Two different types of calcified dolomite, or dedolomite, occur as stratiform and non-stratiform bodies within the Jurassic (Kimmeridgian) upper Jubaila Formation in the Wadi Nisah area of central Saudi Arabia. In the stratigraphically-equivalent subsurface Arab-D reservoir in eastern Saudi Arabia, two types of dolomite, stratiform and non-stratiform, occur which appear to be similar in architecture to the dedolomites examined in this study. However, Wadi Nisah dedolomites exhibit systematic changes in texture and isotopic composition from their precursor dolomites. Non-stratiform dedolomite contains lower oxygen isotope (average δ18O = -10.99‰) and much lower carbon isotope (average δ13C = -7.51‰) values and is much more coarsely crystalline than typical subsurface Arab-D non-stratiform dolomite; in contrast, Wadi Nisah stratiform dedolomite contains similar oxygen isotope values (δ18O = -2.89‰) and only slightly lower carbon isotopes (δ13C = 0.98‰) relative to subsurface Arab-D stratiform dolomites. We suggest that non-stratiform dolomite was more susceptible to late meteoric diagenesis than the horizontally bedded stratiform dolomite intervals. Such differences in character highlight the importance of structural and diagenetic architecture in determining later, post-dolomitization diagenesis and ultimately final reservoir quality.

Calcitization of dolomite in sedimentary carbonate sequences has frequently been reported in the literature, and is often interpreted as a near-surface process reflecting either an erosional unconformity within a sequence (Schmidt, 1965; Goldberg, 1967; Braun and Friedman, 1970; Scholle, 1971) or late, post-burial weathering (Mossler, 1971; Chafetz, 1972; Budai et al., 1984). In addition, experimental work suggests that dedolomitization can only take place at or near the earth’s surface where pCO2 (the partial pressure of carbon dioxide) and temperature are relatively low (Yanat’eva, 1955; De Groot, 1967); further, De Groot (1967) has shown that despite other favorable conditions, dedolomitization becomes ineffective at temperatures higher than 50°C. Based on textural evidence and in spite of this experimental work, some authors have suggested dedolomite can also occur during later burial (Budai et al., 1984), and may post-date stylolitization and accompany hydrocarbon migration. Therefore it may be stated that the majority of occurrences of dedolomite are near-surface related, although dedolomite can be complex and form in a variety of ways.

This study reports the occurrence of dedolomite in the upper Jubaila Formation in outcrops in central Saudi Arabia. Dedolomite in this succession is notable for two major reasons:

  • Although now completely calcitized, dedolomite in these outcrops is analogous to well-studied occurrences of dolomite in the subsurface, in the time-equivalent Arab-D reservoir in eastern Saudi Arabia (Mitchell et al., 1988; Cantrell et al., 2004); and

  • Since these outcrop exposures of dedolomite in central Saudi Arabia occur in rocks of similar age and have similar geometries to dolomites observed in the subsurface in eastern Saudi Arabia, this occurrence of dedolomite allows direct comparison between pre- and post-calcitization dolomite fabrics and characteristics.

Dedolomite occurring in these outcrops represents an additional step in the diagenesis of these rocks beyond what was experienced by the same stratigraphic interval in the subsurface, and thus provides additional insights into the pathway followed by diagenesis in these rocks. By understanding the changes that have occurred in the transformation of dolomite to dedolomite in these outcrops, we can understand how the structural and diagenetic architecture of these rocks continues to provide a template or “grain” during late-stage, post-dolomitization diagenesis at the surface.

The outcrop succession examined in this study occurs in the Wadi Nisah graben in central Saudi Arabia, approximately 50 kilometers south of Riyadh (Figure 1). This east-west trending Wadi Nisah graben extends for about 90 km in length and is 2 to 3.7 km wide (Weijermars, 1998), and connects up to the east with Wadi Sahba which, together with Wadi Nisah, form an east-west striking lineament several hundred kilometers long (Al-Kadhi and Hancock, 1980) (Figure 2). Dip-slip on the normal faults which bound the graben is estimated to range between 100 and 500 meters (m) (Powers et al., 1966; Vaslet et al., 1991). Most of the graben floor is obscured by alluvial fill.

