Integrity of fault and top seal is a key factor that affects hydrocarbon fluid phase distribution in the subsurface. Mapping fluid property distribution can therefore provide important tools towards assessing seal integrity and trapping mechanisms – two of the most critical elements in petroleum systems analysis towards optimzing exploration and development strategies. This is best achieved by proper fluid characterization that integrates fluid geochemistry and pressure–volume–temperature (PVT) data to identify equilibrated and disequilibrated fluid gradients. Fields from Paleozoic and Mesozoic basins in the Arabian Peninsula at different stages of delineation, appraisal, development and management are discussed as examples to demonstrate the role of fluid characterization in aiding the evaluation of top, lateral and fault seal integrity, and in providing insights into the sealing and buffering effects of reservoir heterogeneity on trapping mechanism, fluid distribution and flow capacity. Examples discussed include (1) reservoir heterogeneity controlling fluid distribution and trapping mechanisms in two neighbouring gas fields, (2) geochemical evidence for a lateral seal separating gas condensate region from oil discovered during field development, (3) solid reservoir bitumen atop a giant gas field, (4) selective gas leakage through anhydrite seal, and (5) geochemical evidence for fault-controlled reservoir compartmentalization rather than a hydrodynamically tilted fluid contact in a field at an appraisal stage where PVT data are limited or inconclusive. Put in structural context, seal-related interpretations of the fluid data are aligned with observations from seismic and geological data on the existence of faults with favourable orientations or facies changes.

Thematic collection: This article is part of the Fault and top seals 2022 collection available at: https://www.lyellcollection.org/topic/collections/fault-and-top-seals-2022

Petroleum can be trapped by top and lateral seals having minimum capillary entry (or displacement) pressure greater than the buoyancy pressure of the petroleum column. Common effective seal lithologies include evaporites, organic-rich rocks, and fine-grained clastics (Brown 2003). Capacity of the seal (seal capillary displacement or entry pressure) to entrap petroleum increases with the increase in hydrocarbon-water interfacial tension, and the decrease in pore throat diameter and wettability. Buoyancy pressure of the petroleum column acts to force hydrocarbons through larger pores in the seal, and is proportional to column height and density difference between petroleum and water (Downey 1984). Lower density and higher diffusivity of gas compared to oil favour gas leakage at the crest where the drop in reservoir pressure is greatest. As indicated by Downey (1984), gas diffusion through a seal depends on gas type and the characteristics of the seal pore network; favouring, for example, methane diffusion through a porous, water-bearing shale, but not through an anhydrite seal.

Seal integrity and trap closure affect fluid phase distribution, which led Sales (1997) to define three trap styles. An efficient seal, in which seal capacity exceeds closure, is considered trap style-I that promotes fill and spill migration, where heavier fluids are spilled updip by gas. A trap that is filled to spill point but also leaks gas is classified as trap style-II, whereas a trap that leaks gas and is not filled to spill point is classified as trap style-III (closure exceeds seal capacity), leaving oil behind. Such an approach of assessing seal integrity based on observed fluid property distribution can overcome oversimplifications made by extrapolating a few mercury injection capillary pressure measurements that may not necessarily represent the entire seal rock fieldwide (Downey 1984).

In this paper, fluid phase distributions in key oil and gas fields in the Arabian Peninsula are discussed in order to assess seal integrity, dynamic fault character (leaking v. sealing) and trapping style in different geological scenarios. Fluid data used include PVT fluid composition and geochemistry, which, in addition to inferring seal integrity, can provide critical details on fluid origin and thermal maturity, migration mechanisms and various alteration processes acting to modify fluid composition within the reservoir (e.g. Blanc and Connan 1994). Case studies discussed include (1) reservoir heterogeneity controlling fluid distribution and trapping mechanisms in two neighbouring gas fields, (2) geochemical evidence for a lateral seal separating gas condensate region from oil discovered during field development, (3) solid reservoir bitumen atop a giant gas field, (4) selective gas leakage through anhydrite seal, and (5) geochemical evidence for fault-controlled reservoir compartmentalization rather than a tilted fluid contact in a field at an appraisal stage where PVT data are limited or inconclusive. Interpretations based on fluid data were backed up, where possible, by seismic observations that variably suggest the presence of faults and/or marked lateral facies heterogeneity, representing possible causes for lateral sealing and buffering, or vertical leakage across top seals.

