Controls on microbial and oolitic carbonate sedimentation and stratigraphic cyclicity within a mixed carbonate‐siliciclastic system: Upper Cambrian Wilberns Formation, Llano Uplift, Mason County, Texas, USA

The upper Cambrian Wilberns Formation in central Texas records deposition on a low‐gradient shelf within a mixed carbonate–siliciclastic tidal‐flat system that changes offshore to subtidal shelf and open‐marine oolitic skeletal shoals with large microbial mounds. Siliciclastic sediment is interpreted to have been delivered to the tidal flat by aeolian processes because of the narrow range in grain size and paucity of clay. Tidal influence is dominant as evidenced by reversing currents and desiccation on the tidal flat, and megaripples with reversing current indicators in offshore shoals. Intraclastic conglomerates were deposited in broad channels on the tidal flats during storm surges. Microbialite deposition is interpreted to be controlled by accommodation favouring amalgamated thin biostromes developed in the tidal flat vs. larger mounds with greater synoptic relief in the offshore, and current energy resulting in preferential elongation of offshore mounds in a NE–SW orientation. Intertidal mounds and biostromes grew in the presence of significant siliciclastic flux and trapped it within their structure, whereas offshore large buildups incorporated little siliciclastic component. Oolite and skeletal grainstone formed in tide agitated shoals associated with large subtidal microbial mounds. Storms extensively recycled and redistributed skeletal and oolitic sands from the offshore shoals across the shelf as thin sand sheets. Spatial mixing of siliciclastic and carbonate sediment occurred across the tidal flat and shelf. Low‐frequency and intermediate‐frequency stratigraphic cycles were driven by shifts in the shoreline and changes in rate of siliciclastic flux in response to relative sea‐level fluctuation. Random facies stacking and the lack of metre‐scale cyclicity are interpreted to reflect stratigraphic incompleteness and an episodic signal introduced by storms.


| INTRODUCTION
Marine microbialite and oolitic carbonate sediments are important proxies for physical, chemical and biological environmental conditions in the rock record. Microbial and oolitic carbonates are also fundamental components of two distinct benthic carbonate factory types, or modes of carbonate sediment production, that reflect environmental parameters, water depth and production potential (Schlager, 2000(Schlager, , 2003Reijmer, 2014). Furthermore, microbial and oolitic facies have received intensive sedimentological study as they are common hydrocarbon reservoirs across the geologic column (e.g. Cambro-Ordovician, Permian, Triassic, Jurassic) hosting some of the world's most prolific hydrocarbon fields (Mancini et al., 2013 and references therein).
The widespread interest in microbial and oolitic carbonates has resulted in improved understanding of the environmental factors controlling their genesis. Studies of modern marine microbial carbonate mounds have constrained their morphologies and microscopic structures and elucidated the physical and chemical factors affecting their morphology, growth substrates, microbial communities and their role in sediment trapping/binding and precipitation, together with the taphonomic processes affecting preservation (Dill et al., 1986;Reid et al., 1995Reid et al., , 2003Ginsburg and Planavsky, 2008;Planavsky and Ginsburg, 2009;Dupraz et al., 2011;Suosaari et al., 2016). Studies of ancient microbial carbonates have identified microscopic, mesoscopic and macroscopic internal structures, secular trends in microbial forms and abundance across geologic history vs. change in environmental factors such as climate, seawater chemistry and biotic evolution, and environmental factors effecting growth forms (Rowland and Shapiro, 2002;Riding and Liang, 2005;Riding, 2006Riding, , 2011Kiessling, 2009;Bosak et al., 2013). Similarly, studies of modern and ancient oolitic carbonates have advanced understanding of the microstructure of ooids, physical and biochemical controls on precipitation and sedimentary structures, and secular trends in morphology and mineralogy across Earth history vs. chemical and environmental factors (Wilkinson et al., 1985;Sumner and Grotzinger, 1993;Trower et al., 2017;Harris et al., 2019). Despite these advances, it is still uncertain what environmental factors (such as seawater carbonate saturation, biological factors, siliciclastic input, accommodation space or current energy) played dominant roles on precipitation, deposition and heterogeneity of microbialite and oolite deposits. Analysis of microbialite and oolite in the context of environmental indicators of associated facies in ancient depositional systems promises to help identify the constraints on genesis.
The occurrence of extensive flat-pebble conglomerates in carbonates preceding the Middle Ordovician has been interpreted as promotion by the high seawater carbonate saturation state and reduced depth of bioturbation prior to the great Ordovician biodiversification event (Wright and Cherns, 2016). However, a variety of mechanisms have been proposed for the formation of flat-pebble conglomerates and there is no consensus on their primary mode of genesis (Demicco and Hardie, 1995;Myrow et al., 2004;Alvaro and Clausen, 2007;Pratt and Bordonaro, 2007). Further detailed analysis of Cambrian examples promises to better constrain their origins.
Cambrian-Ordovician strata of the Great American Carbonate Bank (GACB) are important for evaluating the controls on mixed carbonate-siliciclastic depositional systems in epicratonic seas (Cowan and James, 1993;Spencer and Demicco, 2002;Morgan, 2012;Runkel et al., 2012;Labaj and Pratt, 2016), the role of tides, storms and other hydrodynamic currents such as gravity flows and tsunamis on shelf sedimentation (Pratt, 2002;Tape et al., 2003;Myrow et al., 2004;Eoff, 2014;MacNaughton et al., 2019), and mechanisms impacting the record of high-frequency stratigraphic cyclicity (Demicco, 1985;Koerschner and Read, 1989;Goldhammer et al., 1993;Cowan and James, 1996;Lehrmann and Goldhammer, 1999;Runkel et al., 2007Runkel et al., , 2008. However, previous studies have focussed on geographic areas other than central Texas, and there is a need for quantitative analysis of sediment composition and spatial mixing of carbonate and siliciclastics during deposition and the levels of order vs. randomness in cyclicity. Moreover, there is a need for analysis of the roles of sea-level fluctuation, siliciclastic flux and sediment reworking by currents in such depositional systems. Previous studies of the upper Cambrian carbonates in the Llano area (central Texas) have mainly focussed on regional facies, stratigraphic mapping and biostratigraphy (Bridge et al., 1947;Sliger, 1955;Barnes and Bell, 1977;Miller et al., 2012), with only a few studies on the microbial reef structures and palaeoecology (Ahr, 1967(Ahr, , 1971Ruppel and Kerans, 1987;Spincer, 1997). The authors recent research efforts have included detailed characterization of the smaller microbial mounds and biostromes of the lower Point Peak Member, the larger microbial buildups and associated sediments in the upper Point Peak Member, and detailed characterization of the facies succession of the upper Morgan Creek Member and Point Peak Member of the Wilberns Formation in the area along the Llano and James rivers, and Mill Creek south-west of Mason, Texas Khanna et al., 2019;Proctor et al., 2019).
The purpose of this paper is to provide an analysis of the environmental mechanisms controlling the genesis of | 279 LEHRMANN Et AL.
2002; Kyle and McBride, 2012). The Hickory Sandstone is composed of arkose, subarkose and diagenetic quartz arenite with variable rounding and grains with crescentic shapes indicative of aeolian transport (McBride et al., 2002;Kyle and McBride, 2012). The sand is interpreted to have been sourced from the weathered basement of the Llano Uplift. Detrital zircon geochronology of the Hickory Sandstone provides ages consistent with sediment derivation from the Llano basement (Jones and Surpless, 2018).
The contact between the Riley and Wilberns formations, at the base of the Welge Sandstone Member, is a disconformity throughout the Llano area and is interpreted to be a supersequence boundary defining the base of the Sauk III sequence (Krause, 1996;Morgan, 2012). The Welge Sandstone Member records the initial transgressive deposit above the supersequence boundary. The Morgan Creek, Point Peak and San Saba members also reflect an overall transgression with thick microbial mounds developed in the Point Peak Member and a progressive reduction in siliciclastic content and eventual shift to dominant carbonate content during deposition of the San Saba Member and Ellenberger Group ( Figure 1B; Kyle and McBride, 2012). Presumably, siliciclastic sand of the Lion Mountain and Wedge Sandstone members may have been sourced by the local Llano Precambrian igneous and metamorphic basement, although by the time of deposition of the Morgan Creek Member, much of the Llano basement may have been buried by marine deposits ( Figure 1B).
During the Cambrian, the North American craton straddled the equator and was flanked on all sides by passive continental margins developed as a result of the Neoproterozoic breakup of Rodinia. During the Late Cambrian and Early Ordovician, under greenhouse climate conditions, rising sea level had flooded large parts of the continent forming widespread epeiric seas (Lochman-Balk, 1970). Siliciclastic sediments were shed from low-relief uplands such as the transcontinental arch forming an inner detrital belt, and tropical climate conditions stimulated widespread carbonate deposition of a vast, shallow-marine carbonate platform of the GACB (Ginsburg, 1982;Wilson, 1994;Bohacs et al., 2012;Runkel et al., 2012).
Previous studies have interpreted the palaeogeography of the Llano area during Wilberns deposition to either include a shoreline to the west with offshore deposition towards the east (Ahr, 1967), a shoreline to the north, with a E-W trending microbial reef barrier and offshore sedimentation to the south (Spincer, 1997) or a trend from nearshore to the north-east to offshore to the south-west (Ruppel and Kerans, 1987). Sections within Lampasas and San Saba counties in the north-east part of the Llano area are composed of siliciclastic sand and siltstone and lack carbonate strata ( Figure 1C,D). Furthermore, this area has been interpreted as the largest upland bedrock area of the Llano Uplift and the source of siliciclastic flux to earlier Hickory Sandstone deposition (McBride et al., 2002, p. 3). A local uplift in this area likely persisted as islands contributing siliciclastic sediment and affecting the shoreline configuration during Morgan Creek and Point Peak deposition. Isopach maps of the Wilberns Formation indicate that the Point Peak Member thickens into depocentres to the south-east and south-west of the subject field area (Barnes and Bell, 1977). The Point Peak Member thins and is truncated 160 km to the west on the Texas Arch, an upland extension of the Transcontinental Arch in west Texas (Barnes and Bell, 1977). The Wilberns Formation was thus deposited in a vast, low-gradient epicratonic carbonate ramp with a distant shoreline to the west and a local shoreline and siliciclastic source area in the north-east of the subject field area in the Llano Uplift ( Figure 1C).

