Interglacial Hydroclimate in the Tropical West Pacific Through the Late Pleistocene

See allHide authors and affiliations

Science  08 Jun 2012:
Vol. 336, Issue 6086, pp. 1301-1304
DOI: 10.1126/science.1218340

The Rains of Change

Approximately 500 to 400 thousand years ago, a fundamental change in the nature of glacial cycles occurred, including a shift to larger amplitude variations in atmospheric carbon dioxide concentrations and global temperatures. Meckler et al. (p. 1301, published online 3 May; see the Perspective by Kurita) present a stalagmite record from Borneo which provides a record of precipitation from 550,000 to 200,000 years ago. The amount of regional precipitation was similar during all four of the interglacials during that interval, and drying events during glacial terminations corresponded with cooling events in high northern latitudes.


Records of atmospheric carbon dioxide concentration (Pco2) and Antarctic temperature have revealed an intriguing change in the magnitude of interglacial warmth and Pco2 at around 430,000 years ago (430 ka), but the global climate repercussions of this change remain elusive. Here, we present a stalagmite-based reconstruction of tropical West Pacific hydroclimate from 570 to 210 ka. The results suggest similar regional precipitation amounts across the four interglacials contained in the record, implying that tropical hydroclimate was insensitive to interglacial differences in Pco2 and high-latitude temperature. In contrast, during glacial terminations, drying in the tropical West Pacific accompanied cooling events in northern high latitudes. Therefore, the tropical convective heat engine can either stabilize or amplify global climate change, depending on the nature of the climate forcing.

The study of past interglacial climates provides valuable insight into natural climate variability under conditions roughly similar to those of the present day, but with differences in atmospheric carbon dioxide concentration (Pco2) (1) and astronomic forcing. Especially intriguing is a step-like increase in interglacial Antarctic temperature (2), global ice melting (3), and Pco2 (4, 5) that took place at roughly 430,000 years ago (430 ka) and is referred to as the Mid-Brunhes Event (MBE). To understand the underlying cause of this shift toward warmer interglacials, it is crucial to establish the global signature of interglacial climates during that time, but sufficiently long records, especially from low latitudes and terrestrial archives, are sparse. Because Pco2 and ice cover affect the planet’s radiative balance, a global change in interglacial climate at the MBE might be expected. Although many marine sea surface temperature (SST) reconstructions from around the globe show qualitatively similar changes to those observed in Antarctica (6), the few available lower-latitude SST reconstructions (79) from this time period do not show the same consistent increase in post-MBE interglacial temperatures. Given that many models simulate strong dynamical links between high-latitude and tropical climate [e.g., (10)], more tropical records are required to test and improve our understanding of low-latitude climate sensitivity to a range of different forcing factors.

Here, we present stalagmite records from northern Borneo (4°N, 115°E; fig. S1) that span several glacial-to-interglacial cycles and include the MBE. We reconstruct Warm Pool hydroclimate from the oxygen isotope (δ18O) variability recorded in three stalagmites with overlapping growth periods. Borneo lies in the West Pacific Warm Pool, which is an important heat and moisture source for higher latitudes (11). Although there is little seasonality in modern rainfall amount at our site, the isotopic composition of precipitation varies seasonally by up to 2 to 3‰ (12), with the most enriched δ18O values occurring in February and March, when the intertropical convergence zone (ITCZ) is furthest south of our site. However, on interannual time scales, the El Niño–Southern Oscillation (ENSO) phenomenon dominates both precipitation amount and δ18O, with El Niño events resulting in large-scale drying and rainfall δ18O that is roughly 5 to 6‰ higher than average (12). Two recent studies of water isotopes in the same area conclude that regional (much more than local) rainfall amount is the main driver for rainfall δ18O variability on intraseasonal and longer time scales today (13) and during modeled millennial-scale events in the past (14).

