Stable Sr-isotope, Sr/Ca, Mg/Ca, Li/Ca and Mg/Li ratios in the scleractinian cold-water coral Lophelia pertusa

The aragonitic skeletons of scleractinian cold-water corals can serve as valuable archives in paleoceanographic studies. The potential of delta Sr-88/86, Sr/Ca, Mg/Ca, Li/Ca and Mg/Li ratios of the cold-water coral Lophelia pertusa to record intermediate water mass properties has been investigated. Here we used samples from several locations along the European continental margin spanning a large temperature range from 6 to 14 degrees C. Stable strontium isotope measurements were carried out with the recently developed double spike TIMS technique and our results differ from those obtained with less precise methods. In contrast to the strong positive relationship with temperature of previous studies, our results suggest that delta Sr-88/86 measured in scleractinian cold-water corals is not controlled by seawater temperature, but reflects the Sr isotopic composition of seawater with an offset of Delta(ST)-S-88/86 = -0.196 parts per thousand. As found in previous studies, the elemental ratios Sr/Ca, Li/Ca and Mg/Li measured in corals are significantly related to water temperature and do not correlate with salinity. Moreover, Sr/Ca ratios in L pertusa display the expected inverse correlation with temperature. However, the variance in the Sr/Ca data severely limits the accuracy of paleotemperature estimates. The Li/Ca and Mg/Ca ratios reveal other influences besides temperature such as pH and/or growth or calcification rate. However, corresponding Mg/Li ratios in L pertusa are more tightly related to temperature as they remove these secondary effects. In particular, the Mg/Li ratio in L pertusa may serve as a new promising paleotemperature proxy for intermediate water masses. Our dataset represents the most extensive geochemical examination of L pertusa to date, revealing a temperature sensitivity of 0.015 mol/mmol/degrees C for Mg/Li. However, using this temperature dependence and the precision of 5.3% (2SD) only temperature variations larger than similar to 1.5 degrees C can be resolved with 95% confidence. (C) 2013 Elsevier B.V. All rights reserved.