Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Linsley, Brad


This study presents a 271-year (1726–1997) subseasonal oxygen and carbon isotopes (δ18O and δ13 C) records from a coral colony of porites lobata at Rarotonga (21.5°S, 159.5°W) in the southwest subtropical Pacific. A new method is introduced whereby the effects of sea surface temperature (SST) can be separated from those of seawater δ18O composition (δ18 Osw) on coral δ18O by using the coupled coral Sr/Ca and δ18O. The reconstructed δ 18Osw at Rarotonga using this method shows that it contributes significantly to the annual changes of coral δ18O for the period 1726–1997. While changes of δ18O, account for ∼39% of the total coral δ18O variation, changes of SST account for ∼61%. The reconstructed δ18O sw also shows a positive linear correlation with a satellite-based estimated salinity for the period 1980–1997 (r = 0.72). This linear correlation between reconstructed δ18Osw and salinity makes it possible to use the reconstructed δ18O, to estimate the past interannual and decadal salinity changes in this region. Applying a similar method to coral δ13C, the effects of kinetic and metabolic activity on coral δ13C were also quantitatively separated. The results show that the variation of coral δ 13C appears to be mainly caused by variation of metabolic activity rather than that of kinetic activity in both tropical and subtropical regions. For the tropical regions, δ13C variation in corals is predominantly influenced by changes of metabolic activity (∼90%), while for subtropical regions, approximately 70–75% of the total variation of coral δ13C is due to the effects of metabolic activity. The interannual and interdecadal variability in coral δ 18O at Rarotonga for the period 1726–1997 was also examined. The results suggest that although Rarotonga is located outside of the center of action of ENSO, it is generally sensitive to ENSO variability in this region. In addition, the decadal variability (∼12 yr) was further differentiated from the interdecadal-scale variability (∼32 yr) for the period 1726–1997 at Rarotonga. Based on the analysis of both tropical and subtropical coral data and comparisons with the Nino3.4 SST index and PDO index, it was hypothesized that the decadal and interdecadal variability might result from separate forcing mechanisms.