- Society for Sedimentary Geology
Here we document the occurrence of locally common oncoids in the Cedar Mountain Formation of Utah in the Woodside Anticline area of the San Rafael Swell and use them to understand changes in the Early Cretaceous landscape and their effects on the dinosaur fauna. Detailed facies analysis is required to understand the context of these changes within the broader patterns of Mesozoic tectonics and the fossil record. Oncolite crops out in the Cedar Mountain Formation, directly overlying the Buckhorn Conglomerate. Oncolite is not widely distributed outside of the Woodside Anticline area. The oncoids are found in a bimodal population with the majority in the 2–5-cm-diameter range and a smaller population >25 cm in diameter. Nuclei are mostly rounded chert clasts and also include litharenite, polymict conglomerate, limestone, and both abraded and nonabraded dinosaur bone and wood fragments. Cortices are 3–5 mm thick with distinct, penecint (laminae that completely enclose a body; Hofmann, 1969), low-relief laminae. Some laminae are crenulated and comprise ministromatolites. The petrography of the oncoids suggests formation along lake margins where large fragments of reworked sedimentary clasts and dinosaur bones came to rest and were coated by bacterial mats. Caliche cements and coats some of the oncolite; these define a lake shoreline affected by fluctuating lake level. The isotope geochemistry indicates a combination of primary and diagenetic signals consistent with oncoid formation in open, ephemeral freshwater lakes.
Oncoids described from the Cedar Mountain Formation (CMF) above the Buckhorn Conglomerate Member (BCM; see Weiss and Roche, 1988) prove useful for landscape reconstruction. Oncoids are unique forms of microbialite broadly defined as coated grains >2 mm diameter, composed of concentric but typically incomplete laminations (Tucker and Wright, 1990). The fossil record of oncoids stretches back into the Archean, and oncoids form today in a number of marine and freshwater locations (e.g., Freytet and Plaziat, 1979; Peryt, 1983; Ratcliffe, 1988; Davaud and Girardclos, 2001). In most cases, oncoids occur in wholly submerged environments, typically of high energy. Common facies include high-energy tidal and river channels, beaches, and lake shorelines (Awramik et al., 2000; Dixit, 1984; Lanes and Palma, 1998; Rouchy et al., 1993). Concentric laminations form when localized microbial ecosystems coat the exposed surfaces of a clast and drive local cementation through microenvironmental changes in water chemistry. The dominant paradigm has been that oncoids accrete on the upper exposed surface because the cyanobacteria responsible for their formation grow toward the sunlight. Additional surfaces become exposed, and laminae accrete, as the growing oncoid flips over during high-energy movement from waves or streams.
Here we describe a deposit of oncoids from the CMF, directly overlying the BCM within the Woodside Anticline of the eastern San Rafael Swell (Fig. 1). The distribution, sedimentology, petrology, and stable isotope geochemistry of these oncolite beds help constrain the paleoenvironments of the lower CMF to further our understanding of the Early Cretaceous basin, east of the Sevier highlands.
The CMF of Utah and western Colorado is recognized as an important record of the tectonic and landscape evolution that took place from Late Jurassic through Eocene time (Weiss and Roche, 1988; Currie, 1998; Demko et al., 2004; DeCelles and Coogan, 2006; Roca and Nadon, 2007). Contrasting tectonic models rely on detailed facies analysis for deciphering landscape denudation in response to uplift and environmental shifts due to basin subsidence during and after the Sevier Orogeny. This time period is also significant as the strata record the change in dinosaur fauna from a transatlantic to a more endemic population (Cifelli et al., 1997; Kirkland et al., 1997, 1998, 1999) and is a major part of the discussion of the interchange of Asian faunas with the Americas (Kirkland et al., 2005; Britt et al., 2006). Of particular note are the enigmatic therizinosaurs, once thought to be Asian in origin but now potentially originating in the United States (Kirkland et al., 2005; Kirkland and Madsen, 2007). Still another emerging facet of this period is the widely variable climate, as recorded by paleosols and spring carbonates (e.g., White et al., 2005; Zakharov et al., 2005; Dumitrescu et al., 2006; Suarez et al., 2007). Understanding the details of the landscape through facies analysis will help model the complex interplay of tectonics, climate, and sedimentation to address these questions better.
