|Sažetak (engleski)|| |
The Đurdevac Sands are named after its informal type-locality near the town of Đurdevac, where they form a unique terrestrial dune landscape that covers Late Pleistocene loess deposits and Holocene river Drava fluvial and associated marshy deposits (Galović and Posilović, submitted). Based on the superposition principle, Hećimović (1989) concluded that the Đurdevac Sands must be Holocene in age. However, recent 14C dating results of charcoal found in palaeosoils within the dune sands suggest that they were, at least partially, deposited during the Late Pleistocene (Galović and Posilović, submitted). The aim of the current study is to verify those recent findings using optically stimulated luminescence (OSL) dating of the aeolian sands that is covered by and covering the palaeosoils. OSL dating is proven to be a reliable technique to determine the burial ages of Late Pleistocene and Holocene aeolian sands in similar landscapes in northwestern Europe (Vandenberghe et al., 2013). Samples were taken in an abandonned sand quarry, Draganci, in the town of Kalinovac (Figure 1). The quarry is carved out into what seems to have been a massive (hummocky) dune, bordering a partially degraded large-scale parabolic dune created by northern winds. Samples were taken after clearing the excavation wall, turning it partially into a staircase. Stainless steel cylinders were hammered into each of the steps. After removing the steel cylinders and taping the caps until fixed, surrounding sediment was sampled for determining the radioactivity concentrations. After removing the outer material from the cylinders (dark room lab), a portion of each sample was subjected to a 10% HCl, a 15% H2O2 treatment, a 2.77 g/cm³ polytungstate treatment, and, finally, a 45% HF treatment in order to purify the quartz. Equivalent doses were determined on individual aliquots using the 100- 200 µm quartz fraction fixed onto disks using a 4 mm mask. A single aliquot regenerative (SAR) dose procedure (Murray and Wintle, 2003) was used, stimulating the samples for 40s with blue light diodes. A preheat of 240°C was used, and early background subtraction was applied to isolate the fast component. In addition, after drying, 500 g of surrounding sediment was analysed using high- resolution gamma-spectrometry. Dose rates were calculated taking into account beta and gamma contributions from all relevant radionuclides, a fixed internal dose and a depth-dependent cosmic dose contribution. The preheat plateau test, applied on one sample only, did not show any significant dose dependency on preheat, even though the individual data points showed significant scattering. The dose recovery test for the a cutheat/preheat combination of 200°C/240°C yielded a recovery ratio of 1.00 ± 0.02 (23 aliquots) and an overdispersion of 9.0% (± 0.4%). The resulting OSL ages are shown in Figure 1c-d. Overdispersion ranged from 42% for the second lowermost sample to 10% for the uppermost sample. However, given the fact that the distributions are rather symmetrical, and there is no reason to assume significant incomplete bleaching in this depositional setting, the ages are simply based on a central age model. The two lowermost samples yield OSL ages of around 14.5 ka, with a rather large uncertainty, and seem to overestimate the radiocarbon age of the lower palaeosoil (LPS), which is around 12.4 ka BP (Galović and Posilović, submitted). The two uppermost samples, however, seem to be in agreement with the radiocarbon age of ca. 11.7 ka BP for the uppermost palaeosoil (UPS), and seem to confirm the position of the Pleistocene-Holocene boundary in this sediment- soil archive on top of the UPS. A more detailed OSL sampling scheme will be set out in the near future, in the framework of the ACCENT project (Galović et al., submitted), in conjuction with (palaeo)pedological, geomorphological and geophysical techniques to elucidate the evolution of the Đurdevac landscape during the later Pleistocene and the Holocene, against a background of climate, land use and socio-economic change.