Prediction of Thin Shoal Reservoirs Under Reef Controlled by Isochronous Stratigraphic Framework

Abstract

Despite the great success in the global exploration and development of reef reservoirs, research on bioclastic shoals under reefs is still in its infancy. Bioclastic shoal reservoirs are very thin, with multiple vertical levels and fast lateral changes, which makes accurate prediction of the reservoir’s location much tougher. To further implement the reservoir distribution, under the guidance of sequence stratigraphy, the prediction of thin shoals under the control of an isochronous stratigraphic framework was established. Using the combination of spectrum shaping and F-X domain noise suppression techniques and utilizing the signal-to-noise ratio spectrum set as the reference, logging curve as supervision, and well seismic calibration and isochronal amplitude slicing as quality control, the seismic frequency band was extended, and the seismic data resolution and signal-to-noise ratio were improved. After frequency extension, the global optimal seismic automatic interpretation technique was used to construct an isochronal stratigraphic framework model. Through waveform facies-controlled inversion and waveform facies-controlled simulation techniques, the elastic properties of the shoal reservoir were obtained, from which the planar distribution of the reservoir was accurately predicted. The above methods were applied to the prediction of the bioclastic shoal reservoir in the lower submember of the Changxing formation in the Yuanba gas field (China). The plane distribution of bioclastic shoal in the first and second levels was identified, which provides a guideline for the prediction of thin shoal reservoirs.

1. Introduction

To date, notable success has been achieved in the exploration and development of reef reservoirs on an international level. Many scholars have carried out a lot of research work on the oil and gas geologic identification of reefs, reservoir prediction, reservoir formation, and so on [1,2,3,4,5,6]. The growth of reefs is affected by both their paleogeomorphology and changes in the sea level. Controlled by sedimentary facies and the processes of dolomitization and dissolution, the reef reservoirs are mainly dolostone, superimposed in many levels, and developed in rows and belts [7,8,9,10,11,12]. Due to their unique appearance and structure, reefs exhibit apparent seismic characteristics. As a result, a series of techniques, such as forward modeling analysis, multi-scale edge detection and seismic attribute analysis, three-dimensional carving, and seismic inversion, have proven useful for depicting the reefs and their internal structures in a fine manner as well as for predicting the plane distribution of the reservoirs [13,14,15,16,17,18,19,20]. The most common evolution pattern of carbonate sediments is from the carbonate ramp type to the ramp with steeper end type and then to the carbonate platform type. With the rapid accretion and progradation process of carbonate rocks, reefs often inherit and then develop on the foundation of a bioclastic shoal. In-depth research has been conducted on reefs; however, insufficient attention has been paid to the bioclastic shoals that are located below the reefs. The bioclastic shoal gas reservoir of the upper-Permian Changxing formation in the Yuanba area was developed under the paleogeographic platform-margin background of the Yuanba area, which originated from the aulacogen groups formed by Emei ground fissure movement. In the context of Changxing’s transgression, reef shoals were developed near the platform margin in the secondary regression cycle [21,22,23,24]. The reef shoal development in the Yuanba gas field has the characteristics of “early shoal and late reef” and “front reef and back shoal”, the reservoir formation of which has been portrayed by the following characteristics: “near-source enrichment”, “fault-fracture transformation”, “lithologic reservoir control”, and “structure control enrichment” [25,26]. The reef reservoir of the Changxing formation is thick, with obvious seismic response characteristics, which makes it easy to predict using seismic methods [27]. On the contrary, the bioclastic shoal reservoir under the reef has the characteristics of a small scale, multi-level vertical development, dispersed plane distribution, thin reservoir, strong heterogeneity, non-typical seismic responses, and limited seismic resolution; therefore, the bioclastic shoal reservoir offers a strong multi-solution in seismic recognition. Moreover, it is disturbed by carbonaceous mudstone at the bottom of the Changxing formation, which makes reservoir prediction considerably difficult and restricts the efficient exploration and development of the bioclastic shoal. Guided by the sequence stratigraphy, the prediction of thin shoal reservoirs under the control of the isochronous stratigraphic framework was established in this study, aiming at implementing the plane distribution of bioclastic shoals under the reefs of the Changxing formation in the Yuanba gas field. The results provide a reference for thin shoal reservoir prediction under a similar geological background.

2. Geological Background

The Yuanba gas field is located at the confluence of the Jiulongshan anticline, Chixi sag, and Cangxi-Bazhong low-relief tectonic belt, which belongs to the slope zone from the central Sichuan paleo uplift to the northern Sichuan depression (Figure 1), with gentle strata and less fault development. In fact, the area has experienced relatively little tectonic activity. The burial depth of the main target layer Changxing formation was found to be more than 6000 m [28]. Carbonate ramp deposits were developed early in Changxing’s formation and gradually transformed into carbonate platform deposits in the middle and late levels, while multi-level bioclastic shoals and reef reservoirs were developed longitudinally. In previous studies, the Changxing formation in the Yuanba area was divided into two third-order sequences, namely the lower sequence (corresponding to the lower submember of the Changxing formation) and the upper sequence (corresponding to the upper submember of the Changxing formation). In this study, the development and distribution characteristics of shoals from the lower sequence (corresponding to the lower submember of the Changxing formation) were selected as the target. The shoals were divided into two levels from bottom to top, with the first one corresponding to the early level of the lower-sequence transgression system tract and the second corresponding to the middle and late level of the lower-sequence high-water-level system tract.