The CMF was deposited unconformably on the Upper Jurassic Morrison Formation during the Neocomian?–Albian (Early Cretaceous period), though exact confining ages have yet to be determined (Currie, 1998; Kirkland et al., 2005). Previous and ongoing research into the CMF and equivalent formations has distinguished a variety of sedimentary facies that represent continental environments (Kirkland et al., 1997; Sprinkel et al., 1999). Originally defined by Stokes (1949, 1952) and long thought to be a monotonous sequence of unfossiliferous shales, the CMF is now divided into five members (Fig. 1A; see Kirkland et al., 1997). The different members are defined primarily on grain size and secondarily on distribution (Kirkland and Madsen, 2007).
The basal BCM is up to 25 m thick and confined to one or more major incised valleys (Currie, 1998). The majority of clasts within the BCM consist of Paleozoic-aged chert eroded from emergent mountains to the west. The remaining CMF members contain a variety of floodplain (Yellow Cat, Ruby Ranch, and Mussentuchit Members) and channel deposits (Poison Strip Sandstone). The stratigraphic relationship between these members is poorly constrained, though the Ruby Ranch and Mussentuchit Members are the most widely distributed. In this paper, we follow the stratigraphy of Kirkland and Madsen (2007), though we acknowledge that it is difficult to clearly distinguish the Ruby Ranch Member and Poison Strip Sandstone in this area.
GEOLOGICAL SETTING AND MATERIALS
The Woodside Anticline is a low, south-plunging anticline superposed on the larger, broad Laramide-aged anticline that defines the San Rafael Swell of central Utah (Fig. 1). In this area of nearly complete exposure, the indurated BCM holds up ridges and stands out among the more easily weathered mudstones of the underlying Upper Jurassic Morrison Formation or the overlying members of the CMF. In the area of the oncoids, there is not a continuous vertical section of the BCM. Currie (1998) reports a thickness of ∼16 m of BCM in the San Rafael Swell.
Detailed stratigraphic analysis from the Woodside Anticline area has yielded important clues toward interpreting the facies of the oncoid-bearing lower CMF (Fig. 2). Three lithofacies dominate the stratigraphy overlying the BCM and underlying the variegated mudstone of the Ruby Ranch Member: (1) ∼2–4-m-thick, well-sorted, medium-grained sandstone with low to moderate angle cross-beds and very rare dinosaur bones; (2) ∼0.5–2-m-thick, normal-graded beds that range from cobble-sized conglomerate at the base to gravel, or rarely sand-sized at top, with dinosaur bones locally common; and (3) ∼0.5-m-thick lenses of medium gray micritic, massive limestone with rare ostracodes, charophytes, and ruby-red stringers and nodules of chert. These lithofacies are not laterally traceable beyond 10 m. Oncoids are found primarily in the basal 40–100 cm of the conglomeratic units with scattered oncoids higher up. One oncoid was found within the sandy lithofacies. The clast composition of these conglomerates—largely sandstone and limestone—is distinctly different from the underlying BCM (primarily chert), suggesting a hiatus during deposition of the CMF (see also Currie, 1998). We tentatively assign these beds to the Poison Strip Sandstone based on comparison with other CMF outcrops in and around the San Rafael Swell (cf. Currie, 1998; Kirkland and Madsen, 2007).
Collections of oncoid samples for petrographic and stable isotope analysis were made during Summer 2004 from an ∼10-m-thick interval, directly overlying the BCM. Over 50 samples, ranging from fragments to aggradations of oncoids, were collected in the field. Locations were recorded with a handheld global positioning receiver. Orthogonal photographs of outcrops were used for measurements of oncoid sizes. Oncoid nuclei compositions were recorded in the field.
