Hydrochemical controls on aragonite versus calcite precipitation in cave dripwaters

Abstract

Despite the paleoclimatic relevance of primary calcite to aragonite transitions in stalagmites, the relative role of fluid Mg/Ca ratio, supersaturation and CO32− concentration in controlling such transitions is still incompletely understood. Accordingly, we have monitored the hydrochemistry of 50 drips and 8 pools that are currently precipitating calcite and/or aragonite in El Soplao and Torca Ancha Caves (N. Spain), investigating the mineralogy and geochemistry of the CaCO3 precipitates on the corresponding natural speleothem surfaces. The data reveal that, apart from possible substrate effects, dripwater Mg/Ca is the only obvious control on CaCO3 polymorphism in the studied stalagmites and pools, where calcite- and aragonite-precipitating dripwaters are separated by an initial (i.e. at stalactite tips) Mg/Ca threshold at ≈1.1mol/mol. Within the analyzed ranges of pH (8.2–8.6), CO32− concentration (1–6mg/L), supersaturation (SIaragonite: 0.08–1.08; SIcalcite: 0.23–1.24), drip rate (0.2–81drops/min) and dissolved Zn (6–90μg/L), we observe no unequivocal influence of these parameters on CaCO3 mineralogy. Despite the almost complete overlapping supersaturations of calcite- and aragonite-precipitating waters, the latter are on average less supersaturated because the waters having Mg/Ca above ∼1.1 have mostly achieved such high ratios by previously precipitating calcite. Both calcite and aragonite precipitated at or near oxygen isotopic equilibrium, and Mg incorporation into calcite was consistent with literature-based predictions, indicating that in the studied cases CaCO3 precipitation was not significantly influenced by strong kinetic effects. In the studied cases, the calcites that precipitate at ∼11°C from dripwaters with initial Mg/Ca approaching ∼1.1 incorporate ∼5mol% MgCO3, close to the published value above which calcite solubility exceeds aragonite solubility, suggesting that aragonite precipitation in high-relative-humidity caves is favored from a solubility viewpoint. We also show that unaccounted CaCO3 precipitation in intermediate sampling containers and splash effects may in cases result in underestimating dripwater Ca concentration and alkalinity, potentially leading to incorrect conclusions regarding the role of fluid Mg/Ca ratio and supersaturation on CaCO3 mineralogy. A simple way to elude the first effect is by taking water samples directly from stalactites and by titrating alkalinity in the same containers used to collect dripwaters.