Analysis of typical wells shows that about 5 m lagoonal bog facies carbonaceous mudstone was developed at the lower bottom of the Changxing formation; a high natural gamma-ray marsh facies carbonaceous mudstone was found, on top of which the first-level bioclastic shoal reservoir was developed. Its lithology composition includes powder-fine grained dolostone and dolostone-bearing clastic limestone at the bottom, covered by argillaceous limestone, and topped by siliceous limestone, reflecting the rapid rise of the base level. The lithology compositions, from old to new, are siliceous limestone, bioclastic limestone, microcrystalline limestone, bioclastic limestone, and dolostone, indicating that the sea level has been rising slowly. The sedimentary cycle from the upper bioclastic limestone to dolostone indicates that the base level is stable at the highest position. Additionally, the second-level bioclastic shoal is developed, and the lithology includes powder-fine dolostone.

The two levels of bioclastic shoal reservoir in the lower submember of the Changxing formation have a reservoir thickness of 10–20 m, with an average thickness of 13.5 m and a porosity of 5.3–8.2%, with a statistical average core porosity of 5.61%. Compared with the reefs in the upper submember of the Changxing formation, the bioclastic shoal in the lower submember has a reservoir average thickness of 75.7 m, and the average porosity is 5.8%, indicating that the lower part of the bioclastic shoal is buried more deeply, and the reservoir is thinner. In addition, it is also affected by the bottom carbonaceous mudstone with high gamma-ray logging value and central argillaceous limestone, which makes it more difficult to predict the thin bioclastic shoal reservoirs under the reef.

3. Materials and Methods

This study mainly uses the three-dimensional seismic data and drilling data of Yuanba. The 3D seismic data were collected from the 24 lines distributed in Yuanba area, which actually produce 63,427 physical points, covering an area of about 1330 km². The 12L48S240T orthogonal observation system was used for seismic data acquisition. The number of receiving channels of the observation system is 2880 (240 channels × 12 lines), and the dimension of the surface element is 25 m × 25 m. The coverage time is 72, the track distance is 50 m, the receiving line distance is 400 m, the shot point distance is 50 m, and the shot line distance is 500 m. The quality of seismic data is generally good. The characteristics of the seismic reflection wave group are clear, the level is complete, the reflection wave in-phase axis can be continuously compared and tracked, and it has a high signal-to-noise ratio. A total of 30 exploration wells have been drilled through the Changxing formation in the study area, with a depth range of 6800–7200 m. Specifically, of these 30 wells, 22 are located in the platform facies area, and the remaining 8 are located in the shelf area. The spacing between each well is 5–10 km.

According to the characteristics, prediction difficulties, and geological requirements of the bioclastic shoal reservoirs in Changxing formation, the process of investigation in this study started with the processing of seismic data; then, frequency extension of seismic data was carried out, and spectral shaping and F-X domain noise suppression techniques were used to increase the resolution of the data in order to effectively improve their quality. On this basis, combined with geological knowledge and single-well sequence analysis, the isochronal stratigraphic framework was constructed by utilizing the global optimal seismic automatic interpretation technique, and the vertical relationship of the multi-level secondary bioclastic shoal was established. Finally, waveform facies-controlled inversion and simulation were used to predict a high-resolution reservoir and to depict the plane distribution of the multi-level secondary bioclastic shoal. This paper mainly introduces the extension frequency processing of the post-stack seismic number and the establishment of a global optimal time stratigraphic model.

3.1. Extension Frequency Processing of Post-Stack Seismic Data

Many kinds of frequency extension techniques are applied for post-stack seismic data processing, such as inverse Q filtering, spectral shaping, compression sensing, and reflection coefficient inversion [29,30,31]. Spectral shaping and F-X filtering are adopted to improve the resolution of the seismic data. The mixed-phase deconvolution technique is used to expand the frequency band and to make the section plane zero-phased. The key to such technique is wavelet extraction and output definition. The complex cepstrum technique is used in wavelet extraction to make the wavelet amplitude spectrum more objective. The signal-to-noise ratio spectrum is referenced to define the output and reasonably mining the potential of high frequencies. After tests, the expected output wavelet finally selected in the present study was a Yu-wavelet.