| METHODS
Detailed overlapping stratigraphic sections were described bed-by-bed at the centimetre-scale from the Shepard Pavement (Llano River), Lower Mill Creek, Middle Mill Creek, Droxrock, Fallen Blocks and the Bunkhouse localities (Figures 2 and 3). Sections were photographed, sampled and logged for spectral gamma-ray measurements with a RS-125 scintillometer with 0.1-0.3 m vertical spacing. Sections were described in several field visits using an iterative approach in which detail from lithofacies characterization gathered from one visit and calibrated through subsequent petrographic work was integrated into observations made during subsequent visits. Palaeocurrent directions were measured from megaripples. The stratigraphic sections were correlated using distinctive lithofacies marker beds and signatures F I G U R E 2 Satellite image of study area with localities of measured stratigraphic sections (yellow) and additional outcrops studied (red). Locality names are as follows: (1) Shepard Pavement, (2) Lower Mill Creek, (3) Mill Creek Cliff, (4) Droxrock, (5) Fallen Blocks, (6) Bunk House, (7) Shepard Cliff, (8) Zesch Cliff, (9) Zesch Pavement and (10) James River Pavement in gamma ray logs. Markov chain analysis of the composite stratigraphic section was performed using the techniques described in Lehrmann and Goldhammer (1999).
An aerial photography survey was conducted with a drone (quadcopter) equipped with a Sony NEX-7, 24.3 MP camera, gimble and a GPS. The images were georeferenced using  DGPS and total station surveys. Drone images of the Shepard and Zesch Pavements, Shepard, Zesch and Middle Mill Creek cliffs, Droxrock and Fallen Block localities ( Figure 2) were used in this study to quantify macroscopic architectural features such as mound architecture and megaripples. Observations of the architecture of large microbial mounds in the upper Point Peak Member spectacularly exposed on the James River pavement were used to complement field observations from other sections.
Hand samples were used to make thin sections and polished slabs to document facies characteristics and smaller sedimentary structures. Drill cores (2.5 and 8 cm diameter and up to 60 cm long) were taken from a series of transects in Droxrock where lithofacies were difficult to systematically sample with a hammer. Numerous large blocks of entire microbial mounds, cross sections through megaripples and layers of flat-pebble conglomerate were collected and prepared as polished slabs to document larger scale structures. Polished slabs were imaged with a flat-bed scanner or digital camera. Slabs and thin sections were studied using a low-power binocular petrographic scope and a high-power petrographic scope. The nomenclature of Riding (2011) was used for description of microbial fabrics. Point counts were conducted using the grain-solid method (Flügel, 1982), with a variable number of points per sample (minimum of 100) depending on sample heterogeneity, using the program JMicroVision. A total of 248 polished slabs and 222 thin sections were analysed in this study.

| Fine to medium sandstone
Fine to medium sandstone is present in the uppermost Morgan Creek Member and basal Point Peak Member and within the middle Point Peak Member at Mill Creek section (Figures 2 and  3). Sandstone beds are laterally continuous across outcrops and are friable and recessive making it difficult to determine details of the sedimentary structures. Where well exposed, sandstone is massive or horizontally laminated. The sandstone is composed of poorly sorted angular to subrounded grains of quartz, feldspar, mica, and in some cases, a trace of admixed glauconite pellets, skeletal or oolitic grains ( Figure 4A). Intergranular space contains micrite matrix or equant calcite cement.
In the uppermost Morgan Creek and basal Point Peak members, sandstone is interbedded in sharp contact with sandy packstone-grainstone, skeletal packstone-grainstone and oolitic lime packstone-grainstone. In some cases, sandstone occurs in gradational contact with sandy lime-packstone to grainstone (Figure 3 and Figure S1). In the middle Point Peak Member, sandstone occurs within a thick recessive interval dominantly composed of siliciclastic siltstone ( Figure 3 and Figure S1).

| Siltstone to very fine sandstone
Siltstone to very fine sandstone is the dominant siliciclastic facies and occurs as thin intercalations within the carbonate facies throughout the facies succession with the exception of the thick microbial mounds in the upper Point Peak Member ( Figure 3 and Figure S1). It is also included as very thin intercalations within the heterolithic facies described below. The middle of the Point Peak Member is a recessive, relatively poorly exposed interval, dominantly composed of siltstone to very fine sandstone with thin subordinate carbonate intercalations ( Figure 3 and Figure S1).
Siltstone to very fine sandstone beds weather recessively, have sharp contacts with intercalated facies, and are laterally continuous across outcrops except for localized areas where they onlap onto microbial mounds (see Discussion in microbial boundstone section below). An olive green colour in some intervals indicates the presence of disseminated glauconite. Siltstone to very fine sandstone is composed of angular quartz, feldspar and mica with a micritic carbonate matrix or microspar calcite cement ( Figure 4B,C). The dominant sedimentary structure is horizontal lamination ( Figure 4B). Extensively bioturbated, glauconite-rich intervals occur within the uppermost Morgan Creek and lower Point Peak members at the Shepard Pavement and Mill Creek sections (Figures 3, and 4D and Figure S1). Very fine sandstone intervals at 13.7, 14.5 and 15.8 m in the lower Point Peak Member at the Mill Creek section, associated with heterolithic, oolitic and microbial biostrome facies contain symmetrical ripples and interference ripples exposed on bedding surfaces ( Figure 4E). The symmetrical ripples have wavelengths ranging from 2 to 7 cm (av. 5 cm) and heights of 0.5 to 0.6 cm (av. 0.55 cm).  (Figure 3 and Figure S1). The facies is moderately recessive, but exceptionally well-exposed in cliff sections ( Figure 5A). The heterolithic facies is thin bedded and typically laterally continuous across outcrops except for areas where intervals contain interspersed microbial mounds. The heterolithic facies is composed of centimetre-scale to decimetre-scale intercalations of horizontally laminated siltstone and lenticular to continuous, very fine grained peloidal and skeletal carbonate packestone ( Figure 6A). The siltstone is identical to the facies described above; it ranges from tan or olive green in colour indicating the variable presence of disseminated glauconite. Packstone interbeds are composed of very fine sand-sized peloids and trilobite fragments, echinoderm debris, angular quartz silt and glauconite.

|
Packstone interbeds are lenticular to continuous. The lenticular structure results from asymmetric ripple forms, and starved asymmetric ripples, with scour structures and reversing current indicators ( Figure 6A,B). Rippled intervals are draped by siltstone. The base of the carbonate interbeds is commonly sharp and planar, and may be formed of scoured surfaces that truncate laminae in the underlying siltstone. Intervals a few centimetres thick contain bundles of thickening and thinning muddy/grainy carbonate and silt lamination that are interpreted to be tidal bundles ( Figure 6C,D; Tape et al., 2003). V-shaped cracks in the upper part of carbonate beds or carbonate-rich siltstone were observed in vertical cross sections ( Figure 6A,C) within the heterolithic facies in the carbonate-rich portion of the lower Point Peak (Mill Creek section, 13-20 m) and below the upper microbial horizon in the upper Point Peak (Mill Creek section, 44-48 m, Fallen Blocks section, 1.3-5.5 m; Figure 3 and Figure S1). In cross section, the cracks are regularly spaced laterally, have a V-shaped morphology that penetrates downward into the bed and are infilled by overlying sediment (Figure 6A,C). Identical features were observed in bedding-plane exposures where they form polygonal cracks in heterolithic facies below the upper microbial horizon in the upper Point Peak at the Zesch Cliff Locality ( Figure 6E). These features are interpreted to be desiccation cracks.

| Sandy or silty packstone to grainstone
Sandy to silty skeletal packstone to grainstone is a minor component of the stratigraphic section ( Figure 3 and Figure  S1). It is identical to the skeletal packstone to grainstone facies described below except that it contains admixed quartz sand or silt.

