A systematic procedure for the petrophysical identification and interpretation of low-resistivity and low-resistivity-contrast pay zones in intergranular reservoirs is founded upon an analysis of case histories for different reservoir types in diverse areas of the world. The approach acknowledges that a reservoir rock is a coupled physico-chemical system. The proposed method is generic and robust, it is conceptually simple, and it is structured in a manner that is easy to understand. The scheme is modular and it is arranged hierarchically to reflect maturing data scenarios: therefore, it can be progressively refined during the appraisal and development stages. The essence of the method is the definition and calibration of reliable interpretative procedures through quality-assured reference data from key wells by admitting only validated reservoir characteristics. Examples world-wide illustrate how failure to recognize low-resistivity pay can result in much loss of potential value.

A principal thrust is to facilitate the re-evaluation of other wells within the same reservoir system without the need for excessive acquisition of additional data. However, the proposed interpretation framework does allow the incorporation of new logging technology as this becomes established. The end-product is a flexible petrophysical interpretation scheme for these unconventional reservoirs that benefits from cost-effectiveness, portability, a higher degree of exactness and consequently a much reduced uncertainty.


In unfractured reservoirs, low-resistivity pay zones are usually associated with one or more of the following: laminated reservoir/non-reservoir sequences, formations with multimodal pore-size characteristics, sediments with anomalously high surface area, and reservoir components that extend beyond the range of applicability of interpretative algorithms, e.g. electronically conducting minerals. In all these cases, hydrocarbons have been produced with little or no water cut in the presence of high interpreted water saturations, in many different parts of the world. It is therefore important to have a generalized facility for recognizing low-resistivity pay as early as possible in the life of a prospect.

The problem of identifying low-resistivity pay through wireline log analysis has been acknowledged for over 30 years, with much of the early focus on the Texas and Louisiana Gulf Coast of the United States (Tixier et al. 1968), although not exclusively so (Murphy & Owens 1972). During the past three decades an increasing number of examples have come to light in the Gulf Coast region (Table 1). Today there are also documented records of low-resistivity pay in other parts of the conterminous United States, Alaska, Brazil, Canada, Venezuela, Argentina, the North Sea, continental Europe, north Africa, the Middle East, India, southeast Asia, Japan, China and Australia (Table 1). Indeed, the related petrophysical literature has become voluminous during the past ten years (Table 1), presumably because the market-driven requirement to add maximum value with minimum cash outlay has re-emphasized the need for the most effective and efficient well-completion programmes. The subject area therefore remains highly topical, both technically and commercially.

Historical background

Low-resistivity pay is not recognizable through conventional log analysis. Historically, therefore, low-resistivity pay has not been targeted for primary completions, being discovered through core and pressure analysis as late as a higher-risk, third completion stage. Much of the locally applicable, empirical rationale behind the initial overlooking and the subsequent discovering of low-resistivity pay has been lost to the industry with the passage of time. However, the parameter cut-off culture virtually guaranteed that all admitted pay intervals would show fairly consistent reservoir properties with low-resistivity pay going undetected, at least initially.

The concept of what constitutes low-resistivity pay has evolved with time. Originally, it encompassed those intervals with formation resistivities as low as 1–3 Ωm and that subsequently turned out to be commercial hydrocarbon producers (Gauntt et al. 1964; Murphy & Owens 1972; Tripathi et al. 1984). On the basis of an increasing number of case histories, the concept has been expanded to include formations with resistivities less than 0.5 Ωm (Zemanek 1989; Boyd et al. 1995), indicating a progressive awareness that reserves were being lost when higher cut-offs were adopted. It is not known how many low-resistivity pay reservoirs have been abandoned prematurely over the years.

Today, it is recognized that there are no universally acceptable bounds to the resistivity of commercial pay zones. Alternative specifications have been formulated in terms of the contrast in resistivity between the hydrocarbon-bearing reservoir and congenital shales (Boyd et al. 1995). Low-resistivity contrast sands have been specifically defined as, for example, having a resistivity that is less than 1.5 times the resistivity of intraformational shales (J. T. Kulha 1998, pers. comm.). This definition allows a range of limiting resistivities that have been higher than 10 Ωm in some areas.

A more pragmatic basis for low resistivity contrast is the difference in resistivity between the hydrocarbon-bearing interval and the water zone. This approach, too, is not tied to fixed limiting resistivities. However, with this philosophy, the concepts of low resistivity and low resistivity contrast can be merged.

Contemporary definition

From a global perspective, ‘low-resistivity pay’ is better taken as a relative term rather than an absolute descriptor. It exists when there is a lack of useful positive contrast in measured electrical resistivity between zones that contain and produce hydrocarbons in commercial quantities and zones that contain and produce only water, within the same reservoir system. Therefore, as used here, the term ‘low-resistivity pay’ also includes low-resistivity-contrast pay zones and it does not distinguish on the basis of absolute resistivity, in which respect this treatment differs from certain others (e.g. Boyd et al. 1995).

The underlying rationale is that an intra-reservoir quantitative comparison of what constitutes low-resistivity pay is more meaningful than an inter-reservoir analysis, for it is strictly within the setting of the former that completion decisions are made. The philosophy also accommodates those encountered situations in which a pay sand within one reservoir system has a significantly lower absolute resistivity than a clearly identifiable water zone within either the same reservoir system or another. Table 2 lists pertinent examples.

