• Archer, D., A. Winguth, D. Lea, and N. Mahowald, 2000: What caused the glacial/interglacial atmospheric pCO2 cycles? Rev. Geophys, 38 , 159189.

    • Search Google Scholar
    • Export Citation
  • Bender, M., 2003: Climate–biosphere interactions on glacial–interglacial timescales. Global Bigeochem. Cycles,17, 1082, doi: 10.1029/2002GB001932.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., 2000: Tropical cooling at the last glacial maximum: An atmosphere–mixed layer ocean model simulation. J. Climate, 13 , 951976.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., and S. Manabe, 1987: The influence of continental ice, atmospheric CO2, and land albedo on the climate of the last glacial maximum. Climate Dyn, 1 , 8799.

    • Search Google Scholar
    • Export Citation
  • Claquin, T., and Coauthors, 2002: Radiative forcing of climate by ice-age atmospheric dust. Climate Dyn, 20 , 193202.

  • Clement, A. C., R. Seager, and M. A. Cane, 1999: Orbital controls on the El Niño/Southern Oscillation and the tropical climate. Paleoceanography, 14 , 441456.

    • Search Google Scholar
    • Export Citation
  • CLIMAP, 1981: Seasonal reconstructions of the Earth's surface at the last glacial maximum. Geol. Soc. Amer. Map Chart Ser, MC- 36 , 118.

    • Search Google Scholar
    • Export Citation
  • Crowley, T. J., 2000: CLIMAP SSTs re-revisited. Climate Dyn, 16 , 241255.

  • Cubasch, U., G. A. Meehl, G. J. Boer, R. J. Stouffer, A. N. M. Dix, C. A. Senior, S. Raper, and K. S. Yap, 2001: Projections of future climate change. Climate Change 2001: The Scientific Basis, J. T. Houghton, et al., Eds., Cambridge University Press, 525–582.

    • Search Google Scholar
    • Export Citation
  • Cuffey, K. M., and F. Vimeux, 2001: Covariation of carbon dioxide and temperature from the Vostok ice core after deuterium-excess correction. Nature, 412 , 523527.

    • Search Google Scholar
    • Export Citation
  • Dannenmann, S., B. K. Linsley, D. W. Oppo, Y. Rosenthal, and L. Beaufort, 2003: East Asian monsoon forcing of suborbital variability in the Sulu Sea during Marine Isotope Stage 3: Link to Northern Hemisphere climate. Geochem. Geophys. Geosyst.,4, 1001, doi:10.1029/2002GC000390.

    • Search Google Scholar
    • Export Citation
  • Dekens, P. S., D. W. Lea, D. K. Pak, and H. J. Spero, 2002: Core top calibration of Mg/Ca in tropical foraminifera: Refining paleo- temperature estimation. Geochem. Geophys. Geosyst.,3, 1022, doi:10.1029/2001GC000200.

    • Search Google Scholar
    • Export Citation
  • Ganopolski, A., S. Rahmstorf, V. Petoukhov, and M. Claussen, 1998: Simulation of modern and glacial climates with a coupled global model of intermediate complexity. Nature, 391 , 351356.

    • Search Google Scholar
    • Export Citation
  • Genthon, C., J. M. Barnola, D. Raynaud, C. Lorius, J. Jouzel, N. I. Barkov, Y. S. Korotkevitch, and V. M. Kotlyakov, 1987: Vostok ice core: Climatic response to CO2 and orbital forcing changes over the last climatic cycle. Nature, 329 , 414418.

    • Search Google Scholar
    • Export Citation
  • Greene, A. M., R. Seager, and W. S. Broecker, 2002: Tropical snowline depression at the Last Glacial Maximum: Comparison with proxy records using a single-cell tropical climate model. J. Geophys. Res.,107, 4061, doi:10.1029/2001JD000670.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., A. Lacis, D. Rind, G. Russell, P. Stone, I. Fung, R. Ruedy, and J. Lerner, 1984: Climate sensitivity: Analysis of feedback mechanisms. Climate Processes and Climate Sensitivity, J. E. Hansen and T. Takahashi, Eds., Amer. Geophys. Union, 130– 163.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., M. Sato, A. Lacis, R. Ruedy, I. Tegen, and E. Matthew, 1998: Perspective: Climate forcings in the Industrial Era. Proc. Natl. Acad. Sci. USA, 95 , 1275312758.

    • Search Google Scholar
    • Export Citation
  • Hastings, D. W., A. D. Russell, and S. R. Emerson, 1998: Foraminiferal magnesium in Globeriginoides sacculifer as a paleotemperature proxy. Paleoceanography, 13 , 161169.

    • Search Google Scholar
    • Export Citation
  • Hewitt, C. D., and J. F. B. Mitchell, 1997: Radiative forcing and response of a GCM to ice age boundary conditions: Cloud feedback and climate sensitivity. Climate Dyn, 13 , 821834.

    • Search Google Scholar
    • Export Citation
  • Hewitt, C. D., R. Stouffer, A. Broccoli, J. Mitchell, and P. J. Valdes, 2003: The effect of ocean dynamics in a coupled GCM simulation of the Last Glacial Maximum. Climate Dyn, 20 , 203218.

    • Search Google Scholar
    • Export Citation
  • Hoffert, M. I., and C. Covey, 1992: Deriving global climate sensitivity from palaeoclimate reconstructions. Nature, 360 , 573576.

  • Jouzel, J., and Coauthors, 1993: Extending the Vostok ice-core record of palaeoclimate to the penultimate glacial period. Nature, 364 , 407412.

    • Search Google Scholar
    • Export Citation
  • Jouzel, J., F. Vimeux, N. Caillon, G. Delaygue, G. Hoffmann, V. Masson- Delmotte, and F. Parrenin, 2003: Magnitude of isotope/temperature scaling for interpretation of central Antarctic ice cores. J. Geophys. Res.,108, 4361, doi:10.1029/2002JD002677.

    • Search Google Scholar
    • Export Citation
  • Kienast, M., S. Steinke, K. Stattegger, and S. Calvert, 2001: Synchronous tropical South China Sea SST change and Greenland warming during deglaciation. Science, 291 , 21322134.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., T. A. Mashiotta, and H. J. Spero, 1999: Controls on magnesium and strontium uptake in planktonic foraminifera determined by live culturing. Geochim. Cosmochim. Acta, 63 , 23692379.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., D. K. Pak, and H. J. Spero, 2000: Climate impact of late quaternary equatorial Pacific sea surface temperature variations. Science, 289 , 17191724.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., P. A. Martin, D. K. Pak, and H. J. Spero, 2002: Reconstructing a 350 ky history of sea level using planktonic Mg/Ca and oxygen isotopic records from a Cocos Ridge core. Quat. Sci. Rev, 21 , 283293.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., D. K. Pak, L. C. Peterson, and K. A. Hughen, 2003: Synchroneity of tropical and high-latitude Atlantic temperatures over the last glacial termination. Science, 301 , 13611364.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS, 117 pp.

  • Lorius, C., J. Jouzel, D. Raynaud, J. E. Hansen, and H. Le Treut, 1990: The ice-core record: Climate sensitivity and future greenhouse warming. Nature, 347 , 139145.

    • Search Google Scholar
    • Export Citation
  • Manabe, S., and A. J. Broccoli, 1985: The influence of continental ice sheets on the climate of an ice age. J. Geophys. Res, 90 , 21672190.

    • Search Google Scholar
    • Export Citation
  • Martin, P. A., D. W. Lea, Y. Rosenthal, N. J. Shackleton, M. Sarnthein, and T. Papenfuss, 2002: Quaternary deep sea temperature histories derived from benthic foraminiferal Mg/Ca. Earth Planet. Sci. Lett, 198 , 193209.

    • Search Google Scholar
    • Export Citation
  • Nürnberg, D., A. Müller, and R. R. Schneider, 2000: Paleo-sea surface temperature calculations in the equatorial east Atlantic from Mg/ Ca ratios in planktic foraminifera: A comparison to sea surface temperature estimates from UK 37′ , oxygen isotopes, and foraminiferal transfer function. Paleoceanography, 15 , 124134.

    • Search Google Scholar
    • Export Citation
  • Opdyke, B. N., H. J. Spero, D. Lea, M. Gagan, H. Bostock, and K. Fifield, 2001: MD97-2167: A 470,000 year climate record from the Scott Plateau on the southwestern edge of the western Pacific warm pool. 7th Int. Conf. on Paleoceanography, Sapporo, Japan.