In the laboratory, petrographic analysis was carried out on both polished slabs and thick (40 μm) and thin (30 μm) petrographic sections. Photomicroscopy was carried out using a PaxCam mounted on both an Olympus BX-51 petrographic microscope and a Nikon Optiphot microscope with a Reliotron cathodoluminescence chamber. Measurements were performed with PaxIT! software (Midwest Information Systems, Inc., 2006). Samples were taken for analysis from slabbed oncoid samples using a Dremel drill with diamond-tipped bits. Carbon and oxygen isotope ratios of milligram-sized carbonate are reported as δ13C and δ18O values, where R is the ratio of 18O to 16O, and δ is (Rsample/Rstandard − 1)*1000‰, and the standard is Vienna Pee Dee Belemnite for carbon and oxygen. δ18O and δ13C of carbonate were measured using an automated carbonate preparation device (Kiel-III) coupled to a Finnigan MAT 252 isotope ratio mass spectrometer at the University of Iowa. Powdered samples were reacted with dehydrated phosphoric acid under vacuum at 75°C in the presence of silver foil. The isotope ratio measurement was calibrated based on repeated measurements of isotope standards NBS-19, NBS-18, and in-house powdered carbonate standards. Analytical precision is ±0.1‰ for both δ18O and δ13C (1σ). The carbonate-CO2 fractionation for the acid extraction is assumed to be identical to calcite.
The oncoids of the lower CMF are in a variety of sizes, reflecting the size of the nuclei, but retain a consistent microstructure in the cortices. In the field, oncoids are found in a bimodal population (Fig. 3). The majority range from ∼2 cm to 5 cm diameter, with a mean of 3.4 cm. The larger group includes oncoids >25 cm diameter. The oncoids are light gray, subround to oblate, and the external form mimics the shape of the nucleus. Within one oncolite bed, ≤70% of the grains are coated, and random oncoids are found within the overlying conglomerate (Fig. 3B). A 10–15-cm-thick caliche crust overlies several of the oncoid beds. Very rare oncoids are found in cross-bedded sandstone. No oncoids were observed in lacustrine limestone either in the study area or at other locations in the lower CMF.
Nuclei are largely rounded but nonspherical chert clasts. Other nuclei include litharenite, polymict conglomerate, and limestone (Fig. 3C). Of special note are locally common nuclei of both abraded and nonabraded dinosaur bone and wood fragments (Fig. 3D). Dinosaur bones range up to 20 cm long. Most are broken; however, one oncoid coated a vertebral column with extended, fine processes. The dinosaur bones have not been identified but are tentatively assigned to therapods and ankylosaurs based on rough field comparisons to better-identified dinosaur caches elsewhere in the lower CMF.
Cortices are generally 3–5 mm thick, but rinds >30 mm thick were measured. Thickness around the oncoid is not consistent, with one side generally 2–3 times thicker than the opposing margin. Weiss and Roche (1988) described oncoids from the CMF on the Gunnison Plateau as irregular, laminated, and mossy. Laminae are penecint (partially enclosed) and circumscribe the entire oncoid. Individual laminae are either low relief (<0.05 mm high) and crenulated or built up into ministromatolites (Fig. 4). Laminae average 0.24 mm thick and show a moderate degree of inheritance (Figs. 4A–B). Ministromatolites average 4.2 mm wide (range = 2–6 mm wide) and 1.2 cm tall (Figs. 4C–D). The oncoids are nearly devoid of organic remains, like most microbialites. In one thin section, however, a densely interwoven mat of filaments was preserved, albeit poorly (Fig. 4F). The filaments are hollow with no partitions observed. The walls are thick (average = 6.7 μm), but the thickening may be due to diagenesis. The filament diameter averages 30.7 μm. These filaments were most likely produced by bacteria. While it is not possible to assign them to a higher taxonomic grouping, the filaments are putatively of a sulfide-oxidizing bacterium similar to the modern Beggiatoa or Thioplaca based on the orientation and size of the filaments (e.g., Kojima et al., 2003; Nikolaus et al., 2003).
Most of the oncoids show a moderate degree of diagenetic alteration, most notably by the coarsening of minerals. Some laminae have been replaced by fan-shaped crystal aggregates (Freytet and Verrecchia, 1999); the true, finely laminated nature is preserved under cathodoluminescence. Silicification is rare. Overall, the predominant petrofabric is a spongy micrite typical of post-Cambrian oncoids. Disseminated pyrite occurs both along lamination and more randomly scattered throughout the oncoid. Pyrite concentration increases around original voids.
