Lithofacies, diagenesis and reservoir quality of Upper Shu’aiba reservoirs in northwestern Oman Open Access

Reservoirs of the lower member of the Shu’aiba Formation have been extensively studied over the past three decades in many giant oilfields throughout the southeastern Arabian Plate (Frost et al., 1983; Litsey et al., 1983; van Buchem et al., 2010). In contrast, the Upper Shu’aiba Member is more restricted geographically (Figure 1) and contains relatively small oil volumes that have only recently begun to be exploited. Proportionately little information has been published on Upper Shu’aiba reservoirs, other than Boote and Mou (2003) on Safah Field in Oman and the comprehensive study of the Upper Shu’aiba in Abu Dhabi Emirate by Pierson et al. (2010) and Maurer et al. (2010). The present study provides details from Upper Shu’aiba oilfields in northwestern Oman, where exploration and development activities have recently increased. A series of cores were studied together with wireline logs and a reservoir model to characterize the nature and spatial distribution of lithofacies and petrophysical properties in several Upper Shu’aiba depositional cycles.

A particular focus of this study is the hiatus and sequence boundary at the top of the Upper Shu’aiba Member, where the limestones are overlain and sealed by shale from a distant deltaic source. Surfaces representing hiatuses and major paleoenvironmental changes in shallow marine carbonates, termed “discontinuity surfaces” by Immenhauser et al. (2000), are of great scientific and practical interest. They typically form the boundaries of both depositional packages and petroleum reservoirs, as well as being the locus of important diagenetic processes that determine petrophysical properties. Such surfaces are often enigmatic and intriguing: the condensed remnants of possibly complex events of global significance.

The studied reservoirs comprise several clinoform bodies of the Upper Shu’aiba Member that were deposited in Early Cretaceous time (Late Aptian) as the Bab intra-self basin was infilled by a series of low-angle clinoforms that prograded distances of 20–40 km inward from the basin margins (Figures 1 and 2). Upper Shu’aiba strata are confined within the Bab Basin and formed during a major low-stand of sea level that exposed basin margins and surrounding terrain consisting of Lower Aptian limestones of the Lower Shu’aiba Member (Pierson et al., 2010; Maurer et al., 2010). Successive clinoforms may be recognized in individual wells by alternating intervals of higher and lower gamma-ray activity. Maurer et al. (2010) illustrate examples of well data showing basal argillaceous intervals and overlying clean limestones comprising different Upper Shu’aiba clinoforms. In Oman, each clinoform is described as comprising a basal “argillaceous zone” (herein designated “J-arg” for the J clinoform) and an overlying “reservoir zone” of relatively clean limestone (designated “J-res” for the J clinoform).

Al-Husseini and Matthews (2010) and Maurer et al. (2013) propose that each pulse of Upper Shu’aiba clinoform progradation that is resolvable on available seismic data represents a cycle of glacio-eustatic sea-level fluctuation of approximately 400–500 Kyr duration. Well data show that several of these seismic-scale clinoforms are comprised of smaller clinoform units presumably reflecting higher-order cycles of eustatic fluctuations, as illustrated in figure 8 of Maurer et al. (2010).

Seismic imaging shows that individual Upper Shu’aiba clinoforms in Abu Dhabi have remarkable lateral continuity for tens of kilometers, which Pierson et al. (2010) suggest to result from: (1) a continuous narrow zone of high carbonate productivity paralleling the Bab Basin shoreline during each highstand, and (2) a system of persistent wind- or tide-driven longshore drift of siliciclastic fines. Seismic mapping indicates similar lateral continuity in northwest Oman (Figure 1b).

The Upper Shu’aiba clinoforms in northern Oman are informally identified within PDO by letter symbols A through O from younger to older, and an additional five clinoforms younger than the A clinoform were subsequently identified as comprising the final stages of basin infill. The six cored wells used in the present study are from an oilfield with production from the K and J clinoforms. These therefore represent the fifth and sixth clinoform cycles out of a series of twenty depositional cycles of ostensibly similar magnitude that have been recognized in northwest Oman as constituting the progradational infill of the Bab Basin during sequence Ap 5 of Figure 2. Identification of the exact position of the K and J clinoforms within the series of Upper Shu’aiba clinoforms recognized in Abu Dhabi by Pierson et al. (2010) is uncertain. Within the context of Figure 2, the individual Upper Shu’aiba clinoform cycles are schematically represented by the zig-zag contact between the light blue and the brown (Bab Member) colors in third-order sequence Apt 5.

In addition to the main coverage of the K and J clinoforms, one of the present cores (well 5) includes parts of the younger H and I clinoforms, and one core extends a few meters into the older L clinoform. Two additional cores were also described from nearby oilfields that produce from the A and B clinoforms, respectively. These cores were included to provide additional information on the range of lithofacies and petrophysical properties present in the overall Upper Shu’aiba depositional system. The main oilfield studied has an area of roughly 12 km along depositional strike by 4 km in the dip direction. The reservoirs are presently at their maximum burial depth of 1.4–1.5 km at 80°C. Most cores are located above the oil/water contact, but one (from well 4) is from the water zone.

The top of the Upper Shu’aiba Member corresponds to a hiatus along which the duration increases from the central part of the Bab Basin towards the edges. The maximum duration, overlying the oldest clinoform along the basin margins, could be as much as 5 million years (Figure 2). The maximum hiatus at the top of the K and J clinoforms of the present study would be somewhat less than 5 Myr because there are an additional four clinoforms (O through L) older than the K clinoform. The K and J clinoforms are thus the fifth and sixth clinoforms out of a total of 20 Upper Shu’aiba clinoforms identified in northern Oman. If these clinoforms each represent uniform time intervals, then individual durations of 250 Kyr are indicated. In this case, the hiatus at the top of the K clinoform may be around 4.75 Myr. Maurer et al. (2013), however, propose that each clinoform may represent a duration of 400–500 Kyr, in which case the top-K hiatus might be around 3.0 to 2.5 Myr.

