Abstract Plate tectonics forms and destroys sedimentary basins, accumulating organic carbon and converting it into mobile petroleum which may be concentrated in reservoir traps in which, if temperatures are below 80°C, it may become biologically degraded (biodegraded). The biodegradation process produces altered, denser, heavy oils and methane as a primary product. Much of the world’s oil is biodegraded under anaerobic conditions, with methane being a major by-product of the action of the deep biosphere on petroleum when sulphate is not present as an oxidant. A review of the literature relating to destruction of wet gas and the systematics of methane generation during subsurface oil biodegradation concludes that large biodegrading oil fields may be major source systems of dry gas.

Abstract Since the first Geological Society Special Publication on reservoir geochemistry ( Cubitt & England 1995 ), there has been a steady increase in the number of publications and conferences on this topic. Reservoir geochemistry is now a routine tool of the petroleum industry with new companies now providing dedicated reservoir geochemical services. With 95% of future world oil production likely to come from already discovered accumulations this steady increase in research and application of reservoir geochemistry is welcome. Despite this, there remain many challenges and great areas of ignorance and missed opportunities. This volume therefore provides a summary of the current status of reservoir geochemistry and the challenges that remain. Reservoir geochemistry studies the compositional variations of petroleum reservoir fluids (waters, oils and gases) at a variety of spatial and temporal scales. These studies reveal information about petroleum basin development, details of reservoir filling and leaking, and about petroleum mixing and alteration ( Larter & Aplin 1995 ). This information is of interest from both the academic and applied points of view, and provides insights not available from other methods. Reservoir geochemistry is concerned with the 3D compositional variations in reservoir fluids that are commonly observed in petroleum reservoirs as free fluids or as fluid inclusions in mineral cements. It strives to understand the origin of these heterogeneities and to apply them to enhance the understanding of the formation of petroleum accumulations and exploration, appraisal, development and production strategies. Key components of the approach are (1) the analytical methods for determination of fluid compositional heterogeneity, (2) the physical, chemical and numerical models for interpreting the compositional differences in terms of basin history and reservoir connectivity and (3) the case studies of application. By way of introduction, we attempt to summarize these main areas, which provide the basic currency of development within the field of reservoir geochemistry.

Abstract Molar concentration profiles of reservoir fluids reveal accumulation history and alteration. Slope Factors (SF) define rates of exponential decrease in concentration of light n -alkanes (C 3 - n C 5 ) and liquid pseudo-components (P 10+ ) with increasing carbon number. Most petroleum fluids are substantially characterized by SF(C 3 - n C 5 ) and SF(P 10+ ), the former invariably being the greater. Specific paired values of SF(C 3 - n C 5 ) and SF(P 10+ ) are process-diagnostic. In oils, maturation, gas injection, evaporative fractionation and migration depletion involving loss of gas, are recognizable. SF data are interpreted in the light of PVT analyses representing oils (principally from western Canada) and gas-condensates (from numerous basins), also asphaltene pyrolysis experiments and equation of state calculations. Covariant increase of SF(C 3 - n C 5 ) and SF(P 10+ ) during maturation is demonstrated, but correlation is frequently destroyed by modification of the light ends by the admixture of allochthonous gas, increasing only SF(C 3 - n C 5 ). Secondary gas-enrichment is a requisite process for the generation of gas-condensates by evaporative fractionation. Compositional criteria for the recognition of enrichment are provided for the first time, particularly attainment of a value of SF(C 3 - n C 5 ) exceeding 1.69 (a tentative limit). Available data indicate that the process has occurred in a large proportion of oil accumulations, ranging from 23% of 30 reservoirs in the Jurassic Smackover Formation in Alabama, to 78% of 36 in the northern North Sea.

Abstract Essentially all petroleums are mixtures with different components charged from source rocks at different temperatures. This heterogeneous charge is the basis for compositional differences in reservoirs that are the basic elements of reservoir geochemical approaches. Because many classical petroleum geochemical tracers of source facies and maturity, such as the cyclic biomarker hydrocarbons, show several orders of magnitude variation in concentration in petroleum systems these compounds do not reliably track facies or maturity signals in mixed oil situations. Light hydrocarbon and aromatic hydrocarbon parameters are more reliable in this sense but, as mixtures are the norm, the concept of the maturity of oils needs revising. We suggest an alternative approach is needed which tracks the maturity/petroleum mass fraction relationships for reservoired oils (mass fraction maturity) and allows the bracketing of source kitchen maturity. We strongly advise against using compound ratios in reservoir geochemical studies without having knowledge of the compounds concentration range variations within the petroleum system being studied.

