Rock-based 3-D reservoir characterization of a Silurian (Niagaran) reef—Ray gas storage field, Macomb County, Michigan
Published:May 10, 2018
- PDF LinkChapter PDF
Jessica L. Wold, G. Michael Grammer, 2018. "Rock-based 3-D reservoir characterization of a Silurian (Niagaran) reef—Ray gas storage field, Macomb County, Michigan", Paleozoic Stratigraphy and Resources of the Michigan Basin, G. Michael Grammer, William B. Harrison, III, David A. Barnes
Download citation file:
Silurian-age (Niagaran) reefs in the Michigan Basin have long been interpreted as relatively homogeneous units, despite production histories that strongly suggest the reefs are heterogeneous in both lateral and vertical dimensions. In an attempt to better illustrate reservoir heterogeneity in these reefs, a three-dimensional (3-D) sequence stratigraphic model was produced for the Ray Reef field. The resulting 3-D Petrel model incorporates 28 wells in the field using a combination of gamma-ray and neutron logs, porosity and permeability data from whole-core analysis, and facies descriptions from eight cores evenly distributed within the reef complex. Comparison of porosity and permeability values within the diverse depositional facies clearly shows trends related to the individual facies and positioning within the sequence hierarchy. Incorporation of the sequence stratigraphic framework into the 3-D model illustrates the episodic nature of reef growth as exhibited by the stacked nature of reef and capping grainstones, often separated by well-developed exposure horizons. The model also suggests a distinct difference between windward and leeward margins in both the geometry of the reef complex and the distribution of reservoir-prone facies. Windward margins are steeper due to higher rates of aggradational growth, and they typically contain higher percentages of reservoir-quality rock in both the reef core and forereef facies. Utilization of the sequence stratigraphic approach illustrates that the vertical reservoir heterogeneity often predicted from production in these reefs may be controlled in large part by the combination of vertical stacking patterns of facies within third- and fourth-order sequences.
Over the past 50 years, there has been significant advancement in understanding of the growth history of the Silurian (Niagaran) pinnacle reefs and reef complexes of the Michigan Basin region (Balogh, 1981; Briggs and Briggs, 1974; Gill, 1973, 1977a, 1977b, 1985; Huh 1974; Huh et al., 1977; Mantek, 1973; Mesolella et al., 1974; Sears and Lucia, 1979; Shaver, 1977, 1991, 1996; Wylie and Wood, 2005). Research findings have varied significantly, involving reef growth patterns, timing of deposition, depositional environments, and internal reef architecture, with a significant portion of the work being directed to better understand these reefs as hydrocarbon reservoirs.
The goal of this research was to evaluate the vertical and lateral lithofacies distribution within the Ray Reef complex using subsurface core data tied to well logs to constrain facies variability within a three-dimensional (3-D) model. A sequence stratigraphic framework was then developed to help relate the original depositional environments to present-day reservoir architecture and distribution. The sequence stratigraphic approach was used to establish time-correlative surfaces (sequence boundaries) based upon exposure intervals, hardgrounds, flooding surfaces, and vertical facies stacking patterns observed within the core data. The sequence stratigraphic framework was used to develop one of the first modern sequence stratigraphically constrained 3-D rock-based models of a Niagaran pinnacle reef in the Michigan Basin geologically constrained by core observations. This model clearly illustrates the complex episodic growth history of the reef complex and the resulting facies and reservoir heterogeneity both laterally and vertically within the reef complex. An enhanced understanding of the internal geometry and variability within the reef complex could lead to a more efficient use of these reefs for hydrocarbon extraction, natural gas storage, and CO2 sequestration.
Silurian reefs have been studied worldwide for over 150 years. The Niagaran Group (Wenlockian stage) reefs of the Michigan Basin have been a research focal point due to their abundance and economic importance. During the Niagaran, the Michigan Basin was a shallow intracratonic sea measuring 155 mi (250 km) wide (Howell and van der Pluijm, 1990) with a water depth of up to 650 ft (200 m) at the basin center (Brett et al., 1993). The regional extent encompasses the Lower Peninsula of Michigan and parts of Ohio, Illinois, Indiana, Wisconsin, and Ontario, Canada (Howell and van der Pluijm, 1990). Depositional facies and skeletal/biotic contents indicate deposition within 20°S –30°S of the equator in a tropical environment (Briggs and Briggs, 1974). The pinnacle reef structures and reef complexes measure up to 2 mi2 (5 km2) spatially, have a structural thickness of up to 600 ft (182 m), and are stratigraphically located at depths of 3000–7000 ft (900–2100 m) below the surface (Sharma, 1966; Caughlin et al., 1976). As discussed in Jodry (1969), the Niagaran pinnacle reefs most likely grew only a few tens of feet above the surrounding sediment during deposition and grew in maximum water depths of 30–50 ft (9–15 m). Ritter 2008 estimated water depth during deposition of the reef framework lithofacies between 3 and 33 ft (1–10 m) based on the presence of high-energy grainstones and exposure intervals, though reef complex–building stromatoporoids can thrive in water depths up to 230 ft (70 m). The reef intervals are overlain by the carbonate and evaporite deposits of the Salina Group, which make a naturally occurring, regionally extensive seal for these reservoirs (Sharma, 1966).
Oil and Gas Significance and Interest
Over 1100 Silurian pinnacle reefs have been discovered in the Michigan Basin, with ~900 being productive for oil and gas (Gill, 1973). The increase in demand for hydrocarbons over the past 50 years has initiated extensive exploration of the Silurian reefs, resulting in a plethora of data available to assist in understanding these structures. Exploration for the pinnacle reefs has continued since peak exploration during the 1970s, with occasional discoveries. Since the publication of Gill’s instrumental doctoral thesis on the Belle River Mills Gas Storage Field in 1973 (Gill, 1973), the study of Silurian reefs in the Michigan Basin has continued using modern techniques, including evaluation of basinwide controls on reef evolution (Ritter, 2008), 3-D rock-based modeling (Wold, 2008), 3-D modeling of reef reservoir properties (Qualman, 2009), and reef faunal distribution (Trout, 2012).
The Ray gas storage field (Ray Reef) is roughly 2 mi2 (5 km2) in area and is located 25 mi (40 km) northeast of Detroit, Michigan (Fig. 1). In 1961, the discovery well for the Ray Field was drilled and completed based upon a gravity anomaly (Schaefer, 1994) and produced 36 Bcf (billion cubic feet) of gas, out of an estimated 50 Bcf of in-place gas, from 1961 to 1966 (Schaefer, 1994). In 1966, the Ray Field was converted into a gas storage field; the state of Michigan currently has over 40 similarly depleted reservoirs converted to active gas storage fields (U.S. Department of Energy, 2017). According to the Michigan Department of Environmental Quality (MDEQ) online resource database (www.deq.state.mi.us/dataminer/), there are 49 active gas storage wells and 17 active gas storage observation wells within the Ray Reef complex. The Ray gas storage field was one of the largest gas storage fields in Michigan during the 1990s (Schaefer, 1994).