During this time, the tops of the K and J clinoforms could have been subaerially exposed, alternately exposed and reflooded by seawater, or continuously submerged below shallow seawater. The clinoform tops may have been eroded, or may have accumulated minor additional sediment, or both, or may simply have been preserved with no loss or addition. Although the climate is reported to have been warm and humid during Early Cretaceous time, the Late Aptian is interpreted to have been relatively cooler, with evidence for major glaciation in high latitudes and 400–500 Kyr glacio-eustatic fluctuations that caused the Upper Shu’aiba clinoform cyclicity (Maurer et al., 2013).

The database for the present study includes 15 vertical wells from the main oilfield studied, four of which have a total of 127.9 m of whole core, plus 4 horizontal wells having a total of 38.4 meters of core. Wells A and B from two nearby oilfields are vertical and recovered 8.9 m and 7.4 m of core, respectively. Conventional core analyses (CCA) of helium porosity, gas permeability and grain density were performed by commercial laboratories using industry-standard methods.

A series of 195 core samples, mostly corresponding with CCA plugs, was selected to cover the ranges of depth and lithofacies in each of the cored intervals. The samples were examined to remove any material of extraneous origin, such as mud cake. All samples were analyzed for bulk chemical composition by SGS Minerals Services of Toronto, Canada, using a combination of x-ray fluorescence (Si, Ti, Al, Fe, Mn, Mg, Ca, K, P, Rb, Sr, Zr, Ba), neutron activation (Na, Cr, As, Br, Sb, REE, Hg, Th), coulometry (CO₂), specific ion (Cl) and Leco instrument (S and C). Uranium was determined by delayed neutron counting (detection limit = 0.1 ppm; Aldayel et al., 2002). A selection of 28 of these samples were also analyzed by bulk-rock X-ray diffraction (XRD) at Macaulay Scientific Consulting Ltd. of Craigiebuckler, Aberdeen, Scotland.

Bulk-carbonate oxygen- and carbon-isotope ratios were analyzed by the Environmental Isotope Laboratory of the Department of Earth and Environmental Science, University of Waterloo, Canada. Finely ground bulk-rock samples were reacted with 100% phosphoric acid in sealed vials for more than one hour at 90°C to dissolve all calcite and dolomite present. The CO₂ was extracted, purified and then analyzed with a GVI IsoPrime continuous flow isotope ratio mass spectrometer. Standards were run throughout each 60 sample batch, and duplicates were run approximately every tenth sample.

Thin sections were impregnated with blue epoxy to facilitate identification of pore spaces and are polished on both top and bottom surfaces for optimal petrographic resolution. Thin sections containing more than a few percent carbonate cement (thus excluding most mudstone, wackestone and shale samples) were stained over half their surface with alizarin red and potassium ferricyanide. The thin sections were described using a set of standard categories of information recorded in spreadsheet format. Biotic assemblages, cements and stylolites were noted by a system of numerical scoring. Percentages of visible porosity and total cement were estimated using comparator charts.

Cores were described on A3 sheets at a scale of 3.5 cm = 10 cm (roughly 1:3). Data from available thin sections, including 195 prepared for the present study and 118 older thin sections available from PDO, were posted on the core-description sheets, and an atlas of photomicrographs at standardized scale was consulted during core description.

The K and J reservoirs of the main oilfield studied were examined in a digital reservoir model in order to visualize their geometry with respect to the top-Upper Shu’aiba surface and the well locations. This model was made by PDO using data from wells and a three-dimensional seismic survey and was not altered for the present study. Using this model, a subcrop map was made showing where the different clinoform units and the vertical wells intersect the top-Upper Shu’aiba surface (Figure 3).

A series of dip-oriented cross-sections was made through the model showing the well and core locations with respect to the clinoform geometries (Figure 4). The cross-sections were then flattened on the base of the Nahr Umr Formation to better show depositional geometries (Figure 5). Data for the cored intervals and wireline log curves for gamma ray (GR) and total porosity were plotted at true vertical depth so that hole deviations would not affect apparent stratigraphic thicknesses (Figure 6). In most of the reservoir model area, the K clinoform is roughly twice as thick as the J clinoform. However, this relationship varies laterally, with the J clinoform becoming thicker than the K clinoform in the eastern part of the area.

To illustrate the main features of the K and J clinoforms, the individual cross-sections were combined into a composite dip-oriented cross-section, with model artifacts like apparent faults and folds removed (Figure 7). The 15 vertical wells from which log data are available were projected laterally into the composite cross-section in their approximate relative positions from distances of 0 to 8.5 km along strike.

Nine core-scale limestone lithofacies were defined in the Upper Shu’aiba Member based on study of thin sections and slabbed cores, and two siliciclastic lithofacies were defined in short intervals of core in the Nahr Umr Formation (Table 1; Figure 8). Distinctions between the limestone lithofacies are determined by the textural classification of Dunham (1962), as modified by Embry and Kloven (1971) and Lucia (1995), and by biotic assemblages. Care was taken to use the terminology of Lucia (1995) to differentiate between packstones which are “mud-dominated” (all intergranular space filled by mud) and “grain-dominated” (intergranular space filled partly by mud, but with significant proportions of open space or cement). The distribution of lithofacies and the textures determined in the thin sections are shown for each core in Figure 9. This figure may be compared with Figures 10 and 11 to see how lithofacies variations are expressed in the wireline-log curves and CCA data.

Lithofacies LF-7 and LF-9 are not readily distinguishable from LF-6 in thin section because their diagnostic fabric elements are much larger than the 2 to 3.5 cm dimensions of the thin sections used, and because of highly heterogeneous fabrics. In plots of plug-scale petrographic and petrophysical data, therefore, most samples from LF-7 and LF-9 are included with LF-6, except in a few cases where the material included in the thin section has entirely missed the larger fabric elements, resulting in a packstone or wackestone fabric at plug scale (LF-1, LF-2, or LF-3).

The positions of three main lithofacies belts in the clinoforms were interpreted from the core descriptions (Figure 7a). Lithofacies in the reservoir zones vary in a proximal direction from mudstone/wackestone to mud-dominated packstone (shown by comparison of the cores from wells 5 and 5H2 in zone J-res) and from mud-dominated packstone to floatstone, rudstone and boundstone (shown by comparison of well 5H2 versus wells 2H4, 4 and 1 in zone J-res and by comparison of wells 4 and 1 versus wells 1H4, 3H2 and 2 in zone K-res).

