Alkenone δ²H values – a viable seawater isotope proxy? New core-top δ²H_C37:3 and δ²H_C37:2 data suggest inter-alkenone and alkenone-water hydrogen isotope fractionation are independent of temperature and salinity

Abstract

A growing body of work has demonstrated that the δ²H values of alkenones reflect the δ²H values (δ²H_H2O) and / or salinity of the fluid in which they are produced. If so, δ²H_alkenone values would act as a surface seawater isotope / salinity proxy, similar to foraminiferal δ¹⁸O values, but advantaged in locations with poor carbonate preservation and / or high organic content. Nevertheless, laboratory culture, sediment trap, and water column studies have failed to consistently characterize the effects of temperature, alkenone-producing species, and salinity itself on the δ²H_alkenone-salinity and -seawater isotope relationships, and a robust sedimentary alkenone-based calibration remains elusive.

Most δ²H_alkenone datasets report δ²H_C37, i.e., combined δ²H_C37:3 and δ²H_C37:2 values, and differ in how they address inter-alkenone fractionation (i.e., α_C37:3-C37:2). To constrain controls on alkenone hydrogen isotope systematics in the natural environment, we measured δ²H values of C₃₇ and C₃₈ alkenones from 20 open ocean core tops by gas chromatography-stable isotope ratio mass spectrometry after separation of di- and tri-unsaturated forms. Core-top δ²H_alkenone data points are currently concentrated in extreme-salinity regions; in combination with our new values from a more moderate range of open ocean δ²H_H2O / salinity values, for sedimentary alkenones, we show that 1) mean inter-alkenone hydrogen isotope fractionation is negligible (α_C37:3-C37:2 = 1.002 ± 0.006), and therefore that δ²H_C37:3 and δ²H_C37:2 values can be measured in bulk; 2) temperature and salinity have little impact on alkenone-water fractionation (i.e., α_C37-H2O) (mean 0.803 ± 0.010) relative to their expected variability in the ocean; and 3) δ²H_C37 and δ²H_H2O values are correlated such that statistically identical δ²H_C37-, δ²H_C37:3-, and δ²H_C37:2-δ²H_H2O regressions yield a core-top-based calibration of δ²H_C37 = 1.44 (± 0.13) * δ²H_H2O – 191.62 (± 1.13) ‰. This is indistinguishable from water column calibrations, suggesting a consistent response of environmental δ²H_C37 values to changes in δ²H_H2O values.

This calibration still contains a high amount of scatter (∼ 7 ‰), perhaps attributable to irradiance, growth rate, intra- or interspecies variability, or other factors difficult to constrain in sedimentary material. Nevertheless, when applied to the well-constrained Last Glacial Maximum-to-present mean ocean δ²H_H2O change of ∼ 8.8 ‰, it (1.44 ‰ δ²H_C37 per 1 ‰ δ²H_H2O change) reproduces the mean δ²H_{C37 Modern} – δ²H_{C37 LGM} shift observed from the handful of extant down-core records, legitimizing the observed lack of temperature or salinity effects on α_C37-H2O. This suggests that combined δ²H_C37:3 + δ²H_C37:2 values are a valid proxy for δ²H_H2O values in open ocean settings where E. huxleyi and G. oceanica dominate, although additional efforts will be required to refine the core-top calibration for universal use.

Introduction

Seawater isotope (δ¹⁸O_H2O, δ²H_H2O value) proxies are our primary source of information regarding large-scale, long-term shifts in global climate (Emiliani, 1955, Imbrie et al., 1984, Zachos, 2001, Lisiecki and Raymo, 2005). Once such a tool has been calibrated in modern material and proven to reflect present-day conditions, it can be applied down-core to examine temporal paleoceanographic changes. The ideal proxy for δ¹⁸O_H2O or δ²H_H2O values is ubiquitous through space and time, resistant to alteration, and related to ²H/H_H2O or ¹⁸O/¹⁶O_H2O independent of other environmental factors. Alkenones meet the first two of these criteria (Brassell et al., 1986, Prahl et al., 1989); here, we attempt to determine if their δ²H values fulfill the third. If alkenone-water hydrogen isotope fractionation is invariant of variables such as temperature or salinity, then δ²H_alkenone values would be a viable alternative to foraminiferal δ¹⁸O values (δ¹⁸O_calcite), particularly in regions hostile to carbonate preservation or devoid of foraminifera.

