The formation mechanism of oblique extension was first proposed by Withjack and Mckenzie in the study of multi- phase extensional rift basins^[1,2], to describe the development model of structures when the boundary of rift basin was not perpendicular to the late extension direction. Afterwards, different researchers have carried out many physical modeling experiments on oblique rifting extension. Although the experiment models were not identical in setting, including rigid or plastic basement, single or multi-phases of extensions, they all reflected the oblique extension of pre-existing basement faults at the scale of basin boundary, and recorded and described evolution process of basin boundary and internal structures^{[3,4,5,6,7,8,9]}. The modeling experiments of Tron, Clifton and Bonini et al. show that the angle between the strike of pre-existing boundary fault of rift and the late extension direction has an important influence on the development of faults of different natures, and 45 degrees may be the dividing point between the development of normal faults and oblique slip faults and strike-slip faults. Normal faults and a small number of oblique slip faults would develop if the angle exceeds 45 degrees, while strike-slip faults and oblique slip faults would develop if the angle is less than 45 degrees^{[3-4, 6]}. McClay and Bechis also analyzed the development of accommodation zones and transfer zones at different angles between rift basin and late extension direction through modeling experiments^[7,8,9]. In addition, the oblique extension of pre-existing structures in the interior of basins is also an important part of the oblique extension model. This model is similar to the oblique rift boundary model mentioned above in formation mechanism, but lays more stress on the pre-existing faults within the basin and the possibility of pre-existing faults reactivation and the structure development after fault reactivation under different extensional conditions. Morley et al.^[10,11,12] have done many researches on late oblique extension of pre-existing structures within basins, analyzing their structural styles in details and comparing them with extensional and strike-slip structural styles. Tong et al.^[13,14] described the oblique extension of pre-existing structures as "uncoordinated extension" and demonstrated the stress mechanism of selective reactivation of pre-existing faults. The modeling of oblique extension of pre-existing structures laid more stress on the development of pre-existing faults of the internal basin during multi-phase extension and the influence of pre-existing faults produced by early extension on the development of late extensional structures^[14,15,16].

The stress model under oblique extension of pre-existing structures isn’t a simple pure shear model, so the model of fault development doesn’t follow the classical Anderson’s model completely, but is actually a complicated shear model. Because of the existence of pre-existing structures, the triaxial stresses of macroscopic pure shear stress applied at the late stage would change in direction and magnitude, resulting in the formation of stress situation with compound deformations including pure shear and simple shear. This also explains why in some analogue modeling experiments, with the increase of the angle between the strike of pre-existing faults and the late extension direction, the structural styles transform gradually from pure shear extensional deformation to simple shear strike-slip deformation^{[3-5, 16-17]}. In addition, because the pre-existing faults inevitably produce oblique slip effect in the process of oblique extension^{[4, 10, 16-18]}, the reactivated faults of different strikes would have different ratios of strike-slip to extension.

Although the study on the formation mechanism of oblique extension has been mature, there are relatively few studies on the fault activity regularity of oblique extension structures within the basin, and the petroleum geological significance of oblique extension structures was even less mentioned. In this paper, the structural styles and regularity of fault activity in the eastern Bohai Sea are analyzed; the genetic mechanism and development characteristics of oblique extension structures are discussed; by comparing the drilling data of exploration wells, the control of oblique extension structures to differential hydrocarbon accumulation is identified.

The study area is located in the eastern part of sea area of Bohai Bay Basin which is an intracontinental rift-depression basin developed on the North China Craton^[19,20]. During the Late Mesozoic to Paleogene, large-scale mantle upwelling and lithosphere thinning triggered active rifting in Bohai Bay Basin^[21,22]. At the same time, the subduction of Pacific plate and the collision of Indian Ocean plate with the Eurasian plate caused the strike-slip effect of the Tanlu fault zone^[23,24,25], which resulted in the combined effect of strike-slip and extension. The active rifting in Paleogene gave rise to extensional faults of different strikes at different locations^[26], and the superimposed strike-slip movement also produced derivative faults of different strikes near Tanlu fault zone. In a word, in Paleogene, there are a large number of pre-existing faults of different strikes in Bohai Bay Basin, which is particularly remarkable in the study area cut through by Tanlu fault zone. In Neogene, Bohai Bay Basin entered post-rift thermal subsidence^[27], but there is no consistent understanding on the stress state in Bohai Bay Basin during this period. Some researchers argued that the nearly-EW horizontal compression caused the dextral transpressional strike-slip of Tanlu fault^[28,29,30]; others suggested that the back-arc expansion of the Japanese Sea^[31,32] or the trench retreat^[33] resulted from the Pacific plate subduction in this period caused the nearly NS extension stress, while the EW compression has only occurred since Quaternary^[34].