The distal-proximal core comparisons possible for the argillaceous zones are more limited. Zone J-arg varies from mudstone/wackestone in well 5 to mud-dominated packstone and floatstone in wells 4 and 1. This landward-coarsening trend is consistent with the wireline-log trends of decreasing GR and increasing porosity in the proximal direction in zone J-arg (Figure 6). In zone K-arg, however, no such landward trends are apparent. Mudstone/wackestone comprises the K-arg core intervals from both well 4 and well 2, and the wireline logs show no evidence of distal-to-proximal variation in this zone (Figure 6).

Vertical facies relationships observed in the different cores show overall patterns of upward increase in depositional energy and decreasing paleo-water depth from the argillaceous zones into overlying reservoir zones. Within the argillaceous zones, the same general direction of upward changes in facies is observed in some cores, but not in other locations:

• Zone J-arg shows upward transition from mud-dominated packstone to dominantly floatstone in wells 1 and 4.

• Zone J-arg also shows upward transition from mudstone to wackestone in well 5.

• Zone K-arg shows no apparent vertical facies trend in well 2.

• Zone I-arg shows no apparent vertical facies trend in well 5.

• Zone H-arg, corresponding to the top 8.8 m of the core in well 5, comprises two distinct facies trends. The lower 5.4 m of this 8.8 m interval shows an upward-shoaling trend from wackestone to packstone to a thin bed of floatstone. The upper 3.4 m of the core consists of mudstone and wackestone apparently representing a deeper-water setting than the floatstone and packstone.

No distinctive back-shoal lithofacies were identified in the present study, although back-shoal deposits consisting of bioturbated mudstone, wackestone and mud-dominated packstone are likely to be an important feature of these clinoforms, as was documented in Upper Shu’aiba cores from Abu Dhabi (Pierson et al., 2010). A back-shoal facies belt was not delineated in the present study, however, because of the limited core coverage of the most landward parts of the K and J clinoforms and because the cored intervals most likely to represent such deposits do not appear to have such distinctive characteristics as to support a separate classification.

In general, the K and J argillaceous zones are recognized by wireline-log character as having baseline-normalized GR values mostly > 23 API units and total porosity mostly < 10% (Figure 6). In contrast, the K and J reservoir zones have porosity mainly > 10% and comprise large volumes of rock with porosity of 20–30%. In addition, hydrocarbon saturations tend to be much lower in the argillaceous zones, although these values, of course, vary widely depending on position with respect to the oil-water contact.

The zone boundaries defined by the above criteria are also correlative between wells based on wireline-log profiles, especially using GR log shape. These correlations show, however, that the J-arg zone becomes poorly defined in the more proximal, up-dip direction, and is an exception to the above GR and porosity criteria in wells 4, 13, 1, 7, and 19. Here the J-arg zone has relatively low GR (around 20–23 API units or even lower in some intervals) and log porosity values somewhat higher than 10%. Well 10 is also in a proximal position in zone J-arg, but has typically “argillaceous” high GR and low porosity (Figure 6). Zone K-arg, on the other hand, does not display any apparent decrease in argillaceous character in the proximal direction, at least as far as the available well coverage shows.

Examination of the core and thin-section descriptions, in comparison with the zone boundaries defined by the above criteria and correlations, shows that textures and lithofacies overlap extensively between the reservoir and argillaceous zones. Floatstone texture, for example, occurs within both the reservoir zones and argillaceous zones. Furthermore, the top of a reservoir zone or an argillaceous zone in any given core commonly occurs within an interval of a particular lithofacies, rather than at the bed boundary between two different lithofacies (Figures 6 and 9).

Nevertheless, the argillaceous zones contain higher proportions overall of finer-grained, mud-supported lithofacies. Also, where zone boundaries occur within beds of the same texturally defined lithofacies, there are subtle, but statistically important petrologic contrasts coincident with the zone boundary. For example, the boundary between zone K-res and the overlying zone J-arg occurs within an interval of mud-dominated packstone in wells 1 and 4 (Figures 6 and 9). However, the K-res packstones tend to have slightly higher bulk-rock alumina, more stylolites, and lower porosity and permeability than the K-arg packstones (Figure 10).

Bulk-rock XRD analyses show that clay minerals in the Upper Shu’aiba limestones are kaolin, mixed-layer illite-smectite and illite (Figure 12), with traces of chlorite present in some samples. This assemblage is similar to that of the 3 samples analyzed from overlying Nahr Umr Formation. No consistent differences in clay types or relative proportions are observed between reservoir zones and argillaceous zones, but samples from the younger clinoforms A and B and from the Nahr Umr goethite-ooid sandstone in well A tend to have higher ratios of illite/smectite to kaolin than the samples from the main oilfield studied (Figure 12a).

The bulk-rock alumina content of the Upper Shu’aiba limestones should mainly reflect clay abundance because clay is the main aluminous mineral present. The bulk-XRD data also show minor feldspar, but in amounts mostly below 1%. The correlation between bulk-XRD total clay and alumina indicates that a bulk alumina content of around 5 to 6 wt% in these limestones should correspond with a clay content of approximately 20% (Figure 13). The most aluminous and clay-rich samples are from zones K-arg and L-arg in well 2.

Bulk-rock potassium is a second chemical component that can be expected to provide a measure of clay abundance in these limestones. Non-correlation between alumina and potassium, however, indicates variable contamination with potassium chloride drilling fluid, which was used in several of the wells. Sylvite is also recorded in minor amounts (0.1–0.2 wt%) in the bulk XRD analyses of many samples.

GR activity measured by wireline logging is linearly proportional to abundances of the three main elements possessing natural radioactivity (Rider and Kennedy, 2011).

GR calculated from the present bulk-chemical analyses, however, shows little or no correlation with wireline-log GR values corresponding with individual sample depths. This poor correlation could result from the plug-size samples being only poorly representative of the heterogeneous distributions of radioactive elements within the larger rock volumes measured by the in-hole logging tool, as well as the presence of extraneous potassium as KCl or other mud additives.