    • Search Google Scholar
    • Export Citation
  • Oppo, D. W., B. K. Linsley, Y. Rosenthal, S. Dannenmann, and L. Beaufort, 2003: Orbital and suborbital climate variability in the Sulu Sea, western tropical Pacific. Geochem. Geophys. Geosyst.,4, 1003, doi:10.1029/2001GC000260.

    • Search Google Scholar
    • Export Citation
  • Paillard, D., L. Labeyrie, and P. Yiou, 1996: Macintosh program performs time-series analysis. Eos, Trans. Amer. Geophys. Union, 77 , 379.

    • Search Google Scholar
    • Export Citation
  • Pelejero, C., J. O. Grimalt, S. Heilig, M. Kienast, and L. Wang, 1999: High resolution UK 37 temperature reconstructions in the South China Sea over the past 220 kyr. Paleoceanography, 14 , 224231.

    • Search Google Scholar
    • Export Citation
  • Petit, J. R., and Coauthors, 1999: Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature, 399 , 429436.

    • Search Google Scholar
    • Export Citation
  • Pisias, N. G., and A. C. Mix, 1997: Spatial and temporal oceanographic variability of the eastern equatorial Pacific during the late Pleistocene: Evidence from Radiolaria microfossils. Paleoceanography, 12 , 381393.

    • Search Google Scholar
    • Export Citation
  • Ramaswamy, V., and Coauthors, 2001: Radiative forcing of climate change. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 319–416.

    • Search Google Scholar
    • Export Citation
  • Raynaud, D., J. M. Barnola, J. Chappellaz, T. Blunier, A. Indermuhle, and B. Stauffer, 2000: The ice record of greenhouse gases: A view in the context of future changes. Quat. Sci. Rev, 19 , 917.

    • Search Google Scholar
    • Export Citation
  • Rosenthal, Y., and G. P. Lohmann, 2002: Accurate estimation of sea surface temperatures using dissolution corrected calibrations for Mg/Ca paleothermometry. Paleoceanography, 17 .1044, doi: 10.1029/2001PA000749.

    • Search Google Scholar
    • Export Citation
  • Rosenthal, Y., S. Dannenmann, D. W. Oppo, and B. K. Linsley, 2003: The amplitude and phasing of climate change during the last deglaciation in the Sulu Sea, western equatorial Pacific. Geophys. Res. Lett.,30, 1428, doi:10.1029/2002GL016612.

    • Search Google Scholar
    • Export Citation
  • Rühlemann, C., S. Mulitza, P. J. Müller, G. Wefer, and R. Zahn, 1999: Warming of the tropical Atlantic Ocean and slowdown of thermohaline circulation during the last deglaciation. Nature, 402 , 511514.

    • Search Google Scholar
    • Export Citation
  • Schneider, R. R., P. J. Müller, and R. Acheson, 1999: Atlantic alkenone sea-surface temperature records. Reconstructing Ocean History: A Window into the Future, F. Abrantes and A. Mix, Eds., Plenum, 33–55.

    • Search Google Scholar
    • Export Citation
  • Shackleton, N. J., 2000: The 100,000-year ice-age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science, 289 , 18971902.

    • Search Google Scholar
    • Export Citation
  • Shackleton, N. J., J. Imbrie, and M. A. Hall, 1983: Oxygen and carbon isotope record of East Pacific core V19-30: Implications for the formation of deep water in the last Pleistocene North Atlantic. Earth Planet. Sci. Lett, 65 , 233244.

    • Search Google Scholar
    • Export Citation
  • Siddall, M., E. J. Rohling, A. Almogi-Labin, C. Hemleben, D. Meischner, I. Schmelzer, and D. A. Smeed, 2003: Sea-level fluctuations during the last glacial cycle. Nature, 423 , 853858.

    • Search Google Scholar
    • Export Citation
  • Stott, L., C. Poulsen, S. Lund, and R. Thunell, 2002: Super ENSO and global climate oscillations at millennial time scales. Science, 279 , 222226.

    • Search Google Scholar
    • Export Citation
  • Visser, K., R. Thunell, and L. Stott, 2003: Magnitude and timing of temperature change in the Indo-Pacific warm pool during deglaciation. Nature, 421 , 152155.

    • Search Google Scholar
    • Export Citation
  • Waelbroeck, C., L. Labeyrie, E. Michel, J. C. Duplessy, J. F. McManus, K. Lambeck, E. Balbon, and M. Labracherie, 2002: Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records. Quat. Sci. Rev, 21 , 295305.

    • Search Google Scholar
    • Export Citation
  • Watanabe, O., S. Johnsen Jouzel, F. Parrenin, H. Shoji, and N. Yoshida, 2003: Homogeneous climate variability across East Antarctica over the past three glacial cycles. Nature, 422 , 509512.

    • Search Google Scholar
    • Export Citation
  • Weaver, A. J., M. Eby, A. F. Fanning, and E. C. Wiebe, 1998: Simulated influence of carbon dioxide, orbital forcing and ice sheets on the climate of the last glacial maximum. Nature, 394 , 847853.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Comparison of the Cocos Ridge SST record (2°15.5′N, 90°57.1′W, this study) with previously published SST records from the western tropical Pacific. The records are from the Makkassar Straits (4°41.3′S, 117°54.17′E; Visser et al. 2003), the Sulu Sea (8.8°N, 121.3°E) (Dannenmann et al. 2003; Oppo et al. 2003; Rosenthal et al. 2003), and the Ontong Java Plateau (0°19.1′N, 159°21.7′E; Lea et al. 2000). The comparison indicates that the amplitude of glacial–interglacial temperature variations is similar between sites

  • View in gallery

    Alignment of benthic oxygen isotopes (Uvigerina spp.) from Cocos Ridge core TR163-19 (Lea et al. 2002) to the oxygen isotope record of the same species from core V19-30 on the southeastern flank of the Carnegie Ridge (Shackleton et al. 1983). The time scale for core V19-30 is based on Shackleton et al. (2000). Alignment was done graphically by establishing tie points at well-defined transitions and clear climate events. Points used for alignment are shown by squares—a minimum number of tie points was used, resulting in a relatively constant sedimentation rate throughout the entire sequence (rms = ±25%). Shifts relative to the original chronology of TR163- 19 (Lea et al. 2000) are most significant in MIS 5 and 7. Note that the benthic isotope records from these sites cannot be taken as ice volume records because ∼40% of the amplitude is due to deep-sea temperature changes (Shackleton 2000; Martin et al. 2002)

  • View in gallery

    Comparison of tropical Pacific sea surface temperature (SST), derived from Mg/Ca determinations on surface-dwelling planktonic foraminifera in core TR163-19 on the southwest Cocos Ridge (Lea et al. 2000), with atmospheric CO2 levels determined from the Vostok ice core, east Antarctica (Petit et al. 1999). The time scale for Vostok CO2 was derived from Shackleton (2000); the time scale for the Cocos Ridge core is based on aligning the benthic oxygen isotope record (Lea et al. 2002) from core TR163-19 to nearby core V19-30, which was previously aligned to Vostok (Shackleton 2000). The modern SST at the tropical Pacific site is 26.2°C (Levitus and Boyer 1994). The TR163-19 SST record is based on Mg/Ca analyses published in Lea et al. (2000), supplemented by additional analyses in ∼15% of the intervals. SSTs are calculated from the G. ruber equation in Dekens et al. (2002)

  • View in gallery

    Amplitude spectrum of tropical SST, showing distribution between the three major earth orbital periods: e = eccentricity, t = tilt, p = precession. The fraction of the tropical amplitude that is coherent with Vostok CO2 is shown. The coherencies between the two records at eccentricity and tilt are 0.98 and 0.94, respectively. The lower plot shows the phase relationship between the two time series, where the results indicate that the two records are in phase at eccentricity and tilt. Results based on Blackman-Tukey method (Arand package) with 1/3 lags

  • View in gallery

    Comparison of temperature differences (ΔT), relative to modern sea surface temperature, at the tropical Pacific site, with greenhouse gas forcing calculated from Vostok CO2 and CH4 levels. The forcing is computed from standard radiative forcing relationships (Ramaswamy et al. 2001) relative to preindustrial levels of 280 ppm for CO2 and 700 ppb for CH4; N2O was assumed constant at 280 ppb (Raynaud et al. 2000). For reference, equivalent atmospheric CO2 values are shown on the top axis. Both datasets have been interpolated to a constant sampling interval of 2 kyr, which is the average resolution of the tropical Pacific site (Lea et al. 2000). Regression through the data yields a slope, equivalent to the climate sensitivity parameter λ, of 1.4° ± 0.1°C (W m−2)−1 and the intercept of 0.5° ± 0.2°C (r2 = 0.69; errors are 95% CI). Forcing the regression through zero yields a slope of 1.10° ± 0.07°C (W m−2)−1 (r2 = 0.64). For the 4 W m−2 forcing associated with a doubling of atmospheric CO2, these values of λ equate to a 4.4°–5.6°C warming

