Paleoceanography

Preservation of benthic foraminifera and reliability of deep-sea temperature records: Importance of sedimentation rates, lithology, and the need to examine test wall structure

Authors


Abstract

[1] Preservation of planktic foraminiferal calcite has received widespread attention in recent years, but the taphonomy of benthic foraminiferal calcite and its influence on the deep-sea palaeotemperature record have gone comparatively unreported. Numerical modeling indicates that the carbonate recrystallization histories of deep-sea sections are dominated by events in their early burial history, meaning that the degree of exchange between sediments and pore fluids during the early postburial phase holds the key to determining the palaeotemperature significance of diagenetic alteration of benthic foraminifera. Postburial sedimentation rate and lithology are likely to be important determinants of the paleoceanographic significance of this sediment–pore fluid interaction. Here we report an investigation of the impact of extreme change in sedimentation rate (a prolonged and widespread Upper Cretaceous hiatus in the North Atlantic Ocean) on the preservation and δ18O of benthic foraminifera of Middle Cretaceous age (nannofossil zone NC10, uppermost Albian/lowermost Cenomanian, ∼99 Ma ago) from multiple drill sites. At sites where this hiatus immediately overlies NC10, benthic foraminifera appear to display at least moderate preservation of the whole test. However, on closer inspection, these tests are shown to be extremely poorly preserved internally and yield δ18O values substantially higher than those from contemporaneous better preserved benthic foraminifera at sites without an immediately overlying hiatus. These high δ18O values are interpreted to indicate alteration close to the seafloor in cooler waters during the Late Cretaceous hiatus. Intersite differences in lithology modulate the diagenetic impact of this extreme change in sedimentation rate. Our results highlight the importance of thorough examination of benthic foraminiferal wall structures and lend support to the view that sedimentation rate and lithology are key factors controlling the paleoceanographic significance of diagenetic alteration of biogenic carbonates.

1. Introduction

1.1. Foraminiferal Preservation: Planktics Versus Benthics

[2] The preservation of planktic foraminiferal calcite has received widespread attention in the past decade [Schrag et al., 1995; Wilson and Opdyke, 1996; Norris and Wilson, 1998; Crowley and Zachos, 2000; Wilson and Norris, 2001; Pearson et al., 2001; Norris et al., 2002; Wilson et al., 2002; Zachos et al., 2002; Sexton et al., 2006a, 2006b; Moriya et al., 2007; Pearson et al., 2007]. This attention is primarily a consequence of attempts to accurately reconstruct sea surface temperatures (SSTs), especially for past warm climates (e.g., the early through middle Eocene and mid-Cretaceous). Partly on the basis of oxygen isotope ratios (δ18O) of planktic foraminiferal calcite, we know that these intervals of geologic time are characterized by high-latitude temperatures that are consistently much warmer than those seen today [Stott et al., 1990; Barrera and Huber, 1991; Zachos et al., 1994; Huber et al., 1995]. Yet for many years, foraminiferal δ18O-derived SST estimates from the contemporaneous tropical oceans were thought to have been no warmer or even cooler than the modern tropics [Douglas and Savin, 1973, 1975; Savin, 1977; Shackleton and Boersma, 1981; Boersma et al., 1987; Crowley, 1991; Barrera, 1994; Zachos et al., 1994; Bralower et al., 1995; Price et al., 1998]. These cool δ18O-derived tropical SSTs were at odds with the warmer-than-modern tropical SSTs predicted in numerical modeling experiments [e.g., Manabe and Bryan, 1985; Sloan and Rea, 1996; Bush and Philander, 1997; Huber and Sloan, 2000, 2001; Poulsen et al., 2001]. This so-called “Cool Tropic Paradox” [D'Hondt and Arthur, 1996] has subsequently been explained as an artifact of diagenetic alteration of planktic foraminiferal calcite near the seafloor at tropical sites (where the vertical temperature gradient in the ocean is strong) yielding artificially cool δ18O-derived SSTs [Schrag et al., 1995; Wilson and Opdyke, 1996; Norris and Wilson, 1998; Wilson and Norris, 2001; Pearson et al., 2001; Wilson et al., 2002; Sexton et al., 2006a].

