Dating Middle Pleistocene loess from Stari Slankamen (Vojvodina, Serbia) — Limitations imposed by the saturation behaviour of an elevated temperature IRSL signal

Recent advances in post-IR IRSL dating have led to breakthroughs in dating upper Middle Pleistocene loess sequences. Here, an elevated temperature post-IR IR protocol using a second IR stimulation temperature of 290 °C is applied to eleven polymineral fine-grain (4–11 μm) samples from the lower part of the Middle–Lower Pleistocene Stari Slankamen loess–palaeosol sequence with the aim of refining the siteʹs age model and investigating the behaviour of both the IR50 and the pIRIR290 signals in material close to or in luminescence signal saturation. Both signals from the 8 samples below the prominent erosion surface represented by a gravel layer at ~ 13 m were found to be in field saturation (i.e. no systematic increase in burial dose with depth). The average ratio of the sensitivity-corrected natural signal to the laboratory saturation level for the pIRIR290 is 1.00 ± 0.03 (n = 8) indicating that field saturation is equal to laboratory saturation for this signal. Minimum equivalent dose estimates were calculated from 2*D0 values, giving minimum age estimates of ~ 230–390 ka; this result suggests an upper limit for dating these loess deposits of ~ 300 ka. The age estimate of the younger sample SSK2 is in good agreement with the quartz OSL age, showing that the pIRIR290 signal is bleachable in nature and can be used to date material as young as ~ 20 ka. Our data suggest that the loess unit V-L2 accumulated during Marine Isotope Stage (MIS) 6 and that the erosional event marked by an unconformity and gravel layer between loess units V-L2 and V-L3 has a minimum age of ~ 170 ka. Pedocomplex V-S1 can be correlated with the complete MIS 5 period. Furthermore we suggest a minimum age of ~ 250 ka for the loess unit V-L3 and palaeosols V-S3, V-S4 and V-S5; considerably older than proposed by many previous studies.