Middle Jurassic-Early Cretaceous high-latitude sea-surface temperatures from the Southern Ocean

a of the there is a range of isotopic and biotic evidence that favours the concept of discrete “cold snaps”, marked particularly by migration of certain towards lower Extension of the use of the palaeotemperature proxy TEX 86 back to the middle Jurassic indicates that relatively warm sea-surface conditions (26–30 ◦ C) existed from this interval ( ∼ 160 Ma) to the Early Cretaceous ( ∼ 115 Ma) in the Southern Ocean. The Jurassic and Cretaceous “cold snaps” represent falls of only a few degrees. Belemnite δ 18 O 10 data give palaeotemperatures that are consistently lower by ∼ 14 ◦ C than does TEX 86 and these molluscs likely record conditions below the thermocline. Such long-term warm climatic conditions would only be compatible with the existence of continental ice if appreciable areas of high altitude existed on Antarctica, and/or in other polar regions, during the Mesozoic Era.


Introduction
In order to understand Jurassic and Cretaceous climate, the reconstruction of seasurface temperatures at high latitudes, and their variation over different time scales, is of paramount importance. A basic division of Phanerozoic climatic modes into "icehouse" and "greenhouse" periods is now commonplace (Fischer, 1981). However, a Bice et al., 2003;Jenkyns, 2003;Moriya et al., 2007;Dera et al., 2011;Littler et al., 2011). Evidence for cool climates comes from problematic oxygen-isotope data from well-preserved foraminifera from one Upper Cretaceous Atlantic ODP site (Bornemann et al., 2008), changes in nannofossil assemblages from both low and high latitudes (Mutterlose et al., 2009) and putative Cretaceous glacial deposits and so-called glen-5 donites that formed from the cold-temperature hydrated form of calcium carbonate, ikaite (Kemper, 1987;Frakes and Francis, 1988;de Lurio and Frakes, 1999;Alley and Frakes, 2003). The presence of certain plants, fish species and fossil reptiles, however, rather points towards much warmer polar climates, at least at low altitudes (Nathorst, 1911;Tarduno et al., 1998;Friedman et al., 2003;Vandermark et al., 2007), as do 10 oxygen-isotope values of benthonic and planktonic foraminifera (Huber et al., 1995;Bice et al., 2003). Indeed, Moriya et al. (2007) could find no oxygen-isotope evidence for glaciation during the mid-Cenomanian, an interval suggested to have witnessed glacio-eustatic changes in sea level by Gale et al. (2002).
The organic geochemical proxy TEX 86 ("tetraether index of 86 carbon atoms") of-15 fers the advantage of giving estimates of sea-surface temperatures and is applicable to those sediments lacking in carbonate that contain sufficient quantities of immature organic matter (Schouten et al., 2002(Schouten et al., , 2003Kim et al., 2010). TEX 86 data from Aptian and Albian organic-rich sediments suggest low-latitude temperatures in the Atlantic and Pacific Ocean in the range 31-36 • C (Schouten et al., 2003: Forster et al., 2007Du-20 mitrescu et al., 2006;recalibrated after Kim et al., 2010). Upper Berriasian to lower Barremian organic-rich sediments from the peri-equatorial Atlantic Ocean give similar mid-30 high-latitude palaeotemperatures in the Southern Hemisphere is here elucidated by analysing DSDP/ODP sediments retrieved from close to Antarctica (Fig. 1). This report extends the application of the TEX 86 palaeothermometer back into the Callovian (middle Jurassic), the oldest sediments from the World Ocean yet analysed for this proxy.