[3] In comparison to planktics, preservation of benthic foraminifera has received little attention [see Corliss and Honjo, 1981; Widmark and Malmgren, 1988; Murray, 1989; McCorkle et al., 1995; Carman and Keigwin, 2004]. This situation is probably traceable in part to (1) the greater robustness of more heavily calcified benthic tests in comparison to planktics and (2) the realization that diagenetic alteration of benthics will, except at high-latitude sites of deep convection, tend to proceed in waters closer in temperature and composition to the waters in which the tests were secreted than is the case for planktics. Numerical modeling indicates that the carbonate recrystallization histories of deep-sea sections are dominated by events in their early burial history [Rudnicki et al., 2001]. First, in line with the view that clay-rich sediments enhance the preservation of planktic foraminifera [Norris and Wilson, 1998; Wilson and Norris, 2001; Pearson et al., 2001; Wilson et al., 2002; Sexton et al., 2006a], lithological composition should also be an important factor governing the alteration of benthics. Second, postburial sedimentation rate, as well as being fundamental to controlling the impact of diagenetic alteration of bulk carbonate [Schrag et al., 1995], will also likely be important to determining the palaeotemperature significance of alteration of benthic foraminifera. However, because numerical diagenetic models rely on pore fluid chemistry profiles that are controlled by the diagenesis of bulk carbonate (mainly calcareous nannofossils), these models cannot be used to evaluate the alteration pathway of the individually picked foraminifera that are used in palaeoceanographic studies. Here we investigate the impact of extreme change in sedimentation rate (a prolonged and widespread hiatus) and lithological variability on the preservation and δ18O of benthic foraminifera of Middle Cretaceous age (nannofossil zone NC10, uppermost Albian/lowermost Cenomanian, ∼99 Ma ago (Figure 1a)) from multiple drill sites in the North Atlantic Ocean.

Figure 1.

(a) Magnetobiostratigraphic timescale for the Cretaceous epoch (modified from Shipboard Scientific Party [2004]). Gray shading shows the stratigraphic position of calcareous nannofossil biozone NC10, the target interval for this study. A new astrochronology dates the lower and upper boundaries of biozone NC10 at 100.95 and 99.55 Ma, respectively [Watkins et al., 2005]. (b) Paleogeographic reconstruction for the Albian-Cenomanian boundary (biozone NC10, ∼100 Ma; before equatorial Atlantic gateway opening) showing the location of DSDP and ODP drill sites discussed here. All red circles (solid and open) show sites with a hiatus through the Upper Cretaceous. Solid red circles denote DSDP/ODP sites (numbered on the map) studied here, whereas open red circles (not numbered to avoid clutter) denote those studied by Arthur and Dean [1986]. DSDP site numbers for the open red circles are 101, 105, 144, 384, 386, 387, 390, 391, 392, 417, and 534. Dark gray shading denotes emergent continental fragments. Light gray shading denotes submerged continental shelves. Paleogeographic map is from the Ocean Drilling Stratigraphic Network plate tectonic reconstruction service (http://www.odsn.de/odsn/services/paleomap/paleomap.html).

1.2. Mid-Cretaceous

[4] The mid-Cretaceous represents an important interval in the evolution of Earth's climate. Diverse geological evidence indicates that, during this interval, Earth experienced some of the warmest temperatures of the entire Phanerozoic Era (last ∼545 Ma) [Crowley and North, 1991; Veizer et al., 2000]. This extreme warmth is widely interpreted to result from an increase in greenhouse forcing through elevated levels of atmospheric carbon dioxide [Berner et al., 1983; Barron and Washington, 1985; Schlanger et al., 1981; Larson, 1991; Berner and Kothavala, 2001; Bice and Norris, 2002]. Because of their relevance to climate scenarios for the coming centuries, it is important to understand the forcing mechanisms and dynamic feedbacks operating during these past warm climates. An essential prerequisite to developing this understanding is knowledge of the most basic regional and global features of this extreme warmth (and likewise, for patterns of carbon cycling). We are currently in the early stages of developing this observational database for mid-Cretaceous climates.

2. Materials and Methods

[5] Sediments were disaggregated by soaking in deionized water for 30 min and wet sieving through a 63 μm mesh. The coarse (>63 μm) fraction was dry sieved and benthic foraminifera were picked from the 212 to 350 μm size fraction. Monospecific specimens of either Gyroidinoides infracretacea or Gavelinella sp. were used for stable isotope analyses (paired analyses of these two species within the same sample yield minimal δ18O offsets: G. infracretaceaGavelinella sp. = 0.03‰ [1σ = 0.28‰, n = 33]). Oxygen isotope ratios were analyzed using a Europa Geo 20–20 mass spectrometer equipped with an automatic carbonate preparation system. Between 4 and 10 specimens were analyzed after ultrasonic cleaning in deionized water. δ18O data are reported relative to the Vienna Peedee Belemnite standard (VPDB). Standard external analytical precision, based on replicate analyses of in-house standards calibrated to NBS-19, is ±0.08‰. Scanning electron micrographs were generated using a Leo 1450VP (variable pressure) digital Scanning Electron Microscope (SEM) fitted with a tungsten filament. Prior to SEM analysis, foraminiferal specimens were gold coated. Gold coating optimizes the backscattering of secondary electrons from the sample, providing better topographic imaging.