Scope of the problem

Although the concept of low-resistivity pay has evolved with time, the problem itself has not diminished. This point is well illustrated by the synergetic papers of Duren (1967) and Montgomery et al. (1998), which describe the Glick gas field of Kansas, where the main reservoir rock is a highly microporous tripoli chert. The low-resistivity pay problem remains centred on the evaluation of water saturation Sw. More specifically, the problem arises through the very high values of Sw that can be interpreted from wireline logs over low-resistivity (contrast) intervals where a reservoir does not conform to the assumptions made during the conventional petrophysical evaluation of clean or shaly formations.

Although low-resistivity pay is evidently a world-wide problem, potential solutions reported in the literature are directed at specific reservoirs or particular depositional environments. To some extent, the documented approaches are exclusive in that they may not be directly applicable to other reservoirs where conditions appear to be similar. Yet, experience also indicates that the problem of low-resistivity pay is not age-, formation-, lithology- or location-specific (Table 1). All this points towards the need to establish a generic approach.

This paper is concerned with setting the low-resistivity pay problem within a global context with the primary objective of formulating a generic strategy for addressing it. The approach is directed principally at extracting greater value from existing reservoir databases, but it should also provide a foundation for the evaluation of new wells wherein more advanced logging tools might be deployed. Therefore, it is especially appropriate to the re-interpretation of data from older wells in the light of contemporary understanding.

Two types of reservoir are specifically excluded from the following discussion; fractured reservoirs, for which focused resistivity logs can read anomalously low, and reservoirs that can be evaluated meaningfully using standard clean- or shaly-sand procedures, regardless of their absolute or relative formation resistivities.

Nature of low-resistivity pay

The nature of low-resistivity pay is partly a function of the response characteristics of wireline logging tools. In the context of tool response, two definitions are important. A layer is detected when a wireline log shows a significant deflection in response to that layer. A layer is resolved when the deflection of the log attains the true parametric value for that layer, after corrections for borehole effects (Fig. 1). Following this, the spatial resolution of a logging tool can be defined as the minimum bed thickness needed for the tool to record the true parametric value for that bed. Table 3 lists the limits of detection or resolution, as appropriate, for a range of wireline logging tools, together with details of the reference sources. The key is to sharpen the spatial resolution of the deeper sensing logs, especially resistivity logs, relative to the thickness of the reservoir layers. The objective is to evaluate the porosity and deep resistivity of the reservoir layers without the process being distorted by adjacent non-reservoir beds.

It is important to distinguish at the outset between the manifestation of low-resistivity pay, i.e. how it appears in field data, and its cause, i.e. the physico-chemical phenomenon that gives rise to its occurrence. From a petrophysical perspective, low-resistivity pay can have one of four manifestations and one or more of six causes (Table 4).


The first manifestation is that of a reservoir rock with a combination of physical properties that collectively result in a low measured resistivity that is indicative of water-bearing strata. The reservoir layers within the hydrocarbon leg may or may not be directly resolvable by wireline logging tools. In cases where they cannot be resolved, low-resistivity pay is taken to include only those formations that cannot be interpreted meaningfully through conventional laminated-sand procedures. Where they can be resolved, low-resistivity pay includes only those formations that cannot be interpreted meaningfully through conventional clean-sand or (dispersed) shaly-sand procedures. This is an important point because, in this context, earlier literature on low-resistivity pay referred to the Archie equations (Archie 1942), but it did not include the more recently developed and multifarious shaly-sand algorithms (Worthington 1985).

The second manifestation is that of a reservoir rock that cannot be distinguished electrically from water-bearing strata within the same reservoir system. The resistivities may not be low in absolute terms, but the resistivity contrast between the hydrocarbon and water legs is small. Again, the reservoir layers within the hydrocarbon leg may or may not be resolvable by wireline logging tools. In cases where they can be resolved, the criterion for this manifestation is applied to the individual reservoir layers.

The third manifestation is encountered where the hydrocarbon and water legs can be distinguished by wireline logs, but the quantitative petrophysical interpretation is grossly pessimistic because the physical characteristics of the reservoir rock extend beyond the range of applicability of the available interpretative models. This outcome may not be a sole consequence of the factors causing the low resistivity.

The fourth manifestation occurs where a (high) water saturation can be evaluated correctly from the (low) formation resistivity, but the interpretation is incompatible with production characteristics, which show dry hydrocarbons or a low water-cut. This outcome is usually the result of high capillarity.


The causes of the four possible manifestations of low-resistivity pay take the form of coupled elements of a physico-chemical system that encompasses rock type, matrix properties and texture, clay-mineral properties and texture, grain size and shape, pore size(s) and pore geometry, and saturating water salinity. The causes can therefore be seen as components of the low-resistivity pay problem. Attempts to understand the problem should take account of all these factors, and of the ways in which they interact, rather than select subsets for consideration in isolation. This point is supported by the examples of low-resistivity pay listed in Table 1. Although those cases have been sorted according to their principal cause, many of them also cite other causes that co-exist alongside the primary. Further, some reservoirs show different causes of low-resistivity pay over different intervals. For this reason, it is especially advisable to group those reservoir intervals with common characteristics as early as possible during a field study.

