Cold spells in the Nordic Seas during the early Eocene Greenhouse

The early Eocene (56.0–33.9 Ma) was a period of intense global warmth, characterized by greenhouse conditions that extended from the Mesozoic into the earliest Cenozoic. While a significant body of evidence supports globally warm conditions throughout this epoch, including the absence of polar ice, reports of glacial sediments and glendonites (calcite pseudomorphs after ikaite, CaCO3·6H2O) from high latitudes suggest that episodic cooler intervals may have punctuated this overall warmth. This presents a significant research challenge: reconciling the evidence for a globally warm early Eocene with localized indications of significantly colder conditions. Glendonites, traditionally viewed as cold-climate indicators due to ikaite's temperature dependence, have been found in early and mid-Eocene deposits across various locations, including North America (paleolatitude ~55°N), the Arctic (~77°N), and Denmark (~45°N). Their presence in these regions, frequently associated with glacial deposits, has raised questions about the stability of ikaite and the accuracy of using glendonites as reliable paleotemperature proxies. The ability to grow and stabilize synthetic ikaite at 35°C has further complicated interpretations based solely on the presence of glendonites. The Fur Formation in Denmark, known for exceptionally well-preserved marine and terrestrial fossils, provides a compelling case study. Its flora and fauna suggest local tropical to subtropical climates, yet the Fur Formation contains numerous large glendonites, a seeming contradiction. To definitively address this contradiction, precise quantitative temperature reconstructions of early Eocene glendonites were needed. This study utilizes clumped isotope thermometry—a powerful technique for determining past temperatures—applied to glendonites from the Fur Formation to quantitatively reconstruct bottom water temperatures and resolve the apparent climatic paradox.

Previous studies investigating the early Eocene climate have yielded conflicting results. While evidence supports globally warm conditions, the presence of high-latitude glacial sediments and glendonites suggested the possibility of episodic cooling. Spielhagen and Tripati (2009) presented evidence from Svalbard indicating near-freezing temperatures and climate oscillations in the Arctic during the Paleocene and Eocene. However, the significance of glendonites as paleotemperature indicators has been debated due to the discovery that synthetic ikaite can be grown at temperatures higher than previously thought (Marland, 1975; Tollefsen et al., 2020). Studies on the Fur Formation in Denmark have documented the presence of abundant glendonites within a diatomite sequence (Pedersen et al., 2012; Schultz et al., 2020), contradicting the seemingly warm climate suggested by the associated flora and fauna. This discrepancy highlights the need for quantitative temperature data to resolve the uncertainties surrounding glendonite formation and their role as climate indicators. Other studies investigating early Eocene temperatures have utilized various proxies, including Mg/Ca ratios in foraminifera (Evans et al., 2018), and TEX86 derived from GDGTs (Stokke et al., 2020). These studies provide regional and global temperature estimates, but lack the site-specific detail and direct temperature data from glendonites which are needed to understand the local climatic conditions within the Danish Basin during this time.

This study utilizes a multi-proxy approach integrating clumped isotope thermometry, stable isotope analysis, minor element analysis, and biomarker analysis to reconstruct paleotemperatures and environmental conditions in the early Eocene Danish Basin. Glendonites and associated carbonate concretions from the Fur Formation were carefully sampled, with particular attention to separating different calcite phases within the glendonites. Scanning Electron Microscopy (SEM), cathodoluminescence (CL), and light microscopy were used to characterize the glendonite morphology and identify different calcite phases. Stable isotope (δ¹³C and δ¹⁸O) and minor element (Mg, Mn, P, S, Sr, Fe) analyses were performed on both glendonites and concretions using ICP-OES and Isotope Ratio Mass Spectrometry. Clumped isotope thermometry, employing the Δ47 method, was applied to distinct calcite phases within the glendonites (Type I, ikaite-derived calcite; Type III, late-stage sparry calcite) and to the concretionary calcite (Type II). This technique provided quantitative temperature estimates for the time of calcite formation. Biomarker analysis was conducted on glendonites, concretions, and sediment samples to reconstruct sea surface temperatures (SSTs) using TEX86 (using the BAYSPAR calibration), and to assess the influence of freshwater input and the possibility of methane seep activity. The Long Chain Diol Index (LDI) and UK37 were also considered, but their reliability was assessed given the potential influences of coastal and freshwater environments. Detailed examination of the geological setting of the Fur Formation, including its stratigraphy and paleogeography, was also considered to provide context for the paleotemperature reconstructions. A thorough quality control procedure was performed using sub-samples from Recent carbonates, using the same methodology, to test whether glendonite calcite is reliable for clumped isotope thermometry.