Abstract Over the past 10 years, investigations into the characteristics of the high molecular weight hydrocarbon (HMWHC) fraction in crude oils and, to a lesser extent, source rock extracts have continued to reveal novel information concerning the distribution of hydrocarbons >C 40 . The major impetus for this work has come from the fact that HMWHCs can cause significant production problems in certain geographical regions and particularly deepwater frontier areas. Since these HMWHCs appear to be ubiquitous in crude oils, the primary questions that need to be addressed are: what are these compounds, where do they come from, and how do they affect physical properties of oils? Here, we review our work over the past decade and discuss the significance of these results and their potential application to reservoir and production problems involving paraffins and asphaltenes. It was commonly believed for many years that only oils derived from source rocks containing higher plant source material would have a high paraffin content. However, it is now abundantly clear that oils derived from lacustrine and marine source rocks also contain relatively high concentrations of higher molecular weight hydrocarbons. In addition to developing methods for the qualitative and quantitative separation of HMWHCs from asphaltenes, progress has been made in identifying individual components of the high molecular weight fraction. This fraction is not a simple mixture of n -alkanes but a complex mixture of seven or eight homologous hydrocarbon series, each with significantly different physical properties. A knowledge of these structures is important in predicting crude oil properties such as cloud point and pour point. Series identified to date include alkylcyclopentanes, alkylcyclohexanes, alkylbenzenes and various branched hydrocarbons. In summary, since the 1970s most of the geochemical research emphasis has been placed on compounds below C 40 . Whilst compounds above C 40 may not have the same degree of structural specificity as the traditional biomarkers, the amount of information available from these compounds could be extremely beneficial in the long term, particularly for reservoir characterization and production purposes and other problems involving high molecular weight hydrocarbons.

Abstract Anaerobic processes have only recently been recognized as an important mechanism in the biodegradation of crude oils. They are normally invoked to explain extensively biodegraded oils with little or no possibility of contact by oxygenated waters from an active aquifer. This work with Kuparuk Field indicates that early stages of anaerobic biodegradation can be subtle and easily missed, yet have economic impact. Kuparuk River Field, North Slope of Alaska, comprises two reservoir intervals: vertically stratified and imbricated lower shoreface sandstones (A sands), and overlying shallow marine sandstones with complex permeability structure (C sands). The vertical and lateral distribution of viscous oil (less than 20° API) shows a strong relationship to structure and faulting in the Kuparuk Field. Multiple mechanisms for the origin of tars and viscous oils can be proposed, including early aerobic biodegradation, anaerobic biodegradation, inorganic oxidation and gas deasphalting. This geochemical study, integrated with stratigraphic, structural and production data, was undertaken to help understand the origin and distribution of tar and viscous oil in the field. Obvious depletion of n -alkanes and other paraffins, classically regarded as indicative of early biodegradation, is not observed in examined samples. However, Kuparuk viscous oils show slight to extreme selective depletion in long-chain alkyl aromatic (LCAA) hydrocarbon families (e.g. alkylbenzenes and alkyltoluenes). This is interpreted as indicative of an early stage of anaerobic microbial degradation that likely destabilized the oil to promote subsequent precipitation of asphaltenes as tar. Depletions in LCAAs in core samples in the field are linked to decreased hydrocarbon/nonhydrocarbon ratio and to an increase in the high molecular weight (>C 50+ ) components of Rock-Eval 6 pyrolysates. Using a calibration curve constructed from oil Rock-Eval 6 pyrolysis, the API gravity of core oil plus bitumen can be estimated. Tar-plugged formations with depleted LCAAs have estimated API gravities <8°. Portions of the Kuparuk reservoir with higher iron content tend to show greater depletions in LCAA. Anaerobic biodegradation is likely mediated by dissimilatory iron-reducing bacteria. Biodegradation likely destabilizes the oil with respect to asphaltene precipitation such that later arrival of petroleum leads to tar in the reservoir. Increased tar and depleted LCAAS correspond to intervals with lower productivity indices, thus indicating a significant impact on petroleum producibility.