2 Methods
Powdered and freeze-dried sediments (1-3 g dry mass) were extracted with dichloromethane (DCM)/methanol (2:1) by using the Dionex accelerated solvent extraction technique. The extracts were separated by Al 2 O 3 column chromatography using hexane/DCM (9:1), DCM/methanol (95:5) and DCM/methanol (1:1) as subsequent eluents to yield the apolar, tetraether and polar fractions, respectively. The apolar and desulphurized (using Raney Ni) polar fractions were analysed by gas chromatography and gas chromatography/mass spectrometry. The polar fractions were analysed for GDGTs as described in Schouten et al. (2007) (Ludwig and Krasheninikov et al., 1983;Barker and Kennett et al., 1990). The section drilled on the Falkland Plateau is unusual in that it offers a Middle Jurassic-Lower Cretaceous hemipelagic sedimentary section of black locally 10 laminated organic-rich shale and mudstone, ∼140 m in thickness, locally containing a rich macrofauna of belemnites, ammonites and bivalves Jeletsky, 1983). The section drilled on the Antarctic slope is represented by ∼70 m of Lower Cretaceous hemipelagic black organic-rich silty mudstone (Fig. 2: O'Connell, 1990). The biostratigraphy of the high-latitude Cretaceous sediments is not unambiguous 15 because the ranges of critical taxa are imperfectly known and certain key stage boundaries are not yet rigorously defined. The organic-rich section of ODP Site 693 ( Fig. 2) has yielded planktonic foraminifera of probable late Aptian age (Leckie, 1990); nannofossil data suggest the presence of the uppermost Aptian to lowermost Albian interval, with a stage boundary tentatively fixed at around 453 mbsf (Mutterlose et al., 2009).

20
For Site 511, nannofossil biostratigraphy suggests the presence of the uppermost Callovian, Oxfordian and Kimmeridgian-Tithonian stages, an interpretation that is broadly supported by biostratigraphic determinations of molluscan faunas (Jeletsky, 1983) and strontium-isotope ratios from belemnites that, when compared with the global reference curve, suggest the presence of all four stages (Price and   . Whether this putative hiatus is a function of non-deposition or due to large-scale removal of sediment by slumping is unresolved. However, strontium-isotope ratios give values that suggest that the Hauterivian and possibly the Valanginian are represented at this site, at least by those belemnites yielding age-significant geochemical data (Price and 5 Gröcke, 2002). The Barremian and Aptian intervals are recognized by characteristic planktonic foraminiferal faunas ) and nannofossil data have been used to fix the boundary between the stages at ∼555 mbsf (Bralower et al., 1994), an age-assignment that is at odds with that derived from strontium-isotope dating that indicates a Hauterivian-Barremian age as high as 524 mbsf in the core (Price and Gröcke, 2002). The boundary between the lower and upper Aptian is fixed at 508-513 mbsf on the basis of nannofossil and ammonite biostratigraphy (Jeletsky, 1983;Bralower et al., 1993). However, planktonic foraminiferal faunas fix the boundary between the Aptian and Albian stages at ∼486 mbsf (Huber et al., 1995), although some authors, using nannofossil dating, have put the contact lower in the section, between 15 500 and 510 mbsf Bralower et al., 1993). A generalized stratigraphy, utilizing available biostratigraphic and Sr-isotope data, is utilized in Fig. 3.

TOC and organic carbon-isotope curves from ODP Site 693A and DSDP Site 511
The total organic-carbon (TOC) curve from Site 693A in the Weddell Sea is unremark-20 able, indicating values generally lower than 1.5% over the interval analysed and, apart from peak values at ∼456 mbsf, shows a decreasing trend towards the top of the interval (Fig. 2). TOC values for the Lower Cretaceous dark shales and mudstones of this site average ∼2.5% (O'Connell, 1990 (Herrle et al., 2004). TOC values for Site 551 are typically in the 2-6% range for the majority of samples 5 over the uppermost Jurassic-lowest Cretaceous interval (Fig. 3), dropping abruptly to values close to zero around the boundary of the lower and upper Aptian; the organic matter has a relatively high hydrogen index (200-600 mg hydrocarbons per g organic carbon), indicating that it is dominantly marine in nature (Deroo et al., 1983). Over the same Mesozoic interval, δ 13 C org values are typically in the range −30 to −28‰, rising 10 into an irregular positive excursion (values mostly between −16 and −22‰) close to the lower-upper Aptian boundary, as fixed biostratigraphically, with a peak value of −18.5‰. By comparison with European sections in Italy and Switzerland, this isotopic signature is characteristic of the middle part of the Aptian stage where a positive shift in δ 13 C org of 6-7‰ is observed (Menegatti et al., 1998). Biostratigraphy and carbon-15 isotope stratigraphy are hence in agreement.

