5.2.1 Depth Conversion

Much of the data that we use to describe the subsurface – whether it be for appraisal, exploration, or production – are derived from seismic. Moreover, such seismic data are often displayed using two-way time, not depth, for the vertical axis. In making geologic interpretations using conventional seismic displays, we make a mental assumption that two-way time and depth are positively and simply correlated. In gross terms, this assumption is true on most occasions (Figure 5.3). The interpretation of geologic structures and geometries on normal seismic sections is perfectly valid. However, it is important to be aware that on some occasions the Earth’s velocity field is not constant at a given depth. In consequence, obvious structures in the time domain may not exist in the depth domain, and real structures in the depth domain may not exist in the time domain and thus be invisible on seismic data The gross assumption that two-way time and depth are positively and simply correlated is also commonly inadequate for the purposes of well prognosis, well engineering, and evaluation of petroleum volume in a trap. In order to make each of these calculations, it is necessary to convert time to depth. The aim of this section is to examine what data are used in this process, the various methods that can be employed, the resultant uncertainties, and their effects upon the way in which the subsurface structure is interpreted.

Quite clearly, the relationship between depth and seismic two-way travel time is the velocity of seismic energy in the subsurface. The velocity is controlled by the mineralogy of the rock, the state of cementation, the pore fluid, and the porosity fabric. The seismic velocity is greater in well-indurated rocks such as limestone and dolomite compared with most sandstones. Soft sediments such as muds and coals are particularly “slow.” Rock with fluid-filled pores transmits seismic energy more slowly than nonporous rock, and petroleum-filled pores reduce velocity more than water-filled pores. Small quantities of gas can cause dramatic velocity reductions. The pore fabric is also important; for example, a low-porosity but fractured rock commonly has a lower-velocity transmission than a rock with the same primary porosity. This is because the fracturing weakens the fabric of the rock and lowers its elasticity.

The degree of sediment compaction and cementation commonly increases with depth, and as a consequence seismic velocities commonly increase with depth. Exceptions to this generality occur within zones of fluid overpressure (Chapter 3). In such situations, the fluid supports part of the lithostatic load, porosity is often relatively high, and thus fluid content can be high. Consequently seismic velocity can be lower in overpressured intervals than in similar hydrostatically pressured intervals.

It is clear from the above paragraphs that heterogeneity in the lithology distribution, fluid distribution, and pressure (overpressure) distribution will cause heterogeneity in the velocity field. Such velocity heterogeneity needs to be understood to make a detailed conversion between two-way time and depth.

The data used for depth conversion come from one or more of four sources: calibrated sonic logs, pseudo-velocities, stacking velocities, and regional knowledge. Pseudo-velocity is the term used to describe velocity data calculated from combining well depths with seismic arrival times. Velocity data may be calculated for a point (from a sonic log), as an average for an interval or as the average from the Earth’s surface (Tearpock and Bischke 1991). Such velocity data may then be used to convert to depth using one or more of a time–depth curve, an average velocity calculation, a layer cake conversion, or ray tracing.

An average velocity calculation uses a simple algorithm for a whole well (Figure 5.4a). No attempt is made to incorporate systematic variations in velocity that can occur with depth. The technique is often applied when wells are without sonic logs, check shot data, or vertical seismic profiles (VSPs), or when such data are of suspect quality.

In the interval velocity method for time-to-depth conversion, the geology represented on the seismic section is divided into its constituent layers (sequences). The average velocity for each layer is then calculated and the resultant layer cake of interval velocity data used to make the conversion (Figure 5.4b).

The instantaneous velocity model (Figure 5.4c) represents an improvement on the interval velocity method and can be used when velocity data from Check shots or VSPs are more abundant.

Ray tracing is commonly used to check the validity of time-to-depth conversions. The method uses the generated depth maps and velocity data to generate a synthetic seismic section. The generated section is compared with the real seismic data. If the match is poor, the input data are readjusted until an acceptable match is obtained. The refined input data may then be used for time-to-depth conversion elsewhere in the area of interest (McQuillan et al. 1979).

5.2.2 Mapping Surfaces and Faults

Subsurface maps are a fundamental part of petroleum geoscience. They are used at all stages of exploration, appraisal, and development. Regional geologic maps were examined in Chapter 3, and the use of structural maps to interpret petroleum migration and prospectivity was covered in Chapter 4 (including the case histories). We choose to study mapping more closely in this chapter because at the appraisal stage the need for accurate and reliable maps increases disproportionately relative to the data available. Many of the criteria that are used to determine whether a petroleum pool will be developed or abandoned without development will be based upon subsurface maps, and yet the quantity of data available is likely to be limited to a few wells and a 2D or 3D seismic survey.

Maps can be produced for both surfaces and intervals (or their properties). Here, we look at surfaces. Intervals and interval properties are examined later in this chapter (Section 5.8), although the methodology for construction of the surface and interval maps is similar. Maps are most commonly constructed on stratigraphic (iso-geologic time) surfaces. These may be either conformable surfaces or unconformities. Time-slice extractions from seismic data are also maps, and these can also form the basis for structural and stratigraphic interpretations. It is also possible to produce maps on fault surfaces. In such instances, it is common to plot the intersection of the stratigraphy on the fault surface for both the footwall and hanging-wall surfaces.

Cross-sections can also be constructed from maps. The construction of cross-sections can further aid the interpretation of both structure and stratigraphy.

Both structural and interval maps are 2D representations of 3D data. The elevation, thickness, or property data are commonly portrayed as contours. Contours are lines that connect points of equal value. Relative contour spacing conveys much information about the properties of a surface (Figure 5.5). Equally spaced contours indicate that a surface has constant slope (dip). Closely spaced contours indicate a more steeply dipping surface than more widely spaced contours. Therefore, in producing a map from point source well data or from seismic data, care must be taken during the contouring process to ensure that the map accurately conveys the interpretation intended.

