Abstract The accurate prediction of reservoir quality is, and will continue to be, a key challenge for hydrocarbon exploration and development. Prediction is a logical and critically important extension of the description and interpretation of geological processes. However, in spite of the profusion of publications on sandstone and carbonate diagenesis, relatively few articles illustrate the application of such studies to reservoir quality prediction. This Memoir represents the first attempt to compile worldwide case studies covering some predictive aspects of both siliciclastic and carbonate reservoir characteristics. We have attempted here to focus on the variability due to diagenetic effects in sandstones and carbonates, rather than on sedimentological effects, i.e., the presence or absence of a given reservoir. The chapters cover the spectrum of stages in the explorationexploitation cycle (Table 1). The importance of reservoir quality in pay evaluation has been illustrated by Rose (1987), who analyzed an unnamed company's exploration results over a 1-year period. Of 87 wildcat wells drilled, 27 were discoveries (31 % success rate); incorrect predictions of the presence of adequate reservoir rocks were made in 40% of the dry holes. Importantly, the geologists believed that reservoir quality was the primary uncertainty in 79% of the unsuccessful wells. Similarly, a comparison of predrill predictions with postdrill results by Shell (Sluijk and Parker, 1984) indicated that reservoir quality was seriously overestimated, whereas hydrocarbon charge and retention predictions were more accurate. Although these statistics do not clearly separate drilling failure due to lack of potential reservoir from the lack of adequate reservoir quality, it seems that although explorers are aware of the significance of reservoir quality prediction, generation of predictive models continues to be a formidable task.

Abstract In frontier basins where subsurface data are limited, or absent altogether, the study of reservoir rocks exposed in surface outcrops may be the dominant (or only available) means of predicting subsurface reservoir quality. This chapter provides a systematic, decision-tree-based procedure for using existing tools and techniques to evaluate potential subsurface reservoir quality when only surface outcrops are available . This approach is applicable to both carbonate and terrigenous clastic reservoirs. With this system, outcrop samples are classified into one of ten lithofacies types whose reservoir properties are codependent on common diagenetic or burial processes. The classification subdivides outcrop samples into either “tight” or “porous” lithofacies, depending on the measured porosity relative to economic minimum. “Tight” rocks include six end-member lithofacies that were either cemented or compacted during burial, or were originally tight at the time of deposition. “Porous” rocks include four lithofacies types that are categorized by original depositional fabric and the degree of alteration by recent surface weathering. Risk assessment for each of the ten lithofacies types using existing geological tools and techniques is discussed, along with guidelines for estimating potential subsurface porosity and permeability. Case histories that illustrate the recommended process for assessing risk are described from China, Myanmar (Burma), and Turkey.

Abstract We present a new porosity-depth relationship for clean, rigid grain (quartz, feldspar) sands under hydrostatic burial. This allows the prediction of porosity in uncemented sandstones to an accuracy of ±2.5 porosity units at 95% confidence levels. The relationship was derived using experimental data from laboratory compaction experiments and field data for buried uncemented sandstones from around the world. The equation is: where porosity (φ) is in percentages and depth ( z ) is in meters. By scaling this relationship in terms of effective stress rather than depth, it can be used to provide an equally accurate prediction of porosity for uncemented sands in overpressured settings. This is done using the following equation: where z ′ = effective burial depth (in meters); z = burial depth (in meters); ρ r = density of rock (in Kgm −3 [kilograms per cubic meter]) = typically 2650; ρ w = density of water (Kgm −3 ) = typically 1050; g = gravity (in ms −2 [meters per second squared]) = 9.8; φ Σ = average porosity of overburden = typically 0.2; and u = overpressure (in MPa [megapascals]). We propose that there is considerable value in a “compaction only” porosity-depth relationship. A compaction-only trend allows the accurate prediction of porosity in uncemented sandstones, and gives a maximum porosity baseline to which cement volumes, and resultant cemented sandstone porosities, can be compared. If both cemented and uncemented sandstone data are included to produce a “porosity loss-depth” relationship, the resultant scatter (typically ±5 porosity units for a given depth) in the relationship limits its usefulness. Prior to drilling, the new relationships may be used either to predict the porosity of sands that are known to be uncemented or to place an upper limit on the porosity estimated for sandstones either known or suspected to be cemented