Alkenones—long-chain methyl (Me) or ethyl (Et) ketones produced by two cosmopolitan prymnesiophyceae coccolithophorids (Emiliana huxleyi and Gephyrocapsa oceanica) in the modern ocean (de Leeuw et al., 1980, Volkman et al., 1980b, Volkman et al., 1980a, Volkman et al., 1995, Conte et al., 1994)—are typically used as recorders of marine sea surface temperature (SST) due to the temperature-dependence of the proportion of tri- and diunsaturated 37-carbon alkenones (C_37:3 and C_37:2) (Müller et al., 1998). These molecules are widespread across most of the ocean and advantaged over foraminiferal calcite in several notable ways, including a greater resilience to diagenesis (Marlowe et al., 1984a, Marlowe et al., 1984b, Brassell et al., 1986, Conte et al., 1992, Winter et al., 1994) and limited number of strictly planktic producers, minimizing any species-specific vital effects and habitat-related depth effects involved. The exploration of their δ²H values (δ²H_alkenone) as a paleo-sea surface salinity (SSS) or seawater isotope proxy began once it was shown that δ²H_alkenone values increase linearly with growth water δ²H values over large ranges (∼ 500 ‰) of the latter (Paul, 2002, Englebrecht and Sachs, 2005). Down-core δ²H_alkenone records are significantly correlated with contemporaneous foraminiferal δ¹⁸O measurements (Pahnke et al., 2007, Weiss et al., 2019a), and there has been a proliferation of these records in recent decades (Section 4.6).

Nevertheless, additional factors such as growth rate (Schouten et al., 2006, Wolhowe et al., 2009, Wolhowe et al., 2015, M’boule et al., 2014, Sachs and Kawka, 2015), species-specific effects (Schouten et al., 2006, D’Andrea et al., 2007, Wolhowe et al., 2009, Schwab and Sachs, 2011, Nelson and Sachs, 2014, M’boule et al., 2014), temperature (Schouten et al., 2006, Wolhowe et al., 2009, Wolhowe et al., 2015, Gould et al., 2019), irradiance (van der Meer et al., 2015, Wolhowe et al., 2015, Weiss et al., 2017, Wolfshorndl et al., 2019), and salinity itself (Schouten et al., 2006, M’boule et al., 2014, Wolhowe et al., 2015, Sachs et al., 2016, Gould et al., 2019) may affect alkenone-seawater hydrogen isotopic fractionation (α_{alkenone–H2O}). Some of these confounding factors could be easily corrected for—e.g., temperature, with the built-in U^k’₃₇-based thermometer—but many are difficult to reconstruct on geologic timescales, and would severely limit a seawater isotope proxy if they exert a major influence on δ²H_alkenone values.

These potential effects, not directly related to environmental ²H/H, have primarily been studied in algal cultures or on marine suspended particulate matter, which afford control over ambient conditions and a wide experimental range. Nevertheless, these have their downsides. Despite complete control over experimental conditions, the relationship between haptophyte species (Schouten et al., 2006, Wolhowe et al., 2009) and temperature (Schouten et al., 2006, Wolhowe et al., 2009, van der Meer et al., 2013) and inter-alkenone and alkenone-water hydrogen isotope fractionation have proven difficult to reproduce (4.1 Inter-alkenone fractionation: α, 4.5 Core-top, culture, and water column discrepancies), while the δ²H_C37 values of open-ocean water column / suspended particulate organic matter (SPOM) are highly variable: fluctuations of up to 7 ‰ over the course of a single day have been reported (Wolfshorndl et al., 2019). At present, SPOM (Häggi et al., 2015, Gould et al., 2019) and core-top (Weiss et al., 2019c) δ²H_alkenone-δ²H_H2O calibrations diverge in slope and intercept.