  • View in gallery

    Results from the multivariate analysis. (a) The relative weighting of the three climate forcings, ΣGHF, Jul 65°N insolation, lagged 2 kyr, and 2°N fall insolation (Paillard et al. 1996), that provide the best fit to the observed Cocos Ridge SST record (see Table 2 for the individual weightings). The temperature change that can be attributed to each factor is shown, along with the sum of the three influences (T mean calc). Temperature changes are shown as differences from the mean. (b) The modeled temperature change compared to the observed temperature change. The sum of three climate forcings explains 79% of the total variance in the SST record. Temperature changes (ΔT) shown in differences from present SST

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 439 162 18
PDF Downloads 229 119 12

The 100 000-Yr Cycle in Tropical SST, Greenhouse Forcing, and Climate Sensitivity

View More View Less
  • 1 Department of Geological Sciences, University of California, Santa Barbara, Santa Barbara, California
Full access

Abstract

The key scientific uncertainty in the global warming debate is the equilibrium climate sensitivity. Coupled atmosphere–ocean general circulation models predict a wide range of equilibrium climate sensitivities, with a consequently large spread of societal implications. Comparison of models with instrumental data has not been able to reduce the uncertainty in climate sensitivity. An alternative way to gauge equilibrium climate sensitivity is to use paleoclimatic data. Two recent advances, the development and application of proxy recorders of tropical sea surface temperature (SST) and the synchronization of the deep-sea and Antarctic ice-core time scales, make it possible to directly relate past changes in tropical SST to atmospheric carbon dioxide (CO2) levels. The strong correspondence of a proxy SST record from the eastern equatorial Pacific and the Vostok CO2 record suggests that varying atmospheric carbon dioxide is the dominant control on tropical climate on orbital time scales. This effect is especially pronounced at the 100 000-yr cycle. Calibration of the CO2 influence via tropical SST variability indicates a tropical climate sensitivity of 4.4°–5.6°C (errors estimated at ± 1.0°C) for a doubling of atmospheric CO2 concentration. This result suggests that the equilibrium response of tropical climate to atmospheric CO2 changes is likely to be similar to the upper end of available global predictions from coupled models.

Corresponding author address: Dr. David W. Lea, Dept. of Geological Sciences, University of California, Santa Barbara, Santa Barbara, CA 93106-9630. Email: lea@geol.ucsb.edu

Abstract

The key scientific uncertainty in the global warming debate is the equilibrium climate sensitivity. Coupled atmosphere–ocean general circulation models predict a wide range of equilibrium climate sensitivities, with a consequently large spread of societal implications. Comparison of models with instrumental data has not been able to reduce the uncertainty in climate sensitivity. An alternative way to gauge equilibrium climate sensitivity is to use paleoclimatic data. Two recent advances, the development and application of proxy recorders of tropical sea surface temperature (SST) and the synchronization of the deep-sea and Antarctic ice-core time scales, make it possible to directly relate past changes in tropical SST to atmospheric carbon dioxide (CO2) levels. The strong correspondence of a proxy SST record from the eastern equatorial Pacific and the Vostok CO2 record suggests that varying atmospheric carbon dioxide is the dominant control on tropical climate on orbital time scales. This effect is especially pronounced at the 100 000-yr cycle. Calibration of the CO2 influence via tropical SST variability indicates a tropical climate sensitivity of 4.4°–5.6°C (errors estimated at ± 1.0°C) for a doubling of atmospheric CO2 concentration. This result suggests that the equilibrium response of tropical climate to atmospheric CO2 changes is likely to be similar to the upper end of available global predictions from coupled models.

Corresponding author address: Dr. David W. Lea, Dept. of Geological Sciences, University of California, Santa Barbara, Santa Barbara, CA 93106-9630. Email: lea@geol.ucsb.edu

1. Introduction

The similarity of Antarctic proxy air temperature and atmospheric carbon dioxide (CO2) records from the Vostok ice core is striking (Petit et al. 1999; Cuffey and Vimeux 2001) and suggests that CO2 is playing a direct role in forcing the observed climate variations over the ice sheet. Such past relationships between temperature and greenhouse forcing can be used to determine climate sensitivity—the surface temperature response to a prescribed radiative forcing when all feedbacks, such as water vapor and clouds, are included. But there are inherent difficulties in using the Antarctic climate to gauge global climate sensitivity, because factors unique to the ice sheet, such as elevation, ice albedo, and sea ice, must play an important local role (Genthon et al. 1987; Lorius et al. 1990). Climate sensitivity can be computed from global temperature change and forcing estimates for the last glacial maximum (LGM; Hansen et al. 1984; Hoffert and Covey 1992; Hansen et al. 1998), but this approach cannot be applied to longer time series due to the lack of time-dependent estimates of global temperature change and forcing.

The tropical oceans, because they are removed from the direct climate impact of continental ice sheet build up during glaciation, provide an analog for equilibrium climate response to varying CO2 (Crowley 2000). In addition, modeling results for the glacial oceans support the hypothesis that CO2 variations are the dominant source of radiative forcing in tropical ocean regions, although not all of the studies explicitly separate out the effect of CO2 from other factors (Broccoli and Manabe 1987; Hewitt and Mitchell 1997; Weaver et al. 1998; Broccoli 2000) [but see Ganopolski et al. (1998) for a somewhat different view].

Previous research has demonstrated that temperature changes in the tropical oceans during glaciation were significantly smaller than is observed in oceanic areas proximal to the continental ice sheets (CLIMAP 1981; Crowley 2000). In fact, changes in tropical SST during the ice ages were small enough that they largely escaped detection until the development of two sensitive geochemical temperature proxies—alkenone unsaturation ratios and Mg/Ca in planktonic foraminifera (Pelejero et al. 1999; Lea et al. 2000). Recent research based on these two independent techniques suggests that the tropical oceans cooled by 2.8° ± 0.7°C at the LGM, 21 000 yr before present (BP) (Hastings et al. 1998; Pelejero et al. 1999; Rühlemann et al. 1999; Schneider et al. 1999; Lea et al. 2000; Nürnberg et al. 2000; Kienast et al. 2001; Stott et al. 2002; Lea et al. 2003; Rosenthal et al. 2003; Visser et al. 2003). The most important terrestrial constraint on LGM tropical cooling, the lowering of tropical mountain snowlines, is consistent with a ∼3°C cooling at the sea surface (Greene et al. 2002).

An important aspect of the available tropical paleoclimatic data is that all of the longer records, regardless of method, agree in showing the dominance of the SST signal at a period of 100 000 yr. In addition, a number of the records show that the phasing of the 100-kyr SST signal leads that of the associated ice volume signal (Pisias and Mix 1997; Schneider et al. 1999; Lea et al. 2000; Nürnberg et al. 2000; Visser et al. 2003). The strength and timing of the 100-kyr-SST signal is an important clue to the source of tropical temperature variations because recent work suggests that the 100-kyr signal in late quaternary paleoclimate records comes from the influence of atmospheric carbon dioxide variations (Shackleton 2000).

In this study I use a 360-kyr-long SST record from the Cocos Ridge, in the northeastern tropical Pacific, to evaluate the relationship between greenhouse forcing and tropical temperatures. The similarity of this record to SST records from the western tropical Pacific (Lea et al. 2000; Stott et al. 2002; Rosenthal et al. 2003; Visser et al. 2003) suggests that the Cocos Ridge site records basinwide tropical Pacific SST variations. The two main objectives of this paper are 1) to test the hypothesis that atmospheric CO2 variations are the dominant control on tropical SST on orbital time scale, and 2) to compute a (tropical) climate sensitivity based on the tropical SST response to known greenhouse forcing.