The study area consists of the southern part of Bodong Low Uplift, Bodong Sag and part of Bozhong Sag (Fig. 1). Tanlu fault zone in NNE strike runs through the study area, and its strike-slip and derivation in Paleogene produced faults of various strikes. The west boundary fault trending NE and the east boundary fault trending NNE of the southern Bodong Low Uplift intersect on the north of the uplift, controlling jointly the Paleogene sedimentation in the sags on both sides of Bodong Low Uplift (Fig. 1a). In Neogene, different from Paleogene, the stress field changed obviously: some NE- trending uplift boundary faults continued to develop, but reducing in control over sedimentation; while at the end of the boundary faults, curved faults with strike gradually changing from NE in the north to NEE-trending in the south developed (Fig. 1b). Briefly, a large number of pre-existing faults of various strikes in Paleogene and obvious stress field transformation in Neogene provided conditions for the development of oblique extension structures in Neogene.

The Cenozoic in study area is relatively complete: from the bottom to the top, there are Paleogene Shahejie and Dongying Formation (1^st-3^rd member), Neogene Guantao (lower member and upper member) and Minghuazhen Formation (lower member and upper member) and Quaternary Pingyuan Formation. Adjacent to the subsidence center of Bozhong Sag, the deposition thickness in study area is up to nearly 4 000 m thick. In recent years, several large-and-medium-sized oil fields have been discovered in Neogene, proving this area has great potential of petroleum exploration^[35,36].

According to the difference in structural styles, the study area can be divided into three parts: north, middle and south. In the northern region (Fig. 2a), the development of the low uplift is controlled by the east-dipping eastern branch fault (F1) and the west-dipping western branch fault (F2). The south part (Fig. 2c) is located at the southern pitching end of Bodong Low Uplift, where, the western branch fault developed continuously, but compared with the north region, the east-dipping eastern branch fault had no obvious controlling effect on the development of the low uplift in Neogene, and another west-dipping fault (F3) became the main fault of the eastern branch. The middle part (Fig. 2b) is a transitional position, where the east-dipping fault less active in Neogene and west-dipping branch fault more active in Neogene intersect.

2.1.1. Type and distribution of faults

According to the types of activity, faults in the study area can be divided into extinct type, newly-formed type and inherited type. The extinct type faults are those active in Paleogene but not active anymore in Neogene. These faults only reach up to the lower second member of Dongying Formation, for example, F4, F5 and F6 (Fig. 2). They are NNE-trending or near NS-trending, mostly found in the uplift boundary or in the two sags of Bodong and Bozhong, and usually have good lateral continuity, running through the three parts. The newly- formed faults refer to those started developing in Neogene,with fault throw larger in the middle part, and smaller in both ends, or larger in the upper part, and smaller in the lower part. They cut down to the third member of Dongying Formation at most (Fig. 2), for example, F7 and F8. Trending NE or nearly EW, most of them occur in the interior of sag and the boundary of the uplift, and are widely distributed in en-echelon formation in the three parts. The inherited type faults have developed continuously from Paleogene to Neogene and run vertically through the whole Cenozoic. Most of them developed along the uplift boundary in NE-trending, for example, F1, F2 and F3 (Fig. 2). At the end of the inherited type faults, there are usually newly- formed faults connecting with them, forming curved faults on the plane (Fig. 1b).

