A NEW TRIASSIC TIMESCALE

Transcription

1 366 Tanner, L.H., Spielmann, J.A. and Lucas, S.G., eds., 2013, The Triassic System. New Mexico Museum of Natural History and Science, Bulletin 61. A NEW TRIASSIC TIMESCALE SPENCER G. LUCAS New Mexico Museum of Natural History and Science, 1801 Mountain Road N. W., Albuquerque, New Mexico 87104, spencer.lucas@state.nm.us Abstract The current Triassic chronostratigraphic scale is a hierarchy of three series divided into seven stages, divided further into 15 substages. Ammonoid and conodont biostratigraphies provide the primary basis for chronostratigraphic definition based on global stratotype sections and points (GSSP). I propose that Triassic chronostratigraphic definition should rely entirely on ammonoid biochronological events and thereby reject conodont biostratigraphy and the GSSP methodology. Here I propose a new Triassic chronostratigraphy that divides the Triassic into four series (Scythian, Dinarian, Carnian and Norian), recognizes a four-stage Scythian (Griesbachian, Dienerian, Smithian and Spathian), elevates the Carnian and Norian substages (Julian, Tuvalian, Lacian, Alaunian and Sevatian) to stage rank, and includes the Rhaetian as a stage in the Norian Series. A sparse but growing database of precise radioisotopic ages supports the following calibrations: base of Triassic ~ 252 Ma, base Smithian ~ 251 Ma, base Anisian ~ 247 Ma, base Ladinian ~ 242 Ma, base Carnian ~ 237 Ma, base Norian ~ 221 Ma, base Rhaetian ~ 205 Ma, base Jurassic ~ 201 Ma. Triassic magnetostratigraphy is a series of multichrons at best, and needs vast improvement to make a substantial contribution to the Triassic timescale..a GSSP is intended to be permanent, and thus immune to further changes in the status of its perceived naturalness. Walsh, Gradstein and Ogg (2004, p. 208) The game of science is, in principle, without end. He who decides one day that scientific statements do not call for any further test, and that they can be regarded as finally verified, retires from the game. Karl Popper (1959, p. 53) INTRODUCTION Development of a Triassic timescale has a nearly 200-year-long history that began when Alberti (1834) proposed the Triassic as a formation between the Zechstein and Jura. During the last 40 years, development of the Triassic timescale has been dominated by the work of the Subcommission on Triassic Stratigraphy (STS), part of the I.U.G.S. International Commission on Stratigraphy. This work has been undertaken in the Hedbergian tradition of defining global stratotype sections and points (GSSPs) for the bases of the Triassic stages. Thus, recently proposed Triassic timescales (Lucas, 2010a; Ogg, 2012) are essentially the STS timescale. Here, I discuss the status of the Triassic timescale and reject both conodont biostratigraphy and the GSSP methodology for timescale definition. Instead, I advocate a return to using ammonoid biochronological events to define Triassic chronostratigraphic boundaries. I also provide a critical evaluation of the state of Triassic radioisotopic ages, magnetostratigraphy and an astronomically-calibrated Triassic timescale. I thus present a new Triassic timescale that incorporates a modified chronostratigraphy based on ammonoid biochronological events with what I regard as reliable radioisotopic ages of most Triassic stage boundaries and a multichron-based step toward a Triassic geomagnetic polarity timescale. REJECTION OF THE GSSP METHODOLOGY What has been referred to by some as the Hedbergian stratigraphy is the method by which the geological timescale has been defined (actually redefined) since the 1970s. The method focuses on agreeing on GSSPs---global stratotype sections and points---to define stage boundaries and was driven largely by the efforts of Hollis Hedberg ( ), who was chairman of the International Commission on Stratigraphy. Walsh et al. (2004) provide a clear and brief review of the history of the GSSP concept and methodology. Hedberg argued for the definition of the limits of time stratigraphic units in specific rock successions. This evolved into the current method of defining the beginning of a time interval by fixing a base to a chronostratigraphic unit (stage) at a single point in a single section---the GSSP. Prior to the GSSP method, time boundaries were either defined by equating them to major physical events (typically expressed as unconformities) or to significant biological (evolutionary) events, which I will refer to as biochronological events. All three methods have strengths and weaknesses. The method of equating time boundaries to unconformities came first and produced readily correlateable boundaries. But, these boundaries often embodied a hiatus of variable duration, and the method was justifiably rejected because it produced diachronous and discontinuous boundaries. The biochronological method also has a long history, for, as Schindewolf (1970, p. 18) well observed, competent stratigraphy is only possible with the help of fossils, i.e., on the basis of irreversible organic evolution. Almost all Phanerozoic chronostratigraphic boundaries have long been defined by biochronological events. However, as is well known, biochronological events are subject to diachroneity (nothing happens instantly and globally) and also subject to debate (what is the most significant event?) and can be destabilized by discovery (range extension). The GSSP method, which is also largely based on biochronological events, has been claimed to be superior to both methods. GSSP definitions of stage bases now focus on one biotic event, marked by the FAD (first appearance datum) of a single taxon, in order to unambiguously identify a single point in time at a single location. Unfortunately, the focus on fixing a boundary in one place at one point has brought much politics into the discussion, as many stratigraphers have a favorite place to fix the boundary, usually the place they have worked at and too often in their own country. Furthermore, the criterion to locate the boundary also can be mired in the politics of specialties: ammonoid biostratigraphers want ammonoid-based boundaries, conodont specialists want conodontbased boundaries, etc. And, amazingly, many now believe that GSSPs are fixed points that will not move even in the face of new data and analyses