Table 4 identifies six possible causes of low-resistivity pay, each of which can be seen as a category of the subject area from a petrophysical perspective. The first cause is the occurrence of laminated sand/shale sequences. Thin hydrocarbon-bearing beds can be grouped within low-resistivity (contrast) zones where the geometry of the reservoir layers is such that they cannot be resolved by the logging tools that have been deployed. The nature of the reservoir layers themselves is of secondary consideration in this context. The foregoing illustrates the changing nature of the problem, because as higher-resolution resistivity tools are developed and deployed, some low-resistivity pay sands might become resolvable and cease to be low-resistivity pay.

The second cause is the presence of a low-salinity formation water as the saturating electrolyte. Not only can a low-salinity reservoir lie beyond the range of validation of conventional interpretative algorithms for water saturation, but low electrolyte salinity also allows the shale term to assume relatively large values, which make the accurate quantification of this term especially important.

The third cause is one of electronic conduction within the rock matrix. This phenomenon usually arises where the concentration of a metallic mineral exceeds a critical level, beyond which it can act as a significant additional conductor.

The next three causes, fine grains and internal or superficial microporosity, are all concerned with high capillarity. All three give rise to a high surface area, which can appear as a shale effect, especially if the formation waters are fresh. This last comment emphasizes the coupled nature of the low-resistivity pay problem.

All these six causes are associated with reservoir characteristics. The concept of low-resistivity pay as adopted here does not include cases where borehole environmental effects have artificially lowered the log-derived formation resistivity.

The following treatment of low-resistivity pay is structured in terms of the causes listed in Table 4. The greatest problem is recognizing that low-resistivity pay is actually present. Where that recognition emerges through experience within a particular reservoir, possibly with an element of serendipity, value has already been lost and only a partial restoration of cost-effectiveness is achievable through a recompletion programme. The thrust of what follows is therefore to identify low-resistivity pay as early as possible in the life of a field in order to maximize added value, with the further objective of being able to export remedial interpretative procedures to other wells without incurring disproportionate data-acquisition costs.

Approach to Interpretation

At the outset, reservoir layers will be either resolvable or non-resolvable by wireline logs. If they are resolvable, or they can subsequently be resolved through signal processing, the reservoir layers can be interpreted directly. In this sense, the word ‘direct’ means ‘without reference to a laminated reservoir model’. In the particular case of low resistivity pay, laminated models usually comprise sand/shale sequences, and indirect interpretation would therefore require an appropriate laminated sand/shale model.

Where direct interpretation is possible, this conforms to the structure of Fig. 2, which illustrates four possible avenues of approach to the evaluation of water saturation. A nomenclature of interpretative symbols is included at the end of the text.

The first requirement is to classify the reservoir as an Archie or a non-Archie reservoir. An Archie reservoir satisfies the requirements of low clay(-mineral) content and high-salinity formation water above a limiting water saturation (Archie 1942), and it is conventionally evaluated using classical clean-sand methods. A non-Archie reservoir does not satisfy these conditions, and it is conventionally evaluated using traditional shaly-sand techniques.

Direct interpretation of low-resistivity pay zones has often involved variations on the classical clean-sand and shaly-sand procedures. These variations are termed ‘pseudo-methods’, and they have been applied in both Archie and non-Archie situations, in which contexts they are described as pseudo-Archie and pseudo-non-Archie approaches, respectively. They are used because traditional methods of petrophysical interpretation cannot accommodate low-resistivity pay.

Pseudo-Archie methods entail empirical manipulation of the conventional clean-sand algorithms, primarily through the Archie porosity exponent m and the Archie saturation exponent n, so that these quantities can be set without constraint to strictly functional values that lead to meaningful estimates of water saturation. After adjustment, the exponents become pseudo-Archie parameters.

Pseudo-non-Archie methods involve similar empirical adjustments to the traditional shaly-sand algorithms, primarily through the intrinsic porosity exponent m* and the intrinsic saturation exponent n*. The intrinsic exponents are those corrected for shale effects on the electrical conductivity of the reservoir rock. After adjustment, the intrinsic exponents become pseudo-non-Archie parameters. Pseudo-non-Archie methods also include cases where a shaly-sand algorithm is applied to lithologically clean rocks with a complex pore geometry, such as sands containing granular chert.

Note again that where conventional clean- or shaly-sand procedures can be used, the interpretation exercise moves outside the scope of low-resistivity pay as defined here.

Laminated sand-shale sequences

In sedimentological rock description, the word ‘lamina’ describes a layer that can vary in thickness from 1–30 mm (Tucker 1991). In petrophysical parlance, the adjective ‘laminated’ has been used to describe any alternating sequence of sands and shales that cannot be resolved by the available logging tools. Thus, the term can include layers that are ‘many inches in thickness’ (Dewan 1983). This paper follows conventional petrophysical practice in this respect. The more complicated problems associated with undulating or dipping layers of variable thickness are not considered here.

At the outset, laminations might be suspected but their scale is unknown. It is therefore necessary to quantify the thickness of the smallest significant reservoir layer. This quantification is based on the distribution of all reservoir layer thicknesses within the evaluation interval. If core is available, the distribution can be generated from core photographs. If core has not been cut, the response of high-resolution logging tools holds the key. At the smallest detectable scale and in the presence of a water-base mud, microfocused resistivity, dipmeter and electrical imaging tools have all been deployed successfully. With an oil-base mud, dielectric logs have been the preferred tools. At this stage we are concerned only with bed identification. Tools with a sharp spatial resolution that are ideal for this present purpose might not have a sufficiently large depth of investigation for the subsequent petrophysical evaluation of partially invaded reservoir layers.