Characterizing fluid composition could start with bulk properties measured on in situ representative samples or recombined wellstream fluid composition. Data include gas-oil ratio (GOR), saturation pressure (Psat), heptane-plus fractions, hydrocarbon and non-hydrocarbon compounds, in addition to reservoir pressure and temperature. These data are used collectively to determine fluid properties and phase behaviour (e.g. McCain 1990, 1994). Bulk fluid characterization can then be supplemented by detailed fluid geochemistry, both organic (hydrocarbons) and inorganic (aqueous). However, this publication is focused on petroleum geochemistry, integrated with PVT data, where available, and does not discuss aqueous inorganic geochemistry.

Petroleum geochemistry deals with basin-scale and field-scale compositional variations of hydrocarbon fluids to understand fluid migration and reservoir compartmentalization, among other applications (Peters and Fowler 2002). The cornerstone of petroleum geochemistry is the light hydrocarbons, biomarkers and carbon/hydrogen isotopes, which, collectively, can help infer biological sources, depositional environments, expulsion maturities, and various in-reservoir alteration processes, such as water washing, biodegradation, and thermochemical sulfate reduction (Peters et al. 2005). These data can also be used to assess fluid relationships and reservoir compartmentalization for field development and production allocation (e.g. Hwang et al. 1994). Polar compounds are best suited to assess likely migration pathways (e.g. Larter et al. 1996; Yang and Arouri 2016; Arouri and Panda 2022). Petroleum geochemistry methods used on depressurized oil and gas samples include whole-oil gas chromatography (GC) and detailed GC fingerprinting of light hydrocarbons, with special attention paid to the C7 compounds (Halpern 1995) and C8–C20 compound range, as described in Al-Meshari et al. (2009). The C7 dialkylpentane parameters are especially useful in correlating light and highly mature crudes due to their relative resistance to secondary alteration (Thompson 1983; Halpern 1995). Residual oil (between grains) and fluid inclusions entrapped within grains and cement provide additional details regarding the sequence and timing of trap filling, and offer correlation tools, especially where no PVT, drill-stem-test (DST) or production fluid samples are available. Asphaltene gradients technique (not attempted here) also provide useful details for reservoir compartmentalization assessment (Mullins et al. 2017a, b; Mullins 2020).

Several methods were applied to characterize petroleum fluids in terms of PVT and geochemical variations in order to evaluate the dynamic nature of the seals in different fields (Fig. 1). The overall approach, data and methods are elaborated here. Methodological details specific to individual examples are explained in their corresponding sections.

Key PVT data were collected from either in situ representative downhole or recombined wellstream composition samples. These include gas-oil ratio (GOR), condensate-gas ratio (CGR), saturation pressure (Psat), API gravity, heptane-plus fraction, iC4/nC4, in addition to reservoir pressure and temperature.

Source and maturity biomarkers of the saturated and aromatic hydrocarbon compounds in the liquid hydrocarbon fraction were analysed using selected ion monitoring (SIM) gas chromatography-mass spectrometry (HP 6890 GC interfaced to an HP mass selective detector MSD 5973). Several aromatic biomarkers were monitored (m/z 178, 184, 192, 198 and 231). Of these, the phenanthrene and methylphenanthrene data were used to calculate the vitrinite reflectance equivalent (VRe %), following the method of Radke and Welte (1983). The dibenzothiophene to phenanthrene (DBT/Phen) ratio is used to infer source-rock lithology of the liquids (Hughes et al. 1995). Similarly, saturated hydrocarbon biomarkers were monitored (m/z 191, 217, 218 and 259) but only two ratios reported here (viz. C29 to C30 hopanes, and diasterane to sterane) to differentiate the origin of the liquids.

Stable carbon isotope (δ13C, 13C/12C) analysis (reported in parts-per-thousand, ‰) was carried out for the C1–C5 hydrocarbons and the total saturate and aromatic liquid fractions. The δ13C is a fundamental parameter for the assessment of oil and gas genesis (thermogenic v. microbial) and thermal maturity (e.g. Bernard et al. 1976; Sofer 1984; Whiticar et al. 1986; Schoell 1988; Berner and Faber 1996). The isotope analysis was conducted using an Agilent 6890 GC equipped with a 30-m Poraplot GSQ column and a Thermo GC Combustion III system interfaced to a Delta Plus stable isotope ratio mass spectrometer (reproducibility ±0.2‰). The NBS-19 was used as a standard.