2 (see, for example, the quote at the beginning of this article from an article co-authored by former ICS Chairman Gradstein). In so believing, they have relegated GSSPs to the status of non-scientific results (see the quote by Popper at the beginning of this article). The GSSP method fails primarily because of its quixotic goals, beyond realization, and its reliance on a process confounded by politics. Its claims to provide greater precision and stability to timescale definition than other methods are false. The fact is that many GSSPs still have not been agreed on after nearly a half century of debate, and many of those that received early acceptance, such as the stage boundaries of the Silurian System, are now being redefined. The failings of the Hedbergian stratigraphy are many, and deserve a much longer essay than provided here. Suffice it to say that I reject the GSSP method and advocate a return to agreed on biochronological events as the bases of chronostratigraphic definition (see Teichert, 1958 for a good explanation of the conceptual basis of this approach). This is not to say I am averse to fixing boundary stratotypes, but that process as an end in itself appears to have failed. Therefore, I return to identifying ammonoid biochronological events as the criteria to define Triassic stage boundaries. If we can agree on these events (and I acknowledge that this will not always be easy), then we will have a workable timescale that we can then improve further by studying the synchrony/diachroneity and the correlateability of those events. REJECTION OF CONODONT BIOSTRATIGRAPHY The Triassic chronostratigraphic scale was built on ammonoid biostratigraphy. However, in the 1990s, a movement to define Triassic stage boundaries with conodonts began, and at present the FADs of conodont taxa will probably define as many as five stage boundaries (bases of the Induan, Olenekian, Anisian, Norian and Rhaetian). I view this as problematic because it broke with more than a century of practice (see especially Mojsisovics et al., 1895) that relied on a Triassic chronostratigraphic scale based on ammonoid-defined boundaries (which are not always the same as the conodont-defined boundaries). By thus abandoning priority, conodont-based definitions have not served the stability of the Triassic timescale. Furthermore, I see various problems with conodont-based Triassic biostratigraphy, including: (1) the relative youth of Triassic conodont taxonomy, which remains unstable for many taxa; (2) reworking of conodonts, which is not easily recognized and rarely addressed (Macke and Nichols, 2007); (3) problems of facies restrictions, diachroneity and provinciality, which do affect Triassic conodont distributions (Clark, 1984); and (4) the invisibility of conodonts on outcrop, so that they cannot be used in the field to determine the ages of strata. As an example of the drawbacks of using conodonts to define Triassic stage boundaries, consider the article by Orchard (2013) in this volume on the conodonts across the Carnian-Norian boundary at the Black Bear Ridge section in British Columbia, Canada. Study of conodonts from this GSSP candidate section began in the 1980s, and Orchard s article well reveals the differing and frequently changing taxonomic concepts and stratigraphic ranges of these conodonts (also see Mazza et al., 2011). I believe it is premature to use these conodonts to define a Norian base, as I doubt that this will produce a stable and long agreed on boundary. The ammonoid-based Triassic timescale is underpinned by nearly two centuries of collecting and taxonomic work (e.g., Balini et al., 2010). It provides macrofossil-based age assignments that can be used in the field where ammonoids occur and suffers from very few reworking issues. Moreover, the species-level evolution through time of ammonoids is commonly obvious from the morphology of their shells, which record their entire life history, whereas conodont evolution is necessarily interpreted from morphological change in a particular element among the tooth-like elements that constitute their only commonly fossilized record. To a large extent, Triassic conodont-based GSSPs were an answer to longstanding disagreements over taxonomy and correlation among 367 ammonoid specialists. They were also an important part of developing an integrated chronostratigraphic scale. As a relatively newly studied taxonomic group, Triassic conodonts did not have the perceived excess baggage of ammonoids a long history of taxonomic changes and disagreements, known provinciality (Tozer 1981; Dagys 1988) and the demonstrably diachronous distributions of some taxa. Furthermore, the ubiquity and perceived cosmopolitanism of conodonts as well as the retirement in the 1990s and subsequent demise of the main Triassic ammonoid workers, coupled with the rising ranks of younger conodont workers, fueled the rise of Triassic conodont biostratigraphy. Nevertheless, future studies of Triassic conodonts will reveal that they, too, have all of the excess baggage of the ammonoids and are not inherently superior biostratigraphic tools with which to refine Triassic chronostratigraphy. I thus reject conodont biostratigraphy as a basis for defining Triassic chronostratigraphic units and advocate using ammonoid biochronological events to define all Triassic stage boundaries (Fig. 2). TRIASSIC CHRONOSTRATIGRAPHY Current Triassic Chronostratigraphic Scale I recently reviewed the nearly two-century-long development of the Triassic chronostratigraphic scale (Lucas, 2010b), which is now a hierarchy of three series, seven stages and 15 substages (Fig. 1). The first geological studies of Triassic rocks began in Germany in the late 1700s and culminated when Alberti (1834) coined the term Trias for the Bunten Sandstein, Muschelkalk and Keuper of southwestern Germany, an ~1 km thick succession of strata between the Zechstein (Permian) and the Lias (Jurassic). Recognition of the Trias outside of Germany soon followed, and by the 1860s Austrian geologist Edmund von Mojsisovics began constructing a detailed Triassic chronostratigraphy based on ammonoid biostratigraphy. In 1895, Mojsisovics and his principal collaborators, Wilhelm Waagen and Carl Diener, published a Triassic timescale that contains most of the stage and substage names still used today (Mojsisovics et al., 1895). Tozer (e.g., 1965, 1967, 1984, 1994; also see Silberling and Tozer, 1968) proposed a Triassic ammonoid-based timescale based on North American standards, particularly in the Canadian Arctic islands and the Cordillera of British Columbia and Nevada. Distinctive features of Tozer s timescale included proposal of four Lower Triassic stages (Griesbachian, Dienerian, Smithian and Spathian) and abandonment of the Rhaetian as the youngest Triassic stage. The STS began its work in the 1970s and now recognizes seven Triassic stages in three series (Fig. 1). The 1990s saw the rise of Triassic conodont biostratigraphy so that five agreed on (or nearly agreed on) Triassic GSSPs use conodont events as defining features. However, after more than 40 years of work, the STS has only achieved ratification of the bases of three Triassic stages defined by GSSPs: 1. The base of the Induan Stage (= base of Triassic, = base of Lower Triassic) is defined by the FAD of the conodont Hindeodus parvus at the Meishan section in Guangxi, southern China (Yin et al., 2001). 2. The base of the Olenekian Stage may be defined by the FAD of a conodont at the Mud section in Spiti, India (Krystyn et al., 2007a), but this is still under discussion. Ironically, the Mud section has an outstanding ammonoid record, first monographed by Diener (1897) and Krafft and Diener (1907), so why is a conodont-, instead of an ammonoid-based datum to be chosen here for definition? 3. The base of the Anisian Stage (= base of the Middle Triassic) may be defined by the FAD of a conodont at the Desli Caira section in Romania (Orchard et al., 2007; Gradinaru et al., 2007). But, like the Mud section, the Desli Caira section also has an excellent ammonoid record that is not being used for GSSP definition. 4. The base of the Ladinian Stage is defined by the FAD of the ammonoid Eoprotrachyceras curionii at the Bagolino section in Italy (Brack et al., 2005).