2. The Cocos Ridge SST record

a. Data description and comparison to western tropical Pacific sites

A detailed 360-kyr record of past tropical SST variations in the warm [modern SST = 26.2°C (Levitus and Boyer 1994)] eastern equatorial Pacific (EEP) waters overlying the Cocos Ridge was previously obtained using Mg/Ca determinations on the surface-dwelling foraminifera Globigerinoides ruber from core TR163-19 (2°15.5′N, 90°57.1′W; 2348 m; Lea et al. 2000). The magnitude of the SST variations at the Cocos Ridge site has been validated by demonstrating their compatibility with the companion oxygen isotope record, which contains a partial temperature influence, from the same samples (Lea et al. 2002). The previously published dataset has been augmented with 76 new analyses in 39 intervals, following the method described in Lea et al. (2000) but using a higher-precision Finnigan MAT Element2 inductively coupled plasma mass spectrometer (ICP-MS; Lea et al. 2003). Measured Mg/Ca values are converted to SST using the Pacific G. ruber equation in Dekens et al. (2002). The changes in the new merged dataset are very subtle and mostly consist of slightly reduced variability in marine isotope stage (MIS) 5 and 9.

The Cocos Ridge site lies just north of a region with large SST gradients (Levitus and Boyer 1994), and a reasonable concern would be that a gradient shift in the past due to circulation shifts might amplify the extent of tropical cooling as recorded at this site. Records from several sites in the western tropical Pacific warm pool region are now available for comparison (Fig. 1). Temperature changes over the last two glacial terminations are similar between the Cocos Ridge site and the western Pacific sites, suggesting that the Cocos Ridge SST record is representative of Pacific-wide changes.

b. A new time scale for the Cocos Ridge record

The Cocos Ridge SST record bears a strong resemblance in both form and timing to the Vostok ice deuterium/hydrogen (D/H) record, which is a proxy of air temperature over Antarctica (Petit et al. 1999; Lea et al. 2000). This broad similarity suggests a common climate forcing, with the most likely candidate being atmospheric carbon dioxide, which is well mixed throughout the atmosphere and has the potential to link climate change in regions as remote as the Tropics and poles (Genthon et al. 1987; Lorius et al. 1990). The Mapping Spectral Variability in Global Climate Project (SPECMAP) time scale used previously for the Cocos Ridge SST record (Lea et al. 2000) diverges significantly from the Vostok ice core extended glacialogical time scale prior to 260 kyr BP (Petit et al. 1999). Therefore, a detailed comparison of tropical temperatures and atmospheric CO2 requires a common time scale for the two archives.

Shackleton (2000) developed a common time scale for the deep-sea sedimentary record and the Vostok ice core by independently tuning an eastern equatorial Pacific benthic foraminiferal oxygen isotope record (core V19-30; Shackleton et al. 1983) and the Vostok air bubble oxygen isotope record to orbital targets. Because benthic oxygen isotope records within a given region are expected to be quite similar, it is possible to graphically align the benthic oxygen isotope records of TR163-19 (Lea et al. 2002) and core V19-30 (Fig. 2). This alignment produces an independently aligned time scale that allows, for the first time, a direct comparison of tropical SST and atmospheric CO2 variability (Fig. 3). The alignment of the benthic oxygen isotope records was achieved with a minimum number of tie points, and the resulting sedimentation rates vary by only ±25% (rms). The Cocos Ridge SST data on the new time scale are available from the author and from the NOAA World Data Center for Paleoclimatology (see information online at http://web.ngdc.noaa.goc/paleo/data.html).

c. Comparison to Vostok atmospheric CO2 record

The Cocos Ridge SST and atmospheric CO2 records show a striking correlation throughout the 360-kyr sequence, with an overall cross correlation of r = 0.84 (Fig. 3). The correspondence between the two records is clear at both the longer 100-kyr periodicity, which dominates both records, and at the shorter orbital period of 41 kyr. The only time interval where the records clearly diverge is during MIS 5e and 5d, ∼130–110 kyr BP, when the CO2 record indicates a long plateau of high levels while the SST record indicates a clear dip to colder levels during MIS 5d. The long plateau in 5e in Vostok has previously been singled out as anomalous (Jouzel et al. 1993; Cuffey and Vimeux 2001) and appears to be shorter in a new record from the Dome Fuji Station (Watanabe et al. 2003).

Cross-spectral analysis confirms the coherence with no discernable phase lag of the 100- and 41-kyr periodicity in the two records (Fig. 4; Table 1). The amplitude spectrum of the SST and CO2 records indicates that the shared coherence at 100 kyr is 98%. At this period, the peak coherent amplitudes of the SST and CO2 records are 1.0°C and 28 ppmv, respectively. Comparison of the extracted deep-sea temperature component of the V19-30 benthic oxygen isotope record and the CO2 record using a similar analysis yielded a coherent 100-kyr amplitude of 1°C and 34 ppmv, respectively (Shackleton 2000). The coherent 100-kyr amplitude in the Cocos Ridge SST record is likely to be a consequence of the influence of atmospheric CO2 on tropical climate.

It is important to note that the amplitude of SST change at 100 kyr is not unusually high at the Cocos Ridge site. For example, a longer record from the Ocean Drilling Program (ODP) hole 806B in the western Pacific, in the center of the warm pool, has a slightly higher 100-kyr amplitude of 1.1°C (using the data and age model in Lea et al. 2000; Fig. 1). This site does have clearly reduced amplitude at the other major orbital periods, so the overall impression of the record is of a somewhat reduced amplitude relative to the Cocos Ridge site (Lea et al. 2000).

The amplitude spectrum of the Cocos Ridge SST record indicates that a significant fraction of the variability at shorter orbital periods (41 and 23 kyr) is not coherent with CO2 (Fig. 4). The two records are not statistically coherent at the frequency of precession, and the tropical SST record has significantly higher variance at 23 kyr. This is also apparent in the graphical data comparison (Fig. 3); the three large oscillations of SST during MIS 5 (∼130–80 kyr BP) are either reduced or absent in the CO2 record. The implication is that some of the 23-kyr variability in the SST record comes about from other climate influences; one possible candidate is the influence of precession on the seasonal distribution of radiation in the Tropics, which could change the tendency of the tropical Pacific toward warm El Niño versus cold La Niña phases (Clement et al. 1999). Multivariate analysis, however, suggests that the additional influence is likely to come from the high latitudes, presumably through ice volume (see below).

It has previously been suggested that the tropical temperature changes themselves could have caused the atmospheric CO2 change (Visser et al. 2003). Carbonate equilibrium calculations, box models, and ocean GCMs demonstrate that the expected atmospheric CO2 response to the observed low-latitude glacial cooling is too small to explain the observed ice core CO2 change (Archer et al. 2000). Therefore, warming in the Tropics in response to rising atmospheric CO2 probably provided only a small feedback through the decreased solubility of CO2 in warmer seawater.

3. Deriving equilibrium tropical climate sensitivity estimates from the paleoclimatic data

a. Estimating climate sensitivity by linear regression

Coupled ocean–atmosphere GCMs indicate a 500-yr or longer time interval required for climate to reach full equilibrium with a new forcing level (Cubasch et al. 2001). Because each 2-cm-long sediment interval in core TR163-19 integrates ∼660 yr, the reconstructed SSTs approximate an equilibrium tropical climate value for that particular atmospheric CO2 level. Therefore, it is possible to estimate equilibrium climate sensitivity by directly comparing tropical SSTs to known estimates of greenhouse forcing.

Direct point-to-point comparison of SSTs and CO2 levels is achieved by interpolating both records to a constant 2-kyr sample interval, which is slightly greater than the average resolution of both records. Because the radiative effect of CO2 is first order, changes in tropical temperature are compared to the calculated climate forcing using the expression (Ramaswamy et al. 2001)
i1520-0442-17-11-2170-e1
where C is the atmospheric CO2 value at a particular time and C0 the 280 ppmv preindustrial CO2 level. Because methane also varies on glacial–interglacial time scales (Petit et al. 1999), its effect, which is on average ∼13% of the total forcing, has been included in the climate forcing calculation using the expression (Ramaswamy et al. 2001)
i1520-0442-17-11-2170-e2
where M is the atmospheric CH4 value at a particular time, M0 the 700 ppbv preindustrial CH4 level, and N0 the 275 ppbv preindustrial N2O level (past changes in N2O are ignored because of their relatively small radiative influence and the lack of a continuous record for the time period of interest).