| Skeletal packstone to grainstone
Skeletal packstone to grainstone forms prominent resistant beds, laterally continuous across outcrops, within the carbonate-dominated portion of the section in the upper Morgan Creek-lower Point Peak and in the upper Point Peak (Figure 3 and Figure S1). Skeletal packstone to grainstone beds may be massive bedded, burrow mottled or contain cross-bedding and asymmetric megaripples. Prominent asymmetric megaripples and cross-bedding are well preserved at several Skeletal packstone to grainstone is composed of fragments of trilobite sclerites and eocrinoid ossicles with lesser gastropod and brachiopod fragments, peloids, glauconite pellets and rare ooids ( Figure 7C,D). Skeletal material is entirely fragmental and commonly broken and rounded ( Figure 7C,E), with the most complete skeletal components being rare eocrinoid hold-fasts on hardgrounds, and concentrations of disarticulated brachiopods found on bedding planes and in the troughs between megaripples in the lower Point Peak Member. Trilobite fragments are reworked clasts with adhered micrite preserved in protected inner curves "shepherd crooks" of the shell ( Figure 7D). Likewise, gastropods are preserved as reworked steinkerns ( Figure 7D).
Skeletal beds are dominantly grainstone. Packstone contains lime mud in crude laminae or burrow mottles within dominantly grainy sediment or as mud infiltrated into intergranular porosity ( Figure 7C). Isopachous-bladed cement and poikilotopic calcite spar on skeletal grains is followed by equant calcite spar infilling interparticle porosity ( Figure 7C,E). Skeletal grains are commonly selectively dolomitized. Skeletal packstone to grainstone beds are dominantly of medium bed thickness (av. 0.38 m) with the exception of a few thick beds in the carbonate-rich lower portion of the Point Peak and a very thick bed of skeletal-oolitic grainstone (1.8 m thick) adjacent to and at the base of the thick microbial mound horizon in the upper Point Peak (Mill Creek section 51.7-53.5 m, Figures 3, 8A and 9A). The base of this bed is important as it marks an abrupt shift from siliciclastic-dominated sedimentation below to microbial and oolitic carbonate-dominated sedimentation above. The bed is continuous across the field area but contains a variable proportion of skeletal vs. ooid content. In sections at Mill Creek, the bed is dominantly bioclastic with a conspicuous proportion of ooids, whereas in the James River pavement it is dominantly oolitic (Figure 2).

| Oolitic packstone to grainstone
Oolitic packstone to grainstone forms prominent resistant beds, laterally continuous across outcrops, within the carbonate-dominated portion of the section in the upper Morgan Creek-lower Point Peak and in the upper Point Peak (Figure 3 and Figure S1). Beds are most commonly thin to thick bedded except for the very thick bed of skeletal-oolitic grainstone at the base of the thick microbial mound horizon in the upper Point Peak Member (Figures 3, 8A and 9A). Oolitic beds are stratigraphically associated with skeletal packstone-grainstone and microbial boundstone (Table 1, Figure 3 and Figure S1). At three stratigraphic levels in the lower and upper Point Peak, megarippled oolitic grainstone sharply overlies flat-pebble intraclastic conglomerate ( Figure 10A,B and Figure S1). The very thick bed of skeletal-oolitic grainstone at the base of the thick microbial mound horizon occurs laterally adjacent to and underlying the lower part of microbial buildups ( Figure 9A). Sedimentary structures include cross-bedding and asymmetric megaripples and lenses of dolomitized micrite, sometimes occurring in the troughs between megaripples. The megaripples contain evidence of reversing current directions ( Figure 10A,B). Megaripples have wavelengths ranging from 74 to 150 cm (av. 107 cm) and amplitudes from 6 to 10 cm. Stoss slope angles range from 7 to 15° (av. 12°) and lee slope angles 10-21° (av. 16°). Azimuth data are shown in Figure 3.  the studied sections ranging from 54 to 130 cm (av. 84 cm) and amplitudes from 4 to 8 cm. Stoss slope angles range from 10 to 21° (av. 16°) and lee slope angles 17-28° (av. 23°). Azimuths of ripple crests average 106°-286° ( Figure 3). Megaripples and cross-bedded intervals are grainstone, whereas massive beds may be packstone. Oolitic grainstone in the lower part of the Point Peak Member at the Mill Creek and Shepard Pavement sections contains composite coated grains and rounded, pebble sized, oolite lithoclasts ( Figure 10A). Oolitic grainstone is typically strongly recrystallized or pervasively dolomitized with ooids preserved as ghost fabrics ( Figure 10C). Oolitic packstone contains micrite and peloidal sediment that infiltrated into intergranular pores ( Figure 10D,E). Fabric details are better preserved in packstone. Ooids contain peloid or skeletal grain nuclei and radial cortical fabric, outer margins may be micritized ( Figure 10D,E). No quartz nuclei were observed in any of the samples. Ooids are commonly preferentially dolomitized. Ooids range from 0.5 to 2.2 mm in size, very few grains exceed 2 mm. Oolitic packstone and grainstone may contain admixed trilobite, echinoderm and brachiopod fragments, and lesser rounded intraclasts, peloids and glauconite pellets. A quartz sand component was found only within oolite beds of the upper Morgan Creek Member at the base of Mill Creek section ( Figure 3 and Figure S1).

| Intraclastic conglomerate
Intraclastic conglomerate occurs in thin to medium lenticular channelized beds or sheets that are laterally continuous across outcrops (several tens of metres), dominantly intercalated with heterolithic facies with sharp or erosive contacts, and also commonly underlie microbial boundstone mounds and biostromes and oolitic grainstone-packstone beds (Figures 3, 8, 9A and Figure S1). A layer of intraclastic conglomerate occurs at the base of the very thick bed of skeletal oolitic packstone-grainstone that forms the substrate of large microbial  (Figures 8 and 9A). The channel margins are usually not smooth, but rather step-like, where flat pebbles were plucked from the intervening thinly layered heterolithic sediment.
In the Mill Creek section, two fundamentally different clast types are observed, often segregated into different intraclastic beds, although occasionally they may be mixed. Dominant clasts are rounded to subangular, flat pebbles composed of the heterolithic facies ( Figure 11A,B). The subordinate clasts are strongly micritized, oxidized and equant, with highly irregular, embayed margins ( Figure 11C,D). Petrographic observations indicate that micritized clasts contain micritization fronts that only partially replaced clasts; the preserved component is skeletal material similar to the skeletal packstone-grainstone facies ( Figure 11E). The oxidation is likely oxidized glauconite that had mineralized the material making up the clasts. Irregular embayments may be dissolution features, whereas regular circular embayments are likely to be borings ( Figure 11C). Margins and interiors of the micritized clasts also contain peculiar V-shaped tensional cracks that appear to have resulted from shrinkage during diagenesis ( Figure 11F). The matrix of intraclastic conglomerate is skeletal grainstone containing trilobite, eocrinoid and brachiopod fragments, peloids, quartz silt and glauconite pellets ( Figure 11B).
Flat pebbles are commonly oriented subhorizontally, crudely imbricated ( Figure 11A,B) and are occasionally found with spar-filled shelter pores beneath flat pebbles ( Figure 11D). In the lower Point Peak at the Mill Creek section (17 m) flat-pebble conglomerates locally have vertical "edge wise" orientations where they are stacked up adjacent to microbial mounds.
In the upper Point Peak in the interval a few metres below the upper, thick microbial mound horizon at the Zesch Pavement and below the Zesch Cliffs, a few layers dubbed "pin-wheel conglomerates" by Ahr (1967) have vertically oriented clasts with highly organized, imbricated and radially distributed domains (rosette fabric; Demicco and Hardie, 1995, their fig. 27).  Figure 3 and Figure  S1). Microbial intervals of the upper Morgan Creek and lower Point Peak members consist of small isolated mounds (typically less than one metre across) and laterally extensive biostromes ( Figures 5A and 12A,B). Biostromes are composed of coalesced oval to circular mounds ( Figure 5A,C). Coalesced mounds in biostromes also form elongate and fitted fabric where mounds develop straight, concave or nearly polygonal shapes by interference of growth against adjacent mounds ( Figure 12C). Individual mounds are separated by intermound detrital carbonate or heterolithic facies and change laterally over tens of metres into biostromes composed of coalesced mounds ( Figure 5A). Mounds within biostromes occur as circular or oval thrombolite structures and complexly infolded (invaginated) "meandroid" thrombolites that commonly contain a dense micritic leolitic or stromatolitic rind (Figures 5A,C and 12D,E). Thrombolite clots are one to a few centimetres across ( Figure 12D,E). Detrital sediment in the passages between mounds and "meandroid" structures in the biostromes of the lower Point Peak contain intercalated layers of skeletal-oolitic grainstone, carbonate mud and siliciclastic silt.

