Since the 1990s, under the guidance of the theories of late hydrocarbon accumulation, shallow water delta and oil and gas transfer bins, a number of large and medium oilfields such as Penglai 19-3 and Qinhuangdao 32-6 have been discovered in the shallow layers of the uplift in the Bohai Sea Area^{[1,2,3,4,5,6,7,8,9,10]}. As further exploration is carried out in the shallow layers of the uplift increases, the exploration target begins to shift to the sag area with relatively few exploration wells^{[1,2,3,4,5,6,7,8,9,10]}. Due to the neotectonic movement, a series of complex fault zones were formed in the sag area^[1,2,3,4,5]. The Neogene in the complex fault zones has shallow reservoir and favorable caprock conditions, a large number of traps and vertical migration pathways, and some oilfields with oil reserves more than hundreds of million tons, such as Bozhong 8-4 and Lüda 21-1 already discovered, proving the huge exploration potential of the complex fault zones. However, there are lots of failed wildcat wells targeting the shallow layers in the complex fault zones, with a low success rate of exploration, suggesting great difference in hydrocarbon accumulations. How to sort out the enrichment zones out of numerous fault blocks in the sag area is the major challenge facing the exploration^{[8,9,10,11,12]}.

In the complex fault zones, the hydrocarbon migration from the source rock to the shallow trap is a vertical migration process, including the lateral migration in or near deep source rock through the transport layer, vertical migration of oil through the fault, and lateral migration through the shallow transport layer outside the source rock^{[11,12,13,14]}. Although there are numerous faults in the shallow layers in the complex fault zone, affected by the activity and geometry of fault and destructiveness of regional caprock, only 10%-15% of the fracture surfaces in the shallow faults can act as the dominant pathway for the vertical hydrocarbon migration, thus how to accurately describe the law of vertical hydrocarbon migration has become the key issue in exploration study^{[15,16,17,18,19,20,21,22]}.

Researchers have summarized a series of shallow hydrocarbon accumulation models of the Bohai Sea, including the meshwork-carpet type migration, the hydrocarbon transfer bins, and the penetrative migration, etc.^[23,24,25]. These previous studies focused on the description of the accumulation process and mechanism, and the studies on vertical migration mainly attached importance to analysis of migration faults, rather than looking into the vertical migration pathways^{[15-22, 26-29]}.

Numerical simulation is an important hydrocarbon migration analysis method^{[13, 30-34]}. At present, there are three numerical simulation methods for hydrocarbon migration. First, simulation of hydrocarbon migration along layer inside the source rock, it simulates the migration process of oil and gas generated by source rock moving along beds to the trap. It is suitable for simulating hydrocarbon migration inside the deep source rock layer in the complex fault zones, but not for shallow hydrocarbon migration outside the source rock. Second, the simulation of migration in a uniform flooding pattern, it simulates the hydrocarbon migration pathways under even hydrocarbon expulsion intensity and can be used to predict the direction of hydrocarbon migration out of source rocks. But it doesn’t take into account the differences of faults in vertical migration capacity, thus its analysis results has limited predictive value. Third, the hydrocarbon migration simulation based on charging point, it is mainly used to tell the rough direction of oil and gas charging and migration, and is suitable for the hydrocarbon accumulation out of source rocks. However, as there are plenty of faults in the complex fault zones, it is difficult to identify the charging point, thus this method is seldom used^{[13, 32-34]}. Thus, it can be seen that there is no effective numerical simulation method for the simulation of hydrocarbon migration in the complex fault zones.

Nowadays, we have entered into an era of big data^[33,34], which is essentially a data-driven scientific research which eliminates uncertainties with information^{[33,34,35,36,37]}. Massive data has been accumulated over the exploration of China’s major oilfields. Researchers attach increasing importance to how to extract quantitative parameters related to the fault activity and hydrocarbon migration from the qualitative-semi-quantitative geological analysis, how to serve the fine exploration of the complex fault zone with the exploration big data, how to improve the success rate of exploration with fine three-dimensional structural modeling and numerical simulation of hydrocarbon migration^{[33,34,35,36]}.

