An emerging palaeoceanographic ‘missing link’: multidisciplinary study of rarely recovered parts of deep-sea Santonian–Campanian transition from Shatsky Rise

The Cretaceous deep-sea record of the Santonian–Campanian transition is commonly interrupted by an extensive unconformity (representing <10 Myr of hiatus). The resultant palaeoceanographic gap can now be partly bridged by a recent short core of pelagic ooze from Shatsky Rise (Integrated Ocean Drilling Program (IODP) Site U1348), with precise multidisciplinary age constraints developed herein. New oxygen isotope data from very well-preserved benthic foraminifera, together with accurately compiled comparable benthic data from previous Pacific deep-sea sections, exhibit a large (c. +1‰) early Campanian shift. We propose the Santonian–Campanian climatic transition was not gradual but was the first major cooling step after sustained mid-Cretaceous hothouse conditions. Supplementary material: Detailed analytical methods including biostratigraphic notes and Sr isotopic chronology, supplementary figures (locality map; additional geochemical, isotopic and micropalaeontological results; palaeomagnetic results; global Sr isotope compilation and age model; benthic foraminiferal stable isotope compilation), tables of microfossil occurrences and numerical data are available at http://www.geolsoc.org.uk/SUP18598.

and micropalaeontological results; palaeomagnetic results; global Sr isotope compilation and age model; benthic foraminiferal stable isotope compilation), tables of microfossil occurrences and numerical data are available at http://www.geolsoc.org.uk/ SUP18598.
Unconformities in the pelagic sedimentary record are a major obstacle in reconstructing palaeoceanographic histories, yet they also can be robust physical evidence of major shifts in past deep-water properties. Hence, unconformities can convey crucial palaeoceanographic information as long as their causal mechanism (i.e. hiatus) is reasonably explained by the data from coeval stratigraphically complete sections. However, if the spatiotemporal extent of hiatus was so extensive that a certain age of stratigraphic interval was totally erased from the global pelagic sedimentary record, it becomes impossible to reconstruct any changes in the deep-sea environment at that time. Although generally overlooked, this situation has been a serious problem in Late Cretaceous palaeoceanography.
A pronounced break exists in the deep-sea sedimentary record at the Santonian-Campanian (S-C) transition that cannot be ascribed simply to technical artefacts of drilling. Sliter (1992,1995) summarized the chronostratigraphic integrity of Cretaceous Deep Sea Drilling Project (DSDP)-Ocean Drilling Program (ODP) sites in the Pacific basin based on planktonic foraminifera and illustrated the widespread hiatus around the Santonian/Campanian (S/C) boundary (c. 83.5 Ma). Recent Sr isotopic evidence from DSDP Site 463 showed that the S-C hiatus in the central Pacific potentially lasted up to 10 Myr (Ando et al. 2009). Huber (1992), using calcareous microfossil occurrences of multiple southern high-latitude DSDP-ODP sites, documented a 'Southern Ocean hiatus' more or less coincident with the Pacific S-C hiatus. Furthermore, the imperfect nature of the S-C sedimentary record has been highlighted from several North Atlantic deep-sea sites (Huber et al. 2002;MacLeod et al. 2011;Robinson & Vance 2012).
