Multi-proxy temperature records from a northern Borneo stalagmite reveal sample-specific challenges

Over the last decades, several new methods for quantitative paleo-temperature reconstructions with stalagmites have emerged, further enhancing the value of these powerful paleoclimate archives. Among these innovative stalagmite-based thermometers, fluid inclusion microthermometry (Krüger et al., 2011) is often regarded as the most precise and accurate method (Meckler et al., 2015), but its applicability is restricted to formation temperatures > 10 °C, specific calcite fabrics, and abundant fluid inclusions of appropriate size. Fortunately, other temperature proxies have been proposed that each have different strengths and weaknesses, allowing us to compensate for the limitations of individual methods. Many of them, including fluid inclusion water isotopes and TEX₈₆, are still under active development, with substantial uncertainties remaining in their interpretation (e.g., Affolter et al., 2025; Baker et al., 2019). Applying multiple temperature proxies to the same stalagmite allows a direct comparison of proxy behavior, providing improved constraints on the reconstructed paleoclimate variability.

 

In this study, we reconstruct tropical temperature using stalagmite GC08 from northern Borneo, which spans multiple glacial cycles. Here we investigate the oldest part from approximately MIS 14 to MIS 11 (ca. 570 ka to 360 ka), which covers the Mid-Brunhes Transition (MBT; Yin, 2013) at the end of MIS 12. The MBT marks a fundamental change in the climate system with a significant increase in the amplitude of the glacial-interglacial cycles observed in various climatic archives (e.g., Barth et al., 2018). We use three different temperature proxies for the reconstruction: fluid inclusion microthermometry, fluid inclusion water isotopes, and TEX₈₆. Our records reveal notably different temperature trends among the three proxies. Both fluid inclusion microthermometry and TEX₈₆ indicate surprisingly little temperature change across the study interval. We note that the fabric is not ideal for fluid inclusion microthermometry, as large intervals are characterized by biogenic influence and/or diagenesis, which limit the applicability and accuracy of the method. TEX₈₆ seems to be influenced by soil-derived compounds in part of the stalagmite. Fluid inclusion water isotopes appear to be affected by large evaporation or other fractionating effects, as indicated in a cross-plot of oxygen and hydrogen isotopes. Correction attempts do not yield realistic temperatures, with an unrealistically large amplitude of 20 °C. These findings highlight the limitations of individual stalagmite-based paleo-thermometers and emphasize the critical role of depositional context in their interpretation. We therefore call for caution when interpreting single-proxy temperature evidence in the absence of constraints on in-cave processes in future stalagmite-based paleo-temperature studies.