The correlation between temperature change and total radiative forcing has a slope of 1.4° ± 0.1°C (W m−2)−1 and an intercept of 0.5° ± 0.2°C [errors are 95% confidence interval (CI); Fig. 5]. The positive temperature intercept is a consequence of SSTs warmer than modern conditions during discrete time spans of previous interglacials with similar CO2 levels: MIS 5, 119–131 kyr BP; MIS 7, 233–245 kyr BP; and MIS 9, 325–335 kyr BP. This observation cannot readily be explained by greenhouse forcing alone, and either reflects another climate influence or inaccuracies in the data and/or time scale. One hypothesis to explain these warmer interglacial temperatures is that they reflect reduced ice volume during these time intervals; available data, however, do not suggest a significantly higher-than-modern sea level during peak MIS 5, 7, and 9 (Shackleton 2000; Lea et al. 2002; Waelbroeck et al. 2002; Siddall et al. 2003). An alternative hypothesis (N. Shackleton 2003, personal communication) is that the additional SST amplitude at 100 kyr that cannot be attributed to greenhouse forcing reflects the influence of the eccentricity envelope on precession. In this case, spectral analysis would extract both the 100-kyr SST amplitude due to greenhouse forcing as well as to the eccentricity envelope. The fact that the ratio of coherent amplitude of SST and climate forcing at 100 kyr only, which are 1.0°C and 0.7 W m−2, respectively, is 1.4°C (W m−2)−1, identical to the value of λ derived by fitting the discrete points (Fig. 5), is consistent with this hypothesis.

A convenient solution to the problem of a positive intercept is to force the regression through zero (i.e., assume ΔT is zero for a greenhouse forcing of zero), which yields a slope, equivalent to the climate sensitivity parameter λ (Ramaswamy et al. 2001), of 1.10° ± 0.07°C (W m−2)−1 (errors are 95% CI of the slope; Fig. 5). Forcing the regression through zero yields the minimum value of λ, but with the consequence of reducing the correlation between ΣGHF and SST, from r2 of 0.69 to 0.63. The minimum slope can also be calculated by eliminating the warm points that create the positive intercept, in which case λ = 1.1° ± 0.1°C (W m−2)−1, statistically identical to the prior approach.

b. Estimating climate sensitivity by multivariate regression

A potential weakness of the approach just described is that it does not take into account other factors that could influence tropical SST, such as ice volume or atmospheric dust (Claquin et al. 2002). An alternative approach (Genthon et al. 1987) is to use multivariate regression to derive relative weightings for each of the hypothesized independent factors that control SST variations at the Cocos Ridge site. The following inputs were used for multivariate analysis: total greenhouse forcing (ΣGHF), local insolation (annual, spring, and fall), Northern Hemisphere (NH) insolation (July, 65°N), two different records of ice volume (Shackleton 2000; Lea et al. 2002), and records of Vostok ice core dust (both linear and ln; Petit et al. 1999). These records are not completely independent of each other (Bender 2003), but multivariate analysis provides a starting point to evaluate the relative weighting for each factor in the observed SST record.

In all of the multivariate analyses, ΣGHF accounts for the dominant signal, and its weighting always exceeds all of the other tested factors (Table 2). The best fit is achieved (79% variance explained) when July 65°N insolation, lagged by 2 kyr, is used as an ice volume proxy. For this fit, ΣGHF explains 69%, July 65°N insolation (lagged 2 kyr) explains 9%, and local fall insolation explains 1% of the variance. If ΣGHF is excluded from the analysis, up to 59% of the variance in SST can be explained by a combination of Vostok ln(dust), July insolation at 65°N, ice volume, and fall insolation at 2°N (Table 2). But because the Vostok ln(dust) and CO2 records have a strong linear relationship (Bender 2003), it is difficult to assign any significance to this finding. Judged on an individual basis, ΣGHF explains 69% of the SST variance, Vostok ln(dust) explains 44%, the Shackleton (2000) ice volume record explains 30%, and July 65°N insolation (lagged 2 kyr) explains 21% (Table 2). The significant amount of variance explained by ΣGHF, dust, and ice volume in the individual fits indicates that it may be difficult to statistically separate the effects of these three forcings. For this reason, the dominance of ΣGHF in the multivariate effect does not rule out a potentially important role for the other forcings.

Examination of the relative weightings of each factor for the best multivariate fit demonstrates the dominance of ΣGHF in explaining the SST observations at the Cocos Ridge site (Fig. 6a). The effect of July 65°N insolation is significant, however, especially in explaining the observed SST minima in MIS 5.4 (5d) and MIS 7.4. Reconstructed SST, using the multivariate weighting, provides a good fit to the observed data (Fig. 6b), with the chief discrepancies occurring in MIS 5e (observed SST is too warm) and MIS 8.

The key finding from the multivariate analysis is that the value of climate sensitivity parameter λ for the best fit is 1.3° ± 0.1°C (W m−2)−1, which falls within the range of the simple linear fit and the fit forced through zero (see above). Substitutions [such as using the Shackleton (2000) ice volume record instead of the July 65°N insolation] do not significantly affect this finding.

4. Implications for global climate sensitivity

For an assumed radiative effect of 4 W m−2 for CO2 doubling (Ramaswamy et al. 2001), the minimum value of λ [1.10° ± 0.07°C (W m−2)−1] implies a 4.4° ± 0.9°C change in tropical SST [error calculated from the standard error (SE) of the regression and the estimated error of the SSTs (Lea et al. 2000)]. The maximum value of λ [1.4° ± 0.1°C (W m−2)−1] implies a 5.6° ± 0.9°C increase for doubled CO2. A regression through the data since the LGM only (1–25 kyr BP), which is the best constrained time interval from a chronological viewpoint, yields a sensitivity of 4.8° ± 0.8°C. The value of λ derived from the best-fit multivariate analysis, 1.28° ± 0.12°C (W m−2)−1, yields a temperature response to doubled CO2 of 5.1° ± 0.8°C. This value is probably the most representative estimate that can be derived from the data in this study.

It is instructive to compare this value to values derived by direct comparison of the Vostok D/H air temperature proxy and atmospheric CO2 records (Petit et al. 1999). A multivariate study of the Vostok data (0– 160 kyr BP) found that for the majority of fits CO2 variations accounted for 70%–85% of the air temperature change at Vostok, with amplification factors f [i.e., ratio of observed temperature change to temperature change required to restore radiative equilibrium (Hansen et al. 1984)] of 11–14 (Genthon et al. 1987). This range of amplifications is equivalent to a local climate sensitivity of ∼3°–4°C (W m−2)−1. A subsequent study of the same dataset (Lorius et al. 1990) reported a much smaller amplification factor, f ∼ 3, which is equivalent to a local climate sensitivity of ∼1°C (W m−2)−1. An alternative way to calculate climate sensitivity from the Vostok data is to ratio the 100-kyr amplitude of the D/H record that is coherent with ΣGHF, which is 13‰, equivalent to an air temperature change of ∼2°C (Jouzel et al. 2003), to the coherent 100-kyr amplitude in ΣGHF as determined in this study, which is 0.7 W m−2. This approach yields a local climate sensitivity of ∼3°C (W m−2)−1, in agreement with the Genthon et al. (1987) result. The polar amplification of greenhouse gas forcing relative to the Tropics using the approach is ∼2, as evaluated by taking the ratio of the coherent 100-kyr amplitude of the Antarctic air temperature record (2°C) to the coherent 100-kyr amplitude of the tropical SST record (1°C).

For comparison to the paleoclimatic approach, 15 climate models indicate a global mean equilibrium climate sensitivity of 3.5°C, with a range of 2.0°–5.1°C (Cubasch et al. 2001; Fig. 5). The tropical climate sensitivity from this study, 4.4°–5.6°C, is at the high end of the model-derived global climate sensitivity results. Because virtually all equilibrium climate simulations indicate that tropical temperature changes are smaller than the global mean, the global sensitivity implied by this study might lie above the IPCC range. The tropical climate sensitivity derived in this study is also larger than suggested by model simulations of the LGM climate (Hewitt and Mitchell 1997). This difference appears to be a consequence of the smaller effect that glacial CO2 reduction has on tropical ocean temperatures in these models, which is typically 1°–2°C (Hewitt and Mitchell 1997; Ganopolski et al. 1998; Weaver et al. 1998; Broccoli 2000; Hewitt et al. 2003).

