SEARCH

SEARCH BY CITATION

Keywords:

  • sea surface temperature;
  • terrestrial input;
  • Papua New Guinea;
  • Intertropical Convergence Zone

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] Abrupt climate changes such as Dansgaard–Oeschger (D-O) cycles and Heinrich events (HEs) are prevalent during the last glacial cycle and widely documented in Northern Hemisphere (NH) high latitudes. However, in tropical regions and the Southern Hemisphere (SH) far fewer records exist, especially in the western Pacific warm pool (WPWP) area. Here, we present a 50k archive of U37k′ sea surface temperature (SST), planktic foraminifera oxygen isotopes, and terrestrial input indicators including branched and isoprenoid tetraether (BIT) biomarkers, 232Th activity, and non-biogenic sediment components recorded in core MD052928 from the WPWP (near southern Papua New Guinea, PNG). The planktic foraminifer oxygen isotopes in the core show millennial-scale changes indicating fresher seawater during the NH cold periods (i.e., Heinrich Events, HEs) and suggesting hydrological changes that are most likely linked to the strength of the boreal winter Asian-Australian monsoon (AAM). Our observations are corroborated by evidence from the same core that indicates increased terrestrial input caused by higher precipitation on land and more river runoff from southern PNG during the cold periods. Consistent with other nearby hydrological records from land, our study indicates persistent millennial-scale hydrological changes within the past 50k in the western tropical Pacific and Southeast Asia. The timing of the millennial-scale changes appears to have been determined by the latitudinal displacement of the Intertropical Convergence Zone (ITCZ) that reflects a history of heat transport from the tropics and WPWP.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] The Western Pacific Warm Pool (WPWP), with sea surface temperatures (SST) above 28°C and reaching a maximum of ∼30°C, is thought to be a major source of water vapor and a global heat transport engine [Pierrehumbert, 2000]. This region of high and stable SSTs maintains deep atmospheric convection and therefore supports both meridional and zonal heat and moisture transfer (e.g., Hadley and Walker Circulation). WPWP dynamics are affected by El-Niño Southern Oscillation (ENSO), Asian-Australian monsoon (AAM), Intertropical Convergence Zone (ITCZ) migration, and the Indian Ocean Dipole (IOD), an ensemble of inter-related oceanic and atmospheric oscillations.

[3] SST records from the WPWP region (foraminifera transform function, alkenone unsaturation, and Mg/Ca) with millennial-scale resolution are few. The records indicate temperature decreases of ∼1 to 3°C during glacials [Anderson et al., 1989; Lea et al., 2000; Stott et al., 2002, 2007; Chen et al., 2005; de Garidel-Thoron et al., 2005, 2007]. δ18Oseawater estimates from planktic foraminifers in the WPWP, which are indicators of sea surface salinity (SSS) change [de Garidel-Thoron et al., 2007; Leduc et al., 2009], indicate different responses at millennial and orbital time scales. Lea et al. [2000] found that δ18Oseawater becomes lighter during glacials, which indicates a freshening of the WPWP. Stott et al. [2002] found the δ18Oseawater near the Philippine Sea to be heavier during Northern Hemisphere (NH) cold events and lighter during NH warm events, reflecting dryness during cold events and wet conditions during warm events. Zuraida et al. [2009] interpreted isotope data of planktic foraminifers living at different water depths in the Timor Sea as indicating that the Indonesian Throughflow weakened during Heinrich events (HEs) resulting from NH cooling or local hydrological changes. On land, high resolution humidification records [Turney et al., 2004] and geochemical data [Muller et al., 2008] from Lynch's Crater indicated high frequency dry-wet variation with wetter climate conditions during cold events, also interpreted by as an ENSO-like mechanism or ITCZ migration. Griffiths et al. [2009] analyzed stalagmite oxygen isotope records from the Liang Luar cave located on Flores Island, Indonesia, as revealing higher precipitation during the Younger Dryas (YD) event, also implying a southward migration of the ITCZ, and the influence of stronger winter AAM.