MICHIGAN BASIN MODELS
Several models have been developed to describe the growth history of the Silurian (Niagaran) reefs in the Michigan Basin, and they are generally based upon core observations from multiple wells (Gill, 1973, 1977a, 1977b, 1985; Mesolella et al., 1974) or a single well (Balogh, 1981; Cercone and Lohmann, 1986) from within a reef complex. The master’s thesis research by Balogh (1981) focused on the biodiversity of the Ray Reef complex based upon observations from a single core through the maximum structural thickness of the reef interval (Busch-Tubbs Kuhlman 1-36 well). The growth history was interpreted as punctuated growth separated by rapid sea-level falls resulting in intervals of subaerial exposure. Balogh’s research documented the lithological and structural evolution of the reef associated with the biological changes and identified trends associated with several growth forms of corals, bryozoans, and stromatoporoids. The historical Niagaran reef models portray a relatively simplified model for reef growth based on high-resolution vertical core data with less control on horizontal interwell facies variability, which is important to enhance predictability of depositional facies.
As discussed in Van Buchem et al. (2000), there are two approaches that can enhance the predictability between well data points. The first approach involves the establishment of a sequence stratigraphic framework based upon vertical stacking patterns and exposure surfaces observed within core data, which allows for the construction of a stratigraphic architecture based upon time-correlative boundaries. This approach geologically constrains intervals within the model to better represent the growth history of the reef complex and enhances the predictability between well data points (Williams and Aqrawi, 2006). The second approach utilizes a geostatistical approach to reservoir modeling that analyzes well data (well logs, facies descriptions, lithologies, and the aforementioned sequence boundaries) to produce a geologically constrained 3-D rock-based reservoir model. As discussed in Van Buchem et al. (2000), it is important to incorporate the high-resolution sequence stratigraphic framework, which is based upon core observations, into the reservoir model so that the model is based upon geologic characteristics and not solely upon unconstrained computer algorithms.
DATA AND METHODS
Sixteen whole cores were available within the Ray Reef complex, of which eight complete cores evenly distributed within the entire reef buildup were analyzed in this study, totaling over 1600 linear ft (488 m) and providing coverage of the core reef and off-reef regions, all of which have been dolomitized. The cores reside at the Michigan Geological Repository for Research and Education (MGRRE), a core repository affiliated with Western Michigan University in Kalamazoo, Michigan. Locations within the reef complex, length of core, completeness of core, and condition of the core were deciding factors utilized when selecting the cores to analyze. Condition of the core was determined by several criteria, including the amount of fabric-destructive dolomitization present and whether or not primary depositional characteristics could still be recognized. The amount of fabric-destructive dolomitization is important to include from the diagenetic aspect, but using cores that have primary depositional features preserved provides a more accurate detailed description of reef growth, which was the key focus of this study. Some wells were missing large sections of core material (20–50 ft; 6–15 m) in the middle of the reef interval and were not used because key aspects of the growth history were likely missing.
Core descriptions from this study included details of lithology, facies type, sedimentary structures, diversity and abundances of biota, dominant pore types, and interpretation of depositional environment. The six facies used within the core descriptions (Fig. 2), and ultimately the models, were determined to be consistent with previous research conducted on the Silurian (Niagaran) reefs of the Michigan Basin (Mantek, 1973; Briggs and Briggs, 1974; Mesolella et al., 1974; Huh et al., 1977; Sears and Lucia, 1979; Balogh, 1981; Gill, 1973, 1977b, 1985). The most detailed study of primary facies was the work done by Gill (1973, 1977a, 1977b, 1985) on the Belle River Mills reef complex. Many of the facies that Gill described in his research are similar to facies found within the Ray Reef complex. Additionally, the thesis by Balogh (1981) assisted with identifying features within the Busch-Tubbs-Kuhlman No. 1-36 well.
Conventional Whole-Core Analysis—Porosity and Permeability Data
A conventional core analysis (whole-core analysis) was available for seven of the eight cores used in this study, with porosity and permeability measured at 1 ft (0.3 m) intervals. Some sections of the core were thought to be too brittle or too small to withstand the testing procedure, so fractions of a foot that could be tested were used, including 0.25, 0.4, 0.5, 0.6, and 0.8 ft (0.08, 0.12, 0.15, 0.18, 0.24 m), and the rest of the core was discarded. The significance of this is that during the core description process, key boundaries and facies could be missing from the core and not incorporated into the sequence stratigraphic framework and reservoir model. Only one core (Busch-Tubbs-Kuhlman #1-36) contained the complete cored interval. The porosity and permeability data were converted to digital log curves and imported into Petra and Petrel for populating porosity and permeability models at an interval of 1 ft (0.3 m).
Mud Mound Growth
The incipient bioherm and bioherm core facies represent the establishment of a mud mound buildup thriving with diverse organisms (Fig. 3). Over time, these carbonate mud mounds grew locally to form topographic relief on the seafloor (Wilson, 1975; Sears and Lucia, 1979). Sears and Lucia (1979) hypothesized that the pinnacle reefs and reef complexes in the Michigan Basin initiated as carbonate mud mounds in a carbonate ramp environment. Common organisms associated with the mud mound growth phase include delicate and encrusting bryozoans, echinoderms (crinoids), and brachiopods with sparse stromatoporoids and tabulate corals. These organisms acted as baffles to locally accumulate carbonate mud and sediments. The energy regime is interpreted as very low and below storm wave base due to the abundance of more delicate, binding and encrusting bryozoans. Briggs and Briggs (1974) were the first to establish that crinoids and bryozoans constitute ~70% of the biohermal mud mound sediments. Briggs and Briggs (1974) also suggested that the mud mounds increased in height and lateral extent while rising toward higher-energy conditions, providing a foundation and eventually evolving into the reef core facies with the addition of framework organisms.
The reef core, reef debris, and skeletal grainstone facies identify the reef proper or dominant zone of framework reef growth (Fig. 4). Framework organisms, as well as encrusting and binding organisms, built toward wave base in an open-marine carbonate ramp environment. The change from low-energy to higher-energy conditions is interpreted from the biologic diversification of the reefs compared to the more limited biota of the bioherm (Sears and Lucia, 1979; Ritter, 2008; Trout 2012). Briggs and Briggs (1974) suggested that the Niagaran reefs of the Michigan Basin were stabilized by the growth of tabular corals, stromatoporoids, and branching and encrusting bryozoans. The increased abundance of stromatoporoids and tabulate corals as the dominant reef builders allowed the reef mass to better resist waves and rough-water conditions.
Compared to the bioherm stage of reef growth, the reef core stage records a higher abundance and diversity of organisms with a larger range in size. For example, crinoid stems found within the bioherm interval range are <5 mm in diameter, while crinoid stems measuring up to 20 mm diameter were found within the reef core interval in the Busch-Tubbs-Kuhlman 1-36 core.