According to the big data analysis of hydrocarbon geological conditions in the south central area of the Bohai Sea, the quantitative-semi-quantitative analysis is conducted on the formation conditions of the vertical dominant migration pathway in the sag area. On this basis, through building structural model to support analysis, the big data search of vertical dominant migration pathways and charging points/segments outside the source rock in the East Subsag of Huanghekou Sag is realized. Then, the charging and migration of oil and gas is simulated numerically, thus successfully predicting the hydrocarbon migration pathways to the shallow layers and hydrocarbon accumulation zones in the complex fault zones.

The study area is located in the south central area of the Bohai Sea, including the Bozhong Sag and surrounding areas and the East Subsag of Huanghekou Sag, with an area of 1.2×10⁴ km². The Shijiutuo Bulge, Shaleitian Bulge Chengbei Low Bulge, Bonan Low Bulge and Miaoxibei Bulge in the study area have a high degree of exploration, and the oil-bearing series are mainly Neogene, with reserves exceeding 20×10⁸ t (Fig. 1a). The study area has active neotectonic movement and fault activities. The Tan-Lu fault zone passes through the eastern part of the Bozhong Sag and the East Subsag of Huanghekou Sag, forming a series of complex fault zones and beaded traps^{[1,2,3,4,5,6,7,8,9,10]}, of which the complex fault zones in the sag area have less wells drilled and a low degree of exploration (Fig. 1b), becoming a realistic area for further exploration^{[3, 6-10]}.

The deep Paleogene tectonic layer (hereinafter referred to as the deep layer) and the shallow Neogene tectonic layer (hereinafter referred to as the shallow layer) in the study area. The deep layer is composed of cyclic deposits including sand bodies of the multistage fan delta-braided river delta and the lacustrine mudstone in the rift stage, with Kongdian Formation, Shahejie Formation and Dongying Formation developed from deep to shallow. The shallow layer is made up of sandy conglomerate, sandstone and mudstone of fluvial facies and shallow delta facies in the depression stage^[37,38,39], with Guantao Formation and Minghuazhen Formation developed from deep to shallow^[37,38,39] (Fig. 1c).

Controlled by the large tectonic subsidence of the late stage of rifting^[37,38,39], the mudstone caprocks in the Dongying Formation are developed in the sag area, with a thickness of 300-3000 m. Due to different contact relationship with effective source rock and the separation of the regional caprock, the hydrocarbon accumulation is divided into the deep in- source rock accumulation system and the shallow out-of- source rock accumulation system vertically (Fig. 2). As 70% of the reserves in the Bohai Sea area are developed in the shallow accumulation system^{[1,2,3,4,5,6,7,8,9,10]}, the hydrocarbon charging and migration of the shallow out-of-source rock accumulation system in the complex fault zones is mainly discussed herein.

As it is shallow in depth and has no mature source rocks, the shallow accumulation system is of the out-of-source rock accumulation type^{[11-12,23-25]}. During the neotectonic movement, the caprock of the deep hydrocarbon accumulation and the underlying transfer bins were broken, hydrocarbons migrated along vertical faults to the shallow layers, some of the hydrocarbons accumulated in the fault block traps near the charging points, other hydrocarbons moved along the main transport layers of the shallow layers (the upper member of Guantao Formation) to the higher fault blocks, forming the cascade interception and hydrocarbon accumulation pattern (Fig. 2)^{[11-12,23-25]}. This shallow out-of-source rock accumulation process is called out-of-source rock accumulation model herein.

In the out-of-source rock accumulation model, the vertical dominant migration pathway is defined herein as the narrow fault surface allowing hydrocarbons to flow from the deep in-source rock accumulation system to the shallow accumulation system in large scale. The pathway is the bridge for hydrocarbons moving from deep layers to shallow layers and the key factor for hydrocarbon accumulation in shallow layers. Only by accurately predicting the vertical dominant pathways, can the hydrocarbon migration in shallow layers be better predicted. Due to changes of fault throw, fracture surface shape, and overlapping relationship with the in-source rock accumulation system and the breaking degree of regional caprock along the faults, the vertical hydrocarbon migration capacities of faults vary greatly. Therefore, in-depth study on the distribution of vertical dominant migration pathways is required^{[15-22, 26-29]}.

