A comparison of different methods for speleothem age modelling

Abstract

Speleothems, such as stalagmites and flowstones, can be dated with unprecedented precision in the range of the last 650,000 a by the 230Th/U-method, which is considered as one of their major advantages as climate archives. However, a standard approach for the construction of speleothem age models and the estimation of the corresponding uncertainty has not been established yet. Here we apply five age modelling approaches (StalAge, OxCal, a finite positive growth rate model and two spline-based models) to a synthetic speleothem growth model and two natural samples. All data sets contain problematic features such as outliers, age inversions, large and abrupt changes in growth rate as well as hiatuses. For data sets constrained by a large number of ages and not including problematic sections, all age models provide similar results. In case of problematic sections, the algorithms provide significantly different age models and uncertainty ranges. StalAge, OxCal and the finite positive growth rate model are, in general, more flexible since they are capable of modelling hiatuses and account for problematic sections by increased uncertainty. The spline-based age models, in contrast, reveal problems in modelling problematic sections. Application to the synthetic data set allows testing the performance of the algorithms because the ‘true’ age model is available and can be compared with the age models. OxCal and StalAge generally show a good performance for this example, even if they are inaccurate for a short section in the area of a hiatus. The two spline-based models and the finite positive growth rate model show larger inaccurately modelled sections.