The reef debris facies (Fig. 5) represents debris accumulated within the reef complex or debris aprons shed off of the reef. The reef debris facies contains the same allochems that were found within the reef core facies, but it is classified as a separate depositional facies due to the lack of in situ, growth-position framework organisms. The reef debris aprons generally consist of well-rounded grains that exhibit horizontal bedding characteristics.
Restricted Environment Facies Deposition
Increased basin salinity during decreased water depth and restricted water circulation at the end of the Niagaran resulted in the deposition of restricted environment facies overlying the reef facies (Sears and Lucia, 1979). The restricted facies include laminated or burrowed mudstones, peloidal mudstones, cyanobacteria mats, and digitate stromatolite growth structures (Fig. 6). These restricted facies coincide with the four supratidal island lithofacies established by Briggs and Briggs (1974). Organisms found within these zones include microbial colonies, bryozoans, and brachiopods. According to Sears and Lucia (1979), the restricted facies indicate zones of higher salinity and nutrient circulation, which contributed to the lack of framework organism growth. Additional features found within these facies are mud cracks, cyanobacterial laminations, bird’s-eye porosity, rip-up clasts, and flat pebble conglomerates. All of these features are indicative of a tidal flat, supratidal island, or shallow restricted environment of deposition (Lucia, 1972; Gill, 1985).
The exposure intervals indicate an area where relative sea level dropped, exposing parts of the reef complex, and resulting in a hiatus in the rock record (Figs. 7 and 8). Exposure intervals within the Ray Reef complex are typically found within the reef core interval and the restricted facies; none were observed within the bioherm facies. The exposure intervals commonly contain increased porosity and permeability likely due to freshwater influx, karsting, and dolomitization (Mesolella et al., 1974; Gill, 1985). Friedman and Kopaska-Merkel (1991) observed that freshwater leaching and dolomitization of exposure intervals created vadose fractures, vugs, and channels within the reef core and biohermal mud mound zones, which resulted in increased porosity and permeability in many Silurian reefs.
As discussed in Esteban and Klappa (1983), exposure intervals should be analyzed independently from the other observed lithofacies because they are formed purely through diagenetic processes, compared to the original depositional lithofacies, which are formed in various depositional environments. Depending on the amount of change (amplitude) in relative sea level, subaerial exposure surfaces can occur within shallow- through deep-water depositional environments, where depositional facies such as those described here are depth and energy dependent.
The actual subaerial exposure surface may not be easily observed unless it is present within the cored interval, but the alteration zone that exists below the exposure surface can often be identified within the cored sections. As previously stated, sections of the core that were not used during the conventional core analysis process were discarded and may have contained evidence of exposure surfaces.
Six depositional lithofacies were identified through the analysis of ~1600 linear ft (488 m) of core from eight wells spaced throughout the Ray Reef complex (Figs. 3–8). Similar lithofacies were identified in previous fundamental studies of Niagaran reefs in the Michigan basin (Mantek, 1973; Briggs and Briggs, 1974; Mesolella et al., 1974; Huh et al., 1977; Sears and Lucia, 1979; Balogh, 1981; Gill, 1973, 1977b, 1985; Ritter, 2008). The depositional facies are described next in stratigraphic order by stage of reef growth. Rock types are described using Dunham’s classification of carbonate rocks (Dunham, 1962) and Embry and Klovan’s classification table of skeletal limestones (Embry and Klovian, 1971).
Facies 1—Incipient Bioherm
The incipient bioherm depositional facies (Fig. 3) is a mud-dominated skeletal mudstone to wackestone deposited within the mud mound environment. The lithofacies typically includes abundant crinoids, bryozoans, and brachiopods, with sparse stromatoporoid and tabulate coral fragments. Wispy stylolites and stylolite swarms are common in the muddier zones of the interval, as well as stromatactis fabrics related to formation of shelter porosity (Hladil, 2005). The facies exhibits low porosity (1%–10%, 5.6% avg.) and permeability (1–17 mD, 7.5 mD avg.) as a result of low initial porosity and high mud content in the primary depositional fabric. The incipient bioherm facies is present within the Laskowski No. 4-1 well.
Facies 2—Bioherm Core
The bioherm core depositional facies (Fig. 3) is a mud-dominated skeletal mudstone to wackestone, with local zones of packstones, and it represents the transition from a mud mound to preframework reef environment of deposition. Common organisms include abundant crinoids, bryozoans, and brachiopods, with sparse stromatoporoid and tabulate coral fragments. Intervals contain variable percentages of skeletal fragments versus mud content, but they contain a higher diversity of biota and a larger maximum grain size than the lower incipient bioherm facies. Wispy stylolites and stylolite swarms are common in the muddier intervals, which may indicate small-scale lithologic transition areas. Grain-rich areas can exhibit thick continuous sutured-seam stylolites. The lithofacies exhibits low porosity (4%–8%, 6.2% avg.) and permeability (0–31 mD, 11.6 mD avg.) values, likely due to the low initial porosity and high mud content in the primary depositional fabric. The bioherm core facies is present in three of the eight cores studied: the Busch-Tubbs-Kuhlman No. 1-36, Halmich No. 2-1, and the Laskowski No. 4-1 well.
Facies 3A—Reef Core
The reef core facies (Fig. 4) is composed of skeletal wackestones to packstones, boundstones, and rudstones. Common organisms include framework stromatoporoids and tabulate corals (both with various growth forms), crinoids, bryozoans, and brachiopods. The biota assemblages observed within these facies are more diverse, more abundant, and larger in size than seen in other lithofacies (Sears and Lucia, 1979). Stromatoporoids measure up to 3 ft (0.9 m) vertically and cross the width of the core or consist of fragments measuring <0.04 in. (<1 mm) to over 3.54 in. (90 mm) in width and <0.04 in. (<1 mm) to 1.2 in. (30 mm) in height. These fragments can either follow a horizontal orientation or are randomly orientated. The stromatoporoids have the internal boxwork structure preserved, which allowed for easier identification (Fig. 4A).
Several different types of corals were observed within the cores, in which tabulate corals were the most common type found within the reef core facies. Where present, solitary finger corals ranged in size from 0.04 to 0.40 in. (1–10 mm) diameter and 0.59–1.57 in. (15–40 mm) long, with the internal structure of the corals typically preserved. Large well-preserved tabulate corals were found throughout the reef core interval, with individual corals measuring 0.98–3.35 in. (25–85) mm wide by 0.39–3.15 in. (10–80 mm) high (Fig. 4C). Most of the tabulate corals were preserved in growth position, with a few smaller matrix coral fragments distributed at random orientations.
Depending on the percentage of mud, suture and wispy stylolites are present throughout the reef core interval, which may indicate small-scale lithologic transition areas. Extensive leaching occurred in zones where wispy stylolites are present, often associated with discoloration of the matrix from a gray/brown color to a bright orange/red color. The 0.04-in.-thick (1-mm-thick) black sutured-seam stylolites found within the reef core interval were associated with zones of larger-sized grains, lower mud content, and often associated with pyrite.