The formation conditions of vertical dominant migration pathways were figured out through big data statistical analysis of the shallow hydrocarbon accumulation conditions of about 62 oil and gas bearing structures/oil-gas fields in the sea area of the Bozhong Sag and surrounding areas.

The fault connecting the deep and shallow accumulation systems is the pre-condition for forming vertical dominant migration pathway and large scale out-of-source rock accumulation, and the hydrocarbon accumulation capacity inside the source rock connected by the fault determines the hydrocarbon charging and accumulation scale out of source rocks^{[13-14,18-22]}. The overlapping relationship between the fault cutting through the deep and shallow layers (hereinafter referred to as the through-going fault) with the in-source rock accumulation system determines its capacity of hydrocarbon accumulation to a large extent as well as the formation of vertical dominant migration pathway^[24]. The through-going faults are divided into three types according to their connection mode with the in-source rock accumulation system, including Type I which cuts deep transfer bins, Type II which cuts main hydrocarbon migration pathways, and Type III which cuts the lower part of the deep accumulation system.

Type I through-going fault accounts for 68% of the shallow reservoirs in the sag area, with the reserves abundance up to 1200×10⁴ t/km². The larger the scale, the more transfer bins and the larger the scale of the shallow hydrocarbon accumulation will be.

Type II through-going fault does not cut transfer bins, but overlaps with the main hydrocarbon migration pathways at the lower part. It accounts for 22% of the shallow reservoirs in the sag area, with the average reserves abundance of 320×10⁴ t/km².

Type III through-going fault does not cut the deep transfer bins and the main hydrocarbon migration pathway, but cuts the lower part of the source rock of the deep accumulation system. It is mostly inactive, accounting for 10% of the shallow reservoirs in the sag area, with the average reserves abundance of only 20×10⁴ t/km².

By comparing the shallow out-of-source rock accumulation probability and reserves abundance of the above three types of through-going faults, we can see that Type I and Type II have higher probability of shallow hydrocarbon accumulation and reserves abundance, proving that they allow active hydrocarbon charging towards the shallow layers, and thus they have stronger deep accumulation capacities and is more likely to form vertical dominant migration pathways and shallow reservoirs. In contrast, Type III has lower probability of shallow hydrocarbon accumulation and reserves abundance, proving that they cannot efficiently transport hydrocarbons to the shallow layers, and thus has poorer deep accumulation capacities and is less likely to form vertical dominant migration pathways and shallow reservoir.

Usually, the fault has rugged fracture surfaces. In the area with significant changes of fracture surface curvature, the stress is centrally released, and micro-cracks is easily developed, creating favorable conditions for hydrocarbon migration to the shallow layers^{[15-16, 18, 28-29]}. Under buoyancy, hydrocarbons migrate mainly along the convex surface^{[16, 22, 26-29]}. As long as there is fault in contact with the source rock, the hydrocarbons will not spread evenly around, but mainly move upward along the convex surface of the fault surface. The statistical results of oilfields/oil and gas bearing structures in the Bozhong Sag and surrounding areas indicate that the accumulation probability along the convex surface of fault is 20% higher than that along the concave surface. Therefore, the fracture surface structural ridge is more likely to form the vertical dominant migration pathway.

The thick mudstone at the Lower Member of Dongying Formation is the caprock of the deep accumulation system, which was destroyed by the neotectonic movement and its fault activity, making it possible for hydrocarbons to migrate towards the shallow accumulation system^{[13-14, 21-24]}. Therefore, the rupture of the regional caprock is a necessary condition for the formation of the vertical dominant migration pathway.

