Elucidating the relationship between the later Cambrian end-Marjuman extinctions and SPICE Event

Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology journal homepage: www.elsevier.com/locate/palaeo

Elucidating the relationship between the later Cambrian end-Marjuman extinctions and SPICE Event Angela M. Gerhardt a,b, Benjamin C. Gill a,⁎ a b

Department of Geosciences, Virginia Polytechnic Institute and State University, 4044 Derring Hall, Blacksburg, VA 24061, United States Enervest, Ltd, Houston, TX, United States

a r t i c l e

i n f o

Article history: Received 13 April 2016 Received in revised form 20 August 2016 Accepted 26 August 2016 Available online 29 August 2016 Keywords: Cambrian carbon isotopes extinction carbonates Chemostratigraphy

a b s t r a c t The late Cambrian-early Ordovician transition contains several discrete marine extinction events. The first of these extinctions, the end-Marjuman, occurs in two phases and is thought to coincide with the beginning of the Steptoean Positive Carbon Isotope Excursion, or SPICE, a large excursion in the marine carbon isotope record that represents a large perturbation to the carbon cycle during this time. Additionally, the carbon isotope record from the Deadwood Formation in the Black Hills of South Dakota, USA, displays a small negative δ13C excursion near the end-Marjuman extinctions. Here we examine the carbon isotope stratigraphy of the Upper Cambrian portion of the Conasauga Group of the Southern Appalachians, USA, to determine the relative timing between the extinction events and changes in the carbon cycle represented by excursions within the carbon isotope record. Previous high-resolution biostratigraphic studies have identified a thick stratigraphic record of endMarjuman extinctions within the Conasauga Group, making it an excellent target for a high-resolution chemostratigraphic study. In the Conasauga Group, there is no change in carbon isotope stratigraphy across the first phase of the end-Marjuman extinctions, suggesting no major change occurred in the carbon cycle during this time. Further, a negative δ13C excursion is absent in the Conasauga Group across the interval that contains the end-Marjuman extinctions. This suggests that the excursion in the Deadwood Formation is either a local oceanographic signal or a diagenetic feature. Finally, the onset of the SPICE occurs at the same stratigraphic point as the second phase of the end-Marjuman extinctions and at the appearance of a low diversity, potentially low oxygen tolerant, trilobite fauna. The stratigraphic positions of these biological and geochemical events suggest a role for marine anoxia in the second phase of the end-Marjuman extinctions. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The later Cambrian (Furogonian International Series) was characterized by at least three, two-phase extinction events that affected the shallow marine shelf communities of trilobites and brachiopods (Palmer, 1965a, 1984; Rieboldt, 2005; Taylor, 2006). The first of these extinction events occurred at the boundary between the Marjuman and Steptoean Stages of western North American (here referred to as the end-Marjuman extinctions). The end-Marjuman extinctions appear to coincide with the initiation of a globally expressed excursion in the marine carbon isotope record, referred to as the Steptoean Positive Carbon Isotope Excursion, or SPICE (Figs. 1 and 2: (Brasier, 1993; Glumac and Walker, 1998; Montañez et al., 2000; Saltzman et al., 1998, 2000)). Additionally, the carbon isotope records from stratigraphic sections of the Deadwood Formation in the Black Hills of South Dakota show a small (1‰) negative δ13C excursion that appears to occur near the interval of the end-Marjuman extinctions (Perfetta et al., 1999). ⁎ Corresponding author. E-mail address: [email protected] (B.C. Gill).

http://dx.doi.org/10.1016/j.palaeo.2016.08.031 0031-0182/© 2016 Elsevier B.V. All rights reserved.

However, this negative excursion has yet to be documented in stratigraphic successions elsewhere. The precise timing of the end-Marjuman extinctions relative to these isotopic events has yet to be resolved. In the Southern Appalachians — the target region of this study — high sediment accumulation rates produced an expanded stratigraphic sequence (tens of meters) during the end-Marjuman extinctions; elsewhere in North America, this interval is typically a few meters or less (Palmer, 1979, 1984; Stitt and Perfetta, 2000). Thus, the Upper Cambrian succession in the Southern Appalachians presents a unique opportunity for a high-resolution stratigraphic study of these events. The main goal of the geochemical study presented here is to determine the relative timing between the extinction events and changes in the carbon cycle, represented by the excursions within the carbon isotope record, during this interval of the late Cambrian. 1.1. End-Marjuman extinctions Biostratigraphic studies of the late Cambrian of Laurentia recognized that this interval was punctuated by three, two-phase extinction

Biomeres

North American Trilobite Zones

North American Stages

Ptychaspid

Sunwaptan

Stages Jiangshanian

Saukia Saratogia Taenicephalus Irvingella major

Dunderbergia

SPICE

Prehousia Dicanthopyge

Pterocephaliid

Steptoean

Elvinia

Paibian

Furongian

Epochs

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

Aphelaspis

Crepicephalus

-1 to +1

+4 to +5

Cedaria

Marjumiid

Marjuman

Guzhangian

Series 3

Coosella perplexa

13

δ C (‰, VPDB)

Fig. 1. Upper Cambrian chronostratigraphic and biostratigraphic divisions and schematic carbon isotope stratigraphy. Shown are the regional biostratigraphic zonal schemes and biomeres correlated with the North American (Laurentian) and international stages of Cambrian. Note that most workers place the Coosella perplexa Subzone within Aphelaspis Zone as shown, but some others place it within the Crepicephalus Zone. The two red horizontal dashed lines mark the two-phases end-Marjuman and end-Steptoean extinction events. Figure modified from (Glumac, 2011; Peng et al., 2004, 2012; Saltzman et al., 1995, 2004; Taylor, 2006).

intervals that were followed by rapid diversification events (Longacre, 1970; Palmer, 1965a, 1965b; Stitt, 1971). These extinctions occur in both the trilobite and brachiopod faunas that make up the majority of

363

the fossil diversity in the later Cambrian. In western North America, the first of these extinctions, the end-Marjuman, was initially identified as the boundary between regional Marjuman and Steptoean Stages (Palmer, 1965a, 1965b). Internationally, the interval of the endMarjuman extinctions corresponds to the boundary between the Guzhangian and Furongian International Stages of the Cambrian (Fig. 1 (Peng et al., 2004)). The ecological patterns of the extinction intervals show an abrupt extinction of a diverse assemblage of shallow marine, polymerid trilobite faunas (those faunas inhabiting the shelf) and were later replaced by a low-diversity assemblage of olenimorph trilobite fauna normally found in more distal, deep-water facies (Palmer, 1984; Stitt, 1971; Westrop and Ludvigsen, 1987). The morphology of these olenid fauna suggests that they were also well-adapted to low-oxygen environments (Fortey, 1985; Fortey and Wilmot, 1991; Jell, 1978) and their gills have been suggested to have housed sulfur oxidizing chemoautotrophic microbes (Fortey, 2000). Brachipods also show a similar pattern of extinction and diversification across these extinction intervals (Freeman et al., 2011; Rieboldt, 2005; Rowell and Brady, 1976). In detail, the end-Marjuman polymerid trilobite extinctions appear to have occurred in two phases. The first was an extinction of the majority of the shallow-water trilobite assemblage (Crepicephalus Zone fauna), but notably, a few of the trilobite genera from this assemblage survive this event. This surviving assemblage — the Coosella perplexa Subzone fauna (Palmer, 1979) — later go extinct during a second extinction event. Following the second extinction of the remaining shallowwater trilobites, the shelf environment was occupied by the olenimorph trilobite fauna (Aphelaspis zone fauna) (Palmer, 1965a, 1965b). The Coosella perplexa subzone, located between these two extinction horizons, has been termed the “critical interval” by some authors (Palmer, 1979; Stitt, 1971; Taylor, 2006). Following the extinction, the diversity of the shallow-water trilobite assemblage gradually increased, only again to be affected by another mass extinction exhibiting the same ecological patterns as the previous (Longacre, 1970; Palmer, 1984; Stitt, 1971). This stratigraphic pattern of extinction and diversification in the late Cambrian trilobite

Panthalassic Ocean North China Australia

South China Precordillera

Gondwana

Siberia

Kazakhstan

Laurentia Iapetus Ocean Avalonia

Baltica Armorica

Fig. 2. Paleogeographic reconstruction of the Late Cambrian Earth modified from Blakey (2003, 2008). Black circles indicate locations where the SPICE has been documented: Southern Laurentia (Cowan et al., 2005; Gill et al., 2011; Glumac, 2011; Glumac and Walker, 1998; Hurtgen et al., 2009; Saltzman et al., 2004), Northern Laurentia (Saltzman et al., 1998, 2004), Precordillera (Sial et al., 2008), Siberia (Kouchinsky et al., 2008), Avalonia (Woods et al., 2011), Armorica (Álvaro et al., 2008), Kazakhstan (Saltzman et al., 2000; Wotte and Strauss, 2015), Baltica (Ahlberg et al., 2008), Australia (Gill et al., 2011; Lindsay et al., 2005; Saltzman et al., 2000), South China (Saltzman et al., 2000), North China (Bagnoli et al., 2014; Chen et al., 2012; Ng et al., 2014; Zhu et al., 2004) and South Korea (Chung et al., 2011; Lim et al., 2015). Red circles and box indicate the approximate locations of this study's outcrop areas. Map from ©Ron Blakey, Colorado Plateau Geosystems, used with permission.