The reef core facies exhibited variable porosity (5%–13%, 8.3% avg.) and permeability (0–58 mD, 19.1 mD avg.) values, likely due to variations in the initial porosity values related to the primary depositional fabric (Fig. 9). Porosity and permeability are dependent upon pore types present, the type and amount of biota, amount of cementation, and percentage of mud. When comparing the porosity log (from whole-core analysis) to the core itself, large spikes in porosity occurred in association with large tabulate corals with well-preserved primary growth, framework porosity. In zones with abundant brachiopod molds, porosity values were also high, but the permeability was dependent on the extent to which the molds were interconnected (Fig. 10). For example, porosity was found to be lower where brachiopod molds were suspended in a muddier matrix versus higher permeability where brachiopod molds were dense and interconnected. This comparison is true for the abundance of bryozoans within the reef core intervals as well.
Facies 3B—Reef Debris
The reef debris facies is composed of skeletal wackestones to grainstones and represents debris accumulated within the reef complex or debris aprons shed off of the reef. The dominant grains are peloids as well as bryozoan, coral, crinoid, and brachiopod debris, which are the same allochems found within the reef core facies. Skeletal fragments range from 0.04 to 0.79 in. (1–20 mm) wide, are generally well rounded, are preserved in preferred orientations, and exhibit horizontal bedding characteristics (Figs. 5A and 5B). The internal structure of organisms is typically preserved, seemingly dependent on the extent of diagenetic alteration. Locally, 0.04–0.08 in. (1–2 mm) anhydrite laths and 0.04-in.-thick (1-mm-thick), black continuous sutured-seam stylolites are found within this interval. Porosity values range from 6% to 14% (10.6% avg.), and permeability values range from 0 to 117 mD (44.3 mD avg.) when comparing whole-core data to facies descriptions. The reef debris facies is present in five of the eight studied cores: the Busch-Tubbs Kuhlman No. 1-36, Halmich No. 2-1, Jacob No. 1-36, the Lippert No. 3, and the Percy No. 2-2 wells.
Facies 4—Skeletal Peloidal Oolitic Grainstone
The skeletal grainstone facies was classified based upon very low (<10%) mud content. Most often, the grains were too altered to positively identify the original skeletal grain, but ghosts of peloids or ooid grains measuring <0.04 in (<1 mm) could be identified (Fig. 5C). The facies is dominated by fining-upward grain-size packages that are bounded by <0.04-in.-thick (<1-mm-thick) black sutured-seam stylolites (Fig. 5D). Porosity values range from 6% to 14% (9.2% avg.), and permeability values range from 1 to 28 mD (12 mD avg.) when comparing whole-core analysis data to facies descriptions. The grainstone facies is present in seven of the eight studied reef cores (all except Laskowski No. 5-1) and caps the reef core interval in six of the eight cores.
Facies 5—Restricted Facies
The restricted lithofacies group (Fig. 6) makes up a package of lithofacies deposited above the reef core interval. Four different lithologies were identified within this group of facies, and each are discussed in detail next. Porosity values range from 3% to 13% (8.4% avg.), and permeability values range from 0 to 14 mD (6.7 mD avg.) when comparing whole-core analysis data to facies descriptions (Fig. 9).
Laminated Peloidal Mudstone (5A)
The laminated peloidal mudstone lithofacies consists of massively bedded to finely laminated mudstones (Fig. 6B). The mudstone laminations measure 0.04–0.16 in. (1–4 mm) thick and are typically disrupted or cracked. White to blue pinpoint anhydrite nodules up to 0.04 in. (1 mm) diameter or larger anhydrite nodules measuring up to 0.32 in. (8 mm) are found throughout the interval. The dominant grain type is micritic peloids, which average 0.04 in. (1 mm) diameter and can constitute up to 35% of the rock material. Thin black, continuous, wispy stylolites are present, as well as similar-looking cyanobacteria laminations with a crinkled texture. The laminated peloidal mudstone is present in two of the eight cores studied: the Halmich No. 2-1 and the Percy No. 2-2 wells.
Peloidal-Bioturbated Mudstone (5B)
The peloidal-bioturbated mudstone lithofacies consists of a bioturbated peloidal mudstone (Fig. 6A) that lacks skeletal constituents but contains small well-preserved peloids averaging 0.04 in. (1 mm). Burrows are represented as large mottled intervals with smooth or irregular boundaries marked by a tan to brown coloration. Thin, discontinuous, wispy stylolites are present throughout the interval and are locally leached to a bright orange to tan color. White to blue, fabric-destructive nodular and pinpoint anhydrite is present in local areas of this facies. The peloidal-bioturbated mudstone facies is present in the Percy No. 2-2 well.
Stromatolite-Cyanobacterial Boundstone (5C)
The stromatolite-cyanobacterial boundstone lithofacies is dominated by 1–10-mm-thick laminar cyanobacterial mats or stromatolites (Figs. 6C and 6D). Stromatolites are identified as domal structures as defined by Davies (1970) and measure up to 15–35 mm wide by 20–40 mm high. The cyanobacteria laminations in the stromatolite structures are up to 1 mm thick and are overlain by laminar cyanobacterial mats. Dense packstone to grainstone intervals, 5–70 mm thick, are present between the cyanobacterial laminations and contain skeletal grains measuring up to 2 mm and peloids ranging from 1 to 2 mm. Anhydrite laths 1–10 mm long are commonly found along wispy stylolites and cyanobacterial laminations. Stylolites are either thick black and continuous, or thin, wispy, and discontinuous, and they are found throughout the interval. Fenestral porosity (bird’s-eye porosity) is found throughout the interval, which is commonly associated with cyanobacterial mats (Shinn, 1983). The cyanobacterial boundstone/stromatolite facies is present in six of the eight cores studied: the Busch-Tubbs No. 2-36, Busch-Tubbs-Kuhlman No. 1-36, Halmich No. 2-1, Jacob No. 1-36, Laskowski No. 4-1, and the Laskowski No. 5-1 wells.
Flat Pebble Conglomerate (5D)
The flat pebble conglomerate lithofacies consists of 10–40 mm flat clasts of the adjacent laminated cyanobacterial boundstone (Fig. 6C). The clasts are pseudorectangular in shape, with straight edges and slightly rounded corners. The surrounding matrix is composed of finer fragments of cyanobacterial lamination debris measuring 2–10 mm long. The flat pebble conglomerate facies is present in the Halmich No. 2-1 well.