| Microbial boundstone
Isolated mounds in the lower Mill Creek section are nucleated on flat-pebble intraclastic conglomerate, have columnar to upward expanding morphology and have flanks that intertongue with intermound heterolithic facies ( Figure 12A,B).
Sharp-walled, isolated mounds have dome to flat-topped shapes, circular to elongate shape in plan-view and occur one to several metres apart or locally coalescing in the Shepard Pavement section from 4.7 to 5.6 m ( Figure 13A). The sharpwalled mounds range from 0.3 to 3.5 m across and are 0.4 to 0.9 m in height. The mounds are nucleated on an extensive skeletal-grainstone hardground pavement with adhered eocrinoid holdfasts ( Figure 13A,B). Early lithification of the hardground is indicated by isopachous-bladed marine cement. The mounds are onlapped and draped by siliciclastic siltstone that has been mostly stripped away by erosion.
Internal structures of bioherms and mounds in the lower Point Peak Member include thrombolitic mound interiors that are commonly partially dolomitized and dense undolomitized micritic leolitic or stromatolitic external rinds. Thrombolitic structures are composed of cauliflower or irregular lobate clots of dense or peloidal microbial micrite with intervening entrapped detrital carbonate and siliciclastic silt ( Figure 12E,F). Entrapped detrital carbonate allochems include trilobite and eocrinoid fragments and ooids ( Figure 12F). Within the complex invaginated "meandroid" biostromes microbial micrite shrubs progressively coalesce towards the outer surface of the structure forming a dense micritic rind ( Figure 12E). Microbialites in the upper Point Peak Member are composed of the large complex mounds prominent in the cliffs along the Llano River ( Figure 9A) and accessible for detailed study in the outcrops along Mill Creek and James River (Figure 2). Outcrops in the Fallen Blocks section and drill cores from James River demonstrate that the large microbial mounds developed on a flat-pebble conglomerate layer that also occurs at the base of the very thick and laterally extensive bed of skeletal-oolitic grainstone that formed at the same level and adjacent to the base of the large mounds ( Figures 8A and 9A; Khanna et al., 2019).
Analysis of drone images and ground observations of the James River pavement by Khanna et al. (2019) demonstrates that the microbial buildups in the upper Point Peak consist of several scales of macroscopic stromatolitic structures encapsulated in decimetre to metre-thick, dense micritic and finely clotted thrombolitic rinds that are nested to comprise larger buildups. These include decimetre-scale stromatolitic columns that are clustered into metre-sized intermediate-scale mounds ( Figures 9B and 13C), that are grouped into yet larger mounds tens of metres across, and finally, mound complexes hundreds of metres across . Intermediate to large-scale mounds exposed in the James River and Zesch pavements display plan-view shapes ranging from oval to circular (e.g. Figure 9B). Analysis of drone images by Khanna et al. (2019) demonstrates that elongate intermediate to large-scale mounds and complexes within the James River pavement have a dominant NE-SW orientation. Khanna et al. (2019) demonstrated that large microbial mounds in the upper Point Peak Member contain a threephase architecture including lower and upper buildup phases (phases 1 and 3) with a dense micritic and finely clotted thrombolitic rind and sharp outer wall of the mound onlapped by siliciclastic siltstone (Figure 9A,C), and an intervening phase 2 where the microbial buildup intertongues with detrital intermound facies ( Figure 9A). Scaling of vertical relief and siltstone onlap from drone images indicates that synoptic relief evolved from 1.6 to 4 m in phase 1, decreased to a maximum of 0.5 m in phase 2 and increased again to 3.5-5.4 m in phase 3. The minimum estimates reported here for phases 1 and 2 are based on buildup relief that is directly onlapped by siltstone. Maximum estimates include relief that can be traced on the outer buildup contact above the level of onlapping siltstone. No corrections were made for compaction. Tongues of material shed from microbial phase 2 are mud-poor packstone with angular microbial-micrite clasts, peloids and minor skeletal fragments ( Figure 9E). Intermound facies deposited concurrent with phase 2 microbial growth consists of skeletal and oolitic packstone to grainstone with intercalated beds of heterolithic facies (Figures 3  and 9A). Grainstone is massive to cross-bedded, whereas the packstone contains burrows. Skeletal material is bioclastic trilobite, eocrinoid and brachiopod debris.
Decimetre-scale stromatolitic columns are the dominant component in the interiors of the buildups during all three phases of evolution ( Figures 9B,D, 13C and 14A). The interior of the stromatolitic columns are partially or fully dolomitized and contain centimetre-scale clots and micritized burrow tubes 5 mm in diameter ( Figure 13C,E). In thin section, the fabric contains clots of microbial-micrite and peloidal-micrite masses with fenestral cavities ( Figure 14B,C), and micritized burrow tubes with backfilled sediment containing skeletal fragments in the tube ( Figure 14D). The stromatolitic columns range in plan view from simple circular structures ( Figure 13C) to complexly infolded (invaginated) structures ( Figure 13D). The invagination of these structures is interpreted to be the plan-view expression of branching. In cross section this complexity is represented by vertical columns that diverge and branch and coalesce vertically ( Figure 9D).
Thin sections of the stromatolite columns within the interior of the mounds contain traces of entrapped trilobite and eocrinoid fragments but lack a siliciclastic component ( Figure 14B). The dense micritic and finely clotted thrombolitic rind enveloping intermediate and large mounds ( Figure 9B) is undolomitized  and contains the best-preserved microbial microfabrics. Although the rind appears to be leolite in outcrop ( Figure 9B,C), polished slabs and thin section show the rind contains finely clotted thrombolite with millimetre-scale clots composed of Tarthinia and Girvanella ( Figure 14E,G).

D E C
A spectral gamma-ray log through the interior of the Droxrock buildup has very low total gamma ray and thorium values, with a subtle progressive upward increase indicating very low siliciclastic content (Figure 3). Elemental proxy data for siliciclastic flux (e.g. Al, Si, Ti) have low concentrations with no systematic upward increase through Droxrock, and minor increases along stylolites at phase boundaries (e.g. 7% Al 2 O 3 at the top of phase 1) indicating a possible increase in siliciclastic clay at the phase boundaries (Swartz et al., 2017).

| Stratigraphic cyclicity
Stratigraphic cyclicity in the Mill Creek composite section is interpreted from the facies succession in the stratigraphic log and from spectral gamma ray profiles (Figure 15). Sedimentary structures, texture and biotic composition of facies support the interpretation that siliciclastic and heterolithic facies formed in nearshore environments whereas skeletal, oolitic and microbial carbonate facies formed primarily in offshore marine environments (Table 1). Transgressive-regressive curves were interpreted from the facies associations indicating cyclic shifts between siliciclastic and heterolithic facies representing nearshore regressive stages, and skeletal, oolitic and microbial carbonate representing transgressive stages ( Figure 15).
The analysis yields a low-frequency transgressiveregressive cycle with maximum transgression occurring in the carbonate dominant interval in the uppermost Morgan Creek and lower Point Peak members, followed by  Figure 15). Intermediate-frequency cycles with an average thickness of 7.4 m were also interpreted from changes in facies from the section and gamma ray profile ( Figure 15). The section contains six complete and two partial intermediate-frequency cycles. It was impossible to define high-frequency cycles at the metre scale because there is no consistent fining-upward, coarsening-upward or shallowing-upward facies motif ( Figure 15 and Figure S1). Markov chain analysis indicates that the facies transitions are ordered (rule out null hypothesis of randomness at 95% confidence interval) at the scale of the low-frequency cycles, and for three out of seven of the intermediate cycles ( Figure 15). Random facies transitions are indicated at the metre-scale indicating the lack of statistically ordered facies successions at the scale of high-frequency cycles ( Figure 15). Markov chain analysis at the scale of low to intermediate-frequency cycles indicate significant associations (transitions) between heterolithic-intraclastic-siliciclastic facies and skeletaloolitic-microbial facies supporting the interpretation of a cyclic alternation between facies interpreted to represent nearshore vs. offshore environments ( the null hypothesis of randomness indicates that random processes operated at the higher frequencies ( Figure 15).

| Siliciclastic facies
Fine to medium sandstone and fine sand to siltstone facies are interpreted to represent marine nearshore tidal-flat environments with an aeolian sourcing of siliciclastic detritus to the shoreline. Siliciclastic facies are dominantly composed of laminated siltstone nearly identical in aspect and are associated with heterolithic facies interpreted to represent intertidal deposition across an expansive tidal flat with low wave-energy. The dominant horizontal lamination and angular silt size quartz, feldspar grains and the presence of mica support low-energy deposition. Sandstone is subordinate and concentrated within the upper Morgan Creek and middle Point Peak (Figure 3 and Figure  S1). Traces of marine skeletal fragments demonstrate marine deposition. Well preserved, but relatively rare wave-oscillation ripples occur associated with heterolithic and carbonate facies within the lower Point Peak (Figures 3 and 4E and Figure S1) unambiguously demonstrating wave agitation in a very shallow-water shoreline environment. Rare, thin intervals of glauconitic siltstone are extensively burrowed indicating the presence of a low-energy subtidal environment probably just seaward of the tidal flat. Concentration of medium to fine sand within the upper Morgan Creek may reflect long-term trends in sea-level fluctuation. Sand occurrence generally diminishes upward in the Morgan Creek to Point Peak reflecting transgression of the Sauk III supersequence ( Figure 1B, Morgan, 2012).
The aeolian delivery of siliciclastics to the depositional system is supported by the narrow range of grain size, the paucity of a clay fraction and by analogy to an aeolian interpretation for the deposition of the older Hickory Sandstone in the Llano area (McBride et al., 2002;Kyle and McBride, 2012). Aeolian delivery of siliciclastics to marine nearshore environments has been interpreted in several other studies of the GACB (Dott et al., 1986;Chow and James, 1987a;Myrow et al., 2004;Runkel et al., 2007Runkel et al., , 2012 and should be important as there were expansive, non-vegetated upland areas such as the transcontinental arch and various cratonic arches capable of providing an aeolian supply of sediment to epeiric seas.

