvious characteristics of “controlling reservoir development, hydrocarbon accumulation, and hydrocarbon enrichment,” and the oil and gas exploration often starts from the identification, interpretation, and confirmation of strike-slip faults [12, 15-17]. Accordingly, a number of scholars proposed that the widely developed NNE-trending secondary strike-slip faults in the Ordovician may also be favorable destinations for hydrocarbon migration and accumulation [11, 13] and set a new strategic goal (i.e., controlling main fault zones and exploring new types of targets in secondary fault zones) [16, 18, 19]. However, due to the small extension scale of the secondary strike-slip fault in the plane (≤50 km) and the inconspicuous fault characteristics in the profile, the widely developed Permian igneous rocks in the study area seriously affect its correct imaging.

Currently, the fault and reservoir identification underlying igneous rocks mainly rely on prestack migration data. Prestack time migration (PSTM) is based on the velocity model of a homogeneous medium or horizontal layered medium [20]. When the lateral velocity changes greatly, PSTM fails to image the reflected waves, so the influence of igneous rocks cannot be eliminated [21]. Prestack depth migration (PSDM), developed from the real geological depth model considering lateral velocity variation [20], can truly reflect the lateral change of interval velocity, and theoretically restore the real structural characteristics of strata underlying igneous rocks. The PSDM algorithm is relatively mature and serves as a major tool in eliminating seismic pseudofaults underlying igneous rocks. As to imaging in the depth domain, velocity modeling is the core procedure. Zhang et al. used VSP and other prior information to extract the low-frequency components of seismic imaging velocity accurately, and combined with three-dimensional grid tomography to velocity-depth model global optimization, this method can reduce the multisolution in igneous rock velocity calculation [21]. Ma et al. combined inversion algorithm with geological horizon interpretation on the basis of velocity inversion to further improve the accuracy of velocity model and imaging [18]; Zhu et al., Li et al., and Mu et al. constructed a high-precision velocity model by performing tomographic iteration of the initial velocity from large scale to small scale, and then adding well-logging constraints and structural constraints to the objective function term of velocity modeling, so that the velocity variation characteristics of igneous rocks can be described more precisely [22-24]. Through the continuous improvement of the imaging method and velocity modeling, as well as the comprehensive application of multidisciplinary, the imaging accuracy underlying igneous rock has been continuously improved. However, the above-mentioned velocity modeling methods are effective only when logging velocity is fully used, and they work well in areas with many wells, but not in no-well igneous rock areas. The drilled wells in Shuntuoguole low uplift are basically deployed along the main strike-slip faults. Subject to the discontinuous lithology of Permian igneous rocks, the logging velocity can only reflect the local igneous rock velocity information, and cannot reflect the wide range of igneous rock velocity changes. Therefore, for the secondary fault development area outside the main fault zone, we cannot modify and verify the velocity model due to the lack of drilling well, and thus we cannot eliminate the influence of igneous rocks on the accurate fault imaging completely, which restricts the oil and gas exploration process of secondary strike-slip faults.

In order to address the two effects caused by igneous rocks, “interpretative fault preprocessing” and “fault sensitive attribute optimization” are used to improve the seismic identification accuracy of secondary strike-slip faults. In addition, through the “extreme hypothesis method” to maximize the complex migration velocity, we simulate the velocity distortion maximization of the underlying target layer and conduct quantitative identification of seismic pseudofaults. The results of this paper can break through the bottleneck that the existing igneous rock velocity modeling technology is too dependent on the well-logging velocity, which is of great scientific significance to supplement and improve the existing technical system. As well, it is also of great application value to reduce the exploration risk of secondary strike-slip faults in the study area and expand oil and gas productivity.

