New constraints on the End-Triassic (Upper Norian–Rhaetian) magnetostratigraphy

Earth and Planetary Science Letters 255 (2007) 458 – 470 www.elsevier.com/locate/epsl

New constraints on the End-Triassic (Upper Norian–Rhaetian) magnetostratigraphy Yves Gallet a,⁎, Leopold Krystyn b , Jean Marcoux c , Jean Besse a a

Equipe de Paléomagnétisme, Institut de Physique du Globe de Paris, UMR CNRS 7154, 4 Place Jussieu, 75252 Paris Cedex 05, France b Department of Paleontology, Vienna University, Geozentrum, Althanstrasse 14, A-1090 Vienna, Austria c Equipe de Tectonique, Institut de Physique du Globe de Paris, UMR CNRS 7154, 4 Place Jussieu, 75252 Paris Cedex 05, France Received 6 July 2006; received in revised form 8 December 2006; accepted 1 January 2007 Available online 9 January 2007 Editor: R.D. van der Hilst

Abstract The end-Triassic was marked by one of the five important Phanerozoic global mass extinctions. The construction of a detailed magnetic polarity time scale for this period, that would integrate data from marine and terrestrial realms, is thus of particular interest. We report new magnetostratigraphic data from the Oyuklu section located in southwestern Turkey, which allow one to propose a complete late Upper Norian (Sevatian 2) to Rhaetian magnetic polarity sequence. Two correlations are discussed between the new Tethyan marine sequence and the (continental) magnetic polarity record previously determined from the Newark basin in eastern North America. Both options suggest that the Rhaetian is at least partly missing in the Newark basin, which would reconcile most Late Triassic magnetostratigraphic results and biotic features obtained from marine and continental environments. Following our preferred correlation, the Rhaetian would have a duration as short as ∼2 Myr, and ∼ 4.5 Myr if the Sevatian 2 zone is included as part of the Rhaetian. © 2007 Elsevier B.V. All rights reserved. Keywords: Magnetostratigraphy; Triassic; Rhaetian; Tethys; Newark Basin

1. Introduction During the past 15 years, relatively numerous magnetostratigraphic data were obtained for the Late Triassic from marine and continental sedimentary environments. These results give the opportunity to construct a detailed geological time scale during a period characterized by a high and complex faunal extinction rate, the development of large continental basins and ended at ∼ 200 Ma by the emplacement of the Central Atlantic Magmatic Province (CAMP) that ⁎ Corresponding author. E-mail address: [email protected] (Y. Gallet). 0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2007.01.004

may have caused one of the five important Phanerozoic mass extinctions (e.g. [1,2]). However the integration of data of diverse nature (magnetostratigraphy, evolution of marine and continental faunas, isotope stratigraphy, cyclostratigraphy) is presently hampered by the lack of clear correlation between the magnetostratigraphic results obtained from marine (Tethyan) and terrestrial sequences. This topic has been the subject of a lively debate which led to the relocation the Carnian/Norian boundary in the North American Newark basin (continental) standard sequence [3,5]. No such consensus is achieved yet for the location of the Norian/Rhaetian boundary in the Newark sequence, neither on the completeness of the Rhaetian

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

record, at the top of the Triassic, there. For Krystyn et al. [3] and Gallet et al. [4], the Norian/Rhaetian boundary would be placed very high in the Newark sequence implying a very short or a partly missing Rhaetian. On the other hand, Channell et al. [5] and Muttoni et al. [6] following Kent et al. [7] consider a well recorded and long Rhaetian, with a duration of ∼ 8 Myr. These different views obviously pose a problem when analyzing and interpreting the duration and the abruptness of the Late Triassic biotic turnover or the synchrony between extinction events and the emplacement of the CAMP (e.g. [8–11]). Another problem concerns the controversial chronostratigraphic placement of the base of the Rhaetian due to a conflicting overlap with the top of the Norian stage, in particular the upper zone of the late Norian substage (Sevatian 2). As a result, the various published Late Triassic magnetostratigraphic scales [4–6,12] have adopted different Norian-Rhaetian stage boundaries, a circumstance that has led to serious misunderstanding and misinterpretation of the chronological constraints on the compared magnetic polarity intervals. Only recently, the ammonoid genus Cochloceras (including the closely allied Paracochloceras), a biostratigraphic guide for the Sevatian 2 zone, has officially been choosen as a proxy for a future internationally agreed Norian–Rhaetian boundary [13]. Conodonts of the genus Misikella are named as other boundary options with M. posthernsteini occurring coeval to Paracochloceras or M. hernsteini appearing somewhat earlier [14]. However, a final agreement on the placement of the boundary is not anticipated before 2007/8; we therefore proceed with the traditional ammonoid boundary used in our earlier papers but notice that “our” Sevatian 2 corresponds to the basal Rhaetian of Channell et al. [5] and Muttoni et al. [6]. At least part of the above difficulties could be solved by the acquisition of well-dated marine Rhaetian magnetostratigraphic data that would permit a direct comparison with the Newark magnetostratigraphic sequence. For this reason, we present in this paper the magnetostratigraphy of the Sevatian (Upper Norian) to Rhaetian Oyuklu section from southwestern Turkey. 2. Description of the Oyuklu section The Oyuklu section (N36°50.07′, E32°55.81′) is located ∼ 20 km to the north of the city of Ermenek, on the road to the city of Karaman. The exact place of the section is a road cut opposite to the so-called Kayapınar çesme (çesme meaning spring in Turkish), ∼ 1 km to the west of the Yellibeli Geçidi pass at an altitude of ∼ 1830 m. The name of the section is deduced from the