There are a few important rules to consider when producing a contour map. A contour line cannot cross itself. It must also form a closed surface (although this may be beyond the bounds of the map). Contours appear to merge on vertical surfaces and appear to cross on overhanging surfaces, but in 3D space the contour lines lie one above another. To avoid confusion, contours on overhanging surfaces are commonly shown dashed (Figure 5.6). Repeated contour lines of the same value indicate a change of slope (synclinal or anticlinal culmination; Figure 5.7).

Tearpock and Bischke (1991) provide a simple guide to contouring, to enable easy construction and subsequently easy understanding by the reader. They suggest the following:

• Contour maps on surfaces should have a chosen reference. For depth or height maps, this is commonly mean sea level. For elevations below sea level, the contour values should be negative. For seismic time maps, the reference surface is also commonly sea level for marine surveys or a defined surface datum for land surveys, and the contours are lines of equal two-way time.
  
• The contour interval should be constant, although every fourth or fifth contour, known as an index contour, should be slightly thicker or bolder.
  
• Maps should have a graphic scale, thereby avoiding problems caused by reproducing the map through enlargement or reduction.
  
• Hachured lines are commonly used to indicate closed depressions.
  
• Contouring should begin in the areas with the highest density of data, and several contours constructed in the same area before venturing into areas of sparse data coverage.
  
• Contours should be smooth and the simplest contour solution that honors all points should be constructed.
  
• There are four basic methods for contour construction: mechanical contouring, parallel contouring, equal-spaced contouring, and interpretive contouring. All are methods, not a depiction of reality, and in each case assumptions are made that may not be valid when comparisons are made with reality. In mechanical contouring, the contours are equally spaced between control points. This means that slopes remain uniform between control points and only change at control points. In parallel contouring, the contours remain parallel – or nearly so! Unlike the mechanical method, this method does not assume uniformity of slope or dip angle. Carried to its limit, this method can yield unrealistic maps, with undulose highs and lows that lie outside the control points simply as a result of the contouring method. Equal-spaced contouring assumes a uniform slope or dip angle across whole flanks of structures. This method may also yield numerous highs and lows that are unconstrained by real data. As the name suggests, interpretive contouring is a method whereby the geologist has complete freedom within the constraints of the fixed data points to interpret the shape and spacing of the contours. No assumptions are made about bed dip or constancy of slope. Instead, the geologist will use his or her experience and imagination to develop the contouring. This method is the one used most frequently and preferred; whilst the method yields the most realistic geologic interpretation it is also the most subjective.

For maps produced from 2D seismic data, the data density is commonly greater than that produced solely from well data. Elevation data (two-way time) are transferred from lines interpreted on the seismic section onto shot point maps. In situations in which 3D seismic data are being interpreted, the data density is commonly so large that the map essentially makes itself as the geologic surfaces are interpreted. Indeed, on most workstation software it is possible to pick every 10 or 20 lines in a 3D seismic volume and then “zap” the surface. In this process, the software automatically picks the same reflectors on adjacent lines. In structurally complex areas such automatic picking may break down, since the continuity between lines is insufficient. With such a high density of data, it is commonly possible to produce a map with a similar level of detail to many topographic maps. Moreover, the high density of data may also allow mis-picked surfaces to be easily identified. Computer-aided contouring can suffer from the same problems as hand contouring, although because the choice of algorithm (method) is not always readily apparent to the operator, the detection of errors is rarely easy.

Faults are commonly an important part of overall trap geometry and internal segmentation of a petroleum pool. Care must be taken in construction of the fault surface, because the interpreted position of the fault will control the calculated volume of the trap or field segment. Care is also needed to define accurately the position of a fault surface from the perspective of drilling wells. Since faults are commonly boundaries within fields, wells are often drilled close to them. Production wells are commonly drilled close to fault block crests (Yaliz 1997) and injection wells for gas or water can be drilled at crestal or flank limits. A well drilled on the wrong side of a fault can be a massive waste of money. A detailed analysis of fault construction, fault balancing, and section restoration is beyond the scope of this book. For this, the reader is referred to Suppe (1985). Here, we will examine fault nomenclature, fault surface maps, and the representation of faults on (stratigraphic) horizon maps.

The components of a fault system are shown in Figure 5.8. The footwall of a fault is the volume beneath the fault surface and the hanging wall is the volume above the fault surface. This nomenclature holds for both normal and reverse faults. The vertical separation (B–C) is the vertical component of the bed displacement, measured as the vertical difference between the projections of a marker horizon on either side of the fault onto that vertical plane. The vertical separation is what is measured in vertical wells and on vertical cross-sections. The missing section (or repeated section for a reverse fault) is called the “fault cut” when measured vertically. The “fault throw” (B–C) is the vertical component of the dip slip (B–D). The “fault heave” (C–D) is the horizontal component of the dip slip, measured orthogonal to the strike of the fault.

For maps constructed on stratigraphic surfaces, fault heave appears as gaps in normally faulted terrain (stippled in Figure 5.9) and as overlapping surfaces in reverse-faulted terrain. It is particularly common to find maps in which the gaps and/or overlaps have been ignored. Quite clearly, in these instances the geologist constructing the map has misrepresented the extent of the formations on either side of the fault surface and, by extension, has also misrepresented the volume of rock to either side of the fault. Representation of reverse faults is also problematic in much of the current generation of geocellular modeling software (Section 5.8). When using such software, it is often necessary to construct a separate geocellular model to either side of a reverse fault, to capture accurately the volume and true extent of both the footwall and the hanging wall. To mitigate the shortcomings of such software, some geologists will choose to make their reverse faults vertical and produce a single geocellular model. Naturally, this compromises the volume calculations. Moreover, such a model cannot be used for well placements.