Due to the shielding and absorption of igneous rocks on seismic waves, small-scale secondary strike-slip faults often reflect fuzzy seismic response characteristics. A higher-accuracy identification of secondary strike-slip faults, which are also important hydrocarbon migration pathways and reservoir spaces [12-15], is more conducive to correct reserves estimation. The accuracy of seismic fault identification can be improved by technologies roughly in two categories: (1) the migration imaging technology, which is designed to improve the imaging quality of seismic data by enhancing the resolution and signal-to-noise ratio (SNR) of seismic data [25, 26]; and (2) the fault prediction technology, which is designed to increase the noise immunity and resolution of specific seismic attributes (e.g., coherence, curvature, and likelihood) by improving relevant algorithms, depending on the geologic genesis and characterization of faults [27]. The former is not applicable to no-well areas due to its dependence on logging data, while the latter is performed satisfactorily in no-well areas due to its less time consumption and low cost [28]. Excellent algorithm can improve the fault prediction accuracy obviously, and it is the core of fault prediction technology. However, the algorithm modification needs a high geophysical, mathematical, and computer programming capability, which is difficult for geological researchers. Based on the in-depth analysis of the principle and applicability of existing fault prediction technologies, we find that the prediction result is affected by both fault preprocessing data and subsequent processing of fault attributes. Therefore, in this study, according to the characteristics of different algorithms, the input data and output results are processed respectively before and after the computation of fault attributes, allowing the input data to fit more the specific algorithms and the output results to be further optimized, so that the fault identification accuracy can be improved.

2.1. Interpretative Fault Preprocessing

Dip-steering filter (DSF) is one of the preprocessing techniques that have been successfully applied, and its core is anisotropic diffusion smoothing [29]. Its essence is to smooth the part of seismic data parallel to the seismic event and without surpassing the position (e.g., fault) at which seismic reflection terminates. By de-noising filtering along the strike and trend of seismic reflection surfaces, DSF can improve the continuity of seismic events and also enhance the lateral resolution of the termination position of seismic events. In this way, the fault information is not smoothed and small faults are highlighted. By comparing the data before and after filtering (Figure 2(a)), it is found that DSF can effectively realize profile de-noising and preserve the detailed information of fault and geobody boundary, thereby enhancing the continuity of the seismic event. However, it fails to visibly image the fault point or plane (Figure 2(b)).

In this study, fault focus imaging (FFI) was conducted on the basis of DSF. FFI is a frequency-domain preprocessing technology [23]. To be specific, the seismic data are analyzed in time-frequency after domain conversion; then, taking the frequency of the imaging fault as the focusing frequency, the high- and low-frequency range in the frequency range of seismic data is defined; within this range, frequency division processing is made to highlight the frequency bandwidth energy of fault and suppress the energy in other frequency bandwidth; finally, the fault is highlighted. A comparison of the seismic profiles before and after processing indicates that FFI effectively solves the problem of unsharp and unclear fault points. Moreover, FFI effectively highlights the fault plane in the chaotic fault zone, and also apparently enhances the information on small-scale faults and micro-fractures submerged in signals (Figure 2(c)).

2.2. Fault-sensitive Attribute Optimization

The seismic attribute has a multivariate, multidimensional, and nonlinear complex relationship with the object to be predicted. When geological conditions are simple and seismic signals are high-quality, appropriate seismic attributes can be easily chosen for the object; otherwise, when the geological structure is complex, seismic attributes are difficult to choose, and the extracted seismic attributes are not completely independent of each other—some attributes contain similar information, so it is necessary to select the appropriate attributes from them, that is, seismic attribute optimization. In this paper, seismic attributes are used for fault detection. A lot of experiments find that the eigenvalue coherence attribute is sensitive to large-scale faults, the curvature attribute is sensitive to medium-scale faults, and automatic fault extraction (AFE) and likelihood are sensitive to small-scale faults [30]. Moreover, principal component analysis and color blending are used to fuse the sensitive attributes of faults at different scales, highlighting the distribution characteristics of faults.