The outcrop interval examined in this study is adjacent to two previously measured and described sections of the Jubaila-Arab succession reported in Meyer et al. (1996). According to them, this succession consists initially of coarsening-upward packages of stromatoporoid mudstones to grainstones, which are then overlain by generally grainstone-dominated skeletal bank cycles, which are succeeded in turn by a set of thinning-upward cycles that are dominated by muddy, generally fining-upward carbonates. All of these cycles are interpreted to represent shoaling-upward conditions on a gently eastward-dipping ramp. See Meyer et al. (1996) for a more complete description of the facies and rock types present in this area.

It is important to recognize that the Jubaila in outcrop is not exactly analogous to the Jubaila in the subsurface as defined by most later workers. Originally defined in outcrop as the relatively resistant limestone between the softer (and more calcarenite dominated) beds of the underlying Hanifa and the overlying Arab Formations (Steineke and Bramkamp, 1952a; Steineke and Bramkamp, 1952b), the Jubaila has been informally defined by later subsurface workers as the non-productive interval occurring below the Arab-D reservoir, with the Jubaila-Arab contact occurring at the point of lowest productive oil. As a result, outcropping Jubaila Formation does not directly correlate with the Jubaila in the subsurface (Figure 3; Powers, 1968; Meyer et al., 1996), in that the uppermost beds of the Jubaila in outcrop include the middle part of the Arab-D reservoir in the subsurface (Powers et al., 1966; Powers, 1968). Thus, the interval examined in this study occurs in the upper Jubaila in outcrop, and generally corresponds to the middle Arab-D in the subsurface.

The current reconnaissance study focuses primarily on the distinct types of dedolomite that occur in this section and does not attempt to remap or re-describe this interval in detail. For this study, the section was re-examined to insure consistency with the measured sections of Meyer et al. (1996), photographed and sampled, and samples were analyzed petrographically and geochemically (carbon and oxygen isotopes). Three samples of non-stratiform dedolomite were collected, one sample was collected from the stratiform dedolomite and a final sample was collected from a calcite fracture fill.

Petrographic examination was conducted on thin sections injected with blue-dyed epoxy (to indicate the presence of porosity). Geochemical (stable carbon and oxygen isotope) analyses were conducted using the following method. Samples were air-dried and representative portions ground with a mortar and pestle to fine-sand/silt size. Special care was taken to avoid sampling thin white calcite veins in the non-stratiform dedolomites. One milligram aliquots of ground rock were loaded into individual glass reaction vessels in an automated Thermo-Finnigan Kiel III carbonate preparation system. The tubes were sequentially evacuated and treated with three drops (approximately 0.1 ml) of 100% phosphoric acid at 80°C. Evolved carbon dioxide gas was cryogenically trapped throughout the course of the reaction. Calcite samples were allowed to react for six minutes and dolomite (and dedolomite samples) for 14 minutes. Carbon dioxide was then transferred to the inlet system of a Thermo-Finnigan Delta S mass spectrometer operating in dual-inlet mode. 13C/12C and 18O/16O were measured against reference gas calibrated using the NBS-19 (Toilet Seat) carbonate standard (Coplen, 1996). Ratios are reported in delta notation in parts per thousand (‰) variation relative to VPDB (the Vienna Peedee belemnite isotope standard, which has been established as a replacement for the original PDB standard). Most samples were measured two or more times; by pooling the standard deviations of replicate analyses, overall reproducibility is estimated to be ±0.07‰ for δ13C and ±0.25‰ for δ18O. Values for NBS-19, +1.95‰ and -2.20‰, respectively, are equal to those defined by the Vienna PDB scale (Coplen, 1996) as NBS-19 was used to calibrate the reference gas. Values for the inter-laboratory laboratory standard NBS-18 are within 0.1‰ of the expected values.