One factor that could explain the higher climate sensitivity values derived in this study relative to climate models is a strong cooling effect on the Tropics, via ocean and atmospheric heat transport, of mid- and high- latitude continental ice sheets. Modeling studies demonstrate that the influence of continental ice (i.e., ice volume) is vastly diminished in the Southern Hemisphere (Manabe and Broccoli 1985; Broccoli and Manabe 1987). Therefore, an independent way to evaluate ice volume influence is to compare glacial cooling from Southern Hemisphere warm pool sites. Two sites are presently available: 1) a site in the Makassar Straits (4°41.3′S, 117°54.17′E), which records a cooling of 3.3° ± 0.6°C during the LGM and 4.1° ± 0.6°C during the previous glacial maximum (MIS 6; Visser et al. 2003); and 2) a site in the Timor Trough (13°8.86′S, 121°34.69′E), which records a cooling of 2.8° ± 0.5°C during the LGM and 3.9° ± 0.8°C during MIS 6 (Opdyke et al. 2001). For comparison, the coolings observed at the Cocos Ridge site used in this study are 2.6° ± 0.8°C for the LGM and 3.9° ± 0.3°C for MIS 6. The similar level of cooling observed at all three sites supports the hypothesis that the dominant forcing is the glacial reduction in atmospheric greenhouse gases.

5. Limitations and future refinement of the approach

To what extent should the paleoclimatic results for equilibrium climate sensitivity be taken as representative of possible future changes? The advantage of the paleoclimatic approach is that it provides a series of equilibrium climate sensitivity experiments over a range of atmospheric CO2 levels. It also involves no model assumptions other than the hypothesis that changes in tropical ocean temperatures are dominantly forced by varying CO2 levels. A disadvantage of the paleoclimatic approach is that other climatic influences, such as the size and distribution of the glacial ice sheets, surface albedo, insolation changes, and concomitant changes in ocean and atmospheric circulation all could have played a role in determining the relationship between tropical climate and CO2 in the past in such a way that the past pattern does not apply to the future. But, growing paleoclimatic evidence that the timing of tropical temperature changes leads ice volume changes by several thousand years or more (Pisias and Mix 1997; Schneider et al. 1999; Lea et al. 2000; Stott et al. 2002; Visser et al. 2003), suggests that the effect of continental ice on tropical SSTs is likely to be a very slow feedback, at least compared to the rapid equilibration time of the tropical mixed layer. The other influences are harder to separate, but GCM results do not uniformly identify any of them as likely to be an important component of the glacial climate forcing in the Tropics (Hewitt and Mitchell 1997; Ganopolski et al. 1998; Weaver et al. 1998; Broccoli 2000). The more recent suggestion that atmospheric dust loading could exert a considerable influence (up to 3 W m−2 locally) on tropical climate adds another factor with potentially the same magnitude as greenhouse forcing (Claquin et al. 2002). Dust influence, however, has a very strong spatial pattern mediated by source regions (Claquin et al. 2002), and the presently available paleo-SST data indicate a relatively uniform glacial cooling of ∼3°C for the tropical warm pools, which is probably not consistent with the pattern expected from a strong dust influence.

There are a number of possibilities for future work that can be pursued to evaluate the results of this study. Ideally, any calibration of tropical SST versus atmospheric CO2 would not rely on a single record, because any particular site is likely to reflect some combination of the influence of changing greenhouse forcing as well as local climate factors. The availability of a stacked record of tropical SST combining records from the warm pools of all three ocean basins should average out local effects and improve tropical climate sensitivity determination from paleoclimate data. An additional uncertainty stems from the absolute magnitude of SST change reconstructed from the foraminiferal Mg/Ca proxy; the largest uncertainty in paleotemperature analysis appears to be related to potential changes in foraminiferal preservation in the past due to postdepositional changes (Lea et al. 2000; Dekens et al. 2002; Rosenthal and Lohmann 2002), but secondary environmental influences, such as changing pH, might introduce additional uncertainty (Lea et al. 1999). By comparing and combining records from different basins with different preservational histories, it should be possible to reduce the uncertainty associated with postdepositional changes. From the viewpoint of climate process, the biggest uncertainty lies in the assumption that atmospheric CO2 alone drives tropical SST change at the 100-kyr period. Modeling runs, which could separate the response of tropical temperatures to potential glacial climate forcing other than CO2, would be useful in this regard. In addition, it would be useful if model runs could be used to test the assumption that climate sensitivity determined from a colder past can be applied symmetrically to a warmer future.

6. Conclusions

Tropical SST variations spanning the last half million years are dominated by the 100 000-yr cycle. A proposed hypothesis to explain this observation is that varying atmospheric carbon dioxide abundances, which are a likely source of 100-kyr variability in paleoclimate records (Shackleton 2000), are directly forcing tropical climate on long orbital time scales. This hypothesis is tested by comparing a 360-kyr tropical SST record from the Cocos Ridge, north of the Galapagos Islands, with the Vostok atmospheric CO2 record. The comparison indicates that the two records have a strong overall correspondence (r = 0.84) and are highly coherent with no phase offset at 100 and 41 kyr, supporting a greenhouse gas mechanism as the dominant control of tropical climate change. The phasing provides an explanation for the observed lead of tropical SST relative to ice volume (Lea et al. 2000), because atmospheric CO2 changes early in the climate cycle (Petit et al. 1999; Shackleton 2000). Integrating the available observations suggests that atmospheric CO2, Antarctic temperature, and tropical temperature all changed early in the climate cycle relative to ice volume.

Tropical climate sensitivity is computed from the Cocos Ridge comparison by directly comparing the amplitude of SST variations to that of radiative forcing computed from varying atmospheric CO2 and CH4. The computation indicates a climate sensitivity λ of 1.1°–1.4°C (W m−2)−1 [error estimated at ± 0.1°C (W m−2)−1]. Using this range, a 4 W m−2 radiative forcing associated with a doubling of atmospheric CO2 would equate to a 4.4°–5.6°C increase in temperature (total errors estimated at ±1°C), generally above the currently accepted Intergovernmental Panel on Climate Change (IPCC) range of 1.5°–4.5°C from GCMs (Cubasch et al. 2001). The degree to which climate sensitivities determined from paleoclimatic data apply to future warming depends on how differently climate feedbacks operated in the past compared to the future.

Acknowledgments

This work was supported by U.S. NSF OCE0117886. The John Simon Guggenheim and Leverhulme Foundations provided support for the author's sabbatical year in the Godwin Laboratory at the University of Cambridge, where this research was undertaken. I thank D. Pak and G. Paradis for laboratory work at UCSB. Discussions with N. Shackleton and S. Crowhurst were critical to the development of this work. Reviews and comments on earlier versions by A. Broccoli, A. Weaver, D. Hartmann, M. Cane, T. Crowley, and an anonymous reviewer improved the manuscript.

REFERENCES

  • Archer, D., A. Winguth, D. Lea, and N. Mahowald, 2000: What caused the glacial/interglacial atmospheric pCO2 cycles? Rev. Geophys, 38 , 159189.

    • Search Google Scholar
    • Export Citation
  • Bender, M., 2003: Climate–biosphere interactions on glacial–interglacial timescales. Global Bigeochem. Cycles,17, 1082, doi: 10.1029/2002GB001932.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., 2000: Tropical cooling at the last glacial maximum: An atmosphere–mixed layer ocean model simulation. J. Climate, 13 , 951976.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., and S. Manabe, 1987: The influence of continental ice, atmospheric CO2, and land albedo on the climate of the last glacial maximum. Climate Dyn, 1 , 8799.

    • Search Google Scholar
    • Export Citation
  • Claquin, T., and Coauthors, 2002: Radiative forcing of climate by ice-age atmospheric dust. Climate Dyn, 20 , 193202.

  • Clement, A. C., R. Seager, and M. A. Cane, 1999: Orbital controls on the El Niño/Southern Oscillation and the tropical climate. Paleoceanography, 14 , 441456.

    • Search Google Scholar
    • Export Citation
  • CLIMAP, 1981: Seasonal reconstructions of the Earth's surface at the last glacial maximum. Geol. Soc. Amer. Map Chart Ser, MC- 36 , 118.

    • Search Google Scholar
    • Export Citation
  • Crowley, T. J., 2000: CLIMAP SSTs re-revisited. Climate Dyn, 16 , 241255.

  • Cubasch, U., G. A. Meehl, G. J. Boer, R. J. Stouffer, A. N. M. Dix, C. A. Senior, S. Raper, and K. S. Yap, 2001: Projections of future climate change. Climate Change 2001: The Scientific Basis, J. T. Houghton, et al., Eds., Cambridge University Press, 525–582.

    • Search Google Scholar
    • Export Citation
  • Cuffey, K. M., and F. Vimeux, 2001: Covariation of carbon dioxide and temperature from the Vostok ice core after deuterium-excess correction. Nature, 412 , 523527.

