Introduction

Detailed spatial and temporal patterns of the ammonite extinction at or near the Cretaceous—Paleogene (K—Pg, formerly K–T) boundary are still far from clear (see Landman et al. 2007 for the overview). This is mainly due to the presence of a hiatus at the K—Pg boundary in many ammonite-bearing sections, resulting from the erosional removal of the topmost Maastrichtian and lowermost Danian strata during early, but not earliest, Danian sea level oscillations (Keller and Stinnesbeck 1996; Smit 1997). Clearly reworked, worn and/or mineralised ammonite moulds in the Danian transgressive lags testify to the occurrence of these fossils in the sedimentary units once present (e.g., Machalski and Walaszczyk 1988), but data on their original vertical distribution in relation to the K—Pg boundary have been lost.

The best sections to analyse the last phases of ammonite evolution are those which preserve the iridium-rich, Chicxulub impact-related clay layer, which is sandwiched between the Maastrichtian and Danian, usually carbonate strata (Alvarez et al. 1980; Smit 1999; Schulte et al. 2010). The base of this clay defines the Cretaceous—Paleogene boundary, which is thus determined precisely by the moment of an extraterrestrial body impact (Molina et al. 2009). If preserved, the sedimentary and geochemical record of the impact, corresponding to a globally synchronous fallout of impact products, is the best proof available of continuous geological and palaeontological record across the K—Pg boundary at any given site (Claeys et al. 2002; Racki et al. 2011). Unfortunately, only few sites worldwide with preserved K—Pg boundary clay have yielded an adequate ammonite record. These are exemplified by the celebrated Stevns Klint section, eastern Denmark, where some ammonite taxa range up to the K—Pg boundary clay (Birkelund 1993) and even occur above it, which led to the hypothesis of ammonite survival into the early Danian (Machalski and Heinberg 2005). Another example is provided by ammonite assemblages from the GSSP (Global Stratotype Section and Point) for the base of the Danian at El Kef, Tunisia, where ammonites totally disappear from the fossil record in the 2 m-thick interval just below the boundary clay (Goolaerts et al. 2004; Goolaerts 2010).

Fig. 1.

Location of the study area in the western Kopet Dagh Mountains (Sumbar River region near the village of Kara-Kala, Turkmenistan). An arrow indicates the approximate position of the ammonite-bearing Cretaceous—Paleogene Sumbar River section.

In the present paper we add a new point to the global picture of the ammonite extinction by describing and discussing two terminal Maastrichtian ammonite taxa from a complete K—Pg succession in the Sumbar River area, western Kopet Dagh (southwest Turkmenistan, Central Asia). The completeness of the Sumbar River section was first noted by Alekseev et al. (1988), who recognised a boundary clay layer with a prominent iridium anomaly at the very base of the Danian. The impact-related signature of the clay layer was subsequently confirmed by Wolbach et al. (1990), Meisel et al. (1995), and Heymann et al. (1996). Alekseev et al. (1988) provided biostratigraphical data for the section and preliminary notes on the ammonite assemblage from the top of the upper Maastrichtian; the present paper is based on the material referred to in that paper. It was collected by one of us (ASA) during two visits to the section in 1983 and 1985.

Fig. 2.

The ammonite-bearing Sumbar section, Turkmenistan. A. Log of the section (see text for explanation; modified after Alekseev et al. 1988: fig. 1). B. Field photograph of the section with marked K—Pg boundary.

Institutional abbreviations.—MGUH, Statens Naturhistorisk Museum, København, Denmark (formerly Geologisk Museum, Universitet København); NHMM, Natuurhistorisch Museum Maastricht, Maastricht, the Netherlands; VSEGEI, Vserossijskij Nauchno-Issledovatel'tskij Geologicheskij Institut im. A.P. Karpinskogo, Sankt-Peterburg, Russia.

Geographical and geological setting

The Sumbar River ammonite-bearing section is situated in western Kopet Dagh, southwest Turkmenistan, Central Asia (Fig. 1). The section is located 6 km northwest of the village of Kara-Kala on the slope of the Isak Mountain, on both sides of the road to the city of Khizyl-Arvat (Alekseev et al. 1988). Reference is made to Atabekian and Likhacheva (1961) for general data on the Upper Cretaceous deposits of western Kopet Dagh.

