High seawater carbonate saturation states and the chemically reactive nature of most low-latitude marine carbonate sediments make them prone to substantial diagenetic modification, beginning on the seafloor and continuing during subaerial exposure and subsequent burial. Diagenetic processes have a profound effect on how these rocks are preserved in the geologic record, including their propensity to serve as hydrocarbon reservoirs. For half a century, understanding of carbonate diagenesis has relied heavily on Quaternary–Holocene analogs based largely on studies of photozoan carbonate deposits of the Caribbean (e.g., Matthews, 1974; Halley and Harris, 1979; James and Ginsburg, 1979; Land and Moore, 1980; Longman, 1980; Allan and Matthews, 1982). In these settings, high Ω_aragonite values of seawater promote extensive submarine cementation (Opdyke and Wilkinson, 1990), and polymineralic fossil assemblages make these sediments prone to substantial modification upon subaerial exposure (Lohmann, 1988) or submarine burial (Melim et al., 2002). As a consequence, these deposits typically enter the burial realm in a lithified state with predetermined characteristics that strongly control the evolution of reservoir quality.

Beginning with the work of Lees and Buller (1972), increasing recognition and study of heterozoan carbonate (sensu James, 1997) depositional systems has extended understanding of the spectrum of conditions under which carbonate sediments accumulate and lithify (Nelson, 1988; James and Clark, 1997; Pedley and Carannante, 2006; James and Bone, 2011). In contrast to photozoan counterparts, heterozoan carbonates consist primarily of calcitic benthos that are light-independent and rely on elevated nutrient contents to flourish (Mutti and Hallock, 2003; James and Lukasik, 2010). Although relatively few in number, diagenetic studies of heterozoan carbonates have shown that the constructive early diagenetic processes that are common in photozoan systems are minor and significantly outweighed by destructive processes (Knoerich and Mutti, 2003, 2006; Smith and Nelson, 2003). Carbonate communities are not capable of building rigid frameworks, and thus their deposits accumulate as sands and gravels and are prone to winnowing and reworking by waves and bottom currents. Skeletal debris is infested by endolithic borers, which create intraparticle porosity. Because these bores are not cemented, grains are weakened and become increasingly susceptible to fragmentation and dissolution (Betzler et al., 1997; Frank et al., 2014). The dominance of destructive processes is linked to the environment of formation, typically cool settings where seawater carbonate saturation states are lower and nutrient availability relatively high (James and Lukasik, 2010). As a consequence, heterozoan carbonates enter the burial realm largely unlithified and so are susceptible to substantial modification during deep burial (Dodd and Nelson, 1988; Nelson et al., 1988; Nicolaides and Wallace, 1997a).

In this context, Mississippian carbonates of the Anadarko shelf occupy a somewhat unique niche, having formed in a low-latitude setting in which higher trophic resources related to oceanic upwelling supported a carbonate factory dominated by heterozoan and biosiliceous benthos (Franseen, 2006). Carbonate-replacive chert, formed via silicification of spiculitic facies during marine and meteoric diagenesis, serves as hydrocarbon reservoirs in Kansas (McCoy, 1978; Rogers et al., 1995; Montgomery et al., 1998; Watney et al., 2001; Franseen, 2006; Mazzullo et al., 2009) and north–central Oklahoma (Rogers, 2001). More recently, focus has shifted to development of unconventional resources in “tight” limestones in the subsurface of the Anadarko Basin. With unconventional production associated with horizontal drilling (Matson, 2013), understanding the diagenetic history of the Mississippian limestone will prove important to the development of exploration strategies. Recent diagenetic studies focused on parts of the section that serve as unconventional reservoirs are relatively few. They have, however, revealed significant heterogeneity in the nature, timing, and distribution of diagenetic processes and products, with eogenetic marine and meteoric processes dominating in some areas and late-stage hydrothermal processes in others (Morris et al., 2013; Goldstein and King, 2014; Elmore et al., 2019; Goldstein et al., 2019; Mohammadi et al., 2019a, b). This study uses petrography and stable isotope geochemistry to reconstruct the diagenetic history of Mississippian limestones that along the southern (distal) margin of the Burlington–Anadarko shelf.

