Chronological reference datasets: reasoning, creation and application

  • Sebastian Kreutzer (Siaradwr)

Gweithgaredd: Sgwrs neu gyflwyniadCyflwyniad llafar


Methodological progress in trapped-charge dating methods hardly ever leads to reduced complexity of their applications. Advances usually require additional measurement or protocol parameters to be monitored, new tests to be performed, often on top of multiple signals and models to compare. These requirements increase the number of checks made, while the effect on the confidence in the dating results may remain blurry at best. Various software tools released over the years, alongside new methodological developments, support data treatment and help to visualise results [e.g., 1,2].
However, in recent years, software tools themselves have enabled significant methodological progress [e.g., 2,3]. This development is fuelled by free and advanced open-source software frameworks, the open-access data movement, and abundantly available computation power (for instance, allowing for sophisticated statistical modelling). Unfortunately, this situation introduces new challenges for a dating study, and multiple analytical pathways seem to present themselves when solving a single problem. Understanding the merits and differences of different tools is challenging. Finding the origin of diverging results in complex analytical pathways is time-consuming at best and often impossible even to experts without scrutinising the software code.

Reference datasets are collections of well-known static, artificially generated or measured data, explicitly created as test instances. By providing a common basis for evaluating analytical procedures, these datasets help maintain trust and confidence in the dating process. Moreover, they allow one to test model parameters and benchmark their output without in-depth knowledge of the tool.

Our contribution addresses the reasoning, creation and application of (chronological) reference datasets in the context of luminescence dating. We present sets of reference data generated using the statistical programming environment R and available modelling packages, such as ‘RLumModel’ [4] or ‘RLumCarlo’ [5]. We show how these datasets can be applied to test standard models and assumptions used in luminescence dating studies.


[1] Duller, G.A.T., 2015. The Analyst software package for luminescence data: overview and recent improvements. Ancient TL 33, 35–42.

[2] Kreutzer, S., Schmidt, C., Fuchs, M.C., Dietze, M., Fischer, M., Fuchs, M., 2012. Introducing an R package for luminescence dating analysis. Ancient TL 30, 1–8.

[3] Philippe, A., Guérin, G., Kreutzer, S., 2019. BayLum - an R package for Bayesian analysis of OSL ages: An introduction. Quaternary Geochronology 49, 16–24.

[4] Friedrich, J., Kreutzer, S., Schmidt, C., 2016. Solving ordinary differential equations to understand luminescence: “RLumModel” an advanced research tool for simulating luminescence in quartz using R. Quaternary Geochronology 35, 88–100.

[5] Kreutzer, S., Friedrich, J., Pagonis, V., Laag, C., Rajovic, E., Schmidt, C., 2021. RLumCarlo: Simulating Cold Light using Monte Carlo Methods. The R Journal (accepted article). 1–15.
Cyfnod16 Medi 2021
Teitl y digwyddiad16th International Luminescence and Electron Spin Resonance Dating conference (LED2021)
Math o ddigwyddiadCynhadledd
Graddau amlygrwyddRhyngwladol