Having established that the reservoir is laminated and that the laminations are smaller than the spatial resolutions of the deeper sensing logging tools, the next stage is to enhance those resolutions where possible, so that the corresponding logs might be evaluated directly. Several methods are available and their applicability depends on the ratio of bed thickness to unprocessed tool resolution. The following methods are not exclusive: deployment of high-resolution deep sensing tools if available (Strickland et al. 1987); increased digital sampling to bring out greater detail in the logs (Suau et al. 1984); using the short-spacing measurement of a dual-sensing tool to enhance the spatial resolution of the long-spacing measurement without reducing its depth of investigation (Flaum et al. 1989); and signal enhancement through deconvolution (Lyle & Williams 1986) and forward modelling (Dyos 1987).

Table 3 indicates the minimum bed thicknesses for which direct log interpretation might be feasible. Where a direct evaluation of reservoir characteristics is not possible, even with signal enhancement, recourse can be made to a conventional laminated sand/shale model. In such cases, it is especially important that the interpretation be validated against the saturations of water extracted from low-invasion core plugs cut from reservoir rock (e.g. Dawe & Murdock 1990).

Figure 3 exemplifies the responsive signatures of the SP and gamma ray logs in a laminated reservoir in the Gulf of Mexico Basin, where these logs indicate that the shaly and silty sand laminations vary in thickness from 0.5–2.5 m (Darling & Sneider 1992). The deep induction log cannot detect, let alone resolve, many of these laminations and therefore the recorded deep resistivity is generally an intermediate value averaged over the spatial resolution of the tool, here taken to be about 2.5 m. Although the deep induction log measured an average formation resistivity as low as 0.4 Ωm over the test interval, the latter was tested at 778 BBL of effectively dry oil per day.

Table 5 lists some case histories of laminated sand/shale sequences together with the bed thicknesses encountered, the wireline logging tools used to identify the beds, and the benefits derived in terms of interpretation enhancement. Note that a solution to this particular aspect of the problem can take the reservoir out of this low-resistivity pay category, although it might then enter another category.

Fresh formation Waters

In this and the following sections, it is assumed that direct evaluation of wireline logs is possible, i.e. the reservoir layers are sufficiently thick to be resolved by conventional logging tools. It is also presumed that standard procedures for formation evaluation do not allow the interpreter to distinguish between pay and non-pay. If they do, there is no problem to address in the context of low-resistivity pay.

Fresh formation waters can render hydrocarbon-bearing layers indistinguishable from water-bearing zones. Further, they take a reservoir out of the range of formation-water salinities that were used to calibrate the published algorithms for the petrophysical evaluation of water saturation from resistivity logs, and these algorithms might therefore be inapplicable. Most interpretation failures have arisen from the direct application in unmodified form of the Archie clean-sand equations (Archie 1942), but there is a view that even the more comprehensive shaly-sand models are not suited to the evaluation of low-resistivity pay in this category (Sneider & Kulha 1996). There is no general limiting salinity below which standard interpretation methodologies break down. Indeed, reservoir rocks and their interstitial waters show a continuum of electrical behaviour, whereby a given interpretative algorithm can cease to be valid at any point within a range of formation-water salinities, depending on the electrochemical properties of the porous medium (Worthington 1995). Having stated this, most problematic situations do arise when the formation water salinity is less than about 15 000 ppm equivalent NaCl. Reservoirs that contain fresh formation waters are described as ‘freshwater reservoirs’ for the purposes of this paper.

The key issue is whether or not the available shaly-sand algorithms can function at the prevailing formation-water salinity within the subject reservoir. If they do, the problem is solved, because the interpretation will distinguish between pay and non-pay: if they do not, an alternative procedure must be used. In some freshwater porous media, shaly-sand algorithms do have an application, albeit in modified form (e.g. Guru et al. 1995), but in other reservoirs they can break down even at significantly higher salinities (Diederix 1982).

There are several potentially useful approaches to validating the performance of an interpretative algorithm for water saturation. The first two require core data. If low-invasion coring has been undertaken with the quality assurance of a tracer additive to the drilling mud, a comparison of extracted formation waters with log-derived water saturations offers the soundest approach to validation (Fjerstad et al. 1993). If native-state water saturations are not available, another approach is to benchmark the predictive performance of interpretative algorithms at irreducible water saturation, where the latter has been obtained independently, e.g. from capillary pressure studies (Guru et al. 1995).

If there are no core data and the reservoir contains crude that is immobile to flushing by the mud filtrate, a comparison of shallow-sensed and deep-sensed water saturations can confirm or otherwise the validity of a predictive algorithm in the undisturbed zone (Wharton & Delano 1981). Finally, and more generally, a comparison of predictive performance in the flushed and undisturbed zones of the water leg constitutes the most basic option. If, in the last two cases, the validation is approached by applying the same shaly-sand algorithm to both the flushed and the undisturbed zones, rather than by applying a different measurement technique in the flushed zone, a prerequisite is that the mud filtrate and formation water are dissimilar.

If the above methods confirm that the available predictive algorithms have broken down, recourse to a pseudo-Archie approach is the most common. This technique uses the basic Archie equations for the evaluation of water saturation, but the exponents that characterize these equations are allowed to assume very different values from the values m = n = 2 proposed by Archie (1942). The numerical values of these exponents are often chosen so that the algorithms lead to a correct prediction of water saturation either based on core extraction or at irreducible conditions. The same philosophy has been applied to shaly-sand equations, but less frequently. An alternative method that might be attempted where the mud filtrate is highly saline is to draw upon a broad relationship between the predicted water saturation in the flushed zone (where conventional interpretation might work) and that in the undisturbed zone (where it does not). If successful, this method by-passes low-salinity effects in the reservoir (Spalding 1984).