During the Mississippian, widespread neritic carbonate deposition was common across the North American craton (Gutschick and Sandberg, 1983; Maughan, 1983; Lane, 1984; McKerrow and Scotese, 1990; Keith and Zuppann, 1993; Webb, 1994; Bachtel and Dorobek, 1998; Dorobek and Bachtel, 2001; Wynn and Read, 2007). Carbonate ramps were the dominant platform style across the continent with a wide variety of mud-dominated bioherms (e.g., Waulsortian and Waulsortian-type mounds) and oolitic limestones being common instead of traditional shelf margin reefs (Lane, 1984; Keith and Zuppann, 1993; Webb, 1994). The large carbonate province covering much of North America is often subdivided into regional-scale shelves, such as the Madison, Redwall–Escabrosa, Chappel, and Burlington (Gutschick and Sandberg, 1983; Figure 1). Large-scale coeval convergent orogenic events impacted the eastern, southern, and western margins of the continent (i.e., Acadian, early Ouachita, Antler orogenies, respectively; see Gutschick and Sandberg, 1983; Noble, 1993). Additionally, a large positive northeast- to southwest-trending tectonic feature known as the Transcontinental Arch separated the large carbonate shelves across North America and was likely subaerial during much of the Paleozoic (Lane and DeKeyser, 1980; Franseen, 2006; Figure 1). Associated with the widespread convergent tectonic activity was the formation of large foreland basins and troughs at the continental margins, including the Anadarko, Marathon, Illinois, Ouachita, Michigan, and Deseret (Gutschick and Sandberg, 1983). In some of these basins (e.g., the Anadarko), biosiliceous and heterozoan facies are common, which have been associated with possible upwelling conditions (Gammon et al., 2000; Franseen, 2006).

The Burlington–Anadarko shelf was dominated by neritic carbonate deposition, situated on a carbonate ramp environment (Lowe, 1975; Lane and DeKeyser, 1980; Gutschick and Sandberg, 1983; Watney et al., 2001; Franseen, 2006; Koch et al., 2014). Throughout the Mississippian limestone stratigraphy, there is a distinct lack of a photozoan-rich facies. Rather, carbonate units are composed largely of biosiliceous and heterozoan organisms (crinoids, siliceous sponges, bryozoans, etc.), which has led some authors to propose upwelling of nutrient-rich waters as a controlling mechanism for these unusual facies situated in the tropics (Gammon et al., 2000; Franseen, 2006). Shallow marine conditions occurred until the latest Mississippian when regional tectonic uplift (Ouachita) was associated with the convergence of the North and South American plates. This uplift led to widespread exposure and meteoric diagenetic alteration of the shallow shelf, which transformed large areas of the biosiliceous-rich limestone facies into a porous form of tripolitic chert referred to as “chat” by drillers (Montgomery et al., 1998; Rogers, 2001; Watney et al., 2001; Watney et al., 2008; Mazzullo et al., 2009).

The term “Mississippian limestone” is used informally to represent Kinderhookian–Meramecian siliceous skeletal packstones–grainstones and argillaceous mudstones–wackestones that were deposited across the southern portion of the Burlington–Anadarko shelf in present-day Oklahoma (Rogers, 2001; Watney et al., 2001; Koch et al., 2014). Similar age stratigraphy in Kansas is subdivided into multiple units including the Hannibal Shale, Gilmore City Limestone, Osage Limestone, Warsaw Limestone, Salem Limestone, St. Louis Limestone, and Ste. Genevieve Limestone (Maples, 1994; Montgomery et al., 1998; Rogers, 2001; Watney et al., 2001; Qi et al., 2007; Mazzullo et al., 2009; Koch et al., 2014). Correlation of individual units for long distances is generally difficult because of sparse age control and complex changes in lithology and facies throughout the region (Koch et al., 2014).

Results of this study are based upon detailed petrographic and geochemical examination of the Pan American 1 Albert Severin core from the Anadarko Basin (API: 35047007110000, T22N, R5W, Section 14), Oklahoma (Garfield County), which preserves a section of Kinderhookian through lower Chesterian strata (Koch et al., 2014). Lithologies, color, bedding contacts, sedimentary structures, contained fossils, and other characteristics were logged (Figure 2). A total of 334 core plug samples from across the spectrum of facies were obtained for further analysis. The nature and distribution of biotic components and diagenetic features were determined through standard petrographic and cathodoluminescence (CL) analysis of 50 thin sections (Figures 2, 3). CL petrography was carried out using a cold cathode CITL Mk5 system at a beam energy of approximately 15 kV and a beam current of 280 μA. Samples were also stained with alizarin red-s and potassium ferricyanide to aid in identification of carbonate phases (Dickson, 1966). No poroperm data, which might supplement petrographic observations, are available for this core.