459

Oyuklu Dağ (dağ meaning peak in Turkish) which is the most prominent close-by located mountain. The region of the Oyuklu section belongs to the Central Taurides which was the subject of a number of geological investigations (Fig. 1; e.g. [15–20]). Several nappes from the Beyşehir–Hoyran nappe system crop out there covered transgressively by the neo-autochtonous Miocene [18]. The studied outcrop is part of a Late Triassic cherty limestone sequence intercalated between the greenish Huğlu tuffites of Carnian age [17] and younger Mesozoic deep water sediments named as Mahmut tepese limestones found in the more westerly region of Bozkır–Hadim [18]. This Mesozoic oceanic margin sequence, which has a thickness up to several hundred metres, is generally known as the Huğlu Unit sensu Özgül [18] and it forms a geographically widely recognized nappe structure (the Bozkır nappe [21]) between Bozkır, Mersin and Malatya. Note that the latter sequence was locally called as the Ihsaniye Unit in the Ermenek region by Andrew and Robertson [19]. Recent road enlargement along the studied site offers excellent outcrop conditions with bed-by-bed exposures over at least 100 m in lateral distance. Starting from a tectonically truncated base, the sampled section has a thickness of 30 m. Laterally expanded faults were only recognized within the two first metres and around section-m.7.5. Special attention was paid to the latter disturbance because of its concomitance with a short magnetic polarity interval (interval F-, see above). This fault, which is parallel to the bedding plane, shows laterally no offset or suppression of beds, making the loss of a significant sequential part rather unlikely. The Oyuklu section provides a continuous endTriassic conodont record of five successive zones spanning the top-Norian and a complete Rhaetian (Fig. 2). Despite the observed disappearance of conodonts, no other fossil evidence can be used for a more precise recognition of the Triassic–Jurassic boundary. A one centimetre thin and discontinuous dark grey boundary clay tops the conodont bearing whitish limestones and separates them sharply from the overlying conodont-free platy brownish to greyish chert-rich rocks. A facies change from carbonatic to more siliceous often sponge spicules bearing environments is a general feature marking the Triassic to early Jurassic transition in many Alpine–Mediterranean deepwater sequences [22,23] and may be seen as confirmation of the adopted boundary. In Oyuklu, firm evidence of Jurassic rocks, however, is only proven 6 m higher than the transition, within the nodular red limestones of ammonitico rosso type containing small and badly preserved, unextractable involute smooth

460

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

Fig. 1. Tectonostratigraphic map of South–Western Turkey after Gutnic [16] and Andrews and Robertson [19]. We also indicate the location of the main sections investigated for magnetostratigraphy (Bolücekatasi Tepe, Kavur Tepe, Kavaalani and Oyuklu, this study [12,28,29]).

ammonites attributable to the genus Phylloceras. Liassic records of ammonitico rosso facies in the western Tethys (Italy, Alps, Greece, Turkey) are generally not older than Sinemurian and the underlying cherty beds may well represent the basal most Jurassic though no exact data regarding their age are available. The good access to the Oyuklu outcrop would allow easy collecting of megafossils. However, the sequence is macroscopically completely unfossiliferous, most probably due to the distal pelagic setting in greater, bathyal water depth. Thin sections are also fossil-poor exhibiting a mudstone texture with sporadic radiolarians or ostracod valves. Dating is therefore based on conodonts, which occur not very frequently but relatively continuously through the section (Fig. 2). Repeated collecting between 1996 and 2005 has resulted in 35 productive samples that on an average weight of 2 (rarely 3) kg have delivered 4 to 25 p-elements of the genera Misikella and for the basal 10 m also of Epigondolella. The p-elements are small to very small with respect to age equivalent records from less deep environments (condensed Hallstatt facies) maybe as a result of reduced food supply to tropical offshore bathyal water. Ramiform conodont elements are rare, often fragmented and only the stratigraphically important genus Oncodella has been considered. Except for a short interval around

section-m.24 where highly fragmented Middle Norian E. cf. postera suddenly appear, heterochronous redeposition of conodonts is not observed and also not expected with respect to the autochthonous character of the sediment and the far-margin basin floor position. 3. Biostratigraphy Taxonomic terminology for conodonts adopted in this paper is for species of the genus Misikella from Kozur and Mock [24] and for Epigondolella (including Mockina) bidentata according to Orchard [25]. Increasing platform and size reduction in the latter species during its phylogenetic end phase (Cochloceras interval) leads to a predominance of small platform — less parvigondolellid forms in Epigondolella unfavourable environments. Those forms have been named Parvigondolella andrusovi Kozur and Mock or Parvigondolella lata Kozur and Mock and are described as diagnostic for a time interval younger than that of E. bidentata. In Epigondolella favourable facies as in Oyuklu “parvigondolellids” are, however, either fully (P. andrusovi) or for a major part (P. lata) time equivalent to E. bidentata and therefore considered here as morphological variants or ecostratigraphic morphotypes of the latter species.