3. A Widespread Upper Cretaceous Hiatus in the North Atlantic

[6] A summary of results from the first decade and a half of deep-sea drilling revealed a widespread sedimentary hiatus through the Upper Cretaceous across the western part of the North Atlantic [Arthur and Dean, 1986]. This hiatus, with its base typically occurring in the lower Cenomanian, was observed at 11 of the 18 DSDP drill sites examined in this synthesis [Arthur and Dean, 1986] (Figure 1b). Here we extend these earlier observations by examining (1) additional old DSDP sites from the eastern half of the North Atlantic basin and (2) newer drill sites from more recent ocean drilling.

[7] Figure 1 shows the location of all DSDP and ODP drill sites studied here along with those shown to contain an Upper Cretaceous hiatus in the earlier synthesis (all sites with this hiatus are labeled in red). Figure 2 depicts our new evaluation of the timing, duration and stratigraphic position of the Upper Cretaceous hiatus at our studied sites. Figure 2 shows that the base of the hiatus (colored red) occurs at a broadly similar age across sites, toward the top of NC10 around 99 Ma ago, in the lowermost Cenomanian (biozone NC10 colored gray) (ages referred to here are based on updated calcareous microfossil biozonation chronologies [e.g., Erba et al., 1995; Bralower et al., 1997; Burnett, 1999; Premoli Silva and Sliter, 1999; Watkins et al., 2005]). At DSDP Site 137 and ODP Site 1050, the base of the hiatus occurs significantly later (at 92 Ma ago) than at the other sites. The age for the top of the hiatus is more variable, generally falling anywhere from 88 to 77 Ma ago (Coniacian to middle Campanian), meaning that the duration of the hiatus is typically between 11 and 22 Ma ago (notwithstanding its anomalously long duration at Site 545). At Site 137 the hiatus, or at least an extremely condensed horizon, is inferred to occur on the basis of large differences in the age of sediments either side of a short core gap. At Site 398D, the same time interval represented by a hiatus at other sites is here characterized by a transition from carbonate-bearing sediment to carbonate-free clay with very low sedimentation rates. All DSDP and ODP drill sites exhibiting an Upper Cretaceous hiatus are situated in the North Atlantic except for DSDP Site 363 in the northern South Atlantic and Site 511 (with an unusually short hiatus) in the southern South Atlantic (Figures 1b and 2). The hiatus is absent from Site 763B in the Indian Ocean (Figure 2).

Figure 2.

Timing, duration, and stratigraphic position of the Upper Cretaceous hiatus (marked by red horizontal lines) across multiple drill sites. Gray shaded areas define the stratigraphic range of calcareous nannofossil biozone NC10 at each site. Colors of the DSDP/ODP site numbers refer to the stratigraphic position of the hiatus with respect to NC10 (dark blue, a hiatus immediately overlying NC10; light blue, a hiatus just above NC10; orange, no hiatus or its base occurs further up section). Red numbers indicate the ages (in millions of years ago) of the top and the base of the hiatus at each site. Sites that host severely diagenetically altered NC10 benthic foraminifera (see Figure 3) are marked by asterisks. Note that the sites with a hiatus immediately overlying biozone NC10 (dark blue site numbers) also host diagenetically altered NC10 foraminifera. Biostratigraphies for each site are from Bice et al. [2003], Hayes et al. [1972], Shipboard Scientific Party and Bukry [1978], and Shipboard Scientific Party [1978, 1979, 1983, 1984, 1985, 1990, 1998].

4. Taphonomy: Relationship to Sedimentation Rate and Lithology

[8] Figure 3 shows scanning electron micrographs of representative benthic foraminifera from biozone NC10 for each of the drill sites studied. We show images of whole tests and higher-magnification views of test wall cross sections and chamber interiors. Images of whole tests appear to indicate at least moderate preservation of foraminifera at all sites. However, test wall cross sections and chamber interior views reveal a very different picture: benthic foraminifera from those sites with a hiatus immediately overlying the host NC10 strata (site numbers colored dark blue in Figure 2) are extremely poorly preserved (marked by asterisks in Figure 3). These images reveal abundant, large crystals of inorganic calcite growing on interior walls (Sites 398, 545, 550 and 363) and an apparently less severely altered, “melted” fabric in wall cross section (Site 370) suggestive of neomorphic alteration [e.g., Sexton et al., 2006a]. Other drill sites where the hiatus falls slightly further up section (Sites 1050C and 511) (site numbers colored light blue in Figure 2) yield foraminifera with a range of taphonomies. Sites where the hiatus falls much further up section (Site 137) or not at all (Site 763B) (site numbers colored orange in Figure 2) show noticeably better preserved benthic foraminifera from calcareous nannofossil biozone NC10 (Figure 3), suggesting a link between the hiatus and severe diagenetic alteration of calcareous microfossils in immediately underlying sediments.

Figure 3.

Scanning electron micrographs of benthic foraminifera whole tests and wall cross sections from mid-Cretaceous calcareous nannofossil biozone NC10. Numbers refer to DSDP and ODP sites. Sites that host severely diagenetically altered foraminifera (restricted to the North Atlantic, apart from Site 363) are marked by asterisks. The scale bar for whole tests is 100 μm; the scale bar for wall cross sections is 10 μm.