Stable carbon and oxygen isotope compositions were measured on powders (~100 μg) drilled from polished plugs and thin section blanks using a microscope-mounted drilling assembly at the University of Nebraska–Lincoln fitted with drill bits with 0.5 millimeter tip diameters. Isotopic analysis of powders was conducted at the Keck Paleoenvironmental and Environmental Stable Isotope Laboratory at the University of Kansas using a ThermoFinnigan GasBench II in line with a Finnigan MAT 253 isotope-ratio mass spectrometer. Prior to analysis, samples were roasted under vacuum at $200°C$ for 1 hr to release volatile organic compounds. Samples were then placed in Exetainer vials and flushed with ultrahigh purity (UHP) helium for 5 min, three to four drops of anhydrous phosphoric acid were injected into each vial, and the sample was reacted at $25°C$ for 24 hr to release $CO2$⁠. Oxygen isotope ratios were corrected for ¹⁷O contribution (Craig, 1957). Values are reported in permil (‰) relative to the Vienna Pee Dee belemnite (VPDB) standard (Figure 2). Analytical precision is better than 0.12‰ for δ¹⁸O and 0.06‰ for δ¹³C values and was monitored through daily analyses of National Bureau of Standards (NBS) and other in-house carbonate standards. Tabulated data are available in Koch et al. (2014).

The Pan American 1 Albert Severin core from the Anadarko Basin, Oklahoma (Figure 2), preserves a continuous 537 ft (163.7 m) thick succession of Kinderhookian–lower Chesterian strata. The basal part of the section unconformably overlies the Upper Devonian–Lower Mississippian Woodford Shale. The Mississippian section is unconformably overlain by the Pennsylvanian Cherokee Shale (5590–5900 ft [1704–1798 m]).

Eight lithofacies were recognized by Koch et al. (2014): (1) fossiliferous siltstone (IRa), (2) cross-bedded microbioclastic grainstone (IRb), (3) fossiliferous wackestone–packstone (MRa), (4) massive to bedded chert (MRb), (5) spiculitic wackestone–packstone (MRc), (6) silty mudstone–wackestone (ORa), (7) glauconitic siltstone–sandstone (ORb), and (8) shale (B). Lithologies reflect a spectrum of subenvironments and depositional energies. Environments include low- to high-energy settings on the inner ramp (IR) that extended from shallow shelf areas in Kansas to deeper water settings in Oklahoma (Lowe, 1975; Lane and DeKeyser, 1980; Gutschick and Sandberg, 1983; Watney et al., 2001; Franseen, 2006; Mazzullo et al., 2009), to middle-ramp (MR) and lower energy outer-ramp (OR) and basinal (B) environments (Table 1). Inner-ramp facies, fossiliferous siltstone (IRa), and microbioclastic grainstone (IRb; Figure 4A, B) record a low-energy lagoon environment that developed landward of high-energy skeletal sand shoals, respectively. Moderately bioturbated wackestone–packstones of the middle-ramp facies association (Figure 4C) indicate deposition in a storm-influenced, low- to moderate-energy, normal marine setting, with deposition occurring near fairweather wavebase (Franseen, 2006; Koch et al., 2014). Spiculitic and cherty lithologies (Figure 4D) are interpreted to record the presence of demosponges (Mazzullo et al., 2009). Massive to weakly bedded mudstone–wackestones (ORa) record deposition in a low-energy setting below storm wavebase, with soft-sediment deformation features recording slumping and sliding. Crenulated synaeresis cracks and low-diversity assemblages of minute trace fossils in some intervals may record development of stressed conditions from time to time. Thin glauconitic siltstones and very fine sandstones and siltstones (ORb) record very slow deposition in a low-energy outer-ramp setting, whereas dark gray to black, fissile shale in the underlying Woodford Shale is consistent with deposition in poorly oxygenated basinal setting.

Facies are arranged in an overall shallowing-upward pattern from the base to the top of the core, progressing from mainly basinal facies from the base to approximately 6370 ft (1942 m), middle-ramp facies from approximately 6370–5960 ft (1942–1817 m) and inner-ramp facies in the upper 40 ft (12 m) of the core (Figure 2). The stratigraphic sequence can be described as a prograding system consistent with a 3rd-order depositional sequence consisting of a series of higher frequency 4th- $(>30ft[9m]thick)$ and 5th-order $(~10–20ft[3–5m]thick)$ sequences. Facies dispositions in the Albert Severin core indicate more or less continuous sedimentation, with no clear evidence for significant marine hiatuses or subaerial exposure. Carbon isotope chemostratigraphy is consistent with this interpretation (Figure 2, Koch et al., 2014), suggesting that the Anadarko shelf remained a stable carbonate-producing environment during Kinderhookian–early Chesterian time.