| Heterolithic flaser-bedded siltstonepackstone
Heterolithic siltstone to packstone is interpreted to represent generally low-energy intertidal deposition across an expansive tidal flat. Flaser bedding consisting of intercalations of horizontally laminated siltstone and lenticular to continuous carbonate packstone layers is interpreted to represent repeated flooding and emergence by tides. Silt drapes are interpreted to represent slack tide. Carbonate packstone beds contain peloids, skeletal debris and glauconite reflecting marine influence. Reversing current indicators in ripple cross-laminated lenses reflect flood and ebb tidal currents. Carbonate beds containing thinning and thickening-upward coarse-fine lamination are interpreted to represent spring-neap tidal bundles (Tape et al., 2003). Desiccation cracks indicate emergence of the tidal flats during low tide. Similar flaser-bedded heterolithic facies have frequently been interpreted to be an intertidal component of the Cambro-Ordovician peritidal cycles (Goldhammer et al., 1993;Demicco and Hardie, 1995) and tidal influences have been widely recognized within the interior of the GACB (Tape et al., 2003;Runkel et al., 2012).
Desiccation cracks are relatively rare within the heterolithic facies and were commonly seen only in cross section. The desiccation cracks are preferentially preserved in carbonate interbeds within the heterolithic facies (e.g. Figure 6C,E). The paucity of clay in the system may have prohibited the formation and preservation of desiccation cracks within the siltstone component of the heterolithic facies; carbonate mud content and early lithification of carbonate interbeds enabled the formation and preservation of desiccation cracks. Although desiccation cracks had not been previously reported, Ahr (1967) and Spincer (1997) interpreted the heterolithic facies as intertidal, and Ahr (1967) interpreted associated flat-pebble conglomerates to be composed of clasts reworked from desiccation polygons from tidal flats. Abundant rounded clasts of heterolithic facies within the associated intraclastic conglomerate facies demonstrate early lithification, especially of carbonate-rich layers on the tidal flat. Reworking of clasts into intraclastic layers is interpreted to have occurred during storms.
Mud cracks in the Mesoproterozoic Belt Supergroup with a variety of polygonal, linear and spindle-shaped geometries, previously assumed to be desiccation structures, have been reinterpreted to be seismite deformation structures in which the cracks formed as fluidized sand was injected upward and downward into intercalated mud layers (Pratt and Ponce, 2019). The cracks reported herein from the Mill Creek and Fallen Blocks sections, however, are V-shaped structures that penetrate downward from the tops of beds and are infilled with sediment from above consistent with an origin as desiccation features (Figure 9). Furthermore, the sandstonesiltstone and heterolithic facies contain a suite of structures indicative of very shallow water and peritidal environments (small symmetrical ripples, flaser bedding with reversing current indicators, silt drapes and tidal bundling) consistent with the interpretation that cracks in the Point Peak Member formed by desiccation.

| Skeletal packtone to grainstone
Skeletal packstone to grainstone is interpreted to represent open-marine subtidal environments based on a stenohaline biota of trilobites, eocrinoids and brachiopods. High-energy currents are reflected by a grainy texture and the common occurrence of cross-bedding and megaripples. Skeletal grains in packstone and grainstone are entirely fragmental and often broken and rounded (Figure 7), and commonly contain adhered carbonate mud indicating that they were initially deposited in a lower energy environment and reworked by currents or wave action. Brachiopods are less abundant compared to trilobite and eocrinoid fragments in the skeletal packstone-grainstone (Table 2) and commonly are found disarticulated in local concentrations on bedding planes and troughs between megaripples indicating they may represent local living assemblages or were hydrodynamically concentrated.
Skeletal grainstone is interpreted to have been deposited in offshore subtidal shoals and sheets influenced by tidal currents forming megaripples and cross bedding with reversing current indicators (Figures 5 and 7). Storms are interpreted to have reworked the skeletal detritus into thin laterally extensive sheets across the shelf. Background deposition following storm events allowed carbonate mud to infiltrate the intergranular porosity and for bioturbation to mix in carbonate mud forming packstone. Lenses of dolomitized micrite represent lower energy conditions following storms and during slack tide. Presumably, trilobites and eocrinoids lived and were originally deposited in a low-energy environment further offshore prior to being reworked by high-energy currents and waves. Skeletal material is fragmented, broken and rounded. Either the sections measured in the study area formed entirely shelfward of such a low-energy environment or the sections represent amalgamated storm deposits resulting from repeated storm recycling of sediments and obliteration of low-energy background sedimentation. Bioclastic sands composed of trilobite, eocrinoid and brachiopod debris are common in Cambrian epicratonic sequences and have been commonly interpreted to reflect transport and deposition by storms and other high-energy currents (Palmer, 1989;Myrow et al., 2004;Pratt and Bordonaro, 2007;Labaj and Pratt, 2016;Runkel et al., 2012).

| Oolitic packstone to grainstone
Oolitic packstone to grainstone is interpreted to represent open-marine subtidal, tide-agitated shoals reworked into sheet deposits by storms similar to the skeletal packstone-grainstone facies (Figure 3, Table 1). Megaripples with reversing current indicators provide clear evidence for tidal-current agitation (Figure 10). Megarippled beds are grainstone and contain oolite lithoclasts indicating active tidal agitation of shoals, marine lithification and reworking of clasts, similar to modern analogues (Harris, 2019). Ooids likely formed in high-energy, tide-agitated shoals in an offshore subtidal setting . The very thick bed of oolitic skeletal grainstone formed coeval with the base of the large microbial mounds in the upper Point Peak (Figures 3, 8A and 9A) is interpreted to represent such offshore shoals with associated microbial mounds.
Thin oolite beds (0.16-0.46 cm thick) within the Morgan Creek, lower Point Peak and upper Point Peak (Figure 3) are interpreted to record sediment reworking by storms into sheet deposits across the shelf. After storm reworking, tidal currents resumed forming the megaripples that commonly cap the thin layers. Lower energy conditions in troughs of megaripples, or during pauses in currents allowed the accumulation of lenses of carbonate mud. Thin ooid packstone beds with micritized ooids, admixed skeletal fragments, peloids and carbonate mud infiltrated into intergranular porosity are interpreted to represent the waning energy following storms. Similar aprons of ooid packstone with admixed peloids and carbonate mud have been interpreted from modern analogues to represent oolitic deposits washed into the platform interior from the high-energy shoal environment of active ooid genesis (Harris, 2019). Orientation plots of the crests of megaripples compiled from skeletal and oolitic grainstone beds from the lower and upper Point Peak Member show a relatively wide spread in palaeocurrent directions with a dominant NE-SW trend (Figure 3).
Ooids in the Point Peak Member have nuclei composed of peloids or skeletal fragments and lack quartz nuclei supporting the interpretation that the offshore zone of ooid formation may have been beyond the reach of siliciclastic sand flux. Ooids have well preserved radial cortices indicating Mgcalcite mineralogy (Chow and James, 1987b).

| Intraclastic conglomerate
Intraclastic conglomerate beds within Morgan Creek and Point Peak are interpreted to have several origins. The most common conglomerate beds are sheets and broad, flat channelized deposits composed of rounded, tabular (flat pebble) clasts of heterolithic facies and equant, irregular, embayed and oxidized clasts (Figure 11). The broad, low-relief channels (up to 30 cm deep and several metres across) are cut into heterolithic facies and are interpreted to represent erosion and channelization during storm-surge onto tidal flats. The low relief of the channels and absence of features indicating lateral migration is incompatible with tidal channel origins (Shinn, 1983(Shinn, , 1986. Imbrication of clasts, and shelter porosity beneath tabular clasts are consistent with deposition by traction currents during storms.
The intraclastic conglomerates are intercalated with heterolithic facies (Figure 3 and Figure S1). The rounded flat pebbles are interpreted to represent erosion and reworking of partially lithified clasts from the heterolithic facies. The equant micritized clasts are interpreted to have been reworked from subtidal hardgrounds. The equant aspect reflects origin from a more homogeneous, non-tabular skeletal facies rather than the heterolithic facies. Micritization and oxidation of

