ABSTRACT

At least five distinct types of dolomite occur in the Arab-D Reservoir in Ghawar field, Saudi Arabia – one of which appears to be responsible for high flow or ‘super-k’. These dolomite types are distinct petrographically, geochemically and stratigraphically:

  • a finely-crystalline non-fabric-preserving (NFP) variety of dolomite in the lower Arab-D (Zone 3) with low oxygen isotope values and generally poor reservoir quality;

  • a medium-crystalline NFP dolomite with high oxygen isotope values and very poor reservoir quality in the upper Arab-D (Zone 2);

  • a medium to coarsely-crystalline NFP dolomite with low oxygen isotopic values and very good reservoir quality (‘super-k’) occurring in Zone 2; and

  • a finely-crystalline fabric-preserving (FP) dolomite in the uppermost Arab-D (Zone 1) that contains high oxygen isotope values and has generally fair to poor reservoir quality.

Previous studies have documented a rare fifth type of dolomite, baroque or ‘saddle’ dolomite, that occurs locally in the reservoir as well.

This study also quantified and mapped the abundance and distribution of dolomite across the field, using all available core and log data. Analysis of dolomite distribution map patterns reveals that dolomite occurs in Ghawar as a series of linear trends extending for tens’s of kilometers. These map pattern trends are best-developed in Zone 2B, but are also visible in Zones 2A and 3A as well. Baroque dolomite appears to be limited to a few areas of vertically pervasive dolomite occurring on the same trends of high dolomite content. The linearity of these dolomite trends strongly suggests that some structural element is responsible for controlling their orientation. We interpret these linear patterns to have formed in response to a series of fracturing and/or faulting events that allowed dolomitizing fluids to move up into the reservoir from below, and preferentially dolomitize there.

Both a qualitative and a quantitative analysis of performance data (flowmeters) in southern Ghawar (Haradh) indicate that these trends of high dolomite have a profound influence on fluid flow in the reservoir. A qualitative analysis of occurrences of ‘super-k’ in the Arab-D in Haradh suggests that most ‘super-k’ zones (seen as ‘spikes’ or step profiles on the flowmeter) occur in the high dolomite trend in Haradh. A quantitative analysis of flowmeter data and a comparison of this analysis with dolomite map patterns indicate that most reservoir flow occurs where dolomite is abundant, and suggests that there is a direct relationship between patterns of high flow and high dolomite.

INTRODUCTION

The Arab-D reservoir in Ghawar field (Figure 1) contains a significant amount of dolomite rock (carbonate rocks that are composed of more than 75% of the mineral dolomite). These rocks typically formed during diagenesis of pre-existing limestones and thus their pore systems and reservoir characteristics are fundamentally different from those of limestones. The occurrence of dolomite in the Arab-D has a pronounced impact on reservoir quality, in a number of ways. At times, dolomite is responsible for producing zones of very high flow (‘super-k’) in the reservoir, while at other times these rocks act as permeability barriers or baffles. ‘Super-k’ in the reservoir is typically defined as an interval in which flow of greater than 500 barrels per day per foot of vertical interval occurs.

There are important differences between dolomites and limestones in the reservoir, both in terms of their impact on reservoir quality as well as in terms of the reservoir management strategies that are needed for these rocks. Dolomite typically has a higher grain density and acoustic velocity than limestones, which affects the response of most porosity wireline logging tools to dolomite rocks. In addition, dolomite reacts more slowly to the presence of acids than does limestone, resulting in uneven responses to acidization treatments. Surface chemistry differences may also result in dolomite and limestone differing in their wettablity and adsorption characteristics within a reservoir. Finally, limestone and dolomite differ greatly in their strength and ductility. Abundant experimental (Handin et al., 1963; Hugman and Friedman, 1979) and field data (Stearns, 1967) suggest that dolomite is generally stronger and more brittle than limestone at the same burial depth. The preferential development of fractures in dolomite beds often enhances their permeability (without greatly affecting their porosity) and potentially influences the overall quality of the reservoir.