GR activity in the Upper Shu’aiba limestones reflects two independent factors: (1) potassium and thorium within fine-grained siliciclastics (mainly clay) and (2) uranium of unknown but presumed diagenetic origin. Overall linear correlation of bulk-rock alumina with both thorium (Figure 14a) and potassium supports the interpretation that these components mainly reflect siliciclastic fines content. Most of the analyzed limestones fall along trend A in Figure 14a, connecting zero content with the five Nahr Umr shales analyzed, consistent with the alumina and thorium in the limestones being contained in siliciclastic fines similar in composition to the shales. Examination of the more aluminum-rich mudstones indicates that the clay is included as an integral component of the depositional mud matrix. Several samples, however, plot closer to trend B in Figure 14a, connecting zero content with two Nahr Umr goethite-ooid sandstones from just above the top of the Upper Shu’aiba Member in well A. These sandstones have distinctly higher Th/Al than the shales. The limestones closest to trend B are from 3 and 11 cm below the contact with the sandstone bed and contain clay filling vugs and intergranular spaces, suggestive of infiltration of fines at the time of sandstone deposition.

As has been noted to be a general characteristic of carbonate strata elsewhere (Fiet and Gorin, 2000; Ehrenberg and Svånå, 2001; Ehrenberg et al., 2006; Lucia, 2007), uranium shows no overall correlation with other chemical components, including alumina (Figure 14b) and thorium, indicating that uranium is not primarily associated with clay. Uranium variations are not extreme (0–5 ppm; Figure 15), but are large enough to affect the total GR log to a similar degree as the variations in potassium and thorium. In particular, high GR values at the top of the Upper Shu’aiba Member in wells 1 and 2 (Figure 6) appear to mainly reflect uranium abundance. In these locations, uranium shows overall but somewhat erratic trends of upward increase that could reflect processes related to the top-Upper Shu’aiba surface, as discussed below. Iron and manganese are also enriched in these intervals (Figure 15), although these three elements show very poor mutual correlation, both within these specific intervals and overall.

Diagenetic processes observed or inferred to have affected the Upper Shu’aiba cores are described in Table 2 in approximate paragenetic sequence. Nevertheless, the timing of several processes, such as calcite cementation and chemical compaction, is uncertain and may well have overlapped over extended intervals of time. Petrographic observations indicate that nearly all former macropores (pores large enough to be visible in thin section) in the cores from the K and J clinoforms have been filled by blocky calcite cement (Figure 16). Therefore, the wide range in total porosity of all lithofacies mainly reflects variations in microporosity of mud matrix. Micropores within micritized bioclasts probably also contribute to the total porosity values. The grainstones and grain-dominated packstones of the A clinoform, however, contain abundant primary macropores that are not filled by calcite cement.

Porosity information was added to the composite cross-section (Figure 7b) by first marking intervals with wireline total porosity generally higher than 20% and lower than 10% on each well profile (Figure 6) and then transferring these intervals to the cross-section. Correlation of these intervals allowed mapping of regions with correspondingly high and low porosity.

The cross-plot of CCA data in Figure 11a shows how different lithofacies have somewhat different ranges of porosity and permeability. For the entire dataset, log₁₀(permeability) shows broad linear correlation with porosity, but varies by roughly 1.5 log units for any given porosity. Most samples plot within or below rock-fabric field 3 of Lucia (1995), corresponding to limestones where interparticle space occurs mainly within carbonate mud matrix. The lithofacies with larger and more abundant skeletal clasts have statistically higher permeability for given porosity, but there are large degrees of overlap between all lithofacies. The Lucia rock-fabric fields are intended to be used with interparticle porosity rather than total porosity, but vugs comprise only a minor proportion of total pore space in the present samples because most macropores, whether primary or moldic, have been filled by blocky calcite cement.

Figure 11b shows the contrasting but overlapping ranges of porosity and permeability of the reservoir zones versus the argillaceous zones. Most plug samples from the argillaceous zones have porosity of 10–20%, which contrasts with the observation that the wireline logs show mostly < 10% porosity in the argillaceous zones (Figure 6). This apparent discrepancy reflects the relatively high porosity of the particular intervals of the argillaceous zones that are included in the cored intervals: J-arg in wells 1 and 4 and H-arg and I-arg in well 5. As noted above, these J-arg cores are located in the more proximal part of zone J-arg where clay content is relatively lower and porosity higher than elsewhere in the argillaceous zones. The well 5 cores are also located in relatively proximal positions within the H and I argillaceous zones.

Vertical profiles of petrologic measurements in the cored well sections were examined for evidence of trends that might reveal diagenetic processes related to the top-Upper Shu’aiba surface. For some variables, no trends are apparent. Profiles of porosity and permeability show no evidence of systematic increase or decrease as a function of distance below the top-Upper Shu’aiba surface (Figure 10). Neither are any trends in abundance of calcite cement, vuggy porosity, or other textures indicative of dissolution observed independent of lithofacies variations underlying the top-Upper Shu’aiba surface.

Profiles of bulk-carbonate oxygen- and carbon-isotope composition (Figure 17) reveal no trends of consistent upward decrease immediately underlying the top-Upper Shu’aiba surface, such as have been documented in numerous locations just below the top-Lower Shu’aiba surface (Immenhauser et al., 2000; Vahrenkamp, 2010). A cross-plot of these data (Figure 18) and comparison with published carbon-isotope profiles from the Upper Shu’aiba Member (Figure 19) show the overall similarity between the present values and earlier determinations.

Profiles of bulk-rock strontium were examined (Figure 20) because strontium depletion has been reported as an indicator of meteoric diagenesis below paleo-exposure surfaces in carbonate strata elsewhere (Wagner et al., 1995; Railsback et al., 2003). Well 1 shows a trend of upward-decreasing strontium through the 6.5 m below the top-Upper Shu’aiba surface, and a possible but inconstant trend of overall upward decrease is apparent in zone H-arg of well 5. Similar trends are not seen in the other well cores.

As noted above, uranium displays apparent trends of upward increase just below the top-Upper Shu’aiba surface in the cores from wells 1, 2 and A. These intervals also display possible trends of upward enrichment of iron, manganese and phosphorous (Figure 15).