Figure 3 shows the fault distribution characteristics after data processing of different stages. Compared with other figures, Figure 3(c) provides a clearer description of faults of different scales (Figure 3(a)) and significantly reduces noise (Figure 3(b)). From Figure 3(c), it can be seen that the planar distribution of large-scale faults (yellow arrow) and their fracture edges are clearer, while the planar continuity of medium-scale faults (pink arrow) is better, and the number of small-scale faults (white arrow) is significantly increased. Obviously, the proposed fine fault prediction technology can effectively solve the problem of fuzzy response characteristics of seismic faults underlying igneous rocks. According to the fault multi-attribute fusion map, the plane and profile fine fault interpretation was conducted for the Ordovician target layers, and a total of sixty-five faults were identified (Figure 3(d)). It should be noted that the original seismic data volume is fault-augmented for improving the prediction accuracy, which highlights the true faults and also enhances the seismic responses of pseudofaults. Therefore, it is necessary to identify and eliminate seismic pseudofaults.

3.1. Existence Analysis of Seismic Pseudofaults Underlying Igneous Rocks

According to the seismic and geological characteristics of Permian igneous rocks in Shuntuoguole uplift area (Figure 4(a)), the forward modeling model was designed and the forward modeling research was carried out. The middle of the forward model is high-velocity dacite, and the two sides are low-velocity pyroclastic rocks, so there are lithology and velocity abrupt changes in the horizontal direction of the model (Figure 4(b)). In the forward modeling, a velocity similar to or lower and smoother than that of the pyroclastic rock was used as the migration velocity of the dacite layer, and a fault illusion appears at the point of velocity mutation, which is very similar to the fault on the actual seismic profile. This confirms that the migration velocity error of high-velocity igneous rock is the cause of pseudofault underlying the igneous rocks (Figure 4(c)). In addition, the mutation point of lithology and velocity, that is, the mutation point of high-velocity igneous rock thickness in the formation, is the location of the seismic pseudofaults.

3.2. Quantitative Identification and Elimination of Seismic Fake Faults Underlying Igneous Rocks

It is well known that improving velocity modeling accuracy of igneous rock and reducing drilling risk are essentially contradictory in the early stage of oil–gas exploration. For the former, prior information (e.g., logging data) must be obtained first through drilling to constrain and correct the background velocity field. For the latter, high-precision velocity modeling is a must to obtain accurate migration imaging data for fine interpretation of faults and structures. It is inevitable that the two tasks are done successively but not simultaneously, which is also a bottleneck in the application of existing igneous rock velocity modeling technologies in no-well areas.

Based on the lithofacies velocity analysis of igneous intervals drilled in the main strike-slip faults in the Shuntuoguole uplift, it is considered that although the true value of the migration error of the high-velocity igneous rock without well constraints cannot be accurately solved, the range and maximum value of the migration velocity error can be estimated by “extreme hypothesis method.” Extreme hypothesis is a technique based on the mathematical concept of maximization; that is, a complex problem is analyzed under a hypothetical extreme state to make it simple, extreme, and concise [31]. Igneous rocks have complex genesis and diverse lithofacies. Particularly, the igneous rock of effusive facies is formed by magma after flowing along the surface and gradually condensing under the driving of eruption and its gravity. Compared with the explosive facies and volcanic sedimentary rocks, the igneous rock of effusive facies features higher interval velocity and simpler lithology [32]. Therefore, the interval velocity of igneous rocks of effusive facies is relatively stable at different locations in the study area (Table 1). By using the background velocity of high-velocity igneous rocks derived from original seismic data and the regional interval velocity of effusive facies high-velocity igneous rocks, the range of migration velocity error is estimated, and then the maximum error value is determined. With the maximum error value to replace the background velocity of high-velocity igneous rocks, forward modeling and depth migration are performed to determine the maximum psedofault throw value that can be generated in the underlying strata. When the fault throw of the target fault derived from original seismic data is greater than the simulated maximum pseudofault throw, the target fault can be identified quantitatively as a true fault based on an extreme hypothesis.