364

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

Fig. 3. Middle to Upper Cambrian stratigraphy and depositional sequences of eastern Tennessee and southwestern Virginia modified from Walker et al. (1990) and Glumac and Walker (2000). Note the interfingering relationship between shale and carbonate units of the Conasauga Group, the distribution of subtidal and peritidal carbonate facies, and the predominance of carbonate deposits in the Uppermost Cambrian (Maynardville Formation and Copper Ridge Dolomite). Placement of 3rd-order sequences or Grand Cycles and sequence boundaries are based on Glumac and Walker (2000) and references cited within. Supersequences and supersequence boundaries are those defined by Read and Repetski (2012). Black box represents the approximate position of the studied stratigraphic sections within the Nolichucky and Maynardville Formations.

communities has been termed a biomere, which is a biostratigraphic unit whose boundaries are defined by the extinction events (Palmer, 1984). For a more thorough review of the stages, trilobite zonations,

and the history of the biomere concept, refer to Taylor (2006). In addition to its recognition on Laurentia, the end-Marjuman extinctions have subsequently been shown to be of global significance and have

Fig. 4. Map of southwestern Virginia and northeastern Tennessee. Outcrop locations for this study are marked as closed squares. Other outcrops of the Conasauga Group where the SPICE has been documented are indicated with open squares (Glumac, 2011; Glumac and Walker, 1998).

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

365

Fig. 5. Example photomicrographs used for petrographic analysis and assessment of diagenetic alteration. All figures are unstained and are under cross-polarized light. a.) Peloidal mudstone from the Washburn Section with little to no sparitization (Sample W-24). b.) Flat pebble conglomerate (intraclastic rudstone) with an intraclastic wackestone matrix from the Three Springs Section. Photomicrograph displays large flat pebble clast (top left) in the intraclastic wackestone matrix (bottom right). Note that the boundary between the flat pebble clast and the wackestone matrix is accentuated by a stylolite. Both the matrix and flat pebble clast contain microspar, though the size of spar crystals is much larger in the intraclast. c.) Peloidal mudstone with microspar from the Dickensonville Section (Sample D-10). d.) Trilobite wackestone from the Washburn Section (Sample W-44). Note the trilobite skeletal fragments have been entirely micritized. e.) Trilobite wackestone with abundant microspar from the Three Spring Section (T-40). f.) Carbonate mudstone that has undergone extensive sparitization from the Three Spring Section (T-25). g.) Trilobite packstone from the Duffield Section (Sample Du-27). Note the abundant microspar. h.) Trilobite packstone from the Dickensonville Section (Sample D-36) that contains abundant microspar and glauconite.

been correlated to extinctions in other sedimentary basins on other paleocontinents (Öpik, 1966; Peng, 1992). The mechanism for the extinction of the shelf communities at the end-Marjuman is still the subject of ongoing debate. Proposed mechanisms include global temperature changes (Lochman-Balk, 1970; Öpik, 1966), a rise in the thermocline, which acted to cool the shelf areas (Stitt, 1975), and/or ecospace changes linked to sea level rise

(Ludvigsen, 1982; Westrop, 1988; Westrop and Ludvigsen, 1987). These mechanisms were initially proposed based on paleontological, sedimentological and stratigraphic data. Geochemical studies have also been utilized to attempt to identify the cause of the extinctions. Orth et al. (1984) searched for an iridium anomaly at the boundary for evidence of an extraterrestrial impact, but did not find one. More recent geochemical studies led to the recognition of the SPICE and have made

366

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

b

5

5

4

4

3

3

δ13C (‰, VPDB)

δ13C (‰, VPDB)

a

2 1 0 -1 -2

2 1 0 -1 -2

-3

-3

-4

-4

-5 -13

-12

-11

-10

-9

-8

-7

-5

-6

0

δ18O (‰, VPDB)

d

5

0.3

0.4

0.5

0.6

5

4

4

3

3

2 1 0 -1 -2

2 1 0 -1 -2 -3

-3 56.8

-4

-4 -5

-5 0

2

4

6

8

10 12

14

16 18

20

Mn/Sr (ppm/ppm)

e

0.2

Mg/Ca (mol/mol)

δ13C (‰, VPDB)

δ13C (‰, VPDB)

c

0.1

0

20

40

60

80 100 120 140 160 180 200

Fe/Sr (ppm/ppm)

5

δ13C (‰, VPDB)

4

Dickensonville

3 2

Three Springs

1 0

Washburn

-1 -2 -3

Duffield

-4 -5

0

5

10

15

20

25

30

35

40

45

Ca (wt%) Fig. 6. Cross-plots of δ13C versus geochemical data used to access the degree of diagenetic alteration. Given that none of these indicators for diagenetic alteration can directly assess the degree of preservation of the original marine δ13C compositions, we conservatively flagged individual samples as having potentially altered δ13C if they exceeded the set thresholds for at least three of these geochemical measures. Refer to the supplementary data table for specific notes for each sample. A.) δ13C versus δ18O. Samples were flagged as potentially altered if they had δ18O values that were more negative than 95% of the δ18O samples from that section (N2σ from the median δ18O); this was done by section because of the section specific δ18O data populations. Threshold values for each section were set at: Washburn = −8.5‰; Three Springs = −12.1‰; Duffield = −8.6‰; Dickensonville −9.6‰ B.) δ13C versus Mg/ Ca. Vertical line denotes the threshold (0.1) where samples above this may contain appreciable amounts of dolomite. C.) δ13C versus Mn/Sr. Vertical line denotes the threshold (8.5). Samples above this line had elevated Mn/Sr greater than 65% of all Mn/Sr values of the entire dataset (N1σ from the median Mn/Sr). D.) δ13C versus Fe/Sr. Vertical line denotes the threshold of 40. Samples above this line had elevated Fe/Sr greater than 65% of all the Fe/Sr values of the entire dataset (N1σ from the median Fe/Sr). F.) δ13C versus Ca contents. Vertical line denotes the threshold of 10 wt% Ca. Samples with relatively low Ca contents (used here as a proxy for carbonate contents) may have been more susceptible to having their initial marine δ13C significantly altered by diagenetic processes.

the case for the development of widespread anoxia (Saltzman et al., 1998) and euxinia (Dahl et al., 2014; Gill et al., 2011) as drivers of the SPICE and the end-Marjuman extinctions. Additionally, Perfetta et al.

(1999) used the observation of a short-lived − 1‰ shift in δ13C after the first of the end-Marjuman extinction horizons to suggest a rise in the thermocline led to invasion of 12C-enriched cold, deep water into

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373 367

Fig. 7. Lithostratigraphy and chemostratigraphy of the studied sections of the Conasauga Group. Red-bordered δ13C data points refer to samples that exceeded our thresholds for at least three of geochemical measures that potentially may indicate significant diagenetic alteration or showed siginificant petrographic evidence of alteration. Red lines mark the extinction events and represent the boundaries between the Crepicephalus Zone and Coosella perplexa subzone and Coosella perplexa subzone and Aphelaspis Zone as identified in each of these sections by Rasetti (1965); Derby (1966), and Cuggy (1996). Lithologic scale: Sh = Shale; M = Mudstone, W = Wackestone, P = Packstone, G = Grainstone, B = Boundstone, F = Floatstone, R = Rudstone. Line on the δ13C plots represents a three point running average through the δ13C data considered least altered.