Abstract Log-determined porosities of argillaceous limestone, limestone, dolomitic limestone, and dolomite of the Mississippian Madison Group in the Williston Basin were analyzed to determine the influence of carbonate mineralogy, shale content, and fabric on porosity loss with depth of burial. Carbonate mineralogy and shale content strongly influence the rate of porosity loss. Argillaceous carbonates lose porosity at the greatest rate with burial, followed by clean limestone, dolomitic limestone, and dolomite. Average porosity of grain-supported limestone is not systematically higher than aver-age porosity of mud-supported limestone in the same depth range, but there is a significant difference in the respective porosity range. Moderately to deeply buried (1.5–3 km) limestones with a grain-supported texture have a small percentage of high-porosity samples, whereas porosity distributions in matrix-supported limestones at equal burial depth cluster around the mean porosity and lack a tail of high-porosity samples. This effectively limits eco-nomic porosity in moderately to deeply buried Madison limestones to grain-supported rocks (packstones and grainstones). Results of this study reveal characteristics of basin-scale porosity loss mechanisms. Secondary porosity formed during burial is not evident in the porosity-depth profiles. Porosity loss is strongly influenced by mineralogy; clay content greatly accelerates the rate of porosity loss in limestones. In these rocks, dolomite porosity higher than limestone porosity at a given maximum burial depth is due primarily to selective preservation of dolomite porosity. Porosity decreases with increasing temperature in rocks with otherwise simi-lar burial (effective stress) history. The observed porosity-depth relationships roughly follow an exponential trend; this may indicate that there is some sort of feedback between porosity and the porosity reduction mechanism. Data generated in this study can be used to predict porosity distribution at a given depth in the Mississippian strata of the Williston Basin if no other information is available. Average limestone porosity at moderate to deep burial is significantly less than the porosity required for economic development of unfractured petroleum accumulations, so average porosity cannot be used as an estimate of economic porosity in a prospect. However, the distribution of porosity in a depth range can be used to estimate the risk associated with encountering sufficient thickness of economic porosity. The presence or absence of potentially economic porosity is best evaluated as a risk statement. For this reason, the porosity cumulative frequency distribution in a given depth range is a particularly useful tool because it can be interpreted in terms of expected thickness of porosity higher than a given threshold value. If information about vertical spatial correlation of porosity is available, the distribution can be interpreted in terms of risk of finding a minimum net thickness of carbonate exceeding a threshold porosity level. These methods can be used in other wildcat exploration settings where proper calibration data have been collected. The results of this study can be used as a guide to understanding porosity distribution with depth in other Paleozoic carbonates, and perhaps be directly applied to other late Paleozoic carbonates in cratonic settings.

Abstract A method for predicting the three-dimensional distribution of reservoir attributes has been developed by integrating geological and statistical models. The general method, applicable to carbonate and siliciclastic reservoirs, has been demonstrated by predicting the distribution of dolomite, calcitized dolomite, porosity, and permeability from regional to field scales in the Permian Zechstein 2 Carbonate of northern Germany. The first step in the prediction process consists of identifying factors potentially responsible for reservoir quality distribution. For the Zechstein 2 Carbonate, the resulting geologic model suggested that paleofaults and related fracture systems controlled the distribution of nonporous calcite (cal-citized dolomite) by acting as conduits for calcitizing fluids originating from anhydrites underlying the carbonates. The next step in the prediction process involves determining if the geologic model provides variables that can be used to predict the variable of interest given the predrill data available. If not, then other predictor variables, not necessarily cause-and-effect variables but ones whose values are known predrill, are required. Although a geologic model for Zechstein diagenesis elucidated the probable cause-and-effect relationship regarding the distribution of mineral types, it provided no means for predicting the geographic distribution of mineral types, because data on the distribution of paleofault and paleofracture systems cannot be obtained. For pragmatic purposes, models must both predict the desired parameter at the necessary scale and use predictor vari-ables whose values are known prior to drilling. For the Zechstein 2 Carbonate, linear regression models using facies and location (x-y coordinates and depth) accomplished practical predictions of mineral distribution. The fact that location provides significant predictions indicates that calcite and dolomite occur in a spatially organized manner, reflecting the geologic processes that caused the calcitization of the dolomite. Because paleostructure presumably controlled calcite distribution, separate models were developed for structurally distinct subareas. The use of structural subdivisions provided a way to account for different types of calcite distribution caused by different types of fault and fracture systems. Although mineralogy is a dominant control on reservoir quality in the Zechstein 2 Carbonate, the porosity and permeability distributions reflect additional factors. Like the mineralogy distribution, however, the porosity and permeability distributions have a dominant nonrandom spatial component, and therefore can be predicted reliably using location information. Because the spatial distribution of porosity and permeability in the Zechstein 2 Carbonate is highly complex, a nonparametric predictive technique (an artificial neural network) was implemented. It produced models that surpassed those of linear regression. Although cast here in terms of a particular application, the methodology is general, and such predictive models can be used to generate maps and cross sections of predicted parameters within any reservoir. In addition, sets of point values generated by the models can be loaded into visualization software to provide three-dimensional representations of the predicted parameters.