Middle Jurassic-Early Cretaceous marine sea-surface temperatures in the Southern Ocean
TEX 86 -derived sea-surface temperatures (formula in Kim et al., 2010) for the continental slope off Antarctica (Site 693A), around Aptian-Albian boundary time, fall in the 20 range 24-28 • C and suggest a warming trend into the early Albian (Fig. 2).
The data from the Falkland Plateau (Site 511) give the first TEX 86 palaeotemperature record from the Jurassic and suggest values in the range 26-30 • C, with an overall warming trend, for the latter part of this Period (Fig. 3). Such a general warming trend fits with the overall decline in oxygen-isotope ratios in Upper Jurassic belemnites from 25 Europe and Russia (Jenkyns et al., 2002). Conversely, the Cretaceous section, over the Hauterivian-Aptian interval, shows an overall cooling trend over a closely similar 1345  et al., 1994), similarly arguing for the fact that belemnites do not record sea-surface or even mixed-layer temperatures, despite their long-term application to marine Mesozoic palaeoclimatological studies (Urey et al., 1951;Lowenstam and Epstein, 1954 Moriya et al. (2003), for Cretaceous ammonites, based on their assumed 15 nektobenthonic ecology, but is in line with the relative depth habitat inferred for these two types of cephalopod, with belemnites typically inhabiting deeper, colder water (Anderson et al., 1994). Because oxygen-isotope values from planktonic foraminifera are typically reset by recrystallization on the sea floor, hence producing spuriously low temperatures (Pear-20 son et al., 2001), benthonic foraminiferal records are potentially more reliable indices of ambient conditions. Basal Albian benthic foraminifera from Site 511 suggest subthermocline temperatures of ∼13 • C (Fassell and Bralower, 1999), in line with reconstructed belemnite palaeotemperatures established in the Barremian-Aptian part of the core (Fig. 3)

The early Aptian Oceanic Anoxic Event on the Falkland Plateau
The early Aptian Oceanic Anoxic Event (OAE1a or Selli Event), defining a period of unusually widespread oxygen-depleted waters accompanied by widespread deposition of black organic-rich shales, has been recorded in all major ocean basins (Schlanger and Jenkyns, 1976;Arthur et al., 1990;Jenkyns, 2003Jenkyns, , 2010. The record of the OAE has 5 been identified at Site 511 on the Falkland Plateau on lithological and biostratigraphic grounds (Bralower et al., 1994). A defining characteristic of this OAE is the presence of a negative carbon-isotope excursion followed by a positive excursion that extends into the late Aptian (Menegatti et al., 1998). Typically, the positive δ 13 C excursion extends stratigraphically well above the most organic-rich horizon. The suggested level 10 on the Falkland Plateau that records the early Aptian OAE, as fixed by TOC (note the dramatic drop in values passing up-section) and δ 13 C org stratigraphy (Fig. 3), is illustrated in Fig. 4. The reconstructed palaeotemperatures at this site range from 26 to 29 • C during the OAE and indicate a drop of ∼3 • C at the level where maximum TOC is recorded. Although the early Aptian OAE represents an interval of relative warmth and continental weathering during the OAE (Jenkyns, 2011). Given that the Shatsky Rise occupied a peri-equatorial position during the early Aptian, the Equator-to-pole sea-surface temperature gradient during the OAE was probably close to 10 • C. A drop in temperature of ∼3 • C during the early phase of this event has been determined for a mid-latitude site (southern France), based on oxygen-isotope data from well-preserved Introduction