[4] Though the WPWP dynamics appears to be sensitive to the ENSO, AAM, ITCZ, and IOD, it is unclear what is the key factor driving the regional climate changes at the millennial time scale, and what is the most important process causing most of the dry or wet climate conditions during the NH millennial-scale cold events (e.g., HEs). Modeling studies indicate a southward migration of the ITCZ and high precipitation in the Southern Hemisphere (SH) in response to increased freshwater input to the North Atlantic, weakening the thermohaline circulation and increasing equator-to-pole temperature gradients [Dahl et al., 2005; Zhang and Delworth, 2005; LeGrande et al., 2006]. Stalagmite/speleothem records from Brazil also indicate high precipitation during HEs, supporting the ITCZ southward migration hypothesis [Wang et al., 2004; Cruz et al., 2005]. Furthermore, the study of Greenland δ18O and deuterium-excess records reveals that the Greenland moisture source shifted northward during interstadials and southward during stadials and HEs [Masson-Delmotte et al., 2005]. This pattern could also be explained by the migration of the ITCZ acting as a trigger for abrupt shifts of Northern Hemisphere atmospheric circulation [Steffensen et al., 2008]. The ITCZ migration seems to play an important role in tropical climate dynamics and could be connected to the North Atlantic during the NH cold periods at millennial time scales. The southern margin of the WPWP, located in the tropical SH, would be an ideal place to test the influence of ENSO-like oscillation or ITCZ migration on tropical dynamics during the NH's abrupt cold periods.

2. Data and Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[5] We analyzed a CALYPSO core MD052928 (11°17.26′S, 148°51.60′E, core length: 26.10 m, water depth: 2,250 m) retrieved from the southeastern PNG slope (Figure S1 of the auxiliary material) during the IMAGES (International Marine Past Global Change Study) PECTEN cruise in 2005. Located far away from the major fluvial sediment depocenter, the MD052928 core sediment is composed of a mixture of biogenic carbonate and siliciclastic detritus [Beaufort et al., 2005]. The top 6 m of the core was sampled at 5 cm intervals for a total of 120 samples. We used 10 AMS 14C dates supplemented by a curve of benthic oxygen isotopes (of Uvigerina spp.) to construct the age model. The BIT (branched and isoprenoid tetraether) index, 232Th activity and detritus fluxes were used as the terrestrial proxies. The unsaturated alkenones (U37k′) were used to reconstruct SST and the seawater oxygen isotopic composition (δ18Osw) was taken as a SSS proxy (the residual values after eliminating the temperature [derived from the U37k′-SST] [Bemis et al., 1998] and sea level effects [Thompson and Goldstein, 2005] from the oxygen isotopes of planktic foraminifers Globigerinoides ruber) (detailed information is given in the auxiliary material).

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[6] The SST record ranged from ∼26.5–28.5°C and showed obvious glacial-interglacial variation, with the coldest temperature found in the Last Glacial Maximum (LGM, 19–23 ka) (Figure 1). In the interval of Marine Isotope Stage (MIS) 3 (∼24–50 ka), U37k′-SST was ∼27.5°C on average. After the LGM, there was a short-lived higher temperature interval between ∼18–19 ka, and the U37k′-SST then increased continuously into the Holocene, reaching values above ∼28.5°C. Because the C37:3 alkenone compounds are not effectively synthesized by coccolithophorids at higher SSTs [Pelejero and Calvo, 2003], the C37:3 alkenone concentrations of some Holocene samples are below the detectable limit (U = 1). After eliminating the temperature and sea level effects from the oxygen isotopes of planktic foraminifers, our δ18Osw data do not show obvious glacial-interglacial variation but do reveal high frequency variations at the millennial time scale. We note that in the NH cold intervals (e.g., HEs, YD event), the δ18Osw data clearly indicate fresher sea surface conditions (Figure 1). Some fresh water events, for example, those in 17, 35, 45 ka, are not associated with major North Hemisphere cooling. These fresh water events may be attributed to the local hydrographic changes in the southern Papua New Guinea, and/or combined effects from high latitude North Hemisphere stadial climate.

