Climatic controls on speleothem initial δ234U: evidence from Ejulve Cave over the last 260 ka

The use of δ²³⁴U as a paleoclimatic proxy in stalagmites has remained sporadic, despite uranium isotopes being routinely obtained through U-Th dating. Here, we investigate δ²³⁴U values in six stalagmites from Ejulve cave (northeastern Iberia) spanning the last 260 ka. Elevated δ²³⁴U values are attributed to selective leaching of ²³⁴U from damaged lattice sites and recoil-induced oxidation, with an additional accumulation of ²³⁴U recoils resulting from alpha-decay after growth hiatuses. This selective leaching mechanism weakens under conditions of enhanced bedrock dissolution, resulting in lower δ²³⁴U values.

The mechanisms controlling δ²³⁴U are primarily governed by infiltration frequency and the exposure of mineral surfaces to percolating solutions. However, the efficiency of these processes is strongly modulated by temperature, through its control on soil respiration, soil CO₂ availability, and the intensity of bedrock dissolution. This interpretation is supported by the consistent long-term correlation between δ²³⁴U and sea surface temperatures from the Atlantic Iberian Margin, with lower δ²³⁴U values observed during warmer SST intervals. During stadials and glacial maxima, lower temperatures likely reduced vegetation cover and soil respiration rates, thereby decreasing soil CO₂ concentrations and overall carbonate dissolution rates. Under such conditions, preferential leaching of ²³⁴U from bedrock surfaces is enhanced due to lower bulk rock dissolution. In addition, the high elevation of the study area and the occurrence of frequent winter frosts may have promoted repeated freeze–thaw cycles, inducing microfracturing and increasing the exposure of fresh mineral surfaces to selective leaching. 

Conversely, warmer conditions during interstadials and interglacials promoted higher soil respiration rates and soil CO₂, accelerating bedrock dissolution and yielding low δ²³⁴U values. This coupling between bedrock dissolution intensity and δ²³⁴U is clearly expressed by its correlation with stalagmite growth rate, with important implications. The link between δ²³⁴U, bedrock dissolution, and the initial dripwater oversaturation indicates that δ²³⁴U can serve as a valuable complement to δ¹³C, as both proxies are strongly influenced by soil respiration and soil CO₂, and thus reflect soil and vegetation productivity sensitive to both humidity and temperature. A further implication is that, unlike δ¹³C, uranium isotopes are not fractionated during prior calcite precipitation (PCP). Consequently, δ²³⁴U can be combined with PCP-sensitive proxies such as Mg/Ca or δ⁴⁴Ca to disentangle PCP variations driven by changes in drip rate from those related to shifts in the initial saturation state of dripwater. Finally, we advocate for the broader use of δ²³⁴U as a paleoclimatic proxy in speleothem-based studies from other cave systems.