Abstract We present a computational framework that can be used to estimate phase equilibria, equation of state properties and composition-dependent viscosity, aimed at the geochemical community for modelling reservoir processes, and at the chemical community to quickly estimate continuum properties of known mixtures. The framework presented is an extensible, component-based set of modules that can be used in calculating phase properties (volumes and densities, compositions and conditions of phase separation, Henry’s Law constants and viscosities) at a wide range of pressures, temperatures and starting compositions. There are two pieces to the model: the HCToolkit (a set of Perl modules that act as computational engines) and the EOSInterface (an ActiveX wrapper for the HCToolkit) which allow the models in the HCToolkit to be usable in Microsoft Office programs. An application hosted in Microsoft Excel is included within the distribution. The EOS models are adaptable for mixtures of arbitrary complexity, with number and types of components only limited by hardware. Model runs with mixtures of over 100 components have been tested, and are perfectly feasible. Implemented within the software are four equations of state; additional equations of state can easily be added. Also implemented are G E models and mixture viscosities. Finally, the EOS models provide liquid-vapour flash calculations, liquid-liquid flash (designed for petroleum-aqueous solutions) and the generation of phase diagrams.

Abstract This paper highlights the benefits of using knowledge of the rates of fluid mixing in the interpretation of reservoir fluid data. Comparison of the time it would take for a fluid difference to mix with the actual time available for mixing to occur allows two significant advances over a purely statistical analysis of reservoir fluid data: (1) differentiation of a step in fluid properties, indicative of a barrier to fluid communication, from a gradient indicative of incomplete mixing; and (2) quantitative estimation of the degree of compartmentalization that can readily be adapted into models for prediction of reservoir production performance. We review the existing equations that estimate the mixing times for three main types of variation in fluid properties (fluid contacts, fluid density and fluid chemistry). In addition, a new relationship for fluid pressure mixing is presented. In each case the relationships were validated by comparison with numerical simulation. The different fluid mixing processes were compared by applying the equations to a range of simple fluid scenarios in one simple reservoir description. This shows that mixing times for fluid mixing processes are diffusion > fluid density > fluid contacts > fluid pressure. For each scenario, the processes were analysed in terms of the volume of fluid that must move in order to bring the system to equilibrium and the drive for fluid mixing (pressure difference × permeability/viscosity). Perhaps surprisingly, there is an excellent linear relation between fluid mixing times (a) calculated from the mixing equations and (b) estimated from volume/drive. This indicates that fluid volumes and mixing drive are the main controls on fluid mixing times. This can be used to derive simple interpretation guidelines to estimate mixing rates even in the absence of quantitative modelling. A simple field case study demonstrates how this understanding of fluid mixing times can add value to the interpretation of reservoir fluid data.

Abstract A common assumption is that hydrocarbon charge homogenizes with the petroleum already in a trap: thus, compositional gradients reflect only subsequent segregation of the petroleum under the combined influences of gravity, temperature and diffusion. Since such homogenization would entail an unfavourable generation of potential energy, a more plausible hypothesis is that hydrocarbons stack into traps roughly in the density sequences in which they arrive. The commonly accepted model suggests that homogeneous and gravitationally unstable petroleum columns move towards graded equilibrium ones, whereas the hypothesis preferred here implies that they gradually diffuse towards equilibrium from the opposite direction, i.e. poorly mixed initial states. According to this hypothesis, trends of oil GOR and bubble point ( P b ) are controlled by either (a) the charge GOR or (b) the evolving PT conditions at the GOC, depending on the gassiness of the charge. By contrast, API gravity trends mainly reflect the integrated maturity histories of source rock kitchens during the trap filling regardless of the charge gassiness. Because trap and kitchen histories are usually (very) different, a wide spectrum of API and GOR (and P b ) combinations is possible, from minor API gradients coupled with large GOR gradients and vice versa .