368

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

the shelf, causing the isotope excursion and extinctions. This scenario may be further supported by oxygen isotope data from inarticulate brachiopods from successions in the Great Basin and midcontinent of North America, which suggest an initial cooling near the interval of the extinctions followed by a warming (Elrick et al., 2011). However, the precise timing of these proposed carbon cycle perturbations and the endMarjuman extinctions have yet to be fully resolved. Here, we investigate the details of the carbon isotope record in the Southern Appalachians to assess the relative timing of the changes in the carbon cycle and the extinctions. 2. Geological Setting The passive margin succession in the Southern Appalachians was part of the Great American Carbonate Bank, which was deposited around Laurentia from the Cambrian into the Early Ordovician and represents one of the largest carbonate-dominated platforms of the Phanerozoic (Fritz et al., 2012). In the Southern Appalachians, the interval that contains the end-Marjuman extinctions occurs within the Nolichucky Formation, which is part of the Conasauga Group (Fig. 3). The Nolichucky Formation (also referred to as the Nolichucky Shale or Conasauga Shale) consists of an up-to-300-m-thick succession of interbedded subtidal carbonate and shale facies (Foreman et al., 1991; Markello and Read, 1981, 1982; Walker et al., 1990; Weber, 1988). The Nolichucky is overlain by the Maynardville Formation, which consists of carbonate dominated, lower subtidal to upper peritidal facies (Glumac and Walker, 2000; Weber, 1988). These units are exposed in multiple thrust sheets within the Valley and Ridge Province of the Appalachian Mountains of eastern North America (Hasson and Haase, 1988; Markello and Read, 1981, 1982; Read and Repetski, 2012). This study examines four outcrops in northeastern Tennessee and southwestern Virginia (Fig. 4). Deposition of the Nolichucky Formation occurred in and around the Rome Trough, a depositional low that had developed on the southern passive margin of Laurentia (Thomas, 1991). The Rome Trough was created with the reactivation of rift faults in the Mississippi Valley-Rough Creek-Rome graben system during the middle to later Cambrian (Hasson and Haase, 1988; Thomas, 1991). The Rome Trough was bounded to the north toward the Laurentian craton by the shallow water carbonate and siliciclastic facies and to the south by a rimmed carbonate platform (represented by the Elbrook Formation) (Hasson and Haase, 1988; Markello and Read, 1981) and can be traced from eastern Tennessee and Kentucky and western Virginia northeastward through Ohio and West Virginia into Pennsylvania (Ammerman and Keller, 1979; Webb, 1980; Weber, 1988). Siliciclastic input into the trough is thought to have been sourced from the exposed craton via a large delta system to the northeast of the basin (now in the subsurface of present day Ohio) that is represented by the Kerbel Formation (Banjade, 2011; Janssens, 1973). The Nolichucky and Maynardville Formations represent carbonate ramp and basinal shale facies deposited to the north of continental margin carbonate platform (Foreman et al., 1991; Glumac and Walker, 2000; Markello and Read, 1982; Weber, 1988). Overall, five thirdorder sequences (also referred to as “Grand Cycles”) are recognized within the Conasauga Group, where ramp carbonates interfinger with the basinal shale facies (Fig. 3, (Glumac and Walker, 2000; Hasson and Haase, 1988; Markello and Read, 1982; Read and Repetski, 2012; Walker et al., 1990)). This cyclicity represents the progradation of the carbonate ramp facies into Rome Trough (Hasson and Haase, 1988; Markello and Read, 1982; Read and Repetski, 2012). Smaller-scale, shallowing-upward cycles — typically shale facies transitioning into carbonate facies on 1–10 m scale (parasequences) — can be identified within the Nolichucky and have been attributed to higher frequency, local sea-level fluctuations (Markello and Read, 1981). On an even finer stratigraphic scale, the interbeds shale and flat-pebble conglomerate (intraclastic, rudstone, and floatstones) beds show weak cyclicity,

however the alternation of these beds could also be storm-dominated shoreface deposits (Markello and Read, 1981, 1982; Myrow et al., 2004). 2.1. Biostratigraphy of the Conasauga Group The biostratigraphy of the Nolichucky and Maynardville Formations is unique in that these formations locally record a thick record of the end-Marjuman extinctions. Within the study area, the Nolichucky contains the two trilobite zones and one subzone of interest for this study: in stratigraphic order, the Crepicephalus Zone, Coosella perplexa subzone, and the Aphelaspis Zone (Cuggy, 1996; Derby, 1966; Rasetti, 1965). The data generated by these previous biostratigraphic studies permit the high-resolution investigation of lithofacies and chemostratigraphic changes across the extinction horizon presented here. 3. Methods 3.1. Field work Four outcrops located in southwestern Virginia and northern Tennessee were chosen for this study based on published biostratigraphic data (Cuggy, 1996; Derby, 1966; Rasetti, 1965) and quality and accessibility of the exposures (Fig. 4). Each section was measured and logged at the decimeter scale. Carbonate samples were collected for petrographic and geochemical analysis at approximately one-meter intervals, unless the quality of the outcrop or particular lithology was not suitable for geochemical analysis. 3.2. Sample preparation Samples were cut with a rock saw with a diamond-tipped blade to remove weathered surfaces giving a fresh surface for detailed lithologic identification and geochemical analysis. A handheld Dremel tool was used to produce a powder targeting calcitic mud (micrite) in each sample for isotopic and elemental analysis. The sampled micrite was made up of either the matrix of fine-grained lithologies (i.e., mudstone, wackestone, and packstones) or clasts composed of micrite (e.g., flat pebbles) found in other coarse grained lithologies (i.e., grainstones and rudstones). 3.3. Petrographic analysis All the cut samples had their lithologies characterized and diagenetic fabrics noted on the hand-sample scale. Alizarin red staining was used to help determine the distribution of calcite and dolomite within samples and to guide the geochemical sampling; dolomite, calcite veins, stylolites, and glauconite-rich areas, if present, were avoided when sampling for isotope and elemental analyses of the carbonates. Additionally, 27 samples of representative lithologies were also selected for the creation of petrographic thin sections. The thin sections were examined for further lithological characterization and the identification of diagenetic fabrics. 3.4. Geochemical analyses 3.4.1. Carbonate carbon and oxygen isotope analysis Approximately 0.350 mg of carbonate powder from targeted fabrics in each sample was reacted with 100% ortho-phosphoric acid at 70 °C for 4 to 8 h. Carbon and oxygen isotope data were obtained by measuring the CO2 gas evolved upon acidification of the sample. Samples were measured on an Isoprime 100 isotope-ratio mass spectrometer (IRMS) coupled with a peripheral MultiFlow-Geo headspace sampler in the Stable Isotope Facility in the Department of Geosciences at Virginia Polytechnic Institute and State University. Carbon and oxygen isotope compositions are reported in standard delta notation as per mil (‰) deviations from Vienna Pee Dee Belemnite (VPDB). Eighty percent of