Across the spectrum of depositional textures, limestones in the Albert Severin core tend to be very fine-grained, in which highly fragmented skeletal debris rarely greater than 250 μm in diameter dominates. Compositional information from these microbioclastic carbonates is therefore best obtained from examination of thin sections (Figure 3). Petrographic analysis reveals that fossiliferous siltstones of the inner ramp (IRa) contain scattered crinoid and brachiopod fragments, whereas sandy, skeletal grainstones (IRb) contain peloids and a more diverse array of highly fragmented skeletal grains that include crinoids, brachiopod shells, bryozoans, and rare trilobites. Middle-ramp facies are variably spiculitic. Other common skeletal grains in spiculites, mostly fragmented, include brachiopods, bryozoans, crinoids, and ostracodes. Scattered skeletal material in outer-ramp facies includes spicules, ostracodes, and other microbiofragments. Basinal facies (B) are more or less devoid of skeletal debris. Laminae of sand- and silt-size material consist of quartz and glauconite. Evidence of aragonitic components is very rare $(<1%)$⁠, limited to inner-shelf grainstones within the upper 40 ft (12 m) of the section and are preserved as cement-filled voids bounded by thin micritic envelopes (discussed in the section on Diagenesis).

Grainstone and packstone textures range from open-grain fabrics, especially in crinoid-rich intervals where syntaxial overgrowths are common, to compact or fitted fabrics where crinoids are less abundant. Grain-to-grain contacts in these lithologies are commonly tangential to sutured. Inter- and intraparticle pore space is filled to varying degrees by sparry calcite and fine sediment, as discussed later under Diagenesis. In matrix-supported lithologies, including spiculites and variably argillaceous to silty mudstones, wackestones, and muddy packstones, grain orientation may be either random or aligned parallel to bedding. Silicified versions of these lithologies are characterized by random grain orientations. Larger grains such as crinoids often show evidence (e.g., dragging of surrounding sediment) of having rotated during physical compaction. These mud-rich lithologies are often bioturbated. Mud-filled burrows tend to have ovoid cross sections that align with bedding, suggesting that they have been affected by physical compaction. Within grainy sediment concentrated in burrows, fitted fabrics are common. Wispy solution seams, along which dark insoluble material is concentrated, are common in muddy lithologies.

The paragenetic sequence observed in the Mississippian section of the Albert Severin core, based on observations from 50 thin sections (Figure 3), is separated into 20 diagenetic events with variable effect on porosity (Figure 5). These events are grouped into a series of stages related to timing relative to deposition and diagenetic environment. Calcite cement morphology is described according to Folk (1965), whereas dolomite textural classification is that of Sibley and Gregg (1987). Choquette and Pray (1970) provide the basis for description of pore types.

The synsedimentary marine stage (Figure 5) comprises four events (1–4), including deposition (event 1), which occurred on the seafloor, and below the sediment–water interface to depths where sediments remained in contact with normal seawater. All predate physical compaction. Skeletal grains contain evidence for bioerosion and micritization (event 2) in the form of microendolithic borings (Figure 6A) and micrite envelopes (Figure 6B). Microendolithic borings, most commonly affecting brachiopod shells, are observed throughout the cored section. Borings have diameters of 10–100 μm and are characterized by straight to slightly sinuous simple cylindrical tubes with minimal branching (Figure 6A). Morphologies are characteristic of the Trypanites ichnogenus, interpreted to record the actions of siphunculid worms (Flügel, 2004). In some cases, borings remain empty and create intraparticle porosity. In other cases, they are filled with fine sediment or iron-rich blocky calcite associated with late burial. Rare micrite envelopes preserve the margins of presumed aragonite grains lost to dissolution during event 4 (Figure 6B). Micritic envelopes are restricted in their distribution to inner-ramp facies (IRa, b) in the upper 40 ft (12 m) of section. Inclusion-rich, bladed calcite cement (event 3) forms thin (20–50 μm), isopachous rims on the surfaces and intraparticle pores of brachiopods and bryozoans in inner-ramp grainstones (IRb) in the upper 40 ft (12 m) of the cored section (Figure 6C). This phase is characterized by patchy CL, indicating subsequent alteration during later stages of diagenesis.

