The piedmont of the complex mountain in southwestern Tarim Basin, located at the intersection of the West Kunlun orogenic belt and the Tianshan orogenic belt, is one of the scarce extremely-challenging exploration areas in China. The high-dip angle and fragmentized underground structures induced by strong compression of the two orogenic belts and the complex and varying near-surface seismic geological conditions jointly lead to extremely low signal-to-noise ratio(SNR) of raw seismic data, bringing difficulty to the imaging of the underground structures. For this reason, there has been no substantial breakthroughs in oil and gas exploration in this area. In recent years, great efforts have been devoted to seismic exploration. In view of seismic acquisition, near-surface seismic survey is strengthened, the shooting and receiving parameters are optimized, the geometry design is updated, and the beam seismic and high-density wide-azimuth 3D seismic techniques are upgraded. In view of seismic processing, some attempts are made in constrained first arrival tomography inversion, strong energy noise suppression at "black triangle belt" of the seismic records, true surface velocity modeling, and depth migration imaging. By virtue of these efforts, the quality of seismic data has been continuously improved, and many drilling targets have been discovered. Two major breakthroughs have been made in the Jurassic sandstone and the Carboniferous-Permian carbonate rocks for the first time, opening up a new stage of oil and gas exploration in the piedmont in southwestern Tarim Basin.

Sim-source OBN acquisition is a hot spot in seismic exploration. There are three coding methods: time delay, phase, and scan length, for land vibrator data. Offshore seismic exploration uses air guns, and thus only time delay coding is feasible. The point is how to encode time delay series for de-blending. We propose a quantitative coding criterion for uniform random time distribution based on normal distribution constraints. The optimal time series is constructed using target function constraints, and the features of other three random time distributions are investigated in common detector gathers. Iterative inversion in the FKK domain is performed for de-blending, and single-shot tests show better performance of optimal time series than of other three distributions. A 3D offshore survey shows good results of data separation using the time delay series constructed based on the optimal criterion. Research results offer technical support to source parameter optimization in OBN acquisition.

Considering the difficulty in identifying faults in medium-deep strata with low signal-to-noise ratio (SNR), a multi-scale fault identification method based on Radon transform is developed. Specifically, the poststack seismic data are processed by noise suppression to improve the SNR of raw data. The noise suppression mainly deals with random interference and high-frequency noise, and removes weak boundary information caused by rock disturbance. The seismic responses at different frequency bands to different orders of faults are monitored through frequency division processing, and then sensitivity analysis is conducted to optimize the sensitive seismic attributes of faults of different scales. Finally, the Radon transform is used for edge detection enhancement to improve the fault edge effect. The method is essentially to introduce the Radon transform based on image reconstruction theory into the three-dimensional interpretation of faults, which is different from conventional edge detection algorithms. It is an integral operation with the characteristics of linear enhancement and high noise resistance. This method was applied for the first time in the shallow water area of the Zhu-I Sag, the Pearl River Mouth Basin. The results demonstrate that the proposed method significantly improves the accuracy and resolution of fracture prediction compared to conventional ant body slicing. It reveals that the LF_A buried-hill structure of the Lufeng 13 sag mainly develops two groups of fractures (NNW and NE), guiding the identification of high-quality granite reservoir distribution. The method helps confirm the fault development in the EP_ B structure of Yangjiang East sag, which is believed to be a left echelon structure controlled by strike slip. There is a large NEE-trending central strike slip fault zone in the Lufeng C sag, which, together with the NW-trending regulating fault, controls the formation of central nose shaped belt traps. Under the new model, a number of structural traps have been discovered, improving the exploration potential of the sag.

OBN data have better quality than streamer data, but OBN acquisition is complicated and expensive. Thus, it is necessary to perform a detailed assessment of OBN and streamer data from the perspectives of imaging and structure and reservoir characterization as well as OBN comparative advantages for hydrocarbon exploration and development. The case study deals with deep-water sub-salt formations. We first examine the discrepancies between OBN and streamer data in offset, azimuth, illumination, signal-to-noise ratio, and seismic velocity field in accordance with their different geometries. In view of the challenges in complicated sub-salt formation imaging, the comprehensive analysis involves diffracted-wave imaging, fault imaging, basement imaging, and vertical and lateral resolutions of these two data types and their interpretation results of structure depth and inversion. With respect to abundant low frequencies in OBN data, we use filtered log data to investigate their contribution to the characterization of geologic boundaries and propose a method of resolution calculation using statistical wavelet. Compared with streamer data, OBN data show better quality owing to large-offset wide-azimuth acquisition and consequent better illumination and more accurate velocity field. Similar sub-salt structural features and trends of inversion results could be obtained using streamer and OBN data, but the latter performs better in detailed structure and reservoir characterization. In summary, OBN data feature better imaging, more distinct fault and basement reflections, higher resolution, wider band, more abundant low frequencies, smaller error of structure depth, and more detailed inversion results compared with streamer data. Research results may be used for reference in the study of streamer and OBN data in other oilfields.