The thickness of the caprock minus the throw of the through-going fault is the juxtaposition thickness between fault and caprock. The juxtaposition thickness between fault and caprock can be used to quantitatively evaluate the destructive effect of the fault activity on the caprock. The smaller the juxtaposition thickness between fault and caprock, the greater the destructiveness of the deep accumulation system is^{[3, 13, 40-42]}. The relationship between the juxtaposition thickness between fault and caprock and the ratio of the shallow reservoirs in the Bozhong Sag and surrounding areas indicates (Fig. 3) that when the juxtaposition thickness between fault and caprock is more than 400 m, the deep accumulation is dominant; when the juxtaposition thickness between fault and caprock is less than 400 m, the shallow accumulation is dominant. The smaller the juxtaposition thickness between fault and caprock, more likely the shallow accumulation is to form. Therefore, the juxtaposition thickness between fault and caprock of less than 400 m is necessary for the formation of vertical dominant migration pathway.

When the fault activity reaches a certain level, the sliding fracture zone and induced fracture zone will be generated, which can act as the high-porosity and permeable pathway for hydrocarbon migration and charging to the shallow layer^{[3, 11-12, 15, 20-24, 40-41]}. As the reservoirs in the study area were formed in the late stage, the throw of the fault in the neotectonic movement period is highly important^{[1,2,3,4,5,6,7,8,9,10]}. The statistical results of the relationship between the fault throw and the ratio of shallow geological reserves ratio during the accumulation period show that the greater the fault throw during the accumulation period, the stronger the hydrocarbon migration capacity of the fault would be. When the fault throw is less than 80 m, the deep accumulation is dominant; when the fault throw is greater than 80 m, the shallow accumulation is dominant. Therefore, the minimum fault throw for the formation of the vertical dominant migration pathway during the accumulation period in the sag area is 80 m (Fig. 4).

To sum up, among Type I and II through-going faults cutting into transfer bins of deep deposits of hydrocarbons and main migration pathways, the juxtaposition thickness between deep fault and caprock is less than 400 m, and the vertical dominant migration pathway in the complex fault zones could be easily formed in the fracture surface with fault throw greater than 80 m during the shallow hydrocarbon accumulation period.

The East Subsag of Huanghekou Sag is in the southeast of the Bozhong Sag and surrounding areas. With numerous late-stage faults and faulted blocks, it is a complex fault zone. Early drilling proves that the shallow Neogene is the main accumulation system. However, lots of wells failed and the exploration results were not satisfactory in the area^{[1,2,3,4,5,6,7,8,9,10]}. Unclear shallow dominant hydrocarbon migration pathways hinder the hydrocarbon exploration in shallow layers of the area.

Therefore, based on the out-of-source rock shallow accumulation model and the formation conditions of vertical dominant migration pathways in the Bozhong Sag and surrounding areas, the authors put forward the simulation method for hydrocarbon migration based on analysis of vertical dominant migration pathways creatively to predict the migration pathways of oil and gas out of source rocks in the East Subsag of Huanghekou Sag.

The method mainly consists of 3 steps: (1) to identify charging points/sections out of source rocks on the basis of structural modeling; (2) to analyze interception of shallow faults; (3) to simulate hydrocarbon migration in shallow layers under the constraint of dominant charging points/sections.

In the out-of-source rock charging model of complex fault zone, the place where the vertical dominant migration pathway overlaps with the shallow transport layer is the out-of-source rock charging point/section (Fig. 2). Based on the formation conditions of vertical dominant migration path, the identification criteria of charging point/section have been established: (1) The breakpoint shall be the one where Type I or II through-going fault cutting into transfer bins of deep deposits of hydrocarbons or main migration pathway overlaps with the main shallow transport layers. (2) The breakpoint has a throw of over 80 m and juxtaposition thickness between fault and caprock of less than 400 m during hydrocarbon accumulation. (3) The breakpoint at the structural ridge of fracture surface is more likely to be the charging point.

The charging points/sections in the East Subsag of Huanghekou Sag were searched according to the above identification criteria. The authors attempted to find the charging points/sections by calculating the juxtaposition thickness between fault and caprock and the fault throw of the breakpoint at different positions along the fault strike. If all the faults in the study area were sampled and analyzed at equal intervals, even if the sampling interval was 1 point/km, and only two horizons were calculated, there would be more than 1000 calculation points. If the analysis accuracy, the analyzed horizon and the analyzed items were further improved, the computational analysis would be multiplied in workload, time-consuming, labor-intensive and less efficient.