Saddle dolomite is especially abundant within the upper 7 meters of limestone in well 1, just below the top-Upper Shu’aiba surface (Figure 21). The most Mg-rich of these samples contains 1.48 wt% MgO, which is equivalent to about 7 wt% dolomite having MgO content of 20 wt%, after deduction of a small amount of MgO for minor clay present. Saddle dolomite is also present in most samples from the top 10 m of the Upper Shu’aiba Member in wells 2 and 4, as well as at locations distant from the top surface in wells 1 and 2 and virtually throughout the core from well 4. Bulk-rock magnesium content provides an indication of dolomite abundance, but dolomite is not the only Mg-rich mineral present.

Ideal, pure dolomite has around 22 wt% MgO, but magnesium is also likely to be contained within illite/smectite clay, which comprises around 12 wt% of the most argillaceous mudstones (in zone K-arg of well 2). These mudstones contain as much as 1.1 wt% MgO, but show nearly no dolomite in the one XRD analysis available. Bulk-rock MgO content therefore reflects the combined abundances of saddle dolomite and illite/smectite (Figure 21). In samples with low alumina content, the MgO is mainly held in dolomite, but where alumina is abundant, the MgO is likely to be contained mainly in illite/smectite, especially where no dolomite is observed in thin section.

The most typical mode of occurrence of the dolomite is as euhedral crystals generally 0.2–0.7 mm in maximum dimension enclosed by blocky calcite cement crystals of similar or larger size (Figure 22). Petrographic relationships are ambiguous as to the relative timing of the dolomite and the enclosing coarse calcite cement. Although the calcite surrounds and thus appears to postdate the dolomite crystals, it is also possible that the coarse calcite has been the preferential site for replacement by later dolomite. Similar textures are shown by Scholle and Ulmer-Scholle (2003; p. 386-387), who note the same ambiguity regarding relative timing.

The only calcite inclusions seen within the dolomite appear to have early “dog-tooth” morphology, so these are not diagnostic of the timing with respect to the later coarse calcite spar (Figure 22). In rare cases, dolomite crystals have irregular boundaries with truncated crystal zones against surrounding coarse calcite, but this could result from either dolomite dissolution before later calcite growth or varying replacement of earlier calcite. In only one sample was clear evidence of relative timing observed, where saddle dolomite has grown into a large vug and overlies earlier coarse calcite crystals that partly filled the vug.

Infiltration of clay just below the top-Upper Shu’aiba surface is apparent within the upper one meter of the limestone in well 1, where clay occurs concentrated in a few small (< 1 cm) masses that may fill vugs or fractures. The exact nature of these features is indistinct, however, because of the delicate and crumbly (friable) condition of the core surface. Clay is not observed in the thin sections prepared, except in a highly altered limestone exactly at the top surface, where areas of silty clay are extensively replaced by pyrite. As described above in connection with Figure 14, a second place where clay appears to have infiltrated into limestones just below the top-Upper Shu’aiba surface is in well A, where the top-Upper Shu’aiba surface is overlain by a bed 43 cm-thick of goethite-ooid sandstone having a distinctive potassium/thorium ratio.

The present observations appear entirely consistent with the previous interpretation of the Upper Shu’aiba Member as comprising a series of progradational cycles formed in response to glacio-eustatic fluctuations in sea level (Pierson et al., 2010; Maurer et al., 2010, 2013). In this context, the lesser thickness and volume of the J clinoform, as apparent in Figures 5 and 7, could reflect differences in the magnitude of the sea-level fluctuations responsible for each depositional cycle. As illustrated by Matthews and Frohlich (2002), eustatic sea-level fluctuations driven by orbital dynamics are in general the composite of several interfering frequency oscillations. Some stratigraphic cycles must therefore be of greater magnitude than others. On the other hand, the variation in relative size of the K and J clinoforms along strike (Figure 5) and the limited strike distance included here (Figure 3) indicate that any such inferences within the limits of the present study are highly uncertain.

Insofar as the K and J clinoforms are interpreted as depositional sequences recording fourth-order cycles of sea-level cycles (Maurer et al., 2013), it is of interest to consider how the component transgressive and regressive hemicycles of each clinoform cycle should best be represented; in other words the position of the maximum flooding surface (MFS) within each clinoform. Pierson et al. (2010) show a general cross-section model for Upper Shu’aiba clinoforms in which the MFS is placed roughly in the middle of each package. Although these authors do not use the term “argillaceous zone,” this MFS placement would appear to be approximately at the top of each argillaceous zone. Maurer et al. (2010) in their figures 6–9, however, show transgressive and regressive hemicycles for various Upper Shu’aiba clinoform cores where the MFS position is quite variable with respect to the top of the argillaceous mudstone intervals in each example.

For K and J clinoforms of the present study, placement of the MFS at the top of the argillaceous zone appears inconsistent with the apparent absence of facies trends suggestive of upward-increasing water depth through the argillaceous zones in the studied cores. The J-arg zone of wells 1, 4 and 5 displays a trend of upward-increasing ratio of grains to mud. Neither do any of the vertical wells examined show evidence of a maximum in gamma-ray activity, such as might reflect a flooding event, near the top of any of the argillaceous zones. A possible solution is to place the MFS near the base of each argillaceous zone, such that the transgression initiating each clinoform is represented by only a surface or thin interval, with nearly all sediment volume being deposited during the highstand. This would be consistent with observations from the Pleistocene 100 Kyr sea-level cycles that rising stage of climate-driven sea-level oscillations tends to be more abrupt than the falling stage (Boulton, 1993).

Within the context of the Upper Shu’aiba clinoform model (Pierson et al., 2010), the elevated clay content of the argillaceous zones is consistent with sedimentation following a fall in sea level because each lowering of sea level would increase the gradients of streams draining into the Bab Basin across the exposed Lower Shu’aiba platform, thus increasing their siliciclastic load and the flux of fine suspended detritus entering the Bab Sea. The rather abrupt drop in clay content at the top of each argillaceous zone is suggestive of a threshold effect (non-linear response), whereby the rate of fines influx decreased or the rate of carbonate production increased relatively abruptly after each sea-level highstand was attained.

The lower porosity of the argillaceous zones with respect to the reservoir zones may be inferred to result from processes whose rates are in large degree dependent on clay mineral abundances, based on the overall inverse relationship between total porosity and bulk-rock alumina shown in Figure 23. Correlation does not, of course, necessarily imply causation, but the effect of clay on promoting porosity loss in carbonates during burial has long been appreciated (Bathurst, 1975; Choquette and James, 1987; Brown, 1997).