There are different vintages of seismic data acquired and processed using different parameters in the northern South China Sea, but imaging quality is not good enough for Mesozoic hydrocarbon exploration. It is necessary to conduct an in-depth study of major controls on seismic imaging. Based on available 2D seismic data, we investigate geologic setting and sedimentation in this area and perform forward modeling of the representative geologic framework to dissect geologic and geophysical factors related to seismic imaging of the Mesozoic Erathem and establish major controls on seismic imaging qualitatively and quantitatively in different geologic conditions. The major controls include two strong reflection boundaries at the seabed and Cenozoic bottom, low signal to noise ratio caused by small impedance contrasts inside the Mesozoic Erathem, various disturbing waves such as diffracted waves and multiples that blur effective signals, and scattered illumination energy induced by irregular seabed. The in-depth study of major controls on seismic imaging offers support to seismic data acquisition and processing targeting the Mesozoic Erathem in this area. Corresponding acquisition solutions include increasing excitation energy, folds, sinking depth of air gun cables, and dimensions to obtain more deep-zone reflection energy. Corresponding processing solutions include self-adaptive ghost suppression, multi-domain high signal-to-noise ratio processing, and energy compensation based on frequency resonance to address the issue of weak reflections, broadband integrated filtering to attenuate multiples, and high-precision velocity analysis through along-horizon high-density velocity picking based on the understanding of geology and refracted waves.

In Huizhou H5 Oilfield in the Huizhou Sag of the Pearl River Mouth Basin, the sandstone reservoir in the lower section of the Paleogene Enping Formation is the major oil reservoir. Tight tuffaceous sandstone is developed in the reservoir due to volcanism. Such sandstone has led to the degradation of reservoir performance. The prediction of high-quality reservoirs is challenging. Accurate identification of tuffaceous sandstone and characterization of its distribution are crucial to the prediction of high-quality reservoirs in Huizhou H5 Oilfield. In this paper, petrophysical modeling of tuffaceous sandstone was conducted using available drilling and seismic data, and sensitive petrophysical elastic parameters for identifying tuffaceous sandstone were obtained. Then, quantitative reservoir prediction was performed through seismic data prestack inversion and artificial intelligence deep learning. Furthermore, tight tuffaceous sandstone and its distribution were identified and characterized, highlighting high-quality sandstone reservoirs. Finally, the technology and process of artificial intelligence reservoir prediction driven by petrophysical modeling of tuffaceous rocks and petrophysical models were proposed. This technology was applied to the reservoir evaluation of Huizhou H5 Oilfield, contributing to the successful identification of the tuffaceous sandstone in the H5-3d well area and the lower section of the Enping Formation in the H5-5d well area before drilling, and to the accurate characterization of the tuffaceous sandstone distribution, which provide an important basis for the subsequent drilling of appraisal wells and reserve declaration.

The Huizhou 21 sag of the Pearl River Mouth Basin is known for its complex underground structures and faults. Due to outdated seismic acquisition and processing techniques, there are issues with unclear fault imaging, unconfirmed faulting points, and blurry imaging of the fault's downthrown side. These issues greatly impact the accurate determination of trap morphology, structural highs, and amplitude of structural traps. To address these issues, deep processing techniques for seismic signals in the fractured zones are investigated. Seismic signals are processed finely to maximize the potential of seismic data collected by conventional towed acquisition technique. In this way, the low-to medium-frequency signals and effective seismic reflections within a large incidence angle range that can effectively improve the imaging results for complex fault zones are obtained to improve the signal-to-noise ratio of seismic data in the fractured zones. Meanwhile, high-fidelity seismic signals are utilized for high-precision prestack depth migration velocity modeling constrained by data and geological models. For complex fault zones, large-shot seismic data and low- to medium-frequency effective signals are efficiently utilized, and high-precision grid tomography velocity modeling driven by dominant signals is conducted to greatly improve the velocity picking accuracy. Moreover, the velocity model is effectively constrained with geological models, and coupling with the correlation between imaging effects and velocity models, the final imaging velocity model is determined. This imaging approach significantly improves the structural imaging effect of the complex fault zone in the Huizhou 21 sag. It has been successfully applied in exploration research in the Huizhou 21 sag by reducing the impact of unconfirmed faults on the evaluation of structural traps and improving the accuracy of structural trap identification, thereby driving the exploration research in this area.

