Using stable oxygen isotope dual-inlet isotope-ratio mass spectrometry to elucidate uranium transport and mixed Th/U calcite formation ages at the seminal Devils Hole, Nevada, natural laboratory.

Rationale: Vein calcite in Devils Hole has been precipitating continuously in oxygen-isotope equilibrium at a constant temperature for over 500 000 years, providing an unmatched δO paleoclimate time series. A substantial issue is that coeval calcite (based on matching δO values) has uranium-series ages differing by 12 000 years.

Methods: An unparalleled high-accuracy δO chronology series from continuously submerged calcite was used to correct the published uranium-series ages of non-continuously formed calcite in two cores, cyclically exposed by water-table decline during glacial-interglacial transitions. This method relies on the premise that the δO values of coevally precipitated calcite are identical, allowing matching calcite δO values to establish formation ages.

Results: Exposed calcite can have apparent ages that are 12 000 years too young due to unrecognized uranium mobility and resulting mixed ages identified in over 50 mixed uranium-series ages from previous studies. Secondary uranium in fluids, sourced from the formation or dissolution of porous carbonate deposits (folia) with high uranium-238 (U) concentrations, has migrated up to 10 mm into vein calcite.

Conclusions: The continuously submerged Devils Hole δO chronology is not explained by orbital forcing. Rather, this chronology represents a regional climate record in the southern Great Basin of sea-surface-temperature (SST) variations off California, variations that preceded the last and penultimate deglaciations by 5000 to approximately 10 000 years. Temporal discrepancies between the continuously submerged Devils Hole chronology and other regional δO records (e.g., the Leviathan chronology) can be explained by unrecognized cryptic, pernicious uranium mobility, leading to model estimations that may be thousands of years younger than actual ages. Consequently, paleo-moisture availability, water-table, and groundwater recharge models based on these mixed uranium-series ages are too young by as much as 12 000 years. The potential for post-formation uranium addition in subaerial cores and speleothems underscores the need for caution in uranium-series dating, highlighting δO time-series comparisons as a method for identifying mixed ages.