Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation

PALAEO-08362; No of Pages 11 Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

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

Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation Brooks B. Ellwood a,⁎, Bruce R. Wardlaw b,1, Merlynd K. Nestell b, Galina P. Nestell b, Luu Thi Phuong Lan c a b c

Department of Geology and Geophysics, Louisiana State University, E235 Howe-Russell Geoscience Complex, Baton Rouge, LA 70803, United States Department of Earth and Environmental Sciences, University of Texas at Arlington, Arlington, TX 76019, United States Institute of Geophysics, Vietnamese Academy for Science and Technology, Hanoi, Vietnam

a r t i c l e

i n f o

Article history: Received 30 March 2017 Received in revised form 5 July 2017 Accepted 12 July 2017 Available online xxxx Keywords: GSSP concept PTB extinctions Conodont biostratigraphy Lung Cam Huangzhishan

a b s t r a c t Understanding the timing and correlation of significant global events in Earth history is facilitated by the Global Boundary Stratotype Section and Point (GSSP) concept, along with multi-proxy correlation techniques. As an example, the Permian–Triassic boundary (PTB) GSSP is used herein to correlate three PTB successions in east and southeast Asia. The PTB is defined using the First Appearance Datum (FAD) of the conodont Hindeodus parvus at the Meishan D section in China. By definition then, Meishan D is the only section on Earth where the FAD of H. parvus represents the beginning of the Triassic, at ~251.88 Ma, and thus the end of the Permian. Therefore, when correlating strata in any other section back to the PTB using biostratigraphic data, the local Lowest Observed Occurrence Point (LOOP) of H. parvus will probably not equate precisely to the defined FAD GSSP level (the PTB) for the beginning of the Triassic at Meishan D. The Graphic Correlation method, applied to PTB sites in China and Vietnam, is used herein to demonstrate that LOOPs of H. parvus in other successions are not equivalent in time to the PTB FAD. The LOOP and Highest Observed Occurrence Point (HOOP) for conodont data at two other successions studied, Huangzhishan in China, and Lung Cam in Vietnam, are used to determine the approximate level where the Triassic begins in these successions, resulting in high-resolution correlation among the sections and correlation back to the PTB GSSP level. It is demonstrated that when critical biostratigraphic data are missing, multiple proxy correlation techniques, geochemical, geophysical and, in some regional instances, unique lithostratigraphic information such as coeval ash beds, can be used to aid in locating the boundary in successions that are not the defining GSSP. LOOP and HOOP data are used to establish a Line of Correlation to differentiate between a defining PTB H. parvus FAD versus the H. parvus LOOP in secondary successions, and to project the PTB FAD into secondary sections to define the PTB at these localities. In addition, the timing of H. parvus arrivals at these sections is used to establish rough dispersal rates and patterns in the region. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Global correlation is a difficult problem in deep-time geology, but is critically important if we are to demonstrate that significant events in Earth history, such as large-magnitude extinctions, are either globally synchronous or not, and to determine if such events represent a very short amount of time, or not. To accomplish the correlations among various coeval successions, standards for global comparison are now being clearly established. Global Boundary Stratotype Sections and Points (GSSPs; Cowie, 1986) for all geologic stages have been defined or are in the process of being defined or redefined. It has generally been agreed by the International Commission on Stratigraphy (ICS) that a GSSP section should be defined based on the ⁎ Corresponding author. E-mail address: [email protected] (B.B. Ellwood). 1 Bruce Wardlaw passed away on March 23, 2016.

First Appearance Datum (FAD) of a certain marine species within that succession. However, the ICS does allow other possibilities to be used, including the Last Appearance Datum (LAD; or Last Occurrence Datum [LOD]) of a species, a magnetic reversal, or a geochemical signal in a well-studied stratigraphic succession (Cowie, 1986; Romane et al., 1996). Today, most GSSPs are defined based on the FAD of some organism, although at a few localities, other criteria have been used, e.g., the carbon isotopic anomaly used to define the Paleocene–Eocene boundary (Aubry et al., 2007; Gradstein et al., 2012), and the LAD (or extinction) of representatives of the foraminifer family Hantkeninidae used to define the Eocene–Oligocene boundary (Silva and Jenkins, 1993; Gradstein et al., 2012). A problem with the GSSP concept is that at any locality that is not identified as a GSSP, there has been a tendency for some workers to interpret a Lowest Observed Occurrence Point (LOOP) or Highest Observed Occurrence Point (HOOP) of the GSSP marker fossil within other than the GSSP succession, to be coeval with the equivalent GSSP

http://dx.doi.org/10.1016/j.palaeo.2017.07.012 0031-0182/© 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

2

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

FAD, or LAD when used as a GSSP marker event. Therefore, a LOOP (or HOOP) at localities other than the GSSP may improperly be assumed to be coeval with the FAD (or LAD) identified within the GSSP section (e.g., Landing et al., 2013). A case in point is the important Permian–Triassic boundary (PTB), where the GSSP boundary point defining the beginning of the Triassic has been established by the ICS as the FAD of the conodont Hindeodus parvus within a specific geological succession, Section D at Meishan in China (Yin et al., 2001). This unique GSSP locality is the only place on Earth where the FAD of H. parvus defines the beginning of the Triassic. If elsewhere the LOOP of H. parvus is used to estimate the boundary location, where the arrival of H. parvus in secondary successions represents a delayed or earlier occurrence, the LOOP of H. parvus will not represent the PTB level in the secondary succession. Unfortunately, as a result the of the misunderstanding of the uniqueness of the GSSP FAD, the PTB is often incorrectly placed at the first occurrence of H. parvus in secondary sections (e.g., Nicoll et al., 2002; Krull et al., 2004; Son et al., 2007; Kolar-Jurkovšek et al., 2011; Yan et al., 2013; Zhao et al., 2013; Yin et al., 2014; Xu et al., 2017; and many others). Other serious problems also arise, (a) the boundary-defining fossil/s may not even be found in a studied succession (e.g., Newton et al., 2004; Son et al., 2007; Richoz et al., 2010), (b) the boundary may be based on lithostratigraphy (Heydari and Hassanzadeh, 2003; Son et al., 2007), or (c) the boundary interval may represent a disconformity (Lehrmann et al., 2003; Payne et al., 2007).