Dedolomite occurs in two habits in the Wadi Nisah section, as a bed parallel stratiform type and as a stratigraphically discordant non-stratiform type (Figures 48). In outcrop, stratiform dedolomite occurs as sheet-like beds of dedolomite typically 0.5 m to 2 m thick that continue with no significant change in bed thickness across the entire face of the outcrop (several hundred meters, Figures 4, 5 and 8). Stratiform dedolomite beds display sharp, non-gradational contacts with overlying and underlying limestones, with this transition occurring from limestone to dedolomite within 1 centimeter (cm) or less. No precursor depositional textures are visible within these dedolomite beds, and fracturing is insignificant.

In contrast, non-stratiform dedolomite appears to be much thicker, extending across the entire vertical thickness of the exposure (approximately 20 m thick) but is of much more limited horizontal dimensions, only extending for approximately 5 m in width (Figure 5). Lateral dimensions and orientation (strike) of this dedolomite body were not mapped in this reconnaissance study, since this body extends into the outcrop. Calcite-filled fractures and vugs commonly occur in the outcrop, most notably in the non-stratiform dedolomite intervals (Figures 6 and 7). Fractures occur primarily as short (2–10 cm long), randomly oriented, partially- to completely-calcite cemented gashes that typically have apertures of 0.2 millimeters (mm) to 1 cm; these fractures are locally common to abundant, and may connect-up to form small breccia bodies within these non-stratiform dedolomite bodies.

Similarly oriented intervals of dolomite occur in the subsurface, and have been documented by Mitchell et al. (1988) and Cantrell et al. (2004). Stratigraphic dolomites are thin, lenticular or sheetlike beds that typically range in thickness from 0.3 m to 5 m, and usually occur within fairly well-defined stratigraphic intervals within the Arab-D. These dolomites can be quite areally extensive, covering tens of square kilometers. In contrast, non-stratigraphic dolomite is vertically pervasive in nature, with thick sections of it crosscutting stratigraphic reservoir zones. These dolomites occur when normally dolomite-free parts of the section contain significant amounts of dolomite, and typically crosscut other stratigraphic dolomites. Although intervals of non-stratigraphic dolomite typically have a very limited areal extent, a NE-SW trend has been identified and mapped across Ghawar field using subsurface well-based lithology information (Cantrell et al., 2001; 2004).


Petrographically, all samples are composed entirely of calcite, although the original dolomite fabric is preserved and visible. Relict dolomite crystals are typically on the order of 100–300 microns in size (Figures 915), and locally appear zoned or display cloudy centers and clear rims. Stratiform dedolomite has a fabric of uniformly sized calcitized dolomite crystals (Figure 9); locally, silt to fine sand sized quartz grains also occur, as does very rare intercrystalline porosity. In contrast, non-stratiform dedolomite is distinguished by relict calcitized dolomite crystals that are slightly coarser than those seen in stratiform dedolomite, which are then enclosed in poikilotoptic calcite crystals 1–2 mm in size (Figures 1015). In addition, non-stratiform dedolomite contains abundant fractures that are filled with very coarse (0.2–2.0 mm in size), equant crystals of calcite (Figure 14); commonly, the calcite crystals filling these fractures are in optical continuity with the surrounding coarse poikilotopic crystals (Figures 12 and 13). Locally, these fractures contain breccia fragments derived from the surrounding matrix, as well as probable clay material (Figures 14 and 15). Porosity is rare overall in these non-stratiform dedolomites, occurring as fracture, intercrystalline and vuggy pores.

Previous work (Cantrell et al., 2001, 2004) has identified three petrographically distinct types of dolomite in the Arab-D reservoir in the subsurface: fabric-preserving (FP), non-fabric-preserving (NFP) and baroque dolomite. Fabric-preserving (FP) dolomite is very finely crystalline dolomite (crystal sizes range from 10–50 microns in size) in which details of the original limestone fabric are usually well-preserved. Beds of FP dolomite typically occur as thin, sheet-like or stratigraphic layers that are always intimately associated with the overlying anhydrite. In contrast, NFP dolomite is a medium crystalline (average crystal sizes generally ranging from 50 to greater than 150 microns), non-baroque dolomite in which all traces of the original limestone fabric have been obliterated. This dolomite also typically occurs as stratigraphic beds, although it is not restricted to the uppermost part of the Arab-D but occurs throughout the reservoir. The third type of dolomite, baroque, is a coarsely crystalline dolomite (crystals generally ranging in size from 100 to 700 microns) with “saddle-shaped” crystals displaying undulose extinction in thin section. It is rare in the reservoir and appears to be limited to wells that contain abnormally thick sections of dolomite; in rare cases, baroque dolomite is vertically pervasive.