One consequence of such ubiquitous S-C unconformities is a significant data gap in global benthic foraminiferal oxygen isotope (δ 18 O) compilations (e.g. Cramer et al. 2009). Accordingly, the timing and tempo of S-C palaeoclimatic evolution, and its relationship to palaeoceanographic and biotic changes, remain unknown. The latest benthic δ 18 O compilation, by Friedrich et al. (2012), shows no discernible gap at the S-C transition, but this reconstruction should be viewed with caution because of uncertainties in their age-models, and their subjective treatment of benthic foraminiferal vital effects (see discussion below). It should be noted that the δ 18 O profile of excellently preserved foraminifera at South Atlantic DSDP Site 511 was often depicted to be stratigraphically complete across the S/C boundary; however, examination of the available palaeomagnetic record suggests that deposition at this site predated the S/C boundary (see Huber et al. 2002). Sites with a complete S-C transition have been reported from off NW Australia (ODP Site 762) and on land in the Mediterranean Tethys (e.g. Petrizzo et al. 2011), but the available materials are diagenetically affected. T he Cretaceous deep-sea record of the Santonian-Campanian transition is commonly interrupted by an extensive unconformity (representing <10 Myr of hiatus). The resultant palaeoceanographic gap can now be partly bridged by a recent short core of pelagic ooze from Shatsky Rise (Integrated Ocean Drilling Program (IODP) Site U1348), with precise multidisciplinary age constraints developed herein. New oxygen isotope data from very well-preserved benthic foraminifera, together with accurately compiled comparable benthic data from previous Pacific deep-sea sections, exhibit a large (c. +1‰) early Campanian shift. We propose the Santonian-Campanian climatic transition was not gradual but was the first major cooling step after sustained mid-Cretaceous hothouse conditions. Supplementary material: Detailed analytical methods including biostratigraphic notes and Sr isotopic chronology, supplementary figures (locality map; additional geochemical, isotopic (Aptian-Campanian) sediment cover of pelagic carbonates was found to be unconsolidated and thus described as 'ooze', and to contain varying amounts of chert (Expedition 324 Scientists 2010). At 80 Ma, this site was situated around 2500-3000 m palaeowater-depth assuming normal lithospheric subsidence.

IODP
Site U1348-Core 2R contains 1.4 m of monotonous pale yellow nannofossil ooze, from which very well-preserved matrixfree foraminiferal specimens are readily isolated by gentle spray washing only. Two samples studied onboard yielded a typical Santonian planktonic foraminiferal assemblage at the bottom of this core and a Campanian assemblage at the top (Expedition 324 Scientists 2010). For the post-cruise study, samples were taken at high resolution (10 cm spacing) and analysed to generate micropalaeontological, geochemical and palaeomagnetic data.

Multidisciplinary chronological results.
Detailed examination of well-preserved planktonic foraminifera (often retaining delicate umbilical features) confirms the Santonian to Campanian age of Site U1348-Core 2R, in which a drastic assemblage compositional change is recognized across the boundary between Sections CC and 1 (85.3 m below seafloor (mbsf)) ( Fig. 1). U1348-2R-CC is in the Dicarinella asymetrica Zone, yielding such Santonian representatives as D. asymetrica and Sigalia rugocostata with species of Marginotruncana. The age is further narrowed to the late Santonian based on the presence of Ventilabrella eggeri and Hendersonites carinatus (e.g. Nederbragt 1991). In contrast, the assemblage in U1348-2R-1 contains typical Campanian taxa of Globotruncana and Globotruncanita, both of which are high in abundance, large-sized, and diverse. Biostratigraphic subdivision of U1348-2R-1 can be facilitated by the species of Globotruncanita (in the absence of Radotruncana calcarata), such that the lower part is marked by the common occurrence of G'ta elevata (G'ta elevata Zone (or Contusotruncana plummerae Zone, sensu Petrizzo et al. 2011); early to middle Campanian), whereas the upper part represents the co-occurrence of G'ta stuarti and G'ta subspinosa (late Campanian). Nannofossils from this interval represent no marked changes in the major assemblage composition and preservation state (with common etching and/or fragmentation). Some age-diagnostic taxa allow for zonal assignments (CC14-CC24), which are consistent with the planktonic foraminiferal age.
