A Warm, Stratified, and Restricted Labrador Sea Across the Middle Eocene and Its Climatic Optimum

Abstract

Several studies indicate that North Atlantic Deep Water (NADW) formation might have initiated during the globally warm Eocene (56-34 Ma). However, constraints on Eocene surface ocean conditions in source regions presently conducive to deep water formation are sparse. Here we test whether ocean conditions of the middle Eocene Labrador Sea might have allowed for deep water formation by applying (organic) geochemical and palynological techniques, on sediments from Ocean Drilling Program (ODP) Site 647. We reconstruct a long-term sea surface temperature (SST) drop from ~30°C to ~27°C between 41.5 to 38.5 Ma, based on TEX₈₆. Superimposed on this trend, we record ~2°C warming in SST associated with the Middle Eocene Climatic Optimum (MECO; ~40 Ma), which is the northernmost MECO record as yet, and another, likely regional, warming phase at ~41.1 Ma, associated with low-latitude planktic foraminifera and dinoflagellate cyst incursions. Dinoflagellate cyst assemblages together with planktonic foraminiferal stable oxygen isotope ratios overall indicate low surface water salinities and strong stratification. Benthic foraminifer stable carbon and oxygen isotope ratios differ from global deep ocean values by 1-2‰ and 2-4‰, respectively, indicating geographic basin isolation. Our multiproxy reconstructions depict a consistent picture of relatively warm and fresh but also highly variable surface ocean conditions in the middle Eocene Labrador Sea. These conditions were unlikely conducive to deep water formation. This implies either NADW did not yet form during the middle Eocene or it formed in a different source region and subsequently bypassed the southern Labrador Sea.

Figure 1

Modern and middle Eocene North Atlantic (paleo)geographic setting. (a) Modern‐day geography. Simplified surface ocean currents marked in light blue (cold) and pink (warm): EGC = East Greenland Current; LC = Labrador Current; NAC = North Atlantic Current; SPG = subpolar gyre; STG = subtropical gyre. Ocean basins and seaways marked in white: BB = Baffin Bay; BS = Barents Sea; FS = Fram Strait; LS = Labrador Sea; NGS = Norwegian‐Greenland Sea; NS = Nares Strait. Dark blue fill represents ocean crust, and black lines represent outlines of continental plates. Gray fills indicate modern coastlines. (b) Approximate paleogeographic reconstruction for 40 Ma, together with the paleolocation of ODP Site 647. Map produced with GPlates, using continental polygons and coastlines from Matthews et al. (2016) and the paleomagnetism‐based rotation frame of Torsvik et al. (2012). Selected oceanographic features from panel (a) annotated. Additional features in panel (b): TS = Turgay Strait; GSR = Greenland‐Scotland Ridge. Dark blue fill represents ocean crust, and black lines represent outlines of continental plates. Gray fills indicate coastlines, rotated together with the continental plates to a 40 Ma position. Light blue represents interpreted flooded continental shelf in the middle Eocene.

Figure 2

Light microscope images of representative Site 647 planktic foraminifera analyzed in this study. (a) Hantkenina australis (Sample 647A‐50R‐6, 59–61 cm), (b) Chiloguembelina ototara (Sample 647A‐50R‐6, 59–61 cm), and (c) Pseudohastigerina micra (Sample 647A‐47R‐4, 33–36 cm). Images were taken at Stockholm University with a Leica M205C binocular light microscope equipped with a Leica camera system. Scale bars all 100 μm. Note the small size of P. micra and C. ototara relative to H. australis. The shiny transparent appearance of the test calcite, revealing original fine surface details, signals excellent shell calcite preservation.

Figure 3

Scanning electron microscope (SEM) images of representative Site 647 microfossils. Taxon and sample names annotated in figure. Images (a)–(d) and (f)–(h) were taken with a Thermo Scientific Apreo SEM (uncoated, 10 kV, working distance 10 mm) at the University of California, Santa Cruz. Images (e), (j)–(l), (o), and (p) were taken with a Philips XL30 FEG ESEM (gold coated, 10 kV, spot size −3, working distance 9 mm) at Stockholm University. Images (i), (m), and (n) were taken with a Philips XL30 FEG ESEM (gold coated) at the School of Earth and Ocean Science at Cardiff University.