3 368 FIGURE 1. The Triassic chronostratigraphic scale (from Lucas, 2010b). 5. The base of the Carnian Stage (= base of the Upper Triassic) is defined by the FAD of the ammonoid Daxatina canadensis at the Stuores Wiesen section in Italy (Mietto et al., 2012). 6. The base of the Norian Stage will apparently be defined by a GSSP located either at Black Bear Ridge in British Columbia, Canada or at Pizzo Mondello in Sicily, and it probably will be based on a conodont event close to the base of the Stikinoceras kerri ammonoid zone, which has been the traditional Norian base in North American usage (Orchard, 2010, 2013). 7. The base of the Rhaetian Stage is proposed to be defined by the FAD of the conodont Misikella posthernsteini at the Steinbergkogel section in Austria (Krystyn et al. 2007b), yet this level is essentially the same as the FAD of the ammonoid Paracochloceras suessi, so why not use the ammonoid FAD to define the base of the Rhaetian? 8. The base of the Hettangian Stage (= base of the Jurassic, = base of the Lower Jurassic) is defined by the FAD of the ammonoid Psiloceras spelae at the Kuhjoch section in Austria (Von Hillebrandt et al. 2007; Morton, 2012). These GSSPs define boundaries of the seven Triassic stages recognized by the STS and also define the boundaries of the three Triassic Series and of the Triassic System (Fig.1). However, progress on such definitions has been painfully slow--only three in more than 40 years-- so apparently the STS averages about years per GSSP. Thus, we can expect to wait for at least another half century before the process is complete! Triassic Series I think the most significant thing we have learned from numerical chronology about the Triassic timescale is how uneven in duration the three traditional Triassic series are (see below). The traditional Early Triassic is about 5 million years long, the traditional Middle Triassic is about 10 million years long and the rest of the Triassic (the traditional Late Triassic) is an amazingly 36 million years long! It makes no sense to continue to divide the Triassic into three such unequal series. Indeed, by numerical chronology, the so-called Early and Middle Triassic together make up only about the first third of the period. Therefore, I advocate recognizing four Triassic Series (Epochs) (Fig. 2). Note that Mojsisovics et al. (1895) also divided the Triassic into four series similar to (but not exactly congruent with) those I recognize here. The four Triassic Series already have names: Scythian, Dinarian, Carnian and Norian. The first two names are from Mojsisovics et al. (1895), and the last two are elevation of the very long Carnian and Norian stages to series rank. Some may argue that Dinarian is almost a homophone of Dienerian, so the alternative series name Muschelkalk (or Muschelkalkian) could be used. The Carnian Series includes two stages (Julian and Tuvalian), though recognition of the Cordevolian as a third, lowermost Carnian Stage merits new discussion. The Norian Series encompasses four stages: Lacian, Alaunian, Sevatian and Rhaetian. It could also be argued that the four series can be justified as bounded by major ammonoid biochronological events. Thus, the Scythian begins with the mass extinction of ammonoids at the beginning of the Triassic and the survival of the otoceratids. The Dinarian begins with a major evolutionary turnover in ammonoid eviolution with extinctions and the appearance of many new taxa. The Carnian begins with the rise to dominance of the trachyceratines and sirenitines, and the Norian begins with the extinction of the tropitids and the appearance of new juvavitines and thisbitids. Arguably, the ammonoid events that mark the beginnings of the Carnian and Norian are not as significant as those that mark the beginnings of the Scythian and the Dinarian. Thus, the argument for a Triassic divided into four series is rooted more in numerical chronology than in biochronology. Scythian Series Short stages are superior to long stages because they discriminate shorter time intervals, one of the primary features of a good timescale. The Scythian can be divided into four stages proposed by Tozer, so why settle for two time intervals, Induan and Olenekian, especially given the lengthy (and seemingly endless) disagreements about where and how to define the base Olenekian GSSP? Arguments for a fourfold division of the Lower Triassic have long been rooted in recognizing key ammonoid biotic events that have demonstrable global correlateability (e. g., Tozer, 1978). The beginnings of the four ages of the Scythian can be defined by biochronological events based on ammonoid evolution. Thus, the Griesbachian begins with the FAD of Otoceras; Otoceras is one of two ammonoid genera that survived the end-permian mass extinction. The base of the Dienerian is marked by the appearance of abundant meekoceratids, a globally recognizable event, here defined by the FAD of Proptychites candidus. The base of the Smithian is marked by the appearance of many new ammonoid taxa, such as Flemingites, Kashmirites and Hedenstroemia. This was accompanied by formation of a pronounced latitudinal diversity of ammonoids (Brayard et al., 2006, 2007). The Smithian base can be defined by the FAD of Hedenstroemia hedenstroemi.