Fields A and M are two separate north–south-trending anticlinal traps 20 km apart, where the crest of Field M anticline is 122-meter deeper than that of Field A (Fig. 2). The reservoir in both fields is siliciclastics, deposited in predominantly an aeolian environment, and is capped by a tight carbonate seal. Basin modelling suggests filling of both fields from a common organic-rich shale source rock located about 30 km to the NE of Field M (Fig. 3a). The fields are being developed as gas condensate producers, but with a striking difference in fluid distribution and gas flow capacity. Field A produces gas of up to 67 000 scf/bbl, with best flow capacity located at the crest. Fluid distribution in Field M is more complex, with wet gas (maximum GOR 30 000 scf/bbl) in its downdip north, shifting to gas condensate then volatile oil towards the crest of the structure (Figs 3c and 4). Here, the best gas flow capacity is reported from deeper wells along the northeastern periphery, closer to the fetch area. These differences provided an impetus to evaluate controls over fluid distribution in both accumulations, using available PVT data.

Petrophysical variations and relation to PVT fluid composition

Porosity and permeability data indicate that the reservoir is mostly tight and of low permeability (<1 millidarcy) in Field M. A more conventional porosity-permeability relationship is observed for Field A, where permeability in some samples exceeds 100 millidarcy (Fig. 3b). Better reservoir quality in Field A is manifested petrographically by abundant diagenetic illite coats, apparently preserving porosity, compared with abundant quartz cement in Field M that reduces porosity and permeability.

The GOR-depth profiles for fluids in both fields are presented in Figure 3c, where each datapoint represents a well. Three vertically stacked trends (suggestive of three reservoir units, apparently separated by tighter sand) can be observed in each field, but with contrasting compositional gradients. The shallower Field A is characterized by normal compositional gradients, where GOR increases updip, while Field M is characterized by reversed gradients, where GOR is highest downdip, decreasing successively towards the field's south. The normal compositional trend observed in Field A is interpreted here to reflect good quality reservoir, where each reservoir unit represents a single hydrocarbon column. In conventional reservoirs, buoyancy trapping prevails, where buoyancy forces overcome the entry pressure, and hydrocarbon fluids segregate freely according to density and buoyancy; therefore, resulting in a typical buoyancy-driven fluid arrangement in an anticlinal trap, where leaner gas and best gas flow capacity occur up-section towards the crest of the structure (Fig. 3).

In Field M, the drier fluids (wet gas) occur in the deeper north of the field, transitioning to gas condensate and volatile oil towards the crestal area in the field's south (Fig 4). The reversed fluid distribution reflects greater reservoir heterogeneity, where each reservoir unit is apparently made up of several pockets of sand enclosed within tighter or less permeable sand, as represented schematically in Figure 4. This reversed stacking of oil, condensate and wet gas resembles fill-spill fluid arrangement and necessitates multiple accumulations that are petrophysically/stratigraphically compartmentalized, with no structural closure. In tight lithologies, permeability trapping prevails because entry pressures dominate over buoyancy forces, resulting in gas generally entrapped deeper (Figs 3c, 4). Non-buoyancy non-Darcy flow leads to unfocused migration, diffusing hydrocarbons pervasively over large areas without clear trap boundaries or fluid contacts (as opposed to buoyancy-driven flow that focuses fluids into the anticlinal trap, as in Field A). As fluid migrates to enclosed sand sweet spots (or to a better reservoir shallower), pressure drops, forming conventional gas-water systems; the column height in each is dependent on size and sealing integrity of each sand body (cf. Forsyth and Vargas-Guzmán 2013; Al-Duhailan et al. 2014). Accumulation of hydrocarbons in such tight lithologies, where reservoir heterogeneity and lithological/diagenetic variations (rather than anticlinal closure) control trapping, is described by Zhao et al. (2017) as quasi-continuous, as opposed to the continuous (shale gas) and discontinuous (conventional reservoir) fluid distribution. Drier gas (in the form of basin-centred gas) can be inferred in deeper traps further to the NE of Field M towards the kitchen.

