Late Cretaceous climate signal of the Northern Pekulney Range Flora of northeastern Russia

Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25 – 46 www.elsevier.com/locate/palaeo

Late Cretaceous climate signal of the Northern Pekulney Range Flora of northeastern Russia Helen J. Craggs* Department of Earth Sciences, The Open University, Walton Hall, Milton Keynes MK7 6AA, UK Received 26 November 2003; received in revised form 29 September 2004; accepted 19 November 2004

Abstract Plant-bearing continental deposits from the northern Pekulney mountain range in the Anadyr-Koryak subregion of northeastern Russia (approximately 668N, 1758E) represent a rich and diverse flora growing close to the Late Cretaceous North Pole at a palaeolatitude of ~788N. The Early Coniacian age of this flora is based on biostratigraphic correlation of the plantbearing beds with underlying and overlying marine units. The Northern Pekulney Range Flora comprises two large plant macrofossil assemblages, one from the Tylpegyrgynai Formation on the western slopes of the northern Pekulney Range and the other from the Poperechnaya Formation on its northeastern slopes. The flora is dominated by angiosperms followed by conifers, ferns, cycadophytes, ginkgophytes and sphenophytes, in decreasing order of species diversity. Dicotyledonous angiosperm leaves (57 morphotypes) from the Northern Pekulney Range Flora were subjected to a Climate Leaf Analysis Multivariate Program (CLAMP) physiognomic analysis. Separate analyses of the constituent assemblages from the Tylpegyrgynai and Poperechnaya formations yielded 27 and 39 morphotypes, respectively. Results suggest that the Northern Pekulney Range Flora experienced a mean annual temperature of 8.1F1.2 8C, a cold month mean temperature of 1.5F1.9 8C, a mean monthly growing season precipitation of 78.8F36.9 mm and a growing season of 5.3F0.7 months. This suggests a warm temperate climate lacking pronounced drought, in which the polar light regime constrained the length of the growing season. CLAMP results for the Tylpegyrgynai Formation Flora suggest slightly higher temperatures, with a mean annual temperature of 9.4F1.2 8C and a cold month mean temperature of 0.9F1.9 8C, whilst estimates for the Poperechnaya Formation Flora are cooler, with a mean annual temperature of 7.3F1.2 8C and a cold month mean temperature of 2.7F1.9 8C. Because the floras from the Tylpegyrgynai and Poperechnaya formations also exhibit differences in their taxonomy and physiognomy, they should be considered as two distinct floras. The overall taxonomic composition of these floras, their foliar physiognomy and estimated palaeoclimate parameters suggest a deciduous phenology with a few evergreen elements. Comparisons with other Arctic Coniacian floras are consistent with the high latitudinal position of the Northern Pekulney Range Flora and its proximity to the northern proto-Pacific Ocean. D 2004 Elsevier B.V. All rights reserved. Keywords: Late Cretaceous; Northeastern Russia; Palaeobotany; Palaeoclimate; Leaf physiognomy

* Tel.: +44 1908 654259. E-mail address: [email protected]. 0031-0182/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.palaeo.2004.11.014

26

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

1. Introduction The Cretaceous is widely recognised as a time of pronounced global warmth and therefore a useful source of data on how Earth systems operate and interact under so-called dgreenhouseT conditions (Skelton et al., 2003). In northeastern Russia, the Cretaceous sediments contain an abundance of plant fossil remains providing evidence of a rich flora thriving at a high latitude (Vakhrameev, 1991; Spicer et al., 2002). Fossil floras from this area have been studied together with other floras from northern Russia and Alaska to provide data that can be used to reconstruct Arctic climate and vegetation patterns during the Cretaceous and to improve our understanding of greenhouse world Earth systems. This study concentrates on climate signals derived from angiosperm leaf physiognomy because these provide a quantitative method of determining a wide range of palaeoclimate variables (Wolfe, 1993). One of the richest and most diverse floras from the Upper Cretaceous lies in plant-bearing continental deposits at the northern end of the Pekulney mountain range within the Anadyr-Koryak subregion of north-

eastern Russia. Until now the flora has been known collectively as the Tylpegyrgynai Flora (Terekhova and Filippova, 1983, 1984) and is found at a present day latitude of approximately 668N (Fig. 1a,b). The plant-bearing deposits were discovered in 1956 by staff of the USSR Geological Survey. The resulting collection of plant fossils was examined by Yefimova but was never formally published, Yefimova’s observations existing only as an internal report of the Northeastern Geological Survey, Magadan. A more detailed study of the locality was carried out by Terekhova and Filippova (1983, 1984), who examined the stratigraphy of the western and eastern slopes of the northern Pekulney Range, including the continental plant-bearing beds and the bounding marine units. Late Cretaceous plant fossils were found in the Tylpegyrgynai Formation on the western slopes and in the Poperechnaya Formation on the eastern slopes. The Poperechnaya Formation is dated as Early Coniacian, on the basis of biostratigraphic constraints of overlying and underlying beds containing marine molluscs. For the Tylpegyrgynai Formation, the upper age limit is constrained as Early Coniacian based on the composition of a marine biota in an overlying

Fig. 1. Locality map for plant-bearing deposits in the Northern Pekulney Range. (a) Northeastern Russia with subregions, (b) Northern Pekulney Range with the plant fossil sites: (1) Tylpegyrgynai Formation floral locality (shown in detail in Fig. 2), (2) Poperechnaya Formation floral locality (shown in detail in Fig. 4) (modified from Herman, 1999).

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

deposit, but the lower age limit remains uncertain because of an erosional basal contact with underlying sediments of Valanginian age. Large numbers of plant fossil samples were collected from both formations and these were later analysed by Filippova (1991, 1994; Filippova and Abramova, 1993). The fossil floras from the Tylpegyrgynai and Poperechnaya formations were regarded as being of similar taxonomic compositions and were therefore likely to reflect a single floral association. This association was named the Tylpegyrgynai Flora by Filippova (Terekhova and Filippova, 1983, 1984). Despite its near-polar palaeolatitude of 788N (Smith et al., 1981), the Tylpegyrgynai Flora appears to represent one of the most taxonomically diverse and well-dated Late Cretaceous floras of the Arctic, growing at a critical time in the evolutionary radiation of angiosperms. The present study was therefore undertaken to obtain further insights into the prevailing Late Cretaceous climatic conditions, to make a comparison with climate signals from other Arctic floras of similar age, and to compare the constituent Tylpegyrgynai and Poperechnaya formation fossil floras. To obtain quantitative data on a range of climate variables a foliar physiognomic approach is required, and the most comprehensive of these is the Climate Leaf Analysis Multivariate Program (CLAMP) methodology (Wolfe, 1993), an introduction to which is given in Methods and techniques.