4 The base of the Spathian is also marked by the appearance of many new ammonoid taxa, notably the dinaritines, tirolitines and columbitids. The disappearance of typical Smithian genera (Anasibirites, Wasatchites, etc.) also marks the base of the Spathian. In other words, a major ammonoid extinction occurred at the end of the Smithian/beginning of the Spathian and was followed by a rather rapid major evolutionary radiation (Tozer, 1982; Galfetti et al., 2007). This event corresponds to a major perturbation of the carbon cycle that has been interpreted as a change from warm and equable global climate (Smithian) to a latitudinally-differentiated global climate (Spathian) (Galfetti et al., 2007). I provisionally define the Spathian base by the FAD of Subolenekites pilaticus, though work underway promises to redefine the Spathian base with an older ammonoid event (e.g., Guex et al., 2010). For the most part I am simply advocating a return to Tozer s original definitions of these stages pending further study. Note that the beginning of the Griesbachian predates the conodont-based GSSP for the beginning of Triassic time. This will require moving the base of the Triassic downward, back to its pre-conodontdefined base. Dinarian Series The Dinarian Series seems to be the most stable part of the Triassic chronostratigraphic scale. Anisian and Ladinian are relatively short stages readily divided into six substages. Longstanding effort by the STS to define the Anisian base by a GSSP in Romania based on the FAD of the conodont Chiosella timorensis has been confounded by discovery of Spathian records of this species. The base of the Anisian is marked by a major ammonoid turnover. Most Spathian genera disappear at the beginning of the Anisian, and a variety of taxa (Paracrochordiceras, Japonites, Gymnites as well as danubitids, longobarditids and cladiscitids, among others) first appear during the latest Spathian or at the Anisian base (Tozer 1981, 1984; Brayard et al., 2006). The FAD of the ammonoid Japonites welteri has long been available as a way to define the beginning of the Anisian. The Ladinian begins with the diversification of trachyceratid ammonoids and its base has been defined by the FAD of the oldest trachyceratid, Eoprotrachyceras curionii (Brack et al., 2005) In this case, the GSSP method chose a significant ammonoid biochronological event for chronostratigraphic definition. Carnian and Norian Series Mietto et al. (2012) recently published a ratified GSSP for the base of the Carnian (base of the Julian) the FAD of the ammonoid Daxatina canadensis at the Stuores Wiesen section in northeastern Italy. The Carnian base was long the FAD of Trachyceras aon, and I see no reason why the FAD of Trachyceras aon could have been used to define the beginning of the Carnian. I find it interesting that Mietto et al. (2012, p ) state that the co-occurrence of Trachyceras and Daxatina allows the D. canadensis zone to be considered Carnian and that they somehow know that D. canadensis has a globally synchronous first appearance. However, defining the base of the Carnian using the FAD of Daxatina canadensis is at a point only a little below the FAD of Trachyceras aon, and this is a second point in the Triassic chronostratigraphic scale where the GSSP method has arrived at what I considered a useful boundary definition. In general, the Julian is dominated by Trachyceratinae, in particular Trachyceras and Austrotrachyceras, and by Sirenitinae. The base of the Tuvalian is marked by one of the major changes in the evolution of Triassic ammonoids, namely the near extinction of the Trachyceratinae, as well as the radiation of Tropitidae (e.g., Tropites and closely allied forms) and to a lesser extent Arpaditinae. For this reason, Tozer (1984) argued that it might be better to define the Carnian base at the base of the Tuvalian. Definition of the Norian base by the GSSP method remains plagued 369 FIGURE 2. The new Triassic timescale advocated here (see text for discussion). by problems of conodont taxonomy and biostratigraphy and mired in STS politics. Two sections are the leading candidates---black Bear Ridge in western Canada and Pizzo Mondello in Italy. Both sections have relatively poor ammonoid records but good conodont records, which underscores that their candidacy has been heavily colored by the advocacy of conodont-based definitions. The choice of a conodont-based GSSP for the Norian base has been delayed for years by changing stratigraphic ranges and the fluid taxonomy of the relevant conodonts (QED: Orchard, 2013). The traditional boundary definition was by the FAD of an ammonoid, Stikinoceras kerri, which is used here (Fig. 2). Indeed, the base of the Norian and of the Lacian is characterized by major ammonoid biochronological events: the nearly complete disappearance of Tropitidae and the appearance of new members of Juvavitinae, such as Guembelites and Dimorphites, and of the Thisbitidae such as Stikinoceras. The base of the Alaunian is marked by the appearance of new genera of Cyrtopleuritidae (Drepanites and Cyrtopleurites). Members of this family (including Himavatites, Mesohimavatites, Neohimavatites)

5 370 together with some Haloritinae, such as Halorites and Thisbitidae, such as Phormedites, characterize the Alaunian. The base of the Sevatian is characterized by a decrease in ammonoid diversity and the first heteromorphic ammonoid, Rhabdoceras. Common Sevatian ammonoids are Haloritinae (Gnomohalorites and Catenohalorites) and Sagenitidae (Sagenites ex gr. S. quinquepunctatus). Here, I define the bases of the Alaunian and Sevatian by traditional ammonoid FADs (Fig. 2). The proposed definition of a GSSP for the base of the Rhaetian is at the classic Steinbergkogel section in Austria based on the FAD of the conodont Misikella posthernsteini (Krystyn et al., 2007b). As already stated, the FAD of the ammonoid Paracochloceras suessi is essentially the same level and very correlatable, so it can readily replace the conodont-based boundary criterion (Fig. 2). The base of the Jurassic is defined by the FAD of the ammonoid Psiloceras spelae at Kuhjoch, Austria. As a direct participant in the process of selecting this GSSP, I can attest that politics (of the Subcommission on Jurassic Stratigraphy: SJS) trumped agreed on procedures and good scientific judgment. Thus, after years of study and political wrangling, the SJS Working Group tasked to recommend a GSSP agreed on four candidates for the GSSP location New York Canyon in Nevada, USA; Kunga Island in British Columbia, Canada; the Utcubamba Valley in Peru; and St. Audrie s Bay in the United Kingdom (see Lucas et al., 2007 for a review of these). However, in 2006, when it became clear that the European GSSP candidate (St. Audrie s Bay) was much inferior to the New World candidates, new