Figure 3.

(continued)

5. Impact of Diagenetic Alteration on Benthic Foraminiferal δ18O Across the North Atlantic

[9] In Figure 4 we compare measured δ18O for benthic foraminifera at all sites (data in Table 1) to the estimated value of δ18O of calcite grown in isotopic equilibrium with pore waters at these sites across a range of depths from the seafloor to 1000 m burial for two time slices: (1) the interval during which the foraminifera lived (biozone NC10, solid gray line) and (2) the interval of the hiatus (dashed gray line). We estimate δ18O of “equilibrium calcite” using a gradient in δ18O of pore fluid of −2.5‰/km [Lawrence and Gieskes, 1981], a conservative geothermal temperature gradient of 40°C/km [Rao et al., 2001] and local bottom water temperatures (BWTs) for the two time slices taken from Poulsen et al. [2001] and Huber et al. [2002]. Thus, the offset in the two equilibrium calcite lines simply reflects Cretaceous long-term cooling of deep-sea temperatures from NC10 to the time frame represented by the hiatus. We plot average δ18O values for multiple analyses of benthic foraminifera at each site (see Table 1) against present burial depth (dark blue = sites with a hiatus immediately overlying NC10; light blue = sites with a hiatus just above NC10; orange = sites without a hiatus, or hiatus occurs much further up section). The distance along the horizontal axes between the intersection points of the vertical colored dashed lines and the solid gray lines with the seafloor (0 m below seafloor (mbsf)) represents the offset between measured and estimated equilibrium foraminiferal calcite δ18O for NC10. Thus, the greater this distance, the more diagentically altered we interpret the measured δ18O value to be. Foraminifera from sites outside the North Atlantic (Sites 363, 511, 763) yield δ18O values that are similar to both one another and to equilibrium calcite at the seafloor during NC10. Of the North Atlantic sites, those where the base of the hiatus occurs well above NC10 (orange symbols) host benthic foraminifera with δ18O values that are lower than those for equilibrium calcite at the seafloor during NC10 (Site 137), indicative of warmer paleotemperatures than the model-derived estimates [Poulsen et al., 2001] used in calculating equilibrium δ18O. However, North Atlantic sites where the base of the Upper Cretaceous hiatus immediately overlies NC10 (dark blue) host benthic foraminifera with δ18O values that are generally significantly higher than those for equilibrium calcite at the contemporaneous seafloor. The exception to this pattern is DSDP Site 550, where NC10 benthic foraminifera register much lower δ18O values than any other site, and than equilibrium calcite at the seafloor. Finally, North Atlantic sites where the hiatus falls a little above NC10 (e.g., 1050) give intermediate paleotemperatures.

Figure 4.

Average values for multiple δ18O analyses of benthic foraminifera within NC10 (∼99 Ma) at each site (colored squares) compared to estimated δ18O of calcite grown in equilibrium with pore waters at these sites during two time slices: (1) NC10 (solid gray lines) and (2) the interval of the hiatus (dashed gray lines). Dark blue, hiatus immediately overlying NC10; light blue, hiatus just above NC10; orange, no hiatus or its base is further up section. Error bars denote 1σ of all δ18O analyses within NC10 at each site. The absence of an error bar indicates 1σ error is less than the thickness of the symbol. Length of symbols along the y axis is proportional to the respective stratigraphic length of biozone NC10. The δ18O of equilibrium calcite is estimated using a gradient in δ18O of pore fluid of −2.5‰/km [Lawrence and Gieskes, 1981], a conservative geothermal temperature gradient of 40°C/km [Rao et al., 2001], and “local” bottom water temperatures (BWTs) for the two time slices taken from Poulsen et al. [2001] and Huber et al. [2002]. BWTs used for NC10 are 16°C (North Atlantic sites) and 14°C (Sites 511 and 763). BWTs used for the interval of the hiatus are taken from the age of the top of the hiatus (or, where an interval of extremely low sedimentation rate must be inferred (see Figure 2), from the age of the hiatus' midpoint). Note that NC10 foraminifera from dark blue sites generally yield higher δ18O than either light blue or orange sites and than δ18O of equilibrium calcite at the seafloor (0 mbsf) during NC10 (solid lines). But NC10 foraminifera immediately underlying a hiatus (dark blue) generally yield very similar δ18O to that of equilibrium calcite at the seafloor during the hiatus (dashed lines). Numbers refer to DSDP and ODP sites.