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

samples were replicated at least once and were equal to or better than 0.1‰ for δ13Ccarb (1σ), and ±0.2‰ for δ18Ocarb (1σ). Repeated measurements of the IAEA-CO-1, CO-9 and NBS-18 standards were ±0.04‰ for δ13Ccarb, and δ18Ocarb ± 0.2‰ (1σ). 3.4.2. Elemental analysis Elemental analysis of Ca, Mg, Mn, Fe, Sr, and Al in the carbonate portion of the sample was carried out on an inductively coupled plasma atomic emission spectrometer (ICP-AES) housed in the Department of Crop and Soil Environmental Sciences at Virginia Polytechnic Institute and State University. For this analysis, 100 to 300 mg of powder were dissolved in 0.5 M trace metal-grade acetic acid solution for 6 h to target the carbonate fraction of the sample and to avoid the breakdown of pyrite or other non-carbonate minerals. The samples were then centrifuged, filtered, and diluted in 0.5 M acetic acid before analysis. The reproducibility for all elements was 7% or better. 4. Results 4.1. Petrographic Analysis Of the 27 samples selected for petrographic examination, eight are displayed in Fig. 5 because they represent the most common fabrics in the studied samples. Fig. 5a, b, c, and d illustrate ideal lithologies for δ13Ccarb sampling — carbonate muds with low porosity, which would have limited post-depositional fluid flow. Generally, carbonate mudstones and/or flat pebble clast that are composed of carbonate mudstone show coherent trends and reproducible δ 13Ccarb values. Fig. 5d, e, f, and g illustrate lithologies that contain diagenetic features and may potentially display altered primary δ13Ccarb signals. Carbonate mud can be prone to dissolution and reprecipitation as calcite microspar (a common diagenetic fabric usually four to 30 μm in diameter) through a process called sparitization. Samples with extensive sparitization often had δ 13Ccarb values much more negative than the rest of the population of the data. A number of the samples with larger clasts (packstones and grainstones) also showed signs of sparitization and/ or the addition of post-depositional carbonate cements. 4.2. Interpretation of the preservation potential of original marine carbon isotope signatures A key consideration in any study using carbonates as proxies for ancient seawater is whether the geochemical data are representative of ancient ocean chemistry. No single geochemical or textural parameter has been shown to be a perfect indicator for the degree of preservation of the primary geochemical signatures in ancient carbonates. This is not surprising given the spectrum of depositional and post-depositional histories that carbonate successions experience. Commonly applied geochemical indicators for diagenetic alteration include low δ18Ocarb and high Mn/Sr, Fe/Sr and Mg/Ca (Jacobsen and Kaufman, 1999; Lohmann, 1988; Marshall, 1992; Metzger and Fike, 2013). Linear or asymptotic trends can indicate mixing between primary marine and diagenetic end-members, denoting partial resetting of the carbonate chemistry (Banner and Hanson, 1990; Jacobsen and Kaufman, 1999; Lohmann, 1988). Additionally, samples with values that lie outside the main population of data can also indicate alteration confined to specific lithologic units within a stratigraphic succession. Cross-plots of the δ13Ccarb data with δ18Ocarb, Mn/Sr, Fe/Sr, Mg/Ca and Ca contents can be found in Fig. 6. None of the studied sections show a significant correlation between δ13Ccarb and δ18Ocarb (R2 for Dickensonville: 0.55, Three Springs: 0.44, Washburn: 0.09 and Duffield 0.004) and between δ13Ccarb and the other proposed geochemical proxies for alteration (Mn/Sr, Fe/Sr and Mg/Ca: Fig. 6). Further evidence supporting the primary nature of the majority of the δ13Ccarb data is the consistency in stratigraphy trends relative to the trilobite biostratigraphy in the four studied sections (discussed in further detail below). All the sections show the broad

369

trend of the rising portion of the SPICE excursion seen elsewhere, which initiates at the second extinction horizon at the top of the Coosella perplexa subzone (Gill et al., 2011; Glumac, 2011; Glumac and Walker, 1998; Kouchinsky et al., 2008; Saltzman et al., 1998, 2000, 2004). However, single samples do display δ13Ccarb that deviate from these stratigraphic trends. In all cases, these samples have more 12C-enriched isotope compositions. For all of these samples, either the elemental data δ18Ocarb and/or petrographic evidence suggests that they had undergone some degree of diagenetic alteration, and therefore, their δ13Ccarb is likely not representative of original seawater chemistry. We have conservatively screened the samples using the petrographic and geochemical data to indicate those that are potentially least altered and best reflect primary marine δ13C and those that potentially do not. Samples that either 1) contained significant amount of diagenetic fabrics (extensive sparitization or cements) or 2) have at least three geochemical parameters that suggest diagenetic alteration (elevated Mn/Sr, Fe/Sr, Mg/ Ca, low Ca contents or δ18O) are indicated on Figs. 7, S2, S3, S5, and S5. It is noteworthy that while this screening process did identify all the samples that had 12C-enriched compositions that deviate from the stratigraphic trends, it also included many samples that fit well within the stratigraphic δ13C trends. With this in mind, all the δ13C data points are shown on the stratigraphic figures. However, trends consisting only of the least altered samples — those whose chemistry passed the screening criterion — are considered to best represent the evolution of the marine carbon cycle. 4.3. Stratigraphic Trends in Geochemistry and Lithofacies 4.3.1. Washburn Section (36.30818°N, 83.59893°W) The first 14 m of the Nolichucky Formation exposed at Washburn contains the Crepicephalus and Coosella perplexa biozones (Cuggy, 1996; Rasetti, 1965). The dominant lithofacies in this part of the section consist primarily of flat pebble conglomerates and carbonate mudstones interbedded with gray shales (Figs. 7 and S2). Up section, within the Aphelaspis Zone, the dominant carbonate lithofacies change to trilobite wackestones, packstones and grainstones with subordinate carbonate mudstones and flat-pebble conglomerates. All these lithofacies are interbedded with gray shales. The beds of the shale facies in the Aphelaspis Zone also become thicker as compared with those in the lower portion of the section. The section is capped by the “ribbon rock” carbonate facies —nodular trilobite wackestones, packstones, and grainstones beds that are internally fine upward — of the Maynardville Formation. δ13Ccarb values throughout the Crepicephalus and Coosella perplexa biozones at Washburn stay between − 0.3‰ to 0.8‰ and show no change across the first extinction horizon of the end-Marjuman extinctions. Directly above the second extinction boundary of the end-Marjuman extinctions — the contact between the Coosella perplexa Subzone and base of the rest of the Aphelaspis Zone — δ13Ccarb steadily increases to approximately + 4‰ up section (Figs. 7 and S2). 4.3.2. Three Springs Section The Three Springs Section consists entirely of the Nolichucky Formation with top of the section terminated by a fault. Lithofacies in the lowermost 24 m of the Three Springs Section, which contains the Crepicephalus Zone and Coosella perplexa Subzone, consist of flat pebble conglomerates, carbonate mudstones, and minor trilobite grainstones all interbedded with gray and green shales (Figs. 7 and S3). In this portion of the section, δ13C values remain tightly grouped between −0.5‰ and +1‰ and show no significant change across the first phase of the end-Marjuman extinctions (Figs. 7 and S3). Carbonate lithofacies in the following Aphelaspis Zone (24 to 47 m) are highly variable and consist of carbonate mudstones, flat pebble conglomerates, ooid and skeletal grainstones, and stromatolitic and thrombolitic boundstones. All these carbonate lithofacies are interbedded with gray and green shales on the decimeter to meter scale. Above the second extinction boundary

370

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

at the base of the Aphelaspis Zone δ13Ccarb show overall an increase from approximately 0‰ to +3‰ over the last 19 m of section.

Maynardville and steadily increase to +3.8‰ in the uppermost 9 m of the section.

4.3.3. Duffield section The Duffield section represents the most temporally limited of the studied sections as it only contains portions of the Coosella perplexa Subzone and Aphelaspis Zone (Cuggy, 1996). The first 8 m of the Nolichucky Formation at this location represent the upper portion of Coosella perplexa Subzone and consists of flat-pebble conglomerates and subordinate carbonate mudstones interbedded with grey shales (Figs. 7 and S4). δ13Ccarb in this portion of the section remains between − 0.1‰ and + 1.1‰. Carbonate lithofacies within the overlying Aphelaspis Zone are more variable and consist of mudstones, trilobite wackestones, packstones and grainstones, and interbedded with grey shales. These shale interbeds decrease in thickness and abundance up section. At 23.5 m, the “ribbon rock” facies that marks the basal unit of the Maynardville Formation occurs, and this lithofacies eventually becomes the dominant lithofacies up section. At the Coosella perplexa/Aphelaspis boundary, which marks the second phase of the end-Marjuman extinctions, δ13Ccarb values start at ~ + 1‰ and rise to +3.5‰ over the overlying 26 m of the Aphelaspis Zone.