Abstract The evolution of the petroleum systems in the Tampen Spur area, with main focus on the filling directions of the northern part of Snorre field, was addressed through 2D basin modelling (Petromod V. 4.5 and 7.0). The geochemical classification of the petroleum populations in the area represented the framework for considering the different kitchen areas and migration systems. Results from the basin modelling support, in general terms, the previous geochemical classification and petroleum families in the region. However, a separate well-defined main kitchen area for the Snorre Field was deduced opposed to the multiple kitchen areas having contributed to the filling as proposed in the literature. Our conclusions are based on the quantitative evaluation of the different proposed kitchen areas and the timing and extent of petroleum generation. Modelling of petroleum generation was performed using asphaltene kinetics determined on petroleum asphaltenes from Snorre oils. This approach was chosen in order to avoid problems associated with the kinetic variability encountered in the Draupne formation. The petroleum asphaltene kinetics was used to delineate the extent of the kitchen area, which reached the time/temperature conditions necessary for the generation of the analysed oil phase. The results thus differ from conventional oil window approximations as we utilize kinetic source rock parameters in the migrated oil for tracing out the generative basin. Three 2D lines crossing the main kitchen areas were modelled in this study. The models were calibrated to data from eight wells, consisting of measured vitrinite reflectance, corrected well temperatures and pore pressure. Three main kitchen areas were considered; one to the west and northwest of Snorre field, one directly to the north (Møre basin) and one to the east of the field (34/5 kitchen). Modelling suggests that the kitchen area to the west and northwest of Snorre is largely immature and that the volume of potentially generated petroleum is too small to fill the Snorre structure. In the northern kitchen area, the seismic indicated very thin upper Jurassic deposits, which reaches oil window maturities only at a relatively large distance from the structure. The modelling also demonstrated problems related to the filling of the Snorre structure from the Møre Basin. The combined effect of a thin source rock, which implies a regionally large drainage area to fill the structure, and the large distance to the mature kitchen, lead to the conclusion that the Møre Basin did not contribute significant volumes of petroleum to the Snorre field. In contrast, the kitchen area east of Snorre Field (the 34/5 kitchen) proved in the modelling to be mature and volumetrically large enough to account for the entire filling of the Snorre Field.

Abstract The possibility to model petroleum composition during hydrocarbon generation as well as the PVT behaviour of the fluids during migration has only recently become available in modern basin modelling software packages. While various compositional kinetic models of petroleum generation have been published in the past few years, none of the studies presented have attempted to match the composition, physical properties and phase state of known petroleum accumulations. Using compositional data from closed-system non-isothermal pyrolysis experiments, we developed a compositional kinetic model of hydrocarbon generation for a marine Type II source rock, which uses 13 components to describe the generated fluid. The data format selected is compatible with the compositional resolution used in reservoir engineering, thus allowing a direct comparison of predicted compositions and phase behaviour with PVT data of natural fluids. Compositional predictions of the model were tuned to a well-documented maturity sequence from the Tampen Spur, Norway, and the calibrated model implemented in a 2D basin modelling study of the Snorre Field, Norway. The results of the modelling led to an excellent correlation between predicted and reported reservoir fluid properties (formation volume factor, GOR and saturation pressure) for the present-day situation. The results indicate that the Snorre reservoir has received a continuous charge since the late Cretaceous-early Tertiary and that it most likely contained a two-phase system prior to the latest Plio-Pleistocene burial and overpressuring event.

Abstract There is considerable diversity in petroleum type within the Judy and Joanne Fields of the Central Graben. Superficially, the three major reservoir systems can be considered to contain the following fluid types: gas-condensates in the Pre-Cretaceous, undersaturated black oils in the Chalk and gas-condensates in the Palaeocene. Reality is, however, quite different. This paper presents the results of several geochemical studies undertaken in the area; oil/condensate analyses to identify differences in maturity and source input, a pressure data review, high resolution GC fingerprinting, strontium isotope analyses to investigate reservoir connectivity/compartmentalization, and 2D basin modelling to determine the timing and extent of petroleum expulsion and migration. These components have been synthesized to produce a composite petroleum charge model which adequately explains the differences observed. This, in turn, enables an assessment to be made of the likely impact, if any, on field development. Furthermore, the petroleum charge model can be applied to predict the charge risk, fluid type and likely petroleum-water contacts in untested parts of the field and in the immediate vicinity. The results of recent development wells are reviewed in the light of the charge model. Subsequent reservoir geochemistry studies have confirmed the validity of the model and highlight additional applications for reservoir management.

Abstract The giant oil fields of the Val D’Agri region (Southern Apennines Internal Thrust Zone, Italy) may rank as the largest onshore accumulations in Europe, but these resources pose special technical challenges due to the secondary alteration process identified as compositional grading. This alteration process is attributed to petroleum system elements and processes that lead to the formation of an unstable oil column. These oil columns are studied within a mass balance perspective that allow identification of key molecular fractionations which can be used to properly diagnose this alteration mechanism in other oil accumulations. The molecular signature is defined in the organic sulphur fraction as well as different hydrocarbon classes, and is completely consistent with the engineering criteria for this reservoir process. This discovery is particularly critical to reservoir compartment studies as the failure to recognize this particular signal will invariably lead to the identification of false compartments.

