Cadmium in coral skeleton: Natural variability and in situ calibration
Skeletal cadmium (Cd) in corals is a promising proxy for oceanic upwelling. However neither a systematic evaluation of its natural variability, nor a calibration to environmental conditions has been made. Here, a high-precision, quantitative study of coral Cd using isotope dilution-inductively coupled plasma-mass spectrometry (ID-ICP-MS) is presented. Advantages of this approach over existing methods are: (1) reduced introduction of high-concentration elements into the MS, (2) decreased sample requirements (15 mg), and (3) improved precision of ±2.2% (for Cd/Ca) and ±2.3% (for [Cd]coral). Using this technique, species, depth and upwelling effects on coral Cd were assessed. Multiple colonies of Porites lobata, Pavona clavus and Pavona gigantea were reared at 1 m and 7 m depths during upwelling and nonupwelling intervals in the Gulf of Panamá (Pacific). Mean coral Cd increased during upwelling. However differences between depth and species were significant, with increased Cd in shallow coral and in Pavona spp. Cd in P. lobata was uniformly low, with a reduced range, suggesting that it is a poor choice for paleoupwelling reconstruction. The non-correlation between Cd and maximum linear skeletal extension, δ13C or δ18O implies that Cd incorporation is decoupled from growth rates, metabolism, and sea surface temperature (SST) and salinity. Shallow Pavona spp. colonies are the most reliable, however multiple cores should be used for reconstructions, given the effects of natural variability.
Analyses of [Cd]coral via laser ablation-ICP-MS at sub-monthly resolution are reported. A significant correlation of P. clavus data from LA-ICP-MS and ID-ICP-MS was found, though with consistently lower concentrations using ID. Single LA tracks in all species yielded variable results, however the average of parallel paths mirrored changes in in situ seawater Cd ([Cd]sw), albeit with an offset between peaks. Maximum [Cd]coral occurred after the SST minimum, but before peak [Cd]sw. Assuming a 1-month offset, [Cd]coral is well correlated to [Cd]sw, providing the first direct calibration of this upwelling proxy (distribution coefficient ≈ 1.3–1.7). A three year record of [Cd]coral demonstrated the ability of LA-ICP-MS to rapidly generate long records. Additional refinement of this technique could further improve the resolution and quality of measurements of this upwelling tracer.