F I G U R E 1 5 Mill
Creek Composite stratigraphic section with spectral gamma ray logs and interpretation of stratigraphic cycles. Integrates stratigraphic data from individual sections of Figure 3 into a composite section. Transgressive-regressive curve interpreted from facies associations and depositional environments (Table 1). Statistics from Markov chain analysis shown at right. Green bars indicate ordered facies transitions. Numerical values are chi-squared statistic normalized to chi-squared statistic that must be exceeded to rule out null hypothesis of randomness at 95% confidence interval. Values greater than 1 indicate ordered vertical facies transitions at 95% confidence interval some clasts is only partial, suggesting that the original lithology was skeletal packstone-grainstone ( Figure 11E). Regular rounded embayments in clast margins are interpreted to represent borings. Pitted, micritized, irregularly embayed and oxidized margins of the micritic clasts reflects hiatus, long residence time and diagenetic alteration of the clasts (Figure 11). The clasts are interpreted to have been reworked from micritized, mineralized (possible glauconitization that was later oxidized) and bored hardgrounds. Localized occurrences in the upper Point Peak Member of edge-wise "pin wheel" flat-pebble conglomerates with highly organized, radially distributed imbrication are interpreted to have been deposited by complex eddies in traction currents (Ahr, 1967) or by wave action in beach deposits by analogy to similar modern deposits (Demicco and Hardie, 1995). Finally, the flat-pebble conglomerate unit at the base of the very thick, skeletal oolitic grainstone bed and at the base of the large microbial mounds is interpreted to represent a transgressive lag-deposit composed of clasts eroded from the base of the underlying heterolithic facies (Figures 3, 8A and 9A; Khanna et al., 2019). The stratigraphic position at the base of a widespread transgressive ravinement surface and a major shift in the depositional system supports this interpretation ( Figure 15).
A wide variety of mechanisms have been proposed for the genesis of flat-pebble conglomerates (Demicco and Hardie, 1995;Myrow et al., 2004;Alvaro and Clausen, 2007;Pratt and Bordonaro, 2007). Mechanisms incompatible with field observations made during this study include slumping and gravity flow of partially lithified beds (Myrow et al., 2004), lags within migrating tidal channels (Demicco and Hardie, 1995) and transgressive lags at the base of high-frequency metre-scale cycles (Demicco and Hardie, 1995). Deposition as tsunamites (Pratt and Bordonaro, 2007) is a model compatible with field observations. However, because of the tectonically stable palaeogeographic setting and the lack of any stratigraphically unusual context that would support tsunami deposition, this mechanism is considered unlikely here.

| Microbial boundstone
Microbial boundstone is interpreted to form both in nearshore marine intertidal and shallow-subtidal environments forming thin intervals of mounds and biostromes associated with heterolithic facies in the lower Point Peak, and as larger offshore, open-marine subtidal mounds represented in the upper Point Peak (Figure 15).
Small microbial mounds in the lower Point Peak Member exposed along the Mill Creek cliff interfinger with heterolithic facies demonstrating that they formed in a marine tidal-flat environment concurrent with heterolithic deposition (Figures 5 and 14). Detrital sediment deposited in the passages between small mounds and within biostromes in the Shepard Pavement section contains alternations of skeletal-oolitic grainstone, siliciclastic silt and carbonate mud, indicating fluctuation in current energy. These are interpreted to be oscillations between high and low current energy consistent with flood-ebb and slack tide conditions on a tidal flat. Siliciclastic silt associated with the microbial structures is consistent with siliciclastic flux into the nearshore intertidal environment. Individual mounds coalesce laterally along the Mill Creek cliff outcrop to form laterally continuous biostromes ( Figure 5). The lateral coalescence into thin biostromes is interpreted to reflect low accommodation-space in the nearshore setting. The isolated mounds in the lower Point Peak at the Shepard Pavement section are nucleated on a skeletal grainstone hardground surface ( Figure 13) and have up to 0.9 m of synoptic relief onlapped by siliciclastic siltstone (Figures 3 and 13). These are interpreted to have been deposited in subtidal environments seaward of the tidal flats.
Microbial mounds in the upper Point Peak Member are interpreted to have formed further offshore in an open-marine environment with greater accommodation. Greater accommodation is supported by greater thickness, up to 7.7 m at Droxrock and up to 14 m at James River , as well as greater synoptic relief (up to 5.4 m; Figures 3, 9 and 15). The presence of a stenohaline fauna consisting of trilobite and eocrinoid fragments trapped within the mound and in associated oolitic-skeletal grainstones supports an open-marine interpretation. The low siliciclastic detrital component in the mounds can be explained by an offshore position. Khanna et al. (2019) interpreted a 10-15 m water depth in the environs of the large mounds based on an analysis of the amplitude of subaqueous dune forms in associated carbonate grainstone.
The interpreted three-phase evolution of the large microbial buildups, supported by margin architecture , indicates that the buildups developed greater synoptic relief and sharp-walled margins with a dense micritic and finely clotted thrombolitic rind in phases 1 and 3, separated by an interval of lesser synoptic relief and intertonguing with intermound detrital sediment during phase 2 ( Figure 9). Phase 1 is interpreted to have occurred concurrent with oolitic skeletal grainstone deposition represented by the very thick grainstone bed adjacent to and underlying the microbial buildups (Figures 3, 8A and 9A; Khanna et al., 2019). Upward growth of the buildups resulted in increasing synoptic relief above intermound detrital carbonate sediment and a sharp-walled geometry ( Figure 9A). A pause in carbonate deposition and influx of siliciclastic silt onlapped the sharp margin prior to phase 2 growth that occurred with lesser synoptic relief and concurrent with the accumulation of intermound detrital carbonate and siliciclastic sediment (Figures 3  and 9A; Khanna et al., 2019). This was followed by a sharp walled phase 3 nearly identical to phase 1 (Figure 9).

| Depositional model
From detailed interpretations of depositional environments, facies associations and stratigraphic succession (Figure 15), in the context of deposition on the GACB, the Morgan Creek and Point Peak members are interpreted to have been deposited on a broad, low-angle, mixed carbonate-siliciclastic ramp system ( Figure 16). The inner ramp was dominated by aeolian deposition of siliciclastic sand and silt and tidal-flat environments with siliciclastic-carbonate heterolithic facies gradually changing seaward to low-relief microbial mounds and biostromes. The middle ramp portion was dominated by subtidal shelf deposition with small isolated microbial mounds and sheets of skeletal-oolitic sand. Deposition in the offshore zone consisted of skeletal and oolitic sand shoals and large microbial mounds formed in open-marine environments with vigorous tidal currents and storm agitation ( Figure 16). Most of the biotic component, dominated by trilobite and eocrinoids, lived in and was initially deposited in low energy environments, possibly further offshore in deeper-water environments prior to being reworked as skeletal debris. Using a regional slope gradient of 0.1 m/km for the epicratonic ramp of the GACB (Runkel et al., 2007(Runkel et al., , 2012, the distance across the ramp from the intertidal zone to the zone of subtidal shoals and large microbial mound development is interpreted to have been ca 50-150 km based on minimum and maximum water depths estimated for large mound deposition . Relatively high rates of siliciclastic sediment flux into the inner ramp coupled with vigorous reworking of sediment across the shelf by storms and tidal currents are interpreted to have maintained a graded low-angle ramp profile. The depositional system experienced a dominant tidal influence both on the tidal flat (as evidenced by fluctuating current energy, ripples with reversing current indicators, tidal bundles), and as vigorous agitation of offshore shoals (as evidenced by abundant megaripples with reversing current indicators), as well as extensive wave influence during storm events, as evidenced by reworked sand sheets across the shelf, and broad channel fills of intraclastic conglomerates on the tidal flats. The result is a hybrid tide-dominated depositional system heavily influenced by storms and also a mixed carbonate-siliciclastic system developed on an extremely low-gradient epicratonic shelf that does not easily classify into end-member depositional models (Dalrymple, 2010;Plint, 2010;Pratt, 2010). Evidence of both tidal and wave influence is well documented from Cambrian epicratonic strata despite debate on the extent of wave and tidal currents in vast epicratonic seas (Klein and Ryer, 1978;Driese et al., 1981;Byers and Dott, 1995;Tape et al., 2003;Allison and Wells, 2006;Eoff, 2014;MacNaughton et al., 2019).
Reworked and rounded skeletal debris, ooids and intraclasts all provide evidence of frequent high-energy recycling. Notably, the skeletal packstone-grainstone contains the greatest proportion of glauconite (Table 2). Glauconite occurs as pellets and glauconitized clasts that also may provide evidence of extensive reworking. The glauconite likely mineralized at or just below the seafloor and was reworked as clasts in the skeletal sands (Chafetz and Reid, 2000;Chafetz, 2007). The presence of reworked oxidized clasts (interpreted to have originally been glauconitized) in intraclastic conglomerates also suggests the occurrence of high-energy subtidal hiatuses and long residence times for reworked clasts (Alvaro and Clausen, 2007;Eoff, 2014).