Table 1. List of all Stable Isotope Data Used in Figure 4 Along With Data Sources
SiteCoreSectionTopBaseDepth (mbsf)δ13Cδ18OSpeciesReferencea
37021139.541.5682.9−1.0620.006Gyroidinoides sp.1
37023199.5102.5702.49−5.421−1.179Gyroidinoides sp.1
370233100103705.5−1.628−0.205Gyroidinoides sp.1
3702429396713.43−3.611−1.243Gyroidinoides sp.1
3702446063716.1−0.504−0.898Gyroidinoides sp.1
398D591117120975.170.1720.411Gyroidinoides infracretacea1
398D5924043975.90.7820.081Gyroidinoides infracretacea1
398D5928083976.30.798−0.289Gyroidinoides infracretacea1
398D5933336977.330.7641.124Gyroidinoides infracretacea1
398D60240.543.5985.410.4951.008Gyroidinoides infracretacea1
398D60278.581.5985.780.6200.453Gyroidinoides infracretacea1
398D5924043975.91.0140.269Gyroidinoides infracretacea1
398D5928083976.31.0820.229Gyroidinoides infracretacea1
398D5933336977.330.8570.993Gyroidinoides infracretacea1
398D60240.543.5985.410.6861.461Gyroidinoides infracretacea1
398D60278.581.5985.780.512−0.285Gyroidinoides infracretacea1
5452813235.5255.820.2760.130Gyroidinoides infracretacea1
54528CC47256.360.4550.619Gyroidinoides infracretacea1
5452921417266.350.5101.415Gyroidinoides infracretacea1
5452924044266.610.2480.559Gyroidinoides infracretacea1
54530140.543274.9−0.057−0.120Gyroidinoides infracretacea1
550B1835962620.592.510−2.378Gyroidinoides infracretacea1
550B2024447636.941.017−6.116Gyroidinoides infracretacea1
1371334648323.460.196−1.645Gyroidinoides infracretacea1
1371334648323.460.554−1.94Gavelinella sp.1
1371339193323.910.437−1.484Gyroidinoides infracretacea1
1371339193323.910.527−1.876Gavelinella sp.1
137141107109340.070.142−2.791Gyroidinoides infracretacea1
137142137139341.87−0.182−2.251Gyroidinoides infracretacea1
1371432325342.230.079−2.781Gyroidinoides infracretacea1
1371437678342.760.196−3.374Gyroidinoides infracretacea1
1371449698344.460.399−2.607Gyroidinoides infracretacea1
1371449698344.460.764−3.135Gavelinella sp.1
137144113116344.630.28−1.996Gyroidinoides infracretacea1
137144113116344.630.849−2.348Gavelinella sp.1
137144142145344.920.532−2.1Gyroidinoides infracretacea1
1371451214345.120.341−2.109Gyroidinoides infracretacea1
1371451214345.120.659−2.532Gavelinella sp.1
1371452426345.240.156−2.097Gyroidinoides infracretacea1
1371452426345.240.608−2.399Gavelinella sp.1
1371455759345.570.113−3.362Gyroidinoides infracretacea1
1371457981345.790.363−2.531Gavelinella sp.1
1371459698345.960.349−2.535Gyroidinoides infracretacea1
1371462022346.70.306−2.764Gyroidinoides infracretacea1
1371523638349.860.624−3.664Gyroidinoides infracretacea1
1371523638349.860.798−2.773Gavelinella sp.1
137152116118350.66−0.04−2.841Gyroidinoides infracretacea1
1371634345378.430.856−3.035Gavelinella sp.1
1371645254380.020.703−3.416Gavelinella sp.1
10502536064542.30.861−0.705Berthelina sp.2
10502541620543.361.501−0.749mixed benthics2
1050261139141549.691.45−0.722Gavelinella sp.2
10502615356548.930.675−0.646Gavelinella sp.2
10502615356548.930.194−0.717Berthelina sp.2
10502625659550.461.035−0.549Nuttalides sp.2
1050263120122552.60.077−0.507Berthelina spp.2
10502637376552.130.889−0.586Berthelina sp.2
1050264140142554.3−0.101−0.903Berthelina spp.2
10502648386553.730.504−0.506Berthelina spp.2
10502648386553.730.657−0.619Gyroidina globosa2
10502658386555.230.904−0.837Berthelina sp.2
105026CC910555.570.991−0.538Gavelinella sp.2
10502717376558.730.291−0.768Gavelinella sp.2
10502728588560.350.024−0.924Berthelina sp.2
10502728588560.350.688−0.697Berthelina spp.2
10502737982561.79−0.012−0.852Berthelina sp.2
10502737982561.790.291−0.297Berthelina sp.2
1050274134136563.840.420.15Berthelina spp.2
10502743436562.84−0.196−0.81Berthelina spp.2
10502748083563.30.468−0.727Planulina sp.2
1050275123125565.230.728−0.214Berthelina spp.2
10502753234564.320.111−0.528Berthelina sp.2
10502767478566.240.061−0.597Berthelina sp.2
10502767478566.240.137−0.621Berthelina sp.2
10502817073568.30.260.142Berthelina sp.2
10502826164569.710.42−0.626Berthelina sp.2
10502836063571.2−0.32−0.946Berthelina sp.2
1050284146150573.56−0.279−0.813Berthelina sp.2
10502846669572.760.422−0.717Planulina sp.2
10502855053574.10.018−0.707Berthelina sp.2
10502855053574.10.472−0.639Berthelina flat2
10502915760577.770.017−0.49Berthelina sp.2
10502962124584.910.46−0.64Berthelina sp.2
10502962124584.910.257−0.464Berthelina sp.2
10502957275583.921.701−0.229Epistomina sp.2
10502957275583.92−0.199−0.38Berthelina sp.2
10503027275589.021.786−0.38Epistomina sp.2
10503037275590.521.777−0.299Epistomina sp.2
10503047881592.080.149−0.509Berthelina intermedia2
10503118185597.21−0.009−0.328Berthelina sp.2
10503135053600.2−0.065−0.52Bethelina sp.2
10503148790601.77−0.948−0.727Berthelina sp.2
10503167983604.690.256−0.66Berthelina sp.2
105031CC  6061.594−1.54Berthelina sp.2
3632623944441.392.163−0.846Gyroidinoides infracretacea1
3632713841458.882.393−1.202Gyroidinoides infracretacea1
3632713841458.881.682−0.934Gavelinella sp.1
3632813841478.281.865−1.121Gavelinella sp.1
3632833842481.281.503−0.536Gavelinella sp.1
5114944346427.930.451−0.526Gyroidinoides infracretacea1
5114933034426.3−0.5180.078Gavelinella sp.1
5114944346427.931.237−0.320Gavelinella sp.1
51149565.567.5429.361.086−0.558Berthelina sp.3
51149565.567.5429.361.376−0.448Berthelina sp.3
511495103105430.041.704−0.555Berthelina sp.3
5114952426429.240.94−0.38Gavelinella sp.2
5114962426430.741.38−0.33G. globosa2
5115032628435.761.39−0.24Gyroidinoides globosa2
5114956163429.611.61−0.43Gavelinella (plano-convex)4
5115012729432.771.69−0.36Gavelinella (rounded)4
5115012729432.771.66−0.4Gavelinella (rounded)4
5115012729432.771.51−0.47Gavelinella (plano-convex)4
5115025355434.531.44−0.19Gavelinella (plano-convex)4
5115025355434.531.06−0.66Gavelinella (plano-convex)4
5115025355434.531.22−0.27Gavelinella (plano-convex)4
5115025355434.531.00−0.35Gavelinella (plano-convex)4
5115035254436.021.80−0.21Gavelinella (plano-convex)4
5115042022437.201.75−0.12Gavelinella (plano-convex)4
51150CC57438.601.620.02Gavelinella (plano-convex)4
763B2314043389.9−2.113−0.237Gyroidinoides infracretacea1
763B2534043411.9−0.414−2.068Gyroidinoides infracretacea1
763B2834043440.40.486−0.040Gyroidinoides infracretacea1
763B2314043389.9−1.951−0.159Gavelinella sp.1
763B2534043411.90.783−0.439Gavelinella sp.1
763B2834043440.40.583−1.679Gavelinella sp.1