1.1. Time versus time stratigraphic nomenclature and the critical need to differentiate between FAD versus LOOP There is some confusion in the literature that results from abbreviations used to represent the position in a stratigraphic succession where a fossil organism is found. Stratigraphers think in terms of height, position or ‘point’ in the section where the organism is first observed during sampling, not in terms of specific time. The terms lowest and highest are stratigraphic terms that are independent of time. It is useful within a GSSP, to establish a First Appearance Datum (FAD) to identify the time-stratigraphic fixed point in a succession, as is commonly used to define GSSPs. However, if a lower (earlier in time) occurrence of the marker fossil is found in the GSSP succession, then this new lowest observed point can be differentiated from the defining FAD by using the LOOP to identify this lower occurrence. For this and other reasons, the terms LOOP and HOOP as time independent stratigraphic terms have been introduced (Wardlaw et al., 2015; Nestell et al., 2015). Stratigraphic boundary successions, where these are condensed, where there is reworking of faunal elements, or where there are various depositional settings, also pose significant problems for interpretation and correlation. Therefore, when correlating to a given GSSP section, it is necessary to use all available information in an attempt to identify the boundary level in all non-GSSP secondary successions. This difficult problem is well-known in stratigraphy; thus, additional tools are needed for correlation, but such tools often relate to other fossils that carry with them the same problems as do the defining fossils, or to geochemical or geophysical indices that may not be well constrained in time. The Graphic Correlation method is a useful tool often employed to overcome some of these problems in stratigraphy, and for comparing a GSSP succession to individual secondary successions to identify the boundary level in those sections. Graphic Correlation (Shaw, 1964; Mann and Lane, 1995) has been used for many years for the purpose of developing fossil ranges, but herein, Graphic Correlation is applied to PTB boundary successions only for the purpose of identifying specific PTB boundary locations within secondary PTB successions. Other correlation methods have also been used in identifying the PTB in other than the GSSP succession (Brosse et al., 2016).

2. Previous work It has long been known that time differences exist between the stratigraphic occurrences of the same organism when found at different localities (e.g., Hedberg, 1965). A species originates only at one place and at one point in time. Very simplistically, any new organism must migrate or be dispersed from its source locality to all other locations. This concept is illustrated in Fig. 1, a time-distance cartoon modified from work by Hedberg (1965) and Eicher (1968), and further developed in the recent work of Landing et al. (2013) in their review of Cambrian successions. The point of origin of an organism is labeled as the ‘origination point’ in Fig. 1. This event occurs at a specific time, Time 1 in Fig. 1, and the new organism then begins to disperse/migrate away from the point of origin. In the example presented, the organism reaches Locality B after traveling for a relatively short period of time, Time 2 (Fig. 1). Millions of years later, a geologist collecting the succession at Locality B, can potentially find the LOOP of the organism at Time 2 within the section. Or, the geologist may miss the actual arrival points due to poor fossil preservation, thus introducing a larger difference in time within the B succession than would be expected from the location of the actual origination point in time (Fig. 1). Dispersal of this organism may eventually take it to localities A, C, D, and E, at Times 4, 3, 5, and 8, respectively, but, as discussed above, the actual arrival points within these successions may be missed in the LOOP reported, when not ‘observed’ in the section, although the organism may actually have ‘appeared’ at that locality. Also, because of an unfavorable environmental facies, the organism may seem to disappear from the geological record in the A and B successions at Times 6 and 7, respectively, seemingly representing HOOPs at those times, but then the organism reappears later in these successions at Times 11 and 9, respectively, as ‘Lazarus’ taxa. These comments assume that each succession is collected at a high enough resolution to resolve LOOPs and HOOPs. Due to an unconformity within successions A and B (Fig. 1), the HOOP for this organism will occur at Times 13 and 14, respectively. Local extinctions will create a HOOP for the organism in successions D and E at Times 12 and 10, respectively. If an estimate for geological time through the time-interval of interest is known, using either numerical dates or time-series analysis, then timing for the LOOPs or HOOPs of this organism can be estimated. An important question is, if the LOOPs or HOOPs are not coeval at different sites, how can these sites be correlated back to the FAD point in the GSSP section? This establishment of correlation can be difficult. Ideally, once the boundary point has been identified, a signal or proxy that is globally synchronous can be used for correlation. Earth's magnetic field reversals, changes in eustasy, climate cyclicity, or certain atmospherically controlled processes are essentially instantaneous (in a geologic sense), and therefore, global correlation efforts in some instances, can use geochemical or geophysical proxies, provided a method like Graphic Correlation is used that can account for uncertainties in the data. Included as possible proxies are geochemical events that are known to be near-global in character such as the Cretaceous–Paleogene iridium anomaly attributed to a bolide impact at that time (Alvarez et al., 1980), or the carbon isotope (δ13C) anomaly associated with the onset of the Paleocene–Eocene Thermal Maximum (PETM) that is used to define the base of the Ypresian Stage, and thus the base of the Eocene Series (Aubry et al., 2004). 2.1. Permian–Triassic examples The time interval that includes the latest Permian to earliest Triassic contains massive world-wide extinctions of biota, when more than 95% of terrestrial and marine species went to extinction through this interval (Raup, 1979; Hallam and Wignall, 1997). It is important to correctly identify the boundary in many localities to resolve, with high precision, the succession of events that preceded and succeeded these extinctions. In the case of the latest Permian, extinction affected tabulate and rugose corals, several classes of echinoderms and also some groups of

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

3

Fig. 1. Time-distance diagram modified from Eicher (1968) illustrating a hypothetical evolution point and dispersal of a species throughout its eventual range in the World's ocean through time. Locations A through E represent points where hypothetical geological successions exist that potentially contain the species in question. Barriers to dispersal, unfavorable facies, unconformities and local extinctions are illustrated as possible factors that limit rates of migration/dispersal or result in no recovery of the species of interest (see text).

ostracodes, brachiopods, bryozoans, bivalves, ammonoids, nautiloids, radiolarians, acritarchs, and foraminifers (Hallam and Wignall, 1997). These latest Permian extinctions have been characterized as a global extinction event or horizon (e.g., Sun et al., 2012; Nestell et al., 2015).