3.2. Establishment of Global Optimum Stratigraphic Model

This method was proposed by Pauget et al.; the core idea of this method is summarized in the following procedures [32]: (1) Seismic sampling points are automatically connected, based on the global optimization of minimum cost function; (2) adjacent seismic trace grid nodes are connected by calculating the correlation between them; (3) the nodes with the highest correlation are connected by line segments; (4) and a relative isochronal model is formed with consistent characteristics and clear stratigraphic meaning [32,33,34]. The calculation process is described in the subsequent sentences [32]. First, the seismic data are sampled, and regular data grids are constructed in the x-y plane. Since all the sample points of the seismic traces are required in the implementation process, they may need to be recombined to reduce the time spent in calculation, as shown in Figure 2a. Next, connections between these grid points are determined by calculating the relevant images of each pair of adjacent tracks. The correlation images of two adjacent seismic tracks X1 and X2 are a series of arrays, which comprise the correlation values between each pair of grid nodes on each seismic trace in the X1 (t)-X2 (t) plane. Each grid node on the image corresponds to the connection between the points on X1 and those on X2. Grid nodes with high correlation values have a high possibility of connection. When series of such nodes are connected to each other to form a correlation cluster line segment, it can be used to characterize the connection of the seismic phase axis, as shown in Figure 2b. The largest correlation cluster is obtained by the global optimization algorithm, which provides a series of connections and can be used to calculate the global position of each point on the sampling grid. The initial geological model is then formed by connecting the nodes that meet a certain correlation value, as shown in Figure 2c. Finally, the initial model is constrained by the minimum cost function, which makes the connection of the grid nodes more reasonable, as shown in Figure 2d. The minimum cost function is related to seismic similarity, geological consistency, and distance.

4. Results

4.1. Frequency Extension of Seismic Data

After frequency extension processing, the random noise is amplified, while the energy of the high-frequency signal is increased, so further attenuation of random noise is needed, and thus, the F-X domain noise suppression technique was used to eliminate the random noise in this study. Before frequency extension, the main frequency of the target layer of the seismic data was about 28 Hz, and the band width ranges between 10 and 60 Hz. After frequency extension of the seismic data, the main frequency of the target layer was about 38 Hz, and the bandwidth was 8–80 Hz (Figure 3). Through the comparison of seismic profiles before and after frequency extension, as shown in Figure 3, composite reflections took place among several sets of wave groups in the Changxing formation before frequency extension; moreover, the interior continuity of the in-phase axis was poor. After frequency extension, the composite wave groups were distinguishable, and the continuity of the in-phase axis was enhanced. The shape and structure of the reef shoal was consistent (marked by the red and yellow arrows). On the time slice, the relationship between the amplitude before and after frequency extension can also be seen as consistent. Moreover, the details of the amplitude change are clearer and enriched after the frequency extension, as shown in Figure 4.

As shown in Figure 5, the rationality of well seismic calibration synthetic records was chosen as the quality control standard. Through drilling, it was proven that the two levels of bioclastic shoals are developed in the lower submember of the Changxing formation of the Y2 well. The reservoir of the second-level shoal (marked with a red dotted frame) is about 12 m thick, which is characterized by a large acoustic time difference and low impedance and is related to seismic troughs with a relatively strong amplitude in the synthetic record. The correspondence between seismic data after frequency extension and the synthetic record is, therefore, more reasonable.

4.2. Globally Optimized Isochronous Stratigraphic Model

Two levels of bioclastic shoal reservoirs are developed in the lower submember of the Changxing formation in the Yuanba area, as shown in Figure 5. The first-level shoal is located near the bottom of the lower submember, and the second-level shoal is located near the top of the upper submember; between the two levels, there is an interlayer of 30–40 m. Such a structure makes it difficult to interpret different levels of seismic data before frequency extension. The global optimal seismic interpretation technique was used on the frequency extension data to carry out different levels of horizon interpretation, and the globally optimized isochronous stratigraphic model was constructed.

The isochronous stratigraphic model was the basis for predicting the multi-level thin bioclastic shoal. According to the sedimentary background of the target area in this study, combined with lithology and logging characteristics, a set of carbonaceous mudstones around 5 m thick are developed at the bottom of the Changxing formation, featuring high natural gamma, high sound waves, and low resistivity on the logging curve. With distinct electrical characteristics from either the upper limestone or bioclastic limestone, the bottom of such a set of carbonaceous mudstones serves as the bottom boundary of the Changxing formation. The carbonate rocks at the top of the Changxing formation are generally characterized by low argillaceous content, low natural gamma value, and low potassium curve value, while the overlying Feixianguan formation is characterized by a higher natural gamma value and higher potassium curve value. Therefore, the top of the Changxing formation is defined at the place where the natural gamma and potassium curve values display an obvious increase. Limited by seismic data in this area, the top and bottom interface of the Changxing formation is an unstable, continuous standard reflection wave, and the thin shoal corresponds to the seismic trough, which causes difficulties in seismic tracking. For this reason, the global optimal isochronal seismic interpretation technique was adopted to extract the sequence stratigraphic framework from the seismic data, highlight the characteristics of the thin shoal, and discover the hidden sequence interface in the complex stratigraphic structure [35]. On the conventional seismic profile, combined with well seismic calibration, the seismic in-phase axes corresponding to the top part, the bottom of the upper submember, and the bottom boundary of the Changxing formation are continuous and can be traced regionally. However, other layers, such as the multi-level subshoal of the lower submember and the multi-level reef of the upper submember of the Changxing formation, are affected by the changes of lithology and sedimentary facies belts. With rapid changes of seismic in-phase axes in a chaotic and unstable manner, great difficulties are encountered with manual interpretation, as shown in Figure 6a. Using the global optimization algorithm, a series of stratigraphic groups with isochronous significance were obtained directly from the seismic data, which form the horizontal groups. In addition to accurately identifying the continuous strata on the conventional profile, this model can also identify special geological bodies such as the reef mound structure and insider characteristics of the Changxing formation, as shown in Figure 6b.