5. Discussion

4.3.4. Dickensonville section The Dickensonville Section is the most temporally expanded of the investigated sections and contains thick intervals of the Crepicephalus Zone, Coosella perplexa Subzone and Aphelaspis Zones (Cuggy, 1996; Derby, 1966). The Crepicephalus Zone lies within the first 22 m of the Nolichucky exposed at this section followed by approximately 22 m of Coosella perplexa Subzone. Approximately 4 m of section between the top of the Coosella perplexa Subzone and the rest of the Aphelaspis Zone lack biostratigraphic control; this uncertainly could potentially lower the top of the Coosella perplexa Subzone (and second endMarjuman extinction horizon) to the 41-m mark, instead of its current position at the 45-m mark at the first datum appearance of Aphelaspis walcotti (Cuggy, 1996). Lithofacies in Nolichucky in the Crepicephalus Zone vary considerably but show a clear stratigraphic trend (Figs. 7 and S5). The dominant carbonate lithofacies in the basal 7.5 m of the section are flat pebble conglomerates and minor carbonate mudstones interbedded with grey shales. These lithofacies gradually transition upward into carbonate mudstones with minor flat pebble conglomerates and isolated bioherms of thrombolitic boundstones (7.5 to 16 m). All these carbonate lithofacies are interbedded with grey shales, and overall these shale beds thin up section and eventually disappear in uppermost the Crepicephalus Zone. Continuing up section (16 to 19 m), thrombolitic boundstones become dominant, and the final 4 m of the Crepicephalus Zone (19 to 23 m) consists of massive ooid grainstones. In the overlying Coosella perplexa subzone (23 to 45 m), shale lithofacies dominate with subordinate carbonate interbeds of flat pebble conglomerates, carbonate mudstones, and trilobite wackestones and packstones. Glauconite and pyrite is abundant throughout this portion of the section in both the carbonate and shale lithofacies. This part of the section also contains sedimentary dikes composed of the carbonate lithofacies cutting the shales (these were not sampled for geochemistry). Carbon isotope values remain tightly grouped between − 0.5‰ and + 1.1‰ in the first 47 m of the section that contain the Crepicephalus Zone and Coosella perplexa Subzone (Figs. 7 and S5) and show no significant variations or directionality across the first phase of the end-Marjuman extinctions. Lithofacies in the Nolichucky in the basal Aphelaspis Zone are similar to those in the underlying Coosella perplexa subzone. However, at 60 m, the “ribbon rock” facies occurs with the loss of the shale facies, indicating the basal contact of the overlying Maynardville Formation. Carbon isotope values in the lowest most Aphelaspis Zone (46 to 47.6 m) start at approximately + 1‰. However, the next 12 m lack carbon isotope data due to a deeply weathered and poorly exposed section of the outcrop. After this sample gap, δ13Ccarb values start at +2.5‰ in the basal

5.1. Local lithofacies trends across the end-Marjuman extinctions Here, we interpret the lithofacies trends with the framework of facies model for Nolichucky and Maynardville carbonate ramp from Markello and Read (1981, 1982). Previous studies of the relative sea level history on Laurentia during the Upper Cambrian suggest a shoaling event occurred during the Crepicephalus Zone, which was followed by an abrupt deepening in the Aphelaspis Zone (Osleger and Read, 1993; Read, 1989). Higher-resolution stratigraphic analyses examining the lithofacies trends in the Coosella perplexa subzone have not been attempted, likely due to the limited stratigraphic thickness of this subzone in Laurentian successions. The sections of the Conasauga Group examined here show a variety of local facies changes over the studied interval. This is not surprising given the likely variance in local bathymetry and subsidence at the time of deposition (Hasson and Haase, 1988). Evidence for shoaling followed local deepening across the extinction interval is clearly displayed in the section at Dickensonville. The facies progression through the Crepicephalus Zone displays a clear shallowing trend: flatpebble conglomerate and shale facies in the lower part of the section give way to thrombolite bioherm and ooid grainstone facies in the latest Crepicephalus Zone. This followed up section by the abrupt appearance of shale facies that contains abundant glauconite and pyrite at the base the Coosella perplexa Zone (Figs. 7 and S5). On the other end of the spectrum at Three Springs, the local facies progression suggests progressive shoaling across the interval of the extinctions. Shallow carbonate facies (ooid grainstone and microbial boundstone lithofacies, etc.) appear in the Coosella perplexa subzone and remain prevalent through out this interval and into the succeeding Aphelaspis Zone. Potentially, this section may have been more proximal to the shallow carbonate platform at the time of deposition and carbonate production may have been able to keep up with accommodation. Lithofacies at Washburn and Duffield show consistent outer ramp lithofacies (subtidal flat pebble conglomerate, carbonate mudstone and shale lithofacies) over the extinction interval and thus do not show detectable changes in sea level. While it is not possible to confirm the proposed relationship between sea level history and the end-Marjuman extinctions given the limited stratigraphic sections investigated here, there are two important conclusions for this study that can be drawn from the lithofacies data. One is there appears to not be a strong lithofacies control on the local appearance and stratigraphic range of the trilobite fauna (i.e. Westrop and Ludvigsen, 1987). Second, despite the range of lithofacies that occur in between biozones and sections, the δ13C records show a consistent trend relative to the trilobite zonations; all the sections show an increase in δ13C which starts at the second extinction horizon at the top of the Coosella perplexa subzone. This indicates that there does not appear to be a strong lithofacies control on the δ13C records, and thus they reflect secular trends in seawater chemistry. 5.2. Relationships between the carbon isotope stratigraphy and the endMarjuman Extinctions In all the studied sections, the carbon isotope trends show a remarkable consistency in relation to the biostratigraphy despite the significant differences in the local lithofacies (Fig. 7). The carbon isotope data show no significant negative or positive directional changes throughout the Crepicephalus Zone varying between −0.5‰ and +1‰ (Fig. 7). Additionally, the Coosella perplexa Subzone in each stratigraphic section is also characterized by low variation in δ13C, varying between 0‰ and + 1‰. However, coinciding with the base of the Aphelaspis Zone and second end Marjuman extinction, δ13C values begin to gradually rise

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

in the positive direction reaching +3 to +4‰ over a few tens of meters in all the sections. The magnitude and biostratigraphic position of this positive shift indicates that this is the rising limb of the SPICE. The carbon isotope trends observed the studied sections of the Conasauga Group are significant as they bear on the proposed mechanism for the extinctions. First, none of the studied sections shows evidence for a short-lived negative excursion, as observed in the datasets from the Deadwood Formation of South Dakota (Perfetta et al., 1999). The Three Springs Section (Figs. 7 and S3) does have a single data point that would represent a − 1‰ δ13Ccarb excursion directly after the second phase of the end-Marjuman extinctions and immediately following the Coosella perplexa Subzone. However, petrographic observations from that sample revealed that it has undergone pervasive sparitization and has elevated Mn/Sr and Fe/Sr ratios (Fig. 6). This information — the fact that the shift is constrained by a single data point and that none of the other sections display a similar trend — supports the interpretation that this sample's δ13Ccarb is likely altered. The lack of a negative δ13C excursion in the Nolichucky data suggests that the negative carbon isotope excursion observed in the Black Hills was not a global trend. This is further supported by the lack of a negative excursion during or after the end-Marjuman extinctions in other stratigraphic successions in Laurentia and (Saltzman et al., 1998, 2004) and elsewhere worldwide (Kouchinsky et al., 2008; Saltzman et al., 2000). One potential explanation for the Black Hills data involves the fact that the negative carbon isotope values occur directly below an unconformity (Perfetta et al., 1999). This suggests that meteoric fluids may have altered the original carbonate chemistry during exposure of the carbonates below this horizon. Alternatively, the negative δ13C shift may record a local oceanographic signal representing a perturbation to the regional carbon cycle (i.e. Panchuk et al., 2006). If this is the case, then the mechanism for its generation (i.e. upwelling of cool, 12 C-enriched waters) cannot be invoked as the sole cause for the endMarjuman extinctions across Laurentia and elsewhere. It is important to point out that the first phase of end-Marjuman extinctions, based on the δ13Ccarb data from the Conasauga Group, is not concurrent with the initiation of the SPICE. The absence of a perturbation in δ13Ccarb values during the first extinction marking the Crepicephalus and Coosella perplexa transition is intriguing. This appears to indicate that the carbon cycle perturbation represented by the SPICE and its underlying environmental drivers (e.g. enhanced organic carbon burial and anoxia) did not initiate at the first phase of the endMarjuman extinctions. Therefore, these processes do not appear to be responsible for this event. Alternatively, if the dissolved inorganic carbon (DIC) reservoir was sufficiently large and/or the initial change in the carbon cycle occurred slowly, then the change in carbon cycle would not, at least initially, have been reflected in the carbon isotope composition of DIC in the oceans. This potentially could have resulted in an apparent stratigraphic offset between the first extinction event and the initiation of the SPICE excursion. To explore this idea, an estimate of the minimum and maximum duration of the “critical interval” represented by the Coosella perplexa subzone is needed to determine the time between the first phase of the end Marjuman extinctions and the observed beginning of the SPICE. Here, we utilize the thickness of the Coosella perplexa subzone at the Dickensonville section (24 m) to derive an estimate for the duration of the subzone. We chose this section because the Coosella perplexa subzone is thickest there and thus records the most complete record. Time estimates for the subzone can be calculated based on depositional rates derived by the Osleger and Read (1993), who estimated sediment accumulation rates for the Nolichucky Formation of 1.6 to 3.9 cm/kyr based on the observed thicknesses of and estimates for the duration of the Upper Cambrian trilobite zones contained within the formation. Multiplying the thickness of the Coosella perplexa subzone at Dickensonville by the range of estimates for the depositional rates gives yields estimates for the duration of subzone of 38.4 to 93.6 kyrs. Because these estimates are several times greater than residence time