Abstract Sandstones that share common detrital mineralogies, depositional environments, and burial histories also share common diagenetic histories. A survey of the diagenetic history of 100 sandstones from around the world has recognized five common, repetitive, and predictable styles of diagenesis in which similar diagenetic mineral assemblages have been observed. The five diagenetic styles are: (1) quartz, commonly with lesser quantities of neoformed clays (e.g., kaolinite and/or illite) and late-diagenetic, ferroan carbonate; (2) day minerals (illite or kaolinite) with lesser quantities of quartz or zeolite and late-diagenetic carbonate; (3) early diagenetic (low-temperature) grain-coating clay mineral cements such as chlorite, which may inhibit quartz cementation during later burial; (4) early diagenetic carbonate or evaporite cement, often localized, which severely reduces porosity and net pay at very shallow burial depths; and (5) zeolites, which occur over a wide range in burial temperature, often in association with abundant clay (usually smectite or chlorite) and late-diagenetic, nonferroan carbonates. The quartz diagenetic style is the most common and accounts for 40% of the sample set. It is also most likely to occur in mineralogically mature sand-stones, while early diagenetic carbonates and zeolites dominate in miner-alogically immature sandstones. Presence or absence of clay appears to be independent of both initial sand mineralogy and depositional environment. However, when clay is present, the type appears to vary as a function of ini-tial sand mineralogy and depositional environment. Large quantities of quartz are unusual cements in sequences that have never been hotter than ~75°C, while illite precipitation at temperatures below ~100°C is rare. Zeolite composition changes systematically from clinoptilolite at ~25°C to laumonite at temperatures >100°C. The repetitive nature and simplicity of these five styles can help predict modifications in reservoir quality due to burial. An accurate prediction of the reservoir quality in sandstones forms the basis of an accurate porosity and permeability prediction ahead of drilling wells in petroleum exploration, development, or production.

Abstract The parameter time-depth index (TDI) is applied in this study to quantify empirically the influence of burial history on sandstone porosity evolution. The TDI, expressed in kilometers per million years of age, is defined as the area in the burial history diagram enclosed by the burial curve of the reser-voir and the axes of the diagram. In practice, reservoir depths during burial history are integrated at regular time intervals of 1 m.y. The calculations exclude present-day bathymetry or paleobathymetry. Sandstone reservoirs from several sedimentary basins along the Brazilian continental margin (Santos, Campos, Espirito Santo, Cumuruxatiba, Reconcavo, Sergipe, Alagoas, and Potiguar) were analyzed to investigate the evolution of porosity against TDI. These Upper Jurassic to Tertiary sand-stones lie in depths of 700 to 4900 m, and are hydrocarbon charged (oil or gas). Average porosities of most of these reservoirs were obtained from core analysis, and a few porosity data were taken from well log interpretations. Detrital constituents of the sandstones are mainly quartz, feldspar, and granitic/gneissic rock fragments. Sandstones were grouped into three main reservoir types, based on composition (detrital quartz content) and grain sorting: Type I (average quartz content <50%) are very coarse grained to con-glomeratic, poorly to very poorly sorted lithic arkoses. Rock fragments are mainly granitic/gneissic and coarse grained. Type II (average quartz content ranging from 50% to 70%) are fine- to coarse-grained (pebbles absent or occurring in small percentages), moderately sorted arkoses. Type III (average quartz content >80%) are fine to coarse, moderately to poorly sorted quartz arenites or subarkoses. Plots of average porosity against depth show great dispersion in porosity values; such dispersion is mostly due to differences in the reservoir burial histories. However, plotting porosity values against the TDI for individual reservoir types produces well-defined trends. The decrease in porosity is less marked in Type III reservoirs, intermediate in Type II, and faster in Type I. Such plots suggest that it is possible to make relatively accurate porosity pre-dictions based on reservoir TDI, texture, and composition, within the con-straints of reservoir depth/age and basin tectonics analyzed in this study.

