1 Journal of Biogeography SUPPORTING INFORMATION The De Geer, Thulean and Beringia routes: key concepts for understanding early Cenozoic biogeography Leonidas Brikiatis Appendix S1: The De Geer, Thulean, and Beringia routes: early and recent geotectonic evidence THE DE GEER ROUTE Etymology The Swedish geologist Baron Gerard Jacob De Geer (1858–1943) is known for his papers establishing the glacial isostatic origin (see Mörner, 1979). In 1926, De Geer first referred to the existence of a physiographic lineament following the western continental slope of the Barents Shelf from Norway to Svalbard and the north-eastern continental slope of Greenland and Canada (see Horsfield & Maton, 1970). Subsequent authors suggested that extensive dextral strike-slip movements occurred along this lineament during the continental drift of Greenland and North America away from Eurasia. Harland (1969) named the fault system between Greenland and Spitsbergen Island (the Spitsbergen fracture zone) the ‘De Geer Zone’. In biogeography, Szalay & McKenna (1971) first introduced the concept of the De Geer route. They concluded that: …if a North Atlantic dry-land dispersal route existed, it lay north of the Norwegian Sea, crossing the De Geer dextral transform fault zone (Spitsbergen fracture zone) from Greenland to a juxtaposed Spitsbergen and from there to the rest of Europe via an elevated Barents shelf. (Szalay & McKenna, 1971, p. 283) McKenna (1971) coined the term ‘De Geer route’ by referring to the ‘De Geer North Atlantic dispersal route’ (see also McKenna, 1975 and the references therein). Early evidence Simpson (1947) proposed the notion of the dispersal of land faunas across the early Cenozoic Holarctic landmasses; the exact dispersal mechanisms did not become known, however, until the discovery in the late 1960s of sea-floor spreading and the subsequent acceptance of plate tectonics (Vine & Matthews, 1963; Le Pichon et al., 1973). The initial evidence for a northern Atlantic land corridor was palaeontological data showing that in the early Eocene, the mid-latitude floras and mammalian faunas of Europe and North America shared over half of their genera with one another (Savage, 1971; McKenna, 1975). Beginning in the 1970s, new geological and geophysical data added tectonic foundations supporting the existence of not only one, 2 but two northern Atlantic land bridges: the Thulean route and the De Geer route (Strauch, 1970; Szalay & McKenna, 1971; McKenna, 1975, 1983a,b; Tucholke & McCoy, 1986). According to Szalay & McKenna (1971), the De Geer route was the main terrestrial passage between Europe and North America; because the Thulean route, albeit southerly, ‘would have involved island-hopping at best’ and ‘would not suffice, however, for the large scale faunal transfer that took place in the Sparnacian’ (early Eocene) (Szalay & McKenna, 1971, p. 284). The contribution of the De Geer route to biogeographical evolution was even greater than that of Beringia, because ‘the Bering area would have been about 8 degrees farther north than the Greenland-Barents shelf route’ (Szalay & McKenna, 1971, p. 284). The De Geer route is suggested to have operated during the Ypresian (early Eocene) and ended when the land bridge broke apart near the De Geer fault zone (Spitsbergen fracture zone) at the end of the period (45–48 Ma) (Szalay & McKenna, 1971; McKenna, 1975). Szalay & McKenna (1971) and McKenna (1975) considered the Barents Shelf to be generally subaerial in both the Palaeocene and a large part of the Eocene. The terrestrial continuity of the Barents Shelf was interrupted in the Spitsbergen fracture zone, separating North America and Greenland on one side from the Spitsbergen– subaerial Barents Shelf–Fennoscandia on the other. McKenna (1983a,b) maintained the same view, considering Fennoscandia as an ‘outpost of North America’ that was separated from western Europe south of the Baltic by marine barriers until the end of the Eocene. On the other hand, McKenna also presented updated data supporting a single subaerial Thulean route [i.e. a more continuous and habitable land connection from France and the British Isles to North America via the Greenland-Scotland Ridge (GSR)]. Thus, contrary to Szalay & McKenna (1971), McKenna (1983a,b) considered the Thulean route to be a more significant biogeographical junction than the De Geer route, given its more southerly geographical position. McKenna further suggested that the two land bridges could have simultaneously connected North America to different parts of Europe (i.e. Great Britain and Scandinavia) near the end of the Palaeocene and up to some point very early in the Eocene. Tiffney (1985) generally accepted the palaeogeography proposed by McKenna (1983a). He kept some reservations, however, concerning whether Fennoscandia and western Europe remained in isolation in the early Eocene. Furthermore, he noted that the De Geer route was probably presented as early as the Danian (early Palaeocene) and maintained until the latest Eocene. Recent geotectonic and palaeogeographical evidence Today, the Nares Strait separates Greenland and Arctic Canada (Fig. 2a in the main paper). It is believed, however, that Greenland was connected with Ellesmere Island and north-eastern North America in the early Palaeocene (Dawson et al., 1975; McKenna, 1983a; Blythe & Kleinspehn, 1998; see also Marincovich, 1994; Golovneva, 2002). Indeed, despite the debate concerning the tectonic structure of the Wegener Fault, which runs through the Nares Strait (e.g. Tessensohn et al. 2008), only terrigenous sediments [Eureka Sound Group: Maastrichtian (?) to Eocene strata; Miall, 1986] have been recorded in the southern part of this area during the latest Cretaceous–early Palaeocene (Lee et al., 2008, and the reference therein; see also the land–sea boundary in Ricketts, 1986: Figure 39.1). 3 Palaeogeographical reconstructions show another marine corridor to the west, however, that could constitute a barrier between Greenland-Ellesmere Island and the north-western Canada landmass (see reconstructions of Torsvik et al., 2002, at 60 Ma; Stampfli & Borel, 2004, at 70 Ma, Appendix 3 in CD-ROM; Smelror et al., 2009, in the early Cenozoic). This corridor starts from Baffin Bay, crosses Lancaster Sound, and reaches as far west as Viscount Melville Sound and the North Canada Basin (see Fig. 3b). In any case, the stratigraphy west of Lancaster Sound does not support such a marine continuity. There (as suggested by strata preserved on Devon Island), Mesozoic sediments of the marine Kanguk Formation (Cenomanian–Campanian), which is widespread across the Canadian margin, are apparently overlaid by the terrestrial Expedition Fiord Formation (Maastrichtian–lower Palaeocene), which is preserved locally in fault-bounded troughs on Devon Island (Witkowski et al., 2011, and references therein). These strata result in the lower part of the Expedition Formation of the Eureka Sound Group of Ricketts (1986, 1991, 1994), which is generally considered to be terrestrial. Miall et al. (1980) correlated these strata with strata in the Eclipse Trough and the Bylot Island Formation, thus attributing a more regional range. Smith et al. (1989) noted that such terrestrial sediments, possibly a lacustrine unit, are found only in the western Lancaster Sound Basin and were either not deposited in the east or were completely eroded beneath a regional unconformity above which ‘Tertiary sediments’ were deposited (see stratigraphic charts of Smith et al., 1989, and Brent et al., 2013). These sediments can be attributed to the Selandian transgression (at 61.8 Ma; see the sea-level curve in Fig. 4) that probably transgressed the entire Lancaster Sound Basin interrupting the terrestrial connection (see Fig. 3b). To the east, Svalbard was connected to Greenland during the latest Cretaceous–early Palaeocene (Worsley et al., 1986; Muller and Spielhagen, 1990; Blythe & Kleinspehn, 1998; see supporting floristic evidence in Sweet, 2008, and references therein, and faunal evidence in Lüthje et al., 2010). Since the late Palaeocene, a shallow corridor may have separated the two areas during sea-level highstand (see Fig. 3b). Τhe break up of Svalbard and Greenland started in the latest Eocene (anomaly 13, c. 36 Ma: Schluter & Hinz, 1978) with a shallow marine setting. A continuous deep-water oceanic corridor was not established until the opening of the Fraim Strait in the Neogene (Engen et al., 2008). Eastwards of Svalbard, the Barents Sea served as part of a marine passageway connecting the North Sea and northern Atlantic Ocean to the Arctic Ocean during much of the Mesozoic and Cenozoic. Because of a lack of evidence and the complicated tectonic history of the region, however, the exact palaeogeographical regime of the Barents Sea remains unclear for certain periods. Apart from the southwestern Barents Sea, where deeper basins formed by subsidence since the Early Cretaceous, the rest of the Barents Shelf remained shallow and more tectonically stable since the Late Carboniferous (c. 300 Ma) (Faleide et al., 1993) (Fig. 2b). Nevertheless, several parts of the floor of the Barents Sea may have uplifted and subsided more than once in the Cenozoic. Today, the upper section (most of the Palaeogene section and the entire Neogene section) of the sea floor is absent or eroded in most of the Barents Sea drilling wells (Henriksen et al., 2011). On the other hand, the timing of the major uplift and erosion periods has been a subject of discussion for many years, because different areas of the Barents Sea achieved maximum burial at different times (Henriksen et al. 2011). During the Pliocene–Pleistocene, the entire Barents Shelf was eroded and large amounts of sediment were shed towards the shelf margin (Smelror et al., 2009). As a result, Cenozoic sediments in the Barents Sea are 4 restricted to the south-western basin areas and the western and northern passive continental margins bounding the shelf. In the south-western Barents Sea, the base of the Torsk Formation lies transgressively upon Mesozoic strata, which range in age from Triassic to Maastrichtian (Ryseth et al., 2003; Nagy et al., 2004). A major stratigraphic hiatus is present everywhere in the Barents Sea, spanning from the Maastrichtian (latest Cretaceous) to the late Danian (early Palaeocene) (Faleide et al., 1993; Nagy et al., 2004; Setoyama et al., 2011) (see Fig. S1.1 below). Danian (early Palaeocene) deposits are generally rare except in the south-western basin areas (Torsk Formation, Bjørnøya Basin, Senja Ridge-Vestbakken Volcanic Province, Sørvestsnaget Basin, Tromsø Basin and Hammerfest Basin) (Ryseth et al., 2003; Nagy et al., 2004; Setoyama et al., 2011) (Fig. 2b). In contrast, late Palaeocene (Thanetian) deposits exist in the western, south-western and other parts of the Barents Sea. According to Ryseth et al. (2003), biostratigraphy and log patterns indicate that these depositions seem to form a laterally continuous facies, while their occurrence throughout the Barents Sea supports the idea of Nøttvedt et al. (1988) of a widespread late Palaeocene epeirogenic sea in the region. Nøttvedt et al. (1988) inferred that a broad subsidence in the Barents Shelf and the continental shelf of northern Greenland produced a large epicontinental marine basin during the late Palaeocene. This epicontinental setting, however, terminated in the early Eocene because of rifting and volcanism associated with crustal break-up (Nøttvedt et al., 1988; Ryseth et al., 2003). In such a scenario, a formerly subaerial phase of the Barents Shelf (Fig. 3a) could have been transgressed and drained through the south-western Barrents Sea basins (Fig. 3b,c), suggesting that the latest Cretaceous–earliest Cenozoic hiatus in the continuity of marine depositions probably corresponds to a subaerial exposure of the shelf. A regional subaerial exposure of the areas surrounding the Barents Sea is also predicted by the model of Lyberis & Manby (1993), which constitutes one of the major current structural hypotheses explaining the formation of the West-Spitsbergen Fold-and-Thrust Belt. Under this hypothesis, latest Cretaceous–Palaeogene intercontinental compressional