    • Search Google Scholar
    • Export Citation
  • Dannenmann, S., B. K. Linsley, D. W. Oppo, Y. Rosenthal, and L. Beaufort, 2003: East Asian monsoon forcing of suborbital variability in the Sulu Sea during Marine Isotope Stage 3: Link to Northern Hemisphere climate. Geochem. Geophys. Geosyst.,4, 1001, doi:10.1029/2002GC000390.

    • Search Google Scholar
    • Export Citation
  • Dekens, P. S., D. W. Lea, D. K. Pak, and H. J. Spero, 2002: Core top calibration of Mg/Ca in tropical foraminifera: Refining paleo- temperature estimation. Geochem. Geophys. Geosyst.,3, 1022, doi:10.1029/2001GC000200.

    • Search Google Scholar
    • Export Citation
  • Ganopolski, A., S. Rahmstorf, V. Petoukhov, and M. Claussen, 1998: Simulation of modern and glacial climates with a coupled global model of intermediate complexity. Nature, 391 , 351356.

    • Search Google Scholar
    • Export Citation
  • Genthon, C., J. M. Barnola, D. Raynaud, C. Lorius, J. Jouzel, N. I. Barkov, Y. S. Korotkevitch, and V. M. Kotlyakov, 1987: Vostok ice core: Climatic response to CO2 and orbital forcing changes over the last climatic cycle. Nature, 329 , 414418.

    • Search Google Scholar
    • Export Citation
  • Greene, A. M., R. Seager, and W. S. Broecker, 2002: Tropical snowline depression at the Last Glacial Maximum: Comparison with proxy records using a single-cell tropical climate model. J. Geophys. Res.,107, 4061, doi:10.1029/2001JD000670.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., A. Lacis, D. Rind, G. Russell, P. Stone, I. Fung, R. Ruedy, and J. Lerner, 1984: Climate sensitivity: Analysis of feedback mechanisms. Climate Processes and Climate Sensitivity, J. E. Hansen and T. Takahashi, Eds., Amer. Geophys. Union, 130– 163.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., M. Sato, A. Lacis, R. Ruedy, I. Tegen, and E. Matthew, 1998: Perspective: Climate forcings in the Industrial Era. Proc. Natl. Acad. Sci. USA, 95 , 1275312758.

    • Search Google Scholar
    • Export Citation
  • Hastings, D. W., A. D. Russell, and S. R. Emerson, 1998: Foraminiferal magnesium in Globeriginoides sacculifer as a paleotemperature proxy. Paleoceanography, 13 , 161169.

    • Search Google Scholar
    • Export Citation
  • Hewitt, C. D., and J. F. B. Mitchell, 1997: Radiative forcing and response of a GCM to ice age boundary conditions: Cloud feedback and climate sensitivity. Climate Dyn, 13 , 821834.

    • Search Google Scholar
    • Export Citation
  • Hewitt, C. D., R. Stouffer, A. Broccoli, J. Mitchell, and P. J. Valdes, 2003: The effect of ocean dynamics in a coupled GCM simulation of the Last Glacial Maximum. Climate Dyn, 20 , 203218.

    • Search Google Scholar
    • Export Citation
  • Hoffert, M. I., and C. Covey, 1992: Deriving global climate sensitivity from palaeoclimate reconstructions. Nature, 360 , 573576.

  • Jouzel, J., and Coauthors, 1993: Extending the Vostok ice-core record of palaeoclimate to the penultimate glacial period. Nature, 364 , 407412.

    • Search Google Scholar
    • Export Citation
  • Jouzel, J., F. Vimeux, N. Caillon, G. Delaygue, G. Hoffmann, V. Masson- Delmotte, and F. Parrenin, 2003: Magnitude of isotope/temperature scaling for interpretation of central Antarctic ice cores. J. Geophys. Res.,108, 4361, doi:10.1029/2002JD002677.

    • Search Google Scholar
    • Export Citation
  • Kienast, M., S. Steinke, K. Stattegger, and S. Calvert, 2001: Synchronous tropical South China Sea SST change and Greenland warming during deglaciation. Science, 291 , 21322134.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., T. A. Mashiotta, and H. J. Spero, 1999: Controls on magnesium and strontium uptake in planktonic foraminifera determined by live culturing. Geochim. Cosmochim. Acta, 63 , 23692379.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., D. K. Pak, and H. J. Spero, 2000: Climate impact of late quaternary equatorial Pacific sea surface temperature variations. Science, 289 , 17191724.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., P. A. Martin, D. K. Pak, and H. J. Spero, 2002: Reconstructing a 350 ky history of sea level using planktonic Mg/Ca and oxygen isotopic records from a Cocos Ridge core. Quat. Sci. Rev, 21 , 283293.

    • Search Google Scholar
    • Export Citation
  • Lea, D. W., D. K. Pak, L. C. Peterson, and K. A. Hughen, 2003: Synchroneity of tropical and high-latitude Atlantic temperatures over the last glacial termination. Science, 301 , 13611364.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS, 117 pp.

  • Lorius, C., J. Jouzel, D. Raynaud, J. E. Hansen, and H. Le Treut, 1990: The ice-core record: Climate sensitivity and future greenhouse warming. Nature, 347 , 139145.

    • Search Google Scholar
    • Export Citation
  • Manabe, S., and A. J. Broccoli, 1985: The influence of continental ice sheets on the climate of an ice age. J. Geophys. Res, 90 , 21672190.

    • Search Google Scholar
    • Export Citation
  • Martin, P. A., D. W. Lea, Y. Rosenthal, N. J. Shackleton, M. Sarnthein, and T. Papenfuss, 2002: Quaternary deep sea temperature histories derived from benthic foraminiferal Mg/Ca. Earth Planet. Sci. Lett, 198 , 193209.

    • Search Google Scholar
    • Export Citation
  • Nürnberg, D., A. Müller, and R. R. Schneider, 2000: Paleo-sea surface temperature calculations in the equatorial east Atlantic from Mg/ Ca ratios in planktic foraminifera: A comparison to sea surface temperature estimates from UK 37′ , oxygen isotopes, and foraminiferal transfer function. Paleoceanography, 15 , 124134.

    • Search Google Scholar
    • Export Citation
  • Opdyke, B. N., H. J. Spero, D. Lea, M. Gagan, H. Bostock, and K. Fifield, 2001: MD97-2167: A 470,000 year climate record from the Scott Plateau on the southwestern edge of the western Pacific warm pool. 7th Int. Conf. on Paleoceanography, Sapporo, Japan.

    • Search Google Scholar
    • Export Citation
  • Oppo, D. W., B. K. Linsley, Y. Rosenthal, S. Dannenmann, and L. Beaufort, 2003: Orbital and suborbital climate variability in the Sulu Sea, western tropical Pacific. Geochem. Geophys. Geosyst.,4, 1003, doi:10.1029/2001GC000260.

    • Search Google Scholar
    • Export Citation
  • Paillard, D., L. Labeyrie, and P. Yiou, 1996: Macintosh program performs time-series analysis. Eos, Trans. Amer. Geophys. Union, 77 , 379.

    • Search Google Scholar
    • Export Citation
  • Pelejero, C., J. O. Grimalt, S. Heilig, M. Kienast, and L. Wang, 1999: High resolution UK 37 temperature reconstructions in the South China Sea over the past 220 kyr. Paleoceanography, 14 , 224231.

    • Search Google Scholar
    • Export Citation
  • Petit, J. R., and Coauthors, 1999: Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature, 399 , 429436.

    • Search Google Scholar
    • Export Citation
  • Pisias, N. G., and A. C. Mix, 1997: Spatial and temporal oceanographic variability of the eastern equatorial Pacific during the late Pleistocene: Evidence from Radiolaria microfossils. Paleoceanography, 12 , 381393.

    • Search Google Scholar
    • Export Citation
  • Ramaswamy, V., and Coauthors, 2001: Radiative forcing of climate change. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 319–416.

    • Search Google Scholar
    • Export Citation
  • Raynaud, D., J. M. Barnola, J. Chappellaz, T. Blunier, A. Indermuhle, and B. Stauffer, 2000: The ice record of greenhouse gases: A view in the context of future changes. Quat. Sci. Rev, 19 , 917.

    • Search Google Scholar
    • Export Citation
  • Rosenthal, Y., and G. P. Lohmann, 2002: Accurate estimation of sea surface temperatures using dissolution corrected calibrations for Mg/Ca paleothermometry. Paleoceanography, 17 .1044, doi: 10.1029/2001PA000749.