Stable isotope results (shown in Table 1 and Figure 16) highlight differences between stratiform and non-stratiform dedolomites in Wadi Nisah. Overall, carbon isotopic compositions range from -7.8‰ to +1.0‰ and oxygen isotope compositions from -13.9‰ to -2.9‰, although systematic differences occur between dedolomite types; non-stratiform dedolomites have very light (low) carbon and oxygen isotope values (average δ13C = -7.51‰ and δ18O = -11.00‰), while stratiform dedolomite in Wadi Nisah has much more positive isotope values (δ13C = +0.98‰ and δ18O = -2.89‰). In contrast, the calcite fracture fill has very low isotope values (δ13C = -6.67‰ and δ18O = -13.85‰).

Subsurface dolomite in the stratigraphically equivalent Arab-D interval in eastern Saudi Arabia also shows distinct differences in their structural and diagenetic architecture that are reflected in their geochemical signatures (Figure 16; Cantrell et al., 2004). Non-stratiform, baroque dolomite has fairly high (heavy) carbon isotope values, but very low (light) oxygen isotope values (average δ13C = +2.72‰ and δ18O = -7.37‰); in contrast, non-baroque generally stratiform dolomite has slightly more positive carbon and much more positive oxygen isotope results (average δ13C = +3.28‰ and δ18O = -2.58‰).

Based on analogy to subsurface Arab-D dolomites, outcrop Arab-D dedolomites show distinctive changes in mineralogy, petrography, geochemistry and reservoir quality from original (pre-calcitization) dolomite. Although the overall architecture of these dedolomite bodies still reflects their pre-calcitization geometry, changes have occurred in their mineralogy (dolomite has been altered to calcite), crystal size (recrystallization and crystal growth has occurred) and geochemistry. These changes were not random, however, and appear to have been controlled by the structural and diagenetic architecture of the original dolomite body.

Calcitization of stratiform dolomite appears to have less extensively altered both the fabric and the geochemistry of the original dolomite, relative to the degree of alteration that occurred in non-stratiform dedolomite. The overall fabric and average crystal sizes of stratiform dedolomites are generally similar to those observed in their precursor dolomites. Geochemically, there has only been a slight change in the isotopic composition of the rock during calcitization, if one assumes that the stratiform dedolomite had similar pre-calcitization stable isotope values to those observed in subsurface stratiform dolomite; δ13C has been altered by approximately -2.3‰ (+3.28‰ for original dolomite to +0.98‰ in outcrop dedolomite) and δ18O by about -0.3‰ (-2.58‰ for original dolomite to -2.89‰ in outcrop dedolomite). In contrast, non-stratiform dedolomite displays a much more altered fabric and geochemical signature. These rocks now comprise very coarse poikilotopic calcite crystals, and there has been a significant change in their isotope composition of about -10.2‰ for carbon (+2.72‰ for original dolomite to -7.51‰ in outcrop dedolomite) and -3.6‰ for oxygen (-7.37‰ for original dolomite to –11.00‰ in outcrop dedolomite). This greater degree of isotopic change in the non-stratiform dedolomites suggests that the water/rock ratio was apparently much higher in non-stratiform dedolomites, relative to the stratiform dedolomites.

In all cases, the isotopic data display a negative shift, or depletion in heavy isotopes of both carbon and oxygen. Low δ13C values reflect incorporation of carbonate ions derived from the decomposition or oxidation of organic matter (Anderson and Arthur, 1983). Low δ18O values typically indicate isotopic exchange with normal formation waters at elevated temperatures or exchange with isotopically light meteoric waters at near-surface conditions (Anderson and Arthur, 1983). In the case of these outcrop dedolomites exposed at the surface, we suggest that meteoric fluids have been responsible for the calcitization of the precursor dolomites and for the changes in their geochemistry. The question this raises then is, why did these geochemical changes not occur in a parallel manner in both stratiform and non-stratiform dolomites, such that the amount of change recorded petrographically and geochemically was equivalent? It appears that structural and diagenetic architecture played a key role in controlling how extensively these dolomites were altered during calcitization.