tectonics in the Greenland–Svalbard margin caused compressive orogen resulting in regional uplift of the Barents Shelf area. On the basis of foraminiferal assemblages from five wells drilled in the south-western Barents Sea, a recent study on the palaeobathymetric trends of the area concluded that significant Late Cretaceous–Palaeocene uplift took place prior to the break-up of the Greenland– Norwegian Sea. Specifically, Setoyama et al. (2011) estimated uplift and subsidence rates being between 0.11 mm/year [1000 m/Maastrichtian (70.6 Ma)–Danian (61.1 Ma)] and 0.56mm/year [4000 m/late Maastrichtian (69.1 Ma)–early Danian (62 Ma)]: results that are compatible with the Lyberis & Manby (1993) model. Under these results, Setoyama et al. (2011) did not reject the perspective that the late Maastrichtian–early Danian hiatus in the south-western Barents Sea basinal areas is due to subaerial exposure, even though, on the basis of morphogroup analysis of the foraminiferal assemblages across the hiatus, these basins are generally recognized as being under deep water during this interval (Nagy et al., 2004; Setoyama et al., 2011). Therefore, it seems likely that the marine setting in the basinal areas of the southwestern Barents Sea was uninterrupted. But given the observed shallowing in the regional bathymetry, the surrounding shallow Barrents Shelf was probably exposed subaerially to the east. The evidence presented thus far favours the possibility that the observed late Maastrichtian–early Danian stratigraphic hiatus (Fig. S1.1) corresponds to a subaerial 5 exposure of the Barents Shelf. Some of the current palaeogeographical reconstructions of the region seem to be congruent with such perspective: the reconstruction of the epicontinental seas and straits in northern Eurasia (Akhmetiev & Beniamovski, 2009: Figure 3), the palaeogeographical reconstruction of the Tethys Ocean c. 70 Ma by Stampfli & Borel (2004: Appendix 3 in CD-ROM), and the early ‘Tertiary’ (Palaeocene) reconstruction of the northern Atlantic by Ziegler (1988). The continuous facies deposited locally since the late Danian, but regionally since the Selandian, throughout the western and south-western Barents Shelf could be interpreted as the transgressive remains of the exposure of a seaway connecting the Norwegian-Greenland Seaway to the Arctic Ocean. Indeed, such a seaway is included in many recent palaeogeographical reconstructions (e.g. since the mid-Danian, Akhmetiev & Beniamovski, 2009: Figure 3; late Palaeocene, Akhmetiev et al., 2012: Figure 1b; c. 60 Ma, Torsvik et al., 2002; in the late Thanetian, Brunstad et al., in press: Figure 4; in the early Cenozoic, Smelror et al., 2009; c. 50 Ma, Brinkhuis et al., 2006; Gleason et al., 2009; Eberle & Greenwood, 2012). All these reconstructions run absolutely contrary to McKenna’s (1983a) view that the De Geer route persisted during the latest Palaeocene and early Eocene. The available stratigraphic evidence, which allows direct observation of the past geology, is limited to the southern portions of the Barents Sea floor and extends as far east as the Nordkapp basin (Nagy et al., 2004) (Fig. 2b). This evidence suggests that a marine barrier persisted to the north of Fennoscandia from the late Palaeocene at least until the middle Eocene, but it cannot support the existence of such a barrier eastwards of the Nordkapp basin (see the palaeogeographical reconstruction of the Eocene Barents Sea in Smelror et al., 2009, p. 123). This is because very little is known about the Palaeocene and Eocene palaeogeography of the eastern parts of the Barents Sea and the Pechora and Kara seas (Fig. 2a). Current geological evidence for these areas suggests that they probably constituted a tectonically stable epicontinental mega-region and were either uplifted continental hinterlands or shallow marine seas with very limited net deposition (Smelror et al., 2009). The sediments that may have been deposited would have been subsequently removed due to later Neogene uplift and erosion. Consequently, there is very little direct geological evidence remaining to help establish models of the Palaeocene and Eocene palaeogeography of the Barents Sea platform and the Kara Sea (Smelror et al., 2009). Nevertheless, as discussed below, the remaining stratigraphic sequences include sufficient marine faunal evidence, albeit fragmentary, to improve our knowledge of the true early Cenozoic palaeogeography of the Barents Shelf. Based on the marine faunas, there is no palaeontological evidence for a latest Cretaceous/early Palaeocene marine connection either between the Arctic Ocean and the northern Atlantic Ocean or between the Arctic Ocean and western Greenland (Baffin Bay) (Marincovich, 1994). In contrast, although marine faunas in the Arctic Ocean during the Danian were geographically confined, there was an excellent interchange of marine molluscs and microbiota between the type Danian Denmark fauna and the Danian faunas of western Greenland (Marincovich et al., 1990). The fact that these more southerly and very diverse Danian faunas shared no mollusc species with the Arctic Ocean fauna suggests that the Arctic Ocean was isolated at that time (Marincovich, 1993, 1994). On the other hand, the appearance of the bivalve Cyrtodaria rutupiensis in the Thanet Sands (type Thanetian) of the London Basin (Strauch, 1972) confirms a late Palaeocene marine connection between the Arctic Ocean and the southern North Sea Basin. Cyrtodaria rutupiensis dwelled only in the Arctic Ocean during the Danian, suggesting that a land barrier existed at that time 6 between northern Europe and northern Greenland (Marincovich, 1993, 1994). The Thanetian first appearance of the bivalve Gari (Garum) in western European faunas is further evidence of a northern marine connection, because the oldest records of both Cyrtodaria and Gari appear in the Danian fauna of the Prince Creek Formation in northern Alaska (Marincovich, 1993). Nagy et al. (2000) reported the recognition of three agglutinated foraminiferal species from the late Palaeocene in the Firkanten Formation of Spitsbergen: Reticulophragmium arcticum, Reticulophragmium boreale, and Labrospira turbida. All three species are known from the lower part of the late Palaeocene in the Beaufort-Mackenzie Basin (Arctic North America) (McNeil, 1997) and the late Palaeocene in Hole M0004A at Lomonsov Ridge (Arctic Basin) (Labrospira sp. in the latter) (Expedition 302 Scientists, 2006). Furthermore, the species Psammosphaera eocenica as well the genus Verneuilinoides are shared among the Spitsbergen and Lomonsov Ridge faunas. The above similarities provide clear evidence of marine faunal exchanges between the western Barents Sea and the Arctic Ocean via the Barents Shelf in the late Palaeocene. The genus Reticulophragmium is believed to have evolved in the Northern Hemisphere during the ‘mid’-Palaeocene and is not known from Cretaceous strata. Primitive species of Reticulophragmium are also known from the Palaeocene strata of the North Sea and western Siberia (Expedition 302 Scientists, 2006). Thus, based on the occurrence of Reticulophragmium, Verneuilinoides, and other Boreal-Arctic foraminifera from the western Barents Sea (Torsk Formation), the central Spitsbergen Basin, the Turgai Strait, the western Siberian Basin, and the Polar Arctic, Akhmetiev et al. (2012) concluded that a meridional seaway system between the Arctic and Tethys oceans connected northward to the northern Atlantic Ocean via the Barents Sea (see the late Palaeocene reconstruction in Akhmetiev et al., 2012: Figure 1b). The system started in western Asia in the Late Cretaceous; it was interrupted around the Cretaceous/Palaeogene (K/Pg) boundary and resumed from the mid-Danian to the Lutetian (Akhmetiev et al., 2012). The interruption corresponds to the period during which the De Geer route should have been exposed (see Discussion). THE THULEAN ROUTE The Thulean land bridge consisted of the subaerial exposure of the shallow transverse GSR, a bathymetric sill that hampered the deep-water flow between the NorwegianGreenland Sea and the northern Atlantic, extending from south-eastern Greenland, through Iceland, to the UK continental shelf (e.g. Thiede & Myhre, 1996) (Fig. 2c). The word Thule comes from ancient Greek literature and refers to an island or region of the North Sea that has been associated with Iceland (Sprague de Camp, 1954). The first direct geological evidence of the existence of the Thulean land bridge came from the recovery of subaerially erupted basalt lavas from the northern flank of the IcelandFaeroe Ridge [Deep Sea Drilling Programme (DSDP) borehole 336; Shipboard Scientific Party, 1976; Nilsen, 1978]. Palaeontologists have long considered the Thulean route to be important in the biogeographical history of marine and terrestrial biota (Strauch, 1970, 1972; Szalay & McKenna, 1971; McKenna, 1975, 1983a,b; Akhmetiev et al. 1978; Friedrich & Símonarson 1981; Axelrod, 1983). Morphologically, the GSR can be divided into three parts (Thiede & Myhre, 1996): (1) the Greenland-Iceland Ridge, or Denmark Strait; (2) the Iceland-Faeroe 7 Ridge; and (3) the Faeroe-Shetland Channel. At the western end of the Thulean landbridge, possible links between Greenland and North America are provided by the shallow bathymetric sill at the Davis Strait between southern Baffin Island and southwestern Greenland and, alternatively, by the higher latitude land connection via the Nares Strait (McKenna, 1983a) (Fig. 3d). A route via the Davis Strait would have certain biogeographical advantages over the Thulean route; this is an issue currently under discussion, depending on whether continental or oceanic crust underlies the bottom of the strait (Srivastava & Arthur, 1989; Chalmers & Pulvertaft, 2001). At the eastern end of the Thulean route, the Dover strait may have played its own role in keeping terrestrial continuity. Because of limited data, however, the geotectonic evolution of this area is not well understood (Van Vliet-Lanoë et al., 2004). The GSR is believed to be part of a mantle plume system centred under the Icelandic plateau. High heat flow and low mantle densities under Iceland result in both the subaerial exposure of the plateau and the relatively shallow depths of other proximal features (Nunns, 1983; Wright & Miller, 1996). In the early Palaeogene, magmatic events prior to and during continental separation, and the post-breakup continuous activity of the Iceland melting anomaly, have resulted in one of the largest igneous provinces in the world: the North Atlantic Igneous Province (NAIP) (White & McKenzie, 1989; Jolley & Bell, 2002). Following Morgan (1971), most researchers explain the early Palaeogene volcanism of the NAIP in terms of lithospheric impingement of the proto-Iceland mantle plume, although the mantle plume concept is currently being challenged as an explanation for the NAIP, and alternative models have been suggested (Meyer et al., 2007, and references therein). The NAIP was emplaced in two main magmatic phases. The first phase occurred in the ‘mid’-Palaeocene (c. 62–58 Ma) (e.g. Rousse et al., 2007). The second phase occurred within Chron 24r (late Thanetian to early Ypresian), spanning the Palaeocene/Eocene boundary, and was associated with regional uplifts and extensive volcanic deposits that began at the end of the first phase (Saunders et al., 1997). Thus, the GSR could have existed at its maximum subaerial continuity only during the second phase of the NAIP emplacement. After its accretion, most parts of the GSR transverse are believed to have been continuously subaerial at least until the Oligocene (Fig. S1.2 below) (see also Denk et al., 2011). Then, the ridge subsided below sea level through a combination of thermal subsidence and erosion. The onset of the subsidence of parts of the GSR has been linked to the onset of the exchange of intermediate and deep-water masses (North Atlantic Deep Water (NADW)) between the Nordic Seas and the Atlantic Ocean across the GSR (e.g. Hohbein et al., 2012). Whatever the exact time frame for the onset of the NADW, the linkage between the Faeroe-Shetland channel and the Faeroe Bank channel constituted the main passageway (Stoker & Varming, 2011). The other significant deep-water passageway across the GSR, the Denmark Strait, may also have developed at about the same time as the GSR became fully submerged (Thiede & Eldholm, 1983). In contrast to the other components of the GSR, the Faeroe-Shetland Basin was preserved as a marine property in the early Palaeogene; and as Jones (2011: Box 3.7) notes, according to the evidence available to McKenna (1983a), the land connection of the Thulean route would have extended only from Greenland to the Faeroes during the mammalian dispersal interval. The Faeroe-Shetland Basin between the Faeroes and Scotland would have been a seaway several tens-of-kilometres wide and several hundred meters deep, constituting a significant barrier to the dispersal of 8 land mammals. It was only recently that three-dimensional seismic data combined with drilling-well data from the southern Faeroe-Shetland Basin provided evidence of a closure of the former seaway (Smallwood & Gill, 2002; Shaw Champion et al., 2008; Hartley et al., 2011), and thus, of the establishment of a land connection between North America and Europe at precisely the time of the observed mammalian dispersal (c. 56-54 Ma according to the most recent calibrations of the North American land mammal ages (NALMAs) and European land mammal ages (ELMAs) against the geomagnetic time-scale of Gradstein et al., 2012). Therefore, the southern Faeroe-Shetland Basin has been recognized as the gateway for the Thulean route. In the following section, the correlation of stratigraphy from the southern Faeroe-Shetland Basin with independent data suggests that the first exposure of the Thulean route took place in at least two episodes: c. 57 Ma and c. 56 Ma (see Fig. S1.3 below). Stratigraphy of the Southern Faeroe-Shetland Basin The generalized stratigraphy of the early Palaeogene in the wells in Quadrant 204 of the southern Faeroe-Shetland Basin (Fig. 2d) reveals that the marine property was interrupted by hiatus or terrestrial intervals (e.g. Shaw Champion et al., 2008). Within the quadrant, the lower Eocene is best developed in the Judd Basin area, where up to about 750 m of sediment were deposited (Sørensen, 2003). Wells such as the British Geological Survey borehole 99/03 (Fig. 2d) are considered to be representative of the regional stratigraphy (Stoker & Varming, 2011; Stoker et al., 2012). The unconformities considered below fall into the time interval following the first magmatic phase of the NAIP, when a possible subaerial exposure of the Thulean route would be possible (i.e. after c. 60 Ma). Figure S1.2 identifies the exact ages of these unconformities on the basis of the biostratigraphic markers presented by Shaw Champion et al. (2008: Figure 2) and the recent bio-chronostratigraphic chart of Vandenberghe et al. (2012). Thus, the first regional unconformity (known also from the North Sea; Mudge and Jones, 2004) is dated 59.2 Ma on the basis of the first appearance of the dinoflagellate cyst Areoligera gippingensis. This hiatus marks the base of the Lamda Formation, which marks a change from the deposition of deepwater turbidites to a deltaic succession and the first major progradation of the shoreline into the basin during Thanetian times (Shaw Champion et al., 2008). Although this unconformity reduced the distance between the Shetland and Faeroe land margins (coinciding with the PL4 sea-level low stand shown in Fig. S1.3), it probably did not manage to accrete the Thulean land bridge in full. A second unconformity appears at the end of the Lamda Formation and the base of the lower Flett Formation. The latter is correlated with the lower Forties Formation of the North Sea (Mudge & Bujak, 2001), the base of which is dated to the upper Thanetian (‘intra-Upper Thanetian’: Mudge & Jones, 2004). Therefore, the second unconformity falls into this age. Smallwood & Gill (2002) showed that this unconformity is marked by a surface upon which a branching network of valleys and intervening topographic highs is preserved. They also identified three valley systems that drain northward into the Judd Basin, and they interpreted the unconformity to have formed by subaerial erosion. Shaw Champion et al. (2008) identified the same unconformity in a wider area, noting that an unconformity of this age had also been identified across the deeper parts of the Faroe-Shetland basin and in the northern and central North Sea basins. Therefore, the dating of this unconformity is of special importance because it probably signals the exact time of the first full exposure of the 9 Thulean land bridge. According to stratigraphic column of Shaw Champion et al. (2008: Figure 2), this ‘intra-Upper Thanetian’ unconformity falls within the interval of the first appearances of Apectodinium augustum (56.5 Ma) and Apectodinium nomomorphum (57.2 Ma). Brunstad et al. (in press) attributed this unconformity to the Lista-Sele facies transition, which reflects a dramatic sea level fall, as seen throughout the UK shelf and onshore, and a simultaneous basin restriction with anoxia in the central and deeper parts of the North Sea basin. This sea-level lowstand interval coincides with the PL6 sea-level lowstand shown in Fig. S1.3 (a 56.6–57 Ma hiatus; Kominz et al., 2008: supplementary material). Hence, this interval likely coincided with the hiatus spanning the ‘Intra-Thanetian unconformity’ and corresponds to the first full exposure of the Thulean land bridge (Fig. 3d) (see also the palaeogeographical reconstructions of Stoker & Varming, 2011: Figure 90). This late Thanetian terrestrial phase was terminated by the transgressive lower Flett Formation (Figs S1.2 & S1.3), which was largely deposited beyond the Lamda Formation shelf edge, resulting in shelf bias, erosion, and non-deposition along the southern margin of the Faeroe-Shetland Basin (Stoker & Varming, 2011). The Flett Formation is missing in this area, and the younger Balder Formation rests unconformably on the Lamda Formation (Smallwood & Gill, 2002) because of erosion. Although marine dinoflagellate cysts, especially the genus Apectodinium, are abundant in the lower Flett unit (corresponding to the T40 sequence of Lamers & Carmichael, 1999), terrestrial forms dominate assemblages in the overlying upper Flett unit (T45 sequence) (Shaw Champion et al., 2008; Stoker & Varming, 2011). The boundary between the lower and upper units of the Flett Formation is well restricted by the last occurrence of the dinoflagellate cyst Apectodinium augustum (55.8 Ma). This age matches perfectly a retreat of the sea (the EL1 sea-level lowstand shown in Fig. S1.3). Thus, the onset of the terrestrial upper Flett Formation likely resulted from a second subaerial exposure of the Thulean land bridge coinciding with the EL1 sea-level lowstand shown in Fig. S1.3. BERINGIA Beringia, first proposed by Hultén (1937), is hypothesized to have been a land bridge between Asia and North America during the Plio-Pleistocene. A pre-Quaternary connection was acknowledged only after the broad acceptance of plate tectonic theory (Hopkins, 1967; Cox, 1974). More recently, palaeontological investigations of correlative fossil-bearing rocks in regions of Alaska (Aniakchak National Monument, numerous localities along the Colville River, and the Denali National Park; see Fiorillo, 2008 and references therein) and Utah (Cifelli et al., 1997), combined with revised tectonic reconstructions of the region (Lawver et al., 2002), led to the conclusion that Beringia originated approximately 100 Ma. This hypothesis is congruent with other evidence from marine faunas suggesting that the Bering Strait was closed as early as the late Albian (Jagt-Yazykova, 2012; Iba et al., 2011). Some authors consider Beringia to have been terrestrial from the late Albian until the late Miocene (Marincovich et al., 1990; Marincovich & Gladenkov, 1999; Gladenkov et al., 2002). Because of the obvious faunal similarities found across the early Cenozoic of Holarctica, Beringia is acknowledged as a possible dispersal route for mammals (McKenna, 1983a,b; Beard, 1998; Beard & Dawson, 1999). Marine evidence, however, leads other authors to leave open the possibility of a marine connection between the Arctic and Pacific oceans in the latest Cretaceous (Jagt- 10 Yazykova, 2011) (Fig. 7) and early Cenozoic (Gleason et al., 2009). Thus, Beringia is emergent in some palaeogeographical reconstructions (Eberle & Greenwood, 2012) and submerged in others (Gleason et al., 2009). On the basis of climatic (Appendix S2 and Fig. 4), floristic, and vertebrate evidence (Figs 5a,b & 6), two time windows are likely for biotic exchanges across Beringia during the Palaeocene: Bering route 1 65.5 Ma (Fig. 7d) and Bering route 2 c. 58 Ma. Possible Eocene exposures are not considered here. Figure S1.1 Lithostratigraphy of the south-western Barents Sea. Modified from Setoyama et al. (2011). 11 Figure S1.2 Palaeogeographical continuity of the Thulean land bridge. The sections correspond to the parts of the Thulean land bridge. Horizontal lines note the ages proposed by various authors concerning the subsidence of each part. Most parts of the Thulean land bridge persisted subaerially at least until the Oligocene. Note the exception of the southern Faeroe-Shetland Basin, which was dominated by tectonic and sea-level changes resulting in the gateway of the Thulean route. Schema based on Denk et al. (2011). Biostratigraphy and T-sequences are after Shaw Champion et al. (2008), and chronostratigraphy is after Vandenberghe et al. (2012). 12 Figure S1.3 Correlation of the eustatic sea-level curve from New Jersey with the stratigraphies of the southern Faeroe-Shetland Basin, the Kangerlussuaq Basin (East Greenland), the Turgai Strait, and the mammalian biostratigraphy of North America and Europe. Figure modified from Brikiatis (in preparation). 13 REFERENCES Akhmetiev, M.A., Bratzeva, G.M., Giterman, R.E., Golubeva, L.V. & Moiseyeva, A. I. (1978) Late Cainozoic stratigraphy and flora of Iceland. Transactions of the Academy of Sciences USSR, 316, 1-188. Axelrod, D.I. (1983) Biogeography of oaks in the Arcto-Tertiary province. Annals of the Missouri Botanical Garden, 70, 629-657. Beard, K.C. (1998) East of Eden: Asia as an important center of taxonomic origination in mammalian evolution. Dawn of the age of mammals in Asia (ed. by K.C. Beard and M.R. Dawson. Carnegie Museum of Natural History Bulletin, 34, 5-39. Blythe, A.E. & Kleinspehn, K.L. (1998) Tectonically versus climatically driven Cenozoic exhumation of the Eurasian plate margin, Svalbard: fission track analyses. Tectonics, 17, 621-639. Brent, T.A., Chen, Z., Currie, L.D. & Osadetz, K. (2013) Assessment of the conventional petroleum resource potential of Mesozoic and younger structural plays within the proposed National Marine Conservation Area, Lancaster Sound, Nunavut. Geological Survey of Canada Open File, 6954. Natural Resources Canada, available online at: http://dx.doi.org/10.4095/289615 Chalmers, J.A. & Pulvertaft, T.C.R. (2001) Development of the continental margins of the Labrador Sea: a review. Geological Society, London, Special Publications, 187, 77-105. Cifelli, R.L., Kirkland, J.I., Weil, A., Deino, A.L. & Kowallis, B.J. (1997) Highprecision 40Ar/39Ar geochronology and the advent of North America’s Late Cretaceous terrestrial fauna. Proceedings of the National Academy of Science of the USA, 94, 11,163-11,167. Cox, C.B. (1974) Vertebrate palaeodistributional patterns and continental drift. Journal of Biogeography, 1, 75-94. Dawson, M.R., West, R.M., Ramaekers, P. & Hutchison, J.H. (1975) New Evidence on the Palaeobiology of the Eureka Sound Formation, Arctic Canada. Arctic, 28, 110-116. Denk, T., Grímsson, F., Zetter, R. & Símonarson, L.A. (2011) The biogeographic history of Iceland – the North Atlantic land bridge revisited. Late Cenozoic floras of Iceland (ed. by T. Denk et al.), pp. 647-668. Springer, New York. Eldholm, O., Myhre, A. M., & Thiede, J. (1994) Cenozoic tectono-magmatic events in the North Atlantic: potential palaeoenvironmental implications. Cenozoic plants and climates of the Arctic (ed. by M.C. Boulter and H.C. Fisher), pp. 35-55. NATO ASI Series, 127, Springer, Berlin. Engen, Ø., Faleide, J.I. & Dyreng, T.K. (2008) Opening of the Fram Strait gateway: a review of plate tectonic constraints. Tectonophysics, 450, 51-69. Expedition 302 Scientists (2006) Sites M0001–M0004. Proceedings of Integrated Ocean Drilling Program, 302 (ed. by J. Backman, K. Moran, D.B. McInroy, L.A. Mayer and the Expedition 302 Scientists), pp. 1-169. Integrated Ocean Drilling Program Management International, Inc., Edinburgh. Faleide, J.I., Vågnes, E. & Gudlaugsson, S.T. (1993) Late Mesozoic – Cenozoic evolution of the southwestern Barents Sea. Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference (ed. by J.R. Parker), pp. 933-950. Geological Society, London, London. Fiorillo, A.R. (2008) Cretaceous dinosaurs of Alaska: implications for the origins of Beringia. The Terrane Puzzle: new perspectives on paleontology and stratigraphy from the North American Cordillera (ed. by R.B. Blodgett and G. Stanley), pp. 313-326. Geological Society of America, Special Paper, 442. 14 Friedrich, W.L. & Símonarson, L.A. (1981) Die fossile Flora Islands: Zeugin der Thule- Landbrücke. Spektrum der Wissenschaft, 10, 22-31. Gladenkov, A,Y., Oleinik, A.E., Marincovich, L., Jr & Barinov, K.B. (2002) A refined age for the earliest opening of Bering Strait. Palaeogeography, Palaeoclimatology, Palaeoecology, 183, 321-328. Harland, W.B. (1969) Contribution of Spitsbergen to understanding the tectonic evolution of the North Atlantic region. North Atlantic, geology and continental drift (ed. by M. Kay), pp. 817-851. American Association of Petroleum Geologists, Memoir, 12. Hartley, R.A., Roberts, G.G., White, N. & Richardson, C. (2011) Transient convective uplift of an ancient buried landscape. Nature Geoscience, 4, 562-565. Henriksen, E., Bjørnseth, H.M., Hals, T.K., Heide, T., Kiryukhina, T., Kløvjan, O.S., Larssen, G.B. , Ryseth, A.E., Rønning, K., Sollid, K. & Stoupakova, A. (2011) Uplift and erosion of the greater Barents Sea: impact on prospectivity and petroleum systems. Arctic petroleum geology (ed. by A.M. Spencer, A.F. Embry, D.L. Gautier, A.V. Stoupakova & K. Sørensen), pp. 271-281. Geological Society, London, Memoirs, 35. Hohbein, M.W., Sexton, P.F. & Cartwright, J.A. (2012) Onset of North Atlantic Deep Water production coincident with inception of the Cenozoic global cooling trend. Geology, 40, 255-258. Hopkins, D.M. (1967) The Cenozoic history of Beringia – A synthesis. The Bering land bridge (ed. by D.M. Hopkins), pp. 451-484. Stanford University Press, Stanford. Horsfield, W.T. & Maton, P.I. (1970) Transform Faulting along the De Geer Line. Nature, 226, 256-257. Hultén, E. (1937) Outline of the history of Arctic and Boreal biota during the Quaternary period. Reprint in: Foundations of biogeography: classic papers with commentaries (ed. by M.V. Lomolino, D.F. Sax and J.H. Brown), pp. 464-512. University Of Chicago Press, Chicago. Jagt-Yazykova, E.A. (2011) Palaeobiogeographical and palaeobiological aspects of mid- and Late Cretaceous ammonite evolution and bio-events in the Russian Pacific. Scripta Geologica, 143, 15-121. Jagt-Yazykova, E.A. (2012) Ammonite faunal dynamics across bio-events during the mid- and Late Cretaceous along the Russian Pacific coast. Acta Palaeontologica Polonica, 57, 737-748. Jolley, D.W. & Bell, B.R. (2002) The evolution of the North Atlantic Igneous Province and the opening of the NE Atlantic rift. The North Atlantic Igneous Province: stratigraphy, tectonic, volcanic, and magmatic processes (ed. by D.W. Jolley and B.R. Bell), pp. 1-13. Geological Society, London, Special Publication, 197. Jones, R.W. (2011) Applications of palaeontology: techniques and case studies. Cambridge University Press, Cambridge, UK. Iba, Y., Mutterlose, J., Tanabe, K., Sano, S., Misaki, A. & Terabe, K. (2011) Belemnite extinction and the origin of modern cephalopods 35 m.y. prior to the Cretaceous−Paleogene event. Geology, 39, 483-486. Lamers, E. & Carmichael, S.M.M. (1999) The Paleocene deep water sandstone play west of Shetland. Petroleum Geology of Northwest Europe: Proceedings of the 5th Conference (ed. by. A.J. Fleet and S.A.R. Boldy), pp. 645-659. Geological Society, London. Larsen, M., Knudsen, C., Frei, D., Frei, M., Rasmussen, T. & Whitham, A.G. (2006) East Greenland and Faroe–Shetland sediment provenance and Palaeogene 15 sand dispersal systems. Geological Survey of Denmark and Greenland Bulletin, 10, 29-32. Le Pichon, X., Francheteau, J. & Bonnin, I. (1973) Plate tectonics. Elsevier, Amsterdam. Marincovich, L., Jr (1993) Danian mollusks from the Prince Creek Formation, Nothern Alaska, and implications for Arctic Ocean paleogeography. Paleontological Society, Memoir, 35, 1-35. Marincovich, L., Jr (1994) Earliest Tertiary Cenozoic Paleogeography of the Arctic Ocean. 1992 proceedings, International Conference on Arctic Margins: Anchorage, Alaska, September 1992 (ed. by D.K. Thurston and K. Fujita), pp. 45-58. U.S. Department of the Interior, Minerals Management Service, Alaska. Marincovich, L., Jr & Gladenkov, A.Y. (1999) Evidence for an early opening of the Bering Strait. Nature, 397, 149-151. McKenna, M.C. (1971) Fossil mammals and the Eocene demise of the De Geer North Atlantic dispersal route. Abstracts with Programs (Geological Society of America), 3, 644. McNeil, D.H. (1997) New foraminifera from the Upper Cretaceous and Cenozoic of the Beaufort-MacKenzie Basin of Arctic Canada. Cushman Foundation for Foraminiferal Research, Special Publication, 35, 1-95. Meyer, R., van Wijk, J. & Gernigon, L. (2007) North Atlantic Igneous Province: a review of models for its formation. Plates, Plumes and Planetary Processes (ed. by G.R. Foulger and D.M. Jurdy), pp. 525-552. Geological Society of America, Special Paper, 430. Boulder, CO. Miall, A.D., Balkwill, H.R. & Hopkins, W.S. (1980) Cretaceous and Tertiary sediments of Eclipse Trough, Bylot Island area, Arctic Canada, and their regional setting. Geological Survey of Canada Paper, 79-23. Miall, A.D. (1986) The Eureka Sound Group (Upper Cretaceous–Oligocene), Canadian Arctic Islands. Bulletin of Canadian Petroleum Geology, 34, 240270. Mörner, N.A. (1979) The Fennoscandian uplift and late Cenozoic geodynamics: geological evidence. GeoJournal, 3, 287-318. Morgan, W.J. (1971) Convection plumes in the lower mantle. Nature, 230, 42-43. Mudge, D.C. & Bujak, J.P. (2001) Biostratigraphic evidence for evolving palaeoenvironments in the Lower Paleogene of the Faroe-Shetland Basin. Marine and Petroleum Geology, 18, 577-590. Mudge, D.C. & Jones, S.M. (2004) Palaeocene uplift and subsidence events in the Scotland–Shetland and North Sea region and their relationship to the Iceland Plume. Journal of the Geological Society, 161, 381-386. Muller, R.D. & Spielhagen, R.F. (1990) Evolution of the Central Tertiary Basin of Spitsbergen: towards a synthesis of sediment and plate tectonic history. Palaeogeography, Palaeoclimatology, Palaeoecology, 80, 153-172. Nagy, J., Kaminski, M.A., Kuhnt, W. & Bremer, M.A. (2000) Agglutinated foraminifera from neritic to bathyal facies in the Paleogene of Spitsbergen and the Barents Sea. Proceedings of the Fifth International Workshop on Agglutinated Foraminifera (ed. by M.B. Hart, M.A. Kaminski and C.W. Smart), pp. 333-361. Grzybowski Foundation Special Publication, 7. Grzybowski Foundation, Krakow. Nilsen, T.H. (1978) Tertiary laterite on the Iceland-Faeroe Ridge and the Thulean land bridge. Nature, 274, 786-788. Nunns, A.G. (1983) Plate tectonic evolution of the Greenland-Scotland Ridge and surrounding regions. Structure and development of the Greenland-Scotland 16 Ridge (ed. by M.H. Bott, P.S. Saxov, M. Talwani and J. Theide), pp. 11-30. Plenum Press, New York. Ogg, J.G. (2012) Geomagnetic polarity time scale. The geological timescale 2012 (ed. by F.M. Gradstein, J.G. Ogg, M.D. Schmitz and G.M. Ogg), pp. 85-113. Elsevier, Amsterdam. Poore, R.H., Samworth, R., White, N., Jones, S. & McCave, I. (2006) Neogene overflow of Northern Component Water at the Greenland-Scotland Ridge. Geochemistry, Geophysics, Geosystems, 7, Q06010. Poore, R.H. (2008) Neogene epeirogeny and the Iceland Plume. PhD Thesis, University of Cambridge, Cambridge, UK. Radionova, E.P., Beniamovski, V.N., Iakovleva, A.I., Muzylöv, N.G., Oreshkina, T.V., Shcherbinina, E.A. & Kozlova, G.E. (2003) Early Paleogene transgressions: stratigraphical and sedimentological evidence from the northern Peri-Tethys. Causes and consequences of globally warm climates in the early Paleogene (ed. by S.L. Wing, P.D. Gingerich, B. Schmitz and E. Thomas), pp. 239-261. Geological Society of America, Special Paper, 369. Boulder, CO. Ricketts, B.D. (1986) New formations in the Eureca Sound Group, Canadian Arctic Islands. Geological Survey of Canada Paper, 86-IB, 363-374. Ricketts, B.D. (1991) Delta evolution in the Eureka Sound Group, western Axel Heiberg Island: the transition from wavedominated to fluvial-dominated deltas. Geological Survey of Canada, Bulletin, 402. Natural Resources Canada, Ottawa. Ricketts, B.D. (1994) Basin analysis, Eureka Sound Group, Axel Heiberg and Ellesmere islands, Canadian Arctic Archipelago. Geological Survey of Canada, Memoir, 439. Natural Resources Canada, Ottawa. Saunders, A.D., Fitton, J.G., Kerr, A.C., Norry, M.J. & Kent, R.W. (1997) The North Atlantic Igneous Province. Large igneous provinces: continental, oceanic, and planetary flood volcanism (ed. by J.J. Mahoney and M.F. Coffin), pp. 45-93. Geophysical Monographs, 100. American Geophysical Union, Washington, D.C. Savage, D.E. (1971) The Sparnacian-Wasatchian mammalian fauna, early Eocene, of Europe and North America. Abhandlungen des Hessischen Landesamtes fόr Bodenforschung, 60, 154-158. Schluter, H.U. & Hinz, K. (1978) The continental margin of West Spitsbergen. Polarforschung, 48, 151-169. Shaw Champion, M.E., White, N.J., Jones, S.M. & Lovell, J.P.B. (2008) Quantifying transient mantle convective uplift: an example from the Faroe-Shetland basin. Tectonics, 27, TC1002. Shipboard Scientific Party (1976) Sites 336 and 352. Initial Reports of the Deep Sea Drilling Program, 38, 23-49. Simpson, G.G. (1947) Holarctic mammalian faunas and continental relationships during the Cenozoic. Bulletin of the Geological Society of America, 58, 613688. Smallwood, J.R. & Gill, C.E. (2002) The rise and fall of the Faroe-Shetland basin: Evidence from seismic mapping of the Balder Formation. Journal of the Geological Society London, 159, 627-630. Smith, D.R., Cowan, R.J. & McComb, M. (1989) Geology and resource potential of a proposed national marine park, Lancaster Sound, Northwest Territories. Geological Survey of Canada Open File, 2022. Natural Resources Canada, Ottawa. 17 Sørensen, A.B. (2003) Cenozoic basin development and stratigraphy of the Faroes area. Petroleum Geoscience, 9, 189-207. Sprague de Camp, L. (1954) Lost continents. Gnome Press, New York. Srivastava, S.P. & Arthur, M.A. (1989) Tectonic evolution of the Labrador Sea and Baffin Bay: Constraints imposed by regional geophysics and drilling results from Leg 1051. Proceedings of the Ocean Drilling Program, Scientific Results, 105, 989-1009. Stoker, M & Varming, T. (2011) Cenozoic (sedimantary). Geology of the FaroeShetland Basin and adjacent areas (ed. by J.D. Ritchie, H. Zisca, H. Johnson and D. Evans), pp. 151-208. British Geological Survey, Nottingham, UK. Strauch, F. (1970). Die Thule-Landbrücke als Wanderweg und Faunenscheide zwischen Atlantik und Skandik im Tertiär. Geologische Rundschau, 60, 381417. Strauch, F. (1972). Phylogenese, adaptation und migration einiger nordischer mariner Molluskengenera (Neptunea, Panomya, Cyrtodaria und Mya). Abhandlungen der Senckenberg Gesellschaft für Naturforschung, 531, 1-211. Tessensohn, F., von Gosen, W., Piepjohn, K., Saalmann, K. & Mayr, U. (2008) Nares transform motion and Eurekan compression along the northeast coast of Ellesmere Island. Geology of northeast Ellesmere Island adjacent to Kane Basin and Kennedy Channel, Nunavut (ed. by U. Mayr), pp. 227-243. Geological Survey of Canada, Bulletin, 592. Natural Resources Canada, Ottawa. Thiede, J. & Eldholm, O. (1983) Speculations about the paleodepth of the Greenland – Scotland Ridge during late Mesozoic and Cenozoic times. Structure and development of the Greenland – Scotland Ridge: new methods and concepts (ed. by M.H.P. Bott, S. Saxov, M. Talwani and J. Thiede), pp. 445-456. Plenum Press, New York. Thiede, J. & Myhre, A.M. (1996) Introduction to the North Atlantic-Arctic Gateways: plate tectonic-paleoceanographic history and significance. Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 151 (ed. by J. Thiede, A.M. Myhre, J.V. Firth, G.L. Johnson and W.F. Ruddiman), pp. 3-23. College Station, TX (Ocean Drilling Program). Tucholke, B.E. & McCoy, F.W., (1986) Paleogeographic and paleobathymetric evolution of the North Atlantic Ocean. The western North Atlantic. The geology of North America (ed. by P.R. Vogt and B.E. Tucholke), pp. 589-602. Geological Society of America. Boulder, CO. Van Vliet-Lanoë, B., Mansy, J.-L., Henriet, J.-P., Laurent, M. & Vidier, J.-P. (2004) A tectonic inversion by steps during the Cenozoic: the Dover Strait. Bulletin de la Société Géologique de France, 175, 175-195. Vine, F.J. & Matthews, D.H. (1963) Magnetic anomalies over oceanic ridges. Nature, 199, 947-949. White, R.S. & McKenzie, D. (1989) Magmatism at rift zones: the generation of volcanic continental margins and flood basalts. Journal of Geophysical Research, 94, 7685-7729. Witkowski, J., Harwood, D.M. & Chin, K. (2011) Taxonomic composition, paleoecology and biostratigraphy of Late Cretaceous diatoms from Devon Island, Nunavut, Canadian High Arctic. Cretaceous Research, 32, 277-300. Worsley, D., Aga, O.J., Dalland, A., Elverhøi, A. & Thon, A. (1986) The geological history of Svalbard—Evolution of an Arctic Archipelago. Statoil, Stavanger. Wright, J.D. & Miller, K.G. (1996) Control of North Atlantic Deep Water circulation by the Greenland-Scotland Ridge. Paleoceanography, 11, 157-170. 18 Appendix S2: Palaeo-climatic conditions around the K/Pg boundary INTRODUCTION Climate is a major factor affecting the extension, structure, and composition of bioprovinces. Hence, past climatic oscillations are of special importance for understanding and interpreting past biotic changes. Relatively warm temperatures around the K/Pg boundary are proposed for the peri-Arctic regions. Specifically, in post-Cenomanian times, floral differentiation among near Arctic regions suggests a relatively warm Arctic Ocean [supplied by heat transported northwards along the Western Interior Seaway (WIS)], compared with a relatively cold northern Pacific Ocean gyre (Spicer & Herman, 2010; Zakharov et al., 2011). Perhaps because of the WIS, the temperature gradient during the latest Cretaceous has been interpreted to be less steep (0.4 ± 0.1 °C/°latitude) than it is in the present day (0.6 °C/°latitude). Temperatures in the latest Cretaceous are thought to have decreased from about 30 °C near the equator to about −5 °C at the poles; and air temperatures above 30° palaeolatitude are thought to have been higher than at present (Amiot et al., 2004). During the time interval when dinosaurs were present on the northern slope of Alaska, the mean temperature in that region is reported to have ranged from 2 to 4 °C, for the coldest monthly mean, to 10 to 12 °C, for the warmest monthly mean (Parrish et al., 1987). In general, although the evidence supports an icefree Arctic summer, the presence of intermittent sea ice in the Arctic winter (Davies et al., 2009; Spielhagen & Tripati, 2009; Spicer & Herman, 2010 and references therein) agrees with a wide body of evidence suggesting low winter temperatures in the Late Cretaceous Arctic region (Amiot et al., 2004; Falcon-Lang et al., 2004; Spicer & Herman, 2010 and references therein), making hypotheses of mean annual temperatures higher than 15 °C in the region (Jenkyns et al., 2004; Zakharov et al., 2011) seem doubtful. In any case, today it is well known that the latest Cretaceous (Maastrichtian)–earliest Palaeoegene (early Palaeocene) period was characterized by great climatic variability (Wolfe & Upchurch, 1987; Barrera, 1994; Li & Keller, 1998a, 1999; Francis & Poole, 2002; Nordt et al., 2003; MacLeod & Huber 2005; Twitchett, 2006; Gallagher et al., 2008; Spicer & Herman, 2010 and references therein; Zakharov et al., 2011; Hunter et al., 2008; Flögel et al., 2011) that may have been caused by tectonic reorganization of the oceans (Frank & Arthur, 1999) or/and the effects of the Deccan volcanism (e.g. Keller et al., 2012). Therefore, general warming and cooling periods are more important than the prevailing mean temperatures for our understanding of the palaeoecological and biotic changes that took place in the peri-Arctic region. For the purposes of the current study, the Kominz et al. (2008) eustatic sea-level curve from offshore and onshore New Jersey (Fig. 4 and Fig. S2.1 below) is used as a proxy for the palaeoclimatic changes that took place during the Maastrichtian–Palaeocene period. Because of limitations of the δ18O record (see below), the eustatic New Jersey sea-level curve is climatically more informative and useful for explaining the biotic changes that took place in the Northern Hemisphere (Fig. 4). For example, the discovery of dinosaur remains in the polar regions was unexpected, because dinosaurs were previously regarded as animals adapted to warm climates. In order to explain the discovery of dinosaurs at the higher latitudes, researchers proposed previously unknown differences between the metabolisms of dinosaurs and those of modern reptiles (Clemens & Nelms, 1993; Golovneva, 2000) and questioned whether the 19 dinosaurs overwintered at higher latitudes (Clemens & Nelms, 1993) or instead migrated south with the retreating sunlight (Parrish et al., 1987). The climatic model based on the Kominz et al. (2008) eustatic sea-level curve suggests that dinosaurs appeared in the peri-Arctic regions during the latest Cretaceous as a result of the latitudinal extension of their ranges during a significantly warmer period (Fig. 4). The sea level record from New Jersey as a climatic proxy Marine stable-isotope records provide the basis for much of our understanding of past climates. Oxygen isotope records, in particular, have been used to estimate past water temperatures, ice sheet volumes, and local salinity variations. Separating the effects of ice volume from those of temperature is very difficult, because despite the relative stability of the deep sea, changes in deep ocean temperatures also affect deep-sea benthic foraminiferal δ18O records (e.g. Wright, 2000). For example, the benthic foraminiferal δ18O records of the past 50 Myr show a 4‰ increase that must reflect mostly deep-water cooling; only c. 1.0‰ of the increase can be due to changes in ice volume (Miller et al., 2005; Miller et al., 2011). Thus, the long-term δ18O record of the last 100 Myr is thought to reflect c. 12 °C of cooling, complicating the use of δ18O as a proxy for ice volume beyond the Pliocene–Pleistocene (Miller et al., 2011 and references therein). Alternatively, Mg/Ca ratios have been used to provide a palaeothermometer that accounts for the temperature component in deep-sea benthic foraminiferal δ18O records. Because of the large errors associated with the Mg/Ca approach, however, the δ18O and Mg/Ca records cannot provide an unequivocal record of sea levels prior to the Pliocene. Therefore, we must look exclusively to the sedimentary record of sea-level change (Miller et al., 2011 and references therein). The sea-level record from onshore and offshore New Jersey (Miller et al., 2005; Kominz et al., 2008) is considered to be a good proxy of past eustatic sea-level fluctuations, presupposing the accumulation of at least ephemeral ice sheets in the Antarctic during the greenhouse world of the Late Cretaceous to middle Eocene (Miller et al., 2008). This supposition is currently under discussion; but it is supported by recent climatic models (Flögel et al., 2011). The main principle of the eustatic sea-level theory is the accumulation of surface water volume in the form of ice sheets at the high latitudes with a rhythm that follows long-term climate trends. On the basis of the eustatic sea-level record from New Jersey, Miller et al. (2008) reconstructed the evolution of the Antarctic ice-sheet during the Late Cretaceous–Cenozoic; whereas Cramer et al. (2011) reconstructed trends in ice volume and deep-ocean temperature for the past 108 Myr. Here, the Kominz et al. (2008) sea-level curve is used as a proxy for the large climatic fluctuations that took place during the latest Cretaceous (Maastrichtian)–earliest Palaeogene (early Palaeocene). To indicate the reliability of this climatic model, the major events presupposed and predicted by the model are correlated with independent data (see Figs 4 and S2.1). Correlation of the New Jersey eustatic sea-level curve with independent climatic proxies The Cretaceous greenhouse climatic mode terminated with a gradual global cooling from the late Campanian (c. 73 Ma) until nearly the end the Maastrichtian. Today, however, it is known that the Late Cretaceous cooling trend was interrupted by two episodes of greenhouse warming. The earlier episode, the mid-Maastrichtian event (MME), lasted from about 70 Ma to 68 Ma, while the second episode, the late 20 Maastrichtian event (LME), commenced 450 kyr before the K/T boundary and lasted only 300 kyr (e.g. Abramovich et al. 2007). The events are represented in the New Jersey eustatic sea-level curve as the CH3 and CH2 highstands (Fig. S2.1). The long-term cooling trend that culminated in the early Maastrichtian (CL4 sea-level lowstand in Fig. S2.1), here referred to as the early Maastrichtian cooling (EMC), is also recorded in marine (Barrera, 1994; MacLeod & Huber, 1996; Barrera et al., 1997; Li & Keller, 1998a, 1999; Zakharov et al., 1999; Friedrich et al., 2004, 2009; Cramer et al., 2009) and terrestrial oxygen stable isotopes (Nordt et al., 2003). The formation of ice sheets in high southern latitudes is thought to be a consequence of the EMC (Miller et al., 1999, 2003). Jagt-Yazykova (2011) noted that during the EMC, most ammonite forms disappeared from the Arctic, possibly as a result of shortlived, subfreezing conditions that occasionally occurred in the Northern Hemisphere (Jagt-Yazykova, 2011 and references therein). Furthermore, the retreat of the sea from the Cretaceous Western Interior Seaway (WIS) (Erickson, 1978; Lillegraven & Ostresh, 1990; Boyd & Lillegraven, 2011) at the boundary of the Baculites clinolobatus and Baculites grandis zones (70.5 Ma; Ogg & Hinnov, 2012) match perfectly with the CL4 sea-level lowstand (70.7 Ma). The cooling period was succeeded by a hyperthermal event known as the MME (CH3 sea-level highstand in Fig. S2.1). The MME is supported by both the terrestrial record (Nordt et al., 2003) and the marine record. In the marine record, the MME encompasses the global extinctions of the inoceramid and rudistid bivalves and the latitudinal migrations of some calcareous nannoplankton and planktic foraminifera. The MME is expressed as an increase in the rate of global climate cooling, a decrease in the global range of benthic foraminiferal δ13C values from approximately 3% to less than 1%, and an increase in the rate of change of seawater 87Sr/86Sr isotope ratios (Frank et al., 2005 and references therein). Jagt-Yazykova (2011) noted that such a temperature maximum in the Arctic Basin at the base of the late Maastrichtian resulted in one of the fastest ammonite radiations during the Late Cretaceous, reaching the levels of the ammonite radiations of the late Turonian (JagtYazykova, 2011 and references therein). In the Bering area, such a warm period was recorded in isotopic palaeotemperatures estimated by δ18O values extracted from ammonoid shells in southern Alaska sampled from the late part of the early Maastrichtian (Zakharov et al., 2011: Figure 13). The succession of EMC to MME can be recognized in the Deep Sea Drilling Programme (DSDP) sites 525A and 21 in the southern Atlantic (Li & Keller, 1998a, b) and 463 in the equatorial Pacific (Li & Keller, 1999) where, within the long-term cooling trend of the late Campanian–Maastrichtian, two prolonged deep-water cooling events are observed at Chron C31r and C30n. These events are ‘associated with more gradual cooling in surface waters and separated by temporary warming of 2–3 °C’ (at Chron C31N) in deep waters (Li & Keller, 1998a, p. 82). These cooling events match very well with the CL4 and CL2 sea-level lowstands of the Kominz et al. (2008) sealevel curve (Fig. S2.1), while the ‘temporary warming’ correlates with the CH3 sealevel highstand. The CL2 sea-level lowstand was followed by the CH2 sea-level highstand (Chron late-C30N) and the CL1 sea-level lowstand. Both events are known from the marine record (Li & Keller, 1998a, 1999; Olsson et al., 2001). …near the end of the Maastrichtian (Chron C29R), beginning about 400 kyr before the K/T boundary and lasting about 100–200 kyr, surface and deep waters warmed rapidly by 3-4˚C and then sharply cooled by 2–3˚C during the last 100–200 kyr of the Maastrichtian. 