2. Geological setting During the Late Cretaceous, the palaeogeography of the North Pacific Asian Region consisted of three tectonic regions. To the west lay the uplands and non-flooded lowlands of Verkhoyansk-Chukotka, in the centre were the highlands of the OkhotskChukotka volcanogenic belt, whilst in the east was the coastal environment of the Anadyr-Koryak area (Belyi, 1994). The Late Cretaceous plant-bearing deposits of the Tylpegyrgynai and Poperechnaya formations were laid down in the Anadyr-Koryak Subregion, which comprised a coastal plain with terrestrial deposits, a shallow shelf with intermittent marine deposits and a persistent marine environment. These plant-bearing deposits are now found in the northern part of the

27

Pekulney Range, at a latitude of approximately 668N (Fig. 1a), however a reconstruction of the continental positions for the Late Cretaceous indicates that, during the Coniacian, this area was situated close to the palaeo North Pole, at a latitude of 788N (Smith et al., 1981). Plant fossil remains are present in both the Tylpegyrgynai Formation, which lies on the western slopes of the Pekulney Range, and in the Poperechnaya Formation from the eastern slopes (Fig. 1b). 2.1. The western slopes of the northern Pekulney Range On the western slopes of the northern Pekulney Range, Upper Cretaceous deposits are widely exposed towards the south of the Tylpegyrgynai Mountains and in the Afon’kina River basin (Figs. 1b and 2a). Terrigenous plant-bearing deposits of the Tylpegyrgynai Formation form the lowermost part of this Upper Cretaceous succession. There is an unconformable erosional contact between these units and the underlying marine Pekulneyveem Formation which includes volcanics of Volgian (approximately equivalent to Tithonian) to Valanginian age (Fig. 2). The Tylpegyrgynai Formation is 1000 to 1100 m thick and is divided into three subformations (Terekhova and Filippova, 1984) (Fig. 2b). The Lower Subformation, 300 to 400 m thick, is largely composed of sandstones with subordinate layers of siltstones, mudstones and tuffs. The Middle Subformation, 400 to 600 m thick, is mainly represented by siltstones interbedded with sandstones, later deposits being distinguished by the presence of white tuffs. The Upper Subformation, 100 to 300 m thick, consists of interbedded volcanic lavas, breccias and tuffs, with subordinate layers of tuffaceous sandstones, siltstones and mudstones. All three subformations contain plant fossils, but foliar remains are particularly abundant in the Middle Subformation. The sediments are interpreted to represent a floodplain environment with increasing volcanic influence. The Tylpegyrgynai Formation is overlain by the Yanranay Formation, which has an overall thickness of 700 to 800 m, and is divided into two subformations (Terekhova and Filippova, 1984) (Fig. 2b). The Lower Subformation, 250 to 300 m thick, comprises tuffaceous sandstones with lenses of tuff and inter-

28

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

Fig. 2. Geology of the south of the Tylpegyrgynai Mountains, western slopes of the Northern Pekulney Range (Zakharov’s and Terekhova’s data; modified from Herman, 1999). (a) Geological map, (b) stratigraphy; note: the full thickness of the Pekulneyveem Formation is not shown.

bedded layers of siltstones and mudstones. The Upper Subformation, 400 to 500 m thick, is dominated by interbedded tuffaceous sandstones, siltstones and tuffs. Marine molluscs found in these two subformations include Inoceramus naumanni Yok., Parallelodon sachalinensis Schmidt, Variamussium sp., Limatula sp., Terebratulina sp., Neopuzosia ? cf. ishikawai (Jimbo), together with remains of echinoids, crinoids and crabs. This fossil fauna has been correlated with the Inoceramus yokoyamai zone, suggesting a Late Coniacian to Early Santonian age (Terekhova and Filippova, 1984) (Fig. 3). In association with the marine molluscs, a small number of fossil plants have been found, including Gingko ex gr. adiantoides (Unger) Heer, Nilssonia sp., Cephalotaxopsis intermedia Hollick, Sequoia reichenbachii (Gein.) Heer, Sequoia cf. fastigiata (Sternb.) Heer, Thuja cretacea (Heer) Newb., Elatocladus sp.,

Quereuxia angulata (Newberry) Krysht. and Dicotylophyllum sp. The upper age limit of the Tylpegyrgynai Formation, and hence its contained flora, is therefore constrained by the overlying inoceramid-bearing marine deposits of the Yanranay Formation (Late Coniacian to Early Santonian) and can be defined as Early Coniacian (Fig. 3). The age of the base is less well defined due to its erosional contact, and associated time gap, with underlying Lower Cretaceous deposits of Valanginian age from the Pekulneyveem Formation. This will be considered further after examining the Poperechnaya Formation and its flora. 2.1.1. Composition of the Tylpegyrgynai Formation taphoflora Plant fossil remains are found in all three subformations of the Tylpegyrgynai Formation. Filippova

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

29

Fig. 3. Marine biostratigraphy of the eastern and western slopes of the Northern Pekulney Range (modified from Terekhova and Filippova, 1983, 1984); plant-bearing beds are shown in grey.

(1991, 1994) and Terekhova and Filippova (1984) identified 50 different leaf species and considered that the taxonomic composition of the floras from each subformation was sufficiently similar for them to constitute a single taphoflora, which Filippova named the Tylpegyrgynai Flora. Based on the number of species (Table 1), the Tylpegyrgynai Flora is dominated by angiosperms (40%) followed by conifers and ferns. Other groups of plants, such as cycadophytes, ginkgoales, sphenophytes, lycopods and probable liverworts, are less diverse and less abundant. Ferns, although not prolific, are found throughout the formation and typically include Coniopteris, Osmunda sp., Ochotopteris sp. and the large-leaved Hausmannia bipartita Samyl. et Shczep. The cycadophyte Nilssonia occurs in all subformations, but Ctenis has not been found. Ginkgoales are only present as Ginkgo ex gr. adiantoides. Czekanowskiales are present and, despite the fact that only isolated linear leaves of the form genus Desmiophyllum have been found, Filippova suggested that these might belong to the genus Phoenicopsis (Terekhova and Filippova, 1984). Conifer remains are numerically dominated by Cephalotaxopsis and Sequoia, with subordinate occurrences of Thuja cretacea,

Elatocladus, Pityophyllum and Metasequoia. Among angiosperms, the large-leaved platanoids such as Arthollia, Pseudoprotophyllum and Paraprotophyllum are the most abundant, whilst smaller-leaved Trochodendroides and Zizyphus are also well represented. Other angiosperm genera include Menispermites, Araliaephyllum, Leguminosites, Dalbergites, Celastrinites, bViburnumQ (Viburniphyllum), Grewiopsis and Terechovia. 2.2. The eastern slopes of the northern Pekulney Range On the eastern slopes of the northern Pekulney Range, Upper Cretaceous deposits extend southwards from the Malaya Poperechnaya River, along the banks of the Poperechnaya and Uval’naya rivers, to the Kuiviveem River in the south (Figs. 1b and 4a). The lowermost Upper Cretaceous units in this sequence are represented by marine deposits of the Vesnovannaya Formation (Fig. 4). This is up to 1400 m thick and is divided into two subformations (Terekhova and Filippova, 1983), each being approximately 600 to 700 m thick. The Lower Subformation consists predominantly of conglomerates with rare

30

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

Table 1 Northern Pekulney Range floral distribution (modified from Filippova, 1994) Filippova’s taxa

Tylpegyrgynai Formation

Poperechnaya Formation

Subformation Lower Thallites tchucotica Philipp. Selaginella sp. Equisetites sp. Osmunda sp. Gleichenites onkilonica (Krysht.) Philipp. Birisia jelisejevii (Krysht.) Philipp. Coniopteris aff. bicrenata Samylina Coniopteris cf. anadyrensis Philipp. Arctopteris aff. rarytkinensis Vassilevsk Asplenium aff. dicksonianum Heer Hausmannia bipartita Samyl. et Shczep. Cladophlebis frigida (Heer) Sew. Cladophlebis aff. grandis Samylina Ochotopteris sp. Nilssonia yukonensis Hollick Ctenis sp. Ginkgo ex gr. adiantoides (Unger) Heer Krannera marginata (Heer) Seward Desmiophyllum (Phoenicopsis?) sp. Cephalotaxopsis heterophylla Hollick Cephalotaxopsis intermedia Hollick Cephalotaxopsis sp. Sequoia reichenbachii (Gein.) Heer Sequoia fastigiata (Sternb.) Heer Sequoia obovata Knowlton Sequoia sp. Elatocladus smittiana (Heer) Seward Thuja cretacea (Heer) Newb. Metasequoia ex gr. disticha (Heer) Miki Metasequoia cuneata (Newb.) Chaney Glyptostrobus groenlandicus Heer Cedrus sp. Pityocladus sp. Sciadopitys sp. Pityophyllum nordenskioldii (Heer) Nath. Menispermites efimovae Philipp. Trochodendroides pekulnejensis Philipp. T. sachalinensis (Krysht.) Krysht. T. speciosa (Ward) Berry T. vassilenkoi Iljinsk. et Romanova T. ex gr. richardsonii (Heer) Krysht. Trochodendrocarpus arcticus (Heer) Krysht. Arthollia pacifica Herman Arthollia insignis Herman Paraprotophyllum ignatianim (Krysht et Baik.) Herman Pseudoprotophyllum boreale (Dawson) Hollick Pseudoprotophyllum sp. Araliaephyllum speciosum Philipp. Araliaephyllum pekulneense Philipp. Araliaephyllum arenaria (Philipp.) Philipp.