We suggest that, because of their near-vertical orientation, the non-stratiform dolomites were much more susceptible to alteration during near-surface meteoric diagenesis than were the stratiform dolomites. In this model, vertically-oriented fractures provided pathways for preferential flow of meteoric fluids; such a model would be consistent with the one proposed by Cantrell et al. (2004) for the original formation of these non-stratiform dolomites, where they proposed that vertical fractures and/or faults would have allowed hot fluids from below the Arab-D to move up and preferentially dolomitize the reservoir along these fractures/faults (model summarized in Figure 17, Time 3). As this interval was gradually uplifted and exposed along the wadi, these fractures have allowed later meteoric waters to move down the fractures and preferentially calcitize these vertically-oriented, non-stratiform dolomites (shown schematically in Figure 17, Time 4). The abundance of fractures in non-stratiform dedolomite, as well as the paucity of fractures in stratiform dedolomite, further supports the contention that fracturing was more common in the original non-stratiform dolomite and so played a significant role in later diagenesis there. Moreover, vertical fluid movement and aquifer communication through the widespread and extensive karstic dissolution and collapse features (e.g. caves and sinkholes) in the Wadi Sabha area and localized along structural lineaments have been previously documented in Stenger et al. (2003) (Figure 2). In contrast, the stratiform dolomites were much less susceptible to the penetration of meteoric fluids and so calcitization would have begun later, and probably proceeded at a slower rate, in these horizontally-oriented stratiform dolomites.

The significance of this finding is two-fold:

  • First, it underlines the importance of relatively early-formed structural elements (fractures and/or faults) and reservoir architecture in controlling the later diagenesis of a sediment, even into late-stage diagenesis (telogenesis); and

  • It also highlights how significant late-stage diagenesis (especially dedolomitization) can be in determining the final mineralogy, fabric and geochemical signature of a sediment.

The initial impression on first viewing this section was that the stratiform and non-stratiform dolomites were completely preserved, and so could be used directly as an analogue for the subsurface; it was only after study that the full extent of the alteration of these intervals was noted. In the subsurface, baroque non-stratiform dolomites are not typically good reservoirs, so had we used this template exactly in our study of these outcrops, these would have been considered barriers to flow. It was only after recognition of the extent of the alteration of these “non reservoir” non-stratiform dolomites that it was understood that these intervals really represented the pathways to preferential flow (albeit of meteoric water) into the reservoir here. Had these dedolomites been subsequently reburied after their recent/current exposure to meteoric diagenesis, their post-re-burial reservoir quality would have shown quite different patterns to what has been observed in subsurface dolomites in eastern Saudi Arabia.

While dedolomite can locally constitute important hydrocarbon reservoirs (Zheng and Coniglio, 1998; Purser, 1985), the significance of dedolomite in this study lies more in the insights that it provides about the predictability of later near-surface diagenesis in these rocks. In this example, reservoir quality continued to evolve throughout later diagenesis, but it continued to follow a template that had been established during earlier diagenesis. Knowledge and understanding of this template – and of the structural and stratigraphic elements that form it – is thus a key step in understanding the controls of the variability of reservoir quality in these rocks and in establishing the predictability of these changes.