European GSSP candidates, including Kuhjoch, were added to the slate by the European-dominated working group, long after the choice of four had been made. Dominated by ammonoid specialists, the vote for the GSSP criterion swept aside other criteria, which were a radiolarian biostratigraphic datum or a carbon-isotope excursion. However, the voting group, which was dominated by Europeans, chose a section in Europe, at Kuhjoch in Austria, about which nothing had been published prior to its proposal as a GSSP. Furthermore, the Kuhjoch GSSP is a man-made trench through a saddle in a strike valley in the Austrian Alps (see Morton, 2012 for a photograph), chosen over the much better exposed, more extensive and much more fossiliferous outcrops of the New World candidates. I realize that some will view these criticisms as little more than the complaints of one who voted for a GSSP location not chosen. However, the first important point here is that the choice to use the FAD of the oldest psiloceratid ammonoid to define the Triassic-Jurassic boundary was a good decision; it respected longstanding tradition and identified a significant biochronological event to define the system boundary. The second point is that the choice of the location of the GSSP was a political one that adds little to correlateability of the boundary. Clearly, timescale research needs to depart from the politics of choosing GSSPs. RADIOISOTOPIC AGES Numerical calibration of the Triassic timescale still remains woefully incomplete and imprecise, especially for the Late Triassic. The most recent compilation (Mundil et al., 2010), some new dates (see Ogg, 2012) and a critical re-evaluation and rejection of the long Norian (Lucas et al., 2012) provide the basis for the numerical calibration presented here (Fig. 3). Some of the problems posed for numerical calibration of the Triassic timescale are the same as those that face Triassic magnetostratigraphy (see below). These include the black box nature of laboratory analysis, so that the lab result is not readily evaluated and rarely replicated; disagreements about laboratory standards (including decay constants); changing technology; and differing standards (or thresholds) for how many and what kinds of data support a reliable numerical age determination. All of these factors should make it difficult to accept any newly published numerical age on face value. Indeed, it is important to see numerical ages for exactly what they are approximations that will change with new data and new technology. However, because they are simple to understand as just numbers, many readily embrace newly published numerical ages as reliable and often live to rue their decisions. If we compare the Mundil et al. (2010) compilation to an earlier, authoritative compilation by Forster and Warrington (1985), the most important result of Triassic numerical chronology has been to show the uneven duration of the three traditional Triassic series, with the Late Triassic representing about two-thirds of Triassic time (see comments above). Logically, it is worth using that knowledge to argue for eliminating the division of the Triassic into three very unequal series, as I have done here (Fig. 2). The few Triassic numbers that do calibrate the Triassic timescale are of varied reliability: 1. To calibrate the beginning and end of the Triassic, enough numbers are known and have been discussed, debated and evaluated for long enough (at least a decade) that concluding that the Triassic began about 250 Ma (more precisely Ma) and ended about 200 Ma (more precisely Ma) seem highly reliable statements. These numbers are unlikely to change by 1% or more. 2. There are a few ages that suggest a Dienerian-Smithian boundary of ~ 251 Ma. A Smithian-Spathian boundary at ~ 249 Ma is an estimate. 3. Numerous Dinarian ages well-tied to marine biostratigraphy have been known since the 1990s, especially for the Anisian. Calibration of the beginning of the Anisian at ~247 Ma and the beginning of the Ladinian at ~ 242 Ma seems almost as reliable as the ages of the Triassic system boundaries. 4. Radioisotopic ages for the Carnian-Norian are scarce, and the only reliable and biostratigraphically controlled age is from a Carnian (lower Tuvalian) tuff dated at Ma (Furin et al., 2006). A wealth of detrital zircon ages from nonmarine strata of the Chinle Group in the western USA are less precise because they are maximum ages of the sediments (based on reworked zircons) and because of debate and imprecision relating them to marine biochronology. Lucas et al. (2012) argued that these detrital zircon ages constrain the beginning of the Norian to ~ 221 Ma. 5. The age of the Pedazzo granite and its inferred relationship to marine biostratigraphy supports a Carnian base at ~ 237 Ma (Ogg, 2012), but I consider this an estimate of relatively low reliability. 6. Ages of the bases of the former Carnian and Norian substages cannot be reliably estimated. The Furin et al. (2006) age puts the Tuvalian base close to 231 Ma. The other bases (of the Lacian, Alaunian and Sevatian) cannot be estimated with any precision, and I simply use the estimates of Ogg (2012) here. 7. Evidence that the Rhaetian is a relatively short stage (~ 4 million years long) is complex to evaluate, involving an assessment of Newark cyclostratigraphy, diverse biostratigraphy and detrital zircon ages in nonmarine strata (Lucas et al., 2011, 2012). I estimate the Rhaetian base as ~ 205 Ma based on such assessment, but this is one of the least reliable estimates in the Triassic numerical timescale advocated here (Fig. 2). MAGNETOSTRATIGRAPHY The global polarity timescale for rocks of Late Jurassic, Cretaceous and Cenozoic age provides a valuable tool for evaluating and refining correlations that are based primarily on radioisotopic ages or biostratigraphy. However, there is no agreed on geomagnetic polarity timescale (GPTS) for the Triassic, and my assessment is that current Triassic magnetostratigraphy has been more of a hindrance than a help to timescale definition and correlation. The reasons are several and include inconsistent results from different laboratories, poor age control of many Triassic magnetostratigraphic sections and simple miscorrelation (or unsupportable correlation) of Triassic magnetostratigraphy, such as that which created the long Norian. Hounslow and Muttoni (2010) provided a comprehensive review of Triassic magnetic polarity history. I rely on this review and some

6 FIGURE 3. Magnetostratigraphic correlations of the Pizzo Mondello and Newark sections. On the left, the correlation matches the marine and nonmarine, biostratigraphically-determined Carnian-Norian boundary. On the right is the pattern matched correlation of Muttoni et al. (2004), which became the basis of the long Norian (from Lucas et al., 2012). 371