+

Subformation Middle + + + +

+

+

+

+ + + +

Upper

+ + + + +

+ + +

+

+ +

+

+

+ + +

+ + + +

+ + + + +

+ + + + +

+ + + + + + +

+ + + + + + +

+ +

+ + + +

Upper

+ +

+ +

+ + +

+

+ + +

+ + + +

+ + + + +

+

+

+ +

+ + + +

+

+ + + +

+

+

+

+

+

+

+

+

+ + + + + + + + +

+ +

+ +

+ +

Middle

+ +

+ + + + +

Lower

+ +

+ + +

+ + + + + + + +

+ +

+

+

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

31

Table 1 (continued ) Filippova’s taxa

Tylpegyrgynai Formation

Poperechnaya Formation

Subformation

Subformation

Lower Leguminosites sp. Dalbergites simplex (Newb.) Seward Celastrophyllum sp. Celastrinites zakharovii Philipp. Zizyphus smilacifolia Budants Zizyphus electilis Hollick Zizyphus anadyrensis Philipp. Paliurus aff. visibilis Hollick Viburnum aff. whymperi Heer Viburnum aff. asperum Newberry Grewiopsis nemorosus Philipp. Smilax aff. grandifolia Lesquereux Terechovia anadyrensis Philipp. Terechovia intermedia Philipp. Hollickia quercifolia (Hollick) Krassilov Dicotylophyllum trilobatum Philipp. Dicotylophyllum microphyllum Philipp. Carpolithes grandis Philipp.

sandstone and siltstone beds. These fine-grained deposits yield Inoceramus sp. indet., Gokoyamaoceras sp. indet., and fragments of ammonites and echinoids. The Upper Subformation is largely composed of siltstones with interbedded sandstones, and rare occurrences of conglomerates and tuffs. Fossil mollusc remains include Inoceramus ex gr. korjakensis Ter., I. gradilis Perg., I. aff. tenuis Mant., I. aff. tychljawajamensis Ver., I. multiformis Perg., I. cf. cuvieri Sow., I. cf. hobetsensis Nagao et Mat., I. cf. mametensis Perg. and I. aff. concentricus var. costatus Nagao et Mat. The ammonite Scaphites sp. (S. pseudoaequalis Yabe or S. yonekurai Yabe) is also found. Terekhova (Terekhova and Filippova, 1983) correlated the earlier beds with the Inoceramus nipponicus zone (Cenomanian and Lower Turonian), and the later beds with the Inoceramus iburiensis zone (Upper Turonian), thus indicating that the Vesnovannaya Formation is Cenomanian to Turonian in age (Fig. 3). Conformably overlying the Upper Vesnovannaya Subformation are plant-bearing deposits of the Poperechnaya Formation, which in places is up to 1300 m thick. This is divided into three subformations (Terekhova and Filippova, 1983) (Fig. 4b). The Lower Subformation, approximately 550 m thick, is

Middle

Upper

Lower

+ + + + + + + +

+ +

Middle + + + + + + + + + + + + + + +

Upper

+

+

+

+ + +

composed of sandstones with subordinate layers and lenses of conglomerates with occasional layers of siltstones and tuffs. The upper part of this sequence is dominated by conglomerates with interbedded layers of gravels, and cross-bedded coarse-grained sandstones with rare occurrences of coal. Finer grained deposits are found in the Middle Subformation, 380 to 530 m thick, which contain an abundance of fossil plant remains. This subformation is represented by interbedded sandstones, siltstones, mudstones and coaly siltstones, small lenses of coal and tuffs. The Upper Subformation, 100 to 240 m thick, comprises interbedded conglomerates, gravels, sandstones and rare siltstones and tuffs. Fossil plant remains are found in all the subformations, but the Middle Subformation provides the richest source. The overall suite of facies is indicative of a floodplain environment. The Poperechnaya Formation is conformably overlain by marine deposits of the Otroginskaya Formation, 500 to 700 m thick (Fig. 4b). This is dominated by siltstones with interbedded sandstones and tuffs. Fossil mollusc remains found here include Inoceramus yokoyamai Nagao et Mat., I. cf. naumanni Yok., I. sp. indet. (I. cf. amakusensis Nagao et Mat.), Gaudryceras cf. denseplicatum (Jimbo), Neopuzosia

32

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

Fig. 4. Geology of the Poperechnaya–Kuiviveem interfluve, eastern slopes of the Northern Pekulney Range (Terekhova’s data; modified from Herman, 1999). (a) Geological map, (b) stratigraphy.

ishikawai (Jimbo) and Yokoyamaoceras cf. kotoi (Jimbo). According to Terekhova (Terekhova and Filippova, 1983) these fossils indicate that the locality belongs to the Inoceramus yokoyamai zone of Upper Coniacian to Lower Santonian age (Fig. 3). The Poperechnaya Formation, and hence the Poperechnaya Flora, is thus constrained biostratigraphically by underlying marine deposits of the Vesnovannaya Formation of Late Turonian age and by overlying marine deposits of the Otroginskaya Formation of Late Coniacian to Early Santonian age (Fig. 3). The age of the Poperechnaya Flora is thus Early

Coniacian and correlates with the Inoceramus uwajimensis zone. 2.2.1. Composition of the Poperechnaya Formation taphoflora Plant fossil remains are found in all three subformations of the Poperechnaya Formation, but are most prolific in the Middle Subformation, where they form large leaf accumulations. Filippova (1991, 1994) identified 62 different species, and regarded the taxonomic composition as sufficiently uniform throughout the sequence for it to be considered a

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

single floral association, subsequently named the Poperechnaya Flora by Herman (1999). Based on the number of species (Table 1), the Poperechnaya Flora is dominated by angiosperms (approximately 50%), followed by conifers, ferns and other groups of plants, as in the Tylpegyrgynai Flora. Ferns are rare, but include Osmunda sp. and Coniopteris. Cycadophytes are only represented by Ctenis sp., and ginkgoales by Ginkgo ex gr. adiantoides. Czekanowskiales may be represented by the genus Phoenicopsis, although only isolated linear leaves of Desmiophyllum have been found. Conifers are dominated by Cephalotaxopsis and Sequoia, often associated with Metasequoia and Thuja (Cupressinocladus). Angiosperm leaves are the most abundant and have the greatest taxonomic diversity. Platanoids form the dominant group, genera such as Arthollia, Paraprotophyllum and Pseudoprotophyllum being the most widespread, particularly in the Middle Subformation where they sometimes form large accumulations. Some platanoid leaves are more than 20 cm in length, but smaller leaves with acrodromous venation are represented by several species of Trochodendroides and Zizyphus. Other less frequently found genera include Menispermites, Araliaephyllum, bViburnumQ (Viburniphyllum), Celastrophyllum, Leguminosites, and Terechovia. 2.3. Age of Tylpegyrgynai and Poperechnaya formations The lower age limit of the Tylpegyrgynai Formation cannot be defined precisely due to the unconformable erosional contact with underlying Lower Cretaceous deposits. However, the floras from the Tylpegyrgynai and Poperechnaya formations have a similar taxonomic composition and were considered to represent a single floral association, named the Tylpegyrgynai Flora, by Filippova (Terekhova and Filippova, 1983, 1984). It is therefore believed that the lower age limit of the Tylpegyrgynai Formation is the same as that of the Poperechnaya Formation, i.e. Early Coniacian. As previously described, the upper age limit of both formations is constrained by overlying Upper Coniacian marine deposits belonging to the Inoceramus yokoyamai zone, and is defined as early Coniacian.

