| Summary: | This thesis explores Holocene climate conditions in the western Pacific warm pool (WPWP) using geochemical tracers in fossil corals from the north coast of Papua New Guinea (PNG). The WPWP, with an annual average sea surface temperature (SST) exceeding 28°C, is not only a major heat source driving equator-to-pole atmospheric circulation, but is a fundamental component of the El Nino-Southern Oscillation (ENSO) system. Ocean-atmosphere interactions in the warm pool are instrumental in triggering ENSO warm events (El Nifio events). Given the often severe ENSO-related climate impacts around the globe, the mid-Holocene has emerged as a key period for investigating this system due to subtle differences in the distribution of solar insolation, relative to that of today. An understanding of how the WPWP climate responded to this altered background state can give us clues as to how ENSO may change under other altered background climate conditions, such as that brought about by atmospheric C02 levels during the 2P1 century. To address this issue, fossil and modern Porites corals were sampled from uplifted and living reefs on Muschu and Koil Islands, PNG, to reconstruct the mid-Holocene climate in the WPWP. Skeletal Sr/Ca and oxygen isotope ratios (8180) were measured in seven fossil corals with ages between 7.6-5.4 ka (xlOOO calendar years ago), one 2 ka coral, and three modern corals. The results were used to calculate SST and the oxygen isotope residual (~81 8 0), the two main proxies for warm pool climate used in this study. Three modern corals, two from Muschu Island and one from Koil Island, were analysed for Sr/Ca and 8180 to establish the methodologies to apply to the fossil corals. Sr/Ca and 8180 analyses on annually-resolved samples from the three modern corals were used to assess the level of reproducibility amongst coral records for this region. The spread of 8180 data about the mean value for the three corals was 0.04%o, and for the Sr/Ca data was 0.00001. When converted to SST, the spread for the Sr/Ca data is equivalent to ± 0.7°C. Seasonally resolved Sr/Ca data from a modern Muschu Island coral were found to track SST changes observed in the instrumental records. This was combined with other modern coral Sr/Ca data from the PNG region and used to develop a calibration equation to convert Sr/Ca to SST. The Sr/Ca-SST calibration equation is Sr/Caatomicxl03 = 11.0- 0.075 x SST. The modern Muschu coral was also analysed for 8180 at seasonal resolution. The seasonal 8180 and Sr/Ca data from the Muschu coral were then used to calculate ~81 8 0, by removing the SST component from the 8180 results. ~81 80 was converted to sea surface salinity (SSS) using the following seawater 8180-SSS relationship developed from tandem measurements on water samples from the north coast of PNG: 8180w = -8.7 + 0.26 x SSS. The modern corals record seasonal, inter-annual and decadal scale climate variability. The seasonally resolved Sr/Ca-SST data showed one annual minimum in January, probably due to northwesterly wind driven coastal upwelling. A ~81 80 maximum in January, also likely reflects upwelling, as well as the annual rainfall minimum . .18180 was at a minimum when rainfall was highest from May-August and southeasterly winds prevail. Both coral Sr/Ca-SST and .18180 show relatively cool and more saline conditions (reduced rainfall) respectively, during ENSO warm events. The annually-resolved Sr/Ca and 8180 records for the three modern corals were averaged and used as the baseline for assessing relative changes in mean climate observed in the fossil coral climate records. Climate reconstructions derived from the geochemistry of fossil corals can be corrupted by diagenetic alteration of coralline skeletal aragonite. To understand the impact of vadose-zone diagenesis on coral climate proxies, two of the mid-Holocene Porites corals were analysed for Sr/Ca, 8180, and 813C along transects from 100% aragonite to 100% calcite. Thin-section analysis showed a characteristic vadose zone diagenetic sequence, beginning with leaching of primary aragonite and fine calcite overgrowths, transitional to calcite void filling and neomorphic, fabric selective replacement of the coral skeleton. Average Sr/Ca and 8180 values for calcite were lower than those for coral aragonite, decreasing from 0.0088 