7 372 more recent data and reappraisals (e.g., Lucas et al., 2011, 2012) and also emphasize the multichron concept of Lucas (2011), which recognizes intervals of dominant polarity rather than individual polarity chrons. The reason for this is that I believe we are a long way from a wellestablished succession of Triassic polarity chrons that can receive numbers (or names), like those of the Late Cretaceous-Cenozoic GPTS. We do, however, at least seem to know the polarity of each of the Triassic stage boundaries with some confidence (Hounslow and Muttoni, 2010). One of the largest hindrances to developing a Triassic GPTS is the polarity record of the Newark Supergroup in eastern North America, which has confounded all attempts to correlate it to other Late Triassic magnetostratigraphic records (Fig. 3). Given the great thickness of the Newark section (~ 4 km of section is equivalent to much of the Late Triassic), it arguably captures a more complete polarity history than do the much thinner marine sections in Europe for which a magnetic polarity record is available. That, however, is the only thing to recommend the Newark magnetic polarity record, because age control of this record is highly problematic. Thus, for example, the Triassic-Jurassic boundary was located incorrectly in the Newark, below the CAMP basalt sheets, for decades; this has only recently been corrected (Kozur and Weems, 2005, 2007, 2010; Lucas and Tanner, 2007; Lucas et al., 2011). Biostratigraphic placement of the Carnian-Norian boundary in the Newark (near the base of the Passaic Formation) is based on reinforcing correlations to the Germanic Basin section from palynomorphs, conchostracans and vertebrate biostratigraphy (Lucas et al., 2012). Abandonment of this boundary was based on a demonstrably incorrect correlation of magnetostratigraphy in a marine section at Pizzo Mondelo in Italy with the Newark and, coupled with a supposed astronomicallycalibrated timescale based on Newark cyclostratigaphy, created the notion that the Carnian-Norian boundary is at about 228 Ma, the so-called long Norian (Muttoni et al., 2004). Correct placement of the Carnian- Norian boundary in the Newark section means it and the beginning of the Jurassic are the only reliable biostratigraphic tiepoints for the Newark magnetic polarity stratigraphy. Placement of any divisions of the Carnian and Norian, including identification of the base of the Rhaetian, are difficult to impossible in the Newark section. Recent magnetostratigraphic correlations to a Rhaetian base in the Newark (Muttoni et al., 2010; Hüsing et al., 2011) are based on the same kinds of unsupportable magnetostratigraphic correlations that produced the long Norian, and produce a long Rhaetian of as much as 11 million years duration. From its initial publication, no convincing correlation of the Newark magnetostratigraphy to broadly correlative magnetostratigraphies could be made, simply because it contains approximately 10 times the number of reversals found in correlative marine sections (Fig. 3). Given what I call the rubber ruler effect---sedimentation rate stretches or contracts magnetic polarity chron thicknesses so that matching patterns is highly problematic---and the lack of biostratigraphic tiepoints, how could any unambiguous correlation of the Newark magnetostratigraphy be made to other polarity stratigraphies? And, why use the Newark polarity history as the standard column for the Late Triassic if nothing else can be correlated to it? Indeed, attempts to correlate the Newark polarity record to broadly co-eval records have produced a fractious literature with little agreement on what correlations are reliable. Both Hounslow and Muttoni (2010) and Ogg (2012) have presented the Solomonesque solution of advocating at least two correlations (notably long Carnian and long Norian ), neither of which is defensible (Lucas et al., 2012). The Triassic magnetic polarity timescale I advocate is a set of multichrons (Fig. 2). It is far less detailed that the provisional Triassic GPTS of Hounslow and Muttoni (2010). But, I believe this is a realistic abstraction of what we now reliably know about the Triassic GPTS. Much more needs to be understood before Triassic magnetic polarity history becomes an important part of Triassic correlation and timescale definition. TRIASSIC ASTRONOMICAL TIMESCALE Because the Earth s astronomical parameters have a predictable periodicity, they can be used, in theory, for geochronometric calibration of the geological timescale. However, to do so, we must identify a cyclostratigraphy in which the sedimentary cycles are forced by astronomical parameters with known periods, and demonstrate that these astronomical cycles are recorded faithfully and completely (Hinnov and Hilgren, 2012). Such a cyclostratigraphy-based numerical timescale, called the astronomical timescale (ATS), has been reasonably well-established for much of Cenozoic time, from the beginning of the Oligocene (~ 34 Ma) to the present. Older parts of the timescale have less complete, disconnected cyclostratigraphies that have been referred to as floating astrochronologies (e.g., Hinnov and Ogg, 2007). The Newark cyclostratigraphy has been identified as one such floating astrochronology capable of providing a high resolution geochronometry for most of the Late Triassic and the older part of the Early Jurassic (e.g., Kent and Olsen, 1999; Ogg, 2012). However, it and other Triassic cyclostratigraphies are fraught with problems. Tanner (2010) reviewed the use of cycles in Triassic stratigraphy, and he noted that high frequency (fourth and fifth-order) cyclicity is a common feature of sedimentary sequences in Triassic depositional settings. Tectonism and autocyclicity clearly drove some of this cyclicity, but many Triassic cycles have been attributed to orbitally-forced variations in solar insolation at the Milankovitch frequencies--the precession, as well as the short and long eccentricity cycles--at scales of tens of thousands to hundreds of thousands of years. This orbital forcing is thought to have controlled sedimentation through periodic changes in climate or sea level. Examples of interpreted Milankovitch-frequency cyclicity throughout the Triassic record include much of the Germanic Triassic section, the Newark Supergroup of eastern North America, and parts of the Alpine Triassic. The cyclostratigraphy of these sections has been used as a tool for intrabasinal correlation and for chronostratigraphy. However, conceptual arguments and radioisotopic age data call all of these interpretations into question. Indeed, Triassic cyclostratigraphic studies remain far from the goal of developing a reliable, astronomicallycalibrated Triassic timescale. In particular, the reliability of the cyclostratigraphy of the Upper Triassic of the Newark basin is contradicted by biostratigraphy that indicates a lack of completeness of the Newark stratigraphic record (e.g., Kozur and Weems, 2010; Lucas et al., 2011, 2012). Therefore, this so-called floating astrochronology should sink from consideration. A TRIASSIC TIMESCALE OF THE FUTURE? The Triassic timescale proposed here (Fig. 2) is a way forward that I see as the next logical step toward a timescale that best resolves Triassic time with available data. It embodies the following: 1. Recognition that the three traditional Triassic series very unevenly divide the system. Therefore, the Triassic is divided into four series (epochs). 