With the help of the globally optimized stratigraphic model, combined with the analysis results of the single-well sequence and system tract, the interpretation horizons reflecting the sequence interface were extracted from the horizon stack, and the isochronous sequence stratigraphic model was constructed. The comprehensive seismic study shows that the Changxing formation can be divided into two third-order sequences. Its thin shoal is mainly developed in the first third-order sequence, with two longitudinal levels: The first level belongs to the transgressive system tract, and the second level belongs to the highstand system tract, as shown in Figure 7a. Combined with the results of the drilling sequence division, the sequence boundaries with geological meaning were extracted from the horizon stack to form the sequence body, as shown in Figure 7b,c.

4.3. Inversion of Thin Shoal Reservoirs

To predict thin shoal reservoirs, the post-stack inversion was performed mainly based on geostatistical random simulation to establish a wide-band model, whose low-frequency components are from logging curve interpolation and whose high-frequency components are from the variation function established between the wells. When the distribution of the well number is uneven and insufficient, it is difficult to characterize the rapid changes of sedimentary facies with the inter-well interpolation model. Consequently, the inversion results have high randomness. Spatial variations of the reservoir can be characterized from waveform inversion, utilizing the changes of seismic waveform. Random simulation was realized by combining seismic and logging information. The inversion result was stable and conforms to the law of sedimentary variation. Based on the Markov Monte Carlo random simulation algorithm [36], the correlation of wells was sorted according to seismic waveform and spatial distance. The impedance information of the wells with high-correlation wells was used as prior information to establish the initial model. Then, the initial model was matched with the seismic frequency band impedance filter to obtain the likelihood function. On this basis, a posteriori probability distribution was acquired using Bayesian theory together with the prior information and likelihood function, which was set as the objective function. The parameters of the model were constantly modified to realize the unbiased optimal estimation of the objective function [37]. When applying such a technique, the key parameters are the selection of effective samples and the best cutoff frequency. The number of effective samples mainly characterizes the influence of the spatial variation of seismic waveforms on the reservoir, reflecting the lateral variations of the reservoir and the complexity of the vertical structure. When the sedimentation of the target area is stable, the seismic response characteristics of the reservoir have strong regularity, and the number of effective samples is large; when the lateral changes of the reservoir are fast, and the heterogeneity is strong, the seismic response regularity of the reservoir is poor, and the number of effective samples is small. The optimal cutoff frequency in inversion is mainly used to adjust the deterministic frequency band and random frequency band of inversion results, among which the deterministic frequency band is controlled by seismic simulation, and the random frequency band comes from the Markov Monte Carlo random simulation. Based on the analysis of geological demand and seismic identifiability, the number of effective samples in the study area was five, and the best cutoff frequency was determined to be 180 Hz.

Therefore, to accurately predict the multi-level thin shoals and eliminate interference from the underlying carbonaceous mudstone, the waveform facies-controlled inversion technique was used to improve the resolution of the thin reservoir, and the facies-controlled characteristic curve was used to simulate and eliminate the influence of carbonaceous mudstone.

As the constraint condition of inversion, the horizon information in the isochronous stratigraphic framework was helpful for reducing the uncertainty of inversion results and improving the accuracy of inversion.

The high-resolution inversions of typical wells are shown in Figure 8, among which Figure 8a corresponds to the waveform facies-controlled inversion, and Figure 8b corresponds to the constrained sparse pulse inversion. Compared with the constrained sparse pulse inversion, the facies-controlled waveform indicates that the inversion resolution is significantly improved, and the inversion results are consistent with the drilling results; additionally, the lateral variations of the two kinds of inversion are consistent. Therefore, the waveform facies-controlled inversion is not only able to improve the longitudinal resolution but also maintains the lateral variation, which is in accordance with geological evidence.

5. Discussion

5.1. Advantages of Isochronous Stratigraphic Model

In thin bioclastic shoal reservoir prediction, the construction of the isochronous formation model is very important. Due to the limitation of the resolution and signal-to-noise ratio of seismic data before extension frequency, it is difficult to distinguish the horizon of different periods effectively. The application of global optimal seismic interpretation technology successfully solves this problem, which makes it possible to construct a global optimal isostratigraphic model and provides a reliable horizon constraint for the subsequent thin bioclastic shoal inversion and reservoir prediction.

