Luminescence and ESR dating of the sedimentary infill from the multi-level cave system of Alkerdi-Zelaieta (Navarre, N Spain)

Abstract

The finding of Upper Paleolithic engravings in 2016 triggered a multidisciplinary investigation of the Alkerdi cave system (Urdazubi, N Spain). The study of the speleogenetic processes led to the identification of at least 6 paragenetic cave levels with associated sedimentary infill. In order to unravel the timing of changes in the karst dynamics and to get some insights about sediment origin, two sediment samples were collected from cave levels 4 and 1 for numerical dating purpose, using both Optically Stimulated Luminescence (OSL) and Electron Spin Resonance (ESR) methods. One additional modern-age sample was also taken from the entrance of the karst to evaluate the magnitude of the optical bleaching achieved by the quartz grains before entering the cave system. Last, one sample was collected from a flowstone for U-series dating, providing independent age control.

OSL measurements were carried out using small quartz multi-grain (MG) and single-grain (SG) Single Aliquot Regenerative-dose (SAR) protocol. Initial Equivalent Dose (De) results show evidence of saturated OSL signal for the sample of the higher level (with ∼60% of saturated aliquots). Unlike SA data, SG analyses do not meet the usual quality criteria (e.g., relatively high recuperation ratios, only a few grains with useable signal), suggesting that corresponding dating results should be treated with caution.

Hence, both MG and SG TT-OSL and MG ESR methods were subsequently employed, as the corresponding signals are known to have higher saturation levels than OSL ones. ESR analyses were based on the Multiple Centre (MC) approach using the standard multi-grain multi-aliquot additive (MAA) dose method, while TT-OSL measurements were performed following Demuro et al. (2020). The latter, however, did not return any useful results for both samples. MC ESR data show the usual De pattern (De(Al)>De(Ti op. D)>De(Ti–H)) indicating that among the three ESR signals analyzed, the Ti–H most likely provides the closest estimate to the true burial age. However, low measurement repeatability and goodness-of-of fit indicate that the reliability of the ESR results may be reasonably questioned.

Results derived from this multi-technique dating approach provide the first chronological constraints for the sedimentary infill of the Alkerdi cave system. In particular, they suggest that sediment deposition in Cave Level 4 (ALK-OSL01) occurred around 130 ka, which is in good agreement with the minimum age constraint given by the speleothem (80.5 ± 9.0 ka). In comparison, the young age obtained for sample ALK-OSL02 (5.0 ± 0.9 ka) suggests recent Holocene formation of the lowest cave level 1.

Finally, the MG De value obtained for the modern sample is close to zero (<1 Gy), indicating that the OSL signal is almost fully reset at a multi-grain level before entering the cave. Additionally, despite the large De overdispersion measured in this modern sample, the De values are one to two order of magnitude lower than those obtained with the same model in the other two samples. Therefore, the large OSL De overdispersion (OD) values of up to 64% obtained for the two samples from Cave Level 4 and 1 are most likely related to re-sedimentation processes inside the cave system.