Field G produces gas condensate (CGR < 170 bbl of oil per million standard cubic feet of gas, bbl/MMscf) from a Paleozoic aeolian and fluvial fine- to coarse-grained sandstone reservoir. The field sits on a gentle north–south fault-propagation anticline plunging to the south, and is bounded to the north by a major strike-slip fault striking east–west (Figs 5, 6a). The underlying north–south fault does not vertically offset the reservoir, although seismic attributes show that it could be affected by linear discontinuities that have an overall east–west en-echelon geometry (Arouri et al. 2010). The reservoir is capped by tight carbonate and shale units, and is believed to be charged from a widespread, proven and prolific marine shale source rock located tens of kilometres to the SE. During field development, black oil (400–700 bbl/MMscf) was discovered downdip in the southern region of the field. A detailed geochemical and basin modelling study was conducted and integrated with PVT data to examine whether the newly discovered oil represents an oil rim or a separate fluid system.

PVT fluid composition data

Given the small differences in reservoir pressure and the systematic shift in saturation pressure (Fig. 6b), an oil rimmed gas accumulation was initially proposed, which may also be justified considering the GOR-depth relationships (Fig. 3c). However, the fact that the GOR and Psat are both increasing for the gas condensate suggests distinct accumulations from separate charges of variable maturities that have not mixed (Fig. 6c: scenario 3). The oil is clearly not a fill-spill product (scenario 1) because it is deeper than the gas condensate. Likewise, the gas condensate is not capping the oil (scenario 2) because the Psat is not decreasing with the increase in GOR, but rather increasing.

Geochemistry results

To verify the multiple-charge scenario and lack of mixing inferred from PVT fluid composition data, fluid maturity for the liquid and gas components was measured using molecular and stable carbon isotope (δ13C) geochemistry. Carbon isotope fingerprinting of the C1–C5 gas range shows two clusters of different maturities for the oils and gas condensates, which is matched by a wide difference in the vitrinite reflectance equivalent (VRe%) measurements (Fig. 6d). These differences suggest that the oils and gas condensates were derived from charges of variable maturities, and that fluids have not mixed within the reservoir. Both sets of geochemical data agree with GOR-Psat relationships, confirming separate fluid systems that are not in communication.

Fluid inclusion analysis

Fluid inclusions (micro-sized fluids entrapped in cement and healed fractures during migration and entrapment) were analysed to constrain the time of charge in both regions. Homogenization temperatures of fluid inclusions, extrapolated to burial-thermal history, indicate that the oil and gas condensate represent separate chambers filled at different times. The oil started filling the southeastern part of the field from a different fetch area 20 to 30 million years after the gas condensate filled the northern region (Fig. 6e). Furthermore, the oil phase envelopes overlap with those of the gas condensate (Arouri and Herrera 2023), a testimony to multiple charging and the separation of the two fluid regions.

Possible causes for lateral sealing

Integration of PVT and geochemistry data confirms the compartmentalization of oil and gas condensate in separate fluid systems that need to be developed separately. No clear fault(s) with dip displacement or petrophysical barriers are readily identifiable between the northern and southern regions of the field, which sets the stage to probe the nature of the barrier. Sub-seismic faults or variations in stratigraphic architecture or diagenetic character may be responsible for permeability reduction and lateral sealing. The heterogeneous spatial arrangement of channels within the fluvial system of the reservoir are observed on various seismic volume attributes (e.g. Fig. 7), which could have resulted in discontinuities along fluid flow pathways and reservoir compartmentalization. Furthermore, given the close proximity to a major east–west strike-slip fault forming the northern limit of the field (Figs 5, 6a), it is not unlikely that other, smaller, east–west to NE–SW oriented strike-slip faults may exist in the region. Fluvial channels can be interpreted to be deflected laterally across potential faults (Fig. 7), although no clear discontinuities are observed on structural seismic attributes, such as variance. If such faults exist, they will add to the complexity of the fluvial system by offsetting individual channels locally, which might not be detectable seismically.