image

Figure 1. Down core terrestrial proxies (detritus flux, BIT index, and 232Th activity flux), U37k′-SST, δ18Osw, planktic and benthic δ18O of core MD052928 and the LR04 Stack [Lisiecki and Raymo, 2005]. The red dash indicates the U37k′ = 1 meaning that the C37:3 alkenone is under detectable. The olive bar interval with the δ18Osw variation indicates the estimated error (0.25‰) of the δ18Osw. Black inverted triangles indicate the AMS 14C dates, and the purple asterisk indicates tie to LR04 stack in core MD052928. Cyan bars indicate the Younger Dryas and Heinrich events in the Northern Hemisphere.

Download figure to PowerPoint

[7] We also used 230Th-normalized detritus flux, 232Th activity and the BIT index as terrestrial proxies (Figure 2). The BIT index is the ratio of branched glycerol dialkyl glycerol tetraether (GDGT) derived from soil to crenarchaeol from marine environment; it is used as a proxy for terrestrial organic matter input [Hopmans et al., 2004]. The error of the BIT index in our analysis is 0.014. 232Th mainly comes from the continent and could be use as the terrestrial lithogenic proxy [Huh and Kadko, 1992]. 230Th-normalized method is used in this study for estimating the total sedimentary flux and the detritus flux could be estimated by multiplying the total flux by the detritus content [François et al., 2004].

image

Figure 2. NGRIP δ18O [North Greenland Ice Core Project Members, 2004], Sanbao-Hulu cave δ18O record [Wang et al., 2008], δ18Osw of core MD982181 [Stott et al., 2002], 232Th activity flux and δ18Osw of core MD052928, peat humification absorption from Lynch's Crater [Turney et al., 2004]. The olive bar interval with the δ18Osw variation indicates the estimated error of the δ18Osw. Black inverted triangles indicate the AMS 14C dates, and the purple asterisk indicates tie to LR04 stack in core MD052928. Cyan bars indicate the Younger Dryas and Heinrich events in the Northern Hemisphere. The solid blue and red arrows indicate the wetter and drier climate conditions, respectively.

Download figure to PowerPoint

[8] All these proxies indicate high terrestrial input in MIS 3, the LGM and the last deglaciation with high frequency changes at millennial time scales. The detritus flux values were ∼1.5–4 g cm−2 ky−1 in MIS 2–3 and ∼1–1.8 g cm−2 ky−1 in the Holocene. The fluxes of 232Th were ∼1.5–3.5 dpm cm−2 ky−1 in MIS 2–3 and ∼1–2 dpm cm−2 ky−1 in the Holocene. The BIT index ranged from ∼0.05 to 0.3. In MIS 2 ∼ 3, the BIT index was ∼0.1–0.3 with high frequency variations. In the last deglaciation, there was a high BIT plateau between ∼16–10 ka, where BIT ranged from 0.12 to 0.22 and correlated well with other terrestrial proxies in showing more terrestrial input during that period (Figure 1). Our SSS (δ18Osw) proxy, however, indicates relatively fresher water conditions in the Holocene and in millennial scale events in MIS 2–3. We note that the Holocene values of δ18Osw could be underestimated by the saturated ratio of U37k′, which is close to the detection limit (U37k′ = 1).

