Early diagenesis impact on precise U-series dating of deep-sea corals: Example of a 100–200-year old Lophelia pertusa sample from the northeast Atlantic

Deep-sea coral has proved useful for paleoceanographic reconstructions and for documenting 14C-ages of water masses using 230Th-ages. However, for precise and accurate U-series dating, further information on coral-age structure, growth rate and diagenetic evolution is still needed. To document such processes, we used U-Th-226Ra systematics in a 40 cm diameter, Lophelia pertusa specimen collected in 1912 from the Eastern Atlantic (Sea of the Hebrides). External parts of the specimen are thought to have been alive when collected whereas more internal parts were likely dead. The “live-collected” and “dead” parts of the skeleton were sampled and analyzed for their 230Th, 232Th, 234U, 238U, 226Ra and Ba contents by thermal ionization and multicollector inductively coupled plasma mass spectrometry. 230Th/234U ratios in the most recent parts yielded ages of 86 ± 6 a and 92 ± 9 a, in agreement with the date of recovery. The older parts yielded 230Th ages ranging from 169 ± 15 to 211 ± 10 a (n = 5), but had a 31% higher U content than more recent parts of the skeleton. This raises concerns about the possibility of secondary diagenetic U uptake, although an environmentally controlled U/Ca shift between coral growth stages cannot be ruled out. 226Ra/Ba measurements were made, and model- 226Ra/Ba ages averaging 250 ± 12 yr were calculated for the older part, assuming a constant initial 226Ra/Ba ratio in bottom waters. These ages are slightly older than 230Th-ages, suggesting either that 226Ra/Ba ratios of ambient-seawater changed over time or that a diagenetic phenomenon have affected the U-series system, or both. Scanning electron microscope observations revealed bioerosion and secondary biomineralization in the older part of the coral skeleton, supporting the hypothesis that diagenetic processes may have influenced the ages obtained by the U-series toolbox. Modeled U-series ages for such an open system are discussed. However, a comparison between 14C- and 230Th-ages performed on both pristine and bioeroded parts of the coral gives coherent values (ca 450 a) for the preindustrial 14C-reservoir age of North Atlantic waters. It remains to be determined, however, whether diagenesis occurs rapidly over a short period of time, or whether it continues for longer periods. In the latter case, diagenetic processes would hamper paleoceanographic interpretations as well as the precise calculation of 14C ages of deep-water masses, based on comparative U/Th- and 14C-chronologies.