Fault surface maps are most commonly constructed to enable examination of the relative positions of sealing horizons and reservoir horizons on either side of the fault plane. Such Allan diagrams (Figure 5.10) that is, cross-fault fluid communication, e.g. allow analysis of potential fault seal mechanisms and determination of spill points (Figure 5.10).

5.2.3 Spill Points

The spill point of a structure is the deepest closed contour on that structure. Spill points define the vertical limit of the trap. The volume of rock above the spill point is within the trap and that below the spill point is outside the trap (Figure 5.11). Knowledge of the spill point for a particular trap will enable the vertical closure of the trap to be calculated. Thus spill point identification is an important part of petroleum field definition and ultimately calculation of petroleum in place.

Although both the definition and significance of a spill point are straightforward, the identification of spill points during both exploration and appraisal can be extremely difficult. The density of seismic and well data and its distribution may be insufficient to define adequately any spill point. Some traps have a variety of possible spill points, the integrity of each depending upon different factors, such as four-way dip closure or fault closure (Figure 5.12). Petroleum that is trapped stratigraphically in isolated sandstone or carbonate bodies may be without true spill points.

The relationship between the petroleum/water contact and the spill point is an important one. Much effort is put into understanding the relationship during appraisal. Fields may not be full to spill; that is, the petroleum/water contact may occur above the spill point. Equally, petroleum can be found below the mapped spill point. Of course, this indicates that the mapped spill point is not the true spill point, which must occur deeper.

5.3.1 Fluid Contacts and Transition Zones

A critical part of any appraisal program is determination of the fluid contacts (gas/oil, gas/water, and oil/water) in a petroleum pool. Until a well is drilled into the ground, such contact depths can only be estimated on the basis of a calculated spill point or a direct petroleum indicator from seismic (Section 3.3.3). A common plan is to drill the exploration well on a prospect in such a position that it proves a volume of petroleum (updip) that is deemed to be economically viable. Following the success of that well, the first appraisal well will be drilled in such a position as to penetrate the expected petroleum/water contact and, in consequence, define the size of the pool (Figure 5.13). Although this is a sensible strategy, there is a range of possible outcomes for that appraisal well, only one of which is penetration of an oil/water contact (OWC).

An appraisal well may penetrate petroleum and water intervals in different reservoir units separated by nonreservoir units. From such a well or wells it is possible to define “oil (or gas) down to” and “water up to” elevations. These elevations can also be referred to as “deepest known oil” and “highest known water.” In such instances, the simplest explanation is that the OWC lies between the water up to (WUT) and oil down to (ODT). However, this need only be true if the oil- and water-bearing intervals are in pressure communication (Figure 5.14).

If the appraisal process reveals different contacts in different parts of the field, then clearly the field must be compartmentalized. It can be either layered or faulted, or both. Segmented reservoirs are further investigated later in this section.

Fluid contacts can also be gradational. For a petroleum pool that exists at or above the critical point of the petroleum fluid there will be no gas-oil contact. Instead there will be a continuous gradation due to gravity segregation from gas at the top of the pool to oil at the oil water contact. This was the case for the giant Brent oilfield (UK North Sea) when it was first found. Pressure reduction during production has caused the fluid to no longer be at the critical point and a definable gas cap now exists. The contact between petroleum and water may also be gradational, although the causes are very different. In poor-quality reservoirs, an abundance of small pores and hydrophilic minerals tends to promote high water saturations and gradational petroleum/water contacts. In gas reservoirs, considerable quantities of gas may be dissolved in the formation water underlying the gas leg. Although in this instance the contact may appear sharp, gas can be produced out of solution below the gas/water contact (when pressure is lowered).

The boundary between the petroleum-bearing interval and the water-bearing interval in a field, be it relatively sharp or gradational, is known as the transition zone (Figure 5.15). Within the transition zone there is a downward decrease in oil saturation and a downward increase in water saturation. Although the transition is smooth, the interval can be divided into two distinct zones. Oil can be produced from the upper zone because the petroleum saturation is above the irreducible oil saturation. In the lower zone, the oil is immobile and only water can be produced from here. At a short distance below the observed OWC is the free water level. The relationship between the free water level and the OWC is best illustrated by a laboratory experiment using capillary tubes of different diameters, dipped into a dish of water (Figure 5.16). The wetting phase (water) rises highest in the narrowest tube. Thus, while the free water level is a property of a reservoir system, an OWC is a local phenomenon that is dependent upon the capillary pressure threshold near the wellbore. Fortunately, in high-quality reservoirs, the difference in elevation between the OWC and the free water level is small. However, in poor-quality reservoirs, care must be taken to ensure that differences in elevation in OWCs, measured in different wells, do reflect differences in elevation of free water level and not simply variations in rock properties.

In the absence of a direct identification of a petroleum/water contact in a well, the depth to a contact can be calculated from measurements of pressure data in both the oil or gas legs and in the aquifer. The basic premise is that at the OWC the pressure in the oil zone must equal that in the water zone. In reality, of course, this statement of equilibrium defines the free water level. The pressure in the water (P_W) at the OWC is given by

where C₁ is a constant that represents any degree of overpressure or underpressure.

At depth X_OWC, P_W(OWC) = P_O(OWC). Above the OWC the pressure in the oil leg is simply a function of the pressure of the oil at the contact minus the density head of the oil. Thus estimation of an OWC can be made by extrapolation of the oil and water gradients to the point of crossing (Figure 5.17). The pressure data can be gathered from Repeat Formation Tester (RFT) analysis and well tests. The fluid density data can be measured from formation samples or calculated from wireline logs. The saturation information can be measured from both core and logs, while capillary pressure data are taken from core measurements.