Figure 4 indicates broadly the ranges of formation-water resistivity Rw and shale conductivity term BQv within which the Archie equations and shaly-sand algorithms of the type of Waxman & Smits (1968) are likely to be valid. Outside these ranges, the methods might break down. Therefore such a chart offers an overall guide to the recognition of the low-salinity aspect of the low-resistivity pay problem in the total porosity system of petrophysical interpretation. An equivalent chart can be constructed for the effective porosity system, for example through the modified Simandoux equation (Bardon & Peid 1969), by using as the shale conductivity term the product of wetted shale volume fraction Vsh and wetted shale conductivity Csh. Note once again that if standard clean- or shaly-sand methods of interpretation turn out to be valid, fresh waters alone do not cause a low-resistivity (contrast) pay problem.

As an extreme example, Fig. 5 shows wireline logs run through the Miocene Tipam Sands of the Lakhmani Field of Upper Assam in northeastern India (Jain 1990). These fluvial sands contain formation waters of 1000–2000 ppm equivalent NaCl, and this salinity corresponds to a water resistivity of about 2.5–5.0 Ωm at 25°C. The high water resistivity gives rise to ratios of apparent to intrinsic formation factor Fa/F* that are generally less than 0.5 and to a shale conductivity term that lies within the range 0.2–0.4 S m−1. Therefore the reservoir characteristics extend beyond the region of application of conventional interpretative models for water saturation, encroaching into the pseudo-model area of Fig. 4. At these low salinities, where surface conduction can become highly significant even in the absence of clay minerals, changes in rock texture can have a profound influence on the formation resistivity. Through the surface-conduction effect, fine-grained zones that contain hydrocarbons can appear more conductive than coarser-grained zones that are water-bearing. In freshwater reservoirs, this phenomenon can be accentuated by minor changes in formation-water salinity. One such resistivity reversal is shown in Fig. 5, where the upper oil-producing interval has a measured resistivity that is about half that of the underlying water zone.

Table 6 illustrates the pseudo-Archie and variable shaly-sand exponents that have been applied in studies of low-resistivity pay where the formation-water salinities are low. Note that the severe non-Archie effects in the Tipam Sands drive the saturation exponent n below the theoretical minimum of unity for Archie rocks. Table 6 also indicates the benefits that can be derived from an improved petrophysical interpretation.

Pseudo-Archie exponents will be encountered again where low-resistivity pay can be attributed to one of the three causes of high capillarity, a phenomenon that can also occur in freshwater-bearing hydrocarbon reservoirs. In such cases, the pseudo-Archie exponents generated here might also accommodate high capillarity, to a degree that can be ascertained only by validating the log-derived water saturations.

Conductive Minerals

The presence of conductive minerals is the least cited cause of low-resistivity pay. The root of the problem is electronic conduction due to iron-bearing minerals that occur in clusters and whose concentration exceeds a critical level, taken by Clavier et al. (1976) for the case of pyrite as 7% by volume of the total solids. Another mineral that has been grouped in this category is glauconite, which can in many respects be regarded as a clay mineral. However, here it is seen as an iron-bearing mica with a potential excess conductivity (Cook et al. 1990) over and above any conductivity enhancement due solely to textural effects. Yet again, volcanic tuffs have been cited as showing an excess conductivity beyond that predicted from electrochemical phenomena (Itoh et al. 1982). Conductive minerals can co-exist with other causes of low-resistivity pay.

In the context of conductive minerals, the low-resistivity pay problem reduces to correctly evaluating the water saturation. There is no generally accepted way of handling data from electronically conducting reservoirs. A first step towards recognizing the problem is to measure the conductivity of oven-dried core plugs. If the dry conductivity is finite, there will be a problem to investigate. If the dry conductivity is infinite, there could still be a problem, because conductivity enhancement by clusters of iron-rich minerals might only become significant in the wet state. In this case, the concentration of the metallic minerals should be investigated to see if it is supracritical.

Figure 6 shows wireline logs from the Trimble Field in Smith County, Mississippi (Cook et al. 1990). The logged interval includes the uppermost section of the Stanley Sand (with formation top at 7658 ft), which forms part of the Upper Cretaceous Eutaw play in the Mississippi Interior Salt Basin. This interval was logged with the dual induction device, whose deep sensor is minimally affected by the variations in borehole diameter. The interval contains two perforated zones. The lower test zone has an average deep-induction resistivity of about 0.3 Ωm: it produced only salt water. The upper test zone has an average resistivity of 0.6 Ωm: it flowed gas at a rate of 782×106 ft3 per day with a water cut of 13 BBL per day. The low resistivity of the producing zone is largely attributed to a supracritical content of glauconite in the pore lining of the arenaceous rock matrix. The glauconite provides an additional conducting mechanism for electrical current from resistivity tools and thereby lowers the resistance of the formation. In this example, there are almost certainly co-existing high capillarity effects associated with the overgrowth texture.

Table 7 summarizes the approaches to the conductive mineral problem that have been discerned from the literature. Our knowledge of this particular cause of the low-resistivity pay problem is the least developed.