Conodonts have become the standard for relative age dating of many Paleozoic and early Mesozoic intervals, and are one of the few groups that have a robust record through the PTB. A succession in China, with reasonably well-illustrated and documented conodont PTB faunas, along with a second succession from Vietnam, are compared to the

Fig. 2. Location of the sections discussed in China and northern Vietnam. Also given are estimated travel times for Hindeodus parvus arrival times relative to the Meishan D section.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

4

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

GSSP succession for the PTB at the Meishan D section, China (Fig. 2). There, the beginning of the Triassic is placed at the base of Bed 27c, and is based on the first occurrence of the conodont species Hindeodus parvus in that section. H. parvus was selected to define the PTB because it is found at many globally distributed localities, and its LOOP in these secondary successions appeared to be nearly synchronous (Yin et al., 2001). However, a local LOOP or HOOP is rarely synchronous to the GSSP FAD. Therefore, to stress the localness of each occurrence, and that it is usually not a synchronous datum, usage of LOOP or HOOP, (1) can indicate the localness of the result, and (2) can differentiate between the defining FAD of the GSSP organism. The uniqueness of the GSSP makes it clear that the PTB is defined by the FAD of H. parvus only at Meishan D, but not by the LOOP of H. parvus in other sections. There are several Hindeodus species occurrences through the PTB interval, and also a few Clarkina species below the boundary, with some new occurrences through the transition interval that can be used for correlation to other successions. However, the morphology of many of these conodonts is very similar, species are difficult to differentiate, and therefore can be misidentified. To address this identification problem, Graphic Correlation is used herein to help resolve uncertainties in classification and identification of species. 2.2. Meishan D section, China The GSSP for the base of the Triassic is well documented, initially by Yin (1996) and later established as the GSSP by Yin et al. (2001). Mei et al. (1998) described the Clarkina conodont succession at Meishan, as did Wang and Henderson (2003), and these works have been slightly modified by Zhang et al. (2007). Nicoll et al. (2002) also described the Hindeodus conodont succession in South China. The PTB succession includes limestone beds making up the upper part of the Changxing Formation (Fm.), and limestone, dolomitic limestone, and mudstone beds composing the lower part of the overlying Yinkeng Fm. Lithologies for Beds 24a through 34 are shown in Fig. 3. 2.3. Huangzhishan section, China A detailed description of the PTB succession at Huangzhishan was first presented by Chen et al. (2004). Later, Chen et al. (2008) presented a detailed examination of the contained conodont faunas, and this description is used herein. The boundary succession includes limestone beds of the upper part of the Changxing Fm., and mudstone and argillaceous limestone beds of the lower part of the Yinkeng Fm. (Fig. 3). A thin paleosol layer and infilling burrows separates strata at the top of the Changxing Fm. from the base of the Yinkeng Fm. The upper part of the section, from Bed 7 through Bed 20, is shown in Fig. 3 (modified from Chen et al., 2008). 2.4. Lung Cam section, Vietnam Biostratigraphic work on the Lung Cam section has identified a diverse fauna, including conodonts and foraminifers (Son et al., 2007; Metcalfe, 2012; Nestell et al., 2015; Wardlaw et al., 2015). Magnetic susceptibility and geochemical work has also been reported for the section (Nestell et al., 2015). The boundary succession includes foraminiferal packstone, wackestone and ash beds of the upper part of the Dong Dang Fm., and skeletal carbonate mudstone, ash, and very thin single beds of recrystallized limestone and microgastropodal packstone of the Hong Ngai Fm. (Fig. 3). H. parvus is subdivided into two subspecies, H. parvus parvus and H. parvus erectus (Kozur, 1996) that are often not differentiated (Nicoll et al., 2002; Chen et al., 2008) or used (Chen et al., 2009), because they virtually have the same range. Chen et al. (2009) demonstrated that both subspecies can be consistently differentiated by the number of denticles, length, and height of the blade of the P1 (Platform) element that is usually used to identify the species. These two subspecies are found at Lung Cam (Wardlaw et al., 2015), but for