5.3.2 Intra-Field Variations in Petroleum Composition

Almost all pools of oil and gas are heterogeneous with respect to the composition of their petroleum and associated water. An important part of the appraisal process is to determine how the compositions of petroleum and water vary both across a field and with depth in a field. Knowledge of compositional variation in a field will enable the geoscientist to understand how, and possibly when, the field filled with petroleum and the direction of filling. If the field is to be developed, the compositions of both petroleum and water together with their compositional ranges and spatial distribution will be used as a basis for the facility’s design. Variations in fluid composition also have implications concerning field segmentation (compartmentalization), reservoir simulation, satellite field development, and even unitization.

The properties that commonly vary across fields (Archer and Wall 1986) are the gas to oil ratio (GOR), the condensate to gas ratio (CGR), the oil viscosity, the gas composition, and the content of nonhydrocarbon compounds, such as nitrogen, carbon dioxide, and hydrogen sulfide.

An example of the way in which compositional differences can influence the utility of facilities comes from the Gyda Field in the Norwegian North Sea. The Gyda platform was designed to handle petroleum with a GOR of about 1000 scfbbl⁻¹, defined by the petroleum composition at the field crest. However, one of the early development wells penetrated the southwestern corner of the field, where the GOR was twice that of the crest. Fortunately, the original design of the facilities included the capability to deal with up to 15% more gas than would be expected if the field had a homogeneous GOR of 1000 scfbbl⁻¹. The high GOR from the one well in the southwest of the field pressed the facilities to the limit.

Segmented (compartmentalized) fields (Section 5.4) need special attention when a development plan is conceived. Each segment will require at least one production well and possibly dedicated injection wells if recovery is to be optimized. In many instances, compositional variations in the reservoir fluids give the first clue to the existence of intra-field barriers. Such data may be collected during appraisal, before costly development decisions are made. The Judy Field, Central North Sea has multiple fluid compartments, some black oil, some high volatile oil and others gas condensate (Swarbrick et al. 2000). Most of the field shares a common overpressured aquifer, but faulting leads to hydrocarbon compositional differences during a complex oil and gas filling history.

Reservoir simulation models (Section 5.8) require that a field be divided into regions (volumes) of relatively simple composition, because a reservoir simulator needs to start with an equilibrium pressure–fluid system. Black-oil simulators allow gas and oil mixing under equilibrium P–V–T assumptions. So although it is true that they cannot handle complex fluid gradients, they do routinely handle simple ones. Compositional simulators are capable of handling more complex fluids; however, they are difficult to use.

Gas, condensate, and oil rarely have the same economic value. In fields where the GOR, or the gas to condensate ratio, vary spatially, one part of a field may be more or less valuable than another part of a field. Where such different parts have different owners, fluid compositional differences can become a major part of unitization discussions (equity battles).

The major factors that can cause petroleum fluid variations in a field are filling of reservoirs from one direction, biodegradation of oil in reservoirs, seal failure and vertical leakage, gravitational segregation, and thermal segregation.

For a field filling from a single direction, the end of the field closest to the maturing source rock will receive ever more mature petroleum as time progresses (Figure 5.18a). The maturity of the petroleum will be marked by an increased GOR and a decreased viscosity, and may be accompanied by changes in the content of nonhydrocarbon gases. Fields filled with petroleum from more than one direction and from more than one source rock will also display lateral variations in petroleum composition (Figure 5.18b).

Cool (<70 °C), shallow oil pools are prone to biodegradation, which occurs when bacteria-laden near-surface water comes into contact with the oil. The bacteria attack the short-chain hydrocarbons, and as a consequence the most heavily affected area of the field has the lowest oil gravity, the lowest GORs, and the highest viscosity (Figure 5.18c). The Quiriquire Field in eastern Venezuela (Chapter 4) is a classic example of a field in which petroleum composition variation is a product of biodegradation.

Fields that have been filled from other accumulations, either by spillage or seal failure, commonly show systematic variations in fluid composition (Figure 5.18d). Such fractionation of petroleum is common in oilfields offshore southeastern Trinidad and the Gulf of Mexico (GOM; Thompson 1988). In both areas, the seals are poor, and in consequence gas and light hydrocarbons can leak upward, leaving heavier petroleum behind.

A static column of petroleum, such as that in an oilfield, is affected by the Earth’s gravity. The denser components (molecules with more than six carbon atoms) sink and the less dense molecules, such as methane, rise. Such segregation can often be seen by changing gas to oil and CGRs up and down a petroleum column. Creek and Schraeder (1985) illustrated such a phenomenon in the East Painter Field of Wyoming (Figure 5.19).

The magnitude of gravitational segregation has proved difficult to model from thermodynamic considerations. In general, however, the process seems to be most pronounced in deep, hot, condensate- or gas-rich oil reservoirs. Under such near-critical conditions, gravitational segregation can lead to situations such as that in the Brent Field, UK continental shelf (Tollas and McKinney 1988), in which no distinct gas/oil contact can be observed. The petroleum changes gradually. Gas occurs at the top of the pool, overlies condensate which in turn overlies oil.

Thermal gradients may also induce compositional segregation within petroleum columns. However, the effects are difficult to model and in the real world difficult to differentiate from those of gravity. Modeling performed on the gas cap of the giant Prudhoe Field in Alaska suggested that thermal convection within the laterally extensive gas cap could have led to evaporation of the formation water from parts of the cap. This has produced anomalously low water saturations in some areas of the gas column (pers. comm., W.A. England 1992).