The Sumbar River section (Fig. 2A, B) starts with lightgrey marls, ca. 10 m thick (unit 1 in Fig. 2A), which in the upper part contain rare holasteroid echinoids (Echinocorys sp. of typically Maastrichtian appearance) and common moulds of ammonites (Alekseev et al. 1988). Higher up, there is a layer of reddish-grey clay, 6 cm thick (unit 2 in Fig. 2A), with an anomalous iridium concentration at the base (up to 66.3 ng/g; see Alekseev et al. 1988). Soot particles (Wolbach et al. 1990), osmium and strontium isotopic ratios consistent with the impact scenario (Meisel et al. 1995), as well as the fullerenes C₆₀ and C₇₀ which probably formed during global wildfires in the aftermath of the Chicxulub impact (Heymann et al. 1996), have also been identified in the clay layer. This passes gradually upwards into marls with lens-like limestone intercalations, ca. 15 m thick (unit 3 in Fig. 2A). These strata contain early Danian echinoids, including Cyclaster integer (= C. danicus), which occurs near the base of that unit (Alekseev et al. 1988). Biostratigraphical data for the Sumbar River section rely mainly on the foraminiferal record, which is definitely Boreal in character (Alekseev et al. 1988; ASA unpublished data). These data indicate that unit 1 represents the upper Maastrichtian Pseudotextularia elegans Zone, the base of unit 2 corresponding to the lowermost Danian Guembelitria cretacea Zone, and the remainder of the section being referable to the higher Danian Globigerina eugubina Zone (Alekseev et al. 1988). The clay layer (unit 2) with its unique geochemical anomaly corresponds to the K—Pg boundary clay as recorded elsewhere in the world.

There is no physical evidence of a hiatus around the K—Pg boundary in the Sumbar River ammonite-bearing section, in contrast to other outcrops in the same area which are located between 2 and 11 kilometres to the west. At the latter sites, the clay layer is missing and there is an indurated level (hard-ground) at the top of the Maastrichtian, which is penetrated by burrows (1–2 cm in diameter), filled with overlying sediment, which pipe down to 30 cm. The lowermost part of the Danian contains small limestone pebbles (3–4 cm in size) which clearly must have resulted from erosion and reworking of the underlying Maastrichtian. There is no iridium anomaly at the base of the Danian in these sections, the iridium levels approaching background values only (Alekseev et al. 1988). This picture is reminiscent of other parts of the world where locally complete (with the boundary clay) and incomplete (i.e., terminated by an erosional surface) sections are in close proximity, usually as a result of a complex interplay of local or regional topography, synsedimentary tectonics and sea level changes (e.g., Surlyk 1997; Machalski 2005b).

Ammonite preservation and vertical distribution

The material studied is preserved as yellowish to orange brown, composite moulds, in light grey, indurated marl and, when originating from closer to the K—Pg boundary, in more clayey matrix which shows a tendency to flake after having been soaked in water. Preservation varies widely, from near-complete specimens (Fig. 3F–H), isolated body chambers (e.g., Fig. 4A) and nuclei (e.g., Fig. 3C, E) to associated phragmocones on a bedding plane (Fig. 4E, F). Most specimens are crushed to varying degrees, so that measurements cannot be but approximate. All specimens studied come from the topmost part of the upper Maastrichtian portion of the Sumbar River section (Fig. 2); where known, the exact level is noted in the figure captions (Figs. 3, 4). There is also a single scaphitid record above the boundary clay, in the lower Danian portion of the section (see below).

Systematic palaeontology

Fig. 3.

Scaphitid ammonoid Hoploscaphites constrictus johnjagti Machalski, 2005a, Sumbar River section, Turkmenistan, upper Maastrichtian, Upper Cretaceous, all collected by ASA, with indications in brackets of centimetres below K/Pg boundary (where known). A. NHMM 2011 037, external mould of phragmocone. B. NHMM 2011 039, partial body chamber of macroconch. C. NHMM 2011 036, nucleus. D. NHMM 2011 034, partial phragmocone of ?macroconch (15 cm). E. NHMM 2011 038, nucleus (5 cm). F. NHMM 2011 033, near-complete ?microconch (18 cm). G. NHMM 2011 035, microconch (5 cm). H. NHMM 2011 040, macroconch.

Fig. 4.