this study the LOOP of either H. parvus parvus or H. parvus erectus was used to denote the LOOP of H. parvus. The LOOP of these conodonts is found in Bed +28 at ~11.1 m in the measured section (Fig. 3). 3. Methodology 3.1. Graphic Correlation Given enough biostratigraphic, lithostratigraphic, geophysical or geochemical data from a GSSP and secondary successions, an excellent way to establish a reasonably unambiguous correlation between two successions is by using the Graphic Correlation method (e.g., see Shaw, 1964; Dowsett, 1989; Mann and Lane, 1995; Edwards, 1984, 1989, 1995; Carney and Pierce, 1995). Graphic Correlation is used to compare information from different geological successions, and may be used simply to make a high precision correlation between two stratigraphic successions. Or, if large numbers of sections with abundant, well-studied faunas are available, Graphic Correlation can be used to build a Composite Reference Section that represents the range of geological information available from all sections, such as the ‘total’ range represented by an organism from a given period of time. However, the goal of this manuscript was to identify the PTB locations in the two secondary successions studied, Huangzhistan and in China, and Lung Cam in Vietnam, by Graphically Correlating these successions to the Meishan D GSSP section, and thus projecting the PTB into the secondary sections. To illustrate the Graphic Correlation method as employed herein, a very simple example for two successions, a Hypothetical Reference Section and a Hypothetical Comparison Section, is given in Fig. 4. This correlation model plots the ‘FAD’ (O3 in Fig. 4) as a hypothetical GSSP, as well as LOOPs and HOOPs for three hypothetical species (O1–O3) for the two hypothetical sections (Fig. 4). In addition, because well-studied GSSP successions often contain good geochemical, geophysical and lithostratigraphic information, such hypothetical data are included in Fig. 4 as three additional data points representing ash beds, geochemical (GC; these could be isotopic signatures), and geophysical (GP; these could be polarity or gamma radiation signatures) anomalies that are common to both hypothetical successions. As used herein, the hypothetical LOOPs and HOOPs are presented with uncertainties identified for each hypothetical organism. Therefore, each LOOP or HOOP is represented by a rectangle (Fig. 4), representing sampling uncertainties due primarily to sampling interval used in the two successions sampled. For example, if samples were collected at 0.5 m intervals in the Hypothetical Reference Section, and organism O1 was found at 3.0 m height in the section, but not in the sample at 2.5 m, then the uncertainty for O1 would be ~0.5 m in the Hypothetical Reference Section. If samples were collected at 1.0 m intervals in the Hypothetical Comparison Section, and organism O1 was found at 1.0 m but not in the sample at 0.0 m, then the uncertainty would be ~1.0 m. These LOOPs, with uncertainties, are given as the rectangle labeled O1 in Fig. 4. Rectangles for the LOOPs and HOOPs of hypothetical organisms O1, O2 and the hypothetical FAD organism O3 are also given in Fig. 4. Labeled small squares in Fig. 4 identify the positions of GC and GP anomalies, a hypothetical ash bed and an extinction event, all identified to be common to both successions. After all these data points are plotted, a subjective, ‘best fit’ Primary Line of Correlation (LOC) is fit through these data points to establish a relative time association between the two successions being considered (Fig. 4). This LOC takes into account data quality and uncertainties. Because sediment accumulation rates (SARs) in sections are generally different and may be somewhat variable, the relative stratigraphic position of data points in different sections will be different, even if they represent deposition at exactly the same time. Therefore, the ‘best-fit’ subjective approach adjusts for differences in SARs, and becomes important when very different SARs are represented between sections being compared. When steep LOCs are produced, the uncertainties of projected positions increase in correlated sections.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

5

Fig. 3. Columnar sections for the Permian–Triassic boundary (PTB) interval; Huangzhishan and Meishan China, and Lung Cam, Vietnam, with scale bar for all included; all sections showing the Lowest Observed Occurrence Point (LOOP; coarse dashed line) of Hindeodus parvus and the placement of the PTB graphically correlated to the Meishan GSSP FAD. The solid line is the PTB as determined for the Huangzhishan and Lung Cam sections, by projecting the PTB from Meishan through the main Lines of Correlation (LOCs) into the respective sections using Graphic Correlation. Arrows indicate the top of the extinction event level (event horizon) in each succession and these levels are connected by the fine dashed lines in the figures. Specific heights in each section are given for the intersections of the dashed and solid lines in the diagram. Also included are the geologic formation names for each section and selected bed numbers for each section.

Once the relative positions of LOOPs and HOOPs are established for the two hypothetical successions, the relative ages (stratigraphic positions) for all these data points become visually clear. Two LOCs are illustrated in the model (Fig. 4). The iterative process begins by obtaining a best-fit Primary LOC fitting through as many elements in the plots as possible. Important in these plots are the extinction event level that is assumed to have occurred at approximately the same time in both sections, and the ash bed, again assumed to be coeval. One problem is that very few other data points come near to a fit with the LOC. Therefore, a Secondary LOC is given to show a second possibility where other fitting assumptions are made. Once the Primary LOC is adjusted to the best

possible position, the FAD from the Hypothetical Reference Section is projected through the Primary LOC and into the Hypothetical Comparison Section, to determine the projected location of the boundary position in the secondary section (Fig. 4). The techniques discussed above are used in Figs. 5 and 6 discussed below. 4. Graphic Correlation results Herein, Graphic Correlation is used to compare the known fossil occurrences and events from the PTB GSSP Meishan D section to two secondary PTB sections in a very simple way; each section is compared

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

6

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

Fig. 4. A Graphic Correlation model representing two successions, a Hypothetical Reference Section and a Hypothetical Comparison Section. Data point types are given in the legend; rectangles represent uncertainties due to sample distribution for LOOP – Lowest Observed Occurrence Points and HOOP – Highest Observed Occurrence Points of named fossils O1 to O3. Note that the O3 dataset represents the total range with uncertainties for that organism. Also included are an ash bed, a geochemical anomaly (GC), such as carbon or oxygen isotopic anomalies, and a geophysical anomaly (GP), such as polarity reversal, gamma radiation peaks, or significant magnetic susceptibility changes, all factors that represent possible useful data other than fossil occurrences. The Line of Correlations (LOCs) represents two possible ‘best-fit’ subjective interpretations through the data scatter. O3 was chosen to represent the First Appearance Datum (FAD) of the defining fossil, in this case fossil O3 in the Hypothetical Reference Section. See text for an in-depth explanation.

directly to the GSSP section to define the best-fit LOC between the GSSP and that section. After the LOC has been established, the PTB level at Meishan D is projected through the respective LOC into the secondary section. After Graphic Correlation, the projected PTB level to the Huangzhishan and Lung Cam sections does not correspond to the stratigraphic level of the FAD of H. parvus at Meishan D (Figs. 5 and 6). Even though the identification of the various species is consistent among sections, the sampling interval in these secondary sections is often course, resulting in the uncertainties shown in Table 1 and illustrated in the LOOPs and HOOPs given in Figs. 5 and 6. Therefore, the projected PTB in the secondary sections has some uncertainty. In Figs. 5 and 6, as in the Fig. 4 model, rectangles are used to indicate fossil uncertainties in plotting the respective LOOPS and HOOPS (the uncertainties for the conodont data are reported in Table 1). In the Graphic Correlation diagrams reported herein (Figs. 5 and 6), the scale is the same for both the X and Y axes. As a result, the slopes of the LOCs give an indication of the relative sediment accumulation rate