In a field without barriers, the rate of lateral mixing (diffusion and convective overturn) is insufficient to homogenize petroleum composition in all but the oldest fields. For a medium- to light-gravity oil, diffusion can homogenize a 100 m oil column in about 1 million years for molecules with up to 200 carbon atoms. In contrast, 40 million years would be required to homogenize methane content over a typical well spacing of 2000 m (Table 5.1), and larger molecules would require hundreds of millions of years. It is not that there is any physical difference between the horizontal and vertical diffusion processes but, rather, that the distances involved are larger and the time taken increases as the square of the distance.

Density-driven mixing operates much more quickly than diffusion mixing. If, for example, a field is filled from two directions by oil of different compositions and hence densities, the denser oil will sink to the bottom of the reservoir while the less dense oil will rise to the top. The rate of this process depends mainly upon the viscosity of the oils, although horizontal, and vertical permeability are also important. For a light to medium oil, density-driven mixing can go to completion in 1 million years or less over typical inter-well distances of a few kilometers. More viscous oils will mix more slowly.

5.3.3 Intra-Field Variations in Water Composition

The composition of water, like that of oil, may not be constant in and around an oilfield. This is true for the water trapped within the petroleum leg of the field and that in the underlying aquifer. Variations in the chemistry of oilfield waters result from water and rock interaction (Warren and Smalley 1994). At the time of deposition the interstitial water in sediments will be fresh water, marine water, evaporitic water, or some combination thereof. Such water may begin to be buried with the sediment.

Although many formation waters are derived from meteoric water, few are fresh. The salinity of water and the specific composition of the dissolved salts are controlled by the extent to which the water reacts with the enclosing rock. The rate at which the formation water reacts with the host rock can be slowed by the presence of petroleum, simply because the water becomes trapped and immobile as the petroleum saturation increases. Naturally, such a process does not occur within the aquifer to a petroleum accumulation and, as such, the “water leg” can continue to evolve. In consequence, water trapped within the oil or gas legs of a petroleum accumulation may have a different composition and hence salinity than the surrounding aquifer. In a continuously evolving system, it is likely that the aquifer would be more saline than the water trapped in a petroleum leg. However, it is also possible that following accumulation of petroleum pools, wholesale exchange of aquifers can yield water legs that are less saline than associated water within the petroleum legs.

It is important to measure the salinity and composition of trapped water and aquifers for a number of reasons. Precipitation of mineral salts from formation waters can lead to damage of the reservoir and production pipework (Chapter 6). The salinity of oilfield formation waters is also utilized to calibrate the resistivity logs used for the determination of petroleum saturation (Chapter 2).

In an example from the North Sea, it was discovered late in the life of the field that the aquifer was more saline than the water trapped within the oil leg. Since the aquifer salinity had been used to calibrate the resistivity logs, the water saturations in the oil leg had been underestimated. That is to say, there was less oil in the field than had been calculated.

In the above paragraphs, we have examined the bulk chemical differences between the water chemistry of the oil and water legs in an oilfield. During the 1990s, the development of the technique known as residual salt analysis (RSA, Chapter 2) allowed characterization of formation waters in great detail using core.

The ⁸⁷Sr/⁸⁶Sr isotope ratio of the formation water can be measured from dried core chips simply by washing the evaporated salts from the chips with distilled water (Smalley et al. 1995). The strontium isotope ratio in the solution can then be measured very accurately using a mass spectrometer. The advantage of using an isotope ratio rather than a bulk chemical characterization is that isotope ratios are significantly less affected by disturbance of the water–rock system by drilling, coring, and extracting the samples from the Earth than is the bulk chemistry. The drilling and sampling process will affect the Eh, pH, pressure, and temperature of the formation water, all of which can lead to reaction, precipitation, and dissolution. These processes will affect the bulk strontium composition, but not the isotope ratio.

Once it became possible, using the RSA technique, to characterize the formation waters in great detail, variations in water chemistry within oilfields, and indeed within individual reservoir units, became apparent. The high degree of variability of the isotope ratio within oilfields has in some instances been used to explain the way in which the field filled with petroleum (Figure 5.20). These same data can be used to describe field segmentation (Section 5.4).

5.4.1 Introduction

Rarely is a petroleum field a simple tank – one that could, given sufficient time, be drained with a single well. Most fields are segmented. Fields may contain barriers to flow of fluids either laterally along the stratigraphy or vertically across the stratigraphy. Some fields may be segmented in both ways. Moreover, barriers to flow may not be totally sealing. The flow restrictions may be no more than baffles. During appraisal of a field it will only be possible to examine barriers under virgin conditions. Some barriers may change their properties during production as fluid is extracted from the field and pressure changes occur.

In this section we will examine the geometry of baffles and barriers, and how to identify their presence and location.

5.4.2 Barriers to Lateral Flow

Faults are the most common barriers to lateral flow of petroleum from a field (Figure 5.21), although not all faults seal, and even sealing faults may not always have sealed. Faults may seal under the following conditions:

• Where throw has juxtaposed reservoir and nonreservoir horizons;

  

  
  
• where fault zones are cemented;
  
• where clays have been smeared along the fault plane;
  
• where cataclasis of the rock along the fault plane has produced granulated zones of impermeable rock.

Faults are most commonly identified during interpretation of seismic data. More rarely, faults are observed from wireline log and core data. In order to determine whether a fault is likely to be a juxtaposition seal between segments of an oil- or gasfield, it is necessary to map the intersection between the footwall and hanging wall along the plane of the fault (Figure 5.10). Alternatively, the stratigraphy of the faulted block can be plotted against displacement (Figure 5.22). Such diagrams, developed by Knipe (1997), allow calculation of the sealing capacity of juxtaposition and clay smear faults.

Faults that have associated cemented zones are common but notoriously difficult to identify in the subsurface. Incipient, inactive, and active faults modify the permeability fabric of the host rock and induce stress anisotropy. Such modification causes, at the very least, flow diffraction for moving fluids in the area of the fault. Indeed, many faults act as fluid flow conduits – a process that may itself cause cementation.