2.1.2. The faults

On the seismic section (Fig. 2), most of the extinct faults are steep with large dip angle, and some of them overlap and join with each other; the late newly-formed faults often take on tilted fault blocks, graben-horst structures and rootless flower-like structures; inherited faults continuously controlled the development of uplift and overlap with newly-formed secondary faults on the sides, forming strike-slip-extensional or extensional structural styles such as flower-like structure and multiple "Y" type structure. The curved faults, such as F1, actually consist of two parts (Fig. 3), the inherited NE-trending section cutting through the deep horizons and the newly- formed NEE-trending section cutting only through the shallow horizons.

In different areas, Paleogene faults have different effects on the deformation during Neogene: in the sags, the Paleogene pre- existing faults had hardly any direct effect on the deformation of Neogene, and the deformations, such as tilting, warping, dragging and reversed dragging of Neogene are mainly affected by newly-formed faults; while in the uplift boundaries, Paleogene faults selectively reactivated during Neogene, controlling the deformation of uplift boundary, resulting in the formation of a large number of dragging, reversed dragging and complex fault blocks, which would provide structural traps for later hydrocarbon accumulation. By comparison, it can be found that the factors influencing the reactivation of Paleogene faults are not the size or the dip of the faults but the strike of the early faults, which will be discussed later.

2.1.3. Development mechanism of Neogene faults

The analysis of structural styles in the study area shows that the structural deformation during Neogene is resulted from oblique stretch in nearly-NS direction of the pre-existing faults. During the process of oblique extension of the Paleogene pre-existing faults, the newly-formed faults on the sag boundary are susceptible to the influence of the pre-existing boundary faults, with the strike assimilating to the inherited faults gradually, forming faults in echelon formation; at the end of the inherited faults, the newly-formed faults could connect with the inherited faults, different in strike, and they comprise curved faults on the plane with gradual change strike. These characteristics have been confirmed by the physical modeling experiments of McClay et al.^[7,8]. In the center of the sags (Bodong and Bozhong) where the boundary faults have little influence, the strike of the newly-formed faults tends to be perpendicular to the extension direction but is still affected by the overall strike of the sag, and the faults also appear in en-echelon formation^{[1, 15]}. According to “the criterion of uncoordinated extension”^[13,14], the pre-existing faults with too large dip angle or too small angle between the strike and extension direction are difficult to be reactivated during oblique extension (the specific value depends on the magnitude of the regional principal stress and the friction coefficient in the stratum), which explains why the nearly NS-trending or NNE-trending faults with large dip angle did not continue to develop in the shallow layer in the study area. Oblique extensional structural styles are also reflected on the section: the uplift boundary faults in the study area often have inherited faults as the trunk ones and newly-formed faults connecting with the trunk fault forward or backward, the flower-like structural style is the result of oblique slip of pre-existing fault^[18]; while in the internal sag, where the pre-existing faults have little influence, extensional structural styles turn up, which are the important manifestation of oblique extension.

The static structural styles in the study area accord with the results of oblique extension of Paleogene pre-existing faults in Neogene. We further verify this viewpoint by statistics of the regularity of fault activity.

2.2.1. Fault activity analysis

The active faults in different periods in the study area were counted, and then the active faults of Paleogene (during the sedimentary period of Shahejie Formation) and Neogene (during the sedimentary period of Minghuazhen Formation) were mapped to show the fault activities in different periods (Fig. 4). The figures show that the active faults during Paleogene are mainly NNE to nearly-NS and NE faults (Fig. 4a), which is consistent with the tectonic background of Bohai Bay Basin in Paleogene. During Paleogene, the NNE-trending Tanlu fault strike-slipped dextrally, deriving NE-trending derivative faults, meanwhile rifting happened^[37,38], the main faults were higher in vertical active rate (see the statistics in Fig. 4a), controlling the sags, as a result, a large number of pre-existing faults of different strikes formed. During Neogene, the active faults were mainly NE-trending inherited faults and NEE-trending newly-formed faults, and the NEE-trending faults had a higher activity rate than faults of other trends. This is more obvious in curved faults (see F1, F2, F3 in Fig. 4b): the NEE section of the curved faults has higher rate of vertical activity than the NE section (see the statistics in Fig. 4b).