Field U produces gas condensate from a thick (∼1000 ft) Paleozoic shallow marine sandstone reservoir within an anticline developed on the flank of Carboniferous structures (Wender et al. 1998), and bounded to the west by a major reverse fault (Fig. 8a, b). The reservoir consists of fine- to medium-grained sandstones, weakly cemented by authigenic illite clay, and is capped by a shaly silty unit in its upper section below an unconformity. The topmost several meters of the reservoir immediately below the unconformity are impregnated with solid bitumen, which is mappable petrographically across the field (Arouri and van Dijk 2021). Bitumen occurs in the shallow wells only as pore-lining but mainly as pore-filling in coarser-grained sandstone of intergranular porosity, exemplified by the petrographic image insert in Figure 8b. The effect of solid reservoir bitumen on significantly reducing total effective porosity and permeability, which could lead to the formation of seals, is widely recognized (e.g. Lomando 1992). It is not uncommon for heavy oil or even tar to develop towards the oil–water contact due to density segregation, water washing or biodegradation. It is less common, however, for solid bitumen to form up-section near the crest of the structure. To identify the formation mechanism responsible for bitumen development at shallow levels within the trap, an integrated geochemical and diagenesis study was conducted (Fig. 8c). Methods used to reconstruct the charge history and the paragenetic sequence include thin-section petrography, X-ray diffraction, scanning electron microscopy, sequential extraction of bitumen (cf. Schwark et al. 1997; Arouri et al. 2009), fluid inclusion microthermometry, radiometric (K–Ar) age dating, and basin modelling. Further details on methodologies can be found in Arouri and van Dijk (2021).

Bitumen-development model

The shallow-marine sandstone reservoir shows similar sedimentological and diagenetic features across the field. No direct relationship exists between bitumen occurrences and sedimentology or diagenesis (Arouri and van Dijk 2021). Processes that commonly lead to the development of bitumen, heavy oil or tar, such as water washing, biodegradation, thermochemical sulfate reduction, or thermal cracking (e.g. Blanc and Connan 1994), were discounted as possible mechanisms in this case due to the lack of geochemical evidence.

The proposed model, summarized schematically in Figure 8d, suggests that the reservoir was initially charged with oil during the Early Jurassic, followed in the Early Cretaceous by gas at around 130 Ma, and continues to the present-day. The formation of bitumen started at 129 Ma, shortly after the gas charged the upper part of the structure, most likely via a major reverse fault (Fig. 8d). Evidence for multiple charging is manifested in residual oil extracts and fluid inclusions (Arouri and van Dijk 2021) and by recent fluid data on the overlying sandstone reservoir that show gas escape features, such as abnormally high GOR and fluid maturity, in the fault area, overprinting the pre-existing fieldwide density segregation profiles. Gas influx into the pre-existing oil induced fluid disequilibrium and the subsequent deposition of bitumen across the fault. The process appears to have only affected the upper part of the tilted reservoir immediately below the caprock, as no bitumen is reported in deeper part of the structure that was apparently inaccessible to the late gas charges. Mixing of fluids of different properties destabilizes asphaltene, leading to tar-mat formation, adversely impacting reservoir permeability (e.g. Pfeiffer et al. 2017). It is envisaged that the solid bitumen has improved the efficiency of the shaly silty sandstone caprock in sealing subsequent charges of drier gas currently on production.

Field S is a large oil and gas field producing from two Mesozoic carbonate reservoirs separated by a regional anhydrite seal within an anticlinal closure (Fig. 9). The two reservoirs are charged from two different source rocks and, therefore, represent two petroleum systems separated by the seal (Fig. 10a). Reservoir-A is capped by the anhydrite seal, and produces mainly gas condensate to wet gas sourced from a gas-mature source rock (Petroleum System A). Reservoir-B overlies the seal and produces mainly oil sourced from an oil-mature source rock (Petroleum System B). The seal is tens of feet thick and is laterally expansive regionally, and has traditionally been considered an effective barrier to vertical migration between the two petroleum systems (Murris 1980). However, the late discovery above the seal of localized gas capping the oil in an, otherwise, established oil system has raised questions about its origin and the likelihood that the gas has leaked through the seal, especially given the fact that the maturity of the source rock above the seal is insufficient to generate gas. A geochemical study was therefore conducted on fluids from both reservoirs in order to test the possibility of gas leaking through the seal from Reservoir-A to Reservoir-B.