Barriers to lateral flow are also present as lateral shale-outs of reservoir lithologies. Such changes in lithology may not always be obvious from the limited quantity of data commonly available during the early stages of appraisal. This is particularly the case in reservoir systems with a complex reservoir architecture. Well-to-well reservoir correlation exercises may fail to recognize the true relationships between the reservoir bodies.

5.4.3 Barriers to Vertical Flow

Here, we are concerned with barriers and baffles that impede flow upward and downward across the stratigraphy. This is commonly referred to as vertical flow, although clearly if the reservoir was vertical it would be necessary to take care with nomenclature to avoid confusion.

Barriers to vertical flow are essentially similar to those already described for horizontal flow. They include low-angle faults (thrusts), cemented layers, bitumen layers, and of course the natural stratigraphic composition of the gross reservoir interval (Figure 5.23).

A significant difference between barriers to lateral flow and those to vertical flow is that the latter are commonly penetrated by wells. For that reason, it is often easier to identify potential barriers to vertical flow during appraisal.

5.4.4 Identification of Flow Barriers

We said in opening this section that seismic data are used to map faults, but that faults may or may not be responsible for segmentation of a petroleum accumulation. However, there are a number of tools that can be used to identify the presence of barriers before the field is committed to development and hence major expenditure. Conventionally, data on intra-field barriers have been gained from the analysis of well-test information. A detailed discussion of well-test analysis is beyond the scope of this book. However, the basic methodology is to examine the production behavior of a well as parameters such as pressure drawdown and choke size are changed. Measurements are made of production rate for all fluids (gas, oil, and water). The length of a production test can vary from a few hours to many months; if circumstances and cost permit, the longer the test the better as far as gathering useful information is concerned. A long test at an offshore location may not be feasible because of the cost of the drilling and test facility, and because the petroleum fluids cannot be exported and sold to pay for the test. Flaring may also be a problem because of environmental issues and associated taxes.

Probably more important than the test is the recovery of the reservoir afterwards. The speed and manner of the recovery tell the reservoir engineer a great deal about the local continuity of the reservoir, as well as the approximate distance to baffles and barriers. In order to gather these data, pressure meters are placed in the well and the pattern of pressure buildup recorded. In order to maximize the useful information from such tests, it can be important to “listen” to natural pressure changes in the reservoir before the system is disturbed by the test itself. This can help to filter out unwanted information, such as the effects of lunar and Earth tidal signals.

Most of the geologic tests designed to examine compartmentalization rely on analysis of some component of the fluid system. Measurements can be made on either the petroleum or aqueous components. The basic premise of many such tests is that differences between samples taken from different wells in the same field or different reservoir intervals in the same wells may indicate the presence of barriers. The types of fluid differences that can be observed have been covered in Section 5.3 and the analytic methods introduced in Chapter 2.

5.5.1 Introduction

Reservoirs and reservoir properties are heterogeneous when viewed on a range of scales from the pore to the play fairway (Figure 5.24). Field appraisal provides the opportunity to begin to understand the range of reservoir properties that exist within the discovered petroleum pool. In this section, we will examine the basic description of reservoirs, lithofacies, and lithotypes, and the way in which the lithofacies are joined (reservoir geometry and reservoir correlation).

5.5.2 Lithofacies and Lithotypes

The concept of the “facies” was first used by geologists and miners in the nineteenth century (Reading 1986), who recognized that rock units with particular features were useful in correlating coal seams and ore bodies. Gressly (1838) was the first to use the term. Reading (1986) reports that the term and its usage have been the source of much debate. Here, we will limit ourselves to discussion of the derivative term “lithofacies,” which is used to describe the physical and chemical characteristics of a rock. The criteria that can be used to describe a lithofacies include mineralogy, grain size, sedimentary structures, color, and fossil assemblages. Elsewhere in this book the terms (wireline) log facies and seismic facies are also used. These too are descriptive terms, used to indicate comparable wireline log responses and seismic responses respectively. We avoid terms such as turbidite facies, shallow-marine facies, and post-oregenic facies, all of which are to a greater or lesser extent interpretive.

A lithofacies is the basic unit of outcrop or core description. The physical characteristics of a lithofacies commonly allow some interpretation of sedimentary process, from which may be derived some indication of depositional environment. For example, a lithofacies comprising sandstone with climbing ripples indicates deposition from an unidirectional current. Thus the process is known, but such current activity could occur in a variety of depositional environments, from fluvial to deep-marine. Groups of lithofacies that occur together are commonly termed “lithofacies associations.” Each lithofacies can be interpreted in terms of a sedimentary process and together the different sedimentary processes enable better definition of the depositional environment than does a single lithofacies. For example, a commonly associated group of lithofacies is laminated mudstones, graded and laminated sandstones, and chaotic mud-supported conglomerates. The individual processes for each of these lithofacies could be interpreted as deposition from suspension (mudstones), turbidity current (sandstones), and debris flow (conglomerates). Taken together, these processes are only likely to have occurred in a small number of possible depositional environments, such as deep lakes and deep seas.

In progressing from the description of lithofacies to the interpretation of process and eventually depositional environment, care must be taken to understand the boundaries and thus the relationships between lithofacies. In core or outcrop, the boundaries between lithofacies can be gradational, sharp, or erosive. Only where boundaries are gradational can the vertical succession of lithofacies be unequivocally regarded as indicative of laterally coexisting depositional systems (Walther 1894). Where the boundary between lithofacies is clearly erosional, it may be possible to interpret the degree of induration of the underlying lithofacies, but it may not be possible to determine what has been eroded or not deposited at the lithofacies boundary. Sharp boundaries between lithofacies could represent either conformable or erosional surfaces.