Clumped isotope thermometry yielded average temperatures of 0.9 °C (±4.7 at the 95% confidence level) and 9.1 °C (±3.7) for Type I calcite from the +15 and +60–62 glendonite horizons, respectively. Type III calcite yielded temperatures of 8.5 (±5.1) °C and 13.6 (±4.0) °C for the same horizons, while the concretionary calcite showed mean temperatures of 4.5 (±3.9) °C and 10.3 (±8.3) °C. The near-freezing temperatures from Type I calcite represent bottom-water temperatures, indicating significantly cooler conditions than previously documented for the early Eocene. The higher temperatures for Type III calcite suggest that this sparry calcite grew at a later stage, after the concretion formed around the newly transformed glendonite. Biomarker data from concretions provided SST estimates of 13–15 °C (+15 horizon) and 8–12 °C (+60–62 horizon). Sediment biomarker data exhibited anomalous distributions due to significant terrestrial organic matter input. High P and Mg concentrations in glendonites suggest that these elements played a crucial role in inhibiting calcite formation and promoting ikaite precipitation. The δ¹³C signatures suggest that bacterial sulfate reduction, rather than methane seep activity, likely generated the conditions favorable for ikaite growth. Isotopic analysis revealed that the δ¹⁸O of the calcite reflected early diagenetic conditions, with a potential negative shift due to freshwater influence and diagenetic alteration of volcanic detritus. Comparison of the results from the glendonites and the carbonate concretions, considering the uncertainties, indicates a common early diagenetic phase of calcite precipitation at temperatures consistent with modern marine sedimentary ikaite.

The findings of this study reveal remarkably low bottom-water temperatures in the Danish Basin during the early Eocene, contradicting global temperature reconstructions for this period. The reconstructed near-freezing bottom-water temperatures (approximately 5 °C) along with the higher SSTs derived from concretions (8–15 °C), indicate a significant temperature gradient in the water column. These cold conditions appear to be a regional phenomenon, not reflected in global temperature reconstructions. The proximity of the Danish Basin to the North Atlantic Igneous Province (NAIP) suggests a possible link to volcanic activity. Explosive eruptions could have released sulfur aerosols, leading to regional cooling through atmospheric processes, resulting in much colder winters and triggering density-driven cascading events that brought cold water to the basin floor and facilitated ikaite formation. The paleogeography of the region, particularly the semi-enclosed nature of the Danish Basin, might also have promoted a bias towards colder winter temperatures, facilitating the observed conditions.

This study provides compelling evidence for regionally significant cooling events in the Nordic Seas during the early Eocene greenhouse. The observed near-freezing bottom-water temperatures, reconstructed using clumped isotope thermometry on glendonites from the Fur Formation, highlight the importance of regional climatic variability during the early Eocene. These findings challenge the assumption of uniformly warm conditions throughout the epoch and suggest that volcanic activity from the NAIP played a crucial role in driving regional climatic fluctuations. Further research should focus on higher-resolution studies across glendonite-bearing horizons within the NAIP ash layers to assess the magnitude and frequency of such cool episodes, and high-resolution global climate model (GCM) simulations to better understand the mechanics of regional cooling events in stratified basins.

The study's conclusions are based on a limited number of sample sites within the Fur Formation, potentially limiting the generalizability of the findings to the broader Nordic Seas region. The interpretation of the δ¹⁸O values is complicated by early diagenetic alterations and potential freshwater influence, introducing some uncertainty in the reconstruction of the original bottom-water conditions. Additionally, while the study provides compelling evidence linking regional cooling to NAIP volcanism, a direct causal relationship requires further investigation through more comprehensive modeling efforts and analysis of additional geological records. The reliability of the used biomarker proxies is influenced by the specific depositional environments, which is always a concern when applying these proxies in coastal settings.