2. Ammonoid biochronological events are used to define Triassic chronostratigraphic divisions. Conodonts are not used for timescale definition. 3. The Scythian Series is divided into four stages Griesbachian, Dienerian, Smithian and Spathian. Induan and Olenekian are abandoned. 4. Carnian and Norian are elevated to series rank. Their subdivisions (substages) are elevated to stage rank, and the Rhaetian is regarded as the highest Triassic stage within the Norian Series. 5. The numerical age estimates of the stage boundaries are of variable precision and reliability as discussed earlier. 6. The Triassic GPTS used here is a set of multichrons. My expectation is that those wedded to the current GSSP method of Triassic timescale definition will object to some or all of what has been

8 advocated here. Nevertheless, I challenge all interested in real progress towards an improved Triassic timescale to consider the ideas presented here as a way to move towards that goal. ACKNOWLEDGMENTS 373 I owe a heavy debt to many scientists who I met and (in some cases) collaborated with during the last 20 plus years of my participation in the STS. They taught me much about the Triassic. Adrian Hunt, Jim Jenks, Kathy Nichols and Lawrence Tanner provided helpful reviews of the manuscript. I dedicate this article to the memory of Tim Tozer and Norm Silberling, who, in many ways, got it right decades ago. REFERENCES Alberti, F. von, 1834, Beitrag zu einer Monographie des Bunten Sandsteins, Muschelkalks und Keupers, und die Verbindung dieser Gebilde zu einer Formation. Verlag der J. G. Cotta sschen Buchhandlung, Stuttgart und Tubingen, [Facsimile reprinted in 1998 by the Friedrich von Alberti- Stiftung der Hohenloher Muschelkalkwerke, Ingelfingen, Germany]. Balini, M., Lucas, S.G., Jenks, J.F. and Spielmann, J.A., 2010, Triassic ammonoid biostratigraphy: An overview; in Lucas, S.G., ed., The Triassic timescale: Geological Society, London, Special Publication 334, p Brack, P., Rieber, H., Nicora, A. and Mundil, R., 2005, The global boundary stratotype section and point (GSSP) of the Ladinian Stage (Middle Triassic) at Bagolino (southern Alps, northern Italy) and its implications for the Triassic time scale: Episodes, v. 28, p Brayard, A., Bucher, H., Escarguel, G., Fluteau, F., Bourquin, S. and Galfetti, T., 2006, The Early Triassic ammonoid recovery: Paleoclimatic significance of diversity gradients: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 239, p Brayard, A., Escarguel, G. and Bucher, H., 2007, The biogeography of Early Triassic ammonoid faunas: Clusters, gradients, and networks: Geobios, v. 40, p Clark, D.L., ed., 1984, Conodont biofacies and provincialism: Geological Society of America, Special Paper 196, 340 p. Dagys, A.S., 1988, Major features of the geographic differentiation of Triassic ammonoids; in Wiedmann, J. and Kullmann, J., eds. Cephalopods past and present: Stuttgart, E. Schweizerbart sche Verlagsbuchhandlung, p Diener, K., 1897, Himalayan fossils. The Cephalopoda of the Lower Triassic: Palaeontologica Indica, series 5, v. 2, no. 1, 191 p. Forster, S.C. and Warrington, G., 1985, Geochronology of the Carboniferous, Permian and Triassic; in Snelling, N.J., ed., The chronology of the geological record: Blackwell Scientific Publishing, Oxford, p Furin, S., Preto, N., Rigo, M., Roghi, G., Gianolla, P., Crowley, J.L. and Bowring, S.A., 2006, High-precision U-Pb zircon age from the Triassic of Italy: Implications for the Triassic time scale and the Carnian origin of calcareous nannoplankton and dinosaurs: Geology, v. 34, p Gaetani, M., 2009, GSSP of the Carnian Stage defined: Albertiana, v. 37, p Galfetti, T., Hochuli, P.A., Brayard, A., Bucher, H., Weissert, H. and Os Vigron, J., 2007, Smithian-Spathian boundary event: Evidence for global climatic change in the wake of the end-permian biotic crisis: Geology, v. 35, p Gradinaru, E., Orchard, M.J., Nicora, A., Gallet, Y., Besse, J., Krystyn, L., Sobolev, E.S., Atudorei, N-V. and Ivanova, D., 2007, The Global Boundary Stratotype Section and Point (GSSP) for the base of the Anisian Stage: Desli Caira Hill, North Dobrogea, Romania: Albertiana, v. 36, p Guex, J., Hungerbuhler, A. Jenks, J.F., O Dogherty, L. Atudorei, V., Taylor, D.G., Bucher, H. and Bartolini, A., 2010, New Spathian (Lower Triassic) ammonoids from the western USA (Idaho, California, Utah and Nevada): University of Lausanne, Memoires de Geologie, no. 49, 82 p. Hinnov, L.A. and Hilgren, F.J., 2012, Cyclostratigraphy and astrochronology; in Gradstein, F.M., Ogg, J.G., Schmitz, M.D. and Ogg, G.M., The geologic time scale Volume 1: Cambridge University Press, p Hinnov, L.A. and Ogg, J.G., 2007, Cyclostratigraphy and the astronomical timescale: Stratigraphy, v. 4, p Hounslow, M.W. and Muttoni, G., 2010, The geomagnetic polarity timescale for the Triassic: Linkage to stage boundary definitions; in Lucas, S.G., ed., The Triassic timescale: Geological Society London, Special Publication 334, p Hüsing, S.K., Deenen, M.H.L., Koopmans, J.G. and Krijgsman, W., 2011, Magnetostratigraphic dating of the proposed Rhaetian GSSP at Steinbergkogel (Upper Triassic, Austria): Implications for the Late Triassic time scale: Earth and Planetary Science Letters, v. 302, p Kent, D.V. and Olsen, P.E., Astronomically tuned geomagnetic polarity time scale for the Late Triassic: Journal of Geophysical Research, v. 104, p Kozur, H. and Weems, R.E., 2005, Conchostracan evidence for a late Rhaetian to early Hettangian age for the CAMP volcanic event in the Newark Supergroup, and a Sevatian (late Norian) age for the immediately underlying beds: Hallesches Jahrbuch Geowissenschaft, v. B27, p Kozur, H. and Weems, R.E., 2007, Upper Triassic conchostracan biostratigraphy of the continental rift basins of eastern North America: Its importance for correlating Newark Supergroup events with the Germanic basin and the international geologic timescale: New Mexico Museum of Natural History and Science, Bulletin 41, p Kozur, H. and Weems, R.E., 2010, The biostratigraphic importance of conchostracans in the continental Triassic of the northern hemisphere; in Lucas, S.G., ed., The Triassic