371

of DIC in the modern ocean (~105 yrs), they suggest that if the mechanism that caused the SPICE coincided with the first of the endMarjuman extinctions, the initiation of SPICE (the rise in δ13C) should be observed in the Coosella perplexa subzone. Because it does not, we therefore advocate that the enhanced organic carbon burial associated with the SPICE did not occur with the first phase of the end-Marjuman extinctions. This suggests that the proposed driver for the SPICE (i.e., the expansion of marine anoxia) may potentially be excluded as a driver of the first phase of the end-Marjuman extinctions. However, we note that the application of other geochemical proxies that more directly test for anoxic deposition may better delineate the role of anoxia, if any, in the first phase of the end-Marjuman extinctions. The initiation of the rising limb of the SPICE in the studied sections of the Conasauga Group occurs at end of the Coosella perplexa subzone, which marks the second phase of end-Marjuman extinctions. Therefore, this data set links, in stratigraphy and time, the initiation of the SPICE with the second and final phase of the end-Marjuman extinctions. Further, the overlying Aphelaspis Zone is characterized by the invasion of the low-diversity assemblage of olenimorph trilobites, which were potentially low-oxygen tolerant, onto the shelf areas. Based on their stratigraphic and temporal relationship, these events — the SPICE, second end-Marjuman extinction, and following proliferation of olenimorph trilobites — were conceivably driven by the same or related environmental mechanisms. Other geochemical data, including sulfur and uranium isotopes and molybdenum concentrations, from other sedimentary successions suggest that the SPICE was generated by elevated global rates of organic matter burial (Saltzman et al., 2000) due to the development and spread of anoxic conditions within the oceans (Dahl et al., 2014; Gill et al., 2011). The onset of these environmental changes would have initiated with the start of the SPICE (Dahl et al., 2014; Gill et al., 2011). This scenario is consistent with the relationship between the carbon isotope and biostratigraphic trends observed in the Southern Appalachians: the SPICE initiates at the second end-Marjuman extinction, which was followed by the local invasion of olenimorph trilobites into shallow water environments. Indeed, iron speciation data from subsurface drill cores of the Conasauga Group from Ohio and Kentucky suggest that anoxia locally developed in the Rome Trough with the onset SPICE (LeRoy and Gill, 2015). Further application of redoxsensitive paleoproxies, such as iodine contents of carbonates (Lui et al., 2010) or redox sensitive transition metals to the sections of the Conasauga Group in the Southern Appalachians and elsewhere will further illuminate the relationship between changes in redox state of the ocean and the end-Marjuman extinctions.

6. Conclusions This study is the first high-resolution study of the stratigraphic relationship between the end-Marjuman extinctions and changes in the carbon cycle reflected in the δ13C stratigraphy of the Conasauga Group of the Southern Appalachians. It reveals a lack of a negative δ13C excursion during the end-Marjuman extinctions and suggests that the δ13C shifts seen in the Black Hills are either the product of diagenesis or are a local environmental signal. The absence of this δ13C signal in the Southern Appalachians also suggests that scenarios for its creation cannot be invoked as the sole or universal extinction mechanism for the end-Marjuman extinctions. This study also confirms that the onset of the SPICE does not coincided with first phase of the end-Marjuman extinctions and, thus, the cause of the first phase of the end-Marjuman extinctions remains enigmatic. However, the SPICE does initiate with the second phase of the extinctions. Given the abundant geochemical evidence for the spread of anoxia coinciding with the initiation of the SPICE makes this a likely candidate for a driver of the second phase of the end-Marjuman extinctions and ecological patterns that followed it. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.palaeo.2016.08.031.