| Mixed carbonate-siliciclastic depositional system
Upper Morgan Creek and Point Peak deposition represents a true mixed carbonate-siliciclastic system in which siliciclastic silt and sand were delivered by aeolian deposition on the landward part of the depositional system and mixed with carbonate sediments on the tidal flat and shelf immediately seaward of the tidal flat. All but the oolitic facies contain some siliciclastic silt and sand component ( Table 2). The heterolithic facies contains a nearly equal mix of siliciclastic detrital components (Table 2). Small microbial mounds and biostromes contain siliciclastic grains entrapped in the microbial structures, as well as a mix of siliciclastic and carbonate detritus in the passages between the microbial structures. In the total volumetric breakdown from the Mill Creek composite section (Table 3) siliciclastic and carbonate depositional components are subequal with a slightly larger carbonate fraction (46%) and lesser siliciclastic fraction (35%). The Persian Gulf is the closest modern analogue as a mixed carbonate-siliciclastic, epicontinental tidal flat with silt delivered by aeolian processes (Riegl et al., 2010). In contrast to the Persian Gulf, however, there is no evidence for arid conditions in the Morgan Creek and Point Peak depositional system. A lack of land plants, rather than aridity, likely promoted aeolian delivery of silt to the Cambrian tidal flats. Bradley et al. (2018) point out that modern depositional models are inadequate for interpreting pre-vegetation tidal flat systems.
Fluctuation of siliciclastic input is interpreted to have been modulated by transgressive-regressive shifts in the shoreline driven by sea-level fluctuation with greater input of siliciclastic material during the regressive stage ( Figure 15, see discussion on stratigraphic cyclicity below). Alternatively, fluctuations in climate, potentially coupled with sea-level fluctuation, may have driven patterns of siliciclastic input to the shoreline. Although the transgressive-regressive model ( Figure 15) includes a concept of reciprocal sedimentation (Van Siclen, 1958;Wilson, 1967), carbonate and siliciclastic sediment coexisted and mixed in the tidal flat and shelf, changing only in relative proportion rather than alternating in time as in the traditional model of reciprocal sedimentation (Figures 15 and 16). Similar mixed depositional systems with a landward-dominant siliciclastic component have been described from ancient and modern analogues (Zeccin and Catuneanu, 2017). Spatially mixed systems should be especially common on broad ramps influenced by low-amplitude sea-level fluctuations during green-house climate periods, as such systems would lack the abrupt shutting on and off of carbonate factories from high-amplitude sea-level fluctuation typically invoked in the cyclic reciprocal model.

| Controls on microbialite genesis
The Morgan Creek and Point Peak members contain microbialite mounds and biostromes formed in both proximal intertidal environments and in offshore subtidal environments. A comparative analysis of these microbialites in the context of their depositional environment and associated sediments, as well as a comparison with modern and ancient analogues, illuminates environmental controls on their genesis.
Stromatolites and thrombolites require a stable, lithified or firm substrate to establish and maintain microbial growth. The Ginsburg and Planavsky's (2008) survey of modern microbialites in the Bahamas revealed a variety of "renewable" substrates formed during deposition such as hardgrounds, firmgrounds and skeletal or cemented clasts. Cambro-Ordovician examples in the Conococheague and Arbuckle groups and Bison Creek and Mistaya formations show that lithified clasts and flat-pebble conglomerates commonly form the substrate for growth of stromatolite and thrombolite mounds (Demicco, 1985;Westrop, 1989;Wilson et al., 1991;Fritz et al., 2012).
Substrates for microbialites in the Point Peak Member include subtidal hardgrounds, flat-pebble conglomerates and skeletaloolitic packstone to grainstone beds. It is possible that microbilites underlain by skeletal-oolitic packstone to grainstone beds nucleated on undetected hardground surfaces.
It is notable that although there is an association of microbialites formed on flat-pebble conglomerate layers in the Point Peak, there are also numerous conglomerate layers that lack microbial development (Figure 3). This pattern is especially evident in the upper part of the Point Peak Member, below the level of the large microbial-mound development, which contains numerous flat-pebble conglomerate layers intercalated in heterolithic facies (Figures 3, 8 and 9). Obviously, having an adequate substrate does not foster microbial development unless environmental conditions conducive to microbial growth persist after development of the substrate. In this interval, numerous intraclastic conglomerate layers were buried by heterolithic facies in the tidal-flat before environmental conditions could establish microbial growth. Higher in the section, microbial mounds were instead developed on a flat-pebble conglomerate substrate at the base of the very thick bed of skeletal-oolitic grainstone interpreted to be a transgressive lag deposit that also brought a shift to offshore environments conducive to microbial mound genesis ( Figure 15, Khanna et al., 2019).
Comparison of intertidal biostromes and small mounds of the lower Point Peak with the large mounds of the upper Point Peak indicates that siliciclastic flux played a relatively minor role in the growth of microbialites. Biostromes and small mounds in the lower Point Peak contain siliciclastic silt trapped within the microbialites and in passageways between mounds ( Figure 12). There is no evidence that the siliciclastic flux significantly inhibited growth. Low accommodation space was likely the primary factor that limited vertical growth and encouraged lateral coalescence of mounds into biostromes. The large microbial mounds in the upper Point Peak have a very low siliciclastic content as evidenced by the spectral gamma ray and concentrations of elemental proxies for siliciclastic flux (Figure 3, Swartz et al., 2017). Low siliciclastic content is interpreted to have resulted from their offshore deposition. Development of the three-phase architecture of the large mounds was interpreted by Khanna et al. (2019) to have resulted from pulses of siliciclastic sediment flux driven by falls in relative sea level. Proctor et al. (2019) reported up to 8% clay content in the large microbial mounds of the upper Point Peak and Swartz et al. (2017) reported elemental proxies for greater clay content in stylolites at phase boundaries consistent with such an interpretation. The model of sea-level fluctuation controlling siliciclastic flux and the three-phase architecture of the large mounds ) also interprets lowering of sea level at the end of each growth phase. No petrographic or geochemical evidence for subaerial exposure at the phase boundaries was found at | 303 LEHRMANN Et AL.
Droxrock, indicating that sea-level fall was not great enough to cause emergence of the mounds (Swartz et al., 2017).
Decimetre-scale thrombolitic or stromatolitic mounds and columns are the fundamental elements making up the biostromes and larger microbial mounds (Figures 5, 9 and 13). These features typically form simple columns with circular shape in plan-view but may take on more complexly invaginated shapes (Figures 12 and 13). The invagination seen in plan-view is interpreted to represent vertical branching during growth. Branching and complex-shaped columns have been interpreted to represent lower synoptic relief and sediment infill between columns allowing horizontal growth of columns forming branches, bridges and complex shapes (Bosak et al., 2013). Similar complex maze-like thrombolite structures have been reported from North America and China and have been associated with low synoptic relief (Shapiro and Awramik, 2006;Lee et al., 2014). The complexly invaginated structure is considered to be similar to the maze-like structure although the scale is different with the features in the Point Peak Member being several centimetres to decimetres across rather than centimetres in scale (Shapiro and Awramik, 2006;Lee et al., 2014).
In addition to accommodation space, the rate of intermound detrital infill vs. upward growth of microbial structures affects synoptic relief. The contrast between the sharp-walled mounds that are onlapped by sediment in the small isolated mounds ( Figure 13A) or phases 1 and 3 of the large mounds ( Figure 9) versus intertongued margins of small mounds ( Figure 12) and the intertongued phase 2 of the large mounds ( Figure 9A) reflects different rates of intermound detrital sedimentation. Development of sharp-walled mounds indicates a hiatus developed at the margin due to either high-energy bypassing of detrital sediment in the intermound areas or lack of detrital sediment flux as vertical mound growth developed synoptic relief. The sharp-walled rind of phase 1 development of the large mounds, which grew synchronously with oolitic and skeletal sand intermound deposition (Figure 9), resembles deposition of modern large subtidal mounds in the Bahamas, which are repeatedly covered and uncovered by oolitic sediment (Dill et al., 1986). The generally low content of coarse skeletal and oolitic detrital material trapped within the large microbial mounds in the upper Point Peak Member, however, records a striking dissimilarity between the two models.
Individual decimetre-scale mounds within biostromes and stromatolitic columns in the larger buildups contain a dolomitized clotted interior surrounded by a dense, laminar stromatolitic or micritic rind (Figures 9, 12 and 13; Khanna et al., 2019;Proctor et al, 2019). At the decimetre-scale the rind is formed of merged clots of microbial micrite or is stromatolitic ( Figure 13). In the intermediate to large-scale mounds the rind is a dense micritic, finely clotted thrombolitic structure containing calcified microbial frameworks of Tarthinia and Girvanella (Figures 9 and 14).
The contrast between the more porous thrombolitic interior and the dense non-porous and micritic and layered stromatolitic structure is interpreted to represent a greater hiatus and slower growth during development of the outer rind. Dupraz et al. (2011) emphasized that a key factor in stromatolite genesis is repeated hiatuses and lithification of microbial biofilms, thus resulting in the development of a laminated structure. Furthermore, the outer contact of the rind in phase 1 and 3 of the intermediate to large scale mounds in the Point Peak are consistently onlapped by siliciclastic silt (indicating hiatuses) whereas stage 2 lacks a rind and intertongues with intermound sediment indicating more continuous growth.
Microbial mounds encapsulated by stromatolitic or micritic rinds have been widely reported from upper Cambrian microbialites (Shapiro and Awramik, 2000;Lee et al., 2014;Coulson and Brand, 2016) and have been interpreted to be a global phenomenon (Rowland and Shapiro, 2002).
The biostromes within the lower Point Peak Member contain detrital carbonate and siliciclastic sediment trapped between clots whereas the large mounds of the upper Point Peak have little entrapped detrital material. This indicates that microbial micrite precipitation is the dominant mode of growth, with agglutination of detrital material playing a secondary role only in low-relief forms with a significant flux of detrital material onto the microbialite surface. The microbialites of the Point Peak are dominantly "fine grained" precipitated microbialites typical of the Palaeozoic in contrast to coarse agglutinate forms that are restricted to the Neogene and modern (Riding, 2011). The interior of intermediate and larger buildups is preferentially dolomitized when compared to the outer dense, micritic rind. The faster growth and greater depositional porosity in the mound interior may have led to greater permeability and preferential dolomitization.
Variations in external morphologies of microbial mounds in the Point Peak can be explained by competitive fitted growth of coalesced forms, or alignment of structures with current directions. Fitted structures of the biostromes in the lower Point Peak ( Figure 12C) record column growth close to each other forming "compromise boundaries". Intermediate to large mounds in the upper Point Peak are preferentially elongated in a NE-SW direction ( Figure 9B; Khanna et al., 2019) in alignment with tidal currents and directional oceanic currents (Figure 3; Khanna et al., 2019). Elongation of microbial structures by differential accretion and erosion by tidal currents has been recognized and modelled from modern examples such as the subtidal stromatolites in Shark Bay (Bosak et al., 2013;Suosaari et al., 2016).