Several key technologies have been adopted to overcome the challenges such as weak signals, multiple wave interference, and low signal-to-noise ratio in the seismic data of the Paleogene strata on the western gentle slope of Panyu 4 sag of the Pearl River Mouth Basin. These technologies include 3D τ-p transform ghost wave suppression, shallow-water multiple wave suppression, and high-precision velocity modeling. These technologies can enhance the signal-to-noise ratio of seismic data to improve interpretation accuracy, and achieve breakthroughs particularly in the delineation of intra-sag strata contacts and basement imaging. It is found that effective reflections are typically concentrated in 6~36 Hz, but the absence of low- and high-frequency signals limits the detailed delineation of strata contacts. Hence, the use of the aforementioned technologies has been proposed to improve the quality of seismic data. The ghost wave suppression technology successfully separates and suppresses ghost wave noise, while shallow-water multiple wave suppression technology specifically predicts and attenuates multiples. The integrated application of these techniques has resulted in significant improvements in the imaging of the Paleogene interior structure, clearer wave group characteristics, more reasonable strata contact relationships, and notably enhanced imaging quality of complex faults in the medium-deep strata. The conclusion emphasizes that the application of ghost wave suppression and shallow-water multiple wave suppression technologies has significantly improved the quality of medium-deep seismic data, which is of great importance for oil and gas exploration in the Pearl River Mouth Basin, especially in Panyu 4 sag and its adjacent areas. It provides a new technical approach for improving the quality of seismic imaging in similar depositional environments.

Owing to complex Neogene channel sands stacking and abrupt change of sedimentary facies in the river-lake transition zone in Bohai, microfacies classification using seismic attributes may be quite uncertain. We propose an improved amplitude-spectrum-distance K-center clustering method to mitigate the uncertainties. Microfacies calibrated using log data are set as the initial clustering center and constraints of clustering, and amplitude-spectrum distance is used to measure the differences among seismic wave forms for microfacies prediction. In view of the geologic conditions in Field A, Bohai Bay, a 3D geologic model is built with superposed channel sands. The model test shows a prediction accuracy of 95%, 15% higher than that of the K-mean clustering in the time domain. Six channel styles, i.e. mudstones, single-phase channel edge, single-phase channel body, superposed channel edges, superposed channel edge and body, and superposed channel bodies (or multi-phase superposed channels), could be identified, based on which Neogene channels are classified into 10 microfacies. A large reservoir, S, is classified into 6 microfacies: distributary channel, mouth bar, sand sheet, crevasse channel, inter-distributary bay, and natural levee, among which distributary channel and mouth bar are the promising microfacies for production. The case study demonstrates the feasibility of the improved classification method.

In recent years, the node seismic acquisition technology has advanced rapidly and been widely applied in oil and gas exploration. Compared to cable-based instruments, node seismometers have outstanding advantages in terms of mobility, arrangement, and efficiency. However, they cannot acquire seismic data in a real-time manner, which poses certain challenges to quality control and data evaluation. Ensuring the quality of node seismic acquisition and maintaining the proper functioning of node seismometers have become crucial aspects during operations. In other words, it is necessary to guarantee the equipment stability and safety. Combining with field operation modes and workflows, this paper introduces the design concepts of multi-type node QC mobile App and QC data analysis platform, and discusses key technologies such as node communication, mobile positioning and navigation, node location monitoring, multi-mode inspection, data visualization, and statistical analysis. The multi-type node QC mobile App and QC data analysis platform are developed by using the Visual Studio 2022 integrated development environment and the MVC design pattern, and address the issues such as the limited generality of current node QC software and the low automation in QC data evaluation. In practical applications, the software proposed in this paper demonstrates stable performance, runs smoothly, significantly improves the QC efficiency and effectiveness of node seismometer, and effectively ensures the quality and data integrity of node data acquisition while reducing operation costs.