2.2.2. Statistics of fault activity

The study on fault displacement shows that the maximum displacement of mature normal faults is positively correlated with the length (strike dimension) of faults in a certain range^{[39,40,41,42,43]}, and unlike normal faults, mature strike-slip faults have horizontal movement and longer extension on the plane but no significant vertical displacement^[44]. Based on this principle, the ratio of maximum displacement to length of oblique slip faults should be between normal fault and strike-slip fault, smaller than that of the normal fault but larger than that that of strike-slip fault. The larger the strike-slip component, the smaller the ratio is; the larger the extensional component, the higher the ratio is. This method can be used to determine the relative strike-slip and extension ratio of faults.

The ratio of maximum displacement to length, i.e. the vertical displacement per unit length, of active faults of different strikes during Neogene in study area were counted and compared with the routine rose diagram of the number and strike of faults (Fig. 5a and 5b). It can be seen that the routine rose diagram of active faults in Neogene has two peaks, indicating that the faults mostly trend NE and NEE, while the vertical displacement per unit length of active faults has a single peak, that is the NEE faults have the largest extensional component, and the extensional component progressively decreases and the strike-slip component gradually increases from NEE to NNW. This also shows that, although some NE faults continued to develop in Neogene, all faults including inherited faults and newly-formed faults conform to a general rule: with the decrease of the angle between the fault strike and extension direction, the strike-slip component gradually increases, and the extensional component gradually decreases.

The two kinds of rose diagrams of active faults in Paleogene of study area were also completed for comparison with active faults in Neogene (Fig. 5c and 5d). It can be seen the active faults in Paleogene vary widely from NS to EW, and don’t show obvious regularity between extension component and strike-slip component as those in Neogene, although NE-trending faults have the highest extension component. This shows that the rifting in Paleogene was dominated by the regional SEE-NWW extension^[26], and meanwhile, the NNE extension induced by strike-slip happened, disturbing the overall active regularity and strike of the faults to some extent, which is obviously different from the fault activity regularity in Neogene.

According to the statistics on reactivation probability of Paleogene pre-existing faults in Neogene (i.e., the ratio of Neogene inherited faults to Paleogene active faults in corresponding directions) in all directions (Fig. 6), pre-existing faults in different strikes differ in reactivation ability during oblique extension. The NEE trending pre-existing faults have the highest probability of reactivation, but smaller in number. The NNE pre-existing faults larger in number are lower in probability of reactivation (Fig. 5c). Thus, the NE-trending faults between the former two kinds of faults become the main inherited faults because of their large number. Consequently, the faults active in Neogene were mainly those striking NE (inherited) and NEE (newly-formed faults).

2.2.3. Formation mechanism of oblique extension structure

Based on the activity statistics above, the viewpoint that the pre-existing faults in study area were subjected to nearly NS oblique extension in Neogene was further confirmed. When the active faults in Paleogene in a certain range of strike continued to develop in Neogene under nearly NS oblique extension, the larger the angle between the pre-existing fault and the nearly NS extension direction, the larger the extension component will be, conversely the stronger strike-slip component will be. When the nearly N-S extension is directly exerted on the shallow layer which is not affected by the pre-existing faults, the NEE to nearly EW new faults are formed with mainly extension component (Fig. 7).

Many scholars have carried out experiments on oblique extension of pre-existing faults. Although the basement settings are different, they generally have the following characteristics: (1) When the angle between the strike of pre-existing boundary faults and the direction of late extension is more than 45 degrees or so, the newly-formed faults are mainly normal faults which lie in a position between the direction perpendicular to the extension and the strike of pre-existing boundary faults^[1,15]. If the angle is less than 45 degrees or so, besides dip-slip faults, newly-formed faults dominated by strike- slip components such as oblique-slip faults and strike-slip faults began to form in two main directions^[3,4]. (2) For the reactivated pre-existing faults, with the decrease of the angle between the strike of pre-existing faults and the late extension direction, the oblique slip effect of the faults is obviously enhanced, which is gradually consistent with the structural style of strike-slip faults^[6,17]. (3) The vertical section of the experiments show that the inherited pre-existing faults as the main branches often intersect with the nearby newly-formed faults, forming flower-like structures or multiple "Y" structures^[5,17]_.

