Unveiling the early Eurasian glaciations with new advances of 10Be-26Al burial dating

Abstract

Cosmogenic nuclide burial dating exploits the differential decay rate of a nuclide pair, typically 26Al and 10Be. There are three basic approaches; simple burial dating, the isochron method and P-PINI (Particle Pathway Inversion of Nuclide Inventories). The simple burial age is based on deviation from the surface production ratio assuming quick and deep burial. The isochron method incorporates multiple samples from a single stratigraphic layer and assumes steady erosion in the sediment source area. P-PINI accommodates more complicated situations, such as landscapes undergoing non-steady erosion and an eroding or accumulating sink zone. This source-to-sink model uses a Monte Carlo simulator to create a library of plausible erosional and depositional histories and then compares those histories against measured nuclide data to derive a burial age estimate. In cases involving multiple dated layers in a section or when an independent age estimate is available, a Bayesian tool further constrains the burial age. With the aim of constraining the timing of the Eurasian Ice Sheet at its largest known extent, we applied P-PINI at Wapenveld in the Netherlands and at Szczerców in central Poland. From a drill-core at Wapenveld we collected three sand samples from the Early Pleistocene ‘Hattem’ beds, and derived the P-PINI burial age bracketed by simple burial ages calculated from the units above and below. Szczerców is an open-cast mine exposing multiple Middle Pleistocene till units rich in cobble-sized clasts. Here, P-PINI burial ages show sensitivity to the thickness of the dated layer, a problem we resolve by including feldspar infrared-simulated luminescence dates higher up the sequence. We discuss the pros and cons of our approach at Wapenveld and Szczerców. Burial dating multiple units in section requires considerably more lab effort but also offers important advantages that can be boosted by carefully considering which grain-size to sample, and by incorporating independent age constraints.