Tropical temperature evolution across two glacial cycles derived from speleothem fluid inclusion microthermometry

Abstract. The evolution of tropical temperature across multiple glacial-interglacial cycles is mostly constrained with marine proxy records, which are associated with considerable uncertainties. Here we present a reconstruction of tropical land temperatures derived from fluid inclusions in stalagmite WR5_B from Whiterock Cave (Gunung Mulu National Park, Northern Borneo). The employed paleothermometer – nucleation-assisted microthermometry – is based on the density of the water trapped in fluid inclusions, i.e., on a well-known thermodynamic parameter and yields highly precise temperature estimates. The record consists of 49 temperature data points spanning a 127 kyr period from 460 to 333 ka including the glacial terminations T-V and T-IV. We find that Borneo temperature tracks Southern hemisphere temperature and atmospheric CO₂ concentrations. Deglacial warming is accompanied by relatively dry conditions in Northern Borneo, indicated by pronounced enrichments in calcite δ¹⁸O_cc and reconstructed drip water δ¹⁸O_dw values. The amplitude of glacial-interglacial temperature changes amounts to 4.2 ± 0.4 °C (2SEM) between MIS 12 and the MIS 11 interglacial optimum and 4.3 ± 0.4 °C (2SEM) across T-IV. MIS 11 peak temperature was found to be 0.9 ± 0.4 °C warmer than late Holocene temperatures reconstructed for Whiterock Cave, whereas temperatures during MIS 12 and MIS 10 glacial maxima in our record are indistinguishable from those previously reconstructed for the Last Glacial Maximum. Both the present WR5_B record as well as the recently published record from Løland et al. (2022) covering the last glacial Termination exhibit a clear linear correlation with Antarctic temperature anomalies (R² = 0.89 and 0.97, respectively), with practically identical slopes of the linear regression lines. Depending on the employed Antarctic ΔT reconstruction, Landais et al. (2021) and Jouzel et al. (2007), we found a polar amplification factor of 2.21 ± 0.22 and 2.42 ± 0.23 (95 % CI), respectively.