Dissolution of aragonite (event 4) predates precipitation of blocky calcite (event 5) that subsequently filled moldic porosity preserved by micrite envelopes (Figure 6B). Blocky calcite cement is Fe-poor and characterized by CL zoning (Figure 6D). Like the isopachous bladed calcite cement, events 4 and 5 are only observed in inner-ramp grainstones in the upper 40 ft (12 m) of the section. Although these events could have begun in the synsedimentary marine environment, the presence of bright CL zones in Fe-poor, blocky calcite is typical of precipitates that form in meteoric phreatic systems where redox conditions allow buildup of Mn²⁺ (Choquette and James, 1990).

The early burial stage (Figure 5) comprises six events (6–11) that occurred as sediments were buried beneath the seafloor, beginning at depths where sediments were no longer in contact with unmodified seawater and extending to the onset of chemical compaction. Epitaxial overgrowths (event 6) on crinoid grains (Figure 4A) are observed throughout the stratigraphic section and are best developed in mud-poor lithologies. They are ubiquitous in Facies IRb where crinoid debris dominates. Crinoid debris with epitaxial overgrowths either float or are in point contact with other grains, indicating that cementation began prior to significant physical compaction. Dull CL is consistent with precipitation under reducing conditions where $Mn2+$and $Fe2+$were present in dissolved form (Choquette and James, 1990).

Below 6063 ft (1848 m), wackestones and packstones (middle- to outer-ramp facies) are partially dolomitized (Figure 3). Based on comparison with standard visual estimation charts, the degree of dolomitization is typically 5–10%. The earliest phase (event 7) comprises isolated rhombs (⁠$(<40µm)$ μm) of iron-poor, non-CL planar-s to planar-e dolomite (Figure 7A). This non-CL dolomite predates microcrystalline and cryptocrystalline quartz (event 8), which surrounds it (Figure 7A), and selectively replaces matrix and some fossil grains in middle- to outer-ramp facies (Figure 3). Silicified areas preserve open fabrics (Figure 7B), with randomly oriented grains floating or in point contact, whereas nonsilicified areas show compaction fabrics. In pervasively silicified intervals, primary textures are obscured, or only ghosts of original grains (primarily spicules) are visible. These relationships indicate that silicification began prior to physical compaction and likely continued with progressive burial. Throughout the core, molds after sponge spicules are commonly filled with blocky calcite with dull CL (event 9), both within and outside of silicified areas (Figure 7C). Presuming that the source of silica was spicules, preservation in calcitized form requires that calcitization occurred in close association with spicule dissolution, remobilization of silica, and the early stages of silicification. Dull CL blocky calcite also fills synaeresis cracks, which are crenulated and ptygmatically folded by compaction. Cement precipitation prior to compaction of cracks is suggested by reorientation, breakage, and twinning of the cement crystals within the structures (see also Mohammadi et al., 2019a).

Physical compaction (event 10), which postdates the onset of silicification, is event as fracturing of delicate skeletal grains such as brachiopod shells, grain reorientation, and compression of mud-filled burrows. Packstone and grainstone fabrics, including within burrow fills, are fitted, with grains in tangential contact (Figure 7D). These features indicate that where they are not affected by silicification or synsedimentary cementation, deposits entered the burial realm as unconsolidated sediment. A 2nd phase of iron-poor dolomite (event 11) is present in nonsilicified areas (Figure 7E). These dolomite rhombohedra show bright CL zoning not evident in the earlier phase, indicating that additional dolomitization occurred after silicification. Data in Mohammadi et al. (2019a) suggest that this phase may have precipitated from fluid with a seawater composition. Dolomite texture is crystalline (50–100 μm) planar-s to planar-e.

The late burial stage, beginning with the onset of chemical compaction and continuing to the present, includes four diagenetic events (events 12–15). Within packstone and grainstone with fitted fabrics, sutured grain-to-grain contacts (event 12) mark consumption of grains via pressure solution (Figure 8A). Sutured contacts are themselves crosscut by stylolites (event 13). In muddy lithologies, dark insoluble material is concentrated along stylolites and wispy solution seams. These seams partially cannibalize skeletal grains and CL-zoned dolomite rhombs (event 11), indicating that they postdate these features. Loss of carbonate and concentration of insoluble material along stylolites and seams suggests that these dissolution features continued to develop during further burial. Sharp-walled fractures (event 14) crosscut earlier diagenetic phases (Figure 8B). Whereas most are oriented vertically, rare examples are parallel to bedding. Porosity is evident within some fractures, but most are filled with blocky, nonferroan calcite cement, which appears dull under CL.