Abstract Permeability is a key parameter in determining the economic value of a hydrocarbon accumulation; however, our ability to predict the magnitude and range of permeability in undrilled areas is poor. Traditional methods of permeability prediction are empirical and rely on developing relationships between permeability and other parameters that may be predicted with greater confidence, such as porosity or lithology. These empirical methods may work well in areas where there is sufficient calibration data, but extrapolation away from well data is prone to large errors (often by orders of magnitude). An alternative approach to permeability prediction is to model the effect of geological processes such as burial and cementation on the pore structure of the rock and, hence, calculate the change in permeability. Through under-standing the effect of various geological processes on permeability, it is then possible to predict permeability from geological models. This approach has applications in both data-rich and undrilled areas. The quantitative insight into which factors affect the permeability has been provided by computer modeling, which allows us to focus in on the most important controls, such as grain size and the amount of cement or ductile grains. Our ability to predict permeability in undrilled areas is now more often hampered by our inability to predict the variations in these controlling factors rather than by any lack of understanding of permeability itself.

Abstract Matrix permeability, the permeability associated with measurements on small samples, is controlled by depositional fabric and diagenesis. Prediction of matrix permeability requires: (1) specification of a fabric, (2) specification of the diagenetic state, and (3) a means to assess both factors in a sample set taken from a target basin. The data from the sample set can be used to extrapolate or interpolate within the basin or may be used to calibrate fabric response to basin history data (e.g., thermal history). The effects of fabric and diagenesis on the sample set can be determined using a combination of image analysis data and mercury porosimetry data. Strong correlations exist between permeability and grain size of unconsoli-dated sands and gravels, with permeability increasing exponentially with increasing grain size. Permeability in clastic fabrics is controlled by networks of packing flaws, characterized by large pores connected by large pore throats. Such circuits comprise only a fraction of the porosity and represent the effective flow component of porosity. Diagenesis usually brings about permeability reduction, but preferentially affects the grains in close-packed arrangements that separate the networks of packing flaws. A methodology has been developed over the past decade that quantifies thin-section-based data precisely enough to estimate the effects of grain size and diagenesis on the rock fabric with respect to flow properties. Such rock physics data are necessary for permeability prediction as a function of basin position.

Abstract Rock properties such as lithology and porosity can be obtained from com-parative P- and S-wave traveltimes or velocities measured from multicom-ponent (3-D, 3-C) seismic reflection data. A 3-D, 3-C seismic reflection data survey was acquired by the Colorado School of Mines Reservoir Characterization Project at Joffre field, Alberta, to map the complex porosity distribution in a shelf carbonate reservoir. Velocity ratio analysis, of compressional velocity to shear velocity (Vp/Vs), indicates a linear correlation with porosity in the Devonian Nisku reservoir. Vertical porosity distribution at wells and horizontal porosity distribution derived from seismic reflection data are used to map 3-D porosity distribution using geostatistical methods.The results show enhanced mapping of porosity distribution and better defi-nition of the lateral limits of the reservoir. These results will assist in reser-voir simulation of this field.

