Planktonic Foraminiferal Assemblage Variations of Ontong-Java Plateau during Late Quaternary and Their Implications for Paleotemperature in the Western Pacific Warm Pool

INTRODUCTION

Modern ocean-atmosphere investigation shows that the equatorial Pacific encompasses the warmest sea-surface temperature (SST) (> 28 ℃) area in the world. It has a strong influence on global heat and water vapor transport, and so is the "engine" of the climate system. The western Pacific warm pool (WPWP) and the high northern Atlantic constitute the pivotal areas of global climate change, such as the Asian monsoon and the El Nino-Southern Oscillation. Moreover, the modern equatorial Pacific Ocean shows an east-west gradient in thermocline depth (DOT) and sea-surface temperature (SST). Any fluctuation of the upper ocean structure in the WPWP will result in changes to eradiated heat and water vapor fluxes to the atmosphere, which will affect the global climate system (wang, 1998; Webster et al., 1998).

Previous studies of Ocean Drilling Program (ODP) Leg 130 have been focused on paleoceanographic events since the Tertiary, using proxies such as the stable oxygen isotope of planktonic foraminiferal shells, but never dealing with reconstructions of SST and the upper ocean vertical structure. Here we present new records of the SST and DOT variability at Site 807A for the past 530 ka. These have been reconstructed from the faunal counting of the planktonic foraminifera at this site, and reveal the glacial-interglacial changes of the WPWP climate in the late Quaternary periods.

MATERIAL AND METHODS

ODP Site 807A is located on the northern rim of the huge Ontong-Java plateau, western tropical Pacific (3°36.42′N and 156°37.49′E, water depth 2 803.8 m) (Fig. 1). The sediments are well preserved, mainly consisting of gray-white silty foraminiferal ooze with some bio-disturbances. A total of 165 samples were taken from core 1H and 2H, covering the upper 8.37 m, with a sampling interval of 5 cm. The planktonic foraminifera were only picked from the > 0.154 mm size fraction. In cases of abundant foraminifera in a sample we used the bi-separation method to split the sample, to yield a subsample containing at least 200 specimens. The planktonic foraminifera were identified and counted. Some species, like Globorotalia truncatulinoides and Globorotalia pachyderma, were divided into two groups, the left coiling and the right coiling according to their chambers arrangement. On the basis of the census data the relative abundance of each species and fragmentation of each sample were calculated. We estimated the SST using the FP-12E planktonic foraminiferal transfer function (Wang et al., 1989; thompson, 1981), and the DOT with the thermocline depth transfer function of Andreasen and Ravelo (1997). Equation (1) was used to calculate the fragmentation

Figure  1.  Location of Site 807A of ODP Leg 130.

DownLoad: Full-Size Img PowerPoint

where F is the number of fragments and W is that of well-preserved planktonic foraminifer in the sample (Le and Shackleton, 1992). We used "SPECTRUM" (Schulz and Stattegger, 1997) to do the spectral analyses. The advantage of SPECTRUM is the avoidance of any interpolation of the time series.

CHRONOLOGICAL FRAMEWORK

Based on calcareous nannofossil biostratigraphy (Prentice et al., 1993), we supplemented two planktonic foraminiferal biostratigraphic datum levels to build up the bio-chronological framework of Site 807A: the pink Globigerinoides ruber last appearance datum (LAD) (1.92 cmbsf, at 0.12 Ma) and first appearance datum (FAD) (6.67 cmbsf, at 0.40 Ma). The average sedimentary rate at this site is ~15.6 m/Ma (Prentice et al., 1993), thus the whole record studied spans the past 0.53 Ma. The more detailed chronology is based on the δ¹⁸O stratigraphy of this site (Prentice et al., 1993), as seen in Fig. 2.

Figure  2.  Age-depth plot showing sedimentation rate of upper 8.37 m at Site 807A, compared with the oxygen isotope curve on the right side (from Prentice et al. (1993)). Ⅰ. pink Globigerinoides ruber (F) LAD; Ⅱ. Emiliania huxleyi (N) FAD; Ⅲ. Pseudoemiliania lacunosa (N) LAD; Ⅳ. pink G. ruber (F) FAD; Ⅴ. estimated age of the bottom sample. The numbers on the right denote marine isotope stages.