Evidence for significant hydrothermal alteration of the Mississippian limestone and equivalents has been documented in shelf areas laterally adjacent to the Arkoma and Anadarko basins (King and Goldstein, 2015; Goldstein et al., 2019; Mohammedi et al., 2019a). Although phases typically associated with hydrothermal alteration in the Mississippian limestone are present in the Albert Severin core, they are not prominent, and their distribution is highly localized. Megaquartz occurs in voids formed via dissolution of skeletal material (event 16), whereas iron-rich blocky calcite (event 17) fills moldic porosity after spicules in chert (Figure 9A). Aggregates of euhedral pyrite crystals (up to 1600 μm), identified in reflected light, crosscut phases associated burial diagenesis (event 18). Localized corrosion of event 11 dolomite (event 19) may also be associated with hydrothermal activity. Iron-poor baroque dolomite (event 20) occurs locally as cement in intraparticle pore space (Figure 9B) and as overgrowths on earlier dolomite phases. It is often found in association with silicification. Curved rhombs have dull CL that occasionally display faint zoning.

The microbioclastic nature of limestones in the Albert Severin core made it difficult to separate individual components during microdrilling. Therefore, stable isotope compositions reflect mixtures of matrix, skeletal material, and cement. The δ ¹³C values range from $−2.5‰to+3.5‰$ VPDB, with most values falling between $+1and+3‰$⁠. The δ ¹⁸O values range from $−7.5‰to−1.5‰$ (Figure 10). Examination of the isotopic data in a stratigraphic context (Figure 2) also shows considerable scatter in δ ¹⁸O relative to δ ¹³C values, which are more uniform and show systematic shifts. The lower 60 m (197 ft) of the core is marked by a positive shift in δ ¹³C values, which increase dramatically from $−2.2‰$ to a peak of $+3.3‰$ before decreasing again to $+0.5‰$⁠. Data from the overlying 400 ft (122 m) of core define a long-term increase to values averaging $+2.3‰$⁠, returning to approximately 0.8‰ at 5943.8 ft (1811.7 m; Figure 2). The uppermost 40 ft (12 m) of the core, through the lowermost Chester Group, is marked by a decrease in δ ¹³C values to a minimum value of 0.0‰ at the top of the core. As documented by Koch et al. (2014), the stratigraphic trends in δ ¹³C values through the Albert Severin core correspond to secular shifts that have been documented elsewhere through Lower Mississippian carbonates, including the global type section at Arrow Canyon, Nevada (Saltzman, 2003), the Chesterian section from Idaho (Batt et al., 2007), and the U.S. midcontinent composite section from Mii et al. (1999). On this basis, the carbon isotope record from the Albert Severin core is interpreted to preserve a record of regional–global changes in the carbon cycle during the Kinderhookian–early Chesterian (Koch et al., 2014).

Petrographic and isotopic data from the Pan American 1 Albert Severin core indicate that mesogenetic processes dominated the diagenetic history of this Mississippian limestone section. Petrographic observations show that diagenetic products related to early and late burial are pervasive, whereas those related to synsedimentary, meteoric, and hydrothermal alteration are localized (Figure 3). A pervasive burial diagenetic overprint is also indicated by isotopic data (Figure 10). The trend defined by most (90%) data in the set, namely large decreases in δ ¹⁸O accompanied by relatively small decreases in δ ¹³C values due to buffering by carbonate carbon, is typical of limestones affected by burial diagenesis (Choquette and James, 1990). The smaller population of samples with δ ¹³C values lower than $+1.5‰$⁠, which includes most samples of facies IRb along with select samples from facies IRa, MRb, and ORa, can be explained in the context of global secular shifts in δ ¹³C values (Figure 2; Koch et al., 2014).