372

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373

Acknowledgments We wish to thank editorial guidance of Isabel Montanez and to anonymous reviewers who comments greatly improved the manuscript. We also like to thank Michael Cuggy and Fred Read for advice on selecting field sites, Teddy Them for assistance in the field, and Athena Tilley of the Virginia Tech Soil Testing Laboratory for assistance in ICP analyses. Finally, this work was partially supported by student grants from the Virginia Tech Department of Geosciences. References Ahlberg, P., Axheimer, N., Babcock, L.E., Eriksson, M.E., Schmitz, B., Terfelt, F., 2008. Cambrian high-resolution biostratigraphy and carbon isotope chemostratigraphy in Scania, Sweden: first record of the SPICE and DICE excursions in Scandinavia. Lethaia 42, 2–16. Álvaro, J.J., Bauluz, B., Subías, I., Pierre, C., Vizcaïno, D., 2008. Carbon chemostratigraphy of the Cambrian-Ordovician transition in a midlatitude mixed platform, Montagne Noire, France. Geol. Soc. Am. Bull. 120, 962–975. Ammerman, M.L., Keller, G.R., 1979. Delineation of Rome Trough in Eastern Kentucky by gravity and deep drilling data. AAPG Bull. 63, 341–353. Bagnoli, G., Qi, Y.P., Zuo, J.X., Du, S.X., Liu, S.C., Zhang, Z.Q., 2014. Integrated biostratigraphy and carbon isotopes from the Cambrian Tangwangzhai section, North China. Palaeoworld 23, 112–124. Banjade, B., 2011. Subsurface facies analysis of the Cambrian Conasauga Formation and Kerbel Formation in east-central Ohio. Bowling Green State University, p. 167. Banner, J.L., Hanson, G.N., 1990. Calculation of simultaneous isotopic and trace element variations during water-rock interaction with applications to carbonate diagenesis. Geochim. Cosmochim. Acta 54, 3123–3137. Blakey, R.C., 2003. Carboniferous–Permian paleogeography of the assembly of Pangaea. In: Wong, T.E. (Ed.), Proceedings of the XVth International Congress on Carboniferous and Permian Stratigraphy. Royal Netherlands Academy of Arts and Sciences, Utrecht, the Netherlands, pp. 443–456. Blakey, R.C., 2008. Gondwana paleogeography from assembly to breakup—A 500 m.y. odyssey. In: Fielding, C.R., Frank, T.D., Isbell, J.L. (Eds.), Resolving the late Paleozoic ice age in time and space, pp. 1–28. Brasier, M.D., 1993. Towards a carbon isotope stratigraphy of the Cambrian System: potential of the Great Basin succession. In: Hailwood, E.A., Kidd, R.B. (Eds.), High Resolution Stratigraphy, pp. 341–350. Chen, J., Chough, S.K., Lee, J.-H., Han, Z., 2012. Sequence-stratigraphic comparison of the upper Cambrian Series 3 to Furongian succession between the Shandong region, China and the Taebaek area, Korea: High variability of bounding surfaces in an epeiric platform. Geosci. J. 16, 357–379. Chung, G.S., Lee, J.G., Lee, K.S., 2011. Stable carbon isotope stratigraphy of the Cambrian Machari Formation in the Yeongweol Area, Gangweon Province, Korea. J. Korean Earth Sci. Soc. 32, 437–452. Cowan, C.A., Fox, D.L., Runkel, A.C., Saltzman, M.R., 2005. Terrestrial-marine carbon cycle coupling in ~ 500-my-old phosphatic brachiopods. Geology 33, 661–664. Cuggy, M.B., 1996. Patterns of faunal change at an upper Cambrian trilobite extinction event, Nolichucky Formation, Tennessee and Virginia (Masters Thesis) Brock University, pp. 1–142. Dahl, T.W., Boyle, R.A., Canfield, D.E., Connelly, J.N., Gill, B.C., Lenton, T.M., Bizzarro, M., 2014. Uranium isotopes distinguish two geochemically distinct stages during the later Cambrian SPICE event. Earth Planet. Sci. Lett. 401, 313–326. Derby, J.R., 1966. Paleontology and stratigraphy of the Nolichucky Formation in southwest Virginia and northeast Tennessee. Virginia Polytechnic Institute, pp. 1–593. Elrick, M.B., Rieboldt, S.E., Saltzman, M.R., McKay, R.M., 2011. Oxygen-isotope trends and seawater temperature changes across the Late Cambrian Steptoean positive carbonisotope excursion (SPICE event). Geology 39, 987–990. Foreman, J.L., Walker, K.R., Weber, L.J., Driese, S.G., Dreier, R.B., 1991. Slope and Basinal Carbonate Deposition in the Nolichucky Shale (Upper Cambrian), East Tennessee. Effect of Carbonate Suppression by Siliciclastic Deposition on Basin-Margin Morphology, Mixed Carbonate-Siliciclastic Sequences. SEPM, pp. 511–539. Fortey, R.A., 1985. Pelagic trilobites as an example of deducing the life habits of extinct arthropods. Trans. R. Soc. Edinb. Earth Environ. Sci. 76, 219–230. Fortey, R., 2000. Olenid trilobites: The oldest known chemoautotrophic symbionts? Proc. Natl. Acad. Sci. U. S. A. 97, 6574–6578. Fortey, R.A., Wilmot, N.V., 1991. Trilobite cuticle thickness in relation to palaeoenvironment. Paläontol. Z. 65, 141–151. Freeman, R., Miller, J.F., Holmer, L.E., 2011. Linguate brachiopod extinction and global migration coinciding with three Laurentian trilobite extinction events during the late Cambrian-earliest Ordovician. GSA Abstracts with Programs 43, p. 543. Fritz, R.D., Morgan, W.A., Longacre, S., Derby, J.R., Sternbach, C.A., 2012. Introduction. In: Derby, J.R., Fritz, R.D., Longacre, S.A., Morgan, W.A., Sternbach, C.A. (Eds.), The great American carbonate bank: The geology and economic resources of the Cambrian – Ordovician Sauk megasequence of Laurentia, pp. 1–3. Gill, B.C., Lyons, T.W., Young, S.A., Kump, L.R., Knoll, A.H., Saltzman, M.R., 2011. Geochemical evidence for widespread euxinia in the Later Cambrian ocean. Nature 469, 80–83. Glumac, B., 2011. High-resolution stratigraphy and correlation of Cambrian strata using carbon isotopes: an example from the southern Appalachians, USA. Carbonates Evaporites 26, 287–297.

Glumac, B., Walker, K.R., 1998. A Late Cambrian positive carbon-isotope excursion in the Southern Appalachians; relation to biostratigraphy, sequence stratigraphy, environments of deposition, and diagenesis. J. Sediment. Res. 68, 1212–1222. Glumac, B., Walker, K.R., 2000. Carbonate deposition and sequence stratigraphy of the terminal Cambrian Grand Cycle the southern Appalachians, U.S.A. J. Sediment. Res. 70, 952–963. Hasson, K.O., Haase, C.S., 1988. Lithofacies and paleogeography of the Conasauga Group (Middle and Late Cambrian) in the Valley and Ridge province of east Tennessee. Geol. Soc. Am. Bull. 100, 234–246. Hurtgen, M.T., Pruss, S.B., Knoll, A.H., 2009. Evaluating the relationship between the carbon and sulfur cycles in the later Cambrian ocean: An example from the Port au Port Group, western Newfoundland, Canada. Earth Planet. Sci. Lett. 281, 288–297. Jacobsen, S.B., Kaufman, A.J., 1999. The Sr, C and O isotopic evolution of Neoproterozoic seawater. Chem. Geol. 161, 37–57. Janssens, A., 1973. Stratigraphy of the Cambrian and Lower Ordovician rocks in Ohio. Ohio Div. Geol. Surv. Bull. 64, 1–197. Jell, P.A., 1978. Trilobite respiration and genal caeca. Alcheringa 2, 251–260. Kouchinsky, A., Bengtson, S., Gallet, Y., Korovnikov, I.V., Pavlov, V., Runnegar, B., Shields-Zhou, G.A., Veizer, J., Young, E., Ziegler, K., 2008. The SPICE carbon isotope excursion in Siberia: a combined study of the upper Middle Cambrian–lowermost Ordovician Kulyumbe River section, northwestern Siberian Platform. Geol. Mag. 145, 14. LeRoy, M., Gill, B.C., 2015. Assessing the role of anoxia in the end-Marjuman extinction and SPICE events. Geol. Soc. Am. Abstr. Programs 47, 572. Lim, J.N., Chung, G.S., Park, T.-Y.S., Lee, K.S., 2015. Lithofacies and Stable Carbon Isotope Stratigraphy of the Cambrian Sesong Formation in the Taebaeksan Basin, Korea. J. Korean Earth Sci. Soc. 36, 617–631. Lindsay, J.F., Kruse, P.D., Green, O.R., Hawkins, E., Brasier, M.D., Cartlidge, J., Corfield, R.M., 2005. The Neoproterozoic–Cambrian record in Australia: A stable isotope study. Precambrian Res. 143, 113–133. Lochman-Balk, C., 1970. Upper Cambrian Faunal Patterns on the Craton. Geol. Soc. Am. Bull. 81, 3197. Lohmann, K.C., 1988. Geochemical patterns of meteoric diagenetic systems and their application to studies of paleokarst. In: James, N.P., Choquette, P.W. (Eds.), Paleokarst. Springer-Verlag, New York, pp. 58–80. Longacre, S.A., 1970. Trilobites of the Upper Cambrian Ptychaspid Biomere Wilberns Formation, Central Texas. Paleontol. Soc. Mem. (i-vi-1-70). Ludvigsen, R., 1982. Upper Cambrian and Lower Ordovician trilobite biostratigraphy of the Rabbitkettle Formation, Western District of Mackenzie. Royal Ontario Museum, Toronto. Lui, Z., Jenkyns, H.C., Rickaby, R.E.M., 2010. Iodine to calcium ratios in marine carbonate as a paleo-redox proxy during oceanic anoxic events. Geology 38, 1107–1110. Markello, J.R., Read, J.F., 1981. Carbonate ramp-to-deeper shale shelf transitions of an Upper Cambrian intrashelf basin, Nolichucky Formation, Southwest Virginia Appalachians. Sedimentology 28, 573–597. Markello, J.R., Read, J.F., 1982. Upper Cambrian Intrashelf Basin, Nolichucky Formation, Southwest Virginia Appalachians. AAPG Bull. 66, 860–878. Marshall, J.D., 1992. Climatic and oceanographic isotopic signals from the carbonate rock record and their preservation. Geol. Mag. 129, 143–160. Metzger, J.G., Fike, D.A., 2013. Techniques for assessing spatial heterogeneity of carbonate δ13C values: Implications for craton-wide isotope gradients. Sedimentology 60, 1405–1431. Montañez, I.P., Osleger, D.A., Banner, J.L., Mack, L.E., 2000. Evolution of the Sr and C isotope composition of Cambrian Oceans. GSA Today 10, 1–7. Myrow, P.M., Tice, L., Archuleta, B., Clark, B., Taylor, J.F., Ripperdan, R.L., 2004. Flat-pebble conglomerate: its multiple origins and relationship to metre-scale depositional cycles. Sedimentology 51, 973–996. Ng, T.-W., Yuan, J.-L., Lin, J.-P., 2014. The North China Steptoean Positive Carbon Isotope Event: New insights towards understanding a global phenomenon. Geobios 1–17. Öpik, A.A., 1966. The Early Upper Cambrian crisis and its correlation. J. Proc. R. Soc. NSW 100, 9–14. Orth, C.J., Knight, J., Quintana, L., Gilmore, J., Palmer, A.R., 1984. A search for iridium abundance anomalies at two Late Cambrian biomere boundaries in Western Utah. Science 223, 163–165. Osleger, D.A., Read, J.F., 1993. Comparative analysis of methods used to define eustatic variations in outcrop: Late Cambrian interbasinal sequence development. Am. J. Sci. 293, 157–216. Palmer, A.R., 1965a. Biomere: A new kind of biostratigraphic unit. J. Paleontol. 39, 149–153. Palmer, A.R., 1965b. Trilobites of the Late Cambrian Pterocephaliid Biomere in the Great Basin, United States. Geol. Surv. Prof. Pap. 493, 1–153. Palmer, A.R., 1979. Biomere boundaries re-examined. Alcheringa 3, 33–41. Palmer, A.R., 1984. The Biomere problem: Evolution of an idea. J. Paleontol. 58, 599–611. Panchuk, K.M., Holmden, C., Leslie, S.A., 2006. Local controls on carbon cycling in the Ordovician Midcontinent Region of North America, with implications for carbon isotope secular curves. J. Sediment. Res. 76, 200–211. Peng, S., 1992. Upper Cambrian biostratigraphy and trilobite faunas of the Cili-Taoyuan area, northwestern Hunan, China. Peng, S., Babcock, L.E., Reichart, R.A., Lin, H., Rees, M.N., Saltzman, M.R., 2004. Global Standard Stratotype-section and Point of the Furongian Series and Paibian Stage Cambrian. Lethaia 37, 365–379. Peng, S., Babcock, L.E., Zuo, J., Zhu, X., Lin, H., Yang, X., Qi, Y., Bagnoli, G., Wang, L., 2012. Global Standard Stratotype-Section and Point (GSSP) for the Base of the Jiangshanian Stage (Cambrian: Furongian) at Duibian, Jiangshan, Zhejiang, Southeast China. Episodes 35, 462–477.