(SAR) in each section with respect to the GSSP, which is significantly condensed. Therefore, because the SAR at Meishan D is so low, the LOC slopes in Figs. 5 and 6 are steep, indicating much higher SARs in the Huangzhishan and Lung Cam sections relative to Meishan D. For comparison, in the model example (Fig. 4), the angle of the LOC with respect to the x-axis is ~50°, indicating that the SAR for the Hypothetical Comparison Section is only slightly higher than that for the Hypothetical Reference Section. 5. Discussion Graphic Correlation for the Huangzhishan section in China and Lung Cam in Vietnam, to the GSSP section at Meishan D, is shown in Figs. 5 and 6. The sampling range for each observed occurrence is displayed as a LOOP or HOOP box, with uncertainties given in Table 1. These correlations are based on the FAD of H. parvus, and the ranges of all the identified conodonts common to both sections, which include seven species of Clarkina, and nine species of Hindeodus. After graphically

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

7

Fig. 5. Comparison of the Huangzhishan section (Figs. 2 and 3) to the Meishan D GSSP section using Graphic Correlation. The PTB, as defined by the FAD of Hindeodus parvus at Meishan, is projected through the Main Line of Correlation (LOC) into the Huangzhishan section (defined mainly by LOOPs and HOOPs of conodonts identified in the two sections, with their uncertainties represented by corresponding rectangles; weighted toward conodonts with lower uncertainties), indicating the PTB at ~10.25 m in the Huangzhishan section. The PTB lies ~0.5 m above the LOOP of H. parvus. The extinction event horizon lies at 6.0 m in the section. A slight hiatus is indicated at 6 m in the Huangzhishan section.

correlating among the above mentioned successions, which contain well-documented conodont faunas, neither of the secondary successions (successions to which the GSSP is compared) exhibits a LOOP of H. parvus that is coincident with the Permian–Triassic boundary position as projected from the FAD of H. parvus in the Meishan D GSSP section, to the respective secondary section (Figs. 5 and 6). 5.1. Huangzhishan section, China A small hiatus is indicated in the Huangzhishan section at ~ 6.0 m, where burrows and a paleosol are identified (Fig. 3). This paleosol layer lies within a critical interval and may have produced unrecognized

problems because of missing section at this height (Fig. 5), around which a number of conodont species have been identified. In addition, each section appears to be expanded relative to the Meishan GSSP, and therefore, the uncertainties are relatively large in projecting the boundary FAD of H. parvus at the Meishan D GSSP section, through the Main LOC and into the Huangzhishan section. The uncertainties are unknown, but are estimated to be no more than 0.5 m. In addition to conodont data, an additional aid to constrain the LOC is the top of the major extinction event that is also seen at Meishan D. The PTB in the Huangzhishan section is projected to fall at ~ 10.25 m, ~ 0.5 m above the LOOP of H. parvus in the section. No secondary LOC is given for this data set.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

8

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

Fig. 6. Comparison of the Lung Cam section, Vietnam (Figs. 2 and 3) to the Meishan D GSSP section using Graphic Correlation (slightly modified from Wardlaw et al. (2015) and Nestell et al. (2015) with additional information added). The PTB, as defined by the FAD of H. parvus at Meishan D, is projected through the Main Line of Correlation (LOC) into the Lung Cam section (defined by LOOPs and HOOPs of conodonts identified in the two sections, with their uncertainties represented by corresponding boxes), indicating the PTB lies at ~11.8 m in the Lung Cam section, ~0.8 m above the LOOP of H. parvus. The extinction event horizon lies at 9.06 m in the section.

In the Graphic Correlation comparisons between the two secondary sections and the GSSP at Meishan D, we have also assigned absolute ages from Burgess et al. (2014) to the PTB level, at 251.88 ± 0.031 Ma, and the top of the extinction event, at 251.941 ± 0.037 Ma, yielding a separation of ~61,000 years. These two ages are shown in the Graphic Correlation diagrams presented in Figs. 5 and 6. These dates allow estimates of SARs for the studied sections through the interval between the top of the extinction level and the PTB in the Meishan D section, and within the secondary sections at the stratigraphic levels to which the PTB is projected. For the Meishan D section, that interval is ~ 0.2 m, where ~ 0.06 m of that interval is a single ash making up Bed 25, thus yielding an SAR for the section of ~0.33 cm/kyears. If the ash bed is assumed to represent an instantaneous event, thus artificially thickening this interval, then an alternate SAR using an interval thickness of 0.14 m, yields ~0.23 cm/kyears for that interval in the Meishan D section. In the Huangzhishan section, the projected PTB lies ~4.3 m above

the extinction event level in the section, thus representing an SAR of ~7.05 cm/kyears (Fig. 5). 5.2. Lung Cam section, Vietnam The Lung Cam section is very much expanded when compared to the Meishan D section (Fig. 6), and samples were collected at relatively close spacing through the interval containing the LOOP of H. parvus. In this section, the large uncertainties in two of the three samples with large ranges balance each other (Fig. 6), so that only one LOC seems reasonable. In addition to conodont data, constraining the LOC is the identification of an ash bed in Lung Cam (Bed 13) that appears, due to its position just above the extinction event beds, Bed 10 and 11 at Lung Cam, to be equivalent to the Bed 25 ash bed at Meishan D, also lying just above the extinction event horizon in Meishan D Bed 24 (Nestell et al., 2015). Placement of the projected PTB into the Lung Cam section,

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

9

Table 1 Conodont distribution in studied sections. Conodonts

Meishan

LOOP (m)

HOOP (m)

Lung Cam

LOOP (m)

HOOP (m)

Huang.

