Extending the ESR and OSL dating comparison on coastal dune deposits from the Wilderness-Knysna area (South Africa)

Abstract

Our work follows up on the initial methodological ESR dating study by Ben Arous et al. (2022) on several quartz samples from the Plio-Pleistocene to Holocene aeolian coastal dune deposits of the Wilderness-Knysna area (South Africa) that were previously dated by optically stimulated luminescence (OSL). Here, we extend this first ESR-OSL comparison with five additional optically-bleached quartz samples. We used the Multiple Aliquot Additive Dose (MAAD) method to specifically (i) evaluate the influence of the irradiation dose steps on the determination of low De values (<100 Gy) and (ii) obtain finite ESR ages for older samples that sometimes show saturated OSL signals. Following the Multiple Centre ESR dating approach, the Aluminium (Al) and Titanium (Ti) signals (Ti-H and Ti-Li-H, the latter resulting from a mixture of Ti centres) were systematically measured in all samples, and resulting De values and age estimates were compared with the corresponding OSL data. Our results show that for young samples (<50 ka) showing De values of a few tens of Grays (Gy), the use of smaller irradiation steps spaced by <100 Gy has a noticeable impact on the MAAD dose evaluation from the Ti centres, usually leading to De results closer to the expected values (for 2/3 samples). However, this also makes the ESR measurements somewhat more challenging, with higher experimental uncertainties, lower measurement repeatability and lower goodness-of-fit resulting from the relatively weak Ti ESR intensities, ultimately impacting the robustness of the ESR data collected. In this sense, our study illustrates the limitations of the ESR method to detect very low dose irradiation values <30 Gy with our experimental conditions (i.e., using MAAD procedure, a standard resonator and a measurement temperature of ∼90 K). On the contrary, it also highlights the greater potential of the Ti-H signal to date Late Pleistocene samples, confirming previous studies. Moreover, our results suggest that the transport and bleaching conditions of these aeolianite deposits may not be ideal for the reset of the radiation-induced Al and Ti ESR signals, which is consistent with the very few studies specifically focused on this type of samples, but contrasts with other previous dating applications centered on fluvial environments. Finally, we provide additional chronological constraints to the Landward barrier complex and Coversands deposits, two of the oldest formations in the Wilderness-Knysna area.