The sequence body has the following advantages: (1) It is consistent with the logging data: The well log data were fully considered in the process of model construction to ensure consistency with log data on the sequence scale and improve the geological reliability of the model. (2) It is formed three-dimensionally: Through the global optimization algorithm, the model realizes the transformation from one-dimensional sequence to three-dimensional sequence body and reveals the spatial distribution characteristics of strata more comprehensively. (3) It retains seismic reflection characteristics: The sequence body makes the interpretation results that are more consistent with the actual geological conditions and improves the accuracy of seismic interpretation. (4) It improves the accuracy of reservoir prediction: It provides a fine isochronous stratigraphic framework for inversion, which makes the inversion results more consistent with the objective law of actual formation evolution and helps to improve the accuracy of reservoir prediction. (5) It improves seismic interpretation efficiency: The application of the global optimal stratigraphic model also significantly improves the efficiency of seismic interpretation, reduces the time and cost of manual interpretation, and provides strong support for the rapid and accurate prediction of thin bioclastic shoal reservoirs.

5.2. Distribution of Thin Shoal Reservoirs

Well-logging interpretation indicates that two levels of bioclastic shoal reservoirs are developed in the lower submember of the Changxing formation. Type III reservoirs comprise the highest abundance, followed by type II reservoirs. There is a set of non-reservoir carbonaceous mudstone that is about 5 m thin near the bottom of the first-level shoal. To depict the distribution of the two levels of shoal, lithologic and petrophysical analysis was carried out to determine the sensitive parameters that can reflect the reservoir characteristics, as shown in Figure 9. The cross-plot of gamma and wave impedance demonstrates that both the reservoir and carbonaceous mudstone exhibited medium and low impedance, in which type I and II reservoirs are superimposed with carbonaceous mudstone, and the differentiation between type III reservoir and carbonaceous mudstone was found to be high, as shown in Figure 9a. The gamma curve has a high degree of differentiation on carbonaceous mudstone. When the gamma value is greater than 50, this indicates carbonaceous mudstone.

Drilling revealed that the first level of thin shoal is developed in well Y3, the upper and lower levels of thin shoal are developed in well Y2, and the first-level thin shoal reservoir is developed in well Y4 (marked by the black arrow). A set of thin carbonaceous mudstone with continuous transverse distribution and low wave impedance is developed at the bottom of the first level, which is difficult to predict from the inversion of wave impedance. The facies-controlled characteristic curve was used to simulate the high gamma value characteristics of carbonaceous mudstone, as shown in Figure 8c, in which the curve beside the well is a gamma curve. Simulation of the characteristic curve shows that the high gamma value of the target layer corresponds to the carbonaceous mudstone (marked by the white arrow). The lateral variation is stable and capable of being tracked continuously. Based on the results of the isochronous stratigraphic framework model, the top and bottom layers of both the upper and lower system tracts were extracted in addition to the wave impedance attributes of the interlayer. The reservoir plane distribution of the shoal bodies in different levels was then depicted. The shoal of the first level of the Changxing formation is mainly distributed in the northwest–southeast direction along the platform margin zone, as shown in Figure 10a. Combined with geological knowledge, it is speculated that such a distribution is related to the slope break zone formed by episodic tension in the ancient Kaijiang-Liangping rift in the early level of the Changxing formation. The shoal distribution of the second level of the Changxing formation is scattered, and its range is wider than that of the first-level shoal reservoir, as shown in Figure 10b. Such a distribution is related to the retrogression and paleogeomorphological differences after large area transgression in the early and middle levels of the Changxing formation.

6. Conclusions

• (1). Spectral shaping and F-X domain noise suppression technology were utilized, with the signal-to-noise ratio spectrum as the reference, the well logging curve as supervision, and the well seismic calibration and isochronal amplitude slicing as quality control. The seismic frequency band was extended, and the seismic data resolution and signal-to-noise ratio were improved;

• (2). Based on the global optimal isochronal seismic interpretation technique, the isochronous stratigraphic model of the system tract scale was constructed considering the correlation between nodes, which improved the interpretation accuracy and efficiency;

• (3). Waveform facies-controlled inversion and facies-controlled characteristic curve simulation was proven to be able to effectively identify multi-level thin shoal reservoirs and carbonaceous mudstones, which provides references for studies carried out in areas with the same geological background;

• (4). The shoal of the first level of the Changxing formation in the Yuanba area is mainly distributed in the northwest–southeast direction along the platform margin belt, while the shoal of the second level of the Changxing formation is more scattered, with a wider range than that of the first level.

Author Contributions

S.X., conceptualization, methodology, and writing—original draft; X.H., conceptualization, funding acquisition, and writing—original draft and editing; Z.S., formal analysis, literature investigation, and writing—review and editing; C.Z., literature investigation and writing—review and editing; J.M., investigation and writing—review and editing; B.W., data curation and writing—review and editing. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data are available from authors for a limited time.

Conflicts of Interest

Author Shoucheng Xu was employed by the company Southwest Oil and Gas Company, SINOPEC. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures

Figure 1. Structural map of the wider study area of the Yuanba gas field (left, adapted from [26], with permission) and stratigraphic column map of the study area (right).