LOOP (m)

HOOP (m)

C. yini C. parasubcarinata C. iranica C. zhangi C. meishanensis H. changxingensis C. deflecta H. priscus H. eurypyge H. praeparvus H. latidentatus H. parvus H. n. sp. A C. zhejiangensis H. typicalis H. inflatus Top Extinction Level Top Bed 25/26 ash

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 41.63 m 41.73 m

39.35 ± 0.35 No 39.35 ± 0.35 No 41.66 ± 0.03 41.71 ± 0.03 41.75 ± 0.02 41.75 ± 0.02 41.75 ± 0.02 41.75 ± 0.02 42.20 ± 0.1 41.83 ± 0.02 41.69 ± 0.04 41.72 ± 0.02 No 40.77 ± 0.44 No data No data

41.75 ± 0.03 41.48 ± 0.06 40.31 ± 0.7 41.59 ± 0.05 41.71 ± 0.03 42.7 ± 0.3 No No No No 42.30 ± 0.1 No No 41.72 ± 0.02 44.33 ± 0.4 No No data No data

Yes No No No No Yes No Yes Yes Yes No Yes Yes No No No 9.06 m 9.25 m

No No No No No 11.1 ± 0.03 No 10.82 ± 0.06 10.46 ± 0.09 10.46 ± 0.09 No 11.1 ± 0.03 10.46 ± 0.09 No No No No data No data

10.4 ± 0.09 No No No No No No No No No No No No No No No No data No data

Yes No No Yes Yes Yes Yes No Yes Yes Yes Yes No Yes No Yes 6.0 m No data

No No No No 5.75 ± 0.05 9.04 ± 0.09 No No 9.63 ± 0.06 4.01 ± 0.1 No 9.75 ± 0.06 No 5.92 ± 0.03 No 9.35 ± 0.07 No data No data

5.92 ± 0.03 No No 5.46 ± 0.06 6.03 ± 0.08 9.22 ± 0.1 5.67 ± 0.04 No No No 11.37 ± 0.14 No No 6.03 ± 0.08 No No No data No data

Table 1 footnotes: Conodont species identified in sections studied herein. Column 1 - Conodonts: genera C. (Clarkina) and H. (Hindeodus), with species names given in Column 1. Species H. n. sp. A, is from Nicoll et al. (2002). A ‘yes’ in Columns 2 (labeled Meishan), Column 5 (labeled Lung Cam) and Column 8 (labeled Huang. for Huangzhishan) indicates that the conodont named in Column 1 is present in the respective section; a ‘no’ means it is not found. Columns labeled LOOP and HOOP give the height interval with uncertainties in the sections where the respective conodont is found: Columns 3 and 4 for Meishan; Columns 6 and 7 for Lung Cam; and Columns 9 and 10 for Huangzhishan. The last two lines in Columns 1, 5 and 8 represent (1) the top of the latest Permian mass extinction level in all three sections, and (2) height of ash Bed 25/26 in the Meishan GSSP section, China, which is only correlated to Lung Cam section, Vietnam. (Data from Wardlaw et al., 2015, Nestell et al., 2015, and from our collective field studies.)

using the Graphic Correlation presented in Fig. 6, is at ~11.8 m height in the section, therefore lying at ~ 0.8 m above the LOOP of H. parvus (Fig. 6). In the Lung Cam section, the projected PTB lies ~ 2.8 m above the extinction event level in the section, representing an SAR of ~4.6 cm/kyears at Lung Cam (Fig. 6), lying between Lung Cam Beds 36 and 37 (Fig. 3). Note that this level has been revised slightly from the work of Nestell et al. (2015) and Wardlaw et al. (2015). 5.3. Timing of H. parvus occurrence Fig. 2 gives the location of the studied sections and timing of ‘arrival’ of H. parvus at each of the secondary sites relative to the Meishan D

GSSP. Lung Cam is located ~1700 km from Meishan. Huangzhishan is located ~100 km from Meishan. At Huangzhishan, the LOOP of H. parvus lies ~0.5 m below the projected PTB (Fig. 5). Given the SAR for the section, then H. parvus is found ~ 7000 years (Fig. 2) before H. parvus at Meishan. In the Lung Cam section, the LOOP of H. parvus lies ~ 0.8 m below the projected PTB (Fig. 6). Given the SAR for the section, then H. parvus is found ~ 17,000 years (Fig. 2) after H. parvus at Meishan. These data are interpreted to indicate that H. parvus may have originated at or south of the Lung Cam section during the Permian (Fig. 7). The species then dispersed/migrated northward toward Meishan (Fig. 7), first reaching Huangzhishan ~ 10 kyears after its first appearance at Lung Cam (Fig. 2), and finally arriving at Meishan ~7 kyears after that.

Fig. 7. Paleogeographic map for the Permian–Triassic boundary (PTB) at ~252 Ma (Scotese, personal communication) and modified by the authors. The location of the studied sections in Vietnam (Lung Cam section; Fig. 2) and the Meishan GSSP (Fig. 2) in China are identified by the labeled black dots.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