21 (Li & Keller, 1998a, p. 82) In the terrestrial record, the sequence CH2-CL1-CH1(PH1) was printed on the exported mean annual temperatures from leaf-margin analyses in North Dakota (Wilf et al., 2003); while the intermediate CH2 (Chron late-C30N) warming event was independently recognized and referred to as the late Maastrichtian event (LME) by Nordt et al. (2003). Even the last warming trend of the Cretaceous (leading to the CH1 sea-level highstand shown in Fig. S2.1) is recorded in both the terrestrial (Wilf et al., 2003) and the marine oxygen stable-isotope record (Li & Keller, 1998b). The notable warming led to the PH1 sea-level highstand, referred to here as the early Danian warming (EDW), and is very close to the recently reported hyperthermal Dan-C2 event (Quillévéré et al., 2008). The EDW is recorded in the marine oxygen isotope record in almost all DSDP sites (e.g. Cramer et al., 2009, see the δ18O curve in Fig. 4; see also Tobin et al., 2012 for a recent Antarctic record). In contrast, the PL1 sea-level lowstand shown in Fig. 4, here referred to as the midDanian cooling (MDC), is not represented in the δ18O record to the extent expected. A better image of a Danian cooling just after the warming of the K/Pg boundary, based on planktonic foraminifera and clay mineralogy, is reported from Kazakhstan (Pardo et al., 1999). Jagt-Yazykova (2011) noted that evidence from the Russian Far East suggests a generally abrupt and strong climatic cooling during the Danian with annual temperatures less than 5°C above freezing (Jagt-Yazykova, 2011 and references therein). The PH3 sea-level highstand corresponds perfectly with recent estimations of the latest Danian event (LDE): a hyperthermal event that has been unequivocally identified in benthic foraminiferal isotopes from shelf sediments in Egypt (Bornemann et al., 2009) and in deep-sea material from the Pacific Ocean (Westerhold et al., 2011). Instead of the early estimations for a ‘Top C27n’ age of this hyperthermal event, recent studies suggest an early Chron C26r (Bornemann et al., 2009 and references therein) or 61.75 Ma age for this event (Westerhold et al., 2011), matching perfectly with the 61.8 Ma age of the PH3 sea-level highstand. Recently, Schulte et al. (2013) compared the LDE with the Palaeocene–Eocene Thermal Maximum (PETM; 55.8 Ma) and concluded that both hyperthermal events resulted in significant eustatic sea-level highstands. The arrows in Fig. 4 show the contrasting impression of how a significant climatic event is not obvious in the mean values of the δ18O record but is remarkably clear in the Kominz et al. (2008) sea-level curve. In general, the Kominz et al. (2008) sea-level model is congruent with the changes that took place in the palaeofloristic assemblages, the presence-absence of climatic-indicator taxa such as Cycadophytes (e.g. Golovneva, 2000), and the anatomy of fossilized woods. Thus, the long-term global cooling that culminated during the early Maastrichtian (EMC) (CL4 sea-level lowstand in Fig. S2.1) is also recorded by palaeobotanical (Wolfe & Upchurch, 1987) and fossilized wood-anatomy studies (Francis & Poole, 2002). During the EMC, the occurrence of the Cycadophytes-bearing thermophilous vegetation was very rare in the Bering area (only one flora in the Koryak Upland area); whereas such vegetation was found everywhere in both northern Alaska and north-eastern Russia during the subsequent warm period of the MME (CH3 sea-level highstand in Fig. S2.1) (Zakharov et al., 2011 and references therein). This warming trend is also supported by palaeobotanical evidence from North America (Wolfe & Upchurch, 1987), the presence of dinosaurs in the high-latitude peri-Arctic regions, and possibly the occurrence of growth-ring characteristics in fossilized rare woods in northern Alaska (Spicer & Herman, 2010). 22 Near the Maastrichtian–Danian boundary interval (CL1 sea-level lowstand in Fig. S2.1), Cycadophytes-bearing, thermophilous vegetation are not found in any of the Bering floras (Zakharov et al., 2011 and references therein). Correlation of the New Jersey eustatic sea-level curve with terrestrial climatic proxies The terrestrial δ18O record does not have the restrictions of the corresponding marine record in its proxy representation of past climatic fluctuations. It is, however, governed by another limitation; in the absence of marine faunas, it usually cannot be well dated. Nordt et al. (2003) reported such a record of δ18O and δ13C values measured in palaeosol carbonate from North America. A direct correlation with the Kominz et al. (2008) eustatic sea-level curve, however, encounters two main problems. First, the record is calibrated on an older time-scale; and second, the two recognized hiatuses add uncertainty to the time-scale. Figure S2.1 attempts to recalibrate the record of Nordt et al. (2003: Figure 2) with the Gradstein et al. (2012) geological time-scale and the Kominz et al. (2008) eustatic sea-level curve. For this propose, the three sections A, B, and C of the Nordt et al. (2003: Figure 2) curve are re-aligned by photo editing as follows: Section A: Nordt et al. (2003) correlated the uppermost end of their curve to Chron C28R. The new applied age for this Chron is 64.7 to 65 Ma. A second calibration point is the new age of the K/Pg boundary c. 66 Ma (previously recognized stratigraphically and originally set c. 65 Ma by Nordt et al., 2003). A third calibration point is the alignment of the LME, also recognized by Nordt et al. (2003) (peak 1 in Fig. S2.1), to peak CH2 of the Kominz et al. (2008) eustatic sea-level curve. Section B: The peak of the MME that was recognized by Nordt et al. (2003) (peak 2 in Fig. S2.1) aligned to sea-level highstand CH3 of the Kominz et al. (2008) eustatic sea-level curve. The sequence boundaries (that are recognized as hiatuses by Nordt et al., 2003) correlate with the sequence boundaries of the Kominz et al. (2008) eustatic sea-level curve. Section C: Nordt et al. (2003) set the lowermost section of the curve on the Campanian-Maastrichtian boundary, originally c. 71 Ma, now c. 72 Ma according to the 2012 time-scale. Although the above approach is simplified, it highlights certain similarities between the terrestrial climate proxies of Nordt et al. (2003) and the Kominz et al. (2008) eustatic sea-level curve. This congruency, along with the agreement with the marine proxies mentioned above, suggests that the eustatic sea-level curve from New Jersey is a reliable global climatic proxy. 23 Figure S2.1 Correlation of the Kominz et al. (2008) eustatic sea-level curve with independent proxies from terrestrial and marine records. The sea-level curve is reconstructed with data from Kominz et al. (2008: supplementary material). The curves of the δ18O and δ13C values were measured from paleosol carbonate from North America (Nordt et al., 2003: Figure 2) and recalibrated here (see text). The benthic δ18O isotope curve is from Li & Keller (1999: Figure 2). The grey shadow reflects the uncertainty in the dating of the curve as originally designed by Nordt et al. (2003: Figure 2). 24 REFERENCES Abramovich, S., Benjamini, C. & Almogi-Labin, A. (2007) Global extinction of intermediate-thermocline planktic foraminifera at the mid Maastrichtian warm event. Geophysical Research Abstracts, 9, 05527. Amiot, R., Lécuyer, C., Buffetaut, E.,Fluteau, F.,Legendre, S. & Martineau, F. (2004) Latitudinal temperature gradient during the Cretaceous Upper Campanian– Middle Maastrichtian: δ18O record of continental vertebrates. Earth and Planetary Science Letters, 226, 255-272. Barrera, E. (1994) Global environmental changes preceding the Cretaceous–Tertiary boundary: early–late Maastrichtian transition. Geology, 22, 877-880. Barrera, E., Savin, S.M., Thomas, E. & Jones, C.E. (1997) Evidence for thermohalinecirculation reversals controlled by sea-level change in the latest Cretaceous. Geology, 25, 715-718. Bornemann, A., Schulte, P., Sprong, J., Steurbaut, E., Youssef, M. & Speijer, R.P. (2009) Latest Danian carbon isotope anomaly and associated environmental change in the southern Tethys (Nile Basin, Egypt). Journal of the Geological Society, London, 166, 1135-1142. Boyd, D.W. & Lillegraven, J.A. (2011) Persistence of the Western Interior Seaway. Historical background and significance of ichnogenus Rhizocorallium in Paleocene strata, south-central Wyoming. Rocky Mountain Geology Spring, 46, 43-69. Clemens, W.A. & Nelms, L.G. (1993) Paleoecological implications of Alaskan terrestrial vertebrate fauna in latest Cretaceous time at high paleolatitudes. Geology, 21, 503–506. Cramer, B.S., Miller, K.G., Barrett, P.J. & Wright, J.D. (2011) Late Cretaceous– Neogene trends in deep ocean temperature and continental ice volume: Reconciling records of benthic foraminiferal geochemistry (δ18O and Mg/Ca) with sea level history. Journal of Geophysical Research: Oceans (1978–2012), 116, C12. Davies, A., Kemp, A.E.S. & Pike, J. (2009) Late Cretaceous seasonal ocean variability from the Arctic. Nature, 460, 254-258. Erickson, J.M. (1978) Bivalve mollusk range extensions in the Fox Hills Formation (Maestrichtian) of North and South Dakota and their implications for the Late Cretaceous geologic history of the Williston Basin. North Dakota Academy of Science Annual Proceedings, 32, 79-89. Falcon-Lang, H.J., MacRae, R.A. & Csank, A.Z. (2004) Palaeoecology of Late Cretaceous polar vegetation preserved in the Hansen Point volcanics, NW Ellesmere Island, Canada. Palaeogeography, Palaeoclimatology, Palaeoecology, 212, 45-64. Flögel, S., Wallmann, K. & Kuhnt, W. (2011) Cool episodes in the Cretaceous — Exploring the effects of physical forcings on Antarctic snow accumulation. Earth and Planetary Science Letters, 307, 279-288. Francis, J.E. & Poole, I. (2002) Cretaceous and early Tertiary climates of Antarctica: evidence from fossil wood. Palaeogeography, Palaeoclimatology, Palaeoecology, 182, 47-64. Frank, T.D. & Arthur, M.A. (1999) Tectonic forcings of Maastrichtian ocean-climate evolution. Paleoceanography, 14, 103-117. Friedrich, O., Herrle, J.O., Kössler, P. & Hemleben, C. (2004) Late Campanian/Early Maastrichtian stable isotopes on a N-S transect: deep water and glaciation signal? Geophysical Research Abstracts, 6, 01305. 