Abstract The Middle Jurassic Vajont Limestone of the Venetian Alps, Italy, is pre-dominantly composed of resedimented ooids that were deposited in slope and basin settings. The Vajont Limestone has been partly replaced by massive dolomite that can be mapped at both regional and local scales. Dolomite bodies that are present within or are associated with the Vajont Limestone include: (1) a large-scale wedge, ~ 25 km long, 10-15 km wide, and > 400-500 m thick (50-94 km3), located on the hanging wall of the Alpine-aged, thrust- based Mt. Grappa-Visentin anticline. This dolomite body is located within the axis of the anticline and crosscuts the Stratigraphie section where subver-tical to vertical faults penetrate the crest of the anticline; (2) Isolated, rootless plume-shaped bodies, 100-200 m wide and > 300 m high (> 2 × 10~ 2 km3), which penetrate a footwall syncline within an Alpine-aged thrust sheet. These dolomite “plumes” possess extensively brecciated cores and exhibit sharp to gradational transitions with surrounding Lower to Middle Jurassic basinal limestone; (3) Isolated dolomite “towers” that have partly replaced Cretaceous-age synsedimentary fault breccia. These bodies are found in overlying basinal strata (i.e., the Fonzaso Formation, the Ammonitico Rosso, and the Biancone Formation), but emanate from the underlying dolomitized Vajont; and (4) Small-scale wedge-shaped dolomite bodies on the scale of meters found along small faults and fractures. The connection between these dolomite bodies and Alpine-aged faults and fractures clearly indicates that dolomitization was a late burial process. It is proposed that during the Alpine deformation event, convection-driven fluids derived from Late Tertiary seawater were circulated through subaque-ous Alpine-aged faults and fractures and paleosynsedimentary breccias, thus creating the multitude of dolomite bodies now found in the Vajont and other Mesozoic basinal sediments. Paleogeographic, tectonic, and hydrologic sys-tems, similar to the one proposed for dolomitization of the Vajont, appear to be active in modern subaqueous thrust zones of the Caribbean and Northwest Pacific Coast. Potential reservoir attributes of Vajont dolomite bodies include their large size and medium to coarsely crystalline replacement fabric that is character-ized by significant amounts of partial moldic, intercrystalline, and vug pore space. Visual estimates of porosity within dolomitized grainstone and pack-stone range up to 10% to 15%, with inferred permeabilities of 1-100 md. Permeability of Vajont dolomite replacement fabrics is enhanced through recrystallization and the formation of touching-vug networks (inferred per-meabilities >100 md). >Results of this study indicate that (1) massive replacement dolomitization in thermotectonic (i.e., burial) settings may be much more important than previously thought, and (2) significant reservoirs may be hosted in otherwise tight basinal limestones as the result of late-stage burial dolomitization. Consequently, the geometries of the Vajont dolomite bodies may provide analogs for reservoir characterization and new exploration plays in the subsurface. Exploration methods for analogous dolomite reservoirs in the subsurface may include the mapping of dolomitization fronts using core and log data and seismic reflection identification of crosscutting dolomite bodies. The focus of such efforts should be placed on anticlinal and synclinal struc-tures within buried fold and thrust belts, and along zones of deep-seated tec-tonic fractures and faults within intracratonic basins.

Abstract Prior to BP Exploration’s drilling the well Antufash-1 in the Yemeni waters of the Southern Red Sea, reservoir quality was estimated to be poor; it was dry, plugged, and abandoned. The Miocene sandstones encountered were tight, with a mean porosity of 4% in the cored section and a permeability of only 0.07 md. The prediction of low quality for the reservoir section of Antufash-1 was based on very few core analysis data. The diagenetic history of potential reservoir sands in the Antufash acreage was calculated from data on depth to prospect, burial and thermal history of the area, reservoir sand provenance, and depositional environment. An initial assessment, using limited local well data, led to the conclusion that only at depths <0.5 km was it reasonable to expect high reservoir quality (>100 md). However, at depths <1.5 km, permeability was likely to be as low as 10 md. Throughout this depth range, the chances of halite cementation were also reasoned to be high. The rapid deterioration of reservoir quality with depth was attributed to the instability of the original volcaniclastic detritus. Such detritus was predicted to have converted to a mixture of zeolites and smectitic clay soon after deposition. The reactivity of the assemblage was also predicted to have been exaggerated by the high thermal gradients in the area. The recommendation was to avoid large parts of the license area known to have received input of volcaniclastic sediment, and to develop prospects in the few areas thought to have had arkosic sand input. These sands, it was reasoned, would suffer less degradation of reservoir quality. The Antufash-1 well successfully proved the existence of such arkosic sands in the basin, and their diagenetic history was as predicted. Unfortunately, the sandstones were tight. Halite cement filled, as predicted, all remaining porosity.

