Short-term soil formation events in last glacial east European loess, evidence from multi-method luminescence dating

Here we provide a robust luminescence chronology for Stayky (Ukraine), a reference profile in European Late Pleistocene loess stratigraphy, based on optically stimulated luminescence (OSL) dating on quartz (4-11 mu m, 63-90 mu m) and post infrared-infrared stimulated luminescence (pIR-IRSL)) on polymineral fine grains. For the Bug loess unit, the equivalent of Marine Isotope Stage (MIS 2), results are in agreement between methods, demonstrating that the suite of embryonic soils previously interpreted as reflecting climate variability similar to Greenland interstadials (GI) actually date to similar to 29/27-15 ka, with most emplaced around or after 20 ka. This temporal span is further confirmed by age-depth modelling of available data. Apart from GI-2, no interstadial-type climate events are recorded in Greenland ice core data for that time interval. As short-term pedogenetic phases are also documented in records from central-western Europe, there is a need for more research into the European mid-latitude terrestrial environments response to MIS 2 hydroclimate variability. The dating of Vytachiv paleosol, previously debatably linked to various GI events within MIS 3 resulted in ages of similar to 40 +/- 4 ka and similar to 53 +/- 4 ka at the lower transition, and similar to 26 +/- 2 ka to similar to 30 +/- 2 ka in the overlying loess. These ages indicate that the truncated Vytachiv paleosol is either not continuous, or that it encompasses a broader age range within MIS 3 than previously considered. In both cases, data would not allow for an unambiguous linking of this paleosol with specific GI events as previously attempted. The pIR-IRSL290 dating of the loams immediately underneath Pryluky unit in the range of similar to 120 ka to similar to 168 ka and of the Pryluky mollisol from similar to 90 ka to 126 ka confirm the broad correspondence of this unit with MIS 5, although poor dose recovery results open the possibility for further testing on the degree these ages provide overestimated results. Quartz data severely underestimate the pIR-IRSL290 ages for these samples. The application of pIR-IRSL290 dating for the underlying Dnieper till previously linked to the Saalian glaciation resulted in natural signals at the level of laboratory saturation, yielding minimum ages of c. 700 ka. For the same sample, the natural SAR-OSL signals for 4-11 mu m quartz were found significantly below laboratory saturation level, resulting in finite ages of similar to 250-270 ka interpreted here as underestimates, while coarse quartz (63-90 gm) signals reached about 85% of the laboratory saturation level. These data suggest extreme caution must be taken when dating such old samples using quartz OSL. Results from our high-resolution luminescence dating raises important implications for the chronological representativeness of Stayky as a key loess site in Eastern Europe beyond MIS 2. (C) 2018 The Authors. Published by Elsevier Ltd.