25 Friedrich, O., Herrle, J.O., Wilson, P.A., Cooper, M.J., Erbacher, J. & Hemleben, C. (2009) Early Maastrichtian carbon cycle perturbation and cooling event: implications from the South Atlantic Ocean. Paleoceanography, 24, PA2211. Gallagher, S.J., Wagstaff, B.E., Baird, J.G., Wallace, M.W. & Li, C.L. (2008) Southern high latitude climate variability in the Late Cretaceous greenhouse world. Global and Planetary Change, 60, 351-364. Golovneva, L.B. (2000) Palaeogene climates of Spitsbergen. GFF (Geological Society of Sweden), 122, 62-63. Hunter, S.J., Valdes, P.J., Haywood, A.M. & Markwick, P.J. (2008) Modelling Maastrichtian climate: investigating the role of geography, atmospheric CO2 and vegetation. Climate of the Past Discussions, 4, 981-1019. Jagt-Yazykova, E.A. (2011) Palaeobiogeographical and palaeobiological aspects of mid- and Late Cretaceous ammonite evolution and bio-events in the Russian Pacific. Scripta Geologica, 143, 15-121. Jenkyns, H.C., Forster, A., Schouten, S. & Sinninghe Damsté, J.S. (2004) High temperatures in the Late Cretaceous Arctic Ocean. Nature, 432, 888-892. Keller, G., Adatte, T., Bhowmick, P.K., Upadhyay, H., Dave, A., Reddy, A.N. & Jaiprakash, B.C. (2012) Nature and timing of extinctions in Cretaceous– Tertiary planktic foraminifera preserved in Deccan intertrappean sediments of the Krishna–Godavari Basin, India. Earth and Planetary Science Letters, 341344, 211-221. Li, L. & Keller, G. (1998a) Maastrichtian climate, productivity and faunal turnovers in planktic foraminifera in South Atlantic DSDP sites 525A and 21. Marine Micropaleontology, 33, 55-86. Li, L. & Keller, G. (l998b) Abrupt deep-sea warming at the end of the Cretaceous. Geology, 26, 995-998. Li, L. & Keller, G. (1999) Variability in Late Cretaceous climate and deep waters: evidence from stable isotopes. Marine Geology, 161, 171-190. Lillegraven, J.A. & Ostresh, L.M., Jr (1990) Late Cretaceous (earliest Campanian/Maastrichtian) evolution of western shorelines of the North American Western Interior Seaway in relation to known mammalian faunas. Dawn of the Age of Mammals in the northern part of the Rocky Mountain Interior, North America (ed. by T.M. Bown and K.D. Rose), pp. 1-30. Geological Society of America, Special Paper, 243. Boulder, CO. MacLeod, K.G. & Huber, B.T. (1996) Reorganization of deep ocean circulation accompanying a Late Cretaceous extinction event. Nature, 380, 422-425. MacLeod, K.G. & Huber, B.T. (2005) North Atlantic warming during global cooling at the end of the Cretaceous. Geology, 33, 437-440. Miller, K.G., Barrera, E., Olsson, R.K., Sugarman, P.J. & Savin, S.M. (1999) Does ice drive early Maastrichtian eustasy? Geology, 27, 783-786. Miller, K.G., Sugarman P.J., Browning, J.V., Kominz, M.A., Hernandez, J.C., Olsson, R.K., Wright, J.D., Feigenson, M.D. & Van Sickel, W. (2003) Late Cretaceous chronology of large, rapid sea-level changes: glacioeustasy during the greenhouse world. Geology, 31, 585-588. Miller, K.G., Kominz, M.A., Browning, J.V.,Wright, J.D., Mountain, G.S., Katz, M.E., Sugarman, P.J., Cramer, B.S., Christie- Blick, N. & Pekar, S.F. (2005) The Phanerozoic record of global sea-level change. Science, 310, 1293-1298. Miller, K.G., Wright, J.D., Katz, M.E., Browning, J.V., Cramer, B.S., Wade, B.S. & Mizintseva, S.F. (2008) A view of Antarctic ice-sheet evolution from sea-level and deep-sea isotope changes during the Late Cretaceous–Cenozoic. Antarctica: a keystone in a changing world (ed. by A.K. Cooper, P.J. Barrett, H. Stagg, B. Storey, E. Stump, W. Wise et al.), pp. 55-70. Proceedings of the 26 10th International Symposium on Antarctic Earth Sciences. The National Academies Press, Washington, DC. Miller, K.G., Mountain, G.S., Wright, J.D. & Browning, J.V. (2011) A 180-millionyear record of sea level and ice volume variations from continental margin and deep-sea isotopic records. Oceanography, 24, 40–53. Ogg, J.G. & Hinnov, L.A. (2012) Cretaceous. The geological timescale 2012 (ed. by F.M. Gradstein, J.G. Ogg, M.D. Schmitz and G.M. Ogg), pp. 85-113. Elsevier, Amsterdam. Olsson, R.K., Wright, J.D. & Miller, K.G. (2001) Paleobiogeography of Pseudotextularia elegans during the latest Maastrichtian global warming event. Journal of Foraminiferal Research, 31, 275-282. Pardo, A., Addate, T., Keller, G. & Oberhänsli, H. (1999) Paleoenvironmental changes across the Cretaceous–Tertiary boundary at Koshak, Kazakhstan, based on planktonic foraminifera and clay mineralogy. Palaeogeography, Palaeoclimatology, Palaeoecology, 154, 247–273. Parrish, M.J., Parrish, J.T., Hutchinson, J.H. & Spicer, R.A. (1987) Late Cretaceous vertebrate fossils from the North Slope of Alaska and implications for dinosaur ecology. Palaios, 2, 377–389. Schulte, P., Schwark, L., Stassen, P., Kouwenhoven, T.J., Bornemann, A. & Speijer, R.P. (2013) Black shale formation during the Latest Danian Event and the Paleocene–Eocene Thermal Maximum in central Egypt: two of a kind? Palaeogeography, Palaeoclimatology, Palaeoecology, 371, 9-25. Spielhagen, R.F. & Tripati, A. (2009) Evidence from Svalbard for near-freezing temperatures and climate oscillations in the Arctic during the Paleocene and Eocene. Palaeogeography, Palaeoclimatology, Palaeoecology, 278, 48-56. Tobin, T.S., Ward, P.D., Steig, E.G., Olivero, E.B.,, Hilburn, I.A., Mitchell, R.N., Diamond, M.R., Raub, T.D. & Kirschvink, J.L. (2012) Extinction patterns, δ18O trends, and magnetostratigraphy from a southern high-latitude Cretaceous–Paleogene section: Links with Deccan volcanism. Palaeogeography, Palaeoclimatology, Palaeoecology, 350–352, 180-188. Twitchett, R.J. (2006) The palaeoclimatology, palaeoecology and palaeoenvironmental analysis of mass extinction events. Palaeogeography, Palaeoclimatology, Palaeoecology, 232, 190-213. Westerhold, T., Röhl, U., Donner, B., McCarren, H.K. & Zachos, J.C. (2011) A complete high-resolution Paleocene benthic stable isotope record for the central Pacific (ODP Site 1209). Paleoceanography, 26, PA2216. Wilf, P., Johnson, K.R. & Huber, B.T. (2003) Correlated terrestrial and marine evidence for global climate changes before mass extinction at the Cretaceous– Paleogene boundary. Proceedings of the National Academy of Sciences USA, 100, 599-604. Wolfe, J.A. & Upchurch, G.R., Jr (1987) North American nonmarine climates and vegetation during the Late Cretaceous. Palaeogeography, Palaeoclimatology, Palaeoecology, 61, 33-77. Wright, J.D. (2000) Global climate change in marine stable isotope records. Quaternary geochronology: methods and applications (ed. by J.S. Noller, J.M. Sowers and W.R. Lettis), pp. 671-682. American Geophysical Union, Washington, DC. Zakharov, Y.D., Boriskina, N.G., Ignatyev, A.V., Tanabe, K., Shigeta, Y., Popov, A.M., Afanasyeva, T.B. & Maeda, H. (1999) Palaeotemperature curve for the Late Cretaceous of the northwestern circum-Pacific. Cretaceous Research, 20, 685-697. 27 Appendix S3: Vertebrate evidence for the existence of Northern Hemisphere land bridges DINOSAURS To investigate the exact dispersal course through Beringia or the De Geer route, it would be of special importance to determine whether a specific taxon that migrated from Eurasia to North America was recovered from the western or the eastern bank of the WIS. This is because the WIS separated western (Laramidia) and eastern (Appalachia) North America. Taxa recovered from the western bank would suggest dispersal via Beringia; whereas taxa recovered from the eastern bank would suggest dispersal via the De Geer route. The retreat of the Cretaceous WIS in the early Maastrichtian (corresponding to the CL4 sea-level lowstand shown in Fig. 4), however, resulted in the subaerial reconnection of Laramidia and Appalachia (Erickson, 1978; Lillegraven & Ostresh, 1990; Boyd & Lillegraven, 2011), thus ‘complicating biogeographical interpretations’ (Sampson et al., 2010: e12292). The retreat of the Cretaceous WIS is probably why research on dinosaur communities along the WIS indicates low beta diversity and suggests a single dinosaur community within the entire WIS region in Maastrichtian North America (Vavrek & Larsson, 2010). At least one genus of (probably) theropod dinosaurs is thought to have been shared between Europe and North America during the Maastrichtian. The ‘Euronychodon’ portucalensis from Maastrichtian Portugal and the Paronychodon from Late Cretaceous North America are proposed to be congeneric (e.g. Antunes & Mateus, 2003; Pereda-Suberbiola, 2009). The true affinities of both taxa are difficult to determine, however, because the taxa are known only from fossils of teeth (Antunes & Mateus, 2003; Larson & Currie, 2013). Therefore, although they may be valid taxa, they cannot yet be taken as reliable evidence of intercontinental dispersal. The hadrosaurs were a very diverse and successful dinosaur group that dominated the Laurasian landmasses during the later stages of the Late Cretaceous (Horner et al., 2004). Latest Cretaceous representatives of the Lambeosaurinae subfamily have been recovered from all three Laurasian continents: Asia, North America, and Europe. The closely related taxa suggest land connections in the Northern Hemisphere (Godefroit et al., 2003; Cruzado-Caballero et al., 2011). Two representative pairs of these taxa are the Corythosaurus–Olorotitan from Campanian North America and late (or ‘mid-‘) Maastrichtian Asia (Godefroit et al., 2003; 2012), and the Parasaurolophus–Blasisaurus from Campanian North America and late Masstrichtian Europe (Cruzado-Caballero et al., 2010, 2011) (Fig. 6). Prieto-Marquez & Wagner (2009) demonstrated evidence suggesting that Koutalisaurus kohlerorum from early Maastrichtian to early–late Maastrichtian (Pereda-Suberbiola et al., 2009) Lleida Province in north-eastern Spain is most probably the junior synonym of Pararhabdodon isonensis from the same region (see the review of European taxa in Pereda-Suberbiola et al., 2009), providing conclusive evidence of the presence of the Lambeosaurinae in Europe. The same authors proposed that Tsintaosaurus spinorhinus from Asia forms a clade with Pararhabdodon isonensis, suggesting that this clade originated in Asia during the middle or late Campanian and migrated to the Iberian island of the European archipelago during the Maastrichtian. 28 Another related pair of hadrosaurs is Arenysaurus–Charanosaurus from the Late Cretaceous of Portugal and the Russian Far East (Amur Region) (Godefroit et al., 2003; Cruzado-Caballero et al., 2011). The phylogenetic analysis of Godefroit et al. (2012) suggested a closer relationship between Parasaurolophus and Charanosaurus; their analysis does not, however, include the closely related European genus Blasisaurus. In any case, even with the different topologies, the closed relationships among the lambeosaurins of the three continents of the ‘middle’ or late Maastrichtian Northern Hemisphere constitute clear evidence of faunal exchanges. Within the sauroloph clade, the genus Saurolophus is known both in western Canada (Saurolophus osborni, Horseshoe Formation, lower Maastrichtian) and in Mongolia (Saurolophus angustirostris, Nemegt Formation, lower Maastrichtian) (Bell, 2011 and references therein). Its sister taxon, Prosaurolophus, is only known in western North America and is older (late Campanian, 75 Ma; Gates & Farke, 2009 and reference therein), suggesting dispersal from North America to Asia (Godefroit et al., 2011). Recently, the genus Saurolophus was identified in the Almond Formation of Wyoming dated c. 72 Ma (Gates & Farke, 2009). This discovery further supports the anagenesis of the Prosaurolophus–Saurolophus clade and its later dispersal to Asia via the De Geer route. The Alvarezsaurids are a group of small maniraptoran theropod dinosaurs. Derived members of the Alvarezsaurids (the Parvicursorinae) display many derived features that also occur in some birds, and they were originally interpreted as very basal avians (Xu et al, 2010 and references therein). Alvarezsaurids have been found in Asia, South America, and North America. Previous biogeographical interpretations proposed dispersal from South America to North America and then to Asia (Longrich & Currie, 2009 and references therein). More recent discoveries show, however, that Laurasian taxa can be grouped into a separate derived clade, the Parvicursorinae, with the exclusion of the South American taxa (Xu et al, 2010; 2011). In this scenario, the ‘middle’ and late Maastrichtian representatives in North America (Albertonykus borealis and some other unnamed specimens; see Longrich & Currie, 2009: Figs 1 & 2) are considered to have dispersed from Late Cretaceous Asian populations (Xu et al, 2011). On the basis of recovered teeth, Godefroit et al. (2009) recognized the presence of dinosaur genera such as Dromaeosaurus and Saurornitholestes at Kakanaut in north-eastern Russia. These genera are also known from the Russian Far East (Van Itterbeeck et al., 2005) as well as from the Alaskan (Fiorillo & Gangloff, 2000; Fiorillo, 2008) and mid-latitude sites of North America (e.g. Russell & Manabe, 2002). The North American theropod Troodon formosus is considered to have been present on both sides of the Bering area (in northern Alaska and in Kakanaut, in the Russian Far East), supporting dispersal to Asia via Beringia (Zakharov et al., 2011). This taxon is considered to have been a year-round resident of the Arctic regions that was adapted to the cooler climate and winter darkness (Fiorillo & Gangloff, 