Abstract Successful prior-to-drilling prediction of anomalously good reservoir quali-ty in prospects at deep burial requires an understanding of diagenetic processes and quantitative models on how porosity is related to sandstone composition and to burial history. Quartz cementation and compaction are, in many cases, the most important porosity-reducing processes in quartz- and feldspar-rich arenites, capable of destroying all useful porosity during burial toward 4000 m. Hence, the recognition of factors that may hinder porosity loss by these processes, and thereby preserve good reservoir quality to depths beneath those usually considered as economic basement, is crucial during prospect evaluation of deep structures. In two deep (>4000 m) oil discoveries in Upper Jurassic sandstones in the Norwegian Central Graben, high porosity (>20%) appears to be preserved due to the presence of a ubiquitous microquartz coating on framework grains, and not due to any burial history-dependent factor such as high pore pressure, low thermal maturity, or early oil emplacement. In these sandstones, the microquartz coating has hindered quartz precipitation and late diagenetic chemical compaction. In interbedded sandstones without microquartz coating, the porosity is low (<10%) due to extensive quartz cementation. The microquartz coating appears within specific isochronous layers, and its presence is probably caused by input of amorphous silica (volcanic glass and sponge spicules) during deposition. The recognition of the inhibiting effect of this coating on quartz cementation, combined with quantitative models on the relationship between sandstone composition and diagenetic processes such as compaction and quartz cementation, allows confident porosity predictions. Hence, future porosity prediction in deeply buried Upper Jurassic sandstone in this area should focus on establishing sedimentological models addressing prediction of sandstone facies within intervals deposited during periods with high amorphous silica production and deposition.

Abstract Much of the remaining prospectivity in the Ula trend (Norwegian Central Graben) is deep (<3.5 km). A major risk to successful petroleum exploration in the trend is reservoir effectiveness. A few oil discoveries are not yet com-mercial because they occur in low-permeability sandstone. No simple porosity-depth relationship exists for the whole of the Ula trend. As such, mapping of economic basement is difficult. There are, however, simple porosity vs. depth relationships within the two main producing fields: Ula and Gyda. The porosity-depth relationships in the fields are due to downflank cementation by quartz. Quartz cementation was synchronous with oil emplacement, and evidence from petroleum-filled fluid inclusions has led to the conclusion that cementing fluids and petroleum competed in a “race for space.” The Ula trend displays evidence of all three outcomes of such a race: petroleum emplacement ahead of cementation, synchronous processes, and cementation ahead of petroleum emplacement. Porosity prediction for undrilled prospects and prospect segments was made by risking the three possible outcomes of such a race for space. The reservoir in prospect 7/12-JU4 was predicted to be oil bearing and have a mean porosity of about 16.4%: a function of synchronous petroleum emplacement and cementation. The well, however, was dry. It had a mean porosity of 14%; this compares well with the predicted porosity (13.9%) at the well location for a system in which cementation was completed before oil emplacement (equivalent to a porosity estimate for a dry hole).

Abstract The Mississippian Greenbrier Limestone is a major gas reservoir in the Appalachian basin, but its complex porosity patterns often deter active exploration. In southern West Virginia, the reservoir consists of oolitic tidal bars that are composites of smaller shoals. Porosity trends closely follow the ooidgrainstone facies that occupied shoal crests where coarse-grained, well-sorted ooid sand was generated with either unidirectional or bidirectional crossbeds. Nonporous packstone occurred in adjacent tidal channels, and a transitional grainstone/packstone facies of marginal porosity was situated along the flanks of the shoals. The key to drilling successful wells is in understanding the complex internal geometry of Greenbrier ooid shoals. A well penetrating the oolite with good porosity and bimodal cross-beds should be offset perpendicular to the dip directions; that is, parallel to the shoal axis. However, a well penetrating thin, porous limestone with one dominant crossbed azimuth should be offset opposite to that dip direction; that is, up the flank of the ooid shoal. Shaly interbeds characterize the edges of the shoals and mark the limit of productive wells. Schlum berger’s Formation MicroScanner log, which provides data on both lithology and cross-bedding, has proven to be a useful tool in predicting the distribution of oolite porosity.