    • Search Google Scholar
    • Export Citation
  • Rosenthal, Y., S. Dannenmann, D. W. Oppo, and B. K. Linsley, 2003: The amplitude and phasing of climate change during the last deglaciation in the Sulu Sea, western equatorial Pacific. Geophys. Res. Lett.,30, 1428, doi:10.1029/2002GL016612.

    • Search Google Scholar
    • Export Citation
  • Rühlemann, C., S. Mulitza, P. J. Müller, G. Wefer, and R. Zahn, 1999: Warming of the tropical Atlantic Ocean and slowdown of thermohaline circulation during the last deglaciation. Nature, 402 , 511514.

    • Search Google Scholar
    • Export Citation
  • Schneider, R. R., P. J. Müller, and R. Acheson, 1999: Atlantic alkenone sea-surface temperature records. Reconstructing Ocean History: A Window into the Future, F. Abrantes and A. Mix, Eds., Plenum, 33–55.

    • Search Google Scholar
    • Export Citation
  • Shackleton, N. J., 2000: The 100,000-year ice-age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science, 289 , 18971902.

    • Search Google Scholar
    • Export Citation
  • Shackleton, N. J., J. Imbrie, and M. A. Hall, 1983: Oxygen and carbon isotope record of East Pacific core V19-30: Implications for the formation of deep water in the last Pleistocene North Atlantic. Earth Planet. Sci. Lett, 65 , 233244.

    • Search Google Scholar
    • Export Citation
  • Siddall, M., E. J. Rohling, A. Almogi-Labin, C. Hemleben, D. Meischner, I. Schmelzer, and D. A. Smeed, 2003: Sea-level fluctuations during the last glacial cycle. Nature, 423 , 853858.

    • Search Google Scholar
    • Export Citation
  • Stott, L., C. Poulsen, S. Lund, and R. Thunell, 2002: Super ENSO and global climate oscillations at millennial time scales. Science, 279 , 222226.

    • Search Google Scholar
    • Export Citation
  • Visser, K., R. Thunell, and L. Stott, 2003: Magnitude and timing of temperature change in the Indo-Pacific warm pool during deglaciation. Nature, 421 , 152155.

    • Search Google Scholar
    • Export Citation
  • Waelbroeck, C., L. Labeyrie, E. Michel, J. C. Duplessy, J. F. McManus, K. Lambeck, E. Balbon, and M. Labracherie, 2002: Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records. Quat. Sci. Rev, 21 , 295305.

    • Search Google Scholar
    • Export Citation
  • Watanabe, O., S. Johnsen Jouzel, F. Parrenin, H. Shoji, and N. Yoshida, 2003: Homogeneous climate variability across East Antarctica over the past three glacial cycles. Nature, 422 , 509512.

    • Search Google Scholar
    • Export Citation
  • Weaver, A. J., M. Eby, A. F. Fanning, and E. C. Wiebe, 1998: Simulated influence of carbon dioxide, orbital forcing and ice sheets on the climate of the last glacial maximum. Nature, 394 , 847853.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Comparison of the Cocos Ridge SST record (2°15.5′N, 90°57.1′W, this study) with previously published SST records from the western tropical Pacific. The records are from the Makkassar Straits (4°41.3′S, 117°54.17′E; Visser et al. 2003), the Sulu Sea (8.8°N, 121.3°E) (Dannenmann et al. 2003; Oppo et al. 2003; Rosenthal et al. 2003), and the Ontong Java Plateau (0°19.1′N, 159°21.7′E; Lea et al. 2000). The comparison indicates that the amplitude of glacial–interglacial temperature variations is similar between sites

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Fig. 2.
Fig. 2.

Alignment of benthic oxygen isotopes (Uvigerina spp.) from Cocos Ridge core TR163-19 (Lea et al. 2002) to the oxygen isotope record of the same species from core V19-30 on the southeastern flank of the Carnegie Ridge (Shackleton et al. 1983). The time scale for core V19-30 is based on Shackleton et al. (2000). Alignment was done graphically by establishing tie points at well-defined transitions and clear climate events. Points used for alignment are shown by squares—a minimum number of tie points was used, resulting in a relatively constant sedimentation rate throughout the entire sequence (rms = ±25%). Shifts relative to the original chronology of TR163- 19 (Lea et al. 2000) are most significant in MIS 5 and 7. Note that the benthic isotope records from these sites cannot be taken as ice volume records because ∼40% of the amplitude is due to deep-sea temperature changes (Shackleton 2000; Martin et al. 2002)

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Fig. 3.
Fig. 3.

Comparison of tropical Pacific sea surface temperature (SST), derived from Mg/Ca determinations on surface-dwelling planktonic foraminifera in core TR163-19 on the southwest Cocos Ridge (Lea et al. 2000), with atmospheric CO2 levels determined from the Vostok ice core, east Antarctica (Petit et al. 1999). The time scale for Vostok CO2 was derived from Shackleton (2000); the time scale for the Cocos Ridge core is based on aligning the benthic oxygen isotope record (Lea et al. 2002) from core TR163-19 to nearby core V19-30, which was previously aligned to Vostok (Shackleton 2000). The modern SST at the tropical Pacific site is 26.2°C (Levitus and Boyer 1994). The TR163-19 SST record is based on Mg/Ca analyses published in Lea et al. (2000), supplemented by additional analyses in ∼15% of the intervals. SSTs are calculated from the G. ruber equation in Dekens et al. (2002)

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Fig. 4.
Fig. 4.

Amplitude spectrum of tropical SST, showing distribution between the three major earth orbital periods: e = eccentricity, t = tilt, p = precession. The fraction of the tropical amplitude that is coherent with Vostok CO2 is shown. The coherencies between the two records at eccentricity and tilt are 0.98 and 0.94, respectively. The lower plot shows the phase relationship between the two time series, where the results indicate that the two records are in phase at eccentricity and tilt. Results based on Blackman-Tukey method (Arand package) with 1/3 lags

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Fig. 5.
Fig. 5.

Comparison of temperature differences (ΔT), relative to modern sea surface temperature, at the tropical Pacific site, with greenhouse gas forcing calculated from Vostok CO2 and CH4 levels. The forcing is computed from standard radiative forcing relationships (Ramaswamy et al. 2001) relative to preindustrial levels of 280 ppm for CO2 and 700 ppb for CH4; N2O was assumed constant at 280 ppb (Raynaud et al. 2000). For reference, equivalent atmospheric CO2 values are shown on the top axis. Both datasets have been interpolated to a constant sampling interval of 2 kyr, which is the average resolution of the tropical Pacific site (Lea et al. 2000). Regression through the data yields a slope, equivalent to the climate sensitivity parameter λ, of 1.4° ± 0.1°C (W m−2)−1 and the intercept of 0.5° ± 0.2°C (r2 = 0.69; errors are 95% CI). Forcing the regression through zero yields a slope of 1.10° ± 0.07°C (W m−2)−1 (r2 = 0.64). For the 4 W m−2 forcing associated with a doubling of atmospheric CO2, these values of λ equate to a 4.4°–5.6°C warming

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Fig. 6.
Fig. 6.

Results from the multivariate analysis. (a) The relative weighting of the three climate forcings, ΣGHF, Jul 65°N insolation, lagged 2 kyr, and 2°N fall insolation (Paillard et al. 1996), that provide the best fit to the observed Cocos Ridge SST record (see Table 2 for the individual weightings). The temperature change that can be attributed to each factor is shown, along with the sum of the three influences (T mean calc). Temperature changes are shown as differences from the mean. (b) The modeled temperature change compared to the observed temperature change. The sum of three climate forcings explains 79% of the total variance in the SST record. Temperature changes (ΔT) shown in differences from present SST

Citation: Journal of Climate 17, 11; 10.1175/1520-0442(2004)017<2170:TYCITS>2.0.CO;2

Table 1.

Spectral analysis results for comparison of the Cocos Ridge SST record (TR163-19) to the Vostok atmospheric carbon dioxide record. Confidence limits for phase lags are at the 95% level

Table 1.
Table 2.

Results from the multivariate (MV) analysis, shown as % variance in the Cocos Ridge SST record explained by the various fits. ΣGHF calculated from the sum of CO2 and CH4 (see text). Insolation calculated using Paillard et al. (1996), with fall 2°N insolation integrated from Aug through Sep. The relative weighting for each factor in the multivariate fits was determined in a stepwise fashion, progressing from the strongest to the weakest weighting. Individual fits were determined by linear regression of the evaluated forcing against SST. All records (1–361 kyr BP) were interpolated to 2-kyr spacing

Table 2.
Save