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[9] Our terrestrial proxies showed higher terrestrial input in MIS 2 and 3 than that in the Holocene, with high millennial-scale variability (Figure 1). Potential controlling factors for the changing terrestrial input include sea level, fluvial discharge, and eolian transport. During MIS 2, sea level was ∼120 m below the present level, resulting in increased glacial-age terrestrial input. During MIS 3, sea level was around 80 ∼ 60 m below the present level [Thompson and Goldstein, 2005]. However, the high terrestrial input in MIS 3 and MIS 2 cannot be fully explained by sea level change alone. The Australian continent is the largest eolian dust source in this region, but the eolian dust transport paths are toward the southeast and northwest [Mackie et al., 2008], having little effect on the location of MD052928. Therefore the BIT index, detritus and 232Th activity fluxes likely reflect terrestrial input from fluvial transport. However, based on our δ18Osw estimation the precipitation shows no significant change during the glacial-interglacial time period. Pollen data from the highlands of PNG also indicate cooler climate conditions and the strengthened formation of a cloudy belt during the MIS 2 and MIS 3 associated with the maintenance of the ITCZ close to PNG [Haberle, 1998]. At the glacial-interglacial time scale, the precipitation of PNG may not change much. However, at the millennial time scale, we observe high frequency oscillations in our terrestrial proxies and δ18Osw during MIS 2 and MIS 3 that may indicate more abrupt changes in precipitation, resulting in variations in terrestrial input (Figure 1). In addition, expanded mountain glaciers in PNG during MIS 2 or MIS 3 [Hastenrath, 2009] also may have driven increased fluvial transport during the austral spring. Thus, the high frequency component of terrestrial input changes to MD052928 is likely controlled by fluvial transport related to precipitation.

[10] The BIT index is calculated based on branched GDGTs that are mainly produced by the bacteria from the soil or peat bogs, but also in the coastal marine sediments and lake sediments under anaerobic conditions [Weijers et al., 2006]. We consider that the BIT is possibly complicated by soil erosion rate, the existence of the peat bogs, biodegradation, or the marine crenarchaeota productivity. The complication is shown in the low or a stable BIT value in H3 and H4 (Figure 1), which is inconsistent with those shown by 232Th and detritus flux estimates.

[11] The δ18Osw record shows noticeable high frequency variations at the millennial time scale. We found that MD052928 δ18Osw and terrestrial records are well correlated with Lynch's Crater precipitation-related proxies [Turney et al., 2004; Muller et al., 2008], especially between 30 ∼ 40 ka and at HE2, HE3, HE4 and the Bølling-Allerød interstadial to the YD (Figure 2). These results indicate high precipitation in the southwestern tropical Pacific during high latitude NH cold periods. The modern precipitation at Lynch's Crater is sensitive to ITCZ north-south migration and ENSO events. Under modern ENSO conditions, precipitation in the WPWP area is decreased. Stott et al. [2002] proposed an ENSO-like mechanism to explain the dry climate condition during NH cold events. In contrast, our records and Lynch's Crater indicate wet conditions during NH cold events. Furthermore, the modern PNG highland area has high precipitation year-round, but in the southeastern lowlands the precipitation has strong seasonality, with increased precipitation in the austral summer when the ITCZ was located more southward. Thus based on our data, we suggest that the ITCZ moved southward in this part of the western Pacific during these NH cold periods.

[12] Recently, a study has been conducted [Leduc et al., 2009] for comparing the δ18Osw record of core MD022529 (08°12.33′N; 84°07.32′W, Figure S1) from the East Pacific Warm Pool, MD982181 [Stott et al., 2002] and the humidification record from Lynch's Crater [Turney et al., 2004]. In the modern ENSO, the climate conditions (SST and SSS) at sites MD022529 and MD982181 are out of phase. If the ENSO-like mechanism is the best analogue to explain the climate shifting patterns at the millennial time scale, we should observe opposite climate conditions at these two sites. However, they found that the δ18Osw variability of these 2 cores, distributed on the eastern and western sides of the northern tropical Pacific, indicates wet conditions during warm events and dry conditions during cold events, offering no support to the ENSO-like scenario. In contrast, the record from Lynch's Crater indicates the reverse conditions at the two sites. These results are consistent with an explanation that involves the southward migration of the ITCZ, resulting in wetter tropical SH during the North Atlantic cool periods. Stalagmite δ18O from Liang Luar cave in southern Indonesia indicates high precipitation in the YD event [Griffiths et al., 2009]. The southern Brazilian speleothem records also indicate higher precipitation during the HEs [Wang et al., 2004; Cruz et al., 2005], both supporting the scenario of ITCZ migration.