According to the actual faults in the study area, an experiment device was designed (Fig. 8) to simulate the structural development of Bodong Sag and the southern Bodong Low Uplift. The experiment was done in Physical Modelling Laboratory of China University of Petroleum (East China). The base of the apparatus was set up as a simplified Paleogene pre-existing structure of the study area: the foam and the soft rubber were used to simulate the rigid uplift and the plastic sag respectively, and the boundary between the foam and the rubber simulated the pre-existing nearly-NS strike slip fault in Paleogene; in addition, a foam strip was laid to simulate the NE-trending pre-existing fault generated by the strike slip derivation. In order to reveal the opening and closing effect of fractures clearly, wet clay, like loose quartz sand, used in a large number of scaled modelling experiments^[38,45], was used in this experiment. The wet clay of 3cm thick (about 3 km of stratum deep) was laid on the basement, the fault walls were stretched at a uniform speed of 10 cm/h in NS direction by the driving motor. The experiment was repeated many times to ensure its repeatability. The experimental basement settings and the three stages of the experimental process are shown in Fig. 8a and Fig. 8b-8d, respectively.

The final result of the experiment (Fig. 8d) is consistent with the actual structures of Neogene in southern Bodong Low Uplift and Bodong Sag (Fig. 3a): In the "sag" with plastic formation, newly-formed faults were larger in number and nearly perpendicular to the direction of extension; while in the "uplift" with rigid basement, only little deformation occurred and few faults came about; the pre-existing fault simulated by the foam bar was partially reactivated, forming "inherited fault" which combined with the newly-formed fault at the two ends, into a curved fault in S-shape. In comparison, the nearly NS-trending boundary between the foam and soft rubber reactivated later than the NE-trending foam bar. The modelling is strong evidence that the fault system of the study area in Neogene is the result of oblique extension of pre-existing faults.

The two viewpoints about the Neogene tectonic background of Bohai Sea (sea area of Bohai Bay Basin) are the transpressional dextral strike-slip^[28,29,30] resulted from Pacific plate activity^[46,47] and nearly-NS extension due to the back-arc spreading of the Japanese Sea or the trench retreating of the subduction of Pacific plate^{[31,32,33,34]}. Although this paper does not focus on the tectonic mechanism or dynamic source in Neogene, the static and dynamic characteristics of structures and the modelling experiment results presented in this paper support the latter viewpoint. It is precisely because of the persistent NS extension during Neogene, some of pre-existing faults generated in the Late Mesozoic-Paleogene were reactivated and subjected to oblique extension.

Morley et al.^[10,12] have discussed the difference of oblique extension, pure strike-slip and orthogonal extension. Similar fault structural styles, such as Z-shaped type, en-echelon type, step-over type and curved shaped type, can be formed in these deformation system, so kinematic mechanisms can’t be distinguished according structural types. The origin of structures can be determined by comparing other related structural styles. In this paper, the differences between oblique extension and transpressional strike-slip in Neogene easily confused in Bohai Sea in recent years are discussed.