With the primary microfossil ages constrained, Sr isotopic and palaeomagnetic data serve as precise means of chronology. 87 Sr/ 86 Sr ranges from 0.70747 to 0.70770, with values exhibiting a stepwise trend with three near-invariant 87 Sr/ 86 Sr segments, here called intervals (i)-(iii) (Fig. 1). Jumps in 87 Sr/ 86 Sr at 85.25 and 84.95 mbsf occur at the same levels as the planktonic faunal changes and thus translate to unconformities. Palaeomagnetic results, from which polarity interpretation can reliably be made, indicate that U1348-2R-CC has normal polarity and is probably from Chron C34n (albeit only a single data point), whereas U1348-2R-1 has reversed polarity throughout and should be within Campanian Subchrons C33r and/or C32r2r. Figure 2 shows the probable age ranges of U1348-Core 2R defined with respect to the standard time scale combined with the global seawater Sr isotope curve. The demonstrated coincidence of independent bio-, chemo-and magnetostratigraphic constraints facilitates the numerical age assignments to each of intervals (i)-(iii) with minimal uncertainties. Oxygen isotopic results. The δ 18 O trends for both bulk carbonates and benthic foraminifera document a large +1.0‰ shift over the examined interval (Fig. 1). Benthic foraminiferal specimens are very well preserved with dully translucent 'pearly' tests ( Fig. 1), showing fairly minor surface dissolution and/or recrystallization. The sedimentary CaCO 3 contents are very high (93-97 wt%); in such a case recrystallization is known to be often significant (e.g. Pearson et al. 2007), but this is not the case for U1348 benthic foraminifera.
Of the six benthic taxa selected for analysis, Aragonia is the only group available from almost all samples. This taxon has not been widely used in palaeoceanographic studies, but we document that the Aragonia δ 18 O values are in close agreement with those of Oridorsalis (Fig. 1), the latter being a reliable taxon for early Cenozoic palaeoceanographic analysis (Katz et al. 2003). The relatively low δ 13 C values in Aragonia and Oridorsalis suggest infaunal habitats, which are in line with previous knowledge for the latter taxon (Katz et al. 2003). These facts provide a basis for the use of the Aragonia-Oridorsalis δ 18 O profile as a faithful δ 18 O recorder, because infaunal taxa are subject to slow growth rates in longer life cycles, resulting in δ 18 O equilibrium precipitation of tests (e.g. Friedrich et al. 2006). Three other groups with higher δ 13 C values, including a widely used taxon Nuttallides, are generally interpreted as living epifaunally. Their negative δ 18 O offsets relative to the Aragonia-Oridorsalis δ 18 O profile, an expression of the isotopedisequilibrium vital effects in epifauna characterized by higher metabolic rates (e.g. Friedrich et al. 2006), are compensated for by a correction factor of +0.4‰, which is a reasonable value when compared with the case of a Campanian-Maastrichtian assemblage (Friedrich et al. 2006).

Discussion.
Site U1348-Core 2R (albeit limited stratigraphically) represents the first, superior sedimentary record of the deep-sea S-C transition, with a robust chronology, very good preservation, and a fully open-ocean setting. Most importantly, now we have new definitive control points in the Late Cretaceous deep-sea δ 18 O evolution. Figure 3 shows an updated benthic foraminiferal δ 18 O compilation for central Pacific DSDP-IODP Sites 305, 463 and U1348, for which special attention is given to objective age modelling by means of Sr isotope stratigraphy, proper data corrections for disequilibrium δ 18 O precipitation, and additional corrections for the possible effect of inter-site waterdepth differences.
