Alkenone paleothermometry is a well-established method for estimating past sea surface temperatures (SST) or near-sea surface temperatures (nSST) in the open ocean. Since the 1980s, geoscientists have relied on this effective paleo-proxy, utilizing alkenones—lipid biomarkers produced by specific photosynthetic marine phytoplankton. The initial detection of alkenones in marine sediments coincided with advances in computerized gas chromatography-mass spectrometry, which utilized glass capillary columns that could withstand the high temperatures necessary for elution of long-chained alkenones.

The methodology for extracting and identifying alkenones from deep-sea sediments has remained consistent since its inception. Initially, sediments are freeze-dried and homogenized. A total lipid extract is then generated using lipid-soluble solvents like methylene chloride (DCM). Extraction methods include Soxhlet extraction, repeated sonication at room temperature, or using an Accelerated Solvent Extractor (ASE). Alkenone fractions can be separated from the total lipid extract using thin-layer or column chromatography, which minimizes the co-elution of other compounds. Ultimately, alkenones are extracted from their lipid matrix and quantified using a gas chromatograph (GC), typically paired with a flame ionization detector (FID), although mass spectrometers may also be employed.

Discovery of Alkenones in Sediments and Algae

Before their widespread application among paleoceanographers, alkenones were first identified in deep-sea sediments collected during Leg 40 of the Deep Sea Drilling Project (DSDP). Samples from Site 362, located off the Namibian coast, exhibited prominent peaks at specific mass-to-charge ratios in total lipid extracts. Researchers employed thin-layer chromatography (TLC) to determine the retention factor (Rf) of these organic compounds, identifying them as ketones. Subsequently, long mid-chain ketones were isolated from lipid extracts of diatomaceous ooze in Walvis Bay, and their fractionation patterns were analyzed using TLC, electron ionization mass spectrometry (EI-MS), and gas chromatography. This confirmed their identity as unsaturated long-chain ketones, now recognized as the C37, C38, and C39 ketones that comprise alkenones.

Following the identification of these ketones, geochemical researchers, notably J.K. Volkman from the University of Bristol, theorized that the coccolithophorid Emiliania huxleyi was likely a primary source of alkenones in open-ocean environments. This species is prevalent in modern oceans, thriving across diverse conditions from the equator to subpolar regions.

Volkman and his team later isolated C37, C38, and C39 methyl- and ethyl-ketones and analyzed their distribution during various growth stages of the alga. They observed significant concentrations of alkenones relative to other sediment biomarkers in Miocene and Plio-Pleistocene sediments from the Japan Trench and in recent samples from Walvis Bay and the North Sea. The distribution of ketones in these sediments mirrored those found in E. huxleyi, suggesting their resistance to biological degradation could account for the high levels of this biomarker in ocean sediments.

Research into alkenones continued as scientists sought to identify other algal species that could also contribute these lipids to seafloor deposits. This interest arose partly from the discovery of alkenones in Eocene-aged sediments—much older than the estimated emergence of E. huxleyi. Consequently, researchers hypothesized that other alkenone-producing species likely existed in the geological past and may still be present in modern oceans. This assumption is crucial for applying the alkenone paleothermometer to sediments older than the Quaternary, as it relies on the premise that these ancient producers responded to temperature similarly to E. huxleyi. A few years later, researchers at the University of Bristol reported discovering long-chain unsaturated ketones and esters in four additional algal species, all of which belonged to the haptophyte class but were not coccolithophorids and were found in coastal or saline terrestrial environments.

Additionally, another marine coccolithophore species, Gephyrocapsa oceanica, was identified as a source of alkenones. While its distribution is more restricted than that of E. huxleyi, it is predominantly found in tropical and subtropical waters. Despite evolutionary differences among these species, accumulating molecular and paleontological evidence suggests a significant genetic relationship among marine alkenone producers, bolstering confidence in the alkenone paleothermometer's validity for sediments predating modern taxa.

Development of the Open-Ocean Alkenone-Based Paleotemperature Proxy

Researchers soon began to explore the applications of alkenones as biomarkers, hypothesizing their chemotaxonomic significance. They examined long-chained carbon compound distribution in cultures of thirteen algal species from the class Prymnesiophyceae and found that E. huxleyi exhibited a distinct array of polyunsaturated n-C36 acid esters compared to others. However, it became clear that alkenones were more than just chemotaxonomic indicators; they served as records of historical ocean temperatures. Following laboratory demonstrations that alkenone unsaturation varies with growth temperature, Brassel and colleagues were the first to establish a correlation between alkenone unsaturation and temperature in deep-sea sediments. They found a strong link between the unsaturation ratio of alkenones and planktic foraminiferal oxygen-18 isotopic ratios (?18O) in Quaternary sediments from the eastern equatorial Atlantic. The alkenone unsaturation index (Uk37) was defined as the ratio of di-, tri-, and tetra-unsaturated C37 alkenones: Uk37 = (C37:2–C37:4)/(C37:2 + C37:3 + C37:4).