The relationship in Figure 23 is not linear, however, as many samples have both low porosity and low alumina content. The widest range of alumina occurs in the mudstone/wackestone lithofacies, but the other, less aluminous lithofacies also show similar patterns. Other possible porosity controls are not apparent in the present set of samples, insofar as comparisons between high-porosity samples and low-porosity samples within the same lithofacies have not revealed any consistent differences in depositional textures, biotic content, dissolution features or in chemical components that are not themselves correlated with alumina.

The overall negative relationship in Figure 23 is envisioned to result from two independent processes whereby compaction and cementation of mud matrix should be accelerated by the presence of depositional clay. At shallow depths, early cement growth around carbonate mud particles may be expected to increase rigidity and thus resistance to mechanical compaction. Where clay occurs between carbonate particles, locking cement bridges cannot form, and thus the presence of clay particles dispersed in the carbonate mud should enhance early mechanical compaction by reducing the frequency of cemented interparticle contacts.

The second way in which clay should accelerate porosity loss is by facilitating “pressure dissolution” of adjacent calcite. The positively charged surfaces of clay crystals tend to locally increase the solubility of calcite, resulting in a concentration gradient that drives diffusion of the dissolved carbonate ions toward sites of cement precipitation in the surrounding rock. Porosity loss thereby occurs by both chemical compaction and precipitation of the dissolved calcite in surrounding pore spaces. Many of the Upper Shu’aiba limestones contain stylolites, all of which appear to be lined with clay minerals. Stylolitic dissolution is thought to occur during burial and is a plausible source for the abundant calcite cement present. Clay content can thus be the principal reason for the low porosity of the Upper Shu’aiba argillaceous zones.

The stable-isotope variations observed in the present study may be compared with the values reported for Upper Shu’aiba cores by Vahrenkamp (2010) and Pierson et al. (2010). Vahrenkamp (2010) constructed a composite carbonate carbon-isotope curve for both the Lower and Upper Shu’aiba members using data from different cored intervals for different age segments. For comparison with this curve, data from the present study were plotted on a common depth scale with depths shifted to match reservoir zone boundaries (Figure 19). Vahrenkamp’s curve was then traced onto this plot (the dashed bright-green line in Figure 19) with the vertical (depth) scale adjusted to match the depth scale of the present cores (as shown by the black arrows in Figure 19). Nevertheless, the depth correlations attempted here are highly uncertain. The upper part of the Vahrenkamp curve is from “Field D” in Oman, which contains clinoforms younger than clinoform A according to internal PDO data. The lower part of the trend is from “Field A” in Abu Dhabi, which contains the oldest Upper Shu’aiba clinoforms, perhaps corresponding with clinoforms O, N and M in the PDO nomenclature. The data from the present study may therefore be more properly correlated as lying somewhere in between the depth ranges of Vahrenkamp’s Field D and Field A.

Pierson et al. (2010; their figures 9 and 12) show a profile of stable-isotope data through an Upper Shu’aiba clinoform cored in a well from Abu Dhabi. According to figure 6a of Maurer et al. (2010), the core analyzed is from the oldest clinoform deposited following the sequence boundary that terminated Lower Shu’aiba deposition. The Pierson et al. (2010) profile shows a trend of upward-decreasing δ¹³C through the 38 meter-thick basal argillaceous zone (from around 4‰ at the base to around 2‰ at the top), followed by a return to somewhat higher values (2‰ to 3‰) in the overlying 9 meter-thick reservoir zone of clean limestone. These values coincide with the lower part of the range of δ¹³C in the present data (Figure 19).

According the Vahrenkamp (2010), Shu’aiba diagenesis took place as a closed system with respect to exchange of carbon, but was open with respect to oxygen, as may be expected when diagenesis takes place under conditions of low to moderate water/rock ratios (Lohmann, 1988). The carbon-isotope values should therefore mostly record fluctuations in the isotope composition of Lower Cretaceous seawater. In general the carbon-isotope values of the present study approximately match the Vahrenkamp (2010) composite curve, but show considerably greater variation. The Upper Shu’aiba part of the Vahrenkamp curve ranges from 2.0‰ to 4.5‰, whereas the present data range from roughly 1.5‰ to 6.4‰.

The oxygen-isotope values determined in the present study (mostly -3‰ to -7‰, with a few values in the range -1‰ to -3‰) are similar to the range of δ¹⁸O for the entire Shu’aiba Formation shown in figure 4 of Vahrenkamp (2010). Vahrenkamp (2010) noted that most Shu’aiba oxygen-isotope values are distinctly lower than the range expected for Lower Cretaceous marine carbonate (around -2‰ to -3‰ according to Raven and Dickson, 2007) and concluded that this difference reflects diagenetic alteration throughout the formation. Oman was located near 0° latitude in Early Cretaceous time (Sharland et al., 2001), and meteoric water at low elevations near the equator is expected to have oxygen-isotope composition very similar to seawater (Raven and Dickson, 2007). Therefore, near-surface diagenesis, whether involving seawater or meteoric water, would result in calcite having δ¹⁸O similar to the Lower Cretaceous marine carbonate values of -2‰ to -3‰. The most plausible explanation for the generally lower Upper Shu’aiba δ¹⁸O values (mostly -3‰ to -7‰) is thus precipitation of calcite cement at elevated temperatures during burial.

The patterns of upward-increasing uranium and associated GR just below the top of the Upper Shu’aiba Member in wells 1 and 2 (Figure 15) could reflect enrichment by sea-floor authigenesis. This hypothesis is consistent with the apparent enrichment of manganese and iron within the same intervals. Manganese and uranium both have multiple valence states and are least soluble in their relatively reduced state. Fixation from seawater or from the top of the sediment column is therefore facilitated by decay of organic matter in the sediments, which consumes available oxygen. Uranium solubility decreases with transition from the +6 to +4 valence state (Hostetler and Garrels, 1962; Doi et al., 1975; Langmuir, 1978; Murphy and Shock, 1999). Manganese behaves in a manner similar to U in sea-floor settings (Mangini et al., 2001).