This study deals with the Shun 8 North 3D area in the Shuntuoguole uplift. First, based on the logging data of eighteen main strike-slip faults drilled, the regional interval velocity of effusive dacite was determined to be about 4950 m/s. Second, combined with the background velocity (about 4210–5093 m/s) of dacite in the Shun 8 North 3D area, the range of migration velocity error is estimated to be about 0–15%. Third, two seismic profiles were selected randomly to establish the forward model of igneous rock based on structural constraints (Figure 5(a) and (b)), and the maximum error value (15%) was used to replace the velocity of high-velocity igneous rock in the velocity model for purpose of migration imaging (Figure 6(a) and (b)). The maximum pseudofault throw of the target layer is caused by dacites with different thicknesses under the maximum velocity error (15%) (Figure 6(c) and (d)). Then, the curve fitting was made on the maximum pseudofault throw, and the quantitative relationship between the thickness of dacite and the pseudofault throw of the target layer under the maximum velocity error (15%) as follows: y = 0.0002x2 + 0.1851x − 3.750, where x is the thickness of the Permian dacite and y is the maximum pseudofault throw of the Ordovician target layer (Figure 7). This quantitative relationship can be used to calculate the seismic pseudofault throw value of the Ordovician target layer caused by the maximum migration velocity error of dacite with different thicknesses. Finally, the Permian dacite thickness in the study area was calculated by using the regional interval velocity of dacite and the time window of the top and bottom of dacite on the seismic profile, and the maximum pseudofault throw of the target layer caused by the migration velocity error was obtained by introducing the quantitative relationship. When the fault throw of the target fault on the seismic profile is greater than the maximum pseudofault throw of the target fault, it can be inferred that the fault throw on the seismic profile is not completely caused by the dacite, and the fault is judged to be true based on the extreme hypothesis.

4.1. Failure Analysis of Well SHBP2H

SHBP2H is an exploration well deployed by Sinopec in the secondary strike-slip fault in the Shun 8 North 3D area. It was designed to drill the NNE-trending secondary strike-slip fault F2 on the profile inline 1638 (Figure 3(c)) but actually did not encounter the fault, suggesting a drilling failure. In this paper, the “extreme hypothesis method” is used to quantitatively analyze whether the fault on the profile along Inline1638 is true or not. As shown in Figure 8, the thickness difference of dacite between the two sides of F1 is 165 m, and the boundary thickness of dacite at F2 is 116 m. According to the formula: y = 0.0002x2 + 0.1851x − 3.7506, the maximum pseudofault throw of the target layer is calculated to be 32.23 and 20.41 m under the velocity error of 15% for 165 and 116 m, respectively, while the vertical fault throws of F1 and F2 derived from original seismic data are 40.7 and 11.4 m. According to the principle of “extreme hypothesis,” the fault throw of F1 on the original seismic profile is greater than the simulated maximum pseudofault throw, which indicates that the fault throws on the seismic profile is not completely caused by dacite, so it is a true fault. On the contrary, for F2, the simulated maximum pseudofault throw is greater than the maximum pseudofault throw derived from original seismic data, so F2 is judged as a seismic pseudofault (Table 2), which is consistent with the actual drilling conclusion (Figure 9(a)).

4.2. Fine Analysis of True Faults

The fault of well SHB7 has been identified as true fault by using “extreme hypothesis method.” Figure 9(a) and (b) shows the imaging logging data of well SHBP2H and well SHB7 in the target interval. There are five sections of coring data of SHBP2H (Figure 9(a)), all of which show that the core is tight and the fractures are not developed, which proves that the target seismic fault does not exist. The imaging logging data of SHB7 clearly show that semi-filled karst caves are locally developed in the target interval (Figure 9(b)), which is consistent with the development characteristics of karst reservoirs under the control of strike-slip faults. To sum up, the “extreme hypothesis method” can accurately judge the authenticity of the seismic faults underlying igneous rocks.

Finally, a total of sixty-five faults (Figure 3(c)) in the original seismic data volume of Ordovician target layers in the Shun 8 North 3D area were selected for quantitative identification under the extreme hypothesis. The results show that eleven faults are pseudofaults. After eliminating these pseudofaults, the plane distribution of 54 true faults in the Ordovician target layer was obtained (Figure 10).