The exposure intervals range from 6 in. (0.15 m) to 25 ft (7.62 m) thick in the observed cores (Figs. 7 and 8). The alteration of the rock due to extensive dolomitization makes it difficult to identify most skeletal constituents. Coral and stromatoporoid framework grains measuring up to 2.76 in. (70 mm) are recognizable and indicate original deposition within the reef core facies (see Facies 3a—reef core section). Wispy and sutured-seam stylolites are present in zones where brown to gray parent rock is leached to a bright orange-tan color. Irregular boundaries between the parent rock and the alteration interval are present in several of the cores. Porosity is very high in the alteration zones, which can be seen on the porosity and permeability cross-plots from whole-core analysis (Figs. 9, 10). The dominant porosity types are solution-enhanced vugs and vertical/horizontal fractures. Porosity values range from 8% to 23% (14.2% avg.), and permeability values range from 1 to 355 mD (92.8 mD avg.) when comparing whole-core analysis data to facies descriptions.
As discussed in Vail et al. (1977a), a key component of the application of sequence stratigraphy is the existence of unconformities, or stratal bedding surfaces. These surfaces are important because it is assumed that stratigraphic units are composed of genetically related depositional sequences in which upper and lower boundaries are unconformities or correlative conformities (Vail et al., 1977b). As stated in Posamentier and James (1993), sequence stratigraphy should be used as a tool, not simply as a template, because specific localized factors may control stratigraphy in a given area, including subsidence, sediment influx, eustasy, and local geologic features.
Carbonate rock units are often highly heterogeneous, making it difficult to effectively use lithostratigraphy to predict the horizontal distribution of the rock units. Sequence stratigraphy utilizes depositional bedding surfaces to constrain and represent internal reservoir architecture (Kerans and Tinker, 1997). The generalized facies stacking pattern (Fig. 2) and time-correlative exposure surfaces (Fig. 11) observed in cores formed the basis for the development of the sequence framework presented here that illustrates the complex growth history of the Silurian (Niagaran) reefs in the Michigan Basin. Previous Niagaran reef studies focused on vertical facies variability within the reef structures. The development of the chronostratigraphic sequence framework constrained the growth model, which strengthened facies predictions between data points (boreholes) and is essential in understanding the periodic growth history of the reef.
As discussed in Ross and Ross (1996), one approach for identifying Silurian sea-level fluctuations is to use sequence stratigraphy, where vertical changes in depositional facies and lithologic and sedimentary structures are used to infer relative sea-level fluctuations. This approach, combined with modern analogs that enhance the understanding of lateral depositional facies distribution, makes the resulting sequence stratigraphic framework the most geologically reasonable guide for modeling a reef structure. Three distinct depositional sequences identified from core observations within the Ray Reef complex support the observations of Shaver (1996), Barrick (1997), and Ritter (2008) of three periods of sea-level fluctuations during the Wenlockian stage of the Silurian (Fig. 12).
Subaerial Exposure and Flooding Surfaces
Several exposure intervals were identified within subsurface cores and interpreted to be chronostratigraphic time lines during reef evolution that indicate time-correlative surfaces within the reef complex. Individual exposure and flooding surfaces that were originally present may be missing due to core handling during whole-core analysis procedures or poor initial recovery. The subaerial exposure intervals and vertical facies stacking patterns indicate flooding events and were used to define the main sequence horizons when constructing the sequence stratigraphic framework (Fig. 11).
The flooding surfaces within the reef core facies indicate the start of a transgressive cycle, which overlies a sequence boundary. These surfaces were recorded in 1 in. (0.02 m) to 4 in. (0.10 m) intervals of laminated, mud-rich sediment marking hiatuses in the reef growth zone. The surfaces are interpreted to represent fourth- or fifth-order sequence boundaries, which are relative, and possibly eustatic, flooding events.
According to Kerans and Tinker (1997), high-frequency sequences (fourth order) are the most important to recognize for the application of reservoir-scale sequence stratigraphy. Two orders of depositional sequences (third- and fourth-order) were identified and utilized within this research (Fig. 11). The fifth-order small-scale cycles were observed within the core data, but discerning between natural ecological changes/shifts and relative sea-level fluctuations at this high frequency, especially with the problems of missing core, was not feasible.
Large-Scale Depositional Sequences (Third Order)
The third-order depositional sequences that delineate reef growth are composed of two fourth-order cycles, or high-frequency sequences, which illustrate reef facies variability. The third-order depositional sequences are typically thought to record relative se-level amplitude variations of 164–328 ft (50–100 m) that result from global eustatic sea-level changes controlled by variations in global ice volume as well as seafloor spreading rates (Van Buchem et al., 2000; Read et al., 1995). In this study, the third-order sequence framework was used to geologically constrain the reservoir model, and three third-order sequences were identified within the reef complex. This is consistent with basin-scale unconformities interpreted in core by Ritter (2008) within her basinwide analysis of Silurian sea-level fluctuations in the Michigan Basin. The third-order sequence boundaries were frequently associated with exposure intervals observed in core, but they also were based upon the idealized depositional facies stacking pattern for the Ray Reef complex.
Sequence 1. Sequence 1 represents the formation of the incipient bioherm in a carbonate ramp environment and the transition from bioherm deposition to reef core deposition (Figs. 11 and 13). The lower boundary of sequence 1 is defined as sequence boundary 1 (SB1), and the upper boundary is sequence boundary 2 (SB2). SB1 is the stratigraphic top of the Grey Niagaran (Lockport Dolomite) and is characterized as a flat basinward-dipping carbonate ramp. SB1 was chosen as the base of the depositional model, or datum, because it is a relatively flat surface. Gill (1985) stated that Niagaran reef pinnacle growth was initiated on the Lockport Formation surface, which slopes basinward at ~0.5°–1.5°. The use of a flat-bottom datum for the depositional model allowed for better visualization of aggradation in carbonate complexes (Grammer et al., 2004; Eberli et al., 2004). SB2 is based upon third-order idealized depositional facies stacking patterns observed within core and was differentiated from the next successive sequence by evidence of a flooding or exposure event.
Sequence 2. Sequence 2 represents the major reef growth sequence within the Ray Reef complex and is characterized by the transition to framework builders and reef dwellers as the reef evolved through time (Figs. 11 and 13). The lower boundary of sequence 2 is sequence boundary 2 (SB2), marked by the change from the regressive stage of sequence 1 to the transgressive stage of sequence 2, and the upper boundary is sequence boundary 3 (SB3), which is marked by the end of the regressive stage of sequence 2. SB2 is a transgressive boundary where reef growth continued after a regressive period, or exposure period, and continued to grow toward wave base. SB3 marked the end of the regressive portion of sequence 2 and was observed in several cores as an exposure interval measuring 1–20 ft thick (0.3–6 m; 20 ft exposure interval in Busch-Tubbs No. 2-36). Internally within sequence 2, there are two fourth-order high-frequency sequences that control reservoir facies distribution.