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

461

Fig. 2. Stratigraphy, biostratigraphy and magnetostratigraphy of the Oyuklu section (southwestern Turkey).

Following a basal thrust plane the lowest part of the section is tectonically disturbed leading to a stratigraphic inversion of the first two conodont zones within an interval of two metres (Fig. 2). The bidentata Interval Zone (I.Z.) between ∼ m.1 and m.2 is strongly reduced to less than 1 m and is underlain in faulted contact by one metre of rocks of the younger hernsteini–bidentata I. Z. In normal sequence above, the hernsteini–bidentata I. Z. (m.2–m.3) has again a thickness of 1 m and is characterized by co-occurrence of E. bidentata and M. hernsteini (Fig. 3). Both this zone and the next posthernsteini–bidentata I. Z. (m.3–m.10) are characterized by platform bearing E. bidentata up to the top of the latter zone (sample 96/150). The base of the posthernsteini–bidentata I. Z. is confined to the first appearance (FAD) of M. posthernsteini and is generally marked by a dominance of M. hernsteini over M. posthernsteini. The base of the following posthernsteini–hernsteini I. Z. (m.10–m.16) is drawn

just above the last occurrence (LO) of E. bidentata and shows a distinct frequency change in the represented misikellids, with M. posthernsteini now clearly dominating M. hernsteini by a ratio of 2:1 to 10:1. O. paucidentata is also characteristic for the interval but generally too rare to be used as index form. Cooccurrence of M. rhaetica with M. koessenensis and M. posthernsteini characterizes the rhaetica Range Z. (m.16–m.24.5) whose basal parts yield last and rare representatives of M. hernsteini and O. paucidentata. From m.26 to m.30, M. ultima, together with M. posthernsteini, represents the ultima R.Z. (Fig. 3). This zone has been established in Csövar, Hungary [24] with one more conodont interval (Neohindeodella detrei Z.) of presumed latest Rhaetian or basal Jurassic age still above. Since the conodont record ends with M. ultima just below the boundary clay no space remains for a post — ultima conodont interval in our section.

462

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

4. Paleomagnetic results Paleomagnetic analyses were conducted in the Paleomagnetic laboratory at Institut de Physique du Globe de Paris. Samples were thermally demagnetized in 14 to 16 steps up to 600 °C. Magnetization

measurements were carried out using a 2G three-axis cryogenic magnetometer placed in the shielded room of the laboratory. Examples of thermal demagnetization behavior are shown in Fig. 4. In almost all samples, three magnetic components were isolated [26]. A lowtemperature unblocking component was first isolated

Fig. 3. Conodonts from the Oyuklu section (magnification 150× for Fig. 3.1 and 200× for the others). Fig. 3.1: Epigondolella bidentata Mosher (sample 96/148); Fig. 3.2: Misikella hernsteini (Mostler) (sample 96/154A); Fig. 3.3–4: Misikella posthernsteini Kozur and Mock (samples 98/100A and 96/159, respectively); Fig. 3.5–6: Misikella koessenensis Mostler (sample 96/156); Fig. 3.7–10: Misikella rhaetica Mostler (Fig. 3.7–8: sample 96/154B, Fig. 3.9: sample 96/156B, Fig. 3.10: sample 96/156); Fig. 3.11–14: Misikella ultima Kozur and Mock (Fig. 3.11: sample 96 b :158, Fig. 3.12: sample 96/158A, Fig. 3.13–14: sample 96/159).

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

below 200 °C. Before bedding correction, this soft component isolated from three magnetization steps at most (NRM, 120 °C and 200 °C) has a direction close to that of the present-day geomagnetic field at the site (Fig. 5a), and is likely a transient viscous magnetization (Table 1). A second component was then isolated in the intermediate temperature range, between ∼ 250 °C and 350 °C, although its unblocking spectrum is overlapped by those of the low and high-temperature unblocking components. The directions obtained for this component are reported before and after bedding correction in Fig. 5b and c, respectively. Whatever the time of acquisition relative to the folding, the intermediate component has a reversed magnetic polarity (Table 1). Finally,

463

a high-temperature unblocking component with both magnetic polarities was obtained up to 560 °C–580 °C. Examples of normal (resp. reversed) magnetic polarity are shown in Fig. 4b,d,f (resp. a,c,e; see also Fig. 5d,e). The mean normal and reversed directions obtained for this component yield a positive reversal test (γ = 4.3°, γc = 6.9°; Table 1) [27]. Complete demagnetization was achieved below 600 °C, which indicates that the magnetization is largely dominated by minerals of the magnetite family. The magnetic behavior observed from the Oyuklu section is reminiscent of the one previously obtained from other Triassic sections investigated in the same region (e.g. the Bolücektasi Tepe, Kavur Tepe, Erenkolu

Fig. 4. Orthogonal vector diagrams in stratigraphic coordinates of progressive thermal demagnetization of samples from the Oyuklu section. These examples were chosen from different magnetic polarity intervals (indicated after the name of the samples).