Ammonites from the Sumbar River section, Turkmenistan, upper Maastrichtian, Upper Cretaceous, collected by ASA and Mikhail A. Nazarov, with indications in brackets of centimetres below K—Pg boundary (where known). A, B, E, F. Scaphitid ammonoid Hoploscaphites constrictus johnjagti Machalski, 2005a. A. NHMM 2011 044, microconch body chamber. B. NHMM 2011 043, apertural portion of body chamber, with epizoic craniid brachiopod (arrowed). E. NHMM 2011 042b, pair of phragmocones (10–15 cm). F. NHMM 2011 042a, set of phragmocones (10–15 cm). C. Baculitid ammonoid Baculites cf. vertebralis Lamarck, 1801, NHMM 2011 041a, partial body chamber. D. Scaphitid phragmocone from the Danian portion of the Sumbar River section, NHMM 2011 046, probably Hoploscaphites constrictus johnjagti Machalski, 2005a (= Pachydiscidae indet. sensu Alekseev et al. 1988: fig. 1).

Concluding remarks

In terms of the end-Cretaceous extinction debate, the present study demonstrates that in the western Kopet Dagh area two ammonite taxa, i.e., Baculites cf. vertebralis and Hoploscaphites constrictus johnjagti, representing two families (Baculitidae and Scaphitidae), probably did survive up to the end of the Maastrichtian. Whether or not the single scaphitid mould recovered above the K—Pg boundary clay represents a Danian survivor as suggested for better-preserved and documented material from Stevns Klint, Denmark (Machalski and Heinberg 2005; Machalski et al. 2009) or simply represents a remanié specimen, cannot be determined at this time. Obviously, more extensive collecting across the K—Pg boundary at the Sumbar River section is needed to test this possibility.

As far as the evolution and biostratigraphical potential of scaphitids are concerned, the present study confirms that Hoploscaphites constrictus johnjagti is the last member of the H. constrictus evolutionary lineage (Machalski 2005a, b). At the classic locality of Stevns Klint, Denmark, this temporal subspecies is restricted to the Grey Chalk, a 0–4 m thick unit situated immediately below the K—Pg boundary clay, and its acme is at the top of this unit (Machalski 2005b), much like in the Sumbar River section. Bed-by-bed collecting down the Sumbar River section may be expected to yield additional data which would form an independent test for Machalski's concept of successive chronosubpecies of this lineage (Machalski 2005a). In anticipation of this, the results presented here show that Hoploscaphites constrictus johnjagti may be regarded as a good biostratigraphic proxy for the K—Pg boundary in sections where the boundary clay is absent.

Palaeobiogeographically, the ammonite taxa recorded from the Sumbar River section are essentially Boreal forms (see above). The records of these taxa in the Sumbar River section represent their southeasternmost occurrences known to date. They may have reached the Kopet Dagh region during the late Maastrichtian sea level rise as documented by Mahboubi et al. (2006). The relatively deep-water environment (outer shelf) of the ammonite-bearing marls from the Sumbar River section is confirmed by a high percentage of planktic species in foraminiferal assemblages (76–89%; see Alekseev et al. 1988). Again, more data from the lower part of the succession are needed to obtain a reliable picture of environmental changes towards the K—Pg boundary at this site.

From the perspective of regional ammonoid diversity, the present study complements the rather meagre list of Maastrichtian ammonites known to date from Central Asia (Kazakhstan, Turkmenistan). We are aware of only a handful of records. From Maastrichtian strata in western Kopet Dagh, Moskvin (1959: pls. 3: 3, 4: 1–4) illustrated two heteromorphs, Baculites vertebralis and Discoscaphites constrictus var. niedzwiedzkii (= Hoploscaphites constrictus), the former from the Sumbar River region. Subsequently, Krymgol'ts (1974: pl. 55: 2) recorded Bostrychoceras schloenbachi (= Nostoceras schloenbachi) from the lower Maastrichtian of Kene-Beurme, Turkmenistan. To this, Arkadiev and Bogdanova (1997: pls. 53: 5, 55: 4, 58: 3) (see also Arkadiev et al. 2000: pls. 15: 3a, b; 16: 4–6) added Pseudokossmaticeras galicianum (lower Maastrichtian, Hauericeras sulcatum Zone; Kredin Canyon, western Kopet Dagh) and Hoploscaphites constrictus (upper lower and upper Maastrichtian; Isak, Ayshem and Seitkerderi mountains, western Kopet Dagh).