This Study

This study examined these dolomite rocks and their impact on fluid flow in the reservoir, and included three basic work phases:

  • geochemical and petrographic characterization of dolomite types present in the Arab-D;

  • quantitative evaluation of the abundance and distribution of dolomite in the reservoir; and

  • assessment of the impact of dolomite on reservoir performance and delineation of likely flow units—and especially the high flow or ‘super-k’ units—in the reservoir.

In this paper, we will focus on a qualitative and quantitative characterization of dolomite in the Arab D, and integrate this information with reservoir performance data to improve our understanding of these important reservoir rocks. Since dolomite represents both the very best and the very worst reservoir quality rocks in the Arab-D, it is critical that we accurately describe their properties and distribution in the reservoir. This study thus represents a first step in our efforts to geologically characterize these rocks and, ultimately, develop a predictive model for variability in their reservoir quality.

GEOCHEMICAL AND PETROGRAPHIC CHARACTERIZATION OF DOLOMITE—or Not All Dolomite is Created Equal!

This study identified the occurrence of five different types or styles of dolomite in the Arab-D at Ghawar field. These different dolomite types are distinguished on the basis of their petrographic and geochemical signature, and each has a different impact on reservoir quality. The four main types of dolomite include: Zone 3 non-fabric preserving (NFP) dolomite, Zone 2 non-‘super-k’ NFP dolomite, Zone 2 super-k NFP dolomite, and Zone 1 fabric preserving (FP) dolomite. These four dolomite types are basically restricted to the reservoir zone for which they are named. A rare fifth type of dolomite, baroque or ‘saddle’ dolomite, has been previously documented in the Arab-D (Cantrell and Hagerty, 1988); this type of dolomite is volumetrically insignificant in the reservoir, and will not be discussed further in this paper.

Zone 3 NFP Dolomite

Most examples of dolomite from this zone occur as beds of partly dolomitized mud- to wackestones with considerable amounts of preserved calcite (Figure 2). In these rocks, dolomite typically has obliterated all traces of the original limestone fabric, to create a non-fabric-preserving (NFP) style of dolomitization. Petrographically, this dolomite is finely crystalline, with typical crystal sizes reaching about 100 μm in size. Dolomitization in this part of the reservoir appears to have occurred preferentially in and around non-depositional (hiatal) surfaces and their associated burrows or borings. Geochemically, these rocks display relatively negative (light) oxygen isotope values (-3 to -6‰ δ18O) and positive (heavy) carbon isotope values (2.5 to 3.3‰ δ13C)(Figure 6). Reservoir quality of these rocks is typically very poor, probably reflecting the original mud-rich character of the precursor depositional fabric.

Zone 2 Non-Super-K NFP Dolomite

Non-super-k dolomite in Zone 2 occurs as a more coarsely crystalline NFP dolomite, with typical crystal sizes from about 150 μm to up to 400 μm (Figure 3). Geochemically, this type of dolomite has a much heavier (more positive) oxygen isotopic signature (-2.5 to 0‰ δ18O) than does the Zone 3 NFP dolomite, although its carbon isotopic signature is similar (Figure 6). This type of dolomite is very common, especially in lower Zone 2, and tends to occur as beds of essentially 100 percent dolomite, with little or no reservoir quality.

Zone 2 Super-K NFP Dolomite

While petrographically similar to the Zone 2 non-‘super-k’ NFP dolomite, this type of dolomite has a unique and distinctive geochemical signature—as well as a significantly different impact on reservoir performance (Figure 4). Geochemically, this dolomite has a much lighter (more negative) oxygen and carbon isotopic signature than that seen in the Zone 2 non-‘super-k’ NFP dolomite (Figure 6). Oxygen isotope values range from -3.5 to -5‰ δ18O, while carbon values range from 1.8 to 2.8‰ δ13C (generally the lightest carbon values found during this study).