Underground rocks are generally anisotropic, which affects the velocities, amplitude and phase of seismic waves. The acoustic anisotropic wave equation is a simplification of the anisotropic elastic wave equation. It is widely used in Full Waveform Inversion(FWI) and Reverse Time Migration(RTM). Minimizing boundary reflections is a fundamental aspect of seismic wave numerical modeling. The perfectly matched layer(PML) method is commonly employed to absorb such reflections, but it encounters stability issues in anisotropic media modeling. To address this, in this study, we attempt to apply the stiffness reduction method(SRM) for modeling the acoustic anisotropy wave equations. SRM achieves the reflection absorption by introducing the damping terms into the equations and gradually reducing the stiffness. Firstly, the SRM and PML loading methods and difference schemes in the modeling of acoustic TI wave equations are determined. It is found that SRM is superior to PML in computational storage efficiency and speed. Then, numerical modeling was performed on vertical transversely isotropic(VTI) media, tilted transversely isotropic(TTI) media, layered media and BP model to compare SRM and PML in adsorption attenuation performance. The simulation using VTI media indicates that SRM requires a memory only as a half of PML and reduces the computation time by 25% compared to PML for the same absorbing layer thickness under the given experimental conditions. While SRM demonstrates a slightly lower absorption efficiency than PML in VTI media, it exhibits superior stability in TTI media, layered media, and BP model simulations. These findings prove the practicality and reliability of SRM in modeling the acoustic TI wave equations.

Based on the discussion of the applicability and limitations of existing high-resolution processing techniques, we propose a new method based on the quantitative constraint function of high-frequency noises to improve seismic resolution and signal-to-noise ratio in the context of amplitude preservation. In view of the effective bandwidth of poststack seismic data, we formulate the maximum probability criterion and quantitative constraint function for iterative high-frequency noise reduction to expand effective frequency band step by step and improve the signal-to-noise ratio of full-band data. Reflection coefficients calculated iteratively are convolved with seismic wavelet to obtain high-resolution seismic data. Compared with other high-resolution processing techniques, our method yields a data volume with high vertical resolution, high dominant frequency, and wide band; effective apparent dominant frequency increases from 50 to 100 Hz, and apparent resolution nearly doubles. Based on the frequency data and wave impedance inversion, the qualitative and quantitative prediction of 2~5 meters geological targets is realized, and the verfication test matches with more than 90% of the targets, which effectively guides the well location deployment. This new method performs better than routine techniques in high-frequency signal enhancement within the effective frequency band and thus realizes high- and low-frequency signal reconstruction for full-band data.

In order to obtain realistic seismic fault training samples, a seismic fault training sample synthesis method based on cycle-consistent adversarial networks is proposed.This method takes randomly generated fault labels and real fault data as inputs, and employs an unsupervised adversarial network to learn the relationship between fault labels and fault data and generate seismic fault samples that match the characteristics of the fault labels, thereby obtaining a labeled fault training sample set.This new method alleviates the problem of training dataset shortage for deep learning in seismic fault interpretation.A quantitative analysis of the mean frequency and textural difference between the synthetic and real faults was performed, showing a high similarity between the two.The neural network trained with the fault samples generated by this method was applied to real data for testing and comparison.The results show that these samples are realistic and reliable.Furthermore, this method can generate targeted faults for different work areas, and be flexibly and effectively applied in intelligent seismic fault detection in work areas.

Seismic attributes can be used to predict discontinuous anomalies (e.g.fractures) in unconventional reservoirs.For carbonate fractured-vuggy reservoirs, however, the strong heterogeneity leads to complex seismic responses, making the fine description of reservoir structure very challenging.In this paper, a method for fine prediction of fractures in target reservoirs was proposed.Firstly, the structure-orientated filter (SOF) is used to remove the noise from seismic data, so that the effective information in seismic data is improved while the characteristics of fault structures are preserved.Secondly, the multi-scale characteristics of seismic signals are adequately identified with the support of multiple methods/techniques including 3D multi-scale volumetric curvature (3D-MSVC) attribute, ant tracking, maximum likelihood attribute, and 2D Hilbert transform based multi-scale reservoir fracture edge detection.Finally, the multi-source characteristic information of fractured-vuggy reservoir obtained by these methods/techniques is optimized and fused to mitigate the ambiguity and uncertainty in fine seismic prediction and realize the synchronous detection of featured information of carbonate fractured-vuggy reservoirs at different scales.The application results of actual seismic data showed that the method proposed in this paper can finely predict the fractures, vugs, ancient channels and subtle faults, so as to efficiently support the exploration and development of carbonate fractured-vuggy reservoirs.