33

3. Methods and techniques 3.1. Photographic records Plant fossil remains from the Tylpegyrgynai and Poperechnaya formations collected by Terekhova and Filippova in 1981 and 1982 are stored at the Northeastern Geological Survey, Magadan, N.E. Russia. In 1994, the 252 angiosperm leaf specimens were cleaned and photographed by R.A. Spicer and A.B. Herman. The specimens were placed under low angle incident light and photographed using a Contax 167MT SLR camera with a Zeiss S-Planar 2.8/60 macrolens with 35 mm Ilford FP4 Plus film. Black and white prints (18.624.3 cm) were produced (Fig. 5). Drawings of a subset of 107 leaves were also made, the outlines and venation being drawn directly onto the photographs using black waterproof ink, after which the photographic images were bleached out using an aqueous solution of potassium iodide and iodine. The bleached images were then conventionally fixed, leaving only the line drawings of the leaves (Fig. 6). 3.2. CLAMP methodology 3.2.1. Morphotyping The 252 fossil angiosperm leaf specimens of the Northern Pekulney Range Flora were subjected to a physiognomic analysis after morphotyping based on leaf architecture and vein patterns. Through observation of their general characteristics, they were first divided into 10 broad groups, namely, trochodendroids, zizyphoids, cocculoids, menispermoids, araliaephylls, platanoids, viburniphylls, cissitiphylls, entire margined forms, and a small number of individual specimens that could not be assigned to other groups and which were aggregated into dmiscellaneousT. These groups were then subdivided into separate morphotypes based on a detailed study of leaf shape and venation. Reliable CLAMP results depend upon morphotype categorisation that reflects original species differentiation in the source flora. To base morphotype boundaries solely on the architectural features used in CLAMP analysis would, in a sense, be circular, so venation patterns are also routinely considered when defining morphotypes. Even then, the observed morphological intergradational characteristics of many Late Creta-

34

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

ceous angiosperm leaves frustrate the application of architectural and, in particular, venation description schemes, based on living angiosperm leaves. For this reason, vein pattern analysis was based on a system explicitly designed for the morphological intergradation characteristic of Late Cretaceous leaves (Spicer, 1986), combined with more conventional terminologies where appropriate (Dilcher, 1974). Initial morphotyping yielded 97 distinct leaf forms, but this was regarded as bover-splittingQ because some morphotype

distinctions had been based on single features. Critical analysis, recognising consistent characteristics at a broader scale, reduced this number to 57 (Fig. 6). A full description of the criteria used to define each morphotype, and the audit trail of morphotype aggregation, can be found at the CLAMP website (Spicer and Wolfe, 2001, CLAMP web page). In practice, where assemblages are particularly rich, errors in morphotype differentiation between a few taxa have negligible effect on the overall CLAMP results (Spicer, unpub-

Fig. 5. Selected leaves from the Northern Pekulney Range Flora (m=morphotype). m-1, m-2 trochodendroids; m-12, m-14 zizyphoids; m-17 cocculoid; m-20 menispermoid; m-23 araliaephyll; m-28, m-29, m-31, m-32 platanoids; m-35, m-39 viburniphylls; m-41, m-47 cissitiphylls; m50, m-53 entire margins; m-54 miscellaneous. Scale bars are 1 cm.

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

35

Fig. 5 (continued).

lished data). Furthermore, CLAMP analysis is ideally most reliably performed on material from well-documented sections where taphonomic biases can be assessed. Such data are not available from this remote region. However, the large size of the collection (over 250 specimens), derived from several locations, minimised possible biases (Spicer et al., 2002) and therefore justifies the use of CLAMP in this instance. 3.2.2. The CLAMP technique A full description of the procedure for running a Climate Leaf Analysis Multivariate Program

(CLAMP) is given at the CLAMP website (Spicer and Wolfe, 2001, CLAMP web page). The 57 morphotypes present in the Northern Pekulney Range Flora were analysed using the CLAMP technique of Wolfe (1993), with modifications by Herman and Spicer (1997). Separate CLAMP analyses were also carried out on the floras from each of the two formations contributing to the Northern Pekulney Range Flora, namely the Tylpegyrgynai Formation Flora (27 morphotypes), and the Poperechnaya Formation Flora (39 morphotypes).

36

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

In CLAMP the architecture of woody dicot leaves from modern day vegetation growing under known climatic conditions is used as a reference data set against which to compare the architecture of leaves found in a fossil assemblage. The total reference data set comprises 173 modern Northern Hemisphere vegetation sites scored for 31 leaf characters (Wolfe,

1993; Herman and Spicer, 1997) and correlated with 11 climate variables. The analysis presented in this paper is based on a smaller reference data set of 144 samples (Physg3br) which excludes the so-called dsubalpine nestT, considered to represent an anomalous group derived from 29 sites which experience warm month mean temperatures of less than 16 8C

Fig. 6. Leaf morphotypes from the Northern Pekulney Range Flora (m=morphotype). m-1 to m-7 trochodendroids; m-8 to m-16 zizyphoids; m17 to m-19 cocculoids; m-20 to m-22 menispermoids; m-23 to m-26 araliaephylls; m-27 to m-33 platanoids; m-34 to m-40 viburniphylls; m-41 to m-48 cissitiphylls; m-49 to m-53 entire margins; m-54 to m-57 miscellaneous. Scale bars are 1 cm.

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

37

Fig. 6 (continued).

and cold month means of less than 3 8C. Leaves from subalpine nest sites tend to be small-sized, without teeth, due to exposure to physiological drought resulting from the low temperatures that they experience (Wolfe, 1993). In the Northern Pekulney Range Flora, large leaf sizes and sedimentological evidence such as peat deposits (now coal), suggest that drought was unlikely to have been a significant constraint on growth, so exclusion of the subalpine nest from the reference data set appears to be justified. Eight climate variables are used here: mean annual temperature

(MAT), warm month mean temperature (WMMT), cold month mean temperature (CMMT), mean growing season precipitation (MGSP), mean monthly growing season precipitation (MMGSP), precipitation during the three consecutive wettest months (3WET), precipitation during the three consecutive driest months (3DRY), and length of the growing season (LGS). In order to produce statistically reliable results, the CLAMP technique requires at least 20 leaf morphotypes from any given fossil site, and these criteria were met in the three analyses undertaken.

38

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

Fig. 6 (continued).