DownLoad: Full-Size Img PowerPoint

RESULTS AND DISCUSSIONS

Carbonate Dissolution and Preservation

Planktonic foraminiferal fragmentation and coarse fraction (> 0.63 mm) indicate deep-sea carbonate dissolution and carbonate preservation respectively. Our results show that the coarse fraction percentage at Hole 807A for the past 530 ka is larger during glacials thanduring interglacials, except for MISs (marine isotope stages) 1 and 2 (Fig. 3), which displays a typical Pacific type carbonate preservation pattern and basically agrees with the results of Prentice et al. (1993) at this site (see Fig. 3a).

Figure  3.  Comparison of coarse fraction (a), fragmentation (b) and δ¹⁸O (c) curves. The dashed curve in (a) is based on the data of Prentice et al. (1993), while the dashed line in (b) shows the MIS 7/8 boundary (about 240 ka). Numbers in (c) denote marine isotope stages.

DownLoad: Full-Size Img PowerPoint

The average fragmentation percentage at this site is 16.5 %, indicating good carbonate preservation in the sediments. Figure 3 clearly shows that MIS7/8 (at ~240 ka) is a significant boundary. Above this boundary, the fragmentation percentage from MIS1 to MIS7 is higher during interglacials but lower during glacials, negatively well correlated with the coarse fraction variability (Fig. 3), which indicates that during interglacials the carbonate dissolution was strong but the preservation was bad, whereas the opposite pattern was shown during glacials. Below this boundary from MIS8 to MIS13, the fragmentation percentage is higher during glacials but lower during interglacials, in accord with the coarse fraction variability. The changes during this time interval might result from the productivity enhancement of the sea-surface water during glacials or from the transformation of the ratio of dissolution-accepted and dissolution-resistant species due to changes in sea-water structure.

Sea Surface Temperature (SST)

Among the planktonic foraminifera in the sediments of Hole 807A, 28 species have been identified. Most of these species belong to tropical-subtropical genera and a few are temperate-polar ones. Species with average content > 10 % include Globigerinoides ruber, Globigerinita glutinata, and Pulleniatina obliquiloculata, among which G. ruber has the highest average percentage of up to 17 %.

Planktonic foraminiferal assemblages have been used as a good indicator of paleotemperature variations. The forams have been divided into two suits, the cold-water species and the warm-water species. The cold-water species include Neogloboquadrina dutertrei, Globigerina bulloides, Globorotalia crassaformis and Globorotalia inflate, and the warm-water species include Globigerinoides ruber, Globigerinoides sacculifer, Globigerinita glutinata and Pulleniatina obliquiloculata.

In general, warm-water species except for G. ruber show a high percentage in interglacials but a low percentage in glacials, whereas G. glutinata seems to have had an opposite glacial/interglacial pattern above and below the MIS7/8 boundary. Above this boundary the G. glutinata content was higher during glacials but lower during interglacials, whereas below this boundary its content showed the opposite pattern. The cold-water species content showed an opposite pattern to that of the warm-water species, remaining higher during glacials but lower during interglacials. In addition, G. bulloides showed a higher content during interglacials but a lower content during glacials below the MIS7/8 boundary, as seen in Fig. 4.

Figure  4.  Variations in the relative abundances of planktonic foraminiferal warm-water species (G. ruber, G. sacculifer, G. glutinata, P. obliquiloculata) (a) and cold-water planktonic foraminiferal species (N. dutertrei, G. bulloides, G. crassaformis, G. inflate) (b), compared with the δ¹⁸O curve.

DownLoad: Full-Size Img PowerPoint

The sea-surface temperature at Site 807A including summer SST, winter SST and average annual mean SST, as well as the seasonality, has been reconstructed by using the planktonic foraminiferal transfer function FP-12E (Fig. 5). The standard errors of the FP-12E estimated SST are 1.46 ℃ and 2.48 ℃ for summer and winter respectively. The estimated core top SST at this site is 29.5 ℃ for summer and 27.3 ℃ for winter. Their differences with the present day SST (29.6 ℃ for summer and 29.0 ℃ for winter, Levitus and Boyer, 1994) are within the standard errors of the estimation. In addition, the average communality of 0.85 indicates that the transfer function technique can interpret 92 % of the total variance in the observed faunal information. These demonstrate that the FP-12E-derived SST reconstruction at Site 807A for the past 530 ka is quite reliable.