[13] The tropical SH precipitation records and the Sanbao-Hulu cave stalagmite record [Wang et al., 2008] are out of phase (Figure 2). The cave δ18O record and the tropical north Pacific δ18Osw data [Leduc et al., 2009] all indicate a reverse precipitation pattern relative to our data from MD052928 and Lynch's Crater's records. We may reconcile these out-of-phase hemispheric signals by the north-south migration of the ITCZ. Griffiths et al. [2009] suggest that lighter δ18O from the Liang Luar stalagmite during the YD event reflects an ITCZ southward migration, possibly linked to a stronger winter AAM. MD052928 and Liang Luar cave are located in the same range of latitudes (11°S and 8°S, respectively), which is under the influence of the AAM. Modern observation of the East Asian winter monsoon (EAWM) moisture transport indicates that in the month before the monsoon's onset, the moisture flux between the surface and 450 hPa shows that moderate westerly transport associated with cross-equatorial flow in the Sulawesi Sea and west of Borneo resulted from the EAWM. After the EAWM onset, the strengthened westerly moisture transport was dramatically constrained in a zonal belt stretching from the Timor Sea to the Western Equatorial Pacific, between the latitudes 5°S and 15°S, and associated with the ITCZ southward shift and deepening of the monsoon trough [Godfred-Spenning and Reason, 2002]. In addition, the stronger EAWM is also related to cold events in the North Atlantic linked by the effect of westerly winds [Porter and An, 1995], thus we might link WPWP change to the North Atlantic dynamics through the atmospheric circulation of the westerly-EAWM-AAM system. If so, abrupt climate changes in the WPWP and tropical southwestern Pacific may also be linked to the high latitude NH, coupled through the AAM circulation systems and the ITCZ migration.

5. Conclusion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[14] Terrestrial proxies of core MD052928 indicate higher fluxes due to fluvial discharges in the PNG region during the last glacial period. The δ18Osw and terrestrial proxies during the last glacial period indicate stronger precipitation intensity in the PNG region, which is likely associated with ITCZ migration southward during the NH cold periods (e.g., YD, HE2, HE3, and HE4) on the millennial time scales. We suggest the AAM systems and the ITCZ migration are the major mechanisms connecting the WPWP and NH high latitudes on millennial time scales.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

[15] We thank two anomalous reviewers for their constructive suggestions, and N.R. Wan and H. Li in Department of Earth Sciences, National Cheng Kung University for their help in processing the oxygen isotope data. This research was supported by the National Science Council (NSC97-2611-M-001-002-MY3 to CAH, NSC 97-2611-M-019-001 and 97-2116-M-019-003 to MTC, and NSC 97-2917-I-019-103 to LJS) and National Taiwan Ocean University, Republic of China, to MTC.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Data and Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusion
  8. Acknowledgments
  9. References
  10. Supporting Information

Auxiliary material for this article contains a text file and two figures.

Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select “Save Target As…” (PC) or CTRL-click and select “Download Link to Disk” (Mac).

Additional file information is provided in the readme.txt.

FilenameFormatSizeDescription
grl27562-sup-0001-readme.txtplain text document2Kreadme.txt
grl27562-sup-0002-txts01.pdfPDF document164KText S1. Additional information about the regional setting and methods.
grl27562-sup-0003-fs01.pdfPDF document116KFigure S1. Pacific annual mean precipitation rates derived from the CPC Merged Analysis of Precipitation dataset for a long-term mean from the 1979 to 2000 interval and the annual SST of the Pacific derived from the WOA05.
grl27562-sup-0004-fs02.pdfPDF document183KFigure S2. ENSO precipitation anomalies for all calendar months combined.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.