The CLAMP methodology presented in Wolfe (1993) was used in conjunction with modifications specified by Herman and Spicer (1997), the most significant modification being the use of Canonical Correspondence Analysis (ter Braak, 1986) with the computer program CANOCO (ter Braak, 1987–1992), instead of Correspondence Analysis (CA), as the multivariate statistical engine. This enables assignment of climate vectors in an objective way, and allows the inclusion of fossil data as passive elements (Herman and Spicer, 1997). Canonical Correspondence Analysis is a direct ordination method which was here used to analyse a leaf character versus site matrix together with a climate variable versus site matrix simultaneously in multidimensional space, in order to determine correlations between leaf characters and climate variables. Environmental data are not used to position the sites,

since the sites are ordered by their character scores, and characters are ordered by their distribution among the sites. The sites are therefore positioned relative to each other in multidimensional space by the leaf physiognomic characters pertaining at each particular site. CANOCO identifies vectors for each of the environmental variables and positions them relative to the major axes of variation in the sample space. Thus the correlation between sample sites or leaf characters and environmental variables can be determined directly (Fig. 7). The distribution of vegetation samples from 144 modern sites and 3 fossil sites are positioned relative to each other, based on the physiognomic leaf characters of woody dicots from each site. Fossil leaves are analysed and scored using the same characteristics and criteria as modern leaves, but the data are entered as passive elements. They are therefore included in the final ordination diagram but

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

do not contribute to the determination of the ordination axes. The 31-dimensional leaf character space has been collapsed to two dimensions, in which axes 1 and 2 represent the two axes of greatest variation in the data. The modern vegetation sites are coded according to their mean annual temperature (MAT), and it can be seen that the sites with low MATs plot on the left and those with high MATs on the right. The position of the MAT vector in axis 1/axis 2 space is shown. Also included is the position of the vector for mean precipitation during the growing season (MGSP). Diagrams displaying the six other environmental vectors used in this study are available at the CLAMP website (Spicer and Wolfe, 2001, CLAMP web page). In most cases the environmental vectors are not coincident with the major axes of variation, and it is not appropriate to calibrate axis 1 or 2 in terms of climate variables. Therefore, trigonometric algorithms are used to project site scores onto the environmental vectors. Thus in Fig. 8, which illustrates mean annual temperature calibration, the MAT vector score, in axis 1 versus axis 2 space (rather than

39

Fig. 8. Example of CLAMP analysis using CANOCO. MAT (mean annual temperature) vector scores are plotted against observed MAT values.

axis 1 score) is plotted against observed MAT data from modern sites. The position of a fossil site along the MAT vector can be used to estimate its ancient MAT, by observing where its vector score intercepts the regression line. The scatter of plots about the regression line reflects the statistical uncertainty in estimating MAT, and in this case one standard deviation of the residuals about the regression line is 1.2 Celsius degrees. In previous publications, the regression algorithms were correlated in axis 1/axis 2 space (Herman and Spicer, 1997), but in the spreadsheets available at the CLAMP website regression algorithms correlated in axes 1–4 space are currently used to determine the palaeoclimate parameters.

4. Results

Fig. 7. CLAMP analysis using CANOCO showing distribution of modern and fossil sites in axis 1/axis 2 space as defined by leaf characteristics. Percent variance accounted for by axis 1 is 50.5%, axis 2 is 10.4%. Pe=Northern Pekulney Range Flora; Po=Poperechnaya Formation Flora; Ty=Tylpegyrgynai Formation Flora. MAP vector=mean annual temperature environmental vector; MGSP vector=mean growing season precipitation environmental vector.

The Northern Pekulney Range Flora and its component floras from the Tylpegyrgynai and Poperechnaya formations were subjected to CLAMP analysis. Fig. 7 shows the distribution of modern vegetation sites and fossil sites in respect of axes 1 and 2, with vectors for mean annual temperature (MAT) and mean precipitation during the growing season (MGSP). It can be seen that the fossil site for the Northern Pekulney Range Flora (Pe) plots within the cloud of modern

40

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

reference points, towards the cooler end of the MAT vector, and towards the drier end of MGSP vector. Fig. 9 shows the results for the eight respective environmental vectors predicted for the Northern Pekulney Range Flora and, separately, the results for its two component floras, the Tylpegyrgynai Formation and Poperechnaya Formation floras. Site scores for the MAT vector give estimates of 8.1 8C for the Northern Pekulney Range Flora, 9.4 8C for the Tylpegyrgynai Formation Flora and 7.3 8C for the Poperechnaya Formation Flora (Fig. 9a). The standard deviation of the residuals about the mean of the second order polynomial regression line is 1.2 Celsius

degrees. Although the values are not significantly different, it can be seen that MAT predicted by the Tylpegyrgynai Formation Flora is considerably higher than that by the Poperechnaya Formation Flora. It should be noted that results for the Northern Pekulney Range Flora do not represent an average of results for the Tylpegyrgynai Formation and Poperechnaya Formation floras. For each flora, predicted results depend on the number of leaf specimens preserved, the number and range of morphotypes present, the quality of preservation of each of the 31 characters scored per morphotype, and the association of particular leaf characters with particular environmental vectors.

Fig. 9. Results of CLAMP analysis using CANOCO for the Northern Pekulney Range Flora, the Tylpegyrgynai Formation Flora and the Poperechnaya Formation Flora. (a) Mean annual temperature (MAT), (b) warm month mean temperature (WMMT), (c) cold month mean temperature (CMMT), (d) length of the growing season (LGS), (e) mean growing season precipitation (MGSP), (f) mean monthly growing season precipitation (MMGSP), (g) precipitation during the three consecutive wettest months (3WET), (h) precipitation during the three consecutive driest months (3DRY).

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

41

of 692.4 mm for the Tylpegyrgynai Formation Flora is considerably higher than 373.0 mm for the Poperechnaya Formation Flora. Results for mean monthly growing season precipitation (MMGSP) are also subject to a large uncertainty (S.D. of residuals 36.9 mm), but it can be seen that precipitation predicted for the Tylpegyrgynai Formation Flora, at 83.2 mm, is again higher than that for the Poperechnaya Formation Flora, at 73.1 mm (Fig. 9f). The total precipitation during the three consecutive wettest months (3WET) results are subject to large uncertainty (S.D. of residuals 140.3 mm), but here again the predicted precipitation values are considerably higher for the Tylpegyrgynai Formation Flora, at 358.6 mm, than for the Poperechnaya Formation Flora, with 252.9 mm (Fig. 9g). Results for the total precipitation during the three consecutive driest months (3DRY) exhibit high uncertainty (S.D. of residuals 93 mm), and the floras predict values that are not significantly different from each other (Fig. 9h). Although they do not appear to have suffered from drought, it can be seen that the Tylpegyrgynai Formation Flora has a lower precipitation value during the three driest months, at 125.3 mm, than the Poperechnaya Formation Flora, at 166.5 mm. Table 2 gives a summary of these results including the goodness of fit (r 2) of the data to the regression lines, the standard deviation of the residuals and the completeness of the data. The indications of uncertainty in Table 2 are statistical measures and do not

Predicted warm month mean temperatures (WMMT) for the floras range from 17.1 8C to 18.2 8C (Fig. 9b). These lie within the uncertainty of the predictions, since the standard deviation of the residuals is 1.6 Celsius degrees, and the results may therefore be considered as almost identical. Results for the cold month mean temperature (CMMT) range from 0.9 8C for the Tylpegyrgynai Formation Flora to 2.7 8C for the Poperechnaya Formation Flora (Fig. 9c), although all results lie within the uncertainty of the predictions, the standard deviation of the residuals being 1.9 Celsius degrees. From the above results it can therefore be seen that the Poperechnaya Formation Flora exhibits the lowest temperatures in all categories (MAT, WMMT and CMMT). The length of the growing season (LGS) is defined as the total number of months (or parts thereof) when the mean temperature is higher than 10 8C. Results are tightly clustered (Fig. 9d) and fall within the uncertainty of the predictions (S.D. of residuals 0.7 months), but predictions suggest that the Tylpegyrgynai Formation Flora had a growing season of 5.9 months and that of the Poperechnaya Formation Flora 1 month less. The mean growing season precipitation (MGSP) reflects the adaptation of leaves to environmental conditions during their period of active growth (Fig. 9e). The standard deviation of the residuals about the regression line is 335.9 mm. Despite the large uncertainty, it can be seen that a precipitation value