Fine-Grained sands

In the context of fine-grained sands, the low-resistivity pay problem has two facets. First, the fines can act as a separate mineral even if they are comprised principally of quartz. This facet takes the form of a non-trivial surface conductance arising from the large pore surface area (e.g. Darling & Sneider 1992) and giving rise to a significant excess conductivity that has to be accommodated in the water-saturation algorithms (Rink & Schopper 1974). Second, the fines are associated with a high irreducible water saturation, which is present as a continuous phase and therefore raises the reservoir conductivity still further (e.g. Vajnar et al. 1977). The latter causative factor is more dominant than the former. However, it is important that the water saturation first be evaluated correctly before it is apportioned into immobile and free-fluid components.

The first stage is to validate the proposed water saturation algorithm in the same way as was discussed for freshwater reservoirs above. Interestingly, Kuttan et al. (1980) observed that shaly-sand analysis was inadequate in the presence of a large silt-grade fraction and they designated silt a distinct mineral type, even though X-ray diffraction showed that the silt fraction was principally quartz with only minor amounts of dolomite and siderite. However, it is noteworthy that their case study embraced fresh to brackish waters, which might have compounded the validation issue, emphasizing once again the interrelated nature of the causes of low-resistivity pay.

The second stage is to evaluate the free-fluid fraction of the pore space. This has usually been done by referring the interpreted water saturation to the irreducible water saturation for that particular group of sands or, more directly, by evaluating movable hydrocarbons on a level-by-level basis. The most common method for the evaluation of movable hydrocarbons entails a comparison of the characteristics of the flushed and undisturbed zones, possibly drawing on the self potential (SP) log (Heckel 1985). The direct evaluation of free fluid, which equates to movable hydrocarbons under conditions of irreducible water saturation, is feasible through magnetic resonance logging (Austin & Faulkner 1993). This last application is expected to grow in the future as the capabilities and limitations of the technique become established.

Figure 7 shows logs run in a Tertiary sand in the Gulf of Mexico (Heckel 1985). The formation is described as a fairly clean, fine-grained sand. The hydrocarbon-bearing interval is sufficiently thick to be fully resolved by the deep induction log. This interval has a measured resistivity of about 0.3 Ωm with an implied water saturation of over 80%, and yet it flowed 470 BBL of dry oil per day. The resistivity of the underlying water zone is 0.2 Ωm, the low resistivity contrast being attributed to the high irreducible water saturation associated with the high pore surface area.

Table 8 illustrates some case histories of the fine-grained, low-resistivity pay sand problem, together with the methods of solution and the resulting adjustments to the initial estimate of water saturation. Table 8 illustrates that low-resistivity pay can be indicated where irreducible water saturations are very high, e.g. up to 80%, and that even computed water saturations of up to 100% can be associated with the production of dry hydrocarbons. Similar observations relate to the other forms of high capillarity that follow.

Internal Microporosity

Internal microporosity is within the rock matrix. It is not associated with clay-derived microporosity, which is considered separately as superficial or overgrowth micro-porosity because it can be distinguished petrophysically. Various definitions of microporosity have been put forward, but here the literal meaning is adopted, i.e. pores with a diameter less than 1 μm. Documented examples of internal microporosity are principally concerned with carbonates and granular chert (Table 9). The microporosity of carbonates and chert sandstones can reach 50% of the total interconnected porosity (Dixon & Marek 1990; Worthington & Pallatt 1992). In many cases, conduction during the course of electrical measurement has been presumed to be uniformly ionic, with no correction being applied for surface conduction effects. It is, however, worth noting that excess conductivity has been recorded in chalks, presumably because of the high specific surface area (Barker 1994), as well as in cherts (Swanson 1985).

Once again, the low-resistivity pay problem can be broken down into two parts. The first part is concerned with correctly evaluating the water saturation. The second part involves apportioning the water into the immobile microporosity and the movable macroporosity.

Although a non-Archie approach to the evaluation of water saturation is not usually applied, the presence of dual porosity within the reservoir rock does complicate the application of the Archie equations, particularly as regards the saturation exponent and to a lesser extent the porosity exponent. There are two approaches. The first is directed at establishing separate constant values of these exponents for the micropores and the macropores (Petricola & Watfa 1995), a philosophy that can also accommodate differences in the salinity of micropore and macropore waters. The second and more pragmatic approach to the evaluation of water saturation draws primarily upon the observed variation of saturation exponent with water saturation and therefore uses a value that is different from the classical Archie value (Dixon & Marek 1990). In some cases, the saturation exponent measured conventionally at high water saturations can be as high as 3.0, whereas the value at reservoir water saturations is close to the classical Archie value of n = 2.0 (Swanson 1985). In these cases, the saturation exponent is sometimes taken as that which corresponds to saturations close to irreducible. The value of the porosity exponent is constant for a given sample and is generally closer to the Archie value of 2.0. Some investigators have specified the exponents m and n to be equal and they have evaluated them conjunctively at irreducible conditions (Guillotte et al. 1979).

The second part of the internal microporosity problem, i.e. establishing the immovable fraction of water, follows the corresponding discussion of the previous section.

Internal microporosity is best investigated through pore-size-distribution studies. A unimodal distribution of macroporosity (modal pore diameter typically within the range 5–10 μm) implies that there is no problem of abnormally high capillarity. A strongly bimodal distribution of micropores and macropores suggests that there is a microporosity problem and it might be one of internal microporosity. Of course, if a formation isentirely microporous, e.g. the Smackover Limestone in parts of Texas, it does not flow oil (Kieke & Hartmann 1974). There is a strong potential role for magnetic resonance logging in the evaluation of flow capability.