2000). Nevertheless, Troodon formosus is known from both sides of the WIS in North America from northern Alaska to Wyoming. Ιts presence in Kakanaut in the Russian Far East is based on scarce tooth remains and is tentatively referred to as Troodon cf. formosus. Recently, however, such tooth morphology was concluded to be widely distributed among the troodontids and hence cannot be regarded as diagnostic at the generic level (Godefroit et al., 2009 and references therein). In any case, teeth from 29 troodontids, small theropods, and hadrosaurs have been recovered from all known dinosaur deposits in the peri-Arctic region, i.e. the northern slope of Alaska, Siberia, north-eastern Russia, and Bylot Island (northern Canada) (Rich et al., 2002; Godefroit et al., 2009). Contrary to the arguments of Fiorillo (2008) and Zakharov et al., (2011), the latest Cretaceous dinosaur affinities among the Laurasian continents suggest dispersal via the De Geer route, because the floristic record shows that Beringia was probably transgressed and not exposed during the Maastrichtian (Fig. 7c). This conclusion is congruent with the interpretation of Godefroit et al. …the development of very different kinds of dinosaur communities during the Maastrichtian may reflect some kind of geographical isolation between eastern Asia and western North America during this time, or important differences in climatic or palaeoecological conditions. (Godefroit et al., 2011, p. 184) REPTILES The crocodylians include the clades Crocodyloidea, Gavialoidea, and Alligatoroidea, which together incorporate all of the current species of crocodiles, alligators, caimans, and gharials. The crocodyloids are considered to have a Laurasian origin, and their earliest members are the closely related Prodiplocynodon langi, from a Maastrichtian terrestrial horizon in the Lance Formation of Wyoming (Mook, 1941), and Arenysuchus gascabadiolorum, from late Maastrichtian (within Chron C30n) Spain (Puertolas et al., 2011) (Fig. 6). The gavialoids are also believed to have originated from Laurasia; the oldest known member, Eothoracosaurus mississippiensis, was found in the late Campanian or early Maastrichtian Ripley Formation in Mississippi (Brochu, 2004). By the Late Cretaceous, the species Thoracosaurus neocesariensis and Thoracosaurus macrorhynchus appeared, respectively, in the latest Maastrichtianearliest Palaeocene of New Jersey and the early Palaeocene of France, Poland, and Sweden (Zarski et al., 1998; Brochu, 2004 and references therein). Phylogenetic analyses (Delfino et al., 2005; Puertolas et al, 2011) support the close relationship between the crocodyloid and gavialoid taxa from North America and Europe and thus provide evidence for faunal exchange via the De Geer route around the K/Pg boundary. Late Palaeocene Europe and North America shared another gavialoid genus represented, in each continent respectively, by Eosuchus lerichei and Eosuchus minor (Delfino et al., 2005 and references therein; Brochu, 2006). In Belgium, E. lerichei was found in the same layer as the gavial-like neochoristodere Champsosaurus, which is known also from the late Palaeocene of France and the Late Cretaceous–Palaeocene localites of North America (Matsumoto & Evans, 2010 and references therein). The proposed mid-Thanetian age of the fossil beds in Belgium (Delfino et al., 2005) suggests that both Eosuchus and the Champsosaurus probably migrated via the first exposure of the Thulean route c. 57 Ma (Fig. 3d), rather than via the De Geer route. The turtle genus Compsemys (otherwise restricted to the Late Cretaceous and Palaeocene of North America) is known also from the Cernaysian (late Palaeocene) of France, suggesting dispersal via the Thulean route c. 57 Ma (Godinot & Lapparent de Broin, 2003; Lyson & Joyce, 2011). Godinot & Lapparent de Broin (2003) postulated that early Cenozoic herpetofaunal relations support direct dispersal from Asia to Europe via the Turgai Strait. They note that contrary to mammals, there is no turtle genus among the early 30 Eocene immigrants that is common to Europe and North America. Therefore, the turtles that arrived in Europe in the early Neustrian had to come via a route other than North America, implying possible dispersal from Asia. Such hypotheses, however, cannot currently be verified at the genus and species level, because most of the specimens are not complete enough. The lizard family Agamidae was thought to have reached Europe directly from Asia. The Agamidae are known from the Late Cretaceous of Asia (e.g. Dashzeveg et al., 1995) and have been identified in the late Palaeocene of Germany (Weigelt, 1940), where Agamidae indet. was found together with MP6 faunas such as Arctocyon sp. and Plesiadapis tricuspidens, thus confirming the Cernaysian age of the fossils. Agamid lizards made their first undeniable appearance in Europe in the early Eocene (MP7, locality of Dormaal) in the form of a single genus and species: Tinosaurus europeocaenus (Augé & Smith, 1997). The agamids managed to enter North America only later, during the Eocene, as demonstrated by the presence of the genus Tinosaurus in the middle Eocene (Bridgerian) of Wyoming (Gunnell & Bartels, 2001). Thus, the biogeography of the agamids provides strong evidence for terrestrial access via the Turgai Strait, at least in the earliest Eocene. Some early papers proposed that the snake family Boidae was present around K-Pg times in both Europe and South America (Rage, 1984; Le Loeuff, 1991), suggesting dispersals via the De Geer route. This proposal remains tentative, however, because the evidence is scarce. AMPHIBIANS The neochoristodere Champsosaurus is known from one of the few fossiliforous sites in north-eastern North America: the mid-Palaeocene (Tiffanian) Roche Percée in Saskatchewan, Canada, where the oldest representative of the living salamander genus Cryptobranchus and the family Cryptobranchidae was revealed (C. saskatchewanensis; Naylor, 1981). Recently, Skutschas (2009) concluded that two salamander genera coexisted in the Turonian Byssekty Formation in Uzbekistan: the cryptobranchid Eoscapherpeton and the cryptobranchoid Nesovtriton. According to this interpretation, although stem cryptobranchoids could also have been present in Cretaceous North America, the family Cryptobranchidae originated in Late Cretaceous Asia and dispersed to North America later, around the K/Pg boundary, via the De Geer route (Fig. 6). 31 REFERENCES Antunes, M.T. & Mateus, O. (2003) Dinosaurs of Portugal. Comptes Rendus Palevol, 2, 77-95. Augé, M. & Smith, R. (1997) Les Agamidae (Reptilia, Squamata) du Paleogene d’Europe occidentale. Belgian Journal of Zoology, 127, 123-138. Boyd, D.W. & Lillegraven, J.A. (2011) Persistence of the Western Interior Seaway. Historical background and significance of ichnogenus Rhizocorallium in Paleocene strata, south-central Wyoming. Rocky Mountain Geology Spring, 46, 43-69. Cruzado-Caballero, P., Pereda-Suberbiola, X. & Ruiz-Omeñaca, J.I. (2010) Blasisaurus canudoi gen. et sp. nov., a new lambeosaurine dinosaur (Hadrosauridae) from the Latest Cretaceous of Arén (Huesca, Spain). Canadian Journal of Earth Sciences, 47, 1507-1517. Dashzeveg, D., Novacek, M.J., Norell, M.A., Clark, J.M., Chiappe, L.M., Davidson, A., McKenna, M.C., Dingus, L., Swisher, C. & Altangerel, P. (1995) Extraordinary preservation in a new vertebrate assemblage from the Late Cretaceous of Mongolia. Nature, 374, 446-449. Erickson, J.M. (1978) Bivalve mollusk range extensions in the Fox Hills Formation (Maestrichtian) of North and South Dakota and their implications for the Late Cretaceous geologic history of the Williston Basin. North Dakota Academy of Science Annual Proceedings, 32, 79-89. Fiorillo, A.R. (2008) Cretaceous dinosaurs of Alaska: implications for the origins of Beringia. The Terrane Puzzle: new perspectives on paleontology and stratigraphy from the North American Cordillera (ed. by R.B. Blodgett and G. Stanley), pp. 313-326. Geological Society of America, Special Paper, 442. Boulder, CO. Gates, T.A. & Farke, A.A. (2009) Biostratigraphic and biogeographic implications of a hadrosaurid (Ornithopoda: Dinosauria) from the Upper Cretaceous Almond Formation of Wyoming, USA. Cretaceous Research, 30, 1157-1163. Godefroit, P., Lauters, P., Van Itterbeeck, J., Bolotsky, Y.L., Zhiming, D., Liyong, J., Wenhao, W.U., Bolotsky, I.Y., Shulin, H. & Tingxiang, Y.U. (2011) Recent advances on study of hadrosaurid dinosaurs in Heilongjiang (Amur) River area between China and Russia. Global Geology, 14, 160-191. Godefroit, P., Bolotsky, Y.L. & Bolotsky, I.Y. (2012) Osteology and relationships of Olorotitan arharensis, a hollow-crested hadrosaurid dinosaur from the latest Cretaceous of Far Eastern Russia. Acta Palaeontologica Polonica, 57, 527560. Gunnell, G.F. & Bartels, W.S. (2001) Basin margins, biodiversity, evolutionary innovation, and the origin of new taxa. Eocene biodiversity: unusual occurrences and rarely sampled habitats (ed. by G. F. Gunnell), pp. 403-432. Springer, New York. Horner, J.R., Weishampel, D.B. & Forster, C.A. (2004) Hadrosauridae. The Dinosauria, 2nd edn (ed. by D.B. Weishampel, P. Dodson and H. Osmolska), pp. 438–463. University of California Press, Berkeley, CA. Larson, D.W. & Currie, P.J. (2013) Multivariate analyses of small theropod dinosaur teeth and implications for paleoecological turnover through time. PLoS ONE, 8, e54329. Le Loeuff, J. (1991) The Campano-Maastrichtian vertebrate faunas from southern Europe and their relationships with other faunas in the world; palaeobiogeographical implications. Cretaceous Research, 12, 93-114. 32 Lillegraven, J.A. & Ostresh, L.M., Jr (1990) Late Cretaceous (earliest Campanian/Maastrichtian) evolution of western shorelines of the North American Western Interior Seaway in relation to known mammalian faunas. Dawn of the Age of Mammals in the northern part of the Rocky Mountain Interior, North America (ed. by T.M. Bown and K.D. Rose), pp. 1-30. Geological Society of America, Special Paper, 243. Boulder, CO. Mook, C.C. (1941) A new crocodilian from the Lance Formation. American Museum Novitates, 1128, 1-5. Pereda-Suberbiola, X. (2009) Biogeographical affinities of Late Cretaceous continental tetrapods of Europe: a review. Bulletin de la Société Géologique de France, 180, 57-71. Prieto-Marquez, A. & Wagner, J.R. (2009) Pararhabdodon isonensis and Tsintaosaurus spinorhinus: a new clade of lambeosaurine hadrosaurids from Eurasia. Cretaceous Research, 30, 1,238-1,246. Rage, J.-C. (1984) Serpentes. Handbuch Der Paläoherpetologie (ed. by P. Wellnhofer). Gustav Fischer Verlag, Stuttgart, New York. Rich, T.H., Vickers-Rich, P. & Gangloff, R.A. (2002) Polar dinosaurs. Science, 295, 979-980. Russell, D.A. & Manabe, M. (2002) Synopsis of the Hell Creek (uppermost Cretaceous) dinosaur assemblage. The Hell Creek Formation and the Cretaceous–Tertiary boundary in the northern Great Plains: an integrated continental record of the end of the Cretaceous (ed. by J.H. Hartman, K.R. Johnson and D.J. Nichols), pp. 169-176. Geological Society of America Special Paper, 361. Boulder, CO. Sampson, S.D, Loewen, M.A, Farke, A.A., Roberts, E.M., Forster, C.A., et al. (2010) New horned dinosaurs from Utah provide evidence for intracontinental dinosaur endemism. PLoS ONE, 5, e12292. Van Itterbeeck, J., Bolotsky, Y., Bultynck, P. & Godefroit, P. (2005) Stratigraphy, sedimentology and palaeoecology of the dinosaur-bearing Kundur Section (Zeya-Bureya Basin, Amur Region, Far Eastern Russia). Geological Magazine, 142, 735-750. Vavrek, M.J. & Larsson, H.C.E. (2010) Low beta diversity of Maastrichtian dinosaurs of North America. Proceedings of the National Academy of Science of the USA, 107, 8,265-8,268. Weigelt, J. (1940) The first Paleocene mammalian fauna in Germany. Research and Progress, 6, 117-122. Xu, X., Wang, D.-Y., Sullivan, C., Hone, D.W.E., Han, F.-L., Yan, R.-H. & Du, F.M. (2010) A basal parvicursorine (Theropoda: Alvarezsauridae) from the Upper Cretaceous of China. Zootaxa, 2413, 1-19. Zarski, M., Jakubowski, G. & Gawor-Biedowa, E. (1998) The first Polish find of lower Paleocene crocodile Thoracosaurus Leidy, 1852: geological and palaeontological description. Geological Quarterly, 42, 141-160.