Figure 2. Basic procedures to construct the globally optimized isochronous stratigraphic model (adapted from [34], with permission from The Leading Edge, 2011). (a) Gridded points are linked via a correlation image of neighboring traces. (b) Every point corresponds to a link between a point on X1 and a point on X2 (c). A set of high probability links (A correlation comb) is used to calculate the global position of each point on the sampled grid. (d). The model is constrained by potential seismic signals to ensure consistency.

Figure 3. Seismic profile of cross Y1–Y2 wells before (upper) and after (lower) frequency extension. (The red arrow shows the change of seismic data before and after frequency extension).

Figure 4. Amplitude attribute iso-T0 slice (3084 ms) before (left) and after (right) frequency extension. (Red represents the peaks of the seismic waveform, black represents the troughs of the seismic waveform).

Figure 5. Y2 well seismic calibration before and after frequency extension (The red dotted line frame reflects the location of reservoir development in Changxing Formation).

Figure 6. Typical seismic profile and its globally optimized isochronal interpretation model of the studied area. (a) Typical seismic section. (b) Globally optimized isochronous stratigraphic model.

Figure 7. Isochronous stratigraphic framework in the studied area. (a) Y2 sequence analysis histogram. (b) across Y2 stratigraphic model. (c) Cross-Y2 stratigraphic framework.

Figure 8. Cross-well profile of facies-controlled high-resolution inversion. (a) Waveform facies-controlled inversion. (b) Constrained sparse spoke inversion. (c) Facies-controlled characteristic curve simulation (The arrows in the figure all point to the location of the Changxing Formation bioclastic shoal).

Figure 9. Lithology statistics of shoal reservoirs in the Changxing formation of Yuanba. (a) Thickness of reservoirs; (b) gamma against wave impedance; (c) histogram of wave impedance.

Figure 10. Plan of reservoir prediction of the multi-level shoals of the Changxing formation. (a) First-level shoal reservoir. (b) Second-level shoal reservoir.

References

1. Mancini, E.A.; Llinas, J.C.; Parcell, W.C.; Aurell, M. Upper Jurassic thrombolite reservoir play, northeastern Gulf of Mexico. AAPG Bull.; 2004; 88, pp. 1573-1602. [DOI: https://dx.doi.org/10.1306/06210404017]

2. Lavoie, D.; Chi, G.; Urbatsch, M.; Davis, W.J. Massive dolomitization of a pinnacle reef in the Lower Devonian West Point Formation (Gaspé Peninsula, Quebec): An extreme case of hydrothermal dolomitization through fault-focused circulation of magmatic fluids. AAPG Bull.; 2010; 94, pp. 513-531. [DOI: https://dx.doi.org/10.1306/09080909083]

3. Ma, Y.Z.; Seto, A.; Gomez, E. Depositional facies analysis and modeling of the Judy Creek reef complex of the Upper Devonian Swan Hills, Alberta, Canada: Reply. AAPG Bull.; 2011; 95, pp. 172-174. [DOI: https://dx.doi.org/10.1306/06081010068]

4. Nikitin, Y.I.; Astafyev, E.V.; Akhtyamova, I.R.; Shakirova, G.V.; Shirokovskich, O.A. Exploration and research of oil and gas bearing zones controlled by reef using the regional criteria. Neft. Khozyaystvo-Oil Ind.; 2017; 2017, pp. 64-69. [DOI: https://dx.doi.org/10.24887/0028-2448-2017-9-64-69]

5. Kuznetsov, V.G.; Zhuravleva, L.M. Reef Formations in the West Canada Basin and Their Oil and Gas Potential. Lithol. Miner. Resour.; 2018; 53, pp. 236-251. [DOI: https://dx.doi.org/10.1134/S0024490218030045]

6. Słonka, Ł.; Krzywiec, P. Seismic Characteristics and Development of the Upper Jurassic Carbonate Buildups from the Miechów Trough (Southern Poland). Geosciences; 2020; 10, 239. [DOI: https://dx.doi.org/10.3390/geosciences10060239]

7. Muhammad, M.S.; Bourget, J. Controls on the distribution and growth of isolated carbonate build-ups in the Timor Sea (NW Australia) during the Quaternary. Mar. Pet. Geol.; 2015; 62, pp. 123-143. [DOI: https://dx.doi.org/10.1016/j.marpetgeo.2015.01.014]

8. Kaminskas, D.; Michelevičius, D.; Blazauskas, N. New Evidence of an Early Pridoli Barrier Reef in the Southern Part of the Baltic Silurian Basin Based on Three-Dimensional Seismic survey, Lithuania. Est. J. Earth Sci.; 2015; 64, 47. [DOI: https://dx.doi.org/10.3176/earth.2015.09]