A.M. Gerhardt, B.C. Gill / Palaeogeography, Palaeoclimatology, Palaeoecology 461 (2016) 362–373 Perfetta, P.J., Shelton, K.L., Stitt, J.H., 1999. Carbon isotope evidence for deep-water invasion at the Marjumiid-Pterocephaliid biomere boundary, Black Hills, USA: A common origin for biotic crises on Late Cambrian shelves. Geology 27, 403–406. Rasetti, F., 1965. Upper Cambrian trilobite faunas of northeastern Tennessee. Read, J.F., 1989. Controls on evolution of Cambrian-Ordovician passive margin, U.S. Appalachians. In: Crevello, P., Wilson, J.L., Sarg, J.F., Read, J.F. (Eds.), Controls on Carbonate Platform and Basin Development, pp. 147–166. Read, J.F., Repetski, J.E., 2012. Cambrian–Lower Middle Ordovician Passive Carbonate Margin, Southern Appalachians. In: Derby, J.R., Fritz, R.D., Longacre, S.A., Morgan, W.A., Sternbach, C.A. (Eds.), The Great American Carbonate Bank. American Association of Petroleum Geologists, pp. 357–382. Rieboldt, S.E., 2005. Inarticulate Brachiopods of the Late Marjumiid and Peterocephaliid Biomeres (late Middle-early Late Cambrian) of West-central Utah and East-central, Nevada, U.S.A., University of California-Berkeley. University of California-Berkeley, p. 325. Rowell, A.J., Brady, M.J., 1976. Brachiopods and biomeres. Brigham Young Univ. Geol. Stud. 23, 165–180. Saltzman, M.R., Davidson, J., Holden, P., Runnegar, B., Lohmann, K.C., 1995. Sea-level-driven changes in ocean chemistry at an Upper Cambrian extinction horizon. Geology 23, 893–896. Saltzman, M.R., Runnegar, B., Lohmann, K.C., 1998. Carbon isotope stratigraphy of Upper Cambrian (Steptoean Stage) sequences of the eastern Great Basin: Record of a global oceanographic event. Geol. Soc. Am. Bull. 110, 285–297. Saltzman, M.R., Ripperdan, R.L., Brasier, M.D., Lohmann, K.C., Reichart, R.A., Chang, W.T., Peng, S., Ergaliev, E.K., Runnegar, B., 2000. A global carbon isotope excursion (SPICE) during the Late Cambrian: relation to trilobite extinctions, organic-matter burial and sea level. Palaeogeogr. Palaeoclimatol. Palaeoecol. 162, 211–223. Saltzman, M.R., Cowan, C.A., Runkel, A.C., Runnegar, B., Stewart, M.C., Palmer, A.R., 2004. The Late Cambrian Spice (δ13C) Event and the Sauk II-SAUK III Regression: New Evidence from Laurentian Basins in Utah, Iowa and Newfoundland. J. Sediment. Res. 74, 366–377. Sial, A.N., Peralta, S., Ferreira, V.P., Toselli, A.J., Acenolaza, F.G., Parada, M.A., Gaucher, C., Alonso, R.N., Pimentel, M.M., 2008. Upper Cambrian carbonate sequences of the Argentine Precordillera and the Steptoean C-Isotope positive excursion (SPICE). 13, 437–452.

373

Stitt, J.H., 1971. Repeating evolutionary pattern in Late Cambrian trilobite biomeres. J. Paleontol. 45, 178–181. Stitt, J.H., 1975. Adaptive radiation, trilobite paleoecology, and ex­tinction, Ptychaspidid Biomere, Late Cambrian of Oklahoma. Fossils Strata 4, 381–390. Stitt, J.H., Perfetta, P.J., 2000. Trilobites, biostratigraphy, and lithostratigraphy of the Crepicephalus and Aphelaspis zones Lower Deadwood Formation (Marjuman and Steptoean Stages, Upper Cambrian), Black Hills, South Dakota. J. Paleontol. 74, 199–223. Taylor, J.F., 2006. History and status of the biomere concept. Assoc. Australas. Paleontol. Mem. 32, 247–265. Thomas, W.A., 1991. The Appalachian-Ouachita rifted margin of southeastern North America. Geol. Soc. Am. Bull. 103, 415–431. Walker, K.R., Foreman, J.L., Srinivasan, K., 1990. The Cambrian Conasauga Group of Eastern Tennessee: A preliminary general stratigraphic model with a more detailed test for the Nolichucky Formation. Appalachian Basin Industrial Associates 17, pp. 184–189. Webb, E.J., 1980. Cambrian sedimentation and structural evolution of the Rome Trough in Kentucky. University of Cincinnati, pp. 1–120. Weber, L.J., 1988. Paleoenvironmental analysis and test of stratigraphic cyclicity in the Nolichucky Shale and Maynardville Limestone (Upper Cambrian) in central east Tennessee (Ed.) University of Tennessee, pp. 1–407. Westrop, S.R., 1988. Trilobite diversity patterns in an Upper Cambrian stage. Paleobiology 14, 401–409. Westrop, S.R., Ludvigsen, R., 1987. Biogeographic control of trilobite mass extinction at an Upper Cambrian "biomere" boundary. Paleobiology 13, 84–99. Woods, M.A., Wilby, P.R., Leng, M.J., Rushton, A.W.A., Williams, M., 2011. The Furongian (late Cambrian) Steptoean Positive Carbon Isotope Excursion (SPICE) in Avalonia. J. Geol. Soc. Lond. 168, 851–862. Wotte, T., Strauss, H., 2015. Questioning a widespread euxinia for the Furongian (Late Cambrian) SPICE event: indications from δ13C, δ18O, δ34S and biostratigraphic constraints. Geol. Mag. 152, 1085–1103. Zhu, M.-Y., Zhang, J.-M., Li, G.-X., Yang, A.-H., 2004. Evolution of C isotopes in the Cambrian of China: implications for Cambrian subdivision and trilobite mass extinctions. Geobios 37, 287–301.