to 0.0021 and -5.2 to -8.1 %o, respectively. Diagenesis has a greater impact on reconstructions of SST from Sr/Ca, relative to 8180; the calcite compositions reported here convert to SST anomalies of l15°C and 14°C, respectively. Thus, based on the calcite Sr/Ca compositions analysed in this study, and in the literature, the sensitivity of coral Sr/Ca-SST to vadose-zone calcite diagenesis is l.l-1.5°C per percent calcite. In contrast, the rate of change in coral 8180 SST is relatively small (-0.2 to 0.2°C per percent calcite). The results indicate that large shifts in 8180, reported for mid-Holocene and Last Interglacial corals with warmer than present Sr/Ca-SSTs, cannot be caused by diagenesis. X-ray diffraction and petrographic analysis of fossil coral skeletons used for climate reconstruction in this thesis revealed no significant diagenesis. Fossil corals records of Sr/Ca and 8180 were used to infer mean SST, .18180 and SSS conditions in the WPWP during the mid-Holocene. The U-Th dated fossil corals show that from 7.6 to 6.1 ka SSTs were on average -0.9°C cooler than at present, and .18180 converted to SSS suggests conditions were -1.5 psu more saline than today. Taken together with other tropical Pacific SST proxy records, the overall SST structure is evocative of the modern-day ENSO cool phase (La Nifia). If a mean La Nifia state was operating during this time, then the easterly trade winds were likely to have been stronger, thus increasing evaporation relative to precipitation and raising the SSS of the warm pool. An abrupt shift, particularly in the .18180, was identified between 6.1 and 5.4 ka. This shift, indicating a decrease in SSS of -1.5 psu, may represent the establishment of a modern-like WPWP. The timing of the abrupt shift is similar to abrupt shifts identified in proxy climate records from the Indian sub-continent, Indian Ocean and tropical east Atlantic. The timing is also similar to an enhanced millennialscale climate oscillation in drift-ice rafting and deepwater production, identified in the North Atlantic. This shift points to the WPWP playing a stronger role than previously thought in global climate change during the Holocene. Annually-resolved 8180 data from the seven fossil corals aged from 7.6-5.4 ka showed an El Nifio recurrence interval of nine years, compared to seven years for a 8180 coral record spanning 1911-1997AD. This suggests that El Nifio was slightly suppressed during this time as indicated by models, though not to the same extent as suggested by South American lake sediment records. Four fossil corals dating to 7.3 ka, 6.1 ka, 5.4 ka and 2 ka were analysed at bimonthly resolution for Sr/Ca and 8180 to investigate changes in the seasonal cycle of SST and ~818 0, and changes in the El Nifio signal. The results for the period 7.3 - 6.1 ka suggest locally increased rainfall during the middle of the year, implying strengthened southeasterly winds, consistent with an enhanced La Nifia state during this time. By 5.4 ka, mid-year ~818 0 were at a maximum locally, implying that the southeast trade winds weakened. This is consistent with overall fresher mean surface ocean conditions and suggests that the modern SST and SSS structure of the WPWP may have been established at this time. Analysis of ~8180 for the El Nifio years observed in the 7.3 ka, 6.1 ka and 5.4 ka coral records showed that El Nifio events at these times peaked during the middle of the year. This peak in mid-Holocene El Nifio events is 4-6 months earlier than the peak in El Nifio today (usually at the end of the year). A 7-year protracted El Nifio was identified in the 2 ka coral 8180 records, consistent with increased El Nifio frequency and intensity found in both modelling and proxy studies. High-resolution Sr/Ca and 8180 analysis of one year of the 7-year El Nifio event showed that it peaked at the same time as today. The high-resolution results imply that small changes in background conditions, such as orbital parameters and/or the strength of the tropical monsoons, as proposed by modelling studies, can influence the development of ENSO events. Therefore, it is not possible to rule-out the potential for changes in background conditions, such as the current increase in C02 levels in the atmosphere from the burning of fossil fuels, to cause large changes in the ENSO.
|
|---|