International audienceWe show that the Li/Mg systematics of a large suite of aragonitic coral skeletons, representing a wide range of species inhabiting disparate environments, provides a robust proxy for ambient seawater temperature. The corals encompass both zooxanthellate and azooxanthellate species (Acropora sp., Porites sp., Cladocora caespitosa, Lophelia pertusa, Madrepora oculata and Flabellum impensum) collected from shallow, intermediate, and deep-water habitats, as well as specimens cultured in tanks under temperature-controlled conditions. The Li/Mg ratios observed in corals from these diverse tropical, temperate, and deep-water environments are shown to be highly correlated with temperature, giving an exponential temperature relationship of: Li/Mg (mmol/mol) = 5.41 exp (−0.049 * T) (r2 = 0.975, n = 49). Based on the standard error of the Li/Mg versus temperature correlation, we obtain a typical precision of ±0.9 °C for the wide range of species analysed, similar or better than that of other less robust coral temperature proxies such as Sr/Ca ratios.The robustness and species independent character of the Li/Mg temperature proxy is shown to be the result of the normalization of Li to Mg, effectively eliminating the precipitation efficiency component such that temperature remains as the main controller of coral Li/Mg compositions. This is inferred from analysis of corresponding Li/Ca and Mg/Ca ratios with both ratios showing strong microstructure-related co-variations between the fibrous aragonite and centres of calcification, a characteristic that we attribute to varying physiological controls on growth rate. Furthermore, Li/Ca ratios show an offset between more rapidly growing zooxanthellate and azooxanthellate corals, and hence only an approximately inverse relationship to seawater temperature. Mg/Ca ratios show very strong physiological controls on growth rate but no significant dependence with temperature, except possibly for Acropora sp. and Porites sp. A strong positive correlation is nevertheless found between Li/Ca and Mg/Ca ratios at similar temperatures, indicating that both Li and Mg are subject to control by similar growth mechanisms, specifically the mass fraction of aragonite precipitated during calcification, which is shown to be consistent with a Rayleigh-based elemental fractionation model.The highly coherent array defined by Li/Mg versus temperature is thus largely independent of coral calcification mechanisms, with the strong temperature dependence reflecting the greater sensitivity of the KdLi/Ca partition coefficient relative to KdLi/Ca. Accordingly, Li/Mg ratios exhibit a highly coherent exponential correlation with temperature, thereby providing a more robust tool for reconstructing paleo-seawater temperatures

In order to assess the effects of ocean acidification and warming on the Mediterranean mussel (Mytilus galloprovincialis), specimens were reared in aquarium tanks and exposed to elevated conditions of temperature (+3 °C) and acidity (-0.3 pH units) for a period of 10 months. The whole system comprised a factorial experimental design with 4 treatments (3 aquaria per treatment): control, lowered pH, elevated temperature and lowered pH/elevated temperature. Mortality was estimated on a weekly basis and every 2 months, various biometrical parameters and physiological processes were measured: somatic and shell growth, metabolic rates and body fluid acid-base parameters. Mussels were highly sensitive to warming, with 100 % mortality observed under elevated temperature at the end of our experiment in October. Mortality rates increased drastically in summer, when water temperature exceeded 25 °C. In contrast, our results suggest that survival of this species will not be affected by a pH decrease of ~0.3 in the Mediterranean Sea. Somatic and shell growth did not appear very sensitive to ocean acidification and warming during most of the experiment, but were reduced, after summer, in the lowered pH treatment. This was consistent with measured shell net dissolution and observed loss of periostracum, as well as uncompensated extracellular acidosis in the lowered pH treatment indicating a progressive insufficiency in acid-base regulation capacity. However, based on the present dataset, we cannot elucidate if these decreases in growth and regulation capacities after summer are a consequence of lower pH levels during that period or a consequence of a combined effect of acidification and warming. To summarize, while ocean acidification will potentially contribute to lower growth rates, especially in summer when mussels are exposed to sub-optimal conditions, ocean warming will likely pose more serious threats to Mediterranean mussels in this region in the coming decades.