In 1987, Prahl and Wakeham utilized E. huxleyi cultures to develop a calibration for alkenone unsaturation and temperature that remains in use today. They employed the shortened alkenone unsaturation index (Uk’37), calculated as Uk’37 = (C37:2) / (C37:2 + C37:3). This index has since proven reliable for paleoclimatic studies, especially since the C37:4 ketone predominates in high-latitude waters without improving the unsaturation-temperature calibration in nonpolar regions. After cultivating the alga at temperatures ranging from 8–25°C, the authors found a positive correlation between Uk’37 and growth temperature, along with reproducibility of the unsaturation index at each temperature. Their final calibration was Uk’37 = 0.033T(°C) + 0.043, suggesting that consistent linear relationships in deep-sea sediments could allow for accurate estimations of absolute sea surface paleotemperatures.

Later, comparisons of lab-grown cultures and sediments from two sites (one warm and the other cold) confirmed that their calibration effectively estimated sea surface temperatures. Sikes and colleagues also developed an alkenone thermometer based on core-top Uk37 values and associated sea surface temperatures. They concluded that the 1987 Prahl calibration is well-suited for reconstructing late Quaternary SST records due to its strong alignment with their findings.

Establishing the Alkenone Paleothermometer's Reliability

In addition to establishing a robust laboratory calibration between the alkenone unsaturation index and growth temperature, Prahl's team found minimal alteration in sedimentary alkenones compared to cultures. The stability of unsaturation indices was further tested by Elizabeth Sikes and colleagues, who discovered that the unsaturation ratio remained constant regardless of whether sediments were stored at room temperature or frozen and were resistant to degradation in acidic conditions. This was groundbreaking for paleoceanography, allowing researchers to apply the alkenone proxy in regions where carbonate-based proxies struggle due to poor preservation.

Given the similarities in biomarker conditions between laboratory cultures and marine sediments, Prahl and his collaborators suggested that either E. huxleyi is the sole producer of alkenones in today's oceans (which is inaccurate due to the discovery of G. oceanica) or that other alkenone producers biosynthesize them similarly. This second assumption is particularly important for paleoceanographers applying the alkenone paleothermometer to sediments older than 270,000 years (the estimated date of modern taxa's emergence), necessitating the belief that ancient producers' unsaturation ratios reflected ocean temperatures similarly to those of contemporary species. Subsequent research has indicated this assumption is likely valid. Marlowe and co-authors explored the relationships between alkenone occurrences in deep-sea sediments and the coccolithophorid nannofossil record, finding that only members of the Gephyrocapsaceae family, including both Emiliania huxleyi and Gephyrocapsa oceanica, were present in all alkenone-bearing sediments containing nannofossils. Thus, Gephyrocapsaceae is the most probable biological source of alkenones, with extant species contributing to the modern marine sediment record and extinct genera represented in older sediments.

Refining the Alkenone-Based Paleotemperature Proxy Calibration

Subsequently, Sikes and Volkman developed the first Uk’37–SST field calibration for ocean temperatures below 11°C. Their research involved extracting alkenones from particulate organic matter in the Southern Ocean's surface waters. They observed a linear relationship between the unsaturation index and temperature within a 6–12°C range, but their calibration (Uk’37 = 0.0414T(°C) — 0.156) differed from those established in warmer waters. Below 4°C, they noted little correlation between parameters, indicating the need for further research on the method's lower-temperature limits. The authors also highlighted potential limitations of alkenone paleothermometry as a paleoclimatic tool in regions south of the Polar Frontal Zone, where sea surface temperatures often approach freezing and ice may cover the water. Subsequent studies indicated that alkenone abundances decrease with increasing latitude in the Southern Ocean, necessitating more sensitive instruments, cleaner lipid extracts, and enhanced methods to achieve reliable sea surface temperature estimates in colder areas.

