Corals can be dated over a wide range of time scales (from seasons to periods greater than 100,000 years) by a variety of independent means, including density and fluorescent banding, 230Th/234U ratio, 14C content, stable-isotope fluctuations, and amino acid racemization (Dodge et al., 1974; Isdale, 1984; Edwards et al., 1988; Fairbanks, 1989, 1990; Bard et al., 1990; Cole and Fairbanks, 1990; Goodfriend et al., 1993). Recent work on monthly and even daily coral banding offers the promise for increasingly refined chronological determinations (Barnes and Lough, 1989; Risk and Pearce, 1992).
Corals precipitate a skeleton of aragonite (CaCO3) that can incorporate several independent chemical tracers used to monitor variability in oceanic and atmospheric processes. The most consistently useful of these are the isotopic ratios of carbon and of oxygen in particular, and the concentration of certain lattice-bound trace metals that appear to substitute for skeletal calcium (expressed as Cd/Ca, Ba/Ca, Mn/Ca, and Sr/Ca ratios). In this section, we present an overview of the many tracers now in use for coral-based paleoclimatic reconstruction. Subsequent sections discuss examples of both short records from a variety of sites and longer records that yield insight into the recent history of tropical climate variability.
The stable isotopic content of coral aragonite is expressed in parts per thousand () as d180 and d13C, where the d notation is defined in terms of the isotopic ratios R (either 13C/12C or 18O/16O) in the sample relative to a standard:
d = [(Rsample - Rstandard)/Rstandard] × 1000
For the oxygen and carbon isotopic content of coralline aragonite, the conventional reference standard is the Pee Dee Belemnite, or PDB.
Coral skeletal d18O reflects a combination of local SST and the d18 O of ambient seawater. The d18O of biogenic calcium carbonate precipitated in equilibrium with seawater decreases by about 0.22 for every 1°C rise in water temperature (Epstein et al., 1953). In corals, this effect is biologically mediated such that the d18O of the coral skeleton is offset below seawater d18O. This offset is constant within a coral genus (Weber and Woodhead, 1972) for the rapidly growing central axis of the skeleton (Land et al., 1975; McConnaughey, 1989). Coral d18O records taken along the axis of maximum growth thus track ambient temperatures at subseasonal resolution (Fairbanks and Dodge, 1979; Dunbar and Wellington, 1981; Pätzold, 1984; McConnaughey, 1989). Coral skeletal d18O also records any variations in the d18O of the seawater (Epstein et al., 1953; Fairbanks and Matthews, 1979; Swart and Coleman, 1980). Such variations are small in most regions of the tropical ocean, but in some locations changes in evaporation, precipitation, or runoff may cause pronounced d18O variability (Dunbar and Wellington, 1981; Cole and Fairbanks, 1990). In regions with fairly constant or well-known temperature histories, coral d18O provides a record of past variations in the hydrologic balance (Cole and Fairbanks, 1990). In many cases, coral skeletal d18O reflects a combination of thermal and hydrologic factors.
The interannual d13C signal in coral skeletons is often difficult to decipher in environmental terms, because of complicated interactions with biological processes that involve strong isotopic fractionation. Environment-related controls on skeletal d13C include (1) the isotopic composition of the ambient seawater (Nozaki et al., 1978), (2) coral geometry and growth rate (e.g., apex versus side of coral head) (Land et al., 1975; McConnaughey, 1989), and (3) photosynthesis of endosymbiotic dinoflagellates (Weber, 1974; Goreau, 1977; Fairbanks and Dodge, 1979; Swart, 1983; McConnaughey, 1989). This last parameter depends upon ambient light levels, as regulated by water depth and insolation. Coral skeletal d13C correlates positively with insolation in many contexts, from depth-dependent variation (Weber and Woodhead, 1970; Fairbanks and Dodge, 1979; McConnaughey, 1989) to annual cycles that reflect rainy (i.e., cloudy) seasons (Fairbanks and Dodge, 1979; Pätzold, 1984; McConnaughey, 1989; Cole and Fairbanks, 1990). However, shallow corals may experience reduced photosynthesis during brighter periods, while deeper corals may respond to increased light by increasing photosynthesis (Erez, 1978; McConnaughey, 1989). These responses can produce opposite skeletal d13C signatures in deep and shallow corals (McConnaughey, 1989). Environmental reconstruction from coral d13C records requires more information about growth conditions and physiological responses than is usually available in a paleoceanographic context.
Specific environmental processes, including upwelling, advection, aeolian transport, and runoff, influence the surface-water concentrations of certain trace elements (Boyle, 1988; Martin et al., 1976; Shen and Boyle, 1988; Lea et al., 1989; Shen and Sanford, 1990; Shen et al., 1991, 1992b). Several such metals appear to substitute readily for Ca in the aragonite lattice of coral skeletons. Estimated distribution coefficients between corals and seawater allow the reconstruction of ambient seawater metal concentrations from metal concentrations in the coral skeleton. Trace metal records from corals thus yield a history of the processes that control the local distribution of trace metals.
The most useful metals for coral reconstructions of ENSO variability include Cd, Ba, Mn, and Sr. Many authors (Shen and Boyle, 1988; Lea et al., 1989; Linn et al., 1990; Shen and Sanford, 1990; Shen et al., 1987, 1991; Beck et al., 1992; de Villiers et al., 1993) describe these applications