MG/CA PALEOTHERMOMETRY

Mg/Ca is an independent paleotemperature proxy that is measured on the same biotic carrier as ¹⁸O, which allows reconstruction of ¹⁸O_water and salinity variations without the problems introduced by other temperature proxies like different seasonal signals or different habitat depths of the biotic carriers (Nürnberg, 2000; Lea et al., 2002; Schmidt et al., 2004). Although Mg/Ca paleothermometry is well established for the last glacial–interglacial cycles, few SST_Mg/Ca high-resolution records yet exist for older time periods (Tripati et al., 2003; Shevenell et al., 2004; Ravelo et al., 2006). Here, we present a high-resolution (<5 k.y./sample) Mg/Ca record for the early Pliocene, covering 2.4 m.y.

The application of Mg/Ca paleothermometry to Pliocene sample material demands consideration of possible changes in Mg/Ca_seawater . Brown (1996) showed in a cultivating experiment that a linear dependence between Mg/Ca_solution and Mg/Ca_foram exists (0.1 mol/mol change in Mg/Ca_seawater leads to 0.059 mmol/mol change in Mg/Ca_foram). Recently, Ries (2004) showed that several carbonate-building organisms display an exponential relationship between Mg/Ca_organism and Mg/Ca_solution. The residence time of Mg in seawater is ~13 m.y. (Broecker and Peng, 1982), which suggests that paleo-SST reconstructions on shorter timescales would not be affected by a change in Mg/Ca_seawater . However, concentrations of Mg and Ca in seawater may be affected by several factors, including varying continental weathering rates (Berner et al., 1983; Wilkinson and Algeo, 1989), hydrothermal alteration of basalt at mid-ocean ridges (Mottl and Wheat, 1994; Elderfield and Schultz, 1996), carbonate deposition (Wilkinson and Algeo, 1989), and ion exchange reactions of Mg with clays (Gieskes and Lawrence, 1981). Several models that have attempted to reconstruct the Mg/Ca ratio of Cenozoic seawater suggest lower Mg/Ca_seawater values than today (Wilkinson and Algeo, 1989; Stanley and Hardie, 1998).

Based on the models of Wilkinson and Algeo (1989) and Stanley and Hardie (1998), we assume a maximum lowering of 0.4 mol/mol in Mg/Ca_seawater for the Pliocene time interval from 4.8 to 2.4 Ma. Applying the relationship of Brown (1996) between Mg/Ca_seawater and Mg/Ca_foram, this implies a maximum underestimation of our SSTs by 0.7°–1.0°C. Because this constant offset is within the error estimation for the SST reconstruction, we consider the Pliocene reduction in Mg/Ca_seawater to be of minor importance.