Table 2 CLAMP results of Coniacian floras from Northeastern Russia and Northern Alaska Fossil floras/palaeolatitudes/completeness, C

Environmental variables MAT (8C)

Northeastern Russia Northern Pekulney Range/788N/C=0.74 Tylpegyrgynai Formation/788N/C=0.79 Poperechnaya Formation/788N/C=0.68 Northwest Kamchatka/728N Alaska North Slope/758N Standard deviation r2

WMMT (8C)

CMMT (8C)

LGS (months)

MGSP (mm)

MMGSP (mm)

3WET (mm)

3DRY (mm) 157.9 125.3 166.5 138.5

8.1 9.4 7.3 9.3

17.8 18.2 17.1 17.6

1.5 0.9 2.7 1.1

5.3 5.9 4.9 5.7

484.5 692.4 373.0 464.4

78.8 83.2 73.1 73.9

295.7 358.6 252.9 282.8

13.3 1.2 0.95

18.6 1.6 0.83

8.2 1.9 0.93

7.2 0.7 0.93

398.5 335.9 0.87

54.4 36.9 0.86

232.0 140.3 0.86

77.5 93.0 0.86

MAT=mean annual temperature; WMMT=warm month mean temperature; CMMT=cold month mean temperature; LGS=length of the growing season; MGSP=mean growing season precipitation; MMGSP=mean monthly growing season precipitation; 3WET=precipitation during the three consecutive wettest months; 3DRY=precipitation during the three consecutive driest months. C=completeness (data not available for Northwest Kamchatka or North Slope of Alaska). r 2=goodness of fit.

42

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

reflect taphonomic or evolutionary effects, or the lack of analogues between Late Cretaceous high latitude and modern environments. Evidence supporting the robustness of the CLAMP technique was provided by Herman and Spicer (1997), and it can be seen that the Northern Pekulney Range Flora plots well within the cloud of modern reference points (Fig. 7). Taphonomic effects have been minimised, as the northern Pekulney Range fossil leaves were collected from a variety of sedimentary facies within a geographically restricted area. The high quality of leaf preservation, including some excellent venation, provides evidence of a short transportation period and little decay prior to burial. Since some leaf characters, such as apices or delicate margins, may be vulnerable to damage or loss during burial, diagenesis and collection, a completeness statistic, C, is shown for each flora. The completeness statistic, C, is given by C=F/P where F is the actual number of data matrix cells filled when scoring a fossil assemblage and P is the maximum number of data matrix cells that would be filled if all character states were to be scored for all taxa. The completeness statistic for the Northern Pekulney Range Flora is 0.74, for the Tylpegyrgynai Formation Flora 0.79 and for the Poperechnaya Formation Flora 0.68. These values show that missing characters are unlikely to have significantly degraded palaeoclimate predictions for the overall Northern Pekulney Range Flora or for the Tylpegyrgynai Formation Flora, but may have had a small adverse effect on predictions for the Poperechnaya Formation Flora. Results shown in Table 2 for the Northern Pekulney Range, the Tylpegyrgynai Formation and the Poperechnaya Formation floras suggest a warm temperate climate (climate C of Ko¨ppen, 1936), with winter temperatures around freezing point, and a moist regime with no pronounced dry season. However, the Tylpegyrgynai Formation, with 358.6 mm of precipitation in the three wettest months and 125.3 mm in the three driest months, shows a greater difference between the wet and dry seasons than the Poperechnaya Formation, where precipitation levels are lower than Tylpegyrgynai in the wettest months and marginally higher than Tylpegyrgynai in the driest months. For comparison, results from two additional Coniacian floras, namely Northwest Kamchatka at a palaeolatitude of 728N and the North Slope of Alaska

at 758N (Herman and Spicer, 1997) are shown (Table 2), based on the 144 reference data set (Physg3br). It can be seen that, throughout the year, temperatures for the Alaskan vegetation are higher than those for the Northern Pekulney Range Flora, the mean annual temperature (MAT) and cold month mean temperature (CMMT) being significantly higher. When viewed in three-dimensional space, the Alaskan site is seen to occupy a space devoid of comparable modern sites. This separation is considered to be indicative of the unique environmental conditions experienced by this taphoflora (Herman and Spicer, 1997). For the Northwest Kamchatka Flora, MAT and CMMT are slightly higher than the Northern Pekulney Range Flora, whilst warm month mean temperature (WMMT) is slightly lower, but all values fall within the uncertainty of the predictions. The length of the growing season for the Alaskan vegetation, at 7.2 months, is significantly higher than that predicted for the Northern Pekulney Range Flora. Precipitation values for the North Slope of Alaska and Northwest Kamchatka floras are seen to be consistently lower than those predicted for the Northern Pekulney Range Flora, although all results fall within the uncertainty of the predictions.

5. Discussion The Poperechnaya Formation Flora is similar in taxonomic composition to the Tylpegyrgynai Formation Flora (Table 1), but the Poperechnaya Formation Flora has a greater number of species (as identified by Filippova). Filippova considered that, since almost all the species from the Tylpegyrgynai Formation Flora were present in the Poperechnaya Formation Flora, these two floras comprised a single floral association, which she defined collectively as the Tylpegyrgynai Flora (Terekhova and Filippova, 1984; Filippova and Abramova,1993). However, it is worth noting that from a total of 68 tentatively identified species, only 44 are common to both formations. Six species within the Tylpegyrgynai Formation are not found in the Poperechnaya Formation, including the ferns Hausmannia bipartita and Ochotopteris sp., the cycadophyte Nilssonia yukonensis Hollick, the conifer Sciadopitys sp., and the angiosperm Dicotylophyllum trilobatum Philipp. It should also be noted that

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

eighteen species within the Poperechnaya Formation have not been found in the Tylpegyrgynai Formation. These include the fern Coniopteris cf. anadyrensis Philipp., cycadophyte Ctenis sp., conifers Cephalotaxopsis sp. and Metasequoia cuneata (Newb.) Chaney, together with 13 different species of angiosperm (Table 1). In view of the above differences, separate CLAMP analyses were performed on the floras from the Tylpegyrgynai and Poperechnaya formations in an attempt to obtain further clarification of possible environmental differences between them. The study was based on physiognomic differences, particularly leaf shape and venation, without reference to previous taxonomic classification or pre-assigned species. Of the 57 angiosperm morphotypes in the collective flora, it was possible to assign provenance of 50 to either the Tylpegyrgynai or Poperechnaya formations (for the remaining 7 morphotypes, locality data were incomplete). From the 50 morphotypes, 16 were common to both formations, 11 were found only in the Tylpegyrgynai Formation and 23 only in the Poperechnaya Formation. The CLAMP requirement of a minimum of 20 leaf morphotypes from a given fossil site was thus met by both localities, the Tylpegyrgynai Formation having 27 morphotypes, and the Poperechnaya Formation 39 morphotypes. A set of climate signals was derived from each flora enabling a quantitative analysis to be made of the palaeoclimatic signatures that applied during their growth and an assessment to be made of any differences. It should be noted that although the Tylpegyrgynai Formation Flora has a smaller number of morphotypes, its completeness statistic, C, is higher than that for the Poperechnaya Formation Flora. This suggests that, although Tylpegyrgynai has fewer morphotypes, its specimens are in a better state of preservation and therefore produce a more reliable climate signal. By comparing climate predictions from the floras of the Tylpegyrgynai and Poperechnaya formations (Fig. 9, Table 2), it can be seen that temperatures predicted by the Tylpegyrgynai Formation Flora are consistently higher than those for the Poperechnaya Formation Flora (Fig. 9a,b,c), although all results lie within the uncertainty of the predictions. A possible explanation for this difference in temperatures may be due to the proximity of the coastal plain locality of the