Oil analysis

Results of the GC fingerprinting of the C7 compounds, selected saturated and aromatic hydrocarbon biomarkers, δ13C of saturates and aromatics fractions and of the C1–C5 hydrocarbons are summarized in Figure 10. The C7 compounds and source-specific biomarkers (Fig. 10b) are evidently different, suggesting different sources for the oils. For example, the oil in Reservoir-A is characterized by C29 > C30 hopane, low diasterane/sterane ratios, and dibenzothiophene/phenanthrene ratios greater than 6.2. In contrast, the oil in Reservoir-B is characterized by C29 < C30 hopane, high diasterane/sterane ratios, and much lower dibenzothiophene/phenanthrene ratios <2.7. These contrasting biomarker assemblages suggest a carbonate source for Reservoir-A oil, and a clayey source for Reservoir-B oil. The inference on different sources for the oils is supported by their significantly different stable carbon isotope composition measured on the saturates and aromatics fractions (Fig. 10c, top).

Gas isotope analysis

Unlike the liquid hydrocarbon component that is isotopically different above and below the seal, the C1–C5 gas carbon range in both reservoirs is isotopically similar (Fig. 10c, bottom). The similarity suggests a common source for the total gas range in both reservoirs, hinting to the possibility that the gas leaked freely through the anhydrite seal. Carbon isotope fractionation during fluid migration is reported to range from negligible (e.g. Fuex 1980; Zhang et al. 1995) to significant >5‰ (Prinzhofer and Huc 1995; Prinzhofer and Pernaton 1997). Zhang and Krooss (2001) experimentally investigated carbon isotope fractionation of methane during diffusive gas migration through shale samples and reported a significant carbon isotope fractionation of the diffused methane that increases with organic richness of the shale.

Given that (1) methane is unlikely to diffuse through an intact anhydrite seal (Downey 1984), and (2) leakage/diffusion through shaly intervals possibly intercalating with the anhydrite seal would result in significant depletion of 13C in the diffused methane (Zhang and Krooss 2001), the most feasible mechanism for gas leakage through the seal here is localized disruptions by fracturing or faulting, which reduces the capillary displacement pressure, allowing the entire gas range (C1–C5) to leak. A wholly intact anhydrite seal would result in a lighter carbon isotopic profile for the overlying (less-mature) gas (hypothetically represented by trend C; Fig. 10c, bottom). On the other hand, a seal that is tight enough for wet gas but leaks methane would yield identical methane δ13C composition but different isotope profiles for the wet gas range, exemplified by trend D (Fig. 10c, bottom).

Figure 11 shows seismic interpretation and variance attribute slices that, in light of fluid findings, indeed demonstrate the existence of a fault along the anticlinal structure of the field. The fault can be seen clearly on the variance attribute below the seal and, to some extent, above the seal. Within the seal, however, it is harder to detect the fault as seismic reflectivity becomes more discontinuous. Nonetheless, it is possible to observe a semi-linear feature along a horizon within the seal paralleling the interpreted fault (Fig. 11c). In combination with the fluid-based findings, it is likely that this fault, along with other faults in the region, has allowed gas leakage across the seal onto the shallower reservoir, forming a chemically distinct gas cap. Further work is needed to identify possible weakness zones, where the anhydrite seal is breached and hence gas leak is focused.

Field H is a shallow anticlinal structure <1650-meter deep, currently in the appraisal phase (Fig. 12). Across the 25-km long four-way structural closure of over 122-meter vertical relief, only seven wells have been drilled into the Mesozoic limestone reservoir (Fig. 13), which is sealed by thick anhydrite. Drilling tested heavy to medium oil, ranging in API gravity from 14° to 26°, and thermal maturity from 0.78 to 0.84% VRe (vitrinite reflectance equivalent). The field is interpreted to have been charged via long-range migration from a carbonate source rock located about 100 km to the north (Arouri et al. 2016; Arouri and Panda 2022). Given that the field was charged during the Eocene and Miocene, and that the present-day reservoir temperature is low (54–63°C), the lack of any sign of biodegradation suggests that the reservoir was charged at greater depths and temperatures (>80°C) sufficient to sterilize the reservoir (cf. Wilhelms et al. 2001) and hence prevent biodegradation, before it was uplifted to its current setting.