Despite many wells with high yields from non-structural traps in the Lower Permian Maokou Formation, the Sichuan Basin, which demonstrates that karst reservoirs are an important type in Maokou lithologic gas reservoirs, the genesis and seismic responses of karst reservoirs are uncertain; thus, it is difficult to assess various seismic techniques used in different areas.Based on the study of structural and stratigraphic features and seismic response of karst reservoirs in the Maokou Formation, Zigong area, southern Sichuan, following conclusions are reached.Intra-platform beaches, which mainly turn up in Mao3 and Mao2b, are the material basis of karst reservoirs.Differential uplift caused by the Dongwu movement is the major control on karst reservoirs in the area of interest.Faults occurred after the deposition of the Maokou Formation; karst reservoirs at Maokou top were dominated by later sustained weathering locally, and those inside the Maokou Formation were reworked by later fault activities.Seismic responses of karst reservoirs vary with the distance to Maokou top; there is a negative correlation between epikarst reservoir properties in Mao3 and peak amplitude at Maokou top, and there is a positive correlation between internal karst reservoir properties in Mao2b and interior peak amplitude in the Maokou Formation.The occurrence of epikarst reservoirs of scale is not dependent on faults.

Residual oil analysis is important in the middle and late stages of oilfield development to maintain and increase the production.Time-lapse seismic has significant advantages in residual oil analysis, but it requires a strict consistency of adjoining two or more sets of seismic data in acquisition and processing, which limits its application.In order to overcome the shortcomings of traditional time-lapse seismic in residual oil analysis, this paper proposes the concept of pseudo-time-lapse seismic, which focuses on the responses of seismic data to geological reservoir problems instead of the consistency of seismic data.Firstly, using various data with pseudo-data eliminated, forward analysis is performed to get a reliable and accurate knowledge of geological reservoirs.Then, the differences in the knowledge of geological reservoirs obtained from analysis in two rounds are identified, and accordingly, the change patterns of reservoir development performance are determined to guide the residual oil analysis.In this paper, the pseudo-time-lapse seismic is applied to the analysis of residual oil in Bohai F Oilfield.In this oilfield, the target formation is shallow and composed of meandering river sediments, with good physical properties.The two sets of seismic data of the oilfield are acquired with 12 years apart, and they are apparently different in acquisition and processing.Identifying the differences in knowledge of geological reservoir from the two sets of seismic data allows for an effective diagnosis on the change of oil-water relationship caused by water flooding in the reservoir.The findings are more instructive than the results of reservoir numerical simulation, and can help further delineate the area with rich residual oil, providing a basis for well deployment.

Single borehole reflection imaging is an effective approach to the direct evaluation of geologic structures and reservoirs around a borehole using reflected shear waves extracted from array acoustic logging data.To enhance azimuthal resolution and radial depth of investigation, the key to data processing is complete separation of reflections from original data with the minimum loss of signal energy.Existing separation methods suffer from the problems of modes mixing and energy loss of extracted reflections and consequent small resolution and radial depth.We propose a new algorithm of 2D variational mode decomposition for wave separation without great time consumption of recursive iteration.Original signals are decomposed into several inherent modes with different directions and vibration characteristics, and appropriate modes are used for reconstruction and reflection separation.Reflected waves are extracted from synthetic data with different noise levels and field data using 2D variational mode decomposition, median filtering, and F-K transform; our method performs better than median filtering and F-K transform in effective suppression of mode waves and random noises and extraction of complete reflected signals.Synthetic and field data tests demonstrate the effectiveness and advantages of our method.