The interpreted lithofacies and lithofacies associations are the foundation for the reservoir description. The sedimentologist uses the lithofacies data together with analog information to construct a picture of the reservoir in terms of interlocking “geobodies.” During the appraisal of a field, the description of the individual lithofacies may not change significantly as more wells and cores are added to those obtained from the exploration wells. However, the understanding of the spatial relationships between lithofacies will develop significantly and the description will change from a statistical representation of lithofacies distribution to a more deterministic description.

Before considering reservoir geometries, we must first make a small detour. The lithofacies and lithofacies associations are the basic units for the geologic description of the reservoir; however, they may not be adequate descriptive elements for how the reservoir will behave during petroleum production. The derivative term “lithotype” has been invented to enable characterization of the reservoir in terms of how it is likely to perform under production. The main difference between the lithotype and the lithofacies is that the definition of the lithotype is based upon the permeability characteristics of the rock rather than the full suite of physical and chemical properties. Moreover, from a reservoir performance perspective, lithofacies with moderate to high permeability require more detailed characterization than those with low permeability.

In a typical lithotype subdivision of a reservoir, a single lithotype will have a range in permeability of one order of magnitude or less, but lithologies with less than about 1 mD permeability are commonly lumped together into a single lithotype. In practice, for lithofacies with moderate (1–100 mD) permeability there will be a one-to-one mapping between a lithofacies and a lithotype. For the best lithofacies in a reservoir, the permeability may be hundreds of millidarcies to tens of darcies. If this is the case, it is necessary to subdivide these lithofacies into several lithotypes of, say, 100 mD–1D and 1D–10D. The lithotypes, rather than the lithofacies, will be used in the reservoir simulation.

5.5.3 Reservoir Body Geometry

The relationship between field size and reservoir body size and geometry was introduced in Chapter 4. The small size of individual alluvial fans and pinnacle reefs was recognized as the likely defining size for trap envelopes in such systems. In contrast, large-scale, sand-rich submarine fans were considered to be larger than the footprint of the largest of structurally trapped petroleum fields. The complex stratigraphic trapping geometries of paralic systems were seen as indicative of the interplay between sand-body geometry and trap configuration occurring on similar scales. The difference between exploration and appraisal is that we now need to examine the heterogeneity at the next scale down. At its most simple, the geometry of reservoir bodies can be considered to approximate to one of two basic forms: highly flattened oblate and prolate spheroids. The vernacular descriptions – pancakes and ribbons – probably convey their shapes better to the reader than do the mathematical terms.

Many geometrical parameters could be used to characterize such reservoir geometries. However, we must be careful to use only those that we can be expected to be able to measure and to simulate using analog data. In that respect, the two most commonly used are thickness and width. Thickness can be measured from core and perhaps wireline logs. Width cannot be measured from either of these, although in exceptional circumstances it may be possible to estimate width from seismic data. Whether or not width can be measured, it is useful to be able to understand the relationship between the two parameters for different geobodies deposited in different environments. The geobodies commonly referred to in such geometric classifications are sand bodies, beds, lithotypes, and shales. The sub-lithotype heterogeneity permeability correlation is discussed in the following section.

For different geobodies, there is commonly a correlation between thickness and width (Figure 5.25). The numerical relationship between thickness and width for a given geobody is referred to as the aspect ratio.

Although they are important for reservoir description and reservoir modeling, length-to-width relationships are much more difficult to obtain and estimate. For example, it is commonly possible to make many measurements of channel thickness to channel width from outcrop, for ancient fluvial systems. Only in exceptional circumstances, where exposure is complete and structure conducive, is it possible to measure lengths. Modern analogs can be used in such circumstances, and there has been some success at flume modeling of braided fluvial systems.

Much work has been done over the past decade on the aspect ratios of sand bodies, beds, lithotypes, and shales, and from the statistical data generated it is possible to make a few generalizations. The aspect ratios for sand bodies and beds have a much greater range (typically between 10 : 1 and 10 000 : 1) than do those for lithotypes and permeability correlation (typically 1 : 1–20 : 1). This is not too surprising when one considers that the lithotype, which is derived from the lithofacies, is commonly a product of a distinct physical process, rather than an agglomeration of processes. It also gives hope insofar as the bewildering array of sand-body geometries can be modeled by considering their much more predictable component parts. Aspect ratios of sand bodies, beds, and shales are given in Table 5.2.

Fewer data have been collected on deep-marine and carbonate systems than on fluvial to shallow-marine clastic systems. For deep-marine systems, at least the few data demonstrate aspect ratios that are specific to the system from which they were collected. The lack of global patterns could simply reflect the paucity of the numerical data currently available. Shale width to thickness ratios are shown in Figure 5.26. Aspect ratios vary from 2 : 1 to about 10 000 : 1 for thicknesses between 0.01 and 2 m. In Figure 5.26, there appear to be distinct differences in aspect ratio and width for different depositional environments: fluvial, shallow-marine, and deep-marine.

5.5.4 Reservoir Correlation

The concept and utility of stratigraphic correlation was introduced in Chapter 2 and some of the applications were examined in Chapter 3. The discussion so far has been one of understanding basins and play fairways. However, at the appraisal stage the problem of correlation can become a greater rather than lesser problem compared with what one might suppose. The need is to correlate penetrations of the reservoir in wells that are commonly only a few kilometers apart. The aim is to obtain a correlation between wells that will honor the flow units in the reservoir. The correlation will control the geometric arrangement of the reservoir units used in the reservoir simulator. The simulator will be run and the development scheme and well locations chosen on the basis of the results. Naturally, if the well-to-well correlation used in the simulator is inappropriate, then the reservoir will not behave as the simulation predicts. There may only be minor differences between the model and reality, or there may be large differences. This happened with the Clyde Field (Figure 5.27), where there was sufficient mismatch that development of the field was suspended.

