COMBINING δ13C, RADIOCARBON, AND ELEMENTAL CHEMISTRY TO UNDERSTAND STABLE CARBON LARGE VARIATIONS IN STALAGMITE

Stable carbon isotopes in stalagmite, δ¹³C_c, is one of the most complicated and difficult proxies to interpret in paleoclimate and paleoenvironmental studies because of its sensitivity to various local, regional, and global factors. For example, the large δ¹³C_c shift (~12‰, vs. VPDB) observed in several stalagmites from Anjohibe Cave, in northwestern Madagascar, around the late Holocene remains a subject of scientific study. Although it was previously assumed to reflect C₃ to C₄ vegetation cover change caused by anthropogenic activities, recent modern investigations inside the cave, at different locations, revealed a wide δ¹³C_c range of ~10‰ (vs. VPDB), a value that is comparable in its significance to the 12 ‰ mentioned above.

Here, we combine δ¹³C, radiocarbon (¹⁴C), and major and trace elements (MTE) as a proxy to provide additional insights about the possible non-vegetation drivers of δ¹³C_c. Our study uses two models, the CaveCalc model and the Fohlmeister model, to help our interpretation of the drip water evolution while it percolates down to the cave and precipitates CaCO₃. The radiocarbon results, with values >100 pMC, demonstrate that the bomb peak, i.e., the pronounced increase in atmospheric ¹⁴C signals, are clearly recorded in these modern stalagmites. This suggests that the samples are younger than 1950 CE, and the soil carbon cycling has been fast without significant formation of aged soil organic matter. These radiocarbon results further suggests that stalagmites from Anjohibe Cave might be of use for future efforts to reconstruct past atmospheric ¹⁴C concentrations. Drip water geochemistry, δ¹³C_c and MTE, combined with two model simulations (CaveCalc and Fohlmeister model) suggests that the great δ¹³C_c range is best explained by the combined effects of the dissolved inorganic carbon residence time in the epikarst that affect the water-rock interaction and the rate of prior carbonate precipitation. This study therefore demonstrated how combined modeling and multi-proxy approaches can advance our understanding of the control mechanisms of δ¹³C_c for paleoclimate research.