There are essential differences between dextral transpressional strike-slip and oblique extension. From the perspective of macro-deformation mechanism, the former is from simple shear (non-coaxial) deformation, while the latter is mainly from pure shear (coaxial) deformation, though different extension angles and stress disturbance can also produce local simple shear deformation^{[9, 12]}. The authors believe that the structural deformation of Bohai Sea in Neogene is the result of oblique extension of pre-existing faults rather than transpressional strike-slip, based on the following important features of Bohai Sea: (1) Strike-slip derivative structures such as R shear, P shear, and fold and reverse faults don’t occur in Neogene of Bohai Sea, and there are only a large number of normal faults (Fig. 2). (2) The NEE and nearly-EW faults are distributed widely in Neogene across the whole Bohai Sea area^[25,34], not just at the site where Tanlu fault zone is located, which can be formed only under the regional nearly-NS extension stress field rather than the transpressional strike-slip. (3) The large number of en-echelon normal faults are often considered as R shear at the early stage of the formation of strike-slip fault, but in fact, the angle between their strike and the strike of main fault is obviously larger than that of stan-dard R shear, and the dipping of these faults has not changed like standard R shear^[48,49,50]. They should be graben structures similar to rootless flower structures, formed by extension (Fig. 9a), rather than flower-like structure produced by transpression. (4) Traditionally, flower-like structures controlled by one main fault and overlapped by several secondary faults (F3 fault) are recognized as the products of pure strike-slip or even transpressional strike-slip movement caused by simple shear, but in fact, oblique slip of fault resulted from pure shear or even orthogonal extension can also give rise to this structural style on the section, which has been confirmed by modelling experiments of Schlische and Ma et al.^[17,51]. (5) Some "transpressional" structures in Bohai Sea, such as "back-shaped negative flower" (Fig. 9b), aren’t results of transpressional movement, because the faults are Neogene normal faults and the anticline is not synsedimentary anticline. As the angle unconformity produced by the bending of anticline is near the seabed, it can be inferred that such anticline must be formed very late and even could be the product of nearly EW-trending compression since Quaternary, which obviously doesn’t match with the active period of Neogene faults^[34,52] .

In summary, the widely distributed “en-echelon” faults and curved faults in Neogene of Bohai sea area and even in Bohai Bay Basin are not the result of transpressional strike-slip, but the result of oblique extension of pre-existing structures in late period.

The control of oblique extension on hydrocarbon accumulation in Neogene is mainly manifested in three aspects: (1) The inherited faults served as migration paths connecting the reservoir and the source rock: the NE-trending inherited faults (such as F1, F2 and F3) cutting through the Cenozoic join with the secondary newly-formed faults to connect the Neogene traps with Paleogene source rock in Bodong or Bozhong sag, providing migration channels for deep oil and gas to shallow layers. This is proved by the fact that the oil and gas discoveries in Neogene are concentrated near the major inherited faults. In contrast, NE-trending newly-formed faults or NNE to NS-trending extinct faults, cutting through only shallow or deep layers, don’t contribute to the hydrocarbon accumulation in shallow formation, so hardly any large-scale oil and gas reservoirs have been found near these faults. (2) Oblique extension of pre-existing faults in Neogene resulted in the sedimentary cover deformations such as dragging, reverse dragging, tilting, and warping etc, and the newly-formed faults can overlap or connect with inherited boundary faults in complex patterns, giving rise to a lot of traps including fault noses, fault anticlines and fault blocks in Neogene which can provide sites for hydrocarbon preservation. Previous petroleum exploration also concentrated on these small but numerous structural traps. (3) The Neogene faults formed under oblique extension have different relative components of strike-slip and extension, thus having different migration and sealing capacity to oil and gas. Previous studies on the interior structure, geometry and geostress ^[53,54,55] of the fault zones and the exploration practice in Bohai Sea^[56,57,58] all demonstrated that at the area of geostress release, extensional fault structures tend to be open, conducive to the migration of oil and gas but not to the hydrocarbon sealing; in contrast, pressure-increasing zone induced by strike-slip movement is the geostress increasing area in fault zone, where the rock is ground and compacted, resulting in better sealing ability and poorer migration ability of oil and gas than extensional faults. Similar to this, oblique extension leads to differences in strike-slip and extension components of faults of different strikes: faults with stronger strike-slip component have better sealing ability than those with stronger extension component, while faults with stronger extension component have better migration effect than those with stronger strike-slip component. The modeling experiment result (Fig. 7) also shows the faults with larger strike-slip component have more closed fault planes, while faults with large extensional component are mostly open.