9. Jing, X.Y.; Zhang, X.Q.; Gao, L.; Zhang, M.D.; Li, G.R. Research on the Reservoir Characteristics and Controlling Factors of the Changxing Formation Reef in Yuanba Gas Field of Sichuan Basin, China. Northwestern Geol.; 2021; 54, pp. 39-44. (In Chinese) [DOI: https://dx.doi.org/10.19751/j.cnki.61-1149/p.2021.03.015]

10. Hu, D.F.; Wang, L.J.; Huang, R.C.; Zhang, Q.F. Petroleum Exploration History and Enlightenment in Sichuan Basin: A Case Study on Sinopec Exploration Areas. Xinjiang Pet. Geol.; 2021; 42, pp. 283-290. (In Chinese) [DOI: https://dx.doi.org/10.7657/XJPG20210304]

11. Britta, W.; Hermann, B.; von Hartmann, H.; Tanner David, C. S-wave experiments for the exploration of a deep geothermal carbonate reservoir in the German Molasse Basin. Geotherm. Energy; 2021; 9, 6. [DOI: https://dx.doi.org/10.1186/S40517-021-00189-W]

12. Carl, B.; Heather, B.; Matthew, R.; John, P. Enhancing Paleoreef Reservoir Characterization through Machine Learning and Multi-Attribute Seismic Analysis: Silurian Reef Examples from the Michigan Basin. Geosciences; 2021; 11, 142. [DOI: https://dx.doi.org/10.3390/geosciences11030142]

13. Dahlgren, K.I.T.; Lindberg, B. Buried sedimentary mounds in the Nyk High area, Mid-Norwegian margin: Are they Miocene-Pliocene cold-water coral mounds?. Nor. Geol. Tidsskr.; 2005; 85, 295.

14. Cai, X.Y. The subtly method of reservoir and exploration effects on the organic reef-beach body of Changxing Formation, Yuanba Area, Northeastern Sichuan. Strateg. Study CAE; 2011; 13, pp. 28-33. (In Chinese)

15. Pu, Y. Multi-scale seismic identification of depref-bank reservoirs in the Yuanba area. Nat. Gas Ind.; 2011; 31, pp. 27-31. (In Chinese) [DOI: https://dx.doi.org/10.3787/j.issn.1000-0976.2011.10.006]

16. Tiago, M.A. Polygonal mounds in the Barents Sea reveal sustained organic productivity towards the P–T boundary. Terra Nova; 2016; 28, pp. 50-59. [DOI: https://dx.doi.org/10.1111/ter.12190]

17. Liu, G.P.; You, Y.C.; Feng, Q. Research on reservoir connectivity based on spectral imaging technology for the Changxing ref in the Yuanba area. Geophys. Prospect. Petroleum.; 2017; 56, pp. 746-754. (In Chinese) [DOI: https://dx.doi.org/10.3969/j.issn.1000-1441.2017.05.015]

18. Yan, L.; Cheng, B.; Xu, T.; Wei, S.; Qian, Z. A study and application of high-resolution methods for reef reservoir identification. Acta Geophys.; 2017; 65, pp. 907-917. [DOI: https://dx.doi.org/10.1007/s11600-017-0074-1]

19. Wang, H.; Jian, G.M.; Ke, G.M.; Xu, S.C.; Lv, Q.B. Reef Identification and Reservoir Fine Characterization of Changxing Formation in Yuanba Gas Field, Sichuan Basin. Nat. Gas Ind.; 2019; 39, pp. 107-112. (In Chinese) [DOI: https://dx.doi.org/10.3787/j.issn.1000-0976.2019.S1.018]

20. Hu, X.Q.; Liu, H.; Tan, X.C.; Xiao, C.J.; Tan, Q.; Li, J.H.; Yi, C. Shoal-to-Strata Ratio Technique and Its Use in Describing Carbonate Rock Sedimentary Facies in the Eastern Sichuan Basin. J. Mar. Sci. Eng.; 2022; 10, 825. [DOI: https://dx.doi.org/10.3390/jmse10060825]

21. Zhou, G.; Zheng, R.C.; Luo, R.; Liu, H.Y.; Wen, H.G. Model for Hydrocarbon Migration and Entrapment Within Faulted Structures. AAPG Bull.; 1989; 73, pp. 803-811. (In Chinese) [DOI: https://dx.doi.org/10.1306/44b4a271-170a-11d7-8645000102c1865d]

22. Luo, Z.L. Guidance Function of Emei Taphrogenesis Viewpoint on Discovery of Large Gas Province in Northeastern Sichuan. Xinjiang Pet. Geol.; 2012; 33, pp. 401-407. (In Chinese)

23. Tan, X.C.; Luo, B.; Jiang, X.F.; Liu, H.; Li, L.; Zhang, B.J.; Zhou, J.; Zhao, C.C. Controlling Effect of Basement Fault on Changxing Formation Reef in Sichuan Basin. Geol. Rev.; 2012; 58, pp. 277-284. (In Chinese) [DOI: https://dx.doi.org/10.16509/j.georeview.2012.02.004]