| Controls on oolite genesis
The ooids of the Morgan Creek and Point Peak members have normal sizes (<2 mm) similar to marine ooids from modern depositional systems. The size of ooids has been interpreted to reflect seawater carbonate saturation state, with larger ooid sizes recording higher values, especially in the Proterozoic (Swett and Knoll, 1989;Sumner and Grotzinger, 1993) and Early Triassic (Lehrmann et al., 2012). Extensive development of microbialites and flat-pebble conglomerates have been interpreted to indicate high seawater carbonate saturation during the Cambrian (Riding and Liang, 2005;Wright and Cherns, 2016). Thus, the lack of large ooids indicates that factors other than seawater carbonate-saturation state must have played a greater role in controlling ooid size during Point Peak deposition. Further work is needed to evaluate the numerous possibilities to explain the discrepancy such as biotic control on ooid size, differences in hydrodynamic conditions, the abundance of nuclei or the impact on ooid precipitation of vast epicratonic seas.
Oolite deposition in the Morgan Creek and Point Peak members presents similarities with modern oolite deposition in the Great Bahama Bank. Both systems record deposition in high-energy shoals at the seaward platform edge or outer shelf with vigorous agitation by tide and wave currents, redistribution of oolitic sands as finer-grained sheet-deposits with admixed mud and peloidal sediment , and association of ooid shoals with large subtidal microbial structures (Dill et al., 1986;Reid et al., 1995;Feldmann and McKenzie, 1998;Reid et al., 2003;Dupraz et al., 2011). The two systems differ, however, in original mineralogical composition and microstructure. The oolites of the Morgan Creek and Point Peak members are characterized by radial cortical fabrics indicating an original Mg-calcite mineralogy and formation during the calcite sea conditions of the Late Cambrian (Wilkinson et al., 1985;Chow and James, 1987b).

| Controls on depositional cyclicity
The low-frequency and intermediate-frequency cyclicity in the Morgan Creek and Point Peak members are interpreted to represent transgressive-regressive shifts in the shoreline driven by relative sea-level fluctuation. The approximately 3.5 Myr duration of the Sunwaptan Stage (ICS, 2019) places the cyclicity in the third and fourth temporal orders (Vail et al., 1991). Fluctuation between nearshore and offshore facies associations is interpreted to represent lateral shifts of the shoreline and changes in siliciclastic input into the system and through the cycles (Figure 15). The Morgan Creek and Point Peak members, however, lack any consistent high-frequency (5th order) shallowing-upward peritidal, or graining-upward subtidal cycles at the metre scale similar to those widely reported from Cambro-Ordovician peritidal successions (Demicco, 1985;Koerschner and Read, 1989;Wilson et al., 1991;Cowan and James, 1993;Goldhammer et al., 1993;Fritz et al., 2012).
In contrast, the metre-scale stratigraphic architecture of the Morgan Creek and Point Peak members consists of random shifts between the various facies ( Figure 15 and Figure S1).
In the context of high-energy reworking of skeletal-oolitic-intraclastic sediment by storms, development of hiatuses in hardgrounds and erosion of channels, metre scale cyclicity is interpreted to be obscured by an incomplete record with numerous gaps and an episodic signal introduced by storms ( Figure 15). The Morgan Creek and Point Peak record may be represented more by gap than by sediment accumulation (Ager, 1992). Other mechanisms that have been invoked to explain randomness in peritidal successions, such as "missed beats" in sea-level fluctuation or autocyclicity, do not offer likely explanations for the randomness, as the Cambrian was a green-house climate period with low-amplitude sea-level fluctuations unlikely to produce "missed beats", and autocyclicity, despite introducing randomness, produces shallowing-upward cycles (Lehrmann and Goldhammer, 1999). The interpretation of an incomplete record for Cambrian deposition contrasts with Runkel et al. (2008) who interpreted a relatively complete record of epicratonic sedimentation despite low subsidence rates. Randomness, episodic deposition and gaps at the scale of high-frequency cycles in the Morgan Creek and Point Peak record of central Texas point to local oceanographic factors impacting the stratigraphic record.

| CONCLUSIONS
The facies of upper Morgan Creek and Point Peak members of the Wilberns Formation are interpreted to represent deposition on a broad, tide-dominated, mixed carbonate-siliciclastic ramp system influenced by storms. Aeolian delivery of siliciclastic sediment to the shoreline and spatial mixing of carbonate-siliciclastic sediments on the tidal flat is interpreted from textures and sedimentary structures of the sandstone-siltstone and heterolithic facies. Offshore oolitic and skeletal sand was deposited in shoals associated with large microbial mounds. Oolitic and skeletal sands as well as intraclastic conglomerate were reworked by storms.
Low-frequency and intermediate-frequency stratigraphic cycles were driven by transgressive-regressive shifts in the shoreline and changes in the rate of siliciclastic flux in response to relative sea-level fluctuation. The lack of cyclicity and randomness in facies succession at the metre scale are interpreted to reflect an incomplete record and episodic signal introduced by storms.
Microbialites formed both in the low-accommodation setting of the outer tidal flats and in the greater accommodation setting of offshore open-marine subtidal environments. Substrates for microbial growth are commonly subtidal hardgrounds and flat-pebble conglomerates. Biostromes are composed of coalesced thrombolitic mounds with dense leolite or stromatolite rinds. Larger mounds are composed of stromatolitic columns enveloped in dense micritic and finely clotted thrombolitic rinds. Cauliflower to lobate, centimetre-scale, thrombolitic clots are composed of irregular microbial micrite or peloidal micrite. Millimetre-scale, finely clotted thrombolites are composed of calcimicrobial frameworks. The dense leolite to stromatolitic outer rind of thrombolites in biostromes, and the dense micritic finely clotted rind enclosing larger mounds are interpreted to represent slower growth and a greater hiatus on the outer surface of microbial structures. The internal component of mounds is commonly dolomitized, possibly as a result of greater original porosity.
Dominant controls on microbial genesis are interpreted to be accommodation space and modification by currents. Siliciclastic flux is interpreted to have played a relatively minor role in microbialite growth as biostromes and small mounds grew uninhibited by siliciclastic flux and trapped siliciclastic silt within their thrombolitic structures. The large offshore microbial mounds developed with minor entrapped siliciclastic sediment but contain a three-phase growth history in which pulses of siliciclastic flux impacted growth phases. External mound shape is influenced by competitive growth especially evident in coalesced mounds in biostromes. A dominant north-east orientation of intermediate and largescale mounds in the upper Point Peak is interpreted to have been shaped by tidal currents.
Oolitic sediments show strong similarities to modern, outer platform oolite shoal deposits of the Great Bahama Bank, such as association with large subtidal microbial mounds, occurrence in tide agitated shoals and evidence of reworking of shoal sands into sheet deposits. Ooids have radial cortices indicating original Mg calcite mineralogy typical of a greenhouse calcite sea. Despite the fact that extensive microbialite development and other characteristics of the upper Cambrian lithofacies have previously been interpreted to indicate high seawater carbonate-saturation state, the ooids are similar in size to those of modern settings, with a notable lack of giant ooids, indicating that other factors must have competed with seawater carbonate-saturation state in controlling ooid size in this system.