Evidence for Jurassic-Cretaceous "cold snaps" in the Southern Ocean
The Callovian-Oxfordian boundary interval has been identified in Europe as a relatively cool interval, based on a number of independent criteria. Oxygen-isotope data from English and Russian belemnites indicate a drop in temperature commencing in the latest Callovian (Jenkyns et al., 2002), as do sharks' teeth from England, France and Switzerland (Lécuyer et al., 2003). Accompanying the proposed drop in temperature (∼7 • C in northern hemisphere mid-latitudes from shark-teeth data), there is evidence for simultaneous invasion of boreal ammonite species into lower latitude zones (Dromart et al., 2003). Because regional facies analysis suggests sea-level fall across the stage boundary, it has been suggested that this interval records build-up of continental 10 polar ice (Dromart et al., 2003). The TEX 86 palaeotemperature data from the Falkland Plateau (Fig. 3) indicate an observed minimum of ∼26-27 • C around Callovian-Oxfordian boundary time, followed by a ∼2 • C rise. Because the lowest value lies close to the base of the cored section, neither an absolute minimum nor an absolute rise in temperature can be determined, but clearly the proposed "cold snap" was no more 15 than a minor drop in temperature against a background of extreme polar warmth and fits ill with the notion of sea-level glaciation at this time.
Cold snaps in the Valanginian and around the Aptian-Albian boundary have been proposed on the basis of the presence of glendonites (pseudomorphs of the cooltemperature form of hydrated calcium carbonate, ikaite) in sediments of this age crop-20 ping out in the Sverdrup Basin of Arctic Canada and Svalbard (Kemper, 1987;Price and Nunn, 2010). Glendonites are also reported from Upper Aptian shales in the Eromanga Basin of Australia (Frakes and Francis, 1988;de Lurio and Frakes, 1999). These occurrences are associated with centimetre-scale clasts that have been interpreted as ice-rafted but which could equally well be tree-rafted (Bennett and Doyle, 25 1996). Ikaite typically forms at temperatures no greater than ∼7 • C, although it may be stabilized at higher temperatures in phosphate-rich interstitial waters such as characterize organic-rich sediments (de Lurio and Frakes, 1999). As an early diagenetic product growing by displacement within sediment, however, it clearly offers little in the way of palaeotemperature data for the sea surface as it forms in water depths below the mixed layer. TEX 86 data from the Valanginian of Site 766 (Fig. 1) give sea-surface temperatures consistently in the 25-26 • C range (Littler et al., 2011).
Nannofossil data from both low-and high-latitude sites around the Aptian-Albian 5 boundary show a decline in Tethyan taxa and invasion of more boreal forms, indicative of cooling, and diatoms also appeared in high-latitude sites in both northern and southern hemispheres during this interval (Mutterlose et al., 2010). The cooling trends are indisputable: what is at issue is the absolute value of temperature changes across the Aptian-Albian stage boundary. The TEX 86 data from the Falkland Plateau site suggest a drop in sea-surface temperature of 1.5-2 • C in the latest Aptian, falling to values of ∼27.5 • C (Fig. 3). However, the entirety of the excursion may not have been captured by the TEX 86 profile because the use of this proxy is precluded by the lack of organic-rich black shales extending into the Albian. The Valanginian stage cannot be positively identified in the Cretaceous section from the Falkland Plateau but there is no 15 suggestion of a major drop in temperature in the Lower Cretaceous section of the core, in agreement with data from the Exmouth Plateau, off Australia (Littler et al., 2011).
In conclusion, although accumulation of ice at high altitude on Antarctica, or other polar regions, cannot be ruled out, these reconstructed warm high-latitude palaeotemperatures are difficult to reconcile with the notion of transient "icehouse" interludes for 20 a period extending over ∼40 million years (Middle Jurassic to Early Cretaceous). Such warm palaeotemperatures further imply that mechanisms other than glacio-eustasy must be entertained as explanations for regional changes in sea level during the Mesozoic Era.