Core and thin-section observations indicate that iron has been concentrated by the formation of pyrite at and just below the top-Upper Shu’aiba surface in well 1 and well A. Iron sulfide can form in marine sediments by reaction between dissolved sulfate from seawater and iron derived from the sediments (Berner, 1980). As dissolved oxygen is consumed by decay of organic matter deposited with the sediments, iron in the sediments is reduced to the more soluble ferrous state and enters the pore water. The dissolved Fe tends to diffuse upwards toward the supply of dissolved sulfate from the sediment/water interface.

Elsewhere, however, localized enrichment of uranium has been interpreted to have occurred by flow of burial-diagenetic fluids (Hassan et al., 1975; Fertl et al., 1980; Hoff et al., 1995; Luczaj, 1998) and, in other cases, by evaporative movement of groundwater upwards toward a subaerial exposure surface (Rawson, 1980; Carlisle, 1983; Carr and Lundgren, 1994). Various alternative processes could therefore be responsible for the observed patterns. As documented in Figure 21, saddle dolomite is also especially abundant at the tops of wells 1 and 2, so it is possible that late diagenetic fluid circulation formed both the dolomite and the small uranium anomalies.

Because saddle dolomite is known to form only at temperatures above 60°C, with most documented temperatures above 90°C (Spötl and Pitman, 1998), this must be a late burial feature in the present cores. Because the dolomite is of rather minor abundance, and especially if it formed by replacement of earlier calcite cement, dolomitization has likely had insignificant effects on porosity and permeability. The question of relative timing with respect to the enclosing coarse calcite spar cannot be resolved from available information. However, the replacement hypothesis can be explained by the process of “dolomite-growth-driven pressure solution of the calcite host” proposed by Merino and Canals (2011). By this means, dolomite grows as Mg-rich water infiltrates the limestone. The force of dolomite crystal growth causes pressure-solution of adjacent calcite, releasing dissolved calcium to increase the ion-activity product for dolomite and thus promote further dolomite growth.

The observed occurrence of saddle dolomite in proximity to the top-Upper Shu’aiba surface in well 1 and to lesser degree in well 2 (Figure 21) is suggestive of a genetic relationship. One possible source of magnesium for dolomitization is from illitization of smectite in the Nahr Umr shales. This mechanism has been proposed to explain spatial correlation between dolomite and clay in other strata (McHargue and Price, 1982). The present reservoir temperature in wells 1 through 5 is around 80°C, which is high enough to have enabled at least partial illitization of smectite in shales. Reaction of detrital smectite to form illite typically starts at around 60°C and is completed by 120°C in shales of the United States Gulf of Mexico and the Norwegian continental shelf (Hower et al., 1976; Peltonen et al., 2009; Nadeau, 2011). Smectite can contain varying amounts of both Mg and Fe, which are partly released upon reaction to form illite (Boles and Franks, 1979). An alternative source of Mg for minor late dolomitization is influx of water derived from a deeper source. In the other cores studied, no relationship between saddle dolomite and the top-Upper Shu’aiba surface is apparent, and the dolomite seems to occur more or less haphazardly at various places.

Profiles of bulk-carbonate oxygen- and carbon-isotope composition and bulk-rock strontium content were examined because trends of upward-decrease in these components have been documented as indicators of meteoric diagenesis underlying subaerial exposure surfaces (Allan and Mathews, 1982; Marshall, 1992; Railsback et al., 2003; Theiling et al., 2007). Bulk-rock Sr content of carbonates can be reduced from high marine values when Sr-rich aragonitic components are dissolved and reprecipitated as low-Mg calcite. This can occur under marine phreatic conditions, but is especially rapid during meteoric diagenesis (Friedman and Brenner, 1977; Morrow and Mayers, 1978). Examples of depletion in bulk-rock Sr below paleo-exposure surfaces in carbonate strata are shown by Wagner et al. (1995) and Railsback et al. (2003).

Neither bulk-calcite stable-isotope values nor bulk-rock strontium concentrations show definite trends indicative of upward-increasing meteoric diagenesis below the top-Upper Shu’aiba surface. Well 1 does show a trend of upward-decreasing Sr content in the 6.5 m interval just below the top-Upper Shu’aiba surface (Figure 20), but the range of variation (from around 300 ppm to a minimum of 150 ppm in the topmost sample) seems insignificant in comparison with the range of < 300 to 1,500 ppm in stratigraphic sections elsewhere that have been suggested to reflect meteoric leaching profiles (Railsback et al., 2003).

Porosity and permeability data also show neither increasing nor decreasing vertical trends suggestive of surface-related diagenesis (Figure 10). Neither are vertical trends in CCA values apparent at the top of the K-res zone where it underlies J clinoform beds in wells 1 and 4 (Figure 10).

Despite the absence of evidence for increased meteoric diagenesis immediately below the top-Upper Shu’aiba surface, it is possible that the entire reservoir zone in both the K and J clinoforms has experienced extensive meteoric diagenesis. Virtually all aragonitic skeletal material throughout the studied cores was dissolved, including corals, sclerosponges, gastropod shells and the inner layer of rudist shells, before the moldic macropores thus produced were subsequently filled by blocky calcite cement.

Rameil et al. (2012) describe the surface at the top of the Lower Shu’aiba Member, which has been encountered in wells throughout southeastern Arabia and studied in outcrops in Oman. Immenhauser et al. (2000) presented an earlier version of the same results, but including descriptions of several other, quite similar Cretaceous discontinuity surfaces within and at the tops of the Nahr Umr and Natih formations. The top-Lower Shu’aiba was subaerially exposed by a major fall in sea level at 121–118 Ma and was transgressed between 113–111 Ma by Nahr Umr siliciclastics (Figure 2). The exposure duration should thus be approximately 5–10 million years. This sea-level fall is confirmed by stratal geometries of the Upper Shu’aiba clinoforms in the Bab and Kazdumi basins (Pierson et al., 2010) and by karst fissures and a major siliciclastic-filled incised valley in Qatar (Raven et al., 2010).