Having established that micropores account for a significant fraction of the pore space, their effect on the saturation exponent is best investigated through a method such as controlled-flow continuous injection, which allows sufficiently detailed resistivity index data to be gathered readily (De Waal et al. 1991). At least 20 data points are required and the experiments should extend to water saturations that approach irreducible conditions.

Figure 8 shows a logged interval of the Rodessa Limestone in Houston County, Texas (Kieke & Hartmann 1974). The lower part of this reservoir has a resistivity of less than 1 Ωm and an interpreted water saturation of 0.66. The resistivity of the upper part is at least an order of magnitude greater and this leads to much lower interpreted water saturations. Yet, the lower reservoir produced water-free gas. It was concluded that the high water saturation in the lower reservoir represents immobile water in the micropores within and around the grains, as indicated by scanning electron microscopy and confirmed by mercury injection. Therefore, the intergranular porosity and microporosity contain different fluid phases.

Table 9 illustrates some case histories of the internal microporosity, low-resistivity pay problem, together with the methods of solution and the resulting adjustments to the original estimate of water saturation.

Superficial Microporosity

As used here, the term ‘superficial microporosity’ relates to cases where the micropores are associated with mineral overgrowths which are, by definition, external to the original granular matrix. In this respect it complements internal microporosity, from which it is distinguishable petrophysically. In the context of low-resistivity pay, superficial microporosity encompasses all the issues discussed in the previous two sections.

Superficial or overgrowth microporosity is usually caused by clay minerals coating quartz matrix (Hurst & Nadeau 1995). The effects of these minerals on conduction can go beyond the excess-conductivity phenomena for which shaly-sand algorithms are intended to correct. This point is illustrated by Klein et al. (1993), who used the model of Waxman & Smits (1968). The additional effects are sometimes attributed to surface roughness, and they cause a change in the (shale-corrected and therefore intrinsic) saturation exponent, but not in a way that mirrors the case of internal microporosity (Diederix 1982). Further, superficial microporosity can either be distributed throughout a homogenized rock (Morphy & Thomson 1985) or be confined to alternating layers within a laminated sequence of sands (Bos 1982).

Once again, the problem can be divided into two parts, the evaluation of water saturation and the apportionment of this water between the immobile and free-fluid regions. In this case, the evaluation of water saturation is less straightforward, because the Archie exponents m and n must first be corrected for shale effects to their intrinsic values m* and n*, respectively, unless the formation-water salinity is sufficiently high to render this correction insignificant. The variation of n* with water saturation then becomes the key issue. One approach has been to select n* so that the minimum calculated water saturation is the value at irreducible conditions (Ramburger 1989). Of course, the calculation uses a shaly-sand algorithm, in this case the dual water model (Clavier et al. 1984), because n* here is a pseudo-non-Archie parameter as opposed to a pseudo-Archie parameter. [A related approach (Grimnes 1988) has taken shale resistivity as the tuning parameter in the Indonesia equation (Poupon & Leveaux 1971).] Whatever value of n* is chosen, it must take account of the increasing influence of the microporous region as water saturation is decreased. Where values of m* and n* cannot be established, the fallback is the pseudo-Archie approach.

The second part of the superficial microporosity problem, i.e. establishing the fraction of immovable water, follows the corresponding discussion related to fine-grained sands. However, superficial microporosity is more difficult to study, because pore-size distribution is usually determined through mercury injection into a dry core sample, and conventional drying processes can alter the pore geometry where clays are present (Worthington et al. 1988). For the same reason, detailed measurements of the variation of saturation exponent with water saturation should be made on preserved core, an application to which the continuous injection method can be well suited. Again, there is a strong potential role for magnetic resonance logging.

Figure 9 shows a logged section of thick Miocene sands in the Attaka Field off East Kalimantan, Indonesia (Partono 1992). This reservoir has a clay-mineral content that ranges from 18–43%. Approximately 80% of the clay minerals comprise interlayered illite/smectite that forms honeycombed overgrowths on the quartz grains. It is the irreducible water associated with these surficial clay-mineral structures that suppresses the resistivity of this reservoir, so that conventional shaly-sand equations are not applicable. The effect of superficial microporosity is illustrated in Fig. 9, which shows an upper water zone above a gas-bearing, low-resistivity pay interval. This water zone has an average resistivity of 1 Ωm, which will serve as a reference. The gas-bearing unit has a suppressed resistivity of about 3 Ωm. Below these intervals are a more resistive oil-bearing unit and a second water zone. In contrast to the upper gas zone, the oil-bearing unit has a formation resistivity of around 50 Ωm. This lack of resistivity suppression is attributed to a relative dearth of clay-mineral overgrowth structures within the oil-bearing interval, which therefore cannot be described as low-resistivity pay. The lower water zone is also comparatively unaffected by resistivity suppression. In fact, its resistivity of 5 Ωm is higher than that of the upper gas-bearing interval, another example of resistivity reversal.

Table 10 illustrates some case histories of the superficial microporosity problem, and it also indicates the type of solution and the resulting improvements in the interpreted water saturations. Interestingly, most solutions take the form of a shaly-sand approach with variable parameters, in an attempt to compensate for the effects of high microporosity which, in extreme cases, cause irreducible water saturation to exceed 85%. Note that glauconitic overgrowths also have a microporosity and that the associated resistivity suppression co-exists with and might supersede any conductive-mineral effects.

