The carbonate geochemistry of marine foram shells and corals are often used to provide long, continuous records of paleoclimate.  (Mike adds: The use of such tracers such as Mg/Ca and Sr/Ca is exciting because they are ostensibly univariate water temperature proxies, and may be used to address the controversy over Last Glacial Maximum tropical sea surface temperatures.)  However, the degree to which Sr and Mg replace Ca in carbonates depends on the temperature of the water at the time of formation of the carbonate, the chemistry of the water, and any biological effects of the mediating organism. The biological effects are species-specific, and each species, therefore, must be independently calibrated in culturing experiments. Researchers assume that elemental ratios in ocean water are constant over short time scales, but elemental ratios can change over glacial-interglacial time scales. Differentiating between a changing climate and a changing composition of the ocean is often difficult.

Mg/Ca and Sr/Ca ratios in carbonates are temperature dependent. A 5°C temperature change results in a 0.3% change in the Sr/Ca ratio of inorganic calcite, and a 0.5% change in the Mg/Ca ratio.

Post-depositional alterations to shell chemistry and core stratigraphy are another important consideration. Bioturbation can disrupt the relative position of forams in the sediment, resulting in lowered precision of the paleoclimate record. Shell dissolution preferentially removes Mg, which alters the paleoclimate signal. Increased dissolution would result in lower calculated temperatures. The depth of the lysocline also changes with time, which affects the degree to which dissolution occurs. The lysocline is the depth in the ocean at which carbonate dissolution increases dramatically. A high lysocline would lead to increased dissolution.