Stable Isotope Ratios
Multiple samples from four oncoids, multiple samples from a typical lacustrine limestone, and one dinosaur bone were sampled and analyzed (Table 1). Carbon and oxygen isotope data are characterized by a cluster of relatively high isotope ratios with values ranging from −9.5‰ to −8.5‰ and −4.0‰ to −3.0‰ for δ18O and δ13C, respectively, and a tail of samples characterized by a range of lower isotope ratios (with minimums of −11.0‰ and −8.5‰ for δ18O and δ13C, respectively; see Fig. 5A). Focusing only on samples with the highest isotope ratios, overlap among oncoid-limestone specimens is limited (Fig. 5B). A weak positive correlation between δ18O and δ13C is observed for these same samples (Fig. 5B). The correlation coefficient is 0.70 if all of these samples are included, and it is 0.60 if the oncoid sample with the lowest isotope ratios is excluded.
Oncoids and Landscape Reconstruction
The lack of marine deposits of this age and the association with adjacent lacustrine deposits (e.g., massive limestone) suggest that a lake shoreface is the most likely environment of formation for the CMF oncoids. The asymmetrical laminations and irregular shapes of the oncoids suggests that the oncoids did not form as rolling aggregates in high-energy channels or shorelines but, rather, formed in place along the lake margin where fluvial gravels were deposited from a change in flow velocity (cf. Freytet and Plaziat, 1979; Dahanayake et al., 1985). In some respects, oncoid formation is more akin to pisoid formation in soils or caves. Locally common abraded dinosaur bones and wood fragments within the oncoids suggest transport prior to setting and oncoid formation. As oncolite is not composed entirely of coated grains and some oncoids are found in predominantly conglomeratic beds, it is possible some postaccretion transportation took place. The lack of abraded margins and fracture casts of cortices does not indicate significant postaccretion transport. In some cases, several of the oncoid layers are overlain by a nonmicrobial caliche crust, likely formed during lake drawdown. For an extensive caliche crust to form, the area needed to be subaerial. Whether the entire lake disappeared is unknown. In essence, this crust is analogous to marine beach rock that forms at the high-water mark in tropical regions where evaporation is high.
Considering these inferred environments of oncoid formation in the context of stacked lithofacies it is possible to reconstruct the broader nature of the Early Cretaceous landscape. The lateral discontinuity of facies and aerial distribution suggest a broad plain of ephemeral lakes during oncoid formation. Unit stacking suggests that lakes periodically emptied and were overrun by fluvial processes, which took place several times during the history of the ephemeral lake (Fig. 6). The times of prolonged submergence allowed for the establishment of microbial mats leading to extensive oncoid formation on cobbles that were sometimes wave winnowed.
Stable Isotope Geochemistry and Lake History
The overall pattern in oxygen and carbon ratios for oncoids and associated carbonates most likely represents a combination of primary isotope ratios (i.e., the cluster of higher values) and samples that have undergone diagenetic alteration to varying degrees under different geochemical conditions (Fig. 5A). These data can be compared to stable isotope data from paleosol carbonates of the CMF of the same general area (Skipp, 1997) to provide a more complete interpretation of the CMF landscape.
In general, δ13C values of our unaltered samples are higher than those of paleosol carbonates, whereas δ18O vales are slightly lower (Fig. 5A). Paleosol carbonate δ13C values are typical for the Late Jurassic–Early Cretaceous (e.g., Skipp, 1997; Ekart et al., 1999, and references therein). These values reflect oxidation of terrestrial organic matter in soil, diffusion and atmospheric CO2 mixing, and precipitation of carbonate from groundwater in equilibrium with the soil CO2 reservoir (Cerling, 1991). In contrast, higher δ13C values of oncoids are consistent with formation in open, freshwater lakes (Talbot, 1990). In particular, δ13C are interpreted to reflect multiple sources of carbon, including dissolved Paleozoic marine carbonate rocks in the watershed (δ13C ∼ 0‰) and oxidized C3 plant matter (δ13C ∼ −27‰).