This type of dolomite is especially important because it occurs in association with high-flow or super-permeability (‘super-k’) zones in the reservoir, and has good to excellent reservoir quality. Samples of this type of dolomite have common to abundant intercrystalline porosity, and locally contain leached moldic porosity (typically after Cladocoropsis, a finger-shaped stromatoporoid). Especially in combination, the presence of these two pore types in dolomite typically yields very good reservoir quality and high flow rates. Discovery in this study of the geochemical ‘fingerprint’ for these ‘super-k’ dolomites represents a significant breakthrough in our ability to identify and define ‘super-k’ and high flow units using geologic data. This information, in conjunction with the mapping results detailed in the next section, provide an important new tool for predicting the occurrence of ‘super-k’ ahead of the drill bit.

Zone 1 FP Dolomite

The Zone 1 fabric-preserving (FP) dolomite is rare overall in the Arab-D and only occurs in the transition between the overlying C-D Anhydrite and the underlying Arab-D carbonate. Petrographically, this is a very finely-crystalline dolomite in which details of the original limestone fabric are usually well preserved (Figure 5). Typical crystal sizes range from 50 μm to 80 μm in size. Geochemically, FP dolomite contains moderate oxygen isotope values (-1.5 to - 3.5‰ δ18O) and carbon isotope values similar to the Zone 3 NFP and Zone 2 non-‘super-k’ NFP dolomite (2.5 to 3‰ δ13C) (Figure 6). Beds of FP dolomite typically occur as thin, sheet-like or stratigraphic layers in Zone 1. Reservoir quality of these dolomites generally reflects the character of the original depositional fabric and has not been significantly altered through dolomitization; since most of these rocks were originally deposited as relatively mud-free and partially cemented mud-lean packstones, reservoir quality is fair to poor.

Geochemistry

The wide range of carbon and oxygen isotopic signatures suggests that the mechanism of formation and fluids responsible differed between dolomite types, even though these different dolomite types often occur intimately associated and are fairly similar petrographically. It is this geochemical distinctiveness that allows the development of a genetic—and predictive—model for the occurrence of dolomite in the Arab-D.

Each of the dolomite types discussed above formed at different times and from different fluids (Cantrell et al., 2000). Because of its association with non-depositional (hiatal) surfaces and basically marine δ13C signature, the Zone 3 NFP dolomite is suggested to have formed during the time represented by these non-depositional surfaces by Mg diffusion from the overlying seawater. This mechanism would have been promoted by the oxidation of organic material contained in the sediment at these surfaces. These dolomites might have formed by a mechanism similar to that documented by Swart and Melim (2000). Later burial to present-day depths has probably caused some recrystallization to occur in these dolomites, as they responded to the normal geothermal gradient during burial, to yield their current relatively light δ18O signature.

In contrast, the Zone 2 non-‘super-k’ NFP and Zone 1 FP dolomites have similar δ13C values to the Zone 3 NFP dolomite, but have significantly heavier (more positive) δ18O values. This heavy oxygen isotopic composition and geochemical similarity to the Zone 1 dolomites—which are intimately associated with the overlying evaporites—suggest that both these dolomite types were formed by reflux of hypersaline fluids through the sediment. We propose that these hypersaline fluids were sourced in the overlying evaporites and passed rapidly through the grainstone-dominated Zone 2A, to dolomitize the underlying, less grainstone-dominated Zone 2B. Finally, we interpret the geochemically distinct Zone 2 ‘super-k’ NFP dolomite (with its very light δ18O and relatively light δ13C compositions) to have formed later by hot fluids moving up from below into the reservoir interval and dolomitizing there.

DOLOMITE DISTRIBUTION AND ABUNDANCE

The next step in characterizing dolomite in the Arab-D was to investigate quantitatively the abundance and distribution of dolomitized intervals in the reservoir. For this, we utilized all available data to indicate lithology in Ghawar, including (1) core descriptions, (2) neutron-density wireline logs, and (3) combined sonic-resistivity logs (Figure 7). The total database used to quantify dolomite in the Arab-D contained 1,256 wells from Ghawar field.

Lithology Database

The first priority source of lithology data was provided by inspection of core descriptions. This method of direct observation of dolomite content in cores is the most objective of all methods used but because of the relatively small amount of core material available, was used infrequently to infer the presence of dolomite in the Arab-D.

The second source of dolomite data was neutron-density wireline logs (CNL/FDC or CNL/LDT logs. These logs were utilized to designate dolomite when the neutron log read significantly to the left of the density log.