Figure 4

Light microscope (LM) and scanning electron microscope (SEM) images of representative Site 647 dinocysts. Taxon and sample names in figure. LM images (a–k) taken at Utrecht University with a light microscope equipped with a Leica camera system. SEM images (l–t) taken at Utrecht University with a Philips XL30 FEG ESEM (platinum coated).

Figure 5

Palynomorph and dinocyst assemblages across the middle Eocene at Site 647. (a) Relative abundance of main encountered groups of palynomorphs, stacked area diagram plotted as percentage of total palynomorphs. dino. = dinocysts; prasino. = prasinophytes; mar. = other marine palynomorphs (other algal remains and prasinophytes); bis. pol. = bisaccate pollen; other pol. = other pollen and spores; foram. = organic benthic foraminiferal linings. (b–h) Relative abundance of representative groups plotted as percentage of total dinocysts. (i) Dinocyst content in cysts per gram of dry sediment. Plotted against depth in meters below seafloor.

Figure 6

Middle Eocene surface ocean conditions at Site 647. (a) TEX₈₆‐based sea surface temperature (°C) (exponential calibration of Kim et al., 2010, in red; linear calibration of O'Brien et al., 2017, in blue) and BIT index (green); 3‐point moving averages plotted as thick lines. Open symbols represent TEX₈₆‐based SST values with BIT > 0.4. Propagated analytical plus calibration uncertainty is ±2.6°C for the Kim et al. calibration and ±2.1°C for the O'Brien et al. calibration. (b) Inferred δ¹⁸O of surface ocean (converted to ‰ VSMOW) calculated from ${\mathrm{TEX}}_{86}^{\mathrm{H}}$‐based SST (open symbols: LOESS fit of ${\mathrm{TEX}}_{86}^{\mathrm{H}}$‐based SST; closed symbols: linear interpolation of ${\mathrm{TEX}}_{86}^{\mathrm{H}}$‐based SST) and planktic foraminiferal δ¹⁸O (Figure 7), using the Erez and Luz (1983) (circles) and Kim and O'Neil (1997) (diamonds) δ¹⁸O‐to‐temperature calibrations. Open diamonds are connected with a line. Note that using ${\mathrm{TEX}}_{86}^{\mathrm{H}}$‐based SST likely overestimates the temperature effect on δ¹⁸O_foram, and thus underestimates the salinity effect, as ${\mathrm{TEX}}_{86}^{\mathrm{H}}$ is calibrated based on the modern relationship between GDGTs in core‐top sediments and satellite‐derived SSTs. The studied foraminifera would have lived at deeper, somewhat cooler depths than the depth represented by satellite‐derived SSTs. (c) Relative abundance (% of total dinocyst assemblage) of selected dinocyst species. Low‐salinity tolerant Phthanoperidinium spp. in green, relative abundance plotted from left to right. Low‐latitude/midlatitude Cleistosphaeridium spp. (red) and Cordosphaeridium gracile (purple) relative abundance plotted from right to left. Gray horizontal shading refers to the interval of absent mixed‐layer planktonic foraminifera, with dark gray representing total absence of planktic species. Purple horizontal shading denotes the interval of H. australis incursion. Plotted against depth in meters below seafloor.

Figure 7

Foraminiferal δ¹⁸O and δ¹³C at Site 647. Planktic and benthic δ¹⁸O (left) and δ¹³C (right) (‰ VPDB) for several species. Error bars denote ±1 sd. Error is slightly higher for some samples that had very small mass. Full names of planktic foraminifera are as follows: Pseudohastigerina micra, Chiloguembelina ototara, Acarinina spp., Hantkenina australis, Globoturborotalita cf. ouachitaensis, Globorotaloides quadrocameratus, Turborotalia pomeroli, and Globigerinatheka index. These are mixed‐layer dwellers, except for T. pomeroli and G. quadrocameratus, which were likely thermocline and subthermocline dwellers, respectively. Full name of benthic foraminifera is Oridorsalis umbonatus. Colors as in legend. Gray horizontal shading refers to the interval of absent mixed‐layer planktonic foraminifera, with dark gray representing total absence of planktic species. Purple horizontal shading denotes the interval of the H. australis incursion. Plotted against depth in meters below seafloor.