10

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx

6. Conclusions The conodont distributions Huangzhishan in China, and Lung Cam in Vietnam, are compared to the Meishan D Global Boundary Stratotype Section and Point (GSSP) section, China, using the Graphic Correlation method. The data set presented herein includes 16 different species of conodonts, all observed within the Meishan D section. Eleven of these are identified at Huangzhishan, and 7 at Lung Cam. LOOPs (Lowest Observed Occurrence Point) and HOOPs (Highest Observed Occurrence Points) of these conodonts were compared to the FAD defining the PTB stratigraphic level within the Meishan GSSP section, the point where the Triassic, Induan Stage, begins. A Line of Correlation (LOC) fit through the range of conodont data used in the Graphic Correlation process allowed placement of the stratigraphic position of the PTB in the Huangzhishan and Lung Cam sections. This work resulted in determining the following relative positions within each section studied. 1. At the Huangzhishan section, China, the PTB was determined to lie at the ~10.25 m level within the section, ~4.25 m above the Changxing/ Yinkeng Fm. contact, and ~0.5 m above the LOOP of H. parvus within the section. 2. At the Lung Cam section, Vietnam, the PTB was determined to lie at the ~11.8 m level within the section, ~2.8 m above the Dong Dang/ Hong Ngai Formation contact, and ~ 0.8 m above the LOOP of H. parvus within the section. Paleogeographic comparison of the sections studied here suggests that in the Paleotethys, H. parvus dispersed/migrated toward the north from at least as far south as Lung Cam, arriving at the Meishan GSSP ~17 kyears later, after passing through the Huangzhishan locality. Acknowledgments Partial support on this project was provided by NSF EAR-0745393 to BBE and an endowment to LSU by Robey Clark. Partial funding was provided to LTPL by VAST05.03/17-18. We would like to thank the Dr. Lucy Edwards and other anonymous reviewers for their constructive criticism of an earlier version of this manuscript; we are grateful to Steve Benoist for useful comments on Fig. 1, and to Sue and Amber Ellwood for help in sampling in Vietnam. References Aubry, M-P., Ouda, K., Dupuis, C., Berggren, W.A., Van Couvering, J.A., Ali, J., Brinkhuis, H., Gingerich, P.R., Heilmann-Clausen, C., Hooker, J., Kent, D.V., King, C., Knox, R.O.B., Laga, P., Molina, E., Schmitz, B., Steurbaut, E., Ward, D.R., 2004. The global standard stratotype-section and point (GSSP) for the base of the Eocene Series in Dababiya section (Egypt). Episodes v. 30, 271–286. Alvarez, L.W., Walter Alvarez, W., Asaro, F., Michel, H.V., 1980. Extraterrestrial cause for the Cretaceous–Tertiary extinction. Science 208, 1095–1108. Aubry, M.-P., Ouda, K., Dupuis, C., Berggren, W.A., Van Couvering, J.A., Ali, J., Brinkhuis, H., Gingerich, P.R., Heilmann-Clausen, C., Hooker, J., Kent, D.V., King, C., Knox, R.O.B., Laga, P., Molina, E., Schmitz, B., Steurbaut, E., Ward, D.R., 2007. The Global Standard Stratotype-Section and point (GSSP) for the base of the Eocene Series in Dababiya section (Egypt). Episodes 30, 271–286. Brosse, M., Bucher, H., Goudemand, N., 2016. Quantitative biochronology of the Permian– Triassic boundary in South China based on conodont unitary associations. Earth Sci. Rev. 155, 153–171. Burgess, S.D., Bowring, S., Shen, S.-z, 2014. High-precision timeline for Earth's most severe extinction. Proceedings of the National Academy of Sciences. 111:pp. 3316–3321. http://dx.doi.org/10.1073/pnas.1317692111. Carney, J.L., Pierce, R.W., 1995. Graphic correlation and composite standard databases as tools for the exploration biostratigrapher. SEPM Spec. Publ. 53, 23–43. Chen, Z.Q., Shi, G.R., Kaiho, K., 2004. New ophiuroids from the Permian/Triassic boundary beds of South China. Palaeontology 47, 1301–1312. Chen, J., Henderson, C.M., Shen, S.Z., 2008. Conodont succession around the Permian–Triassic boundary at the Huangzhishan section, Zhejiang and its stratigraphic correlation. Acta Palaeontol. Sin. 47, 91–114. Chen, J., Beatty, T.W., Henderson, C.M., Rowe, H., 2009. Conodont biostratigraphy across the Permian-Triassic boundary at the Dawen section, Great Bank of Guizhou, Guizhou Province, South China: implications for the Late Permian extinction and correlation with Meishan. J. Asian Earth Sci. 36, 442–458.