Finally, sonic and resistivity logs were used to interpret dolomite when no core or neutron-density logs were available. This method utilizes the empirically-observed association between low sonic porosity, high resistivity and dolomite in Zone 2. While delineating dolomite intervals in this manner is more subjective than with the other methods, use of this technique allowed the interpretation of dolomite when no other, more reliable data were available.

Dolomite Statistics and Mapping Results

Using these data, it was possible to analyze and map out the abundance and distribution of dolomite in each of the major reservoir zones, as well as evaluate statistically dolomite variability across the field.

Statistics

A statistical review of the distribution of dolomite across Ghawar revealed that dolomite does not occur randomly or uniformly within each zone. Overall dolomite comprises approximately 14 percent of the Arab-D in Ghawar. Table 1 displays the overall and zonal distribution of dolomite in the Arab-D, as well as the average thickness (in feet) of each zone. For this table, dolomite distribution is listed in terms of both average percent of dolomite comprising each reservoir zone and average thickness (in feet) of dolomite present in each zone.

Some overall conclusions from our analysis of the distribution of Arab-D dolomites include:

  • Although it is the thinnest of the major reservoir zones in the Arab-D, Zone 1 contains the highest proportion of dolomite in the reservoir; Zones 2B and 3A have less dolomite, although still a significant amount. Dolomite is relatively insignificant volumetrically in Zones 3B and 2A (Figure 8).

  • Significant internal variability occurs within each zone (Figure 9), with individual internal layers commonly containing much of the total dolomite present in the entire zone (for example, layer 13 of Zone 3A is typically more dolomitic than other layers in Zone 3A and overall contains most of the dolomite in this zone). Likewise, dolomite content in individual layers in Zone 2B is highly variable.

  • On a Ghawar-wide basis, dolomitization appears to increase systematically from north to south in Zones 2A and 2B, but Zone 3A displays an inverse trend by generally decreasing from north to south (Figure 10).

Map Patterns

Maps of the percent dolomite in the Arab-D and in its major zones are presented in Figures 11 to 16. These maps summarize dolomite content of each zone and of the entire Arab-D.

Analysis of these maps reveals that dolomite occurs over most of Ghawar as a series of parallel linear trends of high dolomite content that extend for ten’s of kilometers in a northeasterly direction (see especially Figures 12, 13 and 16). Locally, these trends connect up with areas of pervasive dolomitization, where much of the section (virtually 100% of Zones 1, 2A, 2B and 3A) has been replaced by dolomite; in places these wells may contain abundant ferroan baroque dolomite. These northeasterly trends become less obvious in southern Ghawar, although their appearance may be compromised somewhat by a slight change in the overall orientation of the field at that point, as well as by a paucity of well control in central and southern Haradh. However, based on the consistency of these trends to the north, we interpret that these trends connect up with high dolomite areas in the south. Of additional significance is the fact that these trends of high dolomite across Ghawar appear to be consistent throughout much of the reservoir; although they are best developed in Zone 2B (Figure 13), they are also somewhat visible in Zones 2A and 3A (Figures 12 and 14), as well as in the overall Arab-D map (Figure 16).

The presence of such pronounced linear trends in the distribution of dolomites strongly suggests that some structural element is responsible for controlling their orientation. We interpret that these linear patterns formed in response to a series of fracturing and/or faulting events that allowed the movement of fluids into the reservoir and preferential dolomitization to occur along these linear features.

Additional evidence exists to support the interpretation of fracture/fault-related dolomitization. Fracture-related dolomitization has been previously reported elsewhere in the literature, in the Albion-Scipio Trend of Michigan (Taylor and Sibley, 1986). In this example, it was postulated that fractures allowed fluids to move vertically upward from depth into the reservoir, and preferentially dolomitize the limestone along the fracture trends. In the Arab-D, an earlier structural-elements study (primarily from image log data) by Western Atlas in northern ‘Uthmaniyah revealed the presence of NE-oriented zones of enhanced fracturing (J. Mattner, 1998, personal communication). The fact that such trends occur in the reservoir in generally the same location as the mapped high dolomite trends demonstrates that structuring generally on-trend with the high dolomite areas did occur, and that open fractures are present which could have allowed the migration of fluids into the reservoir. Additional support for dolomitization related to fault/fracture zones in the reservoir comes from recent horizontal drilling data, in which dolomite is reported to occur only along an identified fault zone (A. Shahri, 1998, personal communication).