In order to improve the resolution and working efficiency of the vertical migration pathway analysis, the full-fault big data structural modeling and its fault attribute analysis were introduced to calculate the fault throw and the juxtaposition thickness between fault and caprock of the faults. The calculation accuracy reached 100 points/km, and the calculation resolution increased by 100 times. The “big data sampling analysis calculation” of more than 350 000 breakpoints in 6 series of strata under the 3D structural modeling was completed, realizing the high-precision search of the out-of-source rock charging points/ sections of the vertical dominant migration pathways.

The specific steps are shown as follows: (1) Carry out the layer modeling using the Petrel software, simulate hydrocarbon migration inside source rocks along the formation to obtain the transfer bins, main hydrocarbon migration pathways and scale of gathered oil and gas in the deep layer (Fig. 5a). (2) Analyze the overlapping relationship between the faults and the deep hydrocarbon accumulation and main hydrocarbon migration pathways, to identify the type of the through-going faults (Fig. 5b). (3) Based on the identification of the oil source fault, calculate the fault throw of the through-going faults, regional caprock thickness and juxtaposition thickness between the through-going faults and caprock in the deep hydrocarbon system by constructing a 3D model, and mark the sections with juxtaposition thickness between the fault and deep caprock of less than 400 m and the fault throw of more than 80 m during the accumulation period of the shallow layers as the dominant charging points/sections (Fig. 5c, 5d). (4) Analyze the fault throw of the shallow layers during the accumulation period by constructing a 3D model, and demarcate the fracture surface with the fault throw greater than 80 m in the potential charging point/section as the dominant charging point/section (Fig. 5d). (5) Analyze the fault morphology by marginal detection, and demarcate the convex surface in the dominant charging points/sections as the strong dominant charging points/sections (Fig. 5e, 5f).

By using the above method, the distribution characteristics of potential charging points/sections, dominant charging points/sections and strong charging points/sections in the East Subsag of Huanghekou Sag were obtained. The dominant and strong charging points/sections account for only 16% of the shallow faults (Fig. 5f), and of the through-going faults, only 25% of the fracture surfaces are charging points/sections (Fig. 5f). It could be seen that although there are many shallow faults in the study area, there are only a few dominant and strong charging points/sections. Moreover, the dominant and strong charging points/ sections are all distributed in the area with active neotectonic movement and late-stage fault activities, indicating the neotectonic movement had strong control on shallow hydrocarbon charging.

After charging at the shallow dominant/strong charging point/section, oil and gas would migrate along the structural ridge of the shallow transport layer^[26,27]. Along the migration pathway, the fault cuts the stratum, the fault with smeared gouge on the fracture surface would break the lateral continuity and connectivity of the transport layer, forming lateral sealing, and the oil and gas would accumulate there after interception^[17,18,19].

Faults in the uplift area mainly play the role of interception, so they can best reflect the coupling relationship between the fault activity and hydrocarbon interception and accumulation. By examining the relationship between fault throw in the accumulation period and oil column thickness in the Bozhong Sag and surrounding areas through statistics, it is found when the fault throw of the uplift area during the accumulation period is more than 15 m, oil and gas can accumulate due to fault interception, and the oil column thickness is 20-60 m (Fig. 6). According to the analysis results of the fault throw of faults in the East Subsag of Huanghekou Sag and the shale smear factor, when the fault throw is greater than 13 m, the shale smear factor of the fault will be ground to 30%, i.e. the fault has certain intercepting capacity^[17,18,19] (Fig. 7), indicating that when the fault throw is about 15 m, the fault can have effective interception.