Many of new U1348 δ 18 O data fall within the c. 7 Myr S-C gap evident in the pre-existing datasets. It appears that the S-C transition marked a major step in δ 18 O (Fig. 3b), when the baseline shifted to 0‰ or greater after c. 25 Myr of universally negative benthic δ 18 O values since the Albian (e.g. Friedrich et al. 2012). In other words, the mid-Cretaceous hothouse persisted to the latest Santonian, and then switched to cool greenhouse (Huber et al. 2002) during the early Campanian. The exact timing and pacing of this climatic transition remain unresolvable, but the shift should be within the first 3-4 Myr of the Campanian. Friedrich et al. (2012) presented a comprehensive benthic δ 18 O compilation showing a gradual trend across the S-C transition. However, their age-model information on the deep-sea sites used for δ 18 O compilation (Friedrich et al. 2012, table DR1) is literature Fig. 2. Graphic summary of probable age ranges of U1348-Core 2R sediments against GTS2004 (Ogg et al. 2004) combined with the standard Sr isotope curve (modified in this study). Superimposed on 87 Sr/ 86 Sr data are a fifth-order polynomial (90-65 Ma) and its confidence limits (= ±1 standard deviation of residuals; grey curve). The Contusotruncana plummerae Zone is adopted from Petrizzo et al. (2011), but its originally defined base (open arrowhead) would require further examination for inter-site diachroneity as discussed elsewhere.  Barrera & Savin (1999);F, Friedrich et al. (2012); L, Li & Keller (1999). Numerical age models for Sites 305 and 463 are based on published Sr isotope data (Barrera & Savin 1999;Mearon et al. 2003;Ando et al. 2009) calibrated against the standard curve (Fig. 2). Site U1348 δ 18 O data are plotted to fill up the 'probable ranges' in Figure 2. Open arrow highlights slight discrepancy in δ 18 O between Site 463 (mid-bathyal; Barrera & Savin 1999) and Site U1348 (2500-3000 m water depth), most probably owing to small bottom-water temperature differences. (b) Same dataset as for (a) but an ad hoc correction factor of +0.3‰ (= c. 1.5 °C) is applied to all Site 463 data to accommodate probable water depth-related δ 18 O offset relative to Site U1348. citations only, without a list of datum events adopted. With specific reference to Sites 305 and 463, it is unclear how Friedrich et al. (2012) could plot the δ 18 O data so evenly through the Santonianmid-Campanian interval. Furthermore, despite the presence of a systematic δ 18 O-δ 13 C offset between the taxa used (Globorotalites v. Praebulimina), those researchers did not make the isotope disequilibrium correction, which is another technical artefact that resulted in somewhat scattered, gradual S-C δ 18 O compilation. Although Friedrich et al. (2012) is commended for adding numerous new data points to the Late Cretaceous δ 18 O database, we conclude that a stepwise δ 18 O evolution for the S-C transition is reasonable and robustly supported by data, especially considering the Sr isotopic inter-site chronology developed herein.
Implications for global change study. The S-C cooling step, as illustrated by our benthic δ 18 O compilation, may be linked to an early Campanian reorganization of global ocean circulation in response to changing palaeoclimates and/or palaeogeography, as suggested on the basis of Nd isotopic data, albeit with lingering age uncertainty (e.g. MacLeod et al. 2011;Robinson & Vance 2012). Now this cooling episode also seems broadly coincident with the resumption of magnetic reversals after 'superchron' C34n (Fig. 3). These observations may elevate the significance of study of the S-C transition to a broader discipline of global change.
In this regard, it should be noted that the U1348 δ 18 O data from bulk carbonates (predominantly coccoliths) exhibit a +1.2‰ shift, paralleling the benthic δ 18 O trend (Fig. 1). The observed bulk δ 18 O variation cannot be explained by the changing nannoflora or a varying extent of early recrystallization (Pearson et al. 2007), because the nannofossil assemblage, coccolith preservation, and lithology are all uniform through the examined interval. If the δ 18 O shift translates to a genuine palaeotemperature signal (in the absence of local palaeoceanographic control), a cooling by 5-6 °C is required (e.g. Ando et al. 2010), although (sub)tropical sea-surface temperatures were probably less sensitive to greenhouse forcings (e.g. Otto-Bliesner et al. 2002). Although speculative, Antarctic glaciation and a resultant whole-ocean δ 18 O shift might have played a role, at least in part, in the observed surface-bottom δ 18 O co-variation (e.g. Miller et al. 2005).
Whatever the likelihood of glaciation, the S-C transition is worthy of more attention as a critical period of global change during the Cretaceous greenhouse. Further scrutiny of available S-C deepsea and continental sedimentary records is necessary in terms of chronology, δ 18 O variation and/or sea-level change. Linking all of the palaeoenvironmental observations with a precise chronology in future work would allow the construction of a unified view on the causes and consequences of one of Earth's major climatic shifts.