Sequence 3. Sequence 3 represents the transition from the reef growth stage into restricted depositional environments capping the reef complex (Figs. 11 and 13). The lower boundary of sequence 3 is sequence boundary 3 (SB3), which is marked by the transition from the regressive phase of sequence 2 to the transgressive phase of sequence 3. SB3 indicates a deepening event where reef growth restabilized after a regressive phase of sequence 2 and continued to grow toward wave base; this boundary is marked by an exposure interval within the reef cores. The upper boundary is sequence boundary 4 (SB4), which marks the end of reef growth and the transition into anhydrite deposition of the Salina A-2 Evaporite. The boundary between the two depositional environments (reef/restricted to evaporite) was not present in every core, most often due to the cored interval not being shallow enough to encompass that interval. If the core did not contain this boundary, SB4 was placed at the distinctive neutron log signature between the Niagaran reef and restricted deposition rocks (1000–2000 API [American Petroleum Institute] neutron units) and the overlying anhydrite of the Salina A-2 Evaporite (3000–4000 API neutron units).
Medium-Scale Depositional High-Frequency Sequence (Fourth Order)
The fourth-order Milankovitch-driven high-frequency cycles have been interpreted to record relative sea-level amplitude changes of 82–165 ft (25–50 m; Van Buchem et al., 2000; Read et al., 1995). The interpreted fourth-order sequences within the Ray Reef model directly control the reservoir distribution and facies variability, and they are based upon the idealized facies stacking pattern (Fig. 2) present within the reef cores. The sequence boundaries are based upon vertical facies changes indicative of flooding events and regional exposure intervals. Each third-order sequence is composed of two fourth-order high-frequency sequences, which range from 25 to 70 ft (7.6–31.3 m) in the studied cores.
Small-Scale Depositional Cycle (Fifth Order)
Small-scale cycles are the smallest set of related depositional facies that can be observed in core data or outcrop and typically have a direct control on reservoir facies distribution in carbonate systems (Kerans and Tinker, 1997). These cycles are also referred to as genetic units (Busch, 1971; Lerat et al., 2000) and are associated with fifth-order (precession, obliquity) cycles. In this study, the fifth-order cycles (one to several meters thick) were grouped into the fourth-order cycles (tens of meters thick) due to limited data availability and extensive dolomitization of the reservoir rocks, which made detailed facies identification difficult. Even if the fifth-order cycles were able to be differentiated, they would not be well represented within the reef model due to the small scale of the cycles (one to several meters thick). Surfaces within the Petrel model have a tested minimum thickness requirement of 15 ft (4.5 m) between surfaces, in which case the fifth-order cycles observed in core data are too small to capture within the model (see also Ritter and Grammer, this volume).
Reservoir characterization incorporates multiple types of well data to improve the understanding and predictability of interwell heterogeneity by producing an integrated 3-D reservoir model. As described by Tinker (1996), a 3-D reservoir model interprets numerical data and portrays them visually. The numerical data for each well point are further constrained by the sequence stratigraphic framework developed from core observations. The use of the sequence framework to geologically constrain the reservoir model in carbonate systems is thought to be the best stratigraphic framework to utilize within a 3-D modeling effort (Tinker, 1996).
The software package used for 3-D modeling in this study was Petrel by Schlumberger. The numerical data input into Petrel consisted of gamma-ray and neutron logs, porosity and permeability data from whole-core analysis, surfaces (sequence boundaries), and detailed facies information. Several types of property models are available within the Petrel modeling software package; a facies model, permeability model, and porosity model were all simulated for the Ray Reef complex (Figs. 14 and 15). The foundation for the models is the third-order sequence boundaries identified from core data or distinctive well-log signatures. Each of the three depositional sequences (S1, S2, S3) were modeled separately (Figs. 14 and 15) as individual zones bounded by the sequence boundary surfaces (SB1, SB2, SB3, and SB4; Fig. 13) to ensure that the time-correlative sequence boundaries were honored. Additionally, subdividing the reef into three smaller zones and modeling each zone separately enabled each zone to be modeled at a higher resolution containing a higher cell count, which in turn, allows for a higher-resolution prediction of facies variability in the model.
Surfaces for each sequence boundary in the Ray Reef model were created using the Sequential Gaussian Simulation (SGS) algorithm, which was determined to best represent the data set as tied to the core. Each surface, or horizon, was interpreted to represent a time-correlative boundary during the growth history of the Ray Reef complex. Zones between each surface, or horizon, were created utilizing three million cells per zone. The benefit of using a large number of cells to describe each zone is that each cell of data within that zone will encompass a smaller area. The smaller cell size for each zone enables the model to contain more detailed information for each zone. For example, if a geologic surface is created that measures 10 m long by 10 m wide and has a cell size of 2 m, the surface model would consist of 25 cells of data. If the same model were simulated with a smaller cell size of 1 m, the surface model would consist of 100 cells of data to represent the geologic surface. The model with the smaller 1 m cell size will contain higher-resolution data values, thus enabling the development of a more detailed geologic surface that is more capable of illustrating the heterogeneity within the rock fabrics.
The facies model was developed from core observations from eight wells located within the reef complex. The subsurface cores encompassed all three major zones of the reef complex; the bioherm, reef, and restricted depositional environments. The 1-in.-scale (2.5-cm-scale) core descriptions for each well in the project were averaged to represent just one facies at a 1 ft (0.3 m) interval, even though multiple facies may exist within that measured foot, and imported into Petrel as a digital facies log from which the facies model was simulated.
Up-scaling is a process by which high-resolution data (e.g., 1 in. [2.5 cm] scale in this study) are converted to a larger scale (1–5 ft [0.3–1.5 m] scale) to be used in broader analysis. As discussed in Pyrcz et al. (2006), the input parameters for a model have to be consistent with the support size of the modeling program. The sample data were at too fine a scale for the program to efficiently process the data, so the data was up-scaled to fit the program processing requirements before the model was simulated. This process is necessary to convert the fine-scale core descriptions taken at 1 in. (2.5 cm) intervals into 1 ft (0.3 m) interval data values for input into the Petrel reservoir models. During this process, much of the fine resolution data are replaced with generalized observations over a higher-resolution interval. The Upscale Logs process built into the Petrel program was used to convert the 1 ft (0.3 m) interval facies descriptions into 3 ft (0.9 m) intervals. The high-resolution data provided enough detail to give an accurate representation of the internal variability within the reef complex, but small enough resolution so that the computer could handle the data processing.
The SGS algorithm was used for all of the reservoir models because it best represented the geologic well data. The SGS algorithm was used in the study by Pranter et al. (2005) to create permeability models using variograms in dolomite outcrops and also in a study by Dull (2004) to model lithofacies in a carbonate ramp environment. Both examples support the algorithm as a good choice for modeling a dolomitized reef complex.
According to Srivastava (1994), the SGS algorithm is a multiple-step calculation for simulating a value for a random grid node that is undefined. During this process, the local conditional probability distribution (lcpd) is estimated based upon the mean value and standard deviation of surrounding known data points. When a classic bell-shaped curve (multi-Gaussian kriging) is used to estimate the lcpd, the algorithm is defined as SGS. After the estimation is complete, the simulated data point is included in the data set (Srivastava, 1994).