In 1998, German scientist Peter J. Müller and his colleagues established a widely-used calibration of alkenone unsaturation to temperature based on C37 methyl ketones for reconstructing near-sea-surface temperature records. Using 149 surface sediments from the eastern South Atlantic, they created a regional sediment-based Uk’37 calibration for the southern Atlantic Ocean, spanning tropical to subantarctic environments. Their results showed a strong correlation between the Uk’37 index and annual mean temperatures, with minimal influence from regional variations in seasonal primary production. They found the best correlation at depths of 0–10 m, indicating that sedimentary Uk’37 reflects mixed-layer temperatures. The authors combined their Uk’37 data with measurements from 370 sites across the Atlantic, Indian, and Pacific Oceans to create a global core-top calibration based on annual mean atlas SST (0–29°C) from 0 m water depth. Their final global core-top calibration was Uk’37 = 0.033T + 0.04, where T represents sea surface temperature in degrees Celsius.

Outstanding Questions, Paleothermometer Corrections, and New Alkenone-Based Proxies

Despite the widespread acceptance of the 1998 Müller calibration, researchers continue to refine alkenone-based biomarker methodologies. While the alkenone paleothermometer is generally reliable in most open-ocean environments, its accuracy may be limited in specific regions, including polar waters, the eastern equatorial Pacific, and Peru's upwelling zone. Factors such as warm- and cold-end bias, seasonal productivity bias, and El Niño effects could contribute to these limitations. To enhance the paleothermometer, new core-top Uk’37 calibrations have been introduced, including a non-linear, global, field-based calibration produced by Conte and colleagues. This calibration, based on over 600 current surface seawater Uk’37 measurements, is expressed as T(°C) = —0.957 + 54.293(Uk’37) — 52.894(Uk’37)^2 + 28.321(Uk’37)^3.

Efforts to address the apparent non-linearity of the alkenone paleothermometer calibration at extreme sea surface temperatures have also been made. One such effort is the 2018 BAYSPLINE paleothermometer, developed by Tierney and Tingley to explore questions regarding the seasonality of alkenone production and the response of alkenone unsaturation to high (>24°C) or low (near-freezing) temperatures. By analyzing over 1,300 core-top measurements, the authors concluded that seasonal production bias of alkenones likely affects paleotemperature reconstructions in the North Atlantic, North Pacific, and Mediterranean regions. They also noted a significant flattening of the Uk’37 response to SST above 24°C. To account for these findings, Tierney and Tingley proposed the BAYSPLINE Uk’37 paleothermometer, utilizing a B-spline fit with Bayesian inference. This model accommodates the reduced sensitivity of alkenone unsaturation to temperature in warmer waters and addresses production seasonality in different ocean basins. The authors claim that their BAYSPLINE calibration corrects past underestimations of tropical SST anomalies and captures uncertainty in the regression. However, this calibration does not account for potential slope attenuation at low temperatures and has yet to gain widespread acceptance, resulting in limited validation compared to established calibrations.

In addition to serving as a reliable paleotemperature proxy, researchers have shown that alkenones may also represent other open-ocean mixed-layer conditions. One of the earliest proposals was to use total alkenone concentration, measured as C37 total, as an indicator of paleo-productivity and marine organic carbon levels. More recent studies suggest that alkenone ?2H values could reflect past seawater hydrogen isotope proportions, independent of temperature and salinity. Other research has indicated that alkenones produced by Group 2i Isochrysidales, a distinct coccolith lineage, coexist with sea ice and may account for the high levels of C37 tetra-unsaturated methyl alkenones (%C37:4) found in high-latitude oceans. Consequently, %C37:4 shows promise for reconstructing historical sea ice distributions. Researchers have also proposed new alkenone-based proxies aimed at enhancing the existing paleothermometer. For instance, studies have demonstrated that the unsaturation index of di- and tri-unsaturated 38-carbon methyl-ketones could extend the reliable range of the alkenone marine paleothermometer to 0–29.5°C, enabling accurate temperature estimates in warmer waters and regions where Group 2 alkenone producers are present, as they do not produce C38 methyl ketones.

Concluding Thoughts

Since the identification of alkenones in the late 1970s, advancements in geochemical techniques and oceanographic research have fostered the rapid development and acceptance of the Uk’37-based paleothermometer as a reliable proxy for open-ocean mixed-layer conditions. While uncertainties remain regarding the method's limitations at extreme temperatures, seasonal biases in temperature reconstructions, and regional variations in alkenone production, alkenones have largely proven to be dependable indicators of nSST and hold the potential for further exploration of paleoceanographic conditions.

The above article was written by Brianna Hoegler, a second-year PhD student studying paleoceanography in the Department of Earth, Environmental, and Planetary Sciences at Brown University. In her graduate studies, Brianna utilizes alkenone paleothermometry to reconstruct sea surface temperatures during potential Pliocene (ca. 5.4–2.6 million years ago) glacial events. This piece was originally crafted as a review paper for an analytical geochemistry seminar and has since been adapted for online publication.