24. Wen, L.; Zhang, Q.; Yang, Y.; Liu, H.Y.; Che, Q.; Liu, W.; Zeng, Y.X.; Wei, X.W.; Xu, L.; Liu, M. et al. Factors controlling reef-bank reservoirs in the Changxing-Feixianguan formations in the Sichuan Basin and their play fairways. Nat. Gas Ind.; 2012; 32, pp. 39-44. (In Chinese)

25. Zhang, Q.F. Geological Characteristics and Enrichment Model of the Reef-Bank model Gas Field in the Sichuan Basin. Acta Geol. Sichuan; 2021; 41, pp. 259-263. (In Chinese) [DOI: https://dx.doi.org/10.3969/j.issn.10060995.2021.02.014]

26. Guo, T.L. Gas acumulation conditions and key exploration & development technologies in Yuanba gas field. Acta Pet. Sin.; 2019; 40, pp. 748-760. (In Chinese) [DOI: https://dx.doi.org/10.7623/syxb201906011]

27. Huaibo, L.; Rigby, J.K.; Guisen, L.; Kedong, X.; Lingshan, L. Upper Permian Carbonate Buildups and Associated Lithofacies, Western Hubei-Eastern Sichuan Provinces, China. AAPG Bull.; 1991; 75, pp. 1447-1467. [DOI: https://dx.doi.org/10.1306/0C9B2961-1710-11D7-8645000102C1865D]

28. Shi, X.C.; Wu, H.Z.; Liu, Y. High Efficient Development Technology and Practice of Yuanba Ultra-Deep High Sulfur Reef Gas Field; China Petrochemical Press: Beijing, China, 2018; pp. 2-3. Available online: https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFDAUTO&filname=TRQG2019S1026&uniplaform=NZKPT&v=pkmUzwcfEOcTyisfdWpcxtMcBZHVsuLnTCTJ64ghjYbmfIUjMc73Z12wP0_PPGy (accessed on 1 May 2018).

29. Wu, Y.J.; Liu, L.P.; Dai, P.; Ren, H.X.; Liu, Y.Y. Gas-water distribution characteristics and early identification of water invasion of Permian Changxing Formation gas reservoir in Yuanba, Sichuan Basin. Mar. Orig. Pet. Geol.; 2021; 26, pp. 1-8. (In Chinese) [DOI: https://dx.doi.org/10.3969/j.issn.1672-9854.2021.03.006]

30. Yuan, H.H. Evaluation of the adaptability of improved resolution technology: A case study of turbidite in Dongying Sag. Geol. Rev.; 2021; 67, pp. 271-272. (In Chinese) [DOI: https://dx.doi.org/10.16509/j.georeview.2021.s1.120]

31. Sun, Y.Z.; Huang, J.; Yu, W.Z.; Hu, W. End-to-end high-resolution seismic processing method based on U-Net network. Prog. Geophys.; 2021; 36, pp. 1297-1305. (In Chinese) [DOI: https://dx.doi.org/10.6038/pg2021EE0349]

32. Pauget, F.; Lacaze, S.; Valding, T. A Global Approach in Seismic Interpretation Based on Cost Function Minimization. Proceedings of the SEG International Exposition and Annual Meeting 2009; Houston, TX, USA, 25–30 October 2009; pp. 2592-2596. [DOI: https://dx.doi.org/10.1190/1.3255384]

33. Ding, L.B.; Shao, D.L.; Wang, H.Q.; Cao, G.W.; Fan, G.Z. Global approach for automated seismic interpretation and its application in deep water exploration of Myanmar Rakhine offshore. Prog. Geophys.; 2018; 33, pp. 2102-2107. (In Chinese) [DOI: https://dx.doi.org/10.6038/pg2018BB0538]

34. Hoyes, J. A review of “global” interpretation methods for automated 3D horizon picking. Lead. Edge; 2011; 30, pp. 38-47. [DOI: https://dx.doi.org/10.1190/1.3535431]

35. Li, S.H.; Shi, G.S.; Jiang, N.C.; Hu, H.; Li, R.; Yu, Y. Identification of Bioclastic Beach of Maokou Formation in Langzhong Area, North Sichuan. Oil Geophys. Prospect.; 2021; 56, pp. 815-825. (In Chinese) [DOI: https://dx.doi.org/10.13810/j.cnkii.ssn.1000-7210.2021.04.015]

36. Wang, D.Y. Pre-Stack Seismic Inversion Based on Markov Chain Monte Carlo Method; China University of Petroleum: Beijing, China, 2012; Available online: https://kns.cnki.net/kcms/detail/detail.aspx?FileName=1014017529.nh&DbName=CMFD2014 (accessed on 30 May 2012). (In Chinese)

37. Wei, M.; Li, S.J.; Shi, X.P.; Wang, S.H.; Yang, G.J.; Wang, Y.T. Application of a seismic motion inversion method for thin reservoir prediction in the Zhuang-3block of the Central Jungar Basin. Geophys. Prospect. Pet.; 2021; 60, pp. 643-651. (In Chinese) [DOI: https://dx.doi.org/10.3969/j.issn.1000-1441.2021.04.013]