Ultimately, δ²H_alkenone values from surficial sediments are the best analogue for down-core δ²H_alkenone values: if the development of the δ²H_alkenone proxy follows the same path as the U^k’₃₇ SST proxy, a salinity / seawater isotope calibration based on core-top data should be the cap on a combined approach featuring culture, water column / SPOM, and mesocosm results in agreement with one another. Unfortunately, this is not yet the case, but the development of the U^k’_{37 proxy} followed a similar course: early tests of the U^k’₃₇-SST relationship in culture also found it to vary with species (E. huxleyi versus G. oceanica) (Volkman et al., 1995), growth phase (exponential, late logarithmic, stationary) (Epstein et al., 1998, Conte et al., 1998, Yamamoto et al., 2000), nutrient levels (Epstein et al., 1998), or otherwise unidentified laboratory conditions (Prahl et al., 1988, Conte et al., 1995, Conte et al., 1998, Sawada et al., 1996, Versteegh et al., 2001). Many of these were irreproducible, however, and the lack of species / growth phase / nutrient level effects in repeated core-top-based U^k’₃₇-SST calibrations (Prahl et al., 1988, Müller et al., 1998, Conte et al., 2006, Tierney and Tingley, 2018) suggests that E. huxleyi and G. oceanica react profoundly differently in culture versus natural media and / or that the aggregate sedimentary pool smooths out many of the effects exhibited in individual cultures.

Finally, inter-alkenone fractionation (particularly between C_37:3 and C_37:2, α_C37:3-C37:2) is difficult to quantify due to the significant effort required to chemically separate the tri- and diunsaturated alkenones and large sample quantities needed for isotope ratio mass spectrometric (IRMS) analysis. When individual alkenone δ²H values (i.e., δ²H_C37:3, δ²H_C37:2) from a mixture of culture, SPOM, and sedimentary settings are reported (D’Andrea et al., 2007, Wolhowe et al., 2009, Schwab and Sachs, 2009, Schwab and Sachs, 2011, van der Meer et al., 2013, Nelson and Sachs, 2014, Sachs and Kawka, 2015, Sachs et al., 2016, Wolfshorndl et al., 2019, Weiss et al., 2019c, Weiss et al., 2019a), they yield varying values for α_C37:3-C37:2. In many cases, α_C37:3-C37:2 ≠ 1; if so, temperature-driven changes in the relative amounts of C_37:3 and C_37:2 could bias the “bulk” / combined δ²H values (i.e., δ²H_C37). On the other hand, others argue that hydrogen isotopes are distributed from the total alkenone pool by the desaturation of C_37:2 to form C_37:3 and therefore that δ²H_C37 values are the more accurate proxy (van der Meer et al., 2013).

Perhaps due to some mixture of the above, existing δ²H_C37-δ²H_H2O calibrations based on environmental (non-culture) samples range from δ²H_C37 = 1.86 (± 0.56) * δ²H_H2O – 202 (± 4.12) (1 σ) (Gould et al., 2019) to 0.84 (± 0.07) * δ²H_H2O – 186 (± 1.37) ‰ (Weiss et al., 2019c). Over a theoretical -12–12 ‰ range of low- to mid-latitude δ²H_H2O values (Fig. 1C), this difference yields a δ²H_C37 offset of up to 24 ‰. To address some outstanding issues with the δ²H_alkenone proxy and produce an updated δ²H_C37-δ²H_H2O calibration, we synthesize existing measurements with a new dataset of globally distributed core-top δ²H_C37:3 and δ²H_C37:2 values with the following objectives:

• Quantify inter-alkenone fractionation factors, particularly α_C37:3-C37:2, and their relationship to several previously proposed environmental variables with high-resolution datasets in the modern (salinity, temperature);

• Constrain alkenone-water fractionation factors, particularly α_C37-H2O, and their dependence on temperature, salinity, and coccolithophore species (E. huxleyi versus G. oceanica);

• Integrate our measurements with existing core-top datasets and determine if the resulting δ²H_alkenone-SSS and / or -δ²H_H2O calibration is robust by applying it to the mean observed down-core change in δ²H_alkenone values since the Last Glacial Maximum (LGM).

To better evaluate hydrogen isotope systematics and minimize competing hypotheses driven by proven species effects (D’Andrea et al., 2007, Schwab and Sachs, 2011, Nelson and Sachs, 2014, M’boule et al., 2014), this study’s scope is limited to environments dominated by Group III Isochryisdales (E. huxleyi, G. oceanica), typically the open ocean.