Oxygen isotope differences between groups of carbonate also point to a lacustrine origin for the oncoids. Paleosol carbonates form in equilibrium with soil waters (Cerling and Quade, 1993; Hsieh et al., 1998) that can have δ18O values higher than mean annual because of (1) evaporative loss of water with 16O to the atmosphere during dry seasons and (2) preferential formation of carbonates during these dry seasons. In contrast to soil water, lakes are much larger reservoirs that collect precipitation from the entire year, and modification of δ18O due to evaporation is only a major factor for smaller ponds located in more arid environments (Talbot, 1990). The lower δ18O of oncoids and associated carbonates compared to paleosols is, thus, consistent with their formation in medium to large lakes rather than smaller ponds or shallow soil settings.
Detailed evaluation of isotopic systematics of oncoids and associated carbonate indicates that evaporation of lake water took place, although evaporation does not appear to have had an extreme influence on lakes of the CMF where oncoids formed. First, for example, the weak positive correlation between δ18O and δ13C for all these samples (Fig. 5B; R2 = 0.47) suggests that evaporation and organic productivity in the lake may have varied through time, with higher isotope ratios reflecting greater productivity during drier time periods (Talbot, 1990). Second, stable isotope ratios of altered samples reflect diagenetic conditions after burial, particularly the sample with the lowest δ18O and δ13C value (Lohmann, 1988). Low δ13C values of diagenetic carbonate are consistent with carbon sourced from the oxidation of soil organic mater and incorporation of this carbon in groundwater. Low δ18O values are again consistent with a meteoric water source, although values are ∼2‰ lower than unaltered carbonate (Fig. 5A). Such an offset between primary lake and diagenetic carbonates could reflect (1) a higher temperature (by ∼10°C) of formation from groundwaters having a δ18O similar to lake water or (2) similar temperatures of formation of both primary and diagenetic carbonate, but with groundwaters having δ18O ∼ 2‰ lower than lake water. The latter would imply that δ18O of lake water was shifted slightly to higher values relative to local meteoric water-precipitation that recharges groundwater. We prefer this latter interpretation given the sedimentological evidence for the limited size of these lakes and for the inferred episodic drying of them in this area during the Early Cretaceous.
Another interesting aspect of the oncoid δ18O data is the relatively low values. δ18O of ancient lake waters can be estimated by assuming temperatures of formation and using known isotopic fractionation factors for calcite and water (O'Neil et al., 1969). For a reasonable range of Early Cretaceous surface temperatures of 20–25°C, lake water is estimated to have been between ∼ −6.5‰ and −7.5‰. A similar calculation using δ18O from diagenetic and meteoric samples indicates that these waters have values of ∼ −8.5‰ and −9.5‰. Regardless of whether precipitation or lake water in this area had δ18O values of −7‰ or −9‰, these values are low compared to low-elevation localities at similar latitudes (∼35°N) during both present-day (Dutton et al., 2005) and hothouse climate states (the Eocene; see Fricke, 2003). The implication of these lower δ18O values is that rivers feeding the CMF lakes captured precipitation from high-elevation areas in the region. This inferred topographic relief is consistent with sedimentological evidence for high-energy rivers (i.e., conglomerates in the BCM) flowing from the Sevier orogenic belt.
The oncoid-bearing lacustrine deposits of the CMF described here help elucidate the Early Cretaceous landscape evolution in the area of the Woodside Anticline. Together with the underlying BCM, these rocks record the change from a classic continental fluvial succession, dominated by channel sandstone and floodplain deposits to lacustrine mudstone and limestone. The details of the oncoids record a unique situation where evaporation of small, ephemeral lakes preserved bacterial carbonates that coated rock fragments along the shore. Included in these fragments are dinosaur bones. Although constrained to the Woodside Anticline, the oncoids of the basal CMF further detail Early Cretaceous depositional environments as a series of ephemeral and evaporative lakes that formed on a braid plain of rivers originated from the Sevier highlands to the west.
We thank Stephen Hasiotis, James Kirkland, and an anonymous reviewer for helpful comments and criticisms that greatly improved the manuscript. The field research was funded through an National Science Foundation Research Experiences for Undergraduates Reconstructing Rivers grant to Dr. Julie Maxson (Metropolitan State University, Minnesota).
↵* Corresponding author
- Accepted July 30, 2008.