Figure  5.  FP-12E-derived summer, winter and annual mean SSTs and seasonality curves of Site 807A.

DownLoad: Full-Size Img PowerPoint

For the past 530 ka, the summer SST has varied between 28.0 ℃ and 30.2 ℃ and the winter SST hasvaried between 22.8 ℃ and 29.3 ℃. The amplitude of the winter SST is ~4.7 ℃, much larger than that of the summer SST. Since MIS2, the largest amplitude of the winter SST has reached up to 4.1 ℃, which is inconsistent with a previous study by Thunell et al. (1994) who reported a relative stability of the SST during glacial-interglacial cycles in the western equatorial Pacific. In our research, the amplitude of the MIS1/MIS2 SST variations reached up to 5.4 ℃, and that of the MIS5/MIS6 SST variations reached up to 6.0 ℃, which is consistent with the Mg/Ca ratio deduced SST variations in the Indo-Pacific warm pool area (Visser et al., 2003). Lea et al. (2000) reconstructed the SST fluctuations in the equatorial Pacific over the last 450 ka by using Mg/Ca ratios of planktonic foraminiferal shells, and found that tropical Pacific sea-surface temperatures (SSTs) were (2.8±0.7) ℃ colder than the present at the last glacial maximum, and the glacial-interglacial temperature differences could be as great as 5 ℃ over the last 450 ka. Such great SST changes are also obvious in our reconstructed SST records. The average glacial/interglacial amplitude of the annual mean SST oscillations was as much as 4.2 ℃, and the average annual mean SST was 26.7 ℃ during glacials, roughly 2.6 ℃ colder than the present SST. Therefore these results indicate strong instability of the western Pacific warm pool climate during glacial/interglacial fluctuations in the late Quaternary period.

Depth of Thermocline (DOT)

Modern tropical ocean studies suggest that planktonic foraminiferal abundance is controlled primarily by the upper ocean temperature and nutrient gradient (Ravelo et al., 1990; Fairbanks and Wiebe, 1980). In low latitude oceans, thermocline-dwelling species such as Globorotalia, Neogloboquadrina and Pulleniatina mainly live in deep water whereas the mixed layer-dwelling species such as Globigerinoides and Globigerinita occupy upper ocean water above the thermocline. When the DOT becomes deeper, the mixed layer-dwelling species increase in abundance while the thermocline-dwelling species decrease in abundance. Therefore, the variations of the abundance of the mixed-layer dwelling species and the thermocline dwelling species can reveal the change of the upper ocean vertical structure (Jian et al., 2000; Ravelo et al., 1990).

As seen in Fig. 6, the glacial/interglacial abundance of the mixed layer-dwelling species at Site 807A has changed in a pattern which is opposite to that of the thermocline-dwelling species for the past 530 ka. Usually during interglacial episodes, the mixed layer-dwelling species decrease while the thermocline-dwelling species increase, which indicates the shoaling of the DOT. During MIS11 the abundance of the mixed layerspecies and the thermocline species reached the minimum and the maximum values respectively. Because G. conglomerata usually lives in an environment with water depth deeper than 100 m (Bé, 1977) and a winter SST of about 26 ℃ (Coulbourn et al., 1980), its abundance variation is often used to imply the change of the thermocline depth. The increasing percentage of this species indicates the shoaling of the DOT. During MIS5, 9, 11 and 13, G. conglomerata had higher abundance values, implying a shallow DOT during these periods (Fig. 6). Comparing the G. conglomerata abundance in MIS5 with the Nansha Islands of the South China Sea (Li et al., 2001) it is clear that this species was relatively abundant in the Ontong-Java plateau of the open western Pacific. This is partly because of the location of ODP Site 807A, which is in the center of the WPWP where the higher winter SST is an advantage to the existence of this species.

Figure  6.  Changes of DOT, relative abundances of shallow-water-dwelling species (G. ruber, G. sacculifer, G. glutinata), deep-water-dwelling species (P. obliquiloculata, G. menardii, G. inflate, N. dutertrei) and G. conglomerata at Site 807A. DOT was estimated by using the transfer function of Andreasen and Ravelo (1997). The dashed line denotes the MIS7/8 boundary (at ~240 ka).