At the appraisal stage, biostratigraphy can remain an important tool, but in many instances identification of reservoir flow units will be beyond the resolution of biostratigraphic methods. In reservoirs that are barren of biostratigraphic material, chemostratigraphy, or magnetostratigraphy may be used.

Sequence stratigraphy (Section 3.6.6), worked within a biostratigraphic, magnetostratigraphic, or chemostratigraphic framework, commonly provides the most powerful reservoir correlation tool. The ease or difficulty encountered in using sequence stratigraphy to solve an appraisal-scale correlation problem depends upon both the well spacing and the depositional environment of the reservoir interval. If the well spacing is larger than the size of the individual elements that make up the reservoir, then accurate correlation will prove difficult. The same will be true if the reservoir interval is highly cyclic and the cycles are all much the same.

The fundamental approach to correlation based on sequence stratigraphy is to use the wireline log responses to define depositional trends. Care must be exercised when doing this to ensure that the trends are products of depositional changes and not due to fluid changes (petroleum versus water). The suite of commonly identified patterns on gamma, spontaneous potential (SP), resistivity, and sonic logs is illustrated in Figure 5.28. It is common, then, to identify maximum flooding surfaces and maximum regression surfaces. The wireline log responses are then used to define the systems tracts between the flooding and regression surfaces (Figure 5.29). Each well is subject to the same treatment, and the combined well logs plus sequence stratigraphic interpretations are used as the basis for inter-well correlation (Figure 5.30).

Correlation based upon rock properties using the above methods may be refined by incorporating data derived from the analysis of both petroleum and water (Sections 5.3 and 5.4). In addition, use of pressure data, specifically where data from overpressured reservoirs in multiple wells plot on common fluid gradients, can be used as indications of reservoir and fluid connectivity.

5.6.1 Introduction

At the pore scale, reservoirs can differ dramatically in their quality. Quality may be measured as either the capacity of the reservoir rock to hold fluids (porosity) or its capacity to transmit them (permeability). We have already covered the basics of porosity, permeability, and fluid saturations in Chapter 4. In this section, we examine intrinsic reservoir properties further. We also look at the post-depositional processes that can modify both porosity and permeability within potential reservoirs: compaction, cementation, and mineral dissolution.

5.6.2 More Intrinsic Reservoir Properties

Porosity, permeability, grain density, and water saturation are measured on core plugs during routine core analysis. Such data, though of prime importance in determining the storage volume and flow potential of a reservoir, do not yield all of the information about how a well or field will behave on production. Further data are generated from a combination of special core analysis and from petrophysical analysis of the wireline logs. We are now at the interface between petroleum geoscience, reservoir engineering, and petrophysics. Nonetheless, it is useful to examine briefly some of these aspects of petroleum science, since they form part of an appraisal program.

5.6.2.1 Relative Permeability

Absolute permeability to air is measured during routine core analysis. Such figures need to be corrected to reservoir conditions in terms of a compaction correction (Figure 2.19), yet such corrected data are still only valid for single-phase flow. A consequence of there being more than one fluid in the pore system is that neither water nor oil will flow as readily as if there were only one phase (Figure 5.31). This is known as relative permeability. As a conceptual device, think of relative permeability in terms of a group of joggers running down a street. If the street is empty, the joggers will be able to move along at a pace determined only by their fitness and capabilities. In a street filled with busy, nonjogging phase (shoppers), the transit time (the relative permeability to joggers) will increase massively as shoppers get in the way. In a similar way, the microscopic viscous drag forces between the oil, water, and mineral surfaces act to reduce the permeability to one of the phases. The importance of accurate determination of relative permeability is high. The relative permeability data obtained during appraisal will be used in the development design process. In detail, such data will be used to calculate the standoff (vertical separation) between production wellbores and petroleum/water or gas/oil contacts, and in calculations associated with the pressure drawdown that will be placed on production wells.

Relative permeability is measured in the laboratory through experimentation. The experiments are run at reservoir temperatures and pressures, using a combination of real or simulated formation water and oil. Sequences of oil and water are flushed through the samples of preserved core, and the relative permeabilities of the two phases at a variety of partial saturations are measured. Relative permeability is highly sensitive to both the detailed rock type (the lithofacies) and the fluid types. Moreover, the transformation of relative permeability data obtained on core plugs or whole core to reservoir-scale behavior carries large uncertainties.

The quality of the relative permeability data will be revealed during field production. Field-scale relative permeability data can be calculated once the undesired phase (usually water or gas) starts to break through into the (oil) production wells. Naturally, one’s hope is that water breakthrough will occur either when or later than predicted using the relative permeability measurements made on core during appraisal.

5.6.2.3 Resistivity, Cementation, and Tortuosity Factors

In Chapters 2 and 4, we introduced wireline logs and described the data that can be derived from them. Petrophysical analysis of wireline log data involves quantification of the properties outlined previously. The key numerical derivatives are the formation resistivity factor (F), the cementation factor (m), and the tortuosity factor (a).

The formation (resistivity) factor is the ratio between the resistance of a fixed volume of formation water and the same volume of formation water plus formation (i.e., nonconducting rock). The unit of measurement is the ohm-meter² per meter. This is commonly abbreviated to “ohm-m” and in practice the measurements are made in milli-ohms.

Naturally, the more any particular rock becomes cemented and loses porosity, so the more the formation resistivity factor increases. The link between F and porosity has been found from experimentation to be a power function corresponding to the form:

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Unconventional Petroleum, Gas Storage, Carbon Storage, and Secondary Products
2. Development and Production
3. 7 ▪ Phase transitions and critical phenomena
4. 8 ▪ Scaling theories and the renormalization group