Cowie, J.W., 1986. Guidelines for boundary stratotypes. Episodes 9, 78–82. Dowsett, H.J., 1989. Application of the graphic correlation method to Pliocene marine sequences. Marine Micropaleontology 14, 3–32. Edwards, L.E., 1984. Insights on why graphic correlation (Shaw's method) works. J. Geol. 92, 583–597. Edwards, L.E., 1989. Supplemented graphic correlation: a powerful tool for paleontologists and nonpaleontologists. PALAIOS 4, 127–143. Edwards, L.E., 1995. Graphic correlation: some guidelines on theory and practice and how they relate to reality. SEPM Spec. Publ. 53, 45–50. Eicher, D.L., 1968. Geologic Time. Prentice-Hall, Inc. (148 pp.). Gradstein, F.M., Ogg, G.G., Schmitz, M.D., Ogg, G.M., 2012. The Geologic Time Scale 2012. vols. 1 and 2. Amsterdam, Boston, Heidelberg, London, New York, Oxford, Paris, San Diego, San Franciso, Singapore, Sydney, Tokyo, Elsevier (1144 pp.). Hallam, A., Wignall, P.B., 1997. Mass Extinctions and Their Aftermath. Oxford University Press, Oxford (320 pp.). Hedberg, H.D., 1965. Earth history and the record in the rocks. Proc. Am. Philos. Soc. 109, 99–104. Heydari, E., Hassanzadeh, J., 2003. Deev Jahi Model of the Permian–Triassic boundary mass extinction: a case for gas hydrates as the main cause of biological crisis on Earth. Sediment. Geol. 163, 147–163. Kolar-Jurkovšek, T., Jurkovšek, B., Aljinović, D., 2011. Conodont biostratigraphy and lithostratigraphy across the Permian–Triassic boundary at the Lukač section in western Slovenia. Riv. Ital. Paleontol. Stratigr. 117, 115–133. Kozur, H., 1996. The conodonts Hindeodus, Isarcicella and Sweetohindeodus in the uppermost Permian and lowermost Triassic. Geologia Croatica 49, 81–115. Krull, E.S., Lehrmann, D.J., Druke, D., Kessel, B., Yu, Y.Y., Li, R., 2004. Stable isotope stratigraphy across the Permian–Triassic boundary in shallow marine carbonate platforms, Nanpanjiang Basin, South China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 204, 297–315. Landing, E., Geyer, G., Brasier, M.D., Bowring, S.A., 2013. Cambrian evolutionary radiation: context, correlation, and chronostratigraphy–overcoming deficiencies of the first appearance datum (FAD) concept. Earth-Sci. Rev. 123, 133–172. Lehrmann, D.J., Payne, J.L., Felix, S.V., Dillett, P.M., Wang, H.M., Yu, Y.Y., Wei, J.J., 2003. Permian–Triassic boundary sections from shallow-marine carbonate platforms of the Nanpanjiang Basin, South China: implications for oceanic conditions associated with the end-Permian extinction and its aftermath. PALAIOS 18, 138–152. Mann, K.O., Lane, H.R. (Eds.), 1995. Graphic correlation. SEPM Society for Sedimentary Geology Spec. Publ. no. 53 (263 pp.). Mei, S., Zhang, K., Wardlaw, B.R., 1998. A refined succession of Changhsingian and Griesbachian neogondolellid conodonts from the Meishan section, candidate of the global stratotype section and point of the Permian–Triassic boundary. Palaeogeogr. Palaeoclimatol. Palaeoecol. 143, 213–226. Metcalfe, I., 2012. Changhsingian (Late Permian) conodonts from Son La, northwest Vietnam and their stratigraphic and tectonic implications. J. Asian Earth Sci. 50, 141–149. Nestell, G.P., Nestell, M.K., Ellwood, B.B., Wardlaw, B.R., Basu, A.R., Ghosh, N., Lan, L.T.P., Rowe, H.D., Hunt, A., Tomkin, J.H., Ratcliffe, K.T., 2015. High influx of carbon in walls of agglutinated foraminifers during the Permian–Triassic transition in global oceans. Int. Geol. Rev. 57:411–427. http://dx.doi.org/10.1080/00206814.2015. 1010610. Newton, R.J., Pevitt, E.L., Wignall, P.B., Bottrell, S.H., 2004. Large shifts in the isotopic composition of seawater sulphate across the Permo–Triassic boundary in northern Italy. Earth Planet. Sci. Lett. 218, 331–345. Nicoll, R.S., Metcalfe, I., Wang, C.Y., 2002. New species of the conodont genus Hindeodus and the conodont biostratigraphy of the Permian–Triassic boundary interval. J. Asian Earth Sci. 20, 609–631. Payne, J.L., Lehrmann, D.J., Follett, D., Seibel, M., Kump, L.R., Riccardi, D.A., Sano, H., Wei, J., 2007. Erosional truncation of uppermost Permian shallow-marine carbonates and implications for Permian–Triassic boundary events. Geol. Soc. Am. Bull. 119, 771–784. Raup, D.M., 1979. Size of the Permo-Triassic bottleneck and its evolutionary implications. Science 206:217–218. http://dx.doi.org/10.1126/science.206.4415.217. Richoz, S., Krystyn, L., Baud, A., Brandner, R., Horacek, M., Mohtat-Aghai, P., 2010. Permian–Triassic boundary interval in the Middle East (Iran and N. Oman): progressive environmental change from detailed carbonate carbon isotope marine curve and sedimentary evolution. J. Asian Earth Sci. 39, 236–253. Romane, J., Bassett, M.G., Cowie, J.W., Gohrbandt, K.H., Lane, H.R., Michelsen, O., Wang, N., 1996. Revised guidelines for the establishment of global chronostratigraphic standards by the International Commission on Stratigraphy (ICS). Episodes 19, 77–81. Shaw, A.B., 1964. Time in Stratigraphy. McGraw-Hill, p. 365. Silva, I.P., Jenkins, D.G., 1993. Decision on the Eocene-Oligocene boundary stratotype. Episodes 16, 379–382. Son, T.H., Koeberl, C., Ngoc, N.L., Huyen, D.T., 2007. The Permian–Triassic boundary sections in northern Vietnam (Nhi Tao and Lung Cam sections): carbon-isotope excursion and elemental variations indicate major anoxic event. Palaeoworld 16, 51–66. Sun, Y., Joachimski, M.M., Wignall, P.B., Yan, C., Chen, Y., Jiang, H., Wang, L., Lai, X., 2012. Lethally hot temperatures during the early Triassic Greenhouse. Science 338, 366–370. Wang, Y., Henderson, C.M., 2003. Report of the workshop “Lopingian Stratigraphy and Events”. Permophiles 43, 5–7. Wardlaw, B.R., Nestell, M.K., Nestell, G.P., Ellwood, B.B., Lan, L.T.P., 2015. Conodont biostratigraphy of the Permian–Triassic boundary sequence at Lung Cam, Vietnam. Micropaleontology 61, 313–334. Xu, G., Deconinck, J.-F., Feng, Q., Baudin, F., Pellenard, P., 2017. Clay mineralogical characteristics at the Permian–Triassic Shangsi section and their paleoenvironmental and/or paleoclimatic significance. Palaeogeogr. Palaeoclimatol. Palaeoecol. 474, 152–163.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012

B.B. Ellwood et al. / Palaeogeography, Palaeoclimatology, Palaeoecology xxx (2017) xxx–xxx Yan, C., Wang, L., Jiang, H., Wignall, P.B., Sun, Y., Chen, Y., Lai, X., 2013. Uppermost Permian to lower Triassic conodonts at Bianyang section, Guihzou Province, South China. PALAIOS 28, 509–522. Yin, H. (Ed.), 1996. The Palaeozoic–Mesozoic Boundary, Candidates of Global Stratotype Section and Point of the Permian–Triassic Boundary. China University of Geosciences Press, Wuhan, China (137 pp.). Yin, H., Zhang, K., Tong, J., Yang, Z., Wu, S., 2001. The Global stratotype Section and Point (GSSP) of the Permian–Triassic boundary. Episodes 24, 102–114. Yin, H., Jiang, H., Xia, W., Feng, Q., Zhang, N., Shen, J., 2014. The end-Permian regression in South China and its implication on mass extinction. Earth Sci. Rev. 137, 19–33.

11

Zhang, K.X., Tong, J.N., Shi, G.R., Lai, X.L., Uy, J.X., He, W.H., Peng, Y.Q., Jin, Y.L., 2007. Early Triassic conodont-palynological biostratigraphy of the Meishan D section in Changxing, Zhejiang Province, South China. Palaeogeogr. Palaeoclimatol. Palaeoecol. 252, 4–23. Zhao, L., Chen, Y., Chen, Z.-Q., Cao, L., 2013. Uppermost Permian to lower Triassic conodont zonation from Three Gorges area, South China. PALAIOS 28, 523–540.

Please cite this article as: Ellwood, B.B., et al., Identifying globally synchronous Permian–Triassic boundary levels in successions in China and Vietnam using Graphic Correlation, Palaeogeogr. Palaeoclimatol. Palaeoecol. (2017), http://dx.doi.org/10.1016/j.palaeo.2017.07.012