HR: 1340h
AN: PP23B-1426    [Abstracts]
TI: Fidelity of Stable Oxygen Isotope Ratios as Environmental Recorders Using Multiple Coral Cores From Coastal Kenya
AU: * Fleitmann, D
EM: fleitman@pangea.stanford.edu
AF: Department of Geological and Environmental Sciences, Stanford University, 325 Braun Hall, Stanford, CA 94305-2115 United States
AU: Dunbar, R B
EM: dunbar@stanford.edu
AF: Department of Geological and Environmental Sciences, Stanford University, 325 Braun Hall, Stanford, CA 94305-2115 United States
AU: Mucciarone, D A
EM: dam@pangea.stanford.edu
AF: Department of Geological and Environmental Sciences, Stanford University, 325 Braun Hall, Stanford, CA 94305-2115 United States
AB: Climate variability in the western equatorial Indian Ocean reflects the combined influence of seasonally changing sea surface temperature (SST), ocean currents, and monsoon circulation, as well as inter-annual to -decadal variability associated with ENSO in the Pacific. However, the Indian Ocean also exhibits variability that appears unrelated to ENSO in the Pacific. The nature of interactions between air-sea variability in the Indian and Pacific oceans is not yet fully resolved, in part because of the lack of long-term, high-resolution SST records from key localities in the Indian Ocean. Such records are now being obtained using oxygen isotope profiles measured on corals from East Africa, Indonesia, Australia, and Indian Oceania (Maldives, Seychelles), with the longest coral-based time series from Malindi, Kenya, covering the last 300 years at near-monthly resolution. The value of these developing coral records depends on the fidelity with which they record regional climate variability. In order to assess the fidelity of oxygen isotope ratios (\delta$^{18}$O) in Indian Ocean corals as a proxy for sea surface temperature, we have generated stable isotopic time series from multiple Porites lutea coral heads collected along the coast of Kenya. Coral-based isotopic paleoclimatology is labor and time-intensive so detailed analyses using multiple coral heads from different sites within a region are extremely rare. Most published records are produced from a single coral head, yet questions have been raised about the accuracy of such records. To address such concerns, near-monthly resolution isotopic profiles, spanning 10 to 50 years prior to 1997, were measured on a total of 8 cores from five sites along a north-south transect between $2\deg$ and $4\deg$S (Kiwayu: $2\deg$2'S, $41\deg$2'E, Malindi: $3\deg$14'S, $40\deg$8'E, Watamu: $3\deg$23'S, $39\deg$52'E, Mombasa: $3\deg$59'S, $39\deg$5'E, and Kisite: $4\deg$43'S, $39\deg$23'E. Correlations among individual \delta$^{18}$O time series (r values range from 0.65 to 0.80) reveal that sample- and site-specific effects do not sufficiently bias \delta$^{18}$O such that regional climate signals cannot be discerned from any of the 5 sites. Furthermore, correlation of all individual oxygen isotope time series with instrumental monthly SST yields r values between -0.70 and -0.75, demonstrating that fluctuations in \delta$^{18}$O primarily reflect variations in SST. Although correlation coefficients between \delta$^{18}$O and instrumental monthly SST are higher using either single-site composite \delta$^{18}$O time series, such as for Malindi (r = -0.76), or a multisite, multi-core composite (r = -0.81), these correlation coefficients are only slightly higher than those for individual coral \delta$^{18}$O time series. We conclude that in this case, the cost of developing replicate coral delta$^{18}$O time series from individual sites may not be warranted in terms of a marginal gain in signal-to-noise ratio. Our work suggests that a 300 year long, near-monthly \delta$^{18}$O record from a single coral head from Malindi, can be used with confidence for environmental reconstructions in the western equatorial Indian Ocean.
DE: 9340 Indian Ocean
DE: 4870 Stable isotopes
DE: 4215 Climate and interannual variability (3309)
DE: 4231 Equatorial oceanography
DE: 4522 El Ni¤o
SC: Paleoceanography and Paleoclimatology [PP]
MN: 2004 AGU Fall Meeting