The “extreme hypothesis method” is to maximize the distortion of the underlying target layer of igneous rocks by maximizing the complex calibration problem of migration velocity in velocity modeling. Seismic pseudofaults can be identified quantitatively without building an accurate velocity-depth model. According to the true faults identified by “extreme hypothesis,” the deployment of wells can not only ensure the success of drilling but also provide necessary constraint information such as logging velocity for subsequent igneous rock velocity modeling. However, the “extreme hypothesis” is still not accurate enough because it quantitatively judges the true or fake faults by maximizing the migration velocity of igneous rocks. Therefore, it is not possible to qualitatively identify all faults, which is also the limitation of this technology at this stage.

To solve the problem above, this paper proposes an igneous rock velocity modeling technology based on “extreme hypothesis” and “dynamic correction.” Due to the limitations of velocity modeling methods in procedure and accuracy, the quantity of constraint information (e.g., logging data), and the processing schedule, error exists inevitably between the simulated velocity model, whether it is obtained by whichever method or method combination, and the real velocity model. Therefore, in igneous rock velocity modeling, the accuracy of the model should be improved gradually by iterative correction to enable it to infinitely approach the real velocity model. First, new wells are deployed at the true faults identified quantitatively by the extreme hypothesis method. Second, the initial velocity model is corrected for the first time by using the logging data (e.g., AC) and horizon information, and PSDM imaging is performed. Third, with the faults identified by extreme hypothesis as a constraint, which are kept consistent with faults from migration data if possible, the velocity model is validated. Since the lithology of igneous rock strata is discontinuous, logging and other data can only reflect the velocity information in local areas, but not in a wide range. Therefore, for pursuing a more accurate velocity model and migration imaging data, the corrected velocity model and migration imaging data are used as the initial velocity model and fault interpretation data for the next time of correction. Finally, the velocity model is enabled to approach the real velocity model by multiple iterative corrections, and the accurate analysis of true faults underlying igneous rocks is completed by repetitive dynamic correction of migration data. The proposed igneous rock velocity modeling technology based on extreme hypothesis and dynamic correction may promote the igneous rock velocity modeling and the interpretation of faults underlying igneous rocks to transit from static to dynamic. It is of scientific significance and application value.

  1. The strong shielding and absorption of igneous rocks on seismic waves can easily lead to poor quality of underlying seismic waveform and fuzzy seismic response characteristics of faults. In no-well igneous rock areas, the seismic identification accuracy of faults can greatly be improved by fault-augmented interpretive preprocessing and fault-augmented sensitive attribute optimization on the original seismic data volume, which is the data basis of quantitative elimination of pseudofaults. It is considered that DSF and FFI can enhance the seismic responses of faults of all scales. The eigenvalue coherence attribute is sensitive to large-scale faults, the curvature attribute is sensitive to medium-scale faults, and AFE and likelihood are sensitive to small-scale faults.

  2. The migration velocity error of high-velocity igneous rocks in PSDM imaging is the cause of seismic pseudofaults underlying igneous rocks, and the logging velocity is the necessary condition to correct the migration velocity error. Therefore, the real error value of migration velocity cannot be accurately solved without well constraints, which directly leads to the inapplicability of existing velocity modeling technologies in no-well areas to accurately confirm the structures underlying igneous rocks. By determining the maximum error of migration velocity from the regional interval velocity and seismic background velocity of high-velocity igneous rocks, the accurate solution of migration velocity can be maximized to simulate the maximum distortion of the target layer underlying igneous rocks, thereby providing a scientific basis for quantitative identification and elimination of seismic pseudofaults underlying igneous rocks. This technology has the technical advantage of quantitatively identifying the true and fake faults without constructing the velocity-depth model and effectively breaks through the technical bottleneck of existing velocity modeling technology, which relies too much on logging velocity.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

There is no conflict of interest regarding the publication of this article.

This work was Sponsored by CNPC Innovation Found (2021DQ02-0104)

Exclusive Licensee GeoScienceWorld. Distributed under a Creative Commons Attribution License (CC BY 4.0).