PERFORMANCE DATA ANALYSIS AND INTEGRATION WITH DOLOMITE DATA

The final phase of this project was to evaluate the relationship between dolomite and fluid flow in the Arab-D reservoir. For this purpose, both qualitative and quantitative analyses were conducted of reservoir performance data in northern Haradh, using flowmeters as the major parameter to describe the location and extent of flow zones.

Qualitative Flowmeter Analysis

In the qualitative analysis of flowmeter data, visual observations of high flow step-profiles or ‘spikes’ on the flowmeter (thin intervals that account for 50 percent or more of the total flow of the reservoir) were identified and tabulated. In the Haradh area, these high flow ‘spikes’ only occurred at three stratigraphic locations within the reservoir interval: (1) predominantly in a dolomite (low porosity) interval in the upper and middle portions of Zone 2B; (2) in a limestone at the base of Zone 2B; (3) and very rarely in a high porosity limestone grainstone in Zone 2A. The high flow ‘spikes’ associated with dolomites in the upper and middle portions of Zone 2B account for the majority of high flow profiles in the Haradh area. Figure 17 is a schematic that illustrates a typical flowmeter profile for an Arab-D well in Haradh, and shows a high flow ‘spike’ in lower Zone 2B.

Comparison of these high flow ‘spikes’ or ‘super-k’ zones with the dolomite maps reveals that, for the most part, their occurrence is not random across the field and in fact appears to be related to dolomite abundance and stratigraphic position. The high flow ‘spikes’ in upper-middle Zone 2B (which represent the vast majority of identified high flow ‘spikes’ in the study area) occur within areas of high dolomite within Zone 2B (Figure 18). Stratigraphically, these high flow zones occur in dolomites in the upper and middle portions of Zone 2B. These are the dolomites that appear to be most often responsible for contributing high flow or ‘super-k’ zones in this portion of Haradh.

Other high flow ‘spikes’ also occur elsewhere in the section and, while they are volumetrically less important than the high ‘spikes’ seen in upper-middle Zone 2B, they also relate to dolomite content and stratigraphic position. The rare high flow ‘spikes’ that occur in lower Zone 2B correspond to relatively low dolomite areas in Haradh (Figure 18) and typically occur in limestones at the base of Zone 2B. Likewise, rare high flow ‘spikes’ occur in limestones in Zone 2A (Figure 19). High flow in these last two instances is thought to be more related to the presence of grain-dominated limestone fabrics than to dolomite.

From a reservoir management perspective, it’s absolutely essential to understand the occurrence of these high flow ‘spikes’ and what controls their distribution and lateral continuity, for several reasons. During initial drilling and development in relatively immature, unswept portions of the field, targeting these high flow intervals can be extremely beneficial, in that wells drilled here tend to produce dry oil at very high rates. Later on in the life of the field, however, these high flow ‘spikes’ tend to produce copious amounts of water, which complicates field operations and surface facility design criteria. The decision to target or avoid these high flow ‘spikes’ is dependent upon several non-geological factors— but makes it imperative that we understand the geological controls on the occurrence of these ‘spikes’ in the reservoir.

Quantitative Flowmeter Analysis

In addition to this qualitative analysis, we also conducted a quantitative analysis of the flowmeter data from all of Haradh. For this procedure, we first merged digital flowmeter profiles with reservoir zone tops, and then calculated the flow percent contributed by each zone in each well. These data were then used to generate statistics and create maps of flow percent for each zone in Haradh.

