Testing the reproducibility of Mg/Ca profiles in the deep-water coral Primnoa resedaeformis: putting the proxy through its paces

Two samples of the calcitic deep-sea coral Primnoa resedaeformis have been analysed for Mg/Ca ratios by micro-beam methods ( laser-ablation ICP-MS and electron microprobe). Continuous profiles of Mg/Ca have been studied with the aim of establishing the reproducibility of the variations in different parts of the coral, and therefore the potential use of Mg/Ca as a paleoceanographic or paleoclimatic tracer.A method of spectral decomposition based on box-filter smoothing has been developed for analysing the signals. This method shows that all profiles measured contain some degree of irreproducibility. Much of this irreproducibility appears to be white-noise added by the analytical methods. The amplitude of this noise is around 15 % of the signal average for the Nd-YAG laser-ablation system at Memorial University of Newfoundland, 5–6 % for the electron microprobe at Dalhousie University, and 4–4.5 % for the ArF Excimer laser-ablation system at the Australian National University.In the case of the LA-ICP-MS analyses, this noise may relate to uneven ablation of material from the coral and/or an uneven size distribution of particles entering the ICP-MS plasma. For the most stable system (ANU’s LA-ICP-MS), features on distance scales smaller than 150–200 µm will be obscured by the noise, and features smaller than 500 µm should be interpreted with caution.There is some evidence that the coral itself contains compositional heterogeneity, although reproducibility is mostly limited by instrumental noise. We recommend that several analytical profiles be compared before any attempt is made to interpret Mg/Ca variations in terms of paleoceanographic changes.