In detail, diagenetic patterns are heterogeneous and vary as a function of lithology and sediment composition. Lithification prior to physical compaction was favored in grainy, crinoid-rich lithologies (IRb) and spiculitic facies (MRc). Epitaxial overgrowths on crinoid ossicles (event 6) occluded a significant amount (up to 70%) of primary porosity in encrinites and promoted resistance to compaction. The significance of epitaxial cement in this regard can be seen in encrinites: whereas crinoid grains are encased in cement and support an open framework, other nearby grains such as bryozoan and brachiopod debris meet along tangential and sutured contacts (Figure 4A). Patchy to pervasive silicification (event 8) of spicule-rich lithologies also led to early lithification. Lithification resulting from silicification is heterogeneous at the scale of a thin section, with silicified areas maintaining a matrix-supported texture of randomly oriented spicules and nonsilicified areas showing evidence of physical (grain reorientation, burrow compression) and chemical (sutured grain contacts, wispy solution seams, stylolites) compaction (Figure 7B).

By contrast, microbioclastic skeletal limestones (MRa) that dominate significant parts of the section (Figures 2, 3) show no evidence of early lithification. Rather, evidence of skeletal grain rotation and reorientation in wackestones and tightly fitted fabrics in skeletal packstones and grainstones indicate that these sediments were unlithified when they entered the burial realm (Figures 7D, 8A). Sutured grain-to-grain contacts, stylolites, and wispy solution seams in these facies record chemical compaction. Carbonate released during chemical compaction is a likely source of carbonate for late-stage cement phases. MR facies are partly dolomitized (5–10% via planar-e/s dolomite of events 7 and 11) at depths greater than 6063.9 ft (1848 m) (Figure 3). Whereas event 7 dolomite is associated with silicification, event 11 dolomite, which shows bright CL zoning, occurs in the matrix of nonsilicified lithologies. Event 11 dolomite is present at depths below 6120 ft (1865 m), more than 200 ft (60 m) below the subaerial unconformity developed at the top of the Mississippian section and within middle-ramp facies. Mohammadi et al. (2019b) report carbon and oxygen isotopic compositions of similar dolomite phases that are consistent with precipitation from Mississippian seawater. The petrographic character of these rhombs in the Albert Severin core indicates that they are certainly early burial phases that formed under reducing conditions prior to physical compaction. Wide spacing between rhombs in lithologies that show evidence for entering the burial realm as unconsolidated sediments indicates that these dolomitization events did not play a strong role in lithification.

All facies are locally overprinted by late-stage features that include cement-filled fractures that crosscut depositional fabrics and chemical compaction features (Figure 8B) and phases suggestive of hydrothermal alteration, including megaquartz, euhedral pyrite, and baroque dolomite (Figure 9B). The sporadic distribution of events 14–20 (Figure 3), in the Albert Severin core relative to observations of similar features on the Nemaha uplift, Cherokee platform, and Burlington shelf (Goldstein and King, 2014; Goldstein et al., 2019; Mohammadi et al. 2019b), suggests that late-stage fracturing and hydrothermal alterations are not uniformly developed across the Anadarko Basin.

Diagenetic events affecting the Mississippian limestone have led to the destruction of most primary and secondary porosity (Figure 5). Where not occluded through eogenetic and early burial lithification, intergranular porosity was lost through both physical (i.e., development of fitted fabrics) and chemical compaction (Figure 8A). Moldic porosity generated via dissolution of sponge spicules and silicification of surrounding matrix was filled soon afterwards with blocky calcite cement (event 9). Likewise, porosity developed during late-stage fracturing was subsequently filled with calcite in all but a few instances. Intraparticle porosity is retained in some cases, for example, in microendolithic borings, but these occurrences are not abundant. The most common type of porosity observed is stylolite-associated microporosity (Figure 8A).

Given the nature of the biotic assemblage and evidence for deposition in a region affected by upwelling, the mesogenetic overprint that characterizes Mississippian carbonates of the Anadarko Basin is perhaps best considered in the context of observations from heterozoan carbonates associated with temperate or cool-water settings (James, 1997) rather than the low-latitude counterparts that were accumulating elsewhere on North America during the Mississippian (Lane and DeKeyser, 1980; Gutschick and Sandberg, 1983; Maughan, 1983; McKerrow and Scotese, 1990; Keith and Zuppann, 1993). Heterozoan-dominated carbonates are known to accumulate in settings where weakly carbonate-saturated waters influence the dominant mineralogy of benthos (calcitic) and limit abiotic fixation of carbonate (i.e., ooids, cement) during deposition, and higher nutrient levels support communities of bioeroders that macerate sediment (James and Lukasik, 2010). As a consequence, synsedimentary diagenetic processes are largely destructive (Smith and Nelson, 2003). In these systems, marine cementation is not extensive, associated dissolution of aragonitic components or hardground development (Nicolaides and Wallace, 1997b; Nelson and James, 2000; Knoerich and Mutti, 2006; Rivers et al., 2008). These characteristics impart a low diagenetic potential on heterozoan carbonates and, with the exception of encrinites, make them prone to entering the burial realm as unconsolidated sediments that are later transformed via chemical compaction and cementation into rocks.