43

Poperechnaya Formation to the northern proto-Pacific Ocean. It has been suggested that during the Late Cretaceous the waters of the northern proto-Pacific Ocean were isolated from those of the warmer Arctic Ocean by a Bering land bridge (Funnell, 1990). In the winter, these waters may have been close to freezing (Herman and Spicer, 1997), and may have influenced the suggested 2.7 8C cold month mean temperature predicted by the Poperechnaya Formation Flora. This difference in temperatures is also reflected in the longer growing season predicted for the Tylpegyrgynai Formation Flora (Fig. 9d). Comparisons of precipitation predictions from the two floras show that, although values lie within the uncertainty of the method, wetter conditions are predicted by the more westerly Tylpegyrgynai Formation Flora during the growing season and during the three wettest months (Fig. 9e,f,g). During the three consecutive driest months, higher precipitation is suggested for the Poperechnaya Formation Flora (Fig. 9h). However, it is difficult to draw any firm conclusions from precipitation data, since the uncertainties are large. Taking into consideration the noted differences in morphology and taxonomy between the floras derived from the Tylpegyrgynai and Poperechnaya formations, and the differences in their predicted climatic conditions, it is here proposed that the floras from the two formations be considered as separate entities, namely the Tylpegyrgynai Formation Flora and the Poperechnaya Formation Flora, to be known collectively as the Northern Pekulney Range Flora. By looking at the collective Northern Pekulney Range Flora and its climate predictions, it is possible to suggest growing season constraints and an overwintering strategy for this vegetation. Factors affecting the length of the growing season include water, light and temperature, but vegetation characteristics such as the large leaves of platanoids and viburniphylls, together with the preservation of peat (now coal) in the same units, indicate that water was not a limiting factor to growth. The most likely constraints were therefore light and/or temperature. Spicer and Parrish (1990) summarised evidence, based on Alaskan flora, suggesting that obliquity in the Late Cretaceous was not significantly different to that of the Present. Therefore, assuming present day obliquity, at a Coniacian palaeolatitude of 788N applicable

44

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

to the Northern Pekulney Range Flora, there would have been a total of 4.2 months of continuous daylight (Anonymous, 1978) bounded by 3 months when the sun was above the horizon for more than 4 h a day. Given the predicted temperatures for the Northern Pekulney Range Flora, for most of the 7.2 months when there was continuous daylight or rapidly changing daylength, the temperature would have been at or above 10 8C (the temperature used to define the lower limit of growing season temperature). This suggests that light, rather than temperature, would have been the limiting factor for the predicted 5.3 month growing season. For the Northern Pekulney Range Flora, the combination of 4.8 months when there was continuous darkness or fewer than 4 h of daylight per 24 h period, and a cold month mean temperature of 1.5 8C, implies that some elements of the vegetation may have adopted an evergreen overwintering strategy. Respiration drain and consequent tissue death would not have been a threat, since respiration becomes minimal at temperatures below 4 8C (Read and Francis, 1992; Herman and Spicer, 1997). Although there are no modern living analogues for this Late Cretaceous vegetation, it is possible that some of the conifers (Table 1) may have adopted an evergreen strategy. For example, Thuja cretacea may be related to the living evergreen Thuja, whilst the extinct Sequoia reichenbachii and Sequoia fastigiata (Sternb.) Heer exhibit morphologies that may have been consistent with an evergreen habit. Elatocladus smittiana (Heer) Seward may also have adopted an evergreen overwintering strategy. However, there is also evidence of a deciduous overwintering strategy, in particular a single bedding plane on which different morphotypes of fossil leaves have been preserved, including small-leaved zizyphoids and cissitiphylls. The fine-grained nature of the sediments precludes transportation by fast-flowing water, and the well-preserved fossil leaf remains, lacking both biological and mechanical degradation, indicate a short transportation period and a depositional environment close to the source vegetation. Burial on the same bedding plane indicates deposition within a few days, or at most weeks, of one another. The minimal exposure to potential predation is therefore consistent with synchronous leaf fall. This suggests that elements of both strategies were present

in the flora, with deciduousness being prevalent. Research by Beerling and Osborne (2002) and Royer et al. (2003) raises the possibility of complex causes for phenological strategies under such combinations of polar light regime and warm winter temperatures, but these studies are inevitably restricted to modern taxa. Results of the CLAMP analysis from the Northern Pekulney Range Flora were compared with those of other Coniacian floras growing at high latitudes. Table 2 includes data from the Northwest Kamchatka Flora at a palaeolatitude of 728N and from the North Slope of Alaska Flora at 758N. Herman and Spicer (1997) suggested that the warmer temperatures of the Alaska site, compared to the more southerly Kamchatka site, could be explained by (a) the effects of a warm Arctic Ocean isolated from the waters of the colder northern proto-Pacific Ocean by a Bering land bridge (Funnell, 1990), or (b) the marginal effects of cold air masses developed in winter over the Asian continental interior (Price et al., 1995), which influenced the Kamchatka site. Temperatures for the northern Pekulney Range are consistently lower than those from the Alaska site, showing significant differences in mean annual temperature (MAT) and cold month mean temperature (CMMT). For the Kamchatka location, MAT and CMMT results are slightly higher than those for the northern Pekulney Range and lie within the uncertainty of the predictions. The results are compatible with the higher latitudinal position of the northern Pekulney Range site compared with the other Coniacian sites, and with its isolation from any influence of a warm polar ocean, both by the Bering land bridge and by the Okhotsk-Chukotka volcanic highlands to the north and west. It may also have been influenced in the winter by the cold waters of the northern protoPacific and by the cold air from the continental interior. A prevailing west to east wind direction during the Late Cretaceous is consistent with conclusions reached by Kelley et al. (1999), as shown by the lack of Cretaceous volcaniclastics in the Vilui Basin to the west of the Okhotsk-Chukotka volcanic belt, but abundant Late Cretaceous bentonites in northern Alaska to the east. Despite large uncertainties, the precipitation values for the northern Pekulney Range site are consistently higher than those for the other Coniacian sites (Table 2).

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

The shorter length of the growing season for the Northern Pekulney Range Flora (5.3 months), compared with the North Slope of Alaska Flora (7.2 months) and the Kamchatka Flora (5.7 months), reflects its higher latitude and consequently more polar light regime, with 4.8 months when there was either continuous darkness or fewer than 4 h each day when the sun rose above the horizon.