Another variable used to estimate the lcpd is a variogram. The variogram aides in constraining the spatial continuity of the data points (Srivastava, 1994) and quantifying reservoir facies continuity (Caers et al., 2003). As defined by Gringarten and Deutsch (2001, p. 508), a variogram is “a measure of variability; it increases as samples become more dissimilar.” The variogram is represented by a graph of data points comparing a specific value versus the distance between data points (Gringarten and Deutsch, 2001; Pranter et al., 2005). The variogram may exhibit random, or nonlinear, behavior. This geologic variation for the normal distribution of data is referred to as the nugget effect. The lower the nugget effect value, the more accurate is the data simulation. The variogram used in an algorithm should best represent the spatial variability of the facies it is defining (i.e., it should be geologically reasonable). According to Gringarten and Deutsch (2001), algorithms such as SGS use the variogram as well as local input data to create a 3-D model.
RESERVOIR MODELS AND DISCUSSION
Two different cross sections through the reef complex will be discussed from the Petrel model for the Ray Reef complex. Cross section A–A′ is strike oriented (east-west), and cross section B–B′ is dip oriented (north-south; Figs. 13–15). Each cross section contains a facies model, porosity model, and permeability model.
Understanding the internal variability within the reef structures is important because the variability of the depositional facies has a direct control on the reservoir quality. A model that lacks internal variability will not assist in understanding the reservoir facies distribution and ultimately make reservoir quality prediction difficult. Additional variables that influence reservoir heterogeneity are pore type, biota present, diagenetic alteration, and amount of cementation.
The whole-core porosity and permeability data available for the Busch-Tubbs No. 2-36 well were plotted against depositional facies observed within the core. The resulting cross-plot, (Fig. 9) shows that an apparent trend exists between porosity and permeability related to primary depositional facies. The mud-rich bioherm (facies 1 and 2) and restricted facies (facies 5) have lower associated porosity and permeability values due to the poor primary depositional reservoir properties. The reef intervals (facies 3A, 3B, and 4) have a mixed range of porosity and permeability values, which can depend on biota present, amount of cementation, and/or mud percentage in the original rock fabric. The exposure interval has enhanced porosity and permeability due to alteration of the originally deposited rocks. Exposure intervals often contain increased porosity and permeability due to freshwater influx, karsting, and dolomitization (Mesolella et al., 1974; Gill, 1985). Friedman and Kopaska-Merkel (1991) observed that freshwater leaching and dolomitization of exposure intervals led to increased porosity and permeability in many Silurian reefs due to the creation of vadose fractures, vugs, and channels within the reef core and biohermal mud mound zones. The higher-resolution porosity and permeability data acquired from whole-core analysis combined with depositional facies from core analysis better demonstrate the variability within depositional facies and represent the known heterogeneity within these reef complexes (Gill, 1973).
Cross Section A–A′
Cross section A–A′ cuts through the reef complex in a strike-oriented, west-east direction (Fig. 13). The facies, porosity, and permeability simulations for this cross section through the reef complex are displayed in Figure 14. The cross-section A–A′ facies model shows internal variability of the depositional facies throughout the growth history of the reef complex. Stacked stratigraphic facies within the model are associated with the fourth-order high-frequency sequences recognized at core observation scale. The porosity and permeability simulations for cross section A–A′ show a range of present-day porosity and permeability values, which are the result of original reservoir parameters, cementation, grain size/type, dolomitization, and porosity type. The models show abundant vertical heterogeneity of facies and reservoir potential within the reef complex, but more importantly, the horizontal variability and distribution of the facies within the reef complex are portrayed in a geologically constrained model.
Comparison of the depositional facies model with the porosity and permeability models shows several relationships between the primary depositional facies and their resultant petrophysical properties. For example, the bioherm and restricted facies have low porosity and permeability values due to low initial depositional fabric porosity and permeability. Exposure intervals exhibit medium to high porosity and permeability values due to alteration of the originally deposited rocks, which correlate to the porosity versus permeability cross-plot in Figures 9 and 10.
The reef core interval exhibits a broad range of porosity and permeability values associated with the original biologic composition of the facies, which can be seen in Figures 9 and 10 and in the cross section A–A′ (Fig. 14). The porosity versus permeability plot for the Laskowski No. 4 well shows no relationship between depositional facies and porosity/permeability values alone for the reef core facies as a whole (facies 3 and 4). When the reef interval is subdivided based upon major skeletal sediment contribution, a conclusive trend between porosity and permeability is evident based upon the original biologic constituents (Fig. 10). As previously discussed, this trend is well illustrated in the Jacob No. 1-36 core, where large spikes on the porosity log correlate to zones containing one or more fragments of tabulate coral (Figs. 4C and 9). In zones with abundant brachiopod molds, porosity values are also high, but the permeability is variable dependent on how well the molds are interconnected. If the brachiopod molds are suspended in a muddier matrix, the porosity is lower. If the brachiopod molds are very dense and interconnected, the permeability is higher.
Cross Section B–B′
Cross section B–B′ cuts through the reef complex in a dip-oriented, south-north direction (Fig. 13). The facies, porosity, and permeability simulations for this cross section through the reef complex are displayed in Figure 15. The facies model for B–B′ shows variability of depositional facies throughout the growth history of the reef complex, with stacked stratigraphic facies associated with the fourth-order high-frequency sequences recognizable at core observation scale. Vertical heterogeneity within the reef complex and the horizontal distribution of facies are geologically constrained by the third-order sequence framework.
The progradation and aggradation evident within the reef complex models indicate that the reef grew both vertically (aggradation) and laterally (progradation) during the growth history of the reef. This variable growth history affects the overall reef geometry, which can be observed when comparing the northern and southern margins of the reef complex (Fig. 16). The south end of the reef complex (left side of Fig. 15; right side of Fig. 16) is interpreted as the windward reef margin based upon the steep reef foreslope, associated thick packages of stacked depositional facies that have high porosity and permeability values, and an aggradational and progradational growth trend. The angular and/or elongate shapes of skeletal grains contribute to the steep marginal reef slopes (35°–45°), which are dominantly composed of skeletal grainstones (Palmer, 1979; Grammer et al., 1993a) with high porosity and permeability values (Fig. 9). The skeletal grains underwent early cementation, which aided in stabilizing the reef slope margin at or above the angles of repose (Grammer et al., 1993b). Grain-flow deposits off of the steep carbonate slope identified on the southern windward margin are generally thin, elongate, homogeneous lenses that are laterally discontinuous, similar to that described by Grammer et al. (1993a).
In contrast, the northern end of the reef complex (right side of Fig. 15; left side of Fig. 16) is interpreted as the leeward margin based upon the gently sloping reef margin, associated thin packages of stacked depositional facies with low porosity and permeability values, and a progradational growth trend rather than an aggradational trend. The leeward margin contains a higher abundance of mud derived from the reef complex as a result of sediment being swept off of the leeward margin, whereas the thick stacked packages of grainstones on the windward margin underwent early cementation making them more resistant to transport (Grammer et al., 1993a). The dominant wind direction proposed for the Ray Reef complex coincides with the Silurian trade winds based upon the location of the Michigan Basin according to global reconstructions published by the Scotese PALEOMAP Project.
