Calibration of Sr/Ca, Li/Mg and Sr‐U Paleothermometry in Branching and Foliose Corals.

Coral skeletons are the most commonly used high‐resolution temperature proxy in the tropical oceans, providing paleoclimate reconstructions dating back centuries to millennia. However, physiological differences in skeletal formation modes together with artifacts arising from coral biomineralization (vital effects) can confound the temperature dependence of single element‐to‐calcium ratios. In efforts to reduce vital effects and isolate temperature, new approaches have been developed based on Sr‐U and Li/Mg, which combine Sr/Ca and U/Ca and Li/Ca and Mg/Ca, respectively. Here we examine the systematics of Sr/Ca, Sr‐U, and Li/Mg paleothermometry in 33 colonies of branching (Acropora, Pocillopora, and Stylophora) and foliose (Turbinaria) genera. To address the calibration of these morphologically complex calcifiers, we conducted repeat field trips every 1 to 3 months and collected the most recent (~1 month) uppermost growth of individual colonies over ~18‐ to 24‐month periods. This enables seasonally resolved calibration of genera that exhibit rapid extension and slower secondary calcification. Based on this experimental design, we show that all three proxies capture seasonal to annual temperature variations for their respective growth intervals, providing calibrations across an 11 °C range. Species effects on the temperature dependence were largest for Sr/Ca (22.7%) yet minor for Li/Mg (7.2%) and Sr‐U (6.3%). Residuals from proxy‐temperature regressions were correlated between Sr/Ca and Li/Mg, indicating similar biological processes may influence Sr/Ca and Li/Mg thermometry. The implications of this study are that Sr‐U and to a lesser extent Li/Mg are applicable to fossil branching coral skeletons identified to genus level without the need for modern‐day calibration. We further show that all three paleothermometers provide complementary temperature constraints, with Li/Mg and the more species‐dependent Sr/Ca showing greater effectiveness at resolving seasonal variability and Sr‐U showing greater reliability at capturing mean annual temperature. Plain Language Summary: In this paper, we report skeletal trace element compositions in seven branching and plating coral species from locations spanning an ~11 °C mean monthly temperature range and latitudes of ~23 to 34 °S along the Western Australian coastline in the southeast Indian Ocean. We compare three coral paleothermometers (Sr/Ca, Li/Mg, and Sr‐U) with respect to performance and species effects. Repeat field trips to collect coral skeletons were undertaken regularly (approximately bimonthly) and skeletons were subsampled to obtain only the most recent ~1 month of new growth, minimizing sampling of secondary skeletal infilling. Using this experimental design, we find that all three paleothermometers record seasonal cycles in temperature, overcoming limitations posed by complex secondary infilling calcification. We find that coral Sr/Ca shows the largest offsets between species (22.7%) with residuals between Sr/Ca and Li/Mg proxy‐temperature regressions being correlated, indicating that similar biological processes influence both Sr/Ca and Li/Mg thermometry. Coral Sr‐U and to a lesser extent Li/Mg exhibit significantly reduced species offsets and appear largely unaffected by the vital effects that influence Sr/Ca. These results support the applicability of coral paleothermometry, especially Sr‐U, for deriving mean annual temperature. We thus provide an approach for reconstructing paleotemperatures using trace elements in common branching fossil coral skeletons without the need for modern calibrations. Key Points: Geochemical temperature proxy calibrations are obtained for branching and foliose corals by repeat (~ bimonthly) collection of new skeletal growthLi/Mg and Sr/Ca were most effective for resolving seasonal temperature while Sr‐U was most effective for capturing mean annual temperatureA multiproxy approach is recommended where Sr‐U is applied to estimate mean annual temperature and Sr/Ca and Li/Mg are used to resolve subannual temperature changes [ABSTRACT FROM AUTHOR]