We analyzed stalagmites from three different caves in Gunung Mulu and Gunung Buda National Parks, each about 5 to 10 km apart. The stalagmites were dated by 238U-234U-230Th measurements (15) and cover the time period from 570 (±20) ka to 210 (±10) ka [from marine isotope stage (MIS) 13 to the beginning of MIS7], with some gaps due to hiatuses. Sample GC08 covers the whole time period, whereas samples Squeeze1 and WR5 yield shorter records that overlap GC08. In comparison to other cave systems, dating of these stalagmites is challenging due to their unusually low 234U/238U ratios and/or low U concentrations (fig. S4 and tables S1 to S3). The old age of the samples exacerbates these dating issues. Furthermore, there is evidence that slight open-system exchange of U has affected sample GC08, requiring us to model the ages in this stalagmite by assuming a constant initial 234U/238U ratio (15). Varying detrital 230Th contamination cannot explain the age reversals in this stalagmite, owing to the short half-life of 230Th compared to the age of our samples (15). The error assigned to the assumed constant initial 234U/238U ratio leads to increased age uncertainty in the GC08 chronology, with individual age errors ranging from ±9200 to ±13,500 years (2σ; table S1), limiting our ability to compare the millennial-scale timing of climate change in Borneo with other records. The age constraints are sufficient, however, to identify each interglacial stage, which is the main focus of this study. In sample WR5, open-system alteration might also have occurred but cannot be corrected for (15). The δ18O record from this stalagmite therefore only serves as a check on the overall structure of the GC08 δ18O record. In contrast, sample Squeeze1 yielded a reliable age model without age reversals and can be used to confirm that the stalagmite δ18O records are not affected by the process(es) that altered the U isotopic composition in GC08 (and potentially WR5).

Samples for oxygen isotope analysis (15) were drilled along the stalagmite growth axes at 1-mm increments, which on average corresponds to 75 years in Squeeze1, 650 years in GC08, and 1000 years in WR5. Equilibrium calcite precipitation with respect to δ18O is supported by several lines of evidence (15) (fig. S2), most importantly by the good reproducibility (within age uncertainty) of the δ18O signal in the three stalagmites from different caves (Figs. 1 and 2 and fig. S3). Hence, the variations in carbonate δ18O are expected to faithfully record changes in the δ18O of drip water (modulated by any changes in the equilibrium carbonate-water fractionation due to changes in cave temperature). The reproducibility of the carbon isotope (δ13C) records from Borneo is generally much poorer (fig. S3), suggesting that this proxy is governed by cave-specific factors (16), with climate exerting a lesser influence.

Fig. 1

Borneo stalagmite δ18O records. Data from this study are from samples GC08 (dark blue), Squeeze1 (light blue), and WR5 (green), with age markers and 2σ error bars shown below. Also plotted is the published record for the last glacial termination from the same location (17) (yellow-orange lines). Interglacial marine isotope stages (MIS) and glacial terminations (T-) are indicated at the top, and H1 is Heinrich event 1. Light and dark gray bars on the right show range of interglacial and glacial values, respectively, and black arrows highlight δ18O maxima interpreted as deglacial drying phases. The age model of GC08 is based on the assumption of constant initial δ234U, to correct for open-system behavior (15). For GC08, the age error bars therefore include the assigned error of the initial δ234U of ±12‰, in addition to analytical errors. In the younger part of the GC08 record, the range of age errors is indicated by smallest and largest error bar below age markers. The orange double-arrow highlights a large age reversal in WR5 (compare to fig. S5). Within age uncertainty, the δ18O signal replicates well among the three different caves. Both trends and absolute values are similar to the record for T-I.

Fig. 2

Comparison of the Borneo stalagmite δ18O record with other climate records. (A) Stalagmite record with adjusted age models for GC08 and WR5 (see below); (B) the precession parameter (30); (C) atmospheric Pco2 (dark gray) (4, 5) and Antarctic δD, a proxy for temperature (light gray) (2); (D) global stack of benthic δ18O (3) as a proxy for global ice volume; (E) West Pacific Warm Pool SST reconstructed from Mg/Ca in foraminifera from the Ontong Java Plateau (dark purple) (7) and off Indonesia (light purple) (9); (F) East Asian Monsoon record from Sanbao/Linzhu caves (18). Indicated on top are interglacial marine isotope stages (MIS) and glacial terminations (T-). The age model of GC08 has been slightly modified around T-IV (+5 ky, within 1σ age errors) to better align the δ18O peak to the decrease in benthic δ18O. The WR5 record, which has the largest age uncertainty of our three records, has been tuned to the GC08 record to facilitate comparison (15). Unlike Pco2, ice volume, and Antarctic temperature, the climate in Borneo does not show a change in the level of peak interglacial values from MIS13 to MIS11. Maximum δ18O values during terminations resemble similar events in China (weak monsoon intervals, WMI), with the signal standing out more in the Borneo record. Note the apparent trend in magnitude of these deglacial signals in Borneo from T-V to T-III, which resembles the trend in ice volume during preceding glacial periods.

