Institute of Oceanography’s contribution to NTU College of Science Research Output 2025

Understanding past ocean temperatures using geochemicals proxies embedded in marine sediments

Ocean temperature plays a fundamental role in Earth’s climate system. To study ocean temperatures beyond the instrumental era, scientists rely on paleotemperature proxies–geochemical signatures found in the remains of marine organisms embedded in sediments. However, accurately interpreting these proxies requires a thorough understanding of the uncertainties associated with their formation, transportation, and analysis. Four recent studies led by Associate Professor Sze Ling Ho’s Paleoproxy Lab at the Institute of Oceanography, National Taiwan University (IONTU), provide key insights into these uncertainties and contribute to refining the reconstruction of past climate conditions (Figure 1).

Study 1: GDGT proxies and methane influence¹

Assoc. Prof. Ho’s team focuses on isoprenoid glycerol dialkyl glycerol tetraethers (GDGTs), organic compounds from archaea used to estimate past ocean temperatures via TEX₈₆. Methane-related influence on TEX₈₆ is assessed using the GDGT-based Methane Index (MI), which also serves as a proxy for methane hydrate dissociation. To test the reliability of these proxies, researchers from IONTU collaborated with three Taiwanese institutions to analyze sediments from methane-rich cold seep sites offshore southwest Taiwan. Using a remotely operated vehicle (ROV), they obtained high-quality samples for gas, pore water, sediment, and biomarker analyses. The study found that methane-related chemistry influenced GDGT composition, and the duration of methane cycling activity affected MI values. Despite MI variations, TEX₈₆-based temperatures remained stable across sites (within 1.5°C) and aligned with modern data. This suggests that TEX₈₆ is reliable even when MI exceeds the conventional threshold. The study enhances understanding of the impact of methane cycling on marine sediment chemistry and proxy signals while showcasing Taiwan’s technical expertise in oceanographic sampling and analysis. This study was covered by TTV News (台視新聞台; https://www.youtube.com/watch?v=7vDJpNxd1QQ).

Study 2: Mg/Ca ratios in foraminifera and cleaning techniques²

Another key proxy is the magnesium-to-calcium (Mg/Ca) ratio in foraminifera shells. Traditionally, multi-specimen analysis is used, but single-specimen studies are increasingly needed to capture fine-scale variability. However, Mg/Ca measurements are prone to contamination, requiring cleaning steps that may alter shell chemistry and lead to sample loss. PhD student Pei-Ting Lee from the Paleoproxy Lab investigated how cleaning intensities affect Mg/Ca ratios across species and sample types. Aggressive treatments like heating and reductive steps lowered Mg/Ca ratios in a species-specific manner. When these steps were omitted, single- and multi-specimen values aligned more closely. This supports the application of calibration equations from multi-specimen data to single-specimen results, improving the reliability of sea surface temperature (SST) reconstructions and enhancing data comparability.

Study 3: Reproducibility of paleotemperature records³

Paleoclimate studies often assume that one sediment core reflects regional trends and that different proxies capture the same season and depth. To test these assumptions, PhD student Ru-Yun Tung from the Paleoproxy Lab and international collaborators analyzed three SST proxies, namely Mg/Ca, U^K’₃₇, and TEX₈₆, from four nearby cores in the northern Okinawa Trough. Despite different coring and analytical methods, each proxy type showed consistent glacial-interglacial trends, validating the assumption that local temperature signals are representative. However, abrupt shifts that did not appear in all cores likely reflect proxy noise or sediment heterogeneity. Foraminifera Mg/Ca records are more variable than U^K’₃₇ records, likely due to the limited number of foraminifera shells included in the analyses, in contrast to the tens of thousands to millions of alkenone molecules underpinning each U^K’₃₇ data point. Notably, TEX₈₆ more closely resembled intermediate-depth temperature records than surface temperatures in this region, suggesting that it may, in certain settings, reflect deeper waters. These results emphasize the need to understand each proxy’s depth and seasonal sensitivities for accurate reconstructions.

Study 4: Evaluating Climate Models with Proxy Data⁴

The tropical Pacific is crucial for global climate, especially through interannual variability in upper ocean temperature patterns associated with the El Niño-Southern Oscillation. Yet many models fail to reproduce observed 20th-century upper ocean temperature patterns in this region. To assess model performance under varying CO₂ levels, postdoctoral researcher Alicia Hou from Paleoproxy Lab compiled proxy-based SST estimates from past climate states and validated them using modern Argo float data. These proxies reliably reflected thermal changes during El Niño and La Niña events. The team compared these proxies with outputs from seven state-of-the-art climate models in the Paleoclimate Modelling Intercomparison Project (PMIP). The models consistently failed to capture cooling in the eastern tropical Pacific under low CO₂ conditions, as they overestimated temperature changes in that region. Glacial-interglacial proxy data also indicated that the western tropical Pacific may warm more rapidly than the eastern Pacific if CO₂ emissions continue. These results highlight the value of proxy data in validating and improving model accuracy, which is crucial for future climate projections.

Conclusions

Together, these studies illustrate how uncertainties in proxy formation, transportation, and analysis can influence paleotemperature records. Despite these complexities, first-order temperature trends are reproducible across nearby sediment cores, validating their use in regional climate reconstructions. Importantly, integrating paleotemperature proxies with model simulations offers a robust framework for improving climate projections, particularly in regions critical to global climate dynamics like the tropical Pacific. Ongoing efforts in refining proxy calibration, cleaning protocols, and multi-proxy comparisons are essential for advancing paleoclimate science and its applications to future climate forecasting.

Further reading

1.     Ho, S. L. et al. Methane Index and TEX₈₆ values in cold seep sediments: Implications for paleo-environmental reconstructions. Geochim et Cosmochim Acta 391, 262–276 (2025).
2.     Lee, P. ‐T. et al. Effect of Cleaning procedures on the Mg/Ca ratio of single‐specimen Planktic foraminifera. Geochem Geophys Geosyst 26, e2024GC011975 (2025).
3.     Tung, R.-Y. et al. Replicability of paleotemperature records in the northern Okinawa Trough and its implications for paleoceanographic reconstructions. Prog Earth Planet Sci 11, 61 (2024).
4.     Hou, A., Jonkers, L., Ford, H. L. & Ho, S. L. El Niño-like tropical pacific ocean cooling pattern during the last glacial maximum. Commun Earth Environ 5, 587 (2024).