[10] The sites with a hiatus immediately overlying NC10 (Sites 370, 398, 545, 550) are the same ones that also host the poorly preserved foraminifera (Figure 3, images with asterisks). The implied link between the hiatus and diagenetic alteration of benthic foraminifera in the immediately underlying sediments gains support from the observation (at Sites 370, 398 and 545) that foraminiferal δ18O values measured at these sites are similar to those of equilibrium calcite at the seafloor during the subsequent interval of nonsedimentation (dashed gray lines in Figure 4). This suggests that the relatively high δ18O signatures of these extremely poorly preserved foraminifera (Figure 3, images with asterisks) are a product of substantial diagenetic alteration while exposed at or near the seafloor during the interval of nonsedimentation represented by the Upper Cretaceous hiatus. We can be confident that diagenetic alteration of these foraminifera was not dominated by late-stage recrystallization at greater burial depths because of the large isotopic offsets between measured δ18O and that of equilibrium calcite grown at the present burial depth (shown by horizontal colored dashed lines). Recrystallization of North Atlantic benthic foraminifera at cooler BWTs during the Late Cretaceous is consistent with global compilations of foraminiferal δ18O data indicating that, following the mid-Cretaceous acme of global warmth, ocean temperatures cooled globally by at least 7°C by the late Campanian [Huber et al., 2002; Shipboard Scientific Party, 2002; Wilson et al., 2002]. Cooler Late Cretaceous BWTs in the North Atlantic are also compatible with numerical modeling experiments predicting that Turonian BWTs in this basin cooled by 2 to 8°C from their Albian equivalents [Poulsen et al., 2003], following the late Cenomanian [Pletsch et al., 2001] connection to the wider, global ocean. Instead of recrystallization at the seafloor, the exceptionally low foraminiferal δ18O values from Site 550 suggest late-stage recrystallization at substantial burial depths (e.g., about 400 m (Figure 4)) under the influence of high “geothermal” temperatures and low pore fluid δ18O.