464

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

Fig. 5. Equal area projection of paleomagnetic directions obtained from the Oyuklu section. Fig. 5a: Paleomagnetic directions isolated in the lower temperature demagnetization steps (b200 °C; in geographic coordinate system). Fig. 5b,c: Paleomagnetic directions isolated in the intermediate temperature range (b450 °C) before (b) and after (c) bedding correction. Fig. 5 d,e: High-temperature magnetic component directions obtained before (d) and after (e) bedding correction.

Mezarlik and Kavaalani sections [12,28–31]). This gives us confidence on the fact that the high-temperature magnetic component was acquired during the sedimentation process at the end of the Triassic. The high-temperature component from the Oyuklu section determines a sequence of 12 magnetic polarity intervals, all defined by several samples (minimum of two samples). However, due to the fault at the base of the section, the interpretable geomagnetic polarity sequence contains only 11 intervals labelled from A+ to K+ in Fig. 2 (from older to younger times). The latter sequence was considered hereafter for comparison with other Late Triassic magnetostratigraphic results and for

the construction of a complete Upper Norian and Rhaetian magnetic polarity time scale. 5. End-Triassic magnetostratigraphy Gallet et al. [12] and Krystyn et al. [3] first constructed a preliminary Upper Norian (Sevatian) magnetic polarity sequence by combining data obtained from the Turkish Kavur Tepe and Austrian Scheiblkogel sections (Fig. 6 [29,31]). More recently, Muttoni et al. [6] obtained additional Sevatian magnetostratigraphic results from the Sicilian Pizzo Mondello section which correlate well with the former sequence (continuous

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

465

Table 1 Mean directions before and after bedding correction obtained from the Oyuklu section in the low, middle and high unblocking temperature ranges Components

N

In situ

Tilt-corrected

Declination

Inclination

K

a95

Declination

Inclination

K

a95

Low Temp. Middle Temp. High Temp. Normal Reverse Normal + Reverse

121 115

342 191.3

62.9 − 17.9

15.1 13.5

3.4 3.7

272.1 196.9

82.4 −42.8

8 16

4.9 3.4

77 29 106

331.6 156 332.8

14 − 14.7 14.2

29.3 12 21.1

3 8.1 3.1

328.1 147.3 327.9

36.8 −41.1 37.9

24.7 15.5 21.1

3.3 7 3.1

lines, Fig. 6). These authors also tentatively correlated their results with those obtained by Channell et al. [5] from the Slovakian Silicka Brezova section (dashed lines, Fig. 6). A straightforward correlation is observed between the Pizzo Mondello and the new Oyuklu magnetostratigraphic data, allowing us to refine the “Tethyan” sequence during the Sevatian 2 zone and to extend it to the entire Rhaetian. A revised composite Sevatian–Rhaetian magnetic polarity time scale combining the data from both sections is shown in Fig. 6, and was then compared with the Newark magnetic polarity record. Channell et al. [5] and Muttoni et al. [6] correlated the Sevatian time interval from the Tethyan sections with the E14r to E17n sequence from the Newark Basin (dashed lines, Fig. 6). The new data from the Oyuklu section do not support this correlation. Unless the Oyuklu magnetostratigraphy is missing several chrons in the lower Rhaetian, it supposes that the reverse polarity dominated during the lower Rhaetian what is not confirmed by the Oyuklu results. It is worth pointing out that the Rhaetian record of Hounslaw et al. [32] from the St. Audrie's Bay shallow water marine section in England, dated from microfloral and palynological assemblages, also shows a predominantly normal polarity interval during the Rhaetian (Fig. 6). The previous correlation [5,6] raises other difficulties. There is now a common agreement to place the Carnian/ Norian boundary at the base of the magnetic interval E7n from the Newark sequence [3–5]. This reasonably allows one to correlate the normal polarity interval observed during the Lacian 3 zone with the Newark interval E13n (Fig. 6). In this case, following Muttoni et al. [6], practically the entire Middle Norian magnetic polarity time scale would contain only three magnetic polarity intervals (E13r–E14n–E14r). Such a possibility cannot be ruled out because the Middle Norian magnetostratigraphic record presently available from Tethyan sections is very fragmentary. However, this would imply a surprisingly short Middle Norian (∼3 Myr), but an extremely long Lower Norian (∼ 11 Myr) and Rhaetian