On the basis of ascertaining the conditions for forming shallow dominant migration pathway and fault interception, the hydrocarbon migration simulation under the constraint of the charging points/sections was carried out for the shallow layer of the East Subsag of Huanghekou Sag. The specific steps are as follows: (1) As the main transport layer in the shallow layer is the upper member of Guantao Formation underlying the mud-rich section of the Lower Member of Minghuazhen Formation, the base of Minghuazhen Formation was selected as the control layer for the routine structural modeling and reservoir modeling. (2) The fault model was built, and all faults with the fault throw of more than 15 m were supposed to have certain intercepting capacity (equivalent to the height of 60 m oil column). (3) The Trinity software was used to carry out the sequential charging and hydrocarbon migration simulation for the strong dominant charging points/sections previously confirmed. The amount of hydrocarbon at each charging point/ section referred to the quantitative scale of deep hydrocarbon accumulation in the vertical dominant migration pathways where the charging point was located.

The simulation results show that (Fig. 5f) the Bozhong 36-A and Penglai structure zones have more charging points/sections and larger hydrocarbon accumulation scale (Fig. 5f); while Penglai 31-B has few charging points/ sections and no large scale hydrocarbon accumulation. The actual drilling shows that Bozhong 36-A has good oil and gas discoveries, while Penglai 31-B has no major discoveries, proving the reliability of the simulation method (Fig. 1).

Meanwhile, the simulation also shows (Fig. 5f) that: (1) The hydrocarbon migration and accumulation mainly happened near the dominant and strong charging points/sections, and after getting into the shallow layers along dominant/strong charging points/sections, the hydrocarbon-bearing fluid mainly accumulates near them as the interception and blocking of the faults in the complex fault zone, e.g. Bozhong 36-A and Penglai 31-B structural zones. (2) The fault blocks with more dominant/ strong charging points/ sections have better hydrocarbon accumulation effect, while those with less have poorer hydrocarbon accumulation effect, e.g. Block Penglai 31-B has better oil and accumulation effect than Penglai 31-A.

In addition, in the areas with stronger neotectonic movement and fault activity (Fig. 5d), charging points/ sections are more likely to occur, and the hydrocarbon migration and accumulation are more active (Fig. 5f). In the areas with weaker neotectonic movement and fault activity, charging points/ sections are less likely to occur, and the hydrocarbon migration and accumulation are weaker (Fig. 5f). Clearly, by controlling the uneven faults cutting into the deep formation, the neotectonic movement and fault activity led to the differential distribution of vertical dominant migration pathways and strong charging points/sections out of source rocks, and determined the differential hydrocarbon accumulation in the shallow layers of complex fault zones.

The hydrocarbon migration simulation method based on the analysis of the out-of-source rock charging points/sections of the vertical dominant migration pathways worked well, the coincidence rate between the hydrocarbon simulation results of the shallow layers and the exploration results of the study area is 90%. The method was used in the exploration evaluation of the East Subsag of Huanghekou Sag, confirming nearly 100 million tons of reserves and showing good prediction effect.

In the complex fault zones around Bozhong subbasin, among the faults overlapping with the oil and gas transfer bins or main migration pathways, the faults having juxtaposition thickness with the deep caprock of less than 400 m and fault throw of more than 80 m during the accumulation period of the shallow layers are likely to be dominant migration pathways and charging points/sections out of source rocks.

Different from hydrocarbon accumulation model inside source rock in the deep formation, in which the starting point of hydrocarbon migration is source rock, in the hydrocarbon accumulation model out of source rocks in the shallow layers of complex fault zones, the starting point of hydrocarbon migration is the place where the vertical dominant migration pathway overlaps with the shallow transport layer, namely, the charging point/section.

Since the hydrocarbon migration simulation of the shallow out-of-source rock accumulation system based on the vertical dominant migration pathways takes into account the control effect of the fault activity, regional caprock, deep formation accumulation capacity, features of the shallow transport layer and fault interception on hydrocarbon migration and accumulation in the shallow layers, it obtained good prediction results.

The simulation of hydrocarbon migration in the East Subsag of Huanghekou Sag shows that the neotectonic movement and associated fault activities determine the hydrocarbon accumulation in the shallow layers of the complex fault zones by controlling vertical dominant migration pathways and strong charging points/sections.