Abstract The Jordan (San Andres) reservoir comprises ~400 ft (120 m) of upward-shoaling subtidal to peritidal carbonate strata, which is now thoroughly dolomitized and partly cemented by sulfates. Subtidal facies include dominant pellet packstone/ grainstone, with local bryozoans, algae, and coral bioherms and associated skeletal grainstone flanking beds. The lower part of the subtidal section is characterized by stratigraphically distinct zones in which permeability has been enhanced by a postburial carbonate-leaching event. These diagenetically altered (leached) zones crosscut subtidal depositional facies. Peritidal facies are nonporous mudstone and generally non- porous pisolite packstone characterized by abundant sulfate cement. The pisolitic rocks are locally porous and permeable where sulfate cement is either leached or absent from fenestrae. Cumulative production is 68 million stock tank barrels (MMSTB) of 218 MMSTB original oil in place, which is a recovery efficiency of 31%. A total of 47 MMSTB of remaining mobile oil occurs as bypassed oil in the contacted upper part of the reservoir, which has been penetrated by well bores; 12 MMSTB of mobile oil is in the uncontacted lower part, which has not been penetrated by well bores. The most prospective areas for increased production by waterflood profile modification in the contacted part of the reservoir are the southwest corner of the field, where low-permeability, diagenetically unaltered subtidal rocks are incompletely swept, and the eastern central part of the field, where heterogeneous permeability in peritidal rocks has resulted in an incomplete sweep. The most prospective areas for increased production through well-bore deepening into the uncontacted part of the reservoir are the southeast corner of the field, where high-permeability, diagenetically altered subtidal rocks are uncontacted, and the central part of the field, where high-permeability, diagenetically altered subtidal rocks are uncontacted. An understanding of diagenetically controlled reservoir properties can be used to predict the locus of remaining resource and to design recovery strategies.

Abstract Litharenites and sublitharenites of the Devonian Lock Haven Formation contain abundant rock fragments of shale and phyllite. These labile grains suffered varying degrees of destruction in several depositional environments; hence, sedimentary processes largely controlled the sandstones’ mineral composition. Fluvial sandstones have a high lithic content, distributary mouth-bar and offshore-shelf sandstones have an intermediate content, and barrier-island sandstones have a low content. Primary porosity relates inversely to compaction of the lithic grains, decreasing from a maximum minus-cement porosity of φ mc = 33% down to zero as lithics increase. The majority of primary porosity, however, has been occluded by cementation. Secondary porosity, created chiefly by dissolution of the chemically unstable rock fragments, is greatest (φ rf = 13%) for sand-stones of a moderate lithic content. Because of these relationships among depositional processes, lithology, and porosity, we predict that sandstones of different sedimentary environments should exhibit distinct porosity volumes and vary in their reservoir potential. Mouth-bar sandstones will have good total porosity, good secondary porosity, and offer the best reservoir quality. Shelf sandstones will have fair total porosity, most of which is secondary, whereas beach sand-stones will have low total porosity, most of which is primary. Fluvial sandstones will be the poorest reservoirs.

Abstract Completely dolomitized Upper Devonian Leduc buildups at depths >4000 m have higher porosities and permeabilities than adjacent limestone buildups; dolostones are more resistant to pressure solution and tend to retain their porosity during burial. Distribution of pore types is controlled by depositional facies, whereas distribution of permeability is largely controlled by diagenetic processes, especially dolomitization. In pool D3A of the Strachan reservoir, porosities and permeabilities are highest in the interior of the buildup where the strata are completely dolomitized. In the reef margin, porous and permeable dolomitized zones are interbedded with nonporous and nonpermeable limestone units. The presence of porous and permeable zones is closely related to the degree of dolomitization, with the greatest porosity and permeability occurring in completely dolomitized rocks. The reservoir character in the Ricinus West buildup closely follows depo-sitional units, despite complete dolomitization. At the reservoir scale, porosity and permeability have relatively similar values throughout the buildup. At the meter to tens of meters scale, the upper buildup interior is characterized by 1- to 2-m-thick, permeable and laterally continuous lagoonal strata. The lower reef interior consists of laterally discontinuous permeable zones. In the reef margin, permeability is controlled by fractures and interconnected vugs. At the millimeter scale, porosity and permeability are controlled by diagenetic processes. Late cementation and dissolution processes have slightly decreased and increased porosity and permeability, mainly in the lower part of the reser-voirs. Bitumen plugging decreased porosity and permeability in the upper part of the reservoirs. Although it is difficult to predict reservoir porosity and permeability trends, the secondary porosities in these deeply buried dolomites are mainly controlled by the primary porosity distribution and the depositional facies. The permeability is mainly controlled by diagenetic processes, especially dolomitization and various phases of cementation and bitumen plugging in the upper part of the reservoirs. Available data from the deep basin and the adjacent Rocky Mountains suggest that these porous dolomites are regionally extensive, and dolomite buildups elsewhere should have porosity and permeability variations similar to the Strachan and Ricinus West reservoirs. However, late-stage dolomite, anhydrite, and bitumen can locally partially to completely fill the pore spaces.