Abstract The Voelkersen field in northwestern Germany produces thermogenic dry gas from Permian Rotliegend sandstone reservoirs located at depths greater than 4700 m. In addition to this considerable depth Zechstein salt overburden plus the structural complexity and heterogeneity of the reservoirs limit the applicability of conventional methods for definition and characterization of field compartmentalization. The Rotliegend gas is sourced from Westphalian coals of semi-anthracite to anthracite rank with conventional geochemical data suggesting little variation in composition on a field-wide scale. However, detailed analysis of associated condensates and wet-gas hydrocarbons shows considerable differences between compartments and producing horizons, despite the apparent high maturity of the gas. These differences support compartment definition and identification of lateral and horizontal seals. Assuming a quasi-uniform gas composition in communicating reservoirs, variation of carbon and hydrogen isotopic data in particular confirms separation of marginal fault blocks. Additionally, these data also suggest reservoir continuity within a large field compartment and over parts of a major fault system for which the sealing properties so far were previously uncertain.

Abstract The Magnolia Field in the deepwater northern Gulf of Mexico is a Plio-Pleistocene age mixed phase reservoir whose fluids are not in compositional equilibrium. Fluid heterogeneities have arisen principally due to (1) variations in maturity of the source rock from which the hydrocarbons were derived, (2) the extent to which biogenic methane has been incorporated into the fluids and (3) phase fractionation effects. These influences express themselves both in terms of bulk fluid properties such as gas/liquid ratio, API gravity and saturation pressure and minor compositional attributes such as hydrocarbon gas isotopic composition and gasoline range molecular ratios. Significant compositional variations that cannot be ascribed to gravitational fluid segregation occur within reservoirs that are demonstrably in pressure communication. These variations challenge the notion that hydrocarbon fluid mixing is geologically instantaneous and underscore the importance of testing assumptions regarding compositional equilibria in conjunction with reservoir studies. Although the state of disequilibrium impedes compartmentalization assessments at Magnolia, it provides both opportunities for fluid property and phase predictions and potentially a development setting in which geochemical surveillance techniques may be profitably employed.

Abstract Jurassic Norphlet Formation sandstone reservoirs in Mobile Bay (offshore Alabama, USA) produce gas from great depths (>6.4 km) and elevated temperatures (>200 °C). Quartz cement is concentrated at the top of these aeolian reservoirs forming a low porosity ‘tight-zone’ of widely variable thickness (3–58 m) above a more porous reservoir section. The extent of the tight-zone is independent of depositional facies and its thickness strongly influences well performance. Intergranular porosity in the Norphlet has been preserved by inhibition of quartz cementation due to the occurrence of robust grain-coating chlorite. Quantitative petrographic data reveal that chlorite grain-coat coverage is less in the tight-zone sands (mean = 92%) than in the reservoir sands (mean = 99%). Burial history and quartz precipitation kinetics modelling indicate that this seemingly minor difference in the completeness of grain coatings is sufficient to produce the observed differences in cementation and porosity. Quartz cementation to form the tight-zone took place under conditions of deep burial and high temperature. It followed in time the onset of pressure solution, emplacement of liquid hydrocarbons, and the precipitation of a solid hydrocarbon film (pyrobitumen) on the walls of the intergranular pores. Fluid inclusion microthermometry data indicate that volumetrically significant quartz cement precipitated at temperatures of 150 °C or greater from highly saline aqueous fluids. Hydrocarbon-bearing inclusions are notably absent in quartz cement of the tight-zone, implying that the pore fluids were predominantly brine during precipitation. Oil and gas associated with pyrobitumen evidently escaped from Norphlet traps prior to tight-zone cementation. Gas presently found in Norphlet reservoirs of Mobile Bay represents a relatively recent accumulation and is not the product of in situ thermal cracking of oil.

Abstract Sandstones of the Upper Permian Bell Canyon Formation were deposited by turbidity currents in a basin-floor setting. The sandstones were deposited in a channel-levee system that terminated in broad lobes; overbank splays filled topographically low inter-channel areas. Diagenesis and reservoir quality of the sandstones were examined in cores from East Ford field, which is undergoing a CO 2 flood. Porosity and permeability are controlled by calcite cement, mainly concentrated in layers ranging from 5 to 40 cm in thickness. In a new infill well, initial production was of a high gas volume that contained a high concentration of CO 2 from the interval beneath several low-permeability, calcite-cemented layers. The CO 2 was most likely from an injector well and was trapped below the calcite layers. Geophysical log correlations support the interpretation that some calcite layers are laterally continuous over a distance of at least 300 m, causing vertical compartmentalization in the reservoir.

