Utilising a portable luminescence reader or rapid age assessment in aeolian sedimentary environments.

Reconstruction the accumulation of sediment at the surface requires geochronological data capturing a wide spatial area and the full temporal range of the preserved record. Collecting such datasets using traditional laboratory-based luminescence dating is extremely time and resource intensive. For example, it has taken from the late 1980s to 2016 to produce the global dataset of ~4000 luminescence ages for desert dunefields1,. Larger datasets, that cover a wide geographical areas to reconstruct dunefield-scale aeolian dynamics helps us to produce a temporally-constrained proxy record from which we can elucidate the mechanisms driving changes in key climate variables, such as precipitation, moisture balance and the wind. Quaternary proxy datasets are also valuable for testing the robustness of model simulations that can be used to predict future change and remobilisations2,3.

Portable luminescence readers (POSL) have an excellent track record of producing relative age information to aid the in situ interpretation of sites and guide targeted field sampling, as well as assisting in informing strategies of sample selection for full-laboratory-based dating approaches. The research presented here extends the utility of the POSL approach to producing rapid age assessments, based on calibrating the POSL signals of sampled which also have published ages using full laboratory-based luminescence dating protocols4. This is based on 144 samples within southern Africa, and represents the first comprehensive approach to do this worldwide, following a pilot approach for contrasting age samples in the Namib Sand Sea5.

This new approach demonstrates that there is no one regression model for all southern African aeolian sediments and that a region-specific approach to calibration is needed, with four broad regions identified: the Namib Sand Sea, northern Kalahari, western Kalahari and southern Kalahari (Figure 1). The strength of linear regressions (r2 of 0.99, 0.93, 081 and 0.52) for samples back to ~120, ~105, ~75 and ~5ka for each region respectively, coupled with the errors on predicted ages (RMSE) generated using k-fold cross validation, demonstrates that this is an excellent approach to rapid age assessment. The main driver of these regional differences is the quartz:feldspar ratio in the bulk sediment (measured using the POSL).

References

[1] Lancaster N. et al. (2016) Quaternary International 410, 3-10. https://www.dri.edu/inquadunesatlas
[2] Thomas, D, et al. (2005) Nature 435, 1218–21
[3] Mayaud, J. et al. (2017) Nature Scientific Reports, 7: 3887. DOI:10.1038/s41598-017-04341-0
[4] Stone, A. et al. (2018) Quaternary Geochronology (in press)
[5] Stone, A, et al. (2015) Quaternary Geochronology 30, 134-140.