As discussed in Grammer et al. (2004), a comparison of modern analogs to ancient environments aids in the understanding of the spatial distribution of the reservoir facies within the reef complex. Tertiary Malampaya and Camago reefs in the Philippines (Grötsch and Mercadier, 1999) provide an analog for the windward versus leeward observations of the Ray Reef complex models, which can enhance understanding of the influence wind direction has on reservoir facies distribution. The Malampaya-Camago is an isolated carbonate platform buildup located offshore of the Philippines between 5°N and 10°N latitude with a northeastern windward margin characterized by a steeper reef margin, higher amounts of aggradation, and a higher percentage of early marine cementation. The leeward margin is characterized by a gently sloping margin and more progradational reef growth (Grötsch and Mercadier, 1999). The overall reef geometry of the Malampaya-Camago carbonate platform is analogous to the reef geometry of the Ray Reef complex, which has a steep southeastern windward margin and a gently sloping northwestern leeward margin (cross section B-B′, Fig. 15). The stacking patterns of the Malampaya and Camago platform indicate aggradation and progradation of the reef complex during transgressive periods, and possible exposure intervals during several periods of regression that may have a control on reservoir quality and distribution (Grötsch and Mercadier, 1999).
Two separate reef mounds are identifiable within sequence 1 (lower section) of cross section B–B′ (Fig. 13) and the surface model for SB2 (Figs. 13 and 15). The two reef mounds ultimately coalesce into a single larger reef complex between SB2 and SB3 (Fig. 13). Several authors, including Gill (1973) and Wylie and Wood (2005), have proposed that the reef mounds coalesced over time, which can be observed in their models for the Belle River Mills reef. Further evidence for coalesced mound growth is the abundance of reef debris and skeletal grainstone facies between the two smaller reef mounds.
Lateral Continuity within Reservoir Modeling
Significant lateral and vertical heterogeneity is apparent in several zones of the reservoir models simulated for the Ray Reef complex (Figs. 14 and 15). The lateral continuity may be explained by two significant parameters input into the models. First, the lateral continuity may be an artifact of limited data points. For example, the facies model is based upon data from eight wells but the porosity and permeability models are based upon data from 19 wells. Additional data with high-resolution facies descriptions from other cores within the reef complex would resolve this problem.
Second, the facies model was based solely upon the depositional facies input from core observations. Primary depositional fabrics were well preserved in some cores, but extensive diagenetic alteration (dolomitization) in several of the cores made identification of these primary depositional fabrics problematic, which could have oversimplified the facies variability in the model. Cores that have primary depositional features preserved were used to construct the framework of the model, but porosity and permeability data were included from 19 wells within the complex. Once again, the inclusion of core descriptions from wells with readily identifiable depositional facies would strengthen the accuracy of the model.
(1) A distinct relationship exists between primary depositional facies and porosity and permeability values (Fig. 9) within the Silurian Ray Reef complex. The mud-rich bioherm (facies 1 and 2) and the restricted facies (facies 5) have lower porosity and permeability values due to the primary depositional fabrics. The reef intervals (facies 3A, 3B, and 4) have a mixed range of porosity and permeability values, which could result from biota present, amount of cementation, and percentage of mud in the original rock fabric. The exposure interval has enhanced porosity and permeability due to freshwater influx, karsting, and dolomitization (Mesolella et al., 1974; Gill, 1985).
(2) The reservoir model created for the Ray Reef complex is constrained by core observations and tied to wireline logs to develop a sequence stratigraphic framework that portrays extensive internal variability within the reef structure compared to previous Silurian reef growth models from the Michigan Basin. The sequence stratigraphic framework, based upon third-order sequence boundaries derived from core analysis, geologically constrains the 3-D reservoir model, resulting in a more geologically reasonable model that enhances the predictability between data points and overall reservoir facies distribution within the reef complex. The 3-D facies model simulated for the Ray Reef complex shows substantial internal variability within the reef structure related to the complex growth history of the reef as it responded to fluctuations in sea level. The internal variability of the depositional facies has a direct control on the reservoir quality; therefore, a model that does not show the internal variability within the reef structure will not assist in understanding the reservoir facies distribution and predictability.
(3) In total, three distinct third-order depositional sequences occur within the Ray Reef complex, which likely coincided with the three periods of global (eustatic) sea-level fluctuations observed in other Silurian studies (Fig. 11). Each third-order depositional sequence is composed of two interpreted fourth-order high-frequency sequences (Fig. 11), which range from 25 to 70 ft (7.6–31.3 m) in the studied cores. The fourth-order sequences within the Ray Reef model directly control the reservoir distribution and facies variability, are based upon the idealized facies stacking pattern (Fig. 2) developed from facies present within the reef cores, and are likely influenced by the windward and leeward margins.
(4) Progradation and aggradation influenced by the prevailing wind direction affected the overall reef geometry and reservoir distribution within the Ray Reef complex, which can be observed when comparing the northern and southern margins of the reef complex (Fig. 15). The comparison of the facies distribution with the porosity and permeability models indicates that potential reservoir intervals are along the windward margin of the Ray Reef complex. The south end of the reef complex is interpreted as the windward reef margin based upon the steep reef foreslope, associated thick packages of stacked depositional facies that have high porosity and permeability values, and an aggradational growth trend rather than a progradational trend. The contrasting northern end of the reef is interpreted as the leeward side of the reef complex based upon the gently sloping reef margin, associated thin packages of stacked depositional facies with low porosity and permeability values, and a progradational growth trend rather than an aggradational trend.
(5) Two separate reef mounds are identifiable within sequence 1 (lower section) of cross section B–B′ (Fig. 15) and on the structural contour map for SB2 (Fig. 13). The two reef mounds ultimately coalesce into a single larger reef complex between SB2 and SB3 (Fig. 13). Further evidence for coalesced mound growth is the abundance of reef debris and skeletal grainstone facies between the two smaller reef mounds, resulting from debris shedding off the reef mounds (Fig. 15).
We wish to thank the Michigan Geological Repository for Research and Education (MGRRE) for providing access to cores and wireline logs. Funding for this project was provided in part from a U.S. Department of Energy grant (DOE: DE-FC26-04NT15513). Matthew Pranter provided valuable reservoir modeling assistance while we were developing the study. Valuable comments from reviewers Charlotte Sullivan and Timothy Brock helped make the manuscript stronger and are acknowledged by the authors. Bill Harrison was a valuable member of Jessica Wold’s M.S. thesis committee, from which this manuscript is based. This is Boone Pickens School of Geology contribution 2017-61.
Figures & Tables
Paleozoic Stratigraphy and Resources of the Michigan Basin
- gamma-ray methods
- gas storage
- heterogeneous materials
- homogeneous materials
- Macomb County Michigan
- Michigan Lower Peninsula
- natural gas
- reef environment
- reservoir properties
- reservoir rocks
- sequence stratigraphy
- three-dimensional models
- United States