6.4.3.2 What Was the Magnitude of Glacial-Interglacial Sea Level Change?

Model-based palaeo-sea level analysis also helps to refine estimates of the eustatic (globally averaged) sea level rise that occurred during the most recent glacial-interglacial transition from the LGM to the Holocene. The extended coral-based RSL curve from the island of Barbados in the Caribbean Sea (Fairbanks, 1989; Peltier and Fairbanks, 2006) is especially important, as the RSL history from this site has been shown to provide a good approximation to the ice-equivalent eustatic curve itself (Peltier, 2002). The fit of the prediction of the ICE-5G(VM2) model to the Fairbanks data set, as shown in Figure 6.8b, constrains the net ice-equivalent eustatic rise subsequent to 21 ka to a value of 118.7 m, very close to the value of approximately 120 m conventionally inferred (e.g., Shackleton, 2000) on the basis of deep-sea O isotopic information (Figure 6.8b). Waelbroeck et al. (2002) produced a sea level reconstruction based upon coral records and deep-sea O isotopes corrected for the influence of abyssal ocean temperature changes for the entire glacial-interglacial cycle. This record (Figure 6.8a) is characterised by a best estimate of the LGM depression of ice-equivalent eustatic sea level that is also near 120 m. The analysis of the Red Sea O isotopic record by Siddal et al. (2003) further supports the validity of the interpretation of the extended Barbados record by Peltier and Fairbanks (2006).

The ice-equivalent eustatic sea level curve of Lambeck and Chappell (2001), based upon data from a variety of different sources, including the Barbados coral record, measurements from the Sunda Shelf of Indonesia (Hanebuth et al., 2000) and observations from the Bonaparte Gulf of northern Australia (Yokoyama et al., 2000), is also shown in Figure 6.8b. This suggests an ice-equivalent eustatic sea level history that conflicts with that based upon the extended Barbados record. First, the depth of the low stand of the sea at the LGM is approximately 140 m below present sea level rather than the value of approximately 120 m required by the Barbados data set. Second, the Barbados data appear to rule out the possibility of the sharp rise of sea level at 19 ka suggested by Yokoyama et al. (2000). That the predicted RSL history at Barbados using the ICE-5G(VM2) model is essentially identical to the ice-equivalent eustatic curve for the same model is shown explicitly in Figure 6.8, where the red curve is the model prediction and the step-discontinuous purple curve is the ice-equivalent eustatic curve.

Figure 6.8. (A) The ice-equivalent eustatic sea level history over the last glacial-interglacial cycle according to the analysis of Waelbroeck et al. (2002). The smooth black line defines the mid-point of their estimates for each age and the surrounding hatched region provides an estimate of error. The red line is the prediction of the ICE-5G(VM2) model for the Barbados location for which the RSL observations themselves provide an excellent approximation to the ice-equivalent eustatic sea level curve. (B) The fit of the ICE-5G(VM2) model prediction (red line) to the extended coral-based record of RSL history from the island of Barbados in the Caribbean Sea (Fairbanks, 1989; Peltier and Fairbanks, 2006) over the age range from 32 ka to present. The actual ice-equivalent eustatic sea level curve for this model is shown as the step-discontinuous purple line. The individual coral-based estimates of RSL (blue) have an attached error bar that depends upon the coral species. The estimates denoted by the short error bars are derived from the Acropora palmata species, which provide the tightest constraints upon relative sea level as this species is found to live within approximately 5 m of sea level in the modern ecology. The estimates denoted by the longer error bars are derived either from the Montastrea annularis species of coral (error bars of intermediate 20 m length) or from further species that are found over a wide range of depths with respect to sea level (longest error bars). These additional data are most useful in providing a lower bound for the sea level depression. The data denoted by the coloured crosses are from the ice-equivalent eustatic sea level reconstruction of Lambeck and Chappell (2001) for Barbados (cyan), Tahiti (grey), Huon (black), Bonaparte Gulf (orange) and Sunda Shelf (purple).