Shale oil resources for exploration in the Xingouzui Formation, the Jianghan Basin were geologically assessed to be 3.6×10⁸t, and oil flow was obtained at several well sites from argillaceous dolomites in the lower Xingouzui Formation, the Chentuokou sag which was evaluated to be Grade-I promising prospect.Economic shale oil production is usually achieved using horizontal well drilling+staged fracturing; thus stress information is important to fracturing stimulation of horizontal wells.A crucial issue in 3D stress modeling is how to improve cross-well prediction.The finite element method routinely used cannot yield good results due to its own limitations; thus our efforts focus on seismic inversion to improve cross-well prediction.In view of the problems related to multi-mineral, local sulfate rocks, and complicated stress conditions in the area of interest, log data-based evaluation is performed under the constraints of lab data for Xingouzui shale beds, and the results are validated using core measurements of X-ray diffraction.Rock physical modeling is achieved using self-consistent approximation to calculate elastic parameters of shale beds; modeled results agree with measured curves.Based on formation pressure, rock mechanical properties measured in the lab, logging interpretation, and dynamic-static relations of different lithologies, effective stress ratio and tectonic strain methods are employed to calculate in-situ stress parameters at well sites.The final 3D stress model of Xingouzui shale beds is established using prestack seismic inversion results and the correlation and parameter constraints equal to those used in 1D geomechanical modeling.By comparison, pseudo-curves extracted from the 3D model are basically consistent with 1D modeled results at most well sites in the area, which demonstrates the feasibility and credibility of this method in stress evaluation for Xingouzui shale beds.

We present a modified barnacles mating optimizer (MBMO) to recuperate model parameters, including density and geometric parameters, from gravity anomalies generated by dipping faults.Barnacles mating optimizer (BMO) is a recently proposed nature-inspired gradient-free global optimizer, which simulates the unique mating and reproduction characteristics of barnacles quantitatively.BMO has been proved to be an effective optimizer strategy in solving various optimization problems.However, its efficiency can be easily affected by the selection of an algorithm-based parameter (the genital length p_l), the methods of evolving barnacles to enhance the stochastic feature, and the processing techniques when solutions exceed the searching bounds.Therefore, we propose a modified barnacles mating optimizer combining a variable p_lvar technique, novel evolving strategy and out-of-bounds correction method.We investigate the influence of genital length p_lvar on BMO using synthetic noise-free gravity anomalies and certify the effectiveness of the variable p_l strategy and the superiority of MBMO.Further comparative study of the accuracy, stability and practicability of BMO, MBMO and particle swarm optimizer (PSO) uses synthetic noise-free gravity anomalies, noise-corrupted gravity anomalies with different noise degrees (10% and 30%), and field data obtained from the Gazelle fault in Egypt.MBMO is demonstrated to have smaller fitting difference, more accurate model parameters, and higher stability than additional two methods.It thus can be applied to a wide range of geophysical inverse problems.

Ancient land form of the passive continental marginal slope is important to deep-water petroleum exploration owing to its influence on gravity-flow reservoirs.In an area with a small number of wells, a routine practice for palaeogeomorphologic restoration is using formation thickness impression method based on 3D seismic data, which suffers from two problems: insufficient sediment supply and consequent semi-starved undercompensation caused by long distance from the slope zone to the sedimentary source and slope gradient leading to morphological distortion of geomorphic units.Inaccurate palaeogeomorphologic restoration has a negative impact on the study of sedimentary system and reservoir distribution.To address these issues, formation thickness impression method is ameliorated from two perspectives according to the features of sedimentary continental slope.One is the feasibility of the method in light of the state of slope compensation obtained using 3D seismic data, and the other is ancient slope correction based on the equilibrium surface of each sequence derived from equal division of sediment volume in accordance with the process of slope evolution in the area of interest.In the case study of the lower continental slope in the Niger Delta Basin, western Africa, there is a good consistency between palaeogeomorphologic restoration and structural evolution in accordance with the idea of tectonic activity domination on geomorphic features.Vigorous thrust-fault activities resulted in faulted anticlines and intra-slope basins, which prevented gravity-flow transportation.Sediments were confined to local areas to form huge lobes.Controlled by mud diapirs, narrow intra-slope basins formed on both sides of the diapiric anticline.Channel-lobe complexes were localized within intra-slope basins and ran subparallel to the inclination of the slope.In addition, there is a good correlation between deep-water sedimentary system derived from seismic attributes and the results of palaeogeomorphologic restoration, which indicates the scientific nature of the modified method and the objectivity of the results.