DownLoad: Full-Size Img PowerPoint

Figure 6 also shows the detailed fluctuations of the DOT derived from Andreasen and Ravelo's transfer function (1997). The estimated DOT of core top samples agrees with the observed modern annual mean DOT in this region (Levitus and Boyer, 1994). The DOT variability has ranged from 102 m to 212 m for the past 530 ka, averaging 164 m. The amplitude of the glacial/interglacial variation is ~110 m, larger than the standard error (±27 m) of the transfer function. The shift between MIS7 and 8 is also an important boundary for DOT variations (Fig. 6). From MIS1 to MIS7, the amplitude of the glacial/interglacial DOT variability has remained small, with an average depth of ~177 m, deep in interglacials but shallow in glacials. From MIS8 to MIS13, the amplitude of the glacial/interglacial DOT variations became bigger, with an average DOT of about 152 m, which is shallower than that in MIS1-7. During these time intervals, the DOT became deep during glacials but shallow during interglacials. After the shift from MIS7 to MIS8, the DOT has gradually become deeper, which hints that the warm-water layer has been getting thicker and thicker in the WPWP area. This certainly highlights this special region's responsibility for the global climate system.

Periodic Variations in SST and DOT

Spectral analyses were performed to study the periodic variability of the winter SST and DOT in the open western Pacific for the past 530 ka. As shown in Fig. 7, the eccentricity-dominated (~100 ka) periodicity signal is evident in both the winter SST and DOT variations, while the obliquity-related (~41 ka) response is weak and only occurs as a 37.4 ka cycle in DOT variability. The precession-related cycle is lacking in both records, taken place by a distinct 30 ka cycle. In the spectrum of the winter SST records, three peaks of 13.5 ka, 11.2 ka and 9.5 ka are particularly significant, and should be regarded as the response of the winter SST to semi-precession (~10 ka). The spectral analyses at Site 807A on the winter SST and DOT records indicate that the SST variations in the study area have been influenced not only by high northern latitude forcing but also by tropical forcing, of which the typical spectrums are 100 ka and 10 ka respectively.

Figure  7.  Spectral analyses of SST (a) and DOT (b) for the last 530 ka. The numbers inside the frame indicate the periodicities (ka), while the horizontal dashed lines denote the 80 % statistic level of the spectrum. Only the peaks above this level are significant. Crosses in lower left corner mark the 6 dB bandwidth (horizontal) and the 80 % confidence interval (vertical).

DownLoad: Full-Size Img PowerPoint

CONCLUSION

(1) Core 807A in the Ontong-Java plateau provides us with continuous planktonic foraminiferal records for the past 530 ka. The coarse fraction and fragmentation records reflect carbonate deposition and dissolution in the western equatorial Pacific. The coarse fraction percentage appears to be higher during glacials but lower during interglacials, representing the typical Pacific carbonate preservation pattern. However, the fragmentation records at this site display opposite glacial/interglacial patterns before and after MIS7/8 boundary (at ~240 ka), which indicates that the glacial carbonate dissolution was much stronger from MIS8 to MIS13, but weak from MIS7 to the present.

(2) Based on the planktonic foraminiferal transfer function FP-12E, we found that the amplitude of glacial-interglacial annual mean SST variability could reach up to 4.2 ℃ for the past 530 ka in the western equatorial Pacific, which reveals the instability of the WPWP SST variability on glacial/interglacial timescale.

(3) From 530 ka to present, the DOT has deepened by about 100 m, which implies the instability of the upper ocean structure variability in the WPWP area in the late Quaternary period. From MIS8 to MIS13, the amplitude of the DOT variability was big and the WPWP showed a deep thermocline during glacial periods; whereas after MIS7, the amplitude of the DOT variability became small and the thermocline has gradually become deeper and deeper.

(4) Spectral analyses of the winter SST and DOT variability in the past 530 ka reveal two dominant cycles in the open western Pacific, the 100 ka late Quaternary glacial-interglacial cycle and the semi-precessional cycle which is dominant in the tropical spectrums. This indicates that the study area has been influenced by both high northern latitude forcing and tropical forcing.