The strong biosiliceous component (Rogers et al., 1995; Montgomery et al., 1998; Rogers, 2001; Watney et al., 2001; Franseen, 2006; Mazzullo et al., 2009) and presence of glauconite (Goebel, 1968) in the heterozoan-dominated deposits that characterize the Anadarko Basin point to oceanographic conditions that differed significantly from those characterizing the Madison shelf and the Redwall–Escabrosa shelf–Marathon Basin margin (Figure 1). Any explanation of the nature of the biotic assemblage requires a mechanism for elevated nutrient input, whereas sparse indicators of synsedimentary lithification require a mechanism for depressing the carbonate saturation state. In this regard, upwelling of deep, nutrient, and silica-rich water from the Anadarko Basin is thought to have served as a primary control (Franseen, 2006). Marine upwelling has been called upon to explain the origin of some source rocks (i.e., Woodford Shale) that underlie the Mississippian rocks of the Anadarko Basin (Parrish, 1982). Continued upwelling during the Mississippian is supported by recent work on potential source rocks in the Mississippian limestone (Al Atwah et al., 2015). Total organic carbon (TOC) concentrations in darker-colored limestones are relatively high (average 2%, ranging up to 6%), consistent with elevated primary productivity and carbon export. Moreover, the kerogen is Type II, indicating a marine origin. Because they have been affected by decomposition of sinking organic matter, upwelling waters are commonly $CO2$-enriched and therefore have depressed carbonate saturation states (Jiang et al., 2015). Such conditions limit the propensity for inorganic carbonate precipitation (Opdyke and Wilkinson, 1990), including synsedimentary cementation.

A distinctive oceanography and biotic assemblage profoundly influenced diagenetic patterns in Mississippian carbonates that accumulated along the southern portion of the Burlington–Anadarko shelf in north–central Oklahoma. The diagenetic potential of deposits was low overall, favoring delayed lithification. The progradational sequence recovered in the Albert Severin core is dominated by microbioclastic skeletal wackestones–packstones, which accumulated in middle- to outer-ramp settings. These facies entered the burial realm essentially unlithified. Packed fabrics, deformed burrows, sutured grain-to-grain contacts, solution seams, and stylolites point to chemical compaction as the driver for cementation and lithification in these facies. Although they accumulated on deeper parts of the ramp alongside carbonate wackestones and packstones, spicule-rich facies experienced yet another diagenetic pathway, with silicification commencing during early burial and limiting the effects of compaction seen in other mud-rich facies in the core. The extent of silicification ranges from partial to complete, with significant spatial heterogeneity at the thin section scale. In crinoid-rich facies of the inner ramp, epitaxial calcite overgrowths lithified sediments prior to physical compaction. These facies are limited in their stratigraphic distribution to near the top of the section. Late-stage alteration through fracturing and hydrothermal processes is not as significant in the Albert Severin core as reported elsewhere (i.e., Nemaha uplift, Cherokee platform, Burlington shelf), suggesting significant regional variation in diagenetic processes relevant to reservoir development.

Outside of the aerially restricted “chat” facies that have long served as conventional reservoirs in the Mississippian lime, diagenesis has destroyed most primary porosity. Lithification during eogenesis (cementation) and early mesogenesis (cementation, silicification) reduced primary intergranular porosity in enrinites and spicule-rich facies. Intergranular porosity in microbioclastic packstones and grainstones was reduced by physical and chemical compaction. Intraparticle porosity is retained in some microendolithic borings, but these occurrences are not significant. Secondary porosity developed through dissolution of skeletal grains is filled by carbonate cement. Most porosity in these rocks is associated with late-stage fractures, developed across the spectrum of facies, and bedding-parallel stylolites, which occur in both inner- and middle-ramp grainstones.

We thank BP for allowing us to publish the results of this study. We thank Vyetta Jordan and Jonathan Green at the Oklahoma Geological Survey for allowing us to systematically sample the Pan American 1 Albert Severin core for this work. We thank Jay Gregg, Jim Hendry, and David Budd for thoughtful reviews, which helped to improve the manuscript.