Abstract Formation water composition data and formation pressure data from oil, gas and gas condensate fields covering an area of approximately 15000 km 2 in the Central Graben Area of the UK North Sea have been analysed. The purpose was to determine whether large-scale barriers to flow influence the compositional distribution of formation waters and how this aids understanding of overpressure distribution and potential hydrocarbon migration pathways. Analyses (Na, K, Mg, Ca, Sr, Ba and Cl) of water and Elemental Residual Salt Analysis (ERSA) samples from both the water and hydrocarbon legs of wells were analysed statistically to reveal eight different water types in the study area. These were located in distinct geographical areas. Aquifer overpressure from formation pressure data for numerous wells in the area were plotted on large-scale regional maps to reveal the distribution of pressure ‘cell’ or compartments. An integrated interpretation of the location of pressure cells, salinity variations and water types provided information on the current and past fluid flow across the pressure cell boundaries, the operational longevity of these boundaries and the nature of the Mesozoic aquifer. The implications of the work are important for understanding: (a) the usefulness of the new ERSA technique for estimating formation water compositions; (b) the effectiveness of integrating fluid composition and pressure data to understand compartmentalization; (c) fluid flow dynamics in the HPHT sector of the Central North Sea; and (d) likely hydrocarbon migration routes and the character and distribution of overpressured basin compartments.

Abstract Petroleum inclusion and geochemical data from core extracts were applied to deduce a model for oil migration, overpressure development and palaeo-leakage of oil from currently dry structures in the Haltenbanken Vest area. The existence of fluorescent oil type inclusions in quartz in the Smørbukk (Åsgard-2) field suggest that oil migrated into this structure 70–50 million years before present (Ma bp). This is also the case for the dry structures 6506/12-4, 6506/11-3 and 6506/11-1, west of the main Smørbukk Fault Zone. Black oil inclusions with medium gas/oil ratio (GOR) occur in these fields together with condensate-type petroleum inclusions. This suggests that the dry structures transformed from containing oil to condensate before leakage. Petroleum extracted from inclusions in these structures and in nearby fields have identical marine type II kerogen signatures. Source rocks at the Spekk Formation level in the current drainage area of Smørbukk and these dry structures, were immature 70–50 Ma bp and the Smørbukk Sør (Åsgard-3) field did not fill at this early time. Thus, oil must initially have entered into Smørbukk from areas to the W-SW, through the currently pressure sealing Smørbukk Fault Zone which today marks the westward limit of the Smørbukk field. Diagenesis in this fault zone caused the much later overpressure development and petroleum was lost from the 6506/12-4, 6506/11-3 and 6506/11-1 structures as overpressure built up regionally. Petroleum loss from these structures with their often thick seals must have occurred via self-propagating open-fracture-induced mechanisms. Lack of petroleum in the Cretaceous strata above these structures suggests that leakage occurred to even shallower strata. This could imply that the Cretaceous strata in Halten Vest were overpressured at the time of leakage. In contrast, the oil in the Cretaceous Lysing and Lange Formation (above the Jurassic reservoirs in Smørbukk and Smørbukk Sør) most likely originated (based on geochemistry and GORs) from the Jurassic reservoirs below and not from Cretaceous strata. This migration event would have been facilitated if it occurred before these sands became overpressured as they are today. Modelling suggests that the Spekk Formation became mature in the Smørbukk Sør region <10 Ma bp and microthermometry of oil inclusions from Smørbukk Sør supports filling during the past 10 Ma. This implies that caprock failure in the Halten Vest structures 6506/12-4, 6506/11-1 and 6506/11-3 most likely occurred after filling of the Smørbukk Sør and 6406/3-1 structures. Rapid regional burial during the past 10 Ma caused local migration of oil into Smørbukk Sør, Smørbukk and 6406/3-1 structures, and generation of high GOR oils in the deeper Halten Vest region. High GOR petroleum inclusions in the Halten Vest structures signify this event and suggests that caprock fracturing occurred after a gas-condensate had replaced oil in these traps. Rapid burial during the past 3 Ma is likely to have caused the current overpressure and associated leakage in Halten Vest. The fact that these traps did not later refill in this progressively subsiding and maturing basin must be related to trap pressures remaining too close to the actual fracture pressures.