A final question to consider is, if it is possible to directly identify these dedolomite intervals (both stratiform and non-stratiform) in the subsurface using seismic data. To date, all attempts to detect stratiform and non-stratiform dolomite (comparable to the dedolomites of this study) in the subsurface have been unsuccessful, for a variety of reasons. As noted earlier, stratiform dolomite beds tend to be very thin (0.3 m to 5 m thick) and, as such, are probably below the limits resolution of conventional surface seismic. In addition, the quality of available seismic data itself is often compromised by such issues as surface-related problems, multiples problems, and evaporites in the section, with the resulting data generally being of fairly poor quality – all of which prevent the detection of either the dolomite intervals directly or of the structural features (faults and fracture trends) that we have called upon to explain the high dolomite trends in the Arab-D. Seismic characterization of faults and fracture trends in the Arab-D reservoir is an area of on-going research at Saudi Aramco and our belief is that future developments will result in an improved ability to image these postulated faults and/or fracture zones in the future.

  • (1) Two different types of calcified dolomite, or dedolomite, occur in the upper Jubaila Formation Wadi Nisah: stratiform and non-stratiform.

  • (2) Wadi Nisah non-stratiform dedolomite contains lower oxygen isotope (average δ18O = -10.99‰) and much lower carbon isotope (average δ13C = -7.51‰) values than do equivalent subsurface non-stratiform dolomite; in addition, Wadi Nisah non-stratiform dedolomite is much more coarsely crystalline than is typical subsurface Ghawar non-stratiform dolomite.

  • (3) Wadi Nisah stratiform dedolomite contains similar oxygen isotope values (δ18O = -2.89‰) and only slightly lower carbon isotopes (δ13C = 0.98‰) relative to subsurface Arab-D stratiform dolomites.

  • (4) We suggest that differences in the structural and diagenetic architecture of these two types of dolomite controlled later diagenesis, with non-stratiform dolomite being more susceptible to late meteoric diagenesis than the horizontally bedded stratiform dolomite intervals. Because of their near-vertical orientation, the non-stratiform dolomites were much more susceptible to alteration during near-surface meteoric diagenesis than were the stratiform dolomites.

Appreciation is given to the Ministry of Petroleum and Mineral Resources and to the Saudi Arabian Oil Company for permission to publish this article. Special thanks goes to H.E. Engineer Y. J. Shinawi, Director-General of the Ministry of Petroleum and Mineral Resources-Eastern Province, who in addition to providing a helpful review of this manuscript, has generously and consistently supported such petroleum related research efforts as this current study. We also extend our thanks to Drs. Steve Franks, Harry Mueller, Wyn Hughes, and Bill Carrigan, who reviewed an earlier version of this manuscript. The final design and drafting by Nestor Niño Buhay is appreciated.


Dave L. Cantrell graduated from the University of Tennessee with an MSc in Geology in 1982, and from the University of Manchester with a PhD in Geology in 2004. He began his industry career with Exxon in 1982, where he conducted numerous reservoir characterization and geological 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. Dave has been responsible for several studies on large carbonate reservoirs since joining Saudi Aramco in 1997. He is presently Chief Geologist in Saudi Aramco’s E&PE Technology Department.

Abdullah Al-Khammash graduated from King Abdulaziz University with a BSc in Petroleum Geology in 1988, and then joined the Ministry of Petroleum and Mineral Resources in Dhahran that same year. He worked on a number of assignments at the Ministry and was the Ministry’s representative to Saudi Aramco’s Southern Area and Red Sea Exploration Department for about 2 years; he also attended a number of technical courses both inside and outside the Kingdom. Abdullah continued his education at the University of Manchester, where he graduated with an MPhil in Carbonate Sedimentology in 2003. He is currently Acting Director-General for the Ministry of Petroleum and Mineral Resources, Partitioned Neutral Zone (PNZ) Branch at Al-Khafji. Abdullah participated in a number of committees formed to create the future vision for the Technical Affairs arm of the Ministry; he also coordinated and was an active member in the study of karst features in the PNZ. He is currently a member of the European Association of Geoscientists and Engineers, the American Association of Petroleum Geologists, and the Dhahran Geoscience Society.

Peter D. Jenden received his PhD from the University of California, Los Angeles, and has 23 years experience applying geochemical principles to petroleum exploration and development. His primary interests include petroleum systems studies, the origin of natural gas, and geochemical indicators of reservoir compartmentalization. Peter joined Saudi Aramco in 2001 and currently supervises the stable isotope laboratory within the Geochemistry Unit of the Research and Development Center.