Calcite Mg and Sr partition coefficients in cave environments: Implications for interpreting prior calcite precipitation in speleothems

Trace element to Ca ratios in speleothems have emerged as important proxies that reflect local environmental conditions. However, interpretations of speleothem trace element records can be challenging due to various processes. Positive correlations between speleothem Mg/Ca and Sr/Ca have often been interpreted to reflect prior calcite precipitation (PCP), a process potentially modulated by rainfall variability. For quantitative interpretation of PCP, the distribution coefficients for Mg and Sr (D-Mg and D-Sr) are required. Here, we use ten cave monitoring calcite and drip water datasets to investigate the influence of temperature and drip water and calcite Mg/Ca and Sr/Ca ratios on speleothem calcite D-Mg and D-Sr. The datasets cover a large range of climatic and geological settings resulting in a large range of drip water Mg/Ca ratios. Speleothem calcite D-Sr shows a positive correlation with the calcite Mg/Ca ratio. Furthermore, D-Mg shows a clear temperature dependence (D-Mg = 0.013*e(0.035)*(T)). Previous work proposed that the slope of a trend line through a plot of ln(Sr/Ca) versus ln(Mg/Ca) of a speleothem trace element dataset is between 0.709 and 1.003 if dominated by PCP. However, this only holds true if the initial drip water Mg/Ca and Sr/Ca ratios as well as D-Sr and D-Mg are constant for the whole dataset. We use an excel-based PCP model (see Electronic Annex) to assess the potential influence of PCP on drip water and speleothem Mg/Ca and Sr/Ca ratios and simulate different initial drip water Ca, Mg, and Sr concentrations corresponding to limestone, dolostone, and mixed host rock compositions. In the case of enhanced PCP and high Mg/Ca ratios, calcite D-Sr increases progressively with the mean Mg/Ca ratio of the speleothem time series resulting in steeper slopes of ln(Sr/Ca) versus ln(Mg/Ca) of up to 1.45. We show that PCP can induce slopes ranging from 0.709 (or even shallower) up to 1.45. This large range suggests that the previously applied criteria to detect PCP in speleothem records were too strict and may lead to unjustified exclusion of PCP as a potential interpretation of speleothem and drip water trace element ratios. Thus, the number of speleothem Mg/Ca and Sr/Ca datasets that potentially reflect past changes in effective rainfall may be larger than previously suggested. (C) 2019 Elsevier Ltd. All rights reserved.