[11] In Figure 5 we show calculated BWTs for each site based on the NC10 benthic foraminifer δ18O data from Figure 4 (colors as in Figure 4). In line with their relatively high measured δ18O values, diagenetically altered benthic foraminifera from North Atlantic sites with a hiatus immediately overlying NC10 (Sites 370, 398, 545) yield anomalously cool estimates of deep-ocean temperature (dark blue numbers in Figure 5). It is notable that Site 370 provides somewhat warmer paleotemperatures, in line with its apparently less severe diagenetic alteration (Figure 3), probably a consequence of its clay-rich lithology (Figure 2). The anomalously cool latest Albian/earliest Cenomanian paleotemperatures (dark blue numbers in Figure 5) are inconsistent both with our contemporaneous warmer deep-ocean paleotemperatures from sites with much better preserved benthic foraminifera (orange numbers in Figure 5) and with predictions of extreme deep-ocean warmth in the Albian North Atlantic from numerical modeling [Poulsen et al., 2001]. The site in the North Atlantic (1050) with a hiatus just above NC10 yields intermediate temperatures (light blue number) similar to those from Site 370. However, the equivalent site (i.e., light blue) outside the North Atlantic (Site 511) registers temperatures that are in line with those from another high-latitude Southern Hemisphere site hosting well-preserved foraminifera (e.g., 763B). Despite the presence of a hiatus at Site 511 just above NC10 (Figure 2), it appears that its clay-rich lithology (Figure 2) inhibited diagenetic alteration (Figure 3).

Figure 5.

Paleogeographic reconstruction (as in Figure 1b) showing mean δ18O-derived deep-ocean paleotemperatures for biozone NC10 at each site (based on δ18O data in Figure 4). Numbers in colored font (colors as in Figures 2 and 4: dark blue, hiatus immediately overlying NC10; light blue, hiatus just above NC10; orange, no hiatus or its base is further up section) refer to paleotemperatures calculated using a δ18Ow value of −1.0‰ (SMOW) (compatible with a deglaciated planet [Shackleton and Kennett, 1975]) and paleotemperature equation (1) of Bemis et al. [1998]. Note the generally much cooler δ18O-derived paleotemperature estimates from North Atlantic sites with a hiatus immediately overlying NC10 (dark blue font).

6. Upper Cretaceous Hiatus and Equatorial Atlantic Gateway Opening

[12] If our interpretation of the δ18O data is correct then we can constrain the timing of the alteration of the poorly preserved foraminifera to fall between the age of the altered foraminifera (earliest possible date) and the top of the hiatus (i.e., sometime between the earliest Cenomanian, ∼99 Ma ago, and the Coniacian to middle Campanian, 86 to 77 Ma ago). This age range for alteration is compatible with independent estimates of the timing of equatorial Atlantic gateway opening [Pletsch et al., 2001; Friedrich and Erbacher, 2006].

[13] Our findings suggest that the extensive Upper Cretaceous hiatus documented in the North Atlantic Ocean was perhaps somehow linked to the opening of the equatorial Atlantic gateway. Following gateway opening between Cenomanian [Pletsch et al., 2001] and Campanian [Friedrich and Erbacher, 2006] time, deep waters of the North Atlantic finally became fully connected with the cooler global ocean. As a consequence, foraminifera of NC10 age at shallow burial depths or exposed at the seafloor came into contact with much cooler, and potentially more corrosive, deep waters than those in which they had lived. Under typical pelagic sedimentary conditions, interstitial pore waters are buffered with respect to carbonate ion concentrations ([CO32−]) owing to the dissolution of biogenic carbonates. However, this was not the case for the poorly preserved North Atlantic foraminifera documented in this study because of the prolonged sedimentary hiatus (although clay-rich lithologies appear to have somewhat reduced the diaganetic impact of the hiatus).