This analysis of flowmeter data indicates that most of the flow (about 60%) in the Arab-D in Haradh is contributed by the highly dolomitized Zone 2B reservoir interval (Figure 20). Further, when the map of flow percent for Zone 2B is compared to the Zone 2B dolomite map, a very striking relationship can be seen between high dolomite and high flow in this zone (Figure 21). Analysis of map patterns seen in these two maps indicates that high flow typically corresponds to intervals of high dolomite, and low flow corresponds to intervals of low dolomite. This high degree of correlation between dolomite and flow percent map patterns suggests that, in Haradh Zone 2B (which contributes most of the flow in the Arab-D in this part of Ghawar), high flow occurs preferentially in areas that contain abnormally high amounts of dolomite. Conversely, where dolomite is relatively insignificant in this zone, high flow intervals are rare.

CONCLUSIONS

  1. Not all dolomite in the Arab-D is the same. There are at least five different types of dolomite present, of which one variety, the Zone 2 ‘super-k’ NFP dolomite, appears to be responsible for high flow or ‘super-k’ intervals in the reservoir. This study documented the geochemical ‘fingerprint’ of this type of dolomite for the first time.

  2. Mapping of the dolomite content of the Arab-D across Ghawar revealed that dolomite does not occur randomly or uniformly across the field. Dolomite occurs as linear NE-oriented belts of high dolomite content across Ghawar.

  3. These areas of high dolomite are important: they tend to be the areas where overall high flow occurs and where super-k ‘spikes’ predominate.

ACKNOWLEDGMENTS

The authors thank the Saudi Aramco for permitting them to publish this paper, the three anonymous GeoArabia reviewers for their helpful comments, and Gulf PetroLink for designing the paper.

ABOUT THE AUTHORS

David Cantrell has over 18 years of worldwide exploration and development experience in the oil industry. He graduated from the University of Tennessee with a MSc in Geology in 1982. Dave began his industry career with Exxon where he conducted numerous reservoir characterization and geologic modeling studies on reservoirs in the Middle East; the Permian, Powder River, Williston, and Gulf of Mexico Basins of the USA; and the Maracaibo and Barinas Basins of Venezuela; among others. He has been responsible for several studies on large carbonate reservoirs since joining Saudi Aramco in 1997.

Peter Swart received his PhD from the University of London in 1980 for work on modern coral reefs. After three years at the University of Cambridge, he started a project on dolomite geochemistry at the University of Miami where he is now Professor of Marine Geology and Geophysics in the Rosensteil School of Marine and Atmospheric Sciences. His professional interests include carbonate geochemistry and diagenesis, hydrology, and paleoclimatology.

Hildegard Westphal studied geology at the universities of Tübingen, Brisbane, and Kiel. She received her PhD in 1997. From 1998 to 1999, she was a postdoctoral researcher at the Rosensteil School of Marine and Atmospheric Sciences, University of Miami. Since 1999, Hildegard has been an Assistant Professor of Carbonate Sedimentology at the University of Hannover. Her interests include carbonate reservoir characterization focusing on diagenesis, petrophysical properties of carbonates, and the application of artificial intelligence methods to prediction problems.

C. Robertson Handford is a consulting sedimentologist and stratigrapher based in Austin, Texas. He received his PhD in Geology from Louisiana State University in 1976. Since then he has spent most of his professional career in petroleum industry research chiefly in the fields of sedimentology and stratigraphy. He has worked for the Texas Bureau of Economic Geology, and the research laboratories of Unocal, Amoco, and ARCO. He has been an invited lecturer to numerous geological societies and universities, and was the AAPG Distinguished Lecturer for 1995–96 on carbonate stratigraphy. He received the SEPM Excellence of Oral Presentation Award on two different occasions, the SEPM Excellence of Poster Presentation Award, second runner-up for the AAPG Jules Braunstein Memorial Award, and SEPM Excellence of Presentation Honorable Mention. He also received Honorable Mention for Best Paper published in SEPM’s Journal of Sedimentary Research in 1995. He is a member of AAPG, SEPM, IAS, and NSS.

Christopher Kendall is Professor of Geology at the University of South Carolina. He has worked extensively in the oil industry and in academia. His research interest and publications have concentrated on the Arabian Gulf. Having started his career in carbonate sedimentology and petrology, Christopher is interested in sedimentary simulations and sequence stratigraphy and has numerous publications on these topics.