A comparative study of cave system Ca isotope ratios with rainfall, δ13C, and trace element data: Implications for quantitative reconstructions of paleorainfall from speleothems

The development of quantitative records of past rainfall is an outstanding goal in the field of speleothem paleoclimatology and represents an essential step for benchmarking paleoclimate model simulations. However, most traditionally-employed speleothem proxies, including δ¹⁸O, δ¹³C, and trace-element-to-calcium ratios, respond to a number of complex climatic and environmental influences and typically provide only qualitative records of paleoclimate change. Variations in speleothem Ca isotope ratios (δ⁴⁴Ca) are thought to be uniquely controlled by carbonate mineral precipitation above a drip site (prior carbonate precipitation, or PCP), which can be modeled as a Rayleigh fractionation process and calibrated with modern rainfall data. Thus, speleothem δ⁴⁴Ca shows promise as a semi-quantitative proxy for past changes in local effective rainfall rates. However, few cave monitoring studies have focused specifically on the ways in which important factors, like host rock δ⁴⁴Ca variability and geology, water flow path geometry, ventilation, and seasonal rainfall distribution affect δ⁴⁴Ca signals in speleothems.

We present a comparative study of δ⁴⁴Ca data and coeval measurements of δ¹³C and trace element ratios, established proxies for water infiltration, from cave drip waters, farmed calcite, and host rocks from three different cave systems in the United States- White Moon Cave (WMC) in coastal California, Lake Shasta Caverns (LSC) in northern California, and Blue Springs Cave (BSC) in east-central Tennessee. These cave systems are characterized by different hydroclimate, geology, flow path geometry, and seasonal infiltration characteristics.

To assess the relationship between Ca isotope variability and effective rainfall, we use Rayleigh fractionation equations to estimate the amount of PCP occurring at each cave site and compare these estimates with local rainfall rates, supplementing with drip rate information when possible.

The comparison of WMC, LSC, and BSC δ⁴⁴Ca, δ¹³C, and trace element data from drip sites with different flow path geometry and from caves in different geologic and climate settings allows for these key factors to be assessed independently. This work, and the direct comparison between δ⁴⁴Ca measurements and measured local rainfall rates in particular, aids in the refinement of speleothem δ⁴⁴Ca as a new, semi-quantitative proxy for paleorainfall.