The stalagmite δ18O time series reveal large variations with maximum peak-to-peak differences of 3.0 to 4.3‰. Both the peak-to-peak differences and the absolute values of δ18O are similar to those of the Borneo stalagmite record spanning the last glacial termination (17) (Fig. 1, yellow-orange lines). During the last deglaciation, the highest δ18O values were observed just after the Last Glacial Maximum, during the time of Heinrich Event 1, and were interpreted as reflecting strong regional drying (17). Although their exact timing and duration are not resolvable by our age model, we interpret the pronounced δ18O maxima (–5.6 to –6.6‰) in the longer record likewise, as deglacial drying phases that occurred after glacial maxima. Apart from these distinct deglacial periods, maximum glacial δ18O values are –7.2 to –7.3‰ (table S4), similar to values for the Last Glacial Maximum. Interglacial periods are characterized by minimum δ18O values of –9.3 to –9.5‰, comparable to values for actively growing stalagmites from northern Borneo caves.

Glacial-to-interglacial (G-IG) changes of 2.0 to 2.2‰ in stalagmite δ18O represent a combination of changes in temperature and the δ18O of precipitation (table S4). G-IG temperature-change estimates in the surrounding warm pool are 1.0° to 4.5°C during the same time interval (7, 9) (Fig. 2E), which would only account for variations in stalagmite δ18O of 0.2 to 0.9‰, if cave temperature change was the same. Therefore, most of the G-IG changes in stalagmite δ18O must be due to changes in the δ18O of precipitation, reflecting changes in source water δ18O, atmospheric moisture transport pathway, or regional precipitation amount. Reconstructed glacial enrichment in source water δ18O was +0.6 to +1.3‰ (7, 9) and is poorly constrained (fig. S9 and table S4). On glacial-interglacial time scales, moisture transport to Borneo was likely influenced by the varying exposure of the adjacent Sunda Shelf (fig. S1), but the effect on the δ18O of rainfall is difficult to constrain. The probably increased length of moisture transport pathways, especially over land, could have led to isotopically more depleted rainfall during glacial maxima, counteracting temperature and source water δ18O changes. Although the magnitude therefore remains uncertain, we interpret the stalagmite δ18O record as reflecting changes in regional hydrology, with drier glacial periods (higher δ18O) compared to interglacials (lower δ18O).

The overall structure of the Borneo record suggests a strong precession influence (Fig. 2), similar to that in records of the East Asian and Australian-Asian summer monsoons (18, 19). The phasing of the Borneo stalagmite δ18O record with respect to precession cannot be constrained with confidence, given the age constraints. Of particular note, the new records clearly reflect a diminished precessional cyclicity between 460 and 360 ka, when precessional forcing was weak. The 100,000-year (100 ky) glacial-interglacial cycle is also apparent in Borneo, most prominently in the pronounced deglacial δ18O maxima. However, the dominance of the precession signal indicates that regional changes in the seasonal distribution of insolation have a large influence on tropical West Pacific hydroclimate, compared to the effects of high-latitude climate change, which is characterized by strong 100-ky “sawtooth” variability (Fig. 2D).

In contrast to records of Pco2 and high-latitude climate, the Borneo stalagmite δ18O record shows no change in the magnitude of interglacial values across the MBE. The average peak interglacial δ18O value for MIS13 (–9.31 ± 0.12‰, 1 SD, N = 11) is indistinguishable from those of MIS11 (–9.33 ± 0.25‰), MIS9 (–9.54 ± 0.11‰), and MIS7 (–9.43 ± 0.12‰). This finding is robust with regard to age uncertainties and will be discussed in more detail below. The duration of MIS11 in the Borneo stalagmite δ18O records seems much longer than in other climate records (Fig. 2), although dating errors are large during this interval. Within MIS11 and MIS9, δ18O varies by 0.6 to 1.2‰ (range of differences between adjacent maxima and minima), similar to the change in δ18O over the course of the Holocene (17). Although our age model does not permit assessment of the length of such suborbital features, both of the stalagmite δ18O records that span MIS11 reveal at least six distinct δ18O excursions, indicating that substantial climate variability occurred in the tropical West Pacific during past interglacials.