The planar S-shaped fault system in Longkou 7-6 oilfield area in the southern end of Bodong Low Uplift (Fig. 10) is a typical result of oblique extension in Neogene. This fault system is taken as an example to demonstrate the control effect of oblique extension on oil accumulation. In this oilfield area, the inherited NE trending faults and the NEE newly-formed faults on both sides grow and connect into a S-shape fault system on the plane. Oil and gas are mainly distributed near the inherited faults or newly formed faults connecting with the inherited faults, and the exploration activities so far are concentrated in fault block and fault nose traps in Neogene. By 2018, 10 exploration wells have been drilled in Neogene traps of this area, but the exploration results in different parts differ greatly: under similar reservoir and trap conditions, for wells less than 5 km apart, some have oil layers of more than 70 m thick, others only 2-3 m thick. Examining the drilling results shows that the difference of fault sealing ability determined by oblique extension is an important factor affecting petroleum accumulation. Take Well A and Well J as examples: Well A has more than 70 m thick oil layer interpreted in Guantao Formation, this is because this well is near the NEE-trending part of the curved fault with large extension component and strong migration ability, and oil and gas could migrate to the NNE-trending part with strong strike-slip component and good sealing ability along the fault plane or structural ridges and accumulated in the trap. In contrast, the migration and sealing faults of Well J to the north are newly-formed NEE-trending faults with strong extension component, which allow strong migration of oil and gas but have much poorer sealing ability, so the oil and gas filling in this site is prone to leak. In this well, mud logging had oil and gas shows in over 100 meters, but only over 10 meters of oil layers in Guantao Formation were interpreted from logging, and many oil-bearing water layers and oil-water layers were interpreted, which is obviously caused by the leakage after oil and gas charging. In addition, on the basis of the analysis above, the angles between the strike of migration and sealing faults and the late extension directions of other eight exploration wells in the oilfield were also counted and the “oblique extension accumulation factor” (F) is proposed to characterize the migration and sealing ability of faults:

where α—the angle between the strike of migration fault and the extension direction, (°); β—the angle between the strike of sealing fault and the extension direction, (°).

When the value F of the traps drilled by exploration wells are compared with the thickness of oil layers interpreted by logging, it can be found that F value is positively correlated with the thickness of oil layer (Fig. 11). The larger the factor F is, the better the matching effect of migration and sealing is, and the thicker the oil layer is. This proves the applicability of this model in evaluating reservoirs in oblique extension areas.

It should be noted that the migration and sealing ability of faults are influenced by many factors, such as the interior structure of fault zone, the lithology of fault rock and the physical properties of host rock, so it is not rigorous to judge the sealing and migration ability only from stress controlled by the faults properties. The authors also found from statistics that “oblique extension accumulation factor” between different regions could not be compared horizontally on a large scale. For example, the F factor of area A is larger than that of area B far from A, but it does not mean that the oil layer of area A is thicker than that of area B, because the migration and sealing of hydrocarbon are influenced by many other factors. But for small areas with uniform lithology, similar late fault activity but huge difference in fault strikes, the effect of stress controlled by fault properties on hydrocarbon accumulation will be more remarkable, and this method of predicting favorable belt for petroleum is more suitable.

The reactivation of pre-existing faults under oblique extension in Neogene in the study area, even in Bohai Bay Basin, led to the formation of many fault blocks, fault noses and curved faults. The migration and sealing capacity of a large number of faults in the petroleum exploration still need to be looked further. In this paper, the stress difference induced by the angle between fault strike and extensional direction is presented to analyze this problem. The research results can guide the petroleum exploration in other areas with similar tectonic background.

The extensional rifting and strike-slip of Tanlu fault zone of Bohai sea area in Paleogene resulted in the formation of a large number of faults of various strikes, and the tectonic stress direction changed in Neogene, providing conditions for oblique extension of Paleogene pre-existing faults.

The fault systems can be divided into inherited type, extinct type and newly-formed type. The inherited faults on the boundary of uplifts are characterized by strike-slip-extension or extension structural style, while the late newly-formed faults in the sag show extension structural style, and mostly appear in en-echelon formation, conforming to the result of oblique extension of the pre-existing faults.

The fault systems in Bodong area in Neogene were the result of oblique extension in nearly-NS direction, rather than transpressional strike-slip caused by simple shear.

In the process of late oblique extension, the pre-existing faults of different strikes had different relative components of strike-slip and extension. The faults with stronger strike-slip component have better sealing ability to hydrocarbon, while the faults with stronger extension component provide better migration conditions for oil and gas, so the oblique extension determined the richness of petroleum in traps.