(∼ 8 Myr). In particular, the duration of the Rhaetian would be roughly similar to that of the Middle–Upper Norian interval, which appears rather unlikely when comparing sedimentation rates and faunal contents. Moreover, the Sevatian Pizzo Mondello section would exhibit three short magnetic polarity intervals that would have not been recovered from the Newark sequence, although the sedimentation rate of the latter was about 10 times higher. We consider that these inconsistencies argue in favour of placing the Norian/ Rhaetian boundary significantly higher, by at least ∼ 450 m, in the Newark basin sequence rather than within E17r. We therefore searched for another possible correlation between the new Sevatian–Rhaetian and Newark Basin sequences that took the following three conditions into account: i) We assume that the Lower Norian lies between intervals E7n and E13n from the Newark Basin sequence [3–5]. ii) The Lower and Middle Norian contain 5 ammonoid subzones each, whereas the Upper Norian and Rhaetian are defined by 6 ammonoid subzones [33]. Despite stratigraphic complexities observed in most Middle Norian sections in the Tethys, it appears that in all studied pelagic sequences, the Lower Norian is considerably thicker than the Middle Norian, suggesting a shorter duration for the latter. Such a comparison is difficult for the Upper Norian and the Rhaetian because of widespread missing of continuous pelagic carbonatic sedimentation. iii) Because the Newark record has a much higher time resolution, there is only a remote chance that short magnetic polarity intervals observed in the Tethyan marine sections would have been missed in the Newark sequence. In contrast, a correlation between the Tethyan and Newark sections leading to the opposite situation would be much more likely.

466 Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470 Fig. 6. Comparison between different Middle Norian (Alaunian) to Rhaetian magnetostratigraphic sequences obtained from Tethyan sections (this study; [3–6,12]). These data are compared to those obtained from the Newark Basin [7] and from the St Audrie's Bay section (UK) [32]. The continuous lines show the correlations proposed in the present study, whereas the dashed lines indicate the correlations previously proposed by Muttoni et al. [6]. The dotted lines between the Newark Basin and St Audrie's bay sequences also show the correlations suggested by Hounslaw et al. [32] and recently modified by Whiteside et al. [11] for the Rhaetian.

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

These different conditions strongly limit the potential correlations between the marine and terrestrial magnetostratigraphic records if no major distortion of one of them is introduced. In this respect, the two correlation frames shown in Fig. 7 better fulfill our criteria. They are quite similar to the one already proposed by Krystyn et al. [3] and Gallet et al. [4], and in both cases, would emphasize the good quality of the Tethyan and Newark magnetic polarity sequences. Note that the possible correlation of the latest magnetic polarity interval (K+) found in the Oyuklu section with interval E24n from the Newark sequence is constrained by new faunal evidence discussed in Kozur and Weems [9] (see below). The reason for these two options results from the discontinuous nature of the available Middle Norian Tethyan magnetostratigraphic results (Fig. 6). Even the most expanded Middle Norian magnetostratigraphic record from Pizzo Mondello [6] is strongly lacunary, and misses ∼ 40% of the sedimentary sequence. This leads to the uncertain correlation of the rather long reversed polarity interval observed during the Alaunian 3 zone (end of the Middle Norian) with either the sole interval E17r (option 1) or the sequence E17r–E20r (option 2) from the Newark sequence (Fig. 7). The agreement would be almost perfect in the case of option 1, while option 2 relies on the possibility that several short magnetic polarity intervals observed in the Newark basin sequence have been missed in Tethyan marine sections because of their much lower time resolution. The main inference from these two options however is that at least the lower part of the “traditional” Rhaetian (option 1) and possibly the Sevatian 2 zone (option 2) would be missing in the Newark sequence. This is clearly a matter of discussion. Kozur and Weems [9] and Kuerschner et al. [10] recently discussed biotic data that further question the completeness of the Newark basin sequence in the uppermost Triassic. While a relatively consistent picture was drawn for the Late Triassic biotic crisis from marine sections (e.g. [34]), the situation is much less clear in the terrestrial realm. An abrupt turnover in tetrapod (reptile and amphibian) and microfloral assemblages has been claimed across the Triassic/Jurassic boundary in the Newark basin (e.g. [35]), but much less pronounced events have been reported in European sections (e.g. [2,36]). Kuerschner et al. [10] also reported the case of sporomorph taxa, which have their last occurrence at the Triassic/Jurassic boundary in the Newark basin but disappear earlier, at the Sevatian–Rhaetian boundary, in the Alpine sections. These authors conclude that the “Triassic/Jurassic boundary” event identified in the Newark basin should be placed within the Sevatian 2