[14] The cause of the widespread hiatus was likely either strengthened vigor in bottom water currents (i.e., physical erosion) or more corrosive deep waters (i.e., chemical dissolution) within the newly opened North Atlantic. Evidence in support of dissolution is found in the stratigraphic record of CaCO3 content of North Atlantic sediments through the Upper Cretaceous. Where Cenomanian to Campanian sediments are found in the North Atlantic they are typically red or multicolored clay with extremely low (≤ a few %) CaCO3 contents (at paleodepths from 5.5 up to at least 2.5 km) [Thierstein, 1979; Tucholke and Vogt, 1979; Arthur and Dean, 1986] and accumulated at extremely low sedimentation rates (<1 m/Ma). This prominence of clay in the North Atlantic persists from the lower Cenomanian (∼99 Ma ago) to the upper Campanian (∼75 Ma ago), a stratigraphic interval corresponding closely to that of the extensive sedimentary hiatus (Figure 2). This preponderance of clay (and absence of CaCO3) suggests a prolonged (∼24 Ma) interval of comparatively shallow (≤2.5 km [Tucholke and Vogt, 1979; Arthur and Dean, 1986]) calcite compensation depths (CCDs). This shallow CCD state is presumably, at least in part, attributable to high eustatic sea levels during the Late Cretaceous [Haq et al., 1987], submerging continental shelves and, through widespread shelf deposition of calcareous microfossils [Roth, 1986], causing enhanced basin to shelf fractionation of CaCO3 [Hays and Pitman, 1973; Berger and Winterer, 1974; Sclater et al., 1979; Opdyke and Wilkinson, 1988]. Yet, this sedimentation regime has also been tentatively linked to the opening of the equatorial Atlantic gateway [Poulsen et al., 2003], presumably through incursion of more corrosive deep waters into the previously isolated North Atlantic basin. Our results, and our documentation of a temporal coincidence between the duration of the hiatus, a CaCO3-poor sedimentation regime and estimates of gateway opening, lend support to this view.

7. Conclusions

[15] Diagenetic alteration of planktic foraminiferal calcite from tropical latitudes at the seafloor is now the widely accepted mechanism to explain a long-standing discrepancy during past “greenhouse” climates between relatively cool foraminiferal δ18O-derived tropical SSTs in comparison to much warmer tropical temperatures predicted by numerical models [Schrag et al., 1995; Wilson and Opdyke, 1996; Norris and Wilson, 1998; Pearson et al., 2001; Wilson et al., 2002; Sexton et al., 2006a]. In contrast, the taphonomy of benthic foraminiferal calcite, and its influence on the deep-sea paleotemperature record, has received little attention. Numerical modeling indicates that the carbonate recrystallization histories of deep-sea sections are dominated by events in their early burial history [Rudnicki et al., 2001]. This means that the effect of benthic foraminiferal diagenetic alteration on the deep-sea paleotemperature record should be strongly influenced by the extent of postburial sediment–pore fluid exchange. The rate of this exchange will likely be determined by sedimentation rate and lithology. We show that an extreme change in sedimentation rate (a prolonged and widespread Upper Cretaceous sedimentary hiatus in the North Atlantic) had a major, detrimental impact on the taphonomy and δ18O of benthic foraminiferal calcite from underlying mid-Cretaceous (nannofossil biozone NC10, ∼99 Ma ago) strata at multiple sites.

[16] Scanning electron microscopy of whole tests gives a false impression of the preservation state of benthic foraminifera. Whole tests appear to show at least moderate preservation at all sites whereas SEM images of cross sections of test walls reveal this to be a misleading picture. We find that, at sites where the hiatus immediately overlies sediments of NC10 age, benthic foraminifera are typically extremely poorly preserved internally, with δ18O values substantially higher than those from contemporaneous better preserved benthic foraminifera at sites without an immediately overlying hiatus. These anomalously high δ18O values are similar to estimated δ18O values for calcite precipitated in equilibrium with cooler, Late Cretaceous deep waters during the subsequent interval of nonsedimentation. We interpret these observations to indicate that benthic foraminifera were heavily altered while exposed at (or very near) the seafloor for a prolonged period of time during the Late Cretaceous. Sediment lithology appears to modulate the intensity of alteration, with clay-rich sediments affording foraminifera most protection.

[17] Our results suggest that sedimentation rate and lithology together determine the extent of diagenetic alteration. The importance of these two variables is probably a function of their mutual influence on the degree of exchange between sediments and pore fluids. Sedimentation rates control sediment–pore fluid exchange via diffusion (the effects of distance, gradients and flux), while lithology regulates exchange via porosity. Furthermore, our results point to the postdepositional alteration of individual foraminifera being a slower and longer-lived process than the maximal time span for alteration of ∼10 Ma suggested by numerical model analysis of bulk carbonate [Rudnicki et al., 2001], with important paleoceanographic implications. Our findings also highlight the importance of detailed examination of the taphonomy of benthic foraminiferal wall structures in paleoceanographic studies.

Acknowledgments

[18] We thank Mike Bolshaw, Matt Cooper, and Richard Pearce for laboratory assistance and Oliver Friedrich, Agostino Merico, and Richard Norris for useful discussions. We also thank Mike Arthur and Brian Huber for constructive reviews and Gerald Dickens for valuable editorial suggestions. This research used samples and data provided by the Ocean Drilling Program (ODP). ODP (now IODP) is sponsored by the U.S. National Science Foundation and participating countries under the management of Joint Oceanographic Institutions, Inc. Financial support was provided by a Natural Environment Research Council (NERC) studentship (to P.F.S.), a European Commission Marie Curie Outgoing International Fellowship (to P.F.S.), and a NERC grant (to P.A.W.).

Ancillary