Figure 8

Crossplot of measured planktic and benthic foraminiferal δ¹³C (‰ VPDB) versus δ¹⁸O (‰ VPDB) for Site 647. Plotted data is from the interval 500–390 mbsf. Species labeling as in Figure 7.

Figure 9

Compilation of middle Eocene TEX₈₆‐based sea surface temperatures. Site 647 (red) data plotted together with published TEX₈₆ data from the Atlantic basin: ODP Site 913, Norwegian‐Greenland Sea (gray) (Inglis et al., 2015, 2020; Liu et al., 2009); Kysing‐4 borehole, North Sea Basin (light blue) (Śliwińska et al., 2019); ODP Site 925, equatorial Atlantic Ocean (pink) (Liu et al., 2009); ODP Site 959, equatorial Atlantic Ocean (orange) (Cramwinckel et al., 2018); Site 1263, subtropical South Atlantic Ocean (purple) (Boscolo‐Galazzo et al., 2014); and South Dover Bridge, Atlantic coastal plain (blue) (Inglis et al., 2015). TEX₈₆ record from Site 1172 (green) (Bijl et al., 2009, 2010) added as a high southern latitude end‐member. We are exclusively plotting temperatures derived using TEX₈₆ ( ${\mathrm{TEX}}_{86}^{\mathrm{H}}$ calibration of Kim et al., 2010) for optimal comparability. Propagated analytical plus ${\mathrm{TEX}}_{86}^{\mathrm{H}}$ calibration uncertainty is ±2.6°C. Age follows GTS2012.

Figure 10

Compilation of middle Eocene Acarinina spp. stable oxygen (left) and carbon (right) isotope records. Site 647 (orange) data together with published data from the North Atlantic Ocean (Site 1051; pink; Edgar et al., 2013), South Atlantic Ocean (Site 1263; green; Boscolo‐Galazzo et al., 2014), Indian Ocean (Site 748; blue; Edgar et al., 2013), Pacific Ocean (Site 865; gray; Henehan et al., 2020), and Tethys Ocean (Baskil section; brown; Giorgioni et al., 2019). Data from Sites 748 and 1051 represent a size range of Acarinina spp. Age models for all sites have been converted to GTS2012.

Figure 11

Compilation of middle Eocene benthic foraminiferal stable oxygen (left) and carbon (right) isotope records. Site 647 (black) data together with published data from the Southern Ocean (Sites 738 and 748; light and dark blue; Bohaty & Zachos, 2003; Bohaty et al., 2009), Atlantic Ocean (Site 1051 pink/purple and Site 1260 light green, Edgar et al., 2010; Site 1263 dark green, Boscolo‐Galazzo et al., 2014), and Tethys Ocean (Baskil section; brown; Giorgioni et al., 2019). Isotope values for Oridorsalis umbonatus (Sites 647 and 1051) and Nuttallides truempyi (Site 1263) have been converted to Cibicidoides‐equivalent values following the isotope correction factors from Katz et al. (2003). Age models for all sites have been converted to GTS2012.

Figure 12

Schematic of hypothesized middle Eocene Labrador Sea circulation. Fresh surface waters are sustained by high precipitation over evaporation and/or surface water connections to the fresh Arctic. These surface waters flow out, likely to the south/east. Denser intermediate and deep waters flow in, likely from the south/east, potentially restricted by sills related to Labrador Sea spreading. Three sources of deep water are hypothesized: winter sinking along the shelf (1), sinking of inflowing intermediate water (2), and/or inflow of deep water (3).