467

zone, implying a sedimentary hiatus of the latest Triassic. From a study of the Newark conchostracan faunas, Kozur and Weems [9] came to the same conclusion. They placed the distinct sporomorph spike previously assigned to the Triassic/Jurassic boundary within the Upper Norian (Sevatian), suggesting the sedimentary absence of the lower Rhaetian (including the Sevatian 2) in the Newark sequence. Going one step further, Kozur and Weems [9] also positioned the Triassic/Jurassic boundary somewhere within the Preakness Basalt, which implies an Upper Rhaetian age for interval E24n. This constraint, in addition to the resemblance between the magnetic polarity sequences, was taken into account in our correlations shown in Fig. 7. But this clearly requires confirmation by additional biostratigraphic data. A definitive and unique correlation with the Newark magnetic polarity sequence is premature because, up to now, only fragmentary Middle Norian magnetostratigraphic data have been obtained from Tethyan sections. However, between the two options proposed here, several arguments together with a better resemblance between the two sequences strongly plead in favor of option 1. Considering the cyclostratigraphy obtained by Kent and Olsen [37] from the Newark basin, the Middle Norian would have a duration of ∼ 6–7 Myr in the case of option 1 and ∼ 11–12 Myr for option 2. Plotted against the duration of ∼11 Myr assumed for the Lower Norian, the first option would be in better agreement with the fact that the Lower Norian in the Tethyan pelagic sections is always thicker than the Middle Norian. Option 2 supposes that several short normal magnetic polarity intervals were missed in the Tethyan Alaunian 3 sections (i.e. Kavur Tepe, Kavaalani etc.). Although not impossible, this scenario seems rather unlikely because a dense and evenly spaced sampling was carried out and no single short normal polarity interval was observed in any of the sections. Our preferred option 1 indicates that the traditional Rhaetian was as short as ∼ 2 Myr. This estimate relies on a rough duration of ∼1 Myr assumed for interval E24n recorded in the Orange Mountain Basalt, the Feltville Formation and in the Preakness Basalt. According to Kozur and Weems [9], the corresponding interval K+ from the Oyuklu section (Fig. 2) represents about half the thickness of the traditional Rhaetian, hence suggesting that the missing part of the Rhaetian in the Newark sequence (i.e. the time interval containing the magnetozones H−, I+ and J− found in Oyuklu) has also a duration of ∼ 1 Myr. If one assumes a stance contrary to Kozur and Weems [9] and claims that interval E24n is lowermost Jurassic in age, then the complete traditional

468 Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470 Fig. 7. Correlations proposed in this study between the new Sevatian–Rhaetian magnetic polarity sequence and the Newark Basin data. Two options are envisaged: option 1 which assumes that at least a part of the traditional Rhaetian is missing in the Newark Basin sequence and option 2 extending the missing part to the Sevatian 2. Uncertainties also exist for the uppermost part of the Triassic because interval E24n from the Newark Basin could either have a Triassic age, according to Kozur and Weems [9], or a Jurassic age as generally considered [7].

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

Rhaetian would be missing in the Newark Basin although its duration could potentially be roughly similar (∼ 2 Myr). Considering the cyclostratigraphy available in the Newark basin sequence, if the Sevatian 2 zone was included as part of the Rhaetian, the duration of the Rhaetian would increase in time up to ∼ 4.5 Myr. The short duration of the traditional Rhaetian agrees well with the absolute age of 201–202 Ma proposed by Pálfy et al. [38] for the Orange Mountain Basalt (Fig. 7; see also [9]) whereas the Triassic/Jurassic boundary is dated at ~ 200 Ma [39]. The Carnian/Norian boundary would then fall around 229 Ma, which suggests a ∼ 8 Myr duration for the Carnian, assuming that the Ladinian–Carnian boundary lies at ∼ 237 Ma [40]. The situation becomes more complex if the missing part is extended to the Sevatian 2 (option 2), i.e. 2 ammonoid zones, within an extremely short duration of ∼ 2 Myr. This would require that the absolute time calibration presently considered for the entire Newark Late Triassic sequence would need to be revised, by lowering the Carnian/Norian boundary (by 1 or 2 Myr?) and reducing the duration of the Carnian by the same amount of time. As a concluding remark, the top-Triassic magnetostratigraphy now appears well established from several data sets ( this study, [6,12,29,31]). The available results suggest that the Norian/Rhaetian boundary should be placed much higher in the Newark sequence than previously thought. The absence of the lower part of the traditional Rhaetian, according to Kozur and Weems [9] or of its integrality, perhaps also including the Sevatian 2 zone, in the Newark supergroup would reconcile most Late Triassic magnetostratigraphic results and biotic features obtained from marine and terrestrial realms.

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

Acknowledgements [13]

We are grateful to the Süleyman Demirel University of Isparta, particularly to Prof. F. Yagmurlu for his help with logistics and H. Kozur for discussion on conodonts. We also thank H. Bouquerel for the help during field work. This is a contribution to IGCP project 467 supported by the Austrian National committee for IGCP (L.K.). This is IPGP contribution no. 2198.