6. Conclusion Plant fossil remains preserved in the Lower Coniacian sediments of the northern Pekulney Range in northeastern Russia represent one of the richest and most diverse Late Cretaceous floras. Sixty-eight different species of plants have been identified from the collective Northern Pekulney Range Flora of which 50 are found in the Tylpegyrgynai Formation Flora on the western slopes of the mountain range and 62 in the Poperechnaya Formation Flora on its eastern slopes, 44 species being common to both formations. A study of the fossil angiosperm leaves from the Northern Pekulney Range Flora identified 57 morphotypes which were subjected to CLAMP analysis. Results suggested a warm temperate climate with a mean annual temperature of 8.1 8C, and a mean annual temperature range of approximately 19 8C, despite the high palaeolatitude (788N) and consequent polar insolation regime. Light constraint is therefore considered to be the main factor determining the short 5.3 month growing season. The combination of restricted light and a cold month mean temperature of 1.5 8C, may have resulted in elements of the vegetation, particularly some conifers, adopting an evergreen overwintering strategy, although there is also abundant evidence of deciduousness. Precipitation values were subject to high uncertainties but yielded a mean growing season precipitation of 484.5 mm, with 157.9 mm during the three consecutive driest months, suggesting a moist regime lacking pronounced drought. Separate CLAMP analyses of 27 morphotypes from the Tylpegyrgynai Formation Flora and 39 from the Poperechnaya Formation Flora indicated that temperatures predicted for the former were consistently higher than those for the latter, with a marked 2.1 8C difference in their mean annual temperatures,

45

and 3.6 8C difference in their cold month mean temperatures. It is suggested that the Poperechnaya Formation may have been influenced by its greater proximity to the cold northern proto-Pacific Ocean which, during the Late Cretaceous was isolated from the warmer Arctic Ocean by a Bering land bridge. Precipitation levels also differed between the two floras, the Tylpegyrgynai Formation Flora generally having higher predicted values than the Poperechnaya Formation Flora. Although the differences in predicted environmental variables between the two floras were not statistically significant, when considered in conjunction with the differences in taxonomy and morphology, it is here proposed that they should be considered as two distinct floras, to be known collectively as the Northern Pekulney Range Flora. Comparisons between the Northern Pekulney Range Flora and other high latitude Coniacian floras were compatible with its higher latitudinal position, isolation from the influence of a warm polar ocean, and proximity to the northern proto-Pacific.

Acknowledgements The valuable assistance and advice given by Bob Spicer, Earth Sciences Department, Open University, and Alexei Herman, Geological Institute, Russian Academy of Sciences, are gratefully acknowledged. Warm thanks go to John Taylor, Earth Sciences Department, Open University, for his help with images, and to Galina Filippova, Northeastern Geological Survey, Magadan, for allowing access to her fossil plant collections.

References Anonymous, 1978. C.I.A. Handbook, Polar Regions Atlas. Natl. Foreign Assess. Center, C.I.A., 66 pp. Beerling, D.J., Osborne, C.P., 2002. Physiological ecology of Mesozoic polar forests in a high CO2 environment. Ann. Bot. 89, 329 – 339. Belyi, V.F., 1994. Geologiya Okhotsko-Chukotskogo Vulkanogennogo Poyasa (Geology of the Okhotsk-Chukotka Volcanogenic Belt). Far East Branch, Russian Acad. Sci., Magadan (76 pp., in Russian). Dilcher, D.L., 1974. Approaches to the identification of angiosperm leaf remains. Bot. Rev. 40 (1), 1 – 157.

46

H.J. Craggs / Palaeogeography, Palaeoclimatology, Palaeoecology 217 (2005) 25–46

Filippova, G.G., 1991. New angiosperms from the Coniacian deposits of the northern part of the Pekulney Range. Palaeontol. J. (Mosc.) 3, 85 – 93 (in Russian). Filippova, G.G., 1994. Coniacian flora of the northern part of the Pekulney Range. Kolyma 3, 13 – 21 (in Russian). Filippova, G.G., Abramova, L.N., 1993. Late Cretaceous Flora of North-Eastern Russia. Nedra, Moscow (348 pp., in Russian). Funnell, B.M., 1990. Global and European Cretaceous shorelines, stage by stage. The global shoreline maps. In: Ginsburg, R.N., Beaudoin, B. (Eds.), Cretaceous Resources, Events and Rhythms. Kluwer Academic Publishers, Dordrecht, pp. 221 – 235. Herman, A.B., 1999. Melovaya Flora Anadyrsko-Koryakskogo Subregiona (Severo-Vostok Rossii): Sistematicheskiy Sostav, Vozrast, Stratigraficheskoye i Florogeneticheskoye Znacheniye (Cretaceous Flora of the Anadyr-Koryak Subregion (Northeastern Russia): Systematics, Age, Stratigraphic and Florogenic Significance). GEOS, Moscow, 122 pp. (Geological Institute, Russian Acad. Sci. Trans. 529) (in Russian). Herman, A.B., Spicer, R.A., 1997. New quantitative palaeoclimate data for the Late Cretaceous Arctic: evidence for a warm polar ocean. Palaeogeogr. Palaeoclimatol. Palaeoecol. 128, 227 – 251. Kelley, S.P., Spicer, R.A., Herman, A.B., 1999. New 40Ar/39Ar dates for Cretaceous Chauna Group tephra, north-eastern Russia, and their implications for the geologic history and floral evolution of the North Pacific region. Cretac. Res. 20, 97 – 106. Kfppen, W.P., 1936. Das geographische System der Klimate. In: Kfppen, W.P., Geiger, R. (Eds.), Handbuch der Klimatologie. Gebrqder Borntraeger, Berlin. Price, G.D., Sellwood, B.W., Valdes, P.J., 1995. Sedimentological evaluation of General Circulation Model simulations of the bGreenhouseQ Earth Cretaceous and Jurassic case studies. Sediment. Geol. 100, 159 – 180. Read, J., Francis, J., 1992. Responses of some Southern Hemisphere tree species to a prolonged dark period and their implications for high-latitude Cretaceous and Tertiary floras. Palaeogeogr. Palaeoclimatol. Palaeoecol. 99, 271 – 290.

Royer, D.L., Osborne, C.P., Beerling, D.J., 2003. Carbon loss by deciduous trees in a CO2-rich ancient polar environment. Nature 424, 60 – 62. Skelton, P.W., Spicer, R.A., Kelley, S.P., Gilmour, I., 2003. In: Skelton, P.W. (Ed.), The Cretaceous World. Cambridge Univ. Press, Cambridge. Smith, A.C., Hurley, A.M., Briden, J.C., 1981. Phanerozoic Paleocontinental World Maps. Cambridge University Press. 102 pp. Spicer, R.A., 1986. Pectinal veins: a new concept in terminology for the description of dicotyledonous leaf venation patterns. Bot. J. Linn. Soc. 93, 379 – 388. Spicer, R.A., Parrish, J.T., 1990. Late Cretaceous–early Tertiary paleoclimates of northern high latitudes: a quantitative view. J. Geol. Soc. (Lond.) 147, 329 – 341. Spicer, R.A., Wolfe, J.A., 2001. CLAMP web page http:// tabitha.open.ac.uk/spicer/CLAMP/Clampset1.html. Spicer, R.A., Ahlberg, A., Herman, A.B., Kelley, S.P., Raikevich, M.I., Rees, P.M., 2002. Palaeoenvironment and ecology of the middle Cretaceous Grebenka flora of northeastern Asia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 184, 65 – 105. ter Braak, C.J.F., 1986. Canonical Correspondence Analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67, 1167 – 1179. ter Braak, C.J.F., 1987–1992. CANOCO—a FORTRAN program for canonical correspondence ordination. Microcomputer Power, Ithaca, NY. 95 pp., +software. Terekhova, G.P., Filippova, G.G., 1983. On the composition and age of the Poperechnaya Formation (Pekulney Range, north-eastern USSR). Rep. USSR Acad. Sci. 269 (4), 911 – 914 (in Russian). Terekhova, G.P., Filippova, G.G., 1984. Tylpegyrgynai Formation and its floral assemblage (Pekulney Range, north-eastern USSR). Rep. USSR Acad. Sci. 278 (6), 1443 – 1446 (in Russian). Vakhrameev, V.A., 1991. Jurassic and Cretaceous Floras and Climates of the Earth. Cambridge Univ. Press, Cambridge. Wolfe, J.A., 1993. A method of obtaining climatic parameters from leaf assemblages. U.S. Geol. Surv. Bull. 2040 (71 pp.).