timescale: Geological Society London, Special Publication 334, p Krafft, A. and Diener, C., 1909, Himalayan fossils. Lower Triassic Cephalopoda from Spiti, Malla Johar and Byans: Palaeontologica Indica, series 15, v. 6, no. 1, 186 p. Krystyn, L., Bhargava, O.N. and Richoz, S., 2007a, A candidate GSSP for the base of the Olenekian Stage: Mud at Pin Valley; Himachal Pradesh (W. Himalaya), India: Albertiana, v. 35, p Krystyn, L., Boquerel, H., Kuerschner, W., Richoz, S. and Gallet, Y., 2007b, Proposal for a candidate GSSP for the base of the Rhaetian Stage: New Mexico Museum of Natural History and Science, Bulletin, 41, p Lucas, S.G., 2010a, The Triassic timescale: an introduction; in Lucas, S.G., ed., The Triassic timescale: Geological Society London, Special Publication 334, p Lucas, S.G., 2010b, The Triassic chronostratigraphic scale: history and status; in Lucas, S.G., ed., The Triassic timescale: Geological Society London, Special Publication 334, p Lucas, S.G., 2011, Multichron: Lethaia, v. 43, p Lucas, S.G. and Tanner, L.H., 2007, The nonmarine Triassic-Jurassic boundary in the Newark Supergroup of eastern North America: Earth Science Reviews, v. 84, p Lucas, S.G., Taylor, D.G., Guex, J., Tanner, L.H. and Krainer, K., 2007, The proposed global stratotype section and point for the base of the Jurassic System in the New York Canyon area, Nevada, USA: New Mexico Museum of Natural History and Science, Bulletin 40, p Lucas, S.G., Tanner, L.H., Kozur, H.W., Weems, R.E. and Heckert, A.B., 2012, The Late Triassic timescale: Age and correlation of the Carnian- Norian boundary: Earth-Science Reviews, v. 114, p Lucas, S.G., Tanner, L.H., Donohoo-Hurley, L.L., Geissman, J.W., Kozur, H.W., Heckert, A.B. and Weems, R.E., 2011, Position of the Triassic- Jurassic boundary and timing of the end- Triassic extinctions on land:

9 374 Data from the Moenave Formation on the southern Colorado Plateau, USA: Palaeogeography, Palaeoecology, Palaeoclimatology, v. 302, p Macke, D.L. and Nichols, K.M., 2007, Conodonts as evolving heavy-mineral grains: New Mexico Museum of Natural History and Science, Bulletin 41, p Mazza, M., Rigfo, M. and Nicora, A., 2011, A new Metapolygnathus platform conodont species and its implications for upper Carnian global correlations: Acta Palaeontologica Polonica, v. 56, p Mietto, P. and 14 co-authors, 2012, The global boundary stratotype section and point (GSSP) of the Carnian stage (Late Triassic) at Prati di Stuores/ Stuores Wiesen section (southern Alps, NE Italy): Episodes, v. 35, p Mojsisovics, E. von, Waagen, W.H. and Diener, C., 1895, Entwurf einer Gliederung der pelagischen Sediments des Trias-Systems: Akademie Wissenschaft Wien, Mathematische-naturwissenschaftliche Klasse Sitzungsberichte, v. 104, p Morton, N., 2012, Inauguration of the GSSP for the Jurassic System: Episodes, v. 35, p Mundil, R., Pálfy, J., Renne, P.R. and Brack, P., 2010, The Triassic time scale: New constraints and a review of geochronological data; in Lucas, S.G., The Triassic timescale: Geological Society London, Special Publication 334, p Muttoni, G., Kent, D.V, Olsen, P.E., Di Stefano, P., Lowrie, W., Bernasconi, S.M. and Hernandez, F.M., 2004, Tethyan magnetostratigraphy from Pizzo Mondello (Sicily) and correlation to the late Triassic Newark astrochronological polarity time scale: Geological Society of America Bulletin, v. 116, p Muttoni, G., Kent, D.V., Jadoul, F., Olsen, P.E., Rigo, M., Galli, M.T. and Nicora, A., 2010, Rhaetian magneto-biostratigraphy from the Southern Alps (Italy): Constraints on Triassic chronology: Palaeogeogrpahy, Palaeoclimatology, Palaeoecology, v. 285, p Ogg, J.G., 2012, Triassic; in Gradstein, F.M., Ogg, J.G., Schmitz, M.D. and Ogg, G.M., eds., The geologic time scale 2012: Amsterdam, Elsevier, p Orchard, M.J., Triassic conodonts and their role in stage boundary definition; in Lucas, S.G., ed., The Triassic timescale: Geological Society London, Special Publication 334, p Orchard, M.J., 2013, Five new genera of conodonts from the Carnian- Norian boundary beds of Black Bear Ridge, northeast British Columbia, Canada: New Mexico Museum of Natural History and Science, Bulletin 61, this volume. Orchard, M.J., Gradinaru, E. and Nicora, A., 2007, A summary of the conodont succession around the Olenekian-Anisian boundary at Desli Caira, north Dobrogea, Romania: New Mexico Museum of Natural History and Science, Bulletin 41, p Popper, K., 1959, The logic of scientific discovery: New York, Basic Books, 458 p. Schindewolf, O., 1970, Stratigraphical principles: Newsletters in Stratigraphy, v. 1, p Silberling, N.J. and Tozer, E.T., 1968, Biostratigraphic classification of the marine Triassic in North America: Geological Society of America Special Paper 110, p Stockar, R., Baumgartner, P.O. and Condon, D., 2012, Integrated Ladinian bio-chronostratigraphy and geochronology of Monte San Giorgio (Southern Alps, Switzerland): Swiss Journal of Geosciences. Tanner, L.H., 2010, Cyclostratigraphic record of the Triassic: A critical examination; in Lucas, S.G., ed., The Triassic Timescale: Geological Society, London, Special Publication 334, p Tozer, E.T., 1965, Lower Triassic stages and ammonoid zones of Arctic Canada: Geological Survey of Canada Paper, Tozer, E.T., 1967, A standard for Triassic time: Geological Survey of Canada Bulletin, v Tozer, E.T., 1978, Review of the Lower Triassic ammonoid succession and its bearing on chronostratigraphic nomenclature: Schriftenreihe Erdwissenschaftlichen Kommissionen Osterreichische Akademie der Wissenschaften, v. 4, p Tozer, E.T., 1981, Triassic Ammonoidea: Geographic and stratigraphic distribution; in House, M.R. and Senior, J.R., eds., The Ammonoidea: Systematics Association Special Volume 18, Academic Press, London, p Tozer, E.T., 1982, Marine Triassic faunas of North America: Their significance for assessing plate and terrane movements: Geologische Rundschau, v. 71, p Tozer, E.T., 1984, The Trias and its ammonoids: The evolution of a time scale: Geological Survey of Canada, Miscellaneous Report 35, 171 p. Tozer, E.T., 1994, Canadian Triassic ammonoid faunas: Geological Survey of Canada, Bulletin 467, 663 p. Von Hillebrandt, A., Krystyn, L. and Kuerschner, W.M., with contributions from Bown, P.R., McRoberts, C., Ruhl, M.S., Simms, M., Tomasovych, A. and Urlichs, M., 2007, A candidate GSSP for the base of the Jurassic in the Northern Calcareous Alps (Kuhjoch section, Karwendel Mountains, Tyrol, Austria: International Subcommission on Jurassic Stratigraphy Newsletter, no. 34, p Walsh, S.L., Gradstein, F.M. and Ogg, J.G., 2004, History, philosophy, and application of the Global Stratotype Section and Point (GSSP): Lethaia, v. 37, p Yin, H., Zhang, K., Tong, J., Yang, Z. and Wu, S., 2001, The global stratotype section and point (GSSP) of the Permian-Triassic boundary: Episodes, v. 24, p