[15]

References

[17]

[1] V. Courtillot, P. Renne, On the ages of flood basalt events, C. R. Geosci. 335 (2003) 113–140. [2] S. Lucas, L. Tanner, Late Triassic extinction events, Albertiana 31 (2004) 31–37. [3] L. Krystyn, Y. Gallet, J. Besse, J. Marcoux, Integrated Upper Carnian to Lower Norian biochronology and implications for the

[14]

[16]

[18]

469

Upper Triassic magnetic polarity time scale, Earth Planet. Sci. Lett. 203 (2002) 343–351. Y. Gallet, L. Krystyn, J. Besse, J. Marcoux, Improving the Upper Triassic numerical time scale from cross-correlation between Tethyan marine sections and the continental Newark basin sequence, Earth Planet. Sci. Lett. 212 (2003) 255–261. J.E.T. Channell, H. Kozur, T. Sievers, R. Mock, R. Aubrecht, M. Sykora, Carnian–Norian biomagnetostratigraphy at Silická Brezová (Slovakia): correlation to other Tethyan sections and to the Newark Basin, Palaeogeogr. Palaeoclimatol. Palaeoecol. 122 (1993) 65–109. G. Muttoni, D.V. Kent, P. Olsen, P. di Stefano, W. Lowrie, S. Bernasconi, F. Martin Hernandez, Tethyan magnetostratigraphy from Pizzo Mondello (Sicily) and correlation to the Late Triassic astrochronological polarity time scale, Geol. Soc. Amer. Bull. 116 (2004) 1043–1058. D.V. Kent, P.E. Olsen, W.K. Witte, Late Triassic-earliest Jurassic geomagnetic polarity sequence and paleolatitudes from drill cores in the Newark rift basin, eastern North America, J. Geophys. Res. 100 (1995) 14965–14998. A. Marzoli, H. Bertrand, K. Knight, S. Cirilli, N. Buratti, C. Vérati, S. Nomade, P. Renne, N. Youbi, R. Martini, K. Allenbach, R. Neuwerth, C. Rapaille, L. Zaninetti, G. Bellieni, Synchrony of the Central Atlantic magmatic province and the Triassic–Jurassic climatic and biotic crisis, Geol. Soc. Amer. Bull. 32 (2004) 973–976 (of America). H. Kozur, R. Weems, Conchostracan evidence for a late Thaetian to Early Hettangian for the CAMP volcanic event in the Newark Supergroup, and a Sevaian (late Norian) age for the immediately underlying beds, Hallesches Jahrb. Geowiss. B27 (2005) 21–51. W. Kuerschner, N. Bonis, L. Krystyn, High resolution carbon isotope stratigraphy and palynology of the Triassic-Jurassic transition in the Tiefenbachgraben –Northern Calcreous Alps (Austria), Palaeogeogr. Palaeoclimatol. Palaeoecol. (in press), doi:10.1016/j.palaeo.2006.06.031. J. Whiteside, P. Olsen, D. Kent, S. Fowell, M. Et-Touhami, Synchrony between the Central Atlantic magmatic province and the Triassic–Jurassic mass-extinction event? Palaeogeogr. Palaeoclimatol. Palaeoecol. (in press), doi:10.1016/j.palaeo.2006.06.035. Y. Gallet, J. Besse, L. Krystyn, J. Marcoux, J. Guex, H. Théveniaut, Magnetostratigraphy of the Kavaalani section (southwestern Turkey): Consequence for the origin of the Antalya Calcareous Nappes (Turkey) and for the Norian (Late Triassic) magnetic polarity timescale, Geophys. Res. Lett. 27 (2000) 2033–2036. J.G. Ogg, Status of divisions of the International Geological Time Scale, Lethaia 37 (2004) 183–199. L. Krystyn, W. Kuerschner, Biotic events around the Norian– Rhaetian boundary, Albertiana 32 (2005) 17–20. A. Koçyigit, Karaman–Ermenek (Konya) bölgesinde ofiyolitli malanj ve diger olusuklar, Geol. Soc. Turkey Bull. 19 (1976) 103–115. M. Gutnic, O. Monod, A. Poisson, J.-F. Dumont, Géologie des Taurides occidentales (Turquie), Mém. Soc. Géol. Fr. 58 (1979) (112 pp.). S. Gökdeniz, Recherches géologiques dans les Taurides occidentales entre Karaman et Ermenek, Turquie. Les séries à “tuffites vertes” triasiques. Stratigraphie, cadre structural, minéralogie, élément de géochimie, Doctoral thesis Université Paris–Sud–Orsay (1981). N. Özgül, Stratigraphy of the tectonostratigraphical units around Hadim–Bozkir–Tashkent region, northern part of the central Taurides, Bull. Miner. Res. Explor. 119 (1997) 113–174.