Interpretation Strategy for low-resistivity pay

In the previous six sections a brief analysis has been made of the different forms of low-resistivity pay according to the adopted definition and groupings. These messages have been synthesized into a hierarchical framework for the recognition and handling of low-resistivity pay zones (Fig. 10). The scheme recognizes that a reservoir rock is a coupled physico-chemical system. It is built around the six causes of low-resistivity pay that have been discussed separately above. In this respect the structure is modular, even though different causes of low-resistivity pay can co-exist (e.g. Barlai 1984). The latter possibility is accommodated by the facility to iterate, an option that also takes account of maturing data scenarios. Because these six causes encompass the vast majority of occurrences of low-resistivity pay, the scheme is generic and, if applied to previously encountered low-resistivity pay situations, it is demonstrably robust. It is noteworthy that at no point does Fig. 10 raise the question: ‘Does the reservoir have low resistivity?’ Indeed, Table 2 aptly demonstrates that absolute values of resistivity are not the issue here.

Figure 10 has the flexibility to be applied during both the appraisal and the development phases of a reservoir. A major potential benefit must be to return to older wells within the same reservoir system in situations where the deployment of more recent logging technology in some of the newer wells has revealed a low-resistivity pay situation. The key to success in this respect is to develop an interpretation scheme that is exportable to those earlier wells for which the database is more limited.

Although core data are required for groundtruthing, the scheme should ideally not require additional logging, but rather the innovative use of existing logs. For example, it is known that logarithmic displays of resistivity can mask subtle variations with depth. A return to the linear conductivity curve would be helpful in this regard. Further, in the case of laminated sands the application of the latest signal-enhancement processing might sharpen the spatial resolution of older tools sufficiently to allow the reservoir layers to be recognized. Where beds can be resolved, appropriate pseudo-Archie parameters might be quantified and exported from key wells to uncored wells. Again, modern nuclear magnetic resonance logs might contribute to the identification of free fluids in current logging programmes, but an even greater added value can be derived if this success can be used to calibrate an interpretation procedure that can be applied to conventional logs, including those in older wells. The ability to export remedial interpretation strategies in this way offers a potentially enormous benefit in mature provinces whose producing life can extend back to the 1960s, when the low-resistivity pay problem was first documented.


Published occurrences of low-resistivity pay, a term that also includes low-resistivity-contrast reservoirs, have been groupedaccording to their primary cause. Six causes have been identified for unfractured reservoirs, and all the recorded cases that have been considered do fit into one or more of these groups. The case histories within each group have been analysed to evaluate the most effective modus operandi for each causative factor. This analysis has formed the basis for the development of a strategy for recognizing and investigating the low-resistivity pay problem. This strategy should be seen against the ever-present need to make optimum use of all the available data, whether these be from core, wireline logs, mud logs or pressure and flow tests.

Although a key target is the recognition of low-resistivity pay before primary completion, a major thrust of this initiative is to facilitate the development of improved interpretative procedures that might lead to cost-effective re-completions of hidden reservoirs within the same producing system. The approach is therefore designed to provide exportable approaches to the evaluation of hydrocarbons in place. Yet, it is sufficiently flexible to incorporate other logging technologies as these become part of established operating practice.

Because the strategy has been established from a global perspective, it has a general relevance to the appraisal and development of reservoirs that contain low-resistivity pay. There are no short cuts to this goal. Application of the strategic modules may require (modest) investment that is recovered through more efficient and effective production against the welcome backdrop of a much reduced uncertainty. The ultimate measure of success remains the ability to predict commercial production rates.

This paper is a transcript of a keynote presentation to the Symposium on Low-resistivity Pay held under the auspices of the Houston Chapter of the Society of Professional Well Log Analysts in Houston, Texas, on 30 April 1998. It constitutes an extended version of a preprinted paper presented at the Society of Petroleum Engineers’ Asia Pacific Oil and Gas Conference held in Kuala Lumpur, Malaysia, during the period 14–16 April 1997. The author acknowledges Gaffney, Cline & Associates for supporting the preparation and presentation of this work.


  • B

    equivalent conductance of (sodium) clay exchange cations (eq−1 litre S m−1) Csh conductivity of wetted shale (S m−1)

  • D

    detection limit of logging tool (cm)

  • Fa

    apparent formation resistivity factor

  • F*

    intrinsic formation resistivity factor

  • K

    permeability (mD)

  • Pc

    capillary pressure (bar)

  • Qv

    cation exchange capacity per unit pore volume (eq litre−1)

  • R

    resolution limit of logging tool (cm)

  • Ro

    formation resistivity in a water zone (Ωm)

  • Rt

    formation resistivity in a hydrocarbon zone (Ωm)

  • Rw

    formation water resistivity (Ωm)

  • Rwa

    apparent formation water resistivity (Ωm)

  • Rxo

    flushed-zone resistivity (Ωm)

  • Sw

    water saturation

  • Swirr

    irreducible water saturation

  • Vsh

    wetted shale volume fraction

  • m

    Archie porosity exponent

  • m*

    shale-corrected Archie porosity exponent

  • n

    Archie saturation exponent

  • n*

    shale-corrected Archie saturation exponent

  • φ