Nevertheless, the top-Lower Shu’aiba surface shows a general lack of evidence for karst formation and only subtle indications of meteoric diagenesis. Minor negative excursions in bulk-carbonate stable-isotope profiles are interpreted to reflect soil formation. Vahrenkamp (2010) reports additional occurrences of negative excursions in bulk-carbonate carbon-isotope composition underlying the top-Lower Shu’aiba surface. The top-Lower Shu’aiba surface was also affected by sea-floor diagenesis, as evidenced by a dark brown/black coating a few millimeters thick and borings by marine organisms penetrating the dark coating material. Rameil et al. (2012) note that these effects indicate only relatively short periods of submarine exposure relative to the major duration of the hiatus. Uranium data have not been reported, but possible evidence of uranium enrichment just below the top-Lower Shu’aiba surface is provided by the GR profiles reported by Vahrenkamp (2010), where two locations show patterns of upward increase in GR (well V in his figure 8 and well 2 in his figure 11) that closely resemble the patterns attributed to uranium enrichment in wells 1 and 2 of the present study.

Comparison between the top-Upper Shu’aiba and top-Lower Shu’aiba surfaces thus shows that both show signs of relatively modest and laterally varying sea-floor diagenesis, although this evidence is more ambiguous in the former. The top-Upper Shu’aiba surface shows no evidence of surface-focused meteoric diagenesis in the limited areas examined, whereas the top-Lower Shu’aiba surface shows locally developed negative stable-isotope excursions, root traces, bleaching and vadose cementation.

The main findings of this study can be summarized as follows.

• The two Upper Shu’aiba clinoforms which are the main focus of this study each consist of a basal argillaceous zone and an overlying reservoir zone, the former having overall higher GR (mostly > 23 API units) and lower porosity (mostly < 10%). Lithofacies vary in a proximal direction from mudstone/wackestone to mud-dominated packstone to mud-rich floatstone, rudstone and boundstone. The reservoir zone of a younger clinoform, cored in a nearby oilfield, comprises well sorted grainstone and grain-dominated packstone, representing a very different, more current-winnowed depositional environment.

• The higher clay abundance in the base of each clinoform likely reflects both increased influx of siliciclastic fines during periods of lower sea level and lowered carbonate productivity during transgression. Low porosity and permeability in these intervals results from the role of clay in facilitating both mechanical and chemical compaction.

• The top-Upper Shu’aiba surface displays local patterns of upward increase in uranium, manganese and iron in the immediately underlying limestone that may record sea-floor mineralization. However, late formation of saddle dolomite has also especially affected the same intervals, so interpretation of the timing of these elemental enrichments is uncertain.

• Profiles of bulk-rock strontium concentration, bulk-carbonate oxygen- and carbon-isotope ratios, petrographic data and conventional core analyses of porosity do not show evidence of increased meteoric diagenesis immediately below the top-Upper Shu’aiba surface in any of the wells examined. Meteoric leaching, forming abundant molds of aragonitic macrofossils, may nevertheless be pervasive throughout the studied reservoir intervals.

We thank the Ministry of Oil and Gas of the Sultanate of Oman and Petroleum Development Oman (PDO) for permission to publish this paper. The work reported here is the basis for two Master of Science theses, wherein more complete documentation is available (Al Habsi, 2013; Al Shukaili, 2013). PDO provided the data used in this research and access to the core samples studied. The reservoir model used was built by Ibrahim Said Ahmed Al-Aghbari of PDO. Schlumberger Oman & Co. LLC generously provided PETREL and TECHLOG software and training for the present research. We thank Felix Schlagintweit (University of Leoben, Department of Applied Geosciences and Geophysics, Leoben, Austria) for identifying the crab legs bioclast type from our thin sections. This information was conveyed thanks to Mariano Parente (Università degli Studi di Napoli Federico II, Dipartimento di Scienze della Terra, Napoli, Italy) and Thomas Stueber (The Petroleum Institute, Abu Dhabi, United Arab Emirates). The two GeoArabia anonymous reviewers are thanked for helpful comments which improved the manuscript. We thank Florian Maurer for managing the review process and Kathy Breining for proofreading the manuscript. GeoArabia’s Production Co-managers, Nestor “Nino” Buhay IV and Arnold Egdane, are thanked for designing the paper for press.

Naima R. Al Habsi received a BSc in Geophysics and an MSc in Petroleum Geosciences from Sultan Qaboos University, Oman. She joined Shell in July 2013 as a Reservoir Engineer for the Land East Sour Gas System (GZI). She received best poster award in the EAGE Forum on Students and Young Professionals in Abu Dhabi in 2013. She is a member of EAGE, SEG and SPE.

[email protected]

Mohammed Al Shukaili holds a BSc in Geology and an MSc in Petroleum Geosciences from Sultan Qaboos University, Oman. He is currently employed with the Ministry of Commerce and Industry, Muscat, Oman.

[email protected]

Sabah Al Tooqi holds a BSc in Geology from Kuwait University and an MSc in Petroleum Geosciences from Sultan Qaboos University, Oman, where she was funded by a scholarship from Petroleum Development Oman. She started her career recently working with static modeling in Saih Nihayda cluster in the Gas Team of PDO. She is a member of SPE.

sabah.sn.tooqi@pdo. com

Stephen Neville Ehrenberg holds a PhD from the University of California at Los Angeles, USA. He is currently working at the The Petroleum Institute, Abu Dhabi, UAE. Steve was Shell Chair in Carbonate Geosciences at Sultan Qaboos University, Oman, from 2010 through 2013 and had previously enjoyed a long career with Statoil in Stavanger, Norway. His research concerns reservoir quality in sandstones and carbonates. He is a member of AAPG.

[email protected]

Michaela Bernecker studied Geology and Paleontology at the University of Erlangen, Germany. For her PhD dissertation on carbonate platforms and reefs (1995) she did field work in the Hajar Mountains on Permian and Triassic formations. She became Assistant Professor at the University of Erlangen and worked on her research project on Paleogene sedimentary environments on the Arabian Shelf funded by the German Research Foundation (DFG). Since 2008 Michaela has been an Associate Professor in Applied Sedimentology and Stratigraphy at the German University of Technology in Oman. Her current projects focus is on carbonate facies characterization and interpretation of outcrop and subsurface data for industrial applications on the Arabian Peninsula.

[email protected]