470

Y. Gallet et al. / Earth and Planetary Science Letters 255 (2007) 458–470

[19] T. Andrew, A. Robertson, The Beysehir–Hoyran–Hadim nappes: genesis and emplacement of Mesozoic marginal and oceanic units of the northern Neotethys in southern Turkey, J. Geol. Soc. (Lond.) 159 (2002) 529–543. [20] O. Masset, P. Moix. Le mélange ophiolitique de Mersin(Turquie du sud), un fragment de marge Neotéthy?, unpub. diplom. University of Lausanne, Institut de Géologie et Paléontologie (2004) 143 pp. [21] N. Özgül, Stratigraphy and tectonic evolution of the Central Taurids, in: O. Tekeli, M. Göncüoglu (Eds.), Geology of the Taurus Belt, Proceedings (MTA) Ankara, 1984, pp. 77–90. [22] F. Amodeo, F. Molisso, H. Kozur, E. Marsella, B. D'Argenio, Age of Transitional Beds from “Cherty Limestones” (calcari con selce) to “Radiolarites” (scisti silicei) in the Lagonegro Domain (Southern Italy). First evidence of Rhaetian conodonts in peninsular Italy, Boll. Serv. Geol. Ital. 110 (1993) 3–22. [23] A. Tollmann, Analyse des klassischen nordalpinen Mesozoikums, Deuticke (Wien), 1976 580 pp. [24] H. Kozur, R. Mock, New Middle Carnian and Rhaetian Conodonts from Hungary and the Alps, Jahrb. Geol. Bundesanst. 134 (1991) 271–297. [25] M.J. Orchard, Upper Triassic conodont biochronology and new index species from Canadian Cordillera, Bull. - Geol. Surv. Can. 417 (1991) 299–335. [26] J.-P. Cogné, Paleomac : a Macintosh application for treating paleomagnetic data and making plate reconstruction, Geochem. Geophys. Geosyst. 4 (2003) 1007, doi:10.1029/2001GC000227. [27] P.L. McFadden, M. McElhinny, Classification of the reversal test in paleomagnetism, Geophys. J. Int. 103 (1990) 725–729. [28] Y. Gallet, J. Besse, L. Krystyn, J. Marcoux, H. Théveniaut, Magnetostratigraphy of the late Triassic Bolücektasi Tepe section (southwestern Turkey): implications for changes in magnetic reversal frequency, Phys. Earth Planet. Inter. 93 (1992) 273–282. [29] Y. Gallet, J. Besse, L. Krystyn, H. Théveniaut, J. Marcoux, Magnetostratigraphy of the Kavur Tepe section (south-western Turkey): a magnetic polarity time scale for the Norian, Earth Planet. Sci. Lett. 117 (1993) 443–456. [30] Y. Gallet, J. Besse, L. Krystyn, H. Théveniaut, J. Marcoux, Magnetostratigraphy of the Mayerling (Austria) and Erenkolu Mezarlik (Turkey) sections: improvement of the Carnian (late

[31]

[32]

[33]

[34] [35]

[36]

[37]

[38]

[39] [40]

Triassic) magnetic polarity time scale, Earth Planet. Sci. Lett. 125 (1994) 173–191. Y. Gallet, J. Besse, L. Krystyn, J. Marcoux, Norian magnetostratigraphy from the Scheiblkogel section, Austria: constraint on the origin of the Antalya Nappes, Turkey, Earth Planet. Sci. Lett. 140 (1996) 113–122. M.W. Hounslaw, P. Posen, G. Warrington, Magnetostratigraphy and biostratigraphy of the Upper Triassic and lowermost Jurassic succession, St. Audrie's Bay, UK, Palaeogeogr. Palaeoclimatol. Palaeoecol. 213 (2004) 331–358. L. Krystyn, Zur Rhät-Stratigraphie der Zlambach-Schichten (vorläuger Bericht), Sitz.ber. - Osterr. Akad. Wiss. Math.-Nat. wiss. Kl. 196 (1987) 21–36. A. Hallam, How catastrophic was the end-Triassic mass extinction? Lethaia 35 (2002) 147–157. P. Olsen, S. Fowell, B. Cornet, The Triassic/Jurassic boundary in continental rocks of eastern North America; a progress report, in: V. Sharpton, P. Ward (Eds.), Global Catastrophes in Earth History: An Interdisciplinary Conference on Impacts, Volcanism and Mass Mortality, Geol. Soc. Am. Spec. Pap., vol. 247, 1990, pp. 585–593. S. Lucas, Triassic tetrapod extinctions and the compiled correlation effect, Can. Soc. Petrol. Geol. Mem. 17 (1994) 869–875. D.V. Kent, P.E. Olsen, Astronomically tuned geomagnetic polarity time scale for the Late Triassic, J. Geophys. Res. 104 (1999) 12831–12841. J. Pálfy, P. Smith, J. Mortensen, Dating the end-Triassic and early Jurassic mass extinction, correlative large igneous provinces, and isotopic events, in: C. Koeberl, K. McLoad (Eds.), Catastrophic Events and Mass Extinctions: Impacts and Beyond, Geol. Soc. Am. Spec. Pap., vol. 356, 2002, pp. 523–532. J. Pálfy, P. Smith, J. Mortensen, A U–Pb and 40Ar/39Ar time scale for the Jurassic, Can. J. Earth Sci. 37 (2000) 923–944. R. Mundil, P. Brack, M. Meier, H. Rieber, F. Oberli, High resolution U–Pb dating of middle volcaniclastics: time-scale calibration and verifications of tuning parameters for carbonate sedimentation, Earth Planet. Sci. Lett. 141 (1996) 137–151.