Wells 1, 6 and 7 drilled in the northern part of the field encountered a fluid contact that is about 40-feet deeper than the contact in the central part of the field. Only three out of the seven boreholes drilled (marked in yellow in Fig. 13) have PVT data, which were inconclusive in determining fluid connectivity. Fortunately, all seven wells recovered oil samples from drill-stem-testing. Routine GC fingerprinting of light hydrocarbons was, however, unfeasible because of the high viscosity of samples with low API gravity. A more detailed molecular analysis was, therefore, conducted to evaluate whether variations in fluid contacts represent compartmentalization or tilting of fluid contact (scenarios 1 and 2, Fig. 13). Because molecules in oil have different densities, a molecular density segregation index was developed in order to assess fluid relationships and potential compartmentalization. An example of molecular density ratios that were used to quantify molecular density segregation is depicted in Figure 13, which represents the ratio of naphthalene (molecular density 1.162 g cm−3) to methylnaphthalene (molecular density 1.013 g cm−3). This and other molecular density segregation ratios are expected to increase with depth due to density segregation at molecular level. In Field H, two distinct molecular density segregation trends are evident for the northern flank and the central structure of the field. The break in the molecular density profile between wells 1 and 5 agrees with different oil–water contacts, possibly due to faulting.

Fluid maturity data from various saturated and aromatic hydrocarbons were also collected, and maturity stratification profiles constructed, an example of which is shown in Figure 13, which represents the 1,6-dimethylnaphthalene/(1,7 + 1,3-dimethylnaphthalene ratio) (cf. Alexander et al. 1994). Similar to the molecular density profiles, a clear separation in fluid maturity stratification profile is evident, suggestive of fluid discontinuity by a lateral seal.

Possible cause for lateral sealing

Analysis of seismic data and attributes in light of the fluid findings showed that the structure is not a simple anticline but is potentially affected by faults oblique to the fold trend (Fig. 14). Most importantly is the presence of a fault or several faults that separate the two fluid systems interpreted from the molecular density segregation and maturity stratification (Fig. 14a, b). Core observations from a later drilling in the fault zone also confirm the presence of a fault system. This fault system is mostly perpendicular to semi-perpendicular to known maximum horizontal compressional stress in the region, which would, therefore, favour the closure of the fault. More analysis, however, could possibly be conducted to evaluate the fault rock sealing potential.

PVT data routinely used by reservoir engineers for identification of fluid phase and for production simulation additionally contain wealth of information for fault and seal integrity assessment, particularly when systematically integrated with fluid geochemistry. Such a multiscale integration of fluid character will provide a more accurate delineation of fluid relationships, which can then translate to useful information on seal integrity and the dynamic fault character, as demonstrated in several examples from the Arabian Peninsula. Furthermore, fluid data and interpretation of sealing potential can be integrated with and/or guide further characterization of reservoirs using seismic data (i.e. attributes), other geophysical data (e.g. images, logs and advanced sonic analysis), and geological data (e.g. core data). This work can reduce exploration risk and help optimize reservoir models and field development. Further research on extracting geochemical information from pre-existing PVT data will prove useful in fault and seal integrity assessment.

The authors acknowledge several fruitful discussions and interactions over the years with present and past colleagues at Saudi Aramco who have collaborated on different fields, notably Carlos G. Herrera, Huzaifa M. Elbushra, Rainer Schmidt, Yogi Priyadi (Fields A and M), Aijaz Shaikh, Hisham Salem, Peter Jenden, Pierre J. Van Laer, Radhay Bansal, Nicoli Garner (Field G), Clemens van Dijk (Field U), Yunlai Yang, Emmanuel Uba, and Nadeem Abbas (Field H). Pan Luo is thanked for general discussions on carbon isotopes. Recommendations from Guest Editors Emma Michie (University of Liverpool, UK) and Graham Yielding (Badley Geoscience Ltd, UK), as well as reviews by Oliver Mullins (Schlumberger, USA) were detailed and greatly improved the organization, content and clarity of this paper.

KRA: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), project administration (lead), writing – original draft (lead), writing – review & editing (lead); SAA: structural analysis (lead), writing – review & editing (supporting)

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated and analysed during this study are included in this published article to the extent allowed by Saudi Aramco.

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License (http://creativecommons.org/licenses/by/4.0/)