The large, steep increases in δ18O during terminations cannot be explained by changes in temperature or source water composition and imply substantial hydrological changes, possibly associated with profound drought. A drastic circulation change could also cause elevated δ18O of precipitation—for example, through a more local vapor source or a strong dominance of northern trajectories, which today bring the most enriched rainfall (12). One possible scenario is that the ITCZ was shifted sufficiently south that it remained south of northern Borneo during most of the year, affecting both water vapor transport and regional rainfall amount. The combination of early Southern Hemisphere warming and millennial-scale cooling in the Northern Hemisphere, accompanied by changes in ocean circulation, could have triggered such conditions (20). This possibility is consistent with the Weak Monsoon Intervals observed in China during the last four terminations (18) (Fig. 2F). It is interesting to note that the magnitude of the stalagmite δ18O maxima that occur during glacial terminations decreases progressively from Termination V to Termination III, similar to the observed trend in glacial ice volume (3) (Fig. 2D) and other marine records (21), but unlike Antarctic temperature and global Pco2. This trend cannot be explained by changes in source water δ18O (table S4). Although it is possible that we missed the most positive δ18O excursions during Terminations IV and III due to hiatuses, the similarity of the trend in stalagmite δ18O maxima to the trend in benthic δ18O, a proxy for glacial ice volume, is pronounced. If it holds true, this observation lends support to studies linking the size and instability of northern hemispheric ice sheets to the severity of deglacial Heinrich events and associated displacements of ITCZ and mid-latitude wind systems (20, 21).

Our data show that in northern Borneo hydroclimate varied substantially within interglacials, but did not change across the MBE, suggesting that the step change in high-latitude interglacial climate did not extend into the deep tropics. This notion is supported by the small change in interglacial temperature in tropical SST records (8) and the absence of the MBE in some terrestrial records from lower mid-latitudes (2224). That the post-MBE increase in interglacial Pco2 did not seem to affect low-latitude climate may be an earlier manifestation of the “polar amplification effect” (25), whereby high-latitude albedo-related feedbacks involving changes in sea ice and/or snow cover translate to a larger sensitivity of high-latitude climate to changes in greenhouse gas forcing (26). Additionally, the geographic pattern of climate responses might have been modulated by variations in Earth’s orbital configuration, as proposed by a recent modeling study (27) suggesting that insolation changes counteracted the MBE change in Pco2 in the Northern Hemisphere. However, existing records from northern mid- to high latitudes are equivocal (28, 29).

With small changes in interglacial temperature suggested by tropical SST records, the latitudinal temperature gradient decreased in post-MBE interglacials, but apparently with no discernible effect on tropical West Pacific hydrology. Instead, overhead insolation remained the dominant forcing factor during interglacials, irrespective of changing Pco2 and latitudinal temperature gradients, providing an important target for climate model simulations. As the region is an important heat and moisture source for the global climate system (11), the lack of response to the high-latitude MBE change could potentially have been transferred to the mid-latitudes. The strong drying observed during terminations, by contrast, may imply less moisture and latent heat export, constituting a potential positive-feedback mechanism during the millennial-scale high-latitude cooling events. This difference in response suggests that with regard to high-latitude climate forcing, tropical West Pacific climate may be more sensitive to transient and/or rapid changes in high-latitude climate and ocean circulation than to differences in steady state.

Supplementary Materials

Materials and Methods

Figs. S1 to S9

Tables S1 to S4

References (3139)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank B. Clark, S. Lejau, and J. Malang of Mulu National Park, as well as J. Partin for assistance in the field. A. Tuen (Universiti Malaysia Sarawak) greatly facilitated fieldwork in Sarawak. N. Sharma and K. Stewart are acknowledged for assistance with stable isotope measurements. D. Lund, A. Subhas, and D. Fernandez are thanked for assistance and advice with U-Th dating. J. Eiler provided access to his facilities at the California Institute of Technology. Support for this work was provided by the Swiss National Science Foundation (SNF) and the German Research Foundation (DFG) through postdoctoral fellowships to A.N.M.; by the US NSF through grants ATM-0318445 and ATM-0903099 to J.F.A, as well as ATM-0645291 to K.M.C.; and by an Edinburgh University Principal’s Career Development Ph.D. Scholarship to M.O.C.
View Abstract

Stay Connected to Science

Navigate This Article