REPORTS

land (13) and between 300 and 600 years for

Ice Core Records of TD (14) during termination I. Further uncer-
tainty is added because the TD CO2 record

Atmospheric CO2 Around the has been dated relative to the Greenland Ice
Sheet Project 2 (GISP2) core (14), whereas

Last Three Glacial Terminations

the Byrd and Vostok isotope temperature
records have been synchronized with respect
to the Greenland Ice Core Project (GRIP) ice
Hubertus Fischer, Martin Wahlen, Jesse Smith, core record (13). This uncertainty is not rel-
Derek Mastroianni, Bruce Deck evant for the interval between 10 and 15 ky
B.P., for which dating of GISP2 and GRIP is
Air trapped in bubbles in polar ice cores constitutes an archive for the recon- in good agreement; however, there is a shift
struction of the global carbon cycle and the relation between greenhouse gases of up to 2000 years between the two Green-
and climate in the past. High-resolution records from Antarctic ice cores show land reference cores at the age of 20 ky B.P.
that carbon dioxide concentrations increased by 80 to 100 parts per million by In Fig. 1, our data and previously pub-
volume 600 6 400 years after the warming of the last three deglaciations. lished CO2 concentration records (1, 6, 9, 11,
Despite strongly decreasing temperatures, high carbon dioxide concentrations 15, 16) are compared with Antarctic isotope
can be sustained for thousands of years during glaciations; the size of this phase (temperature) ice core records (13, 17–19).
lag is probably connected to the duration of the preceding warm period, which Note that the CO2 concentrations represent
controls the change in land ice coverage and the buildup of the terrestrial essentially a global signal. In contrast, the
biosphere. geographical representativeness of isotope

Downloaded from http://science.sciencemag.org/ on February 9, 2021

temperature records may vary from a synop-
Previous studies of Antarctic ice cores (1–3) mate deglaciation (about 70 to 160 ky B.P.) at tical to hemispherical scale and accordingly
revealed that atmospheric CO2 concentra- a high resolution of 100 to 2000 years (8). within different cores with increasing vari-
tions changed by 80 to 100 parts per million This data set was supplemented by a CO2 ability for shorter time scales. The Vostok
by volume (ppmv) during the last climatic record recently derived from the Antarctic and TD CO2 data presented here are in good
cycle and showed, together with continuous Taylor Dome (TD) ice core (6, 9) covering agreement with previous CO2 values. On a
atmospheric measurements (4), that anthro- the last 35,000 years. The internal temporal 10,000-year time scale, CO2 covaries with
pogenic emissions increased CO2 concentra- resolution of ice core air samples is restricted the isotope temperatures with minimum gla-
tions from 280 ppmv during preindustrial by the age distribution of the bubbles caused cial CO2 concentrations of 180 to 200 ppmv,
times to more than 360 ppmv at present, an by the enclosure process (10). This age glacial-interglacial transitions accompanied
increase of more than 80% of the glacial- spread is about 300 years for Vostok (11) and by a rapid increase in CO2 concentrations to
interglacial change. Variations in atmospher- 140 years for the TD ice core (9) at present a maximum of 270 to 300 ppmv, and a
ic CO2 concentrations accompanying glacial- but about three times higher for glacial con- gradual return to low CO2 values during gla-
interglacial transitions have been attributed to ditions (11). The depth–ice age scale used for ciation. On a shorter time scale, however, a
climate-induced changes in the global carbon terminations II and III in the Vostok core is a much more complex picture evolves.
cycle (5, 6), but they also amplify climate recently expanded version of the extended The onset of the atmospheric CO2 in-
variations by the accompanying greenhouse glaciological time scale (12). The dating un- crease during termination I recorded in the
effect. Accordingly, the relation of tempera- certainty (on the order of 10,000 years for TD record is at 19 to 20 ky B.P. The rise in
ture and greenhouse gases in the past derived termination III) is considerable; however, the the long-term trend in CO2 concentrations
from ice core records has been used to esti- absolute time scale is not so important as long seems to be about 1000 years earlier than the
mate the sensitivity of climate to changes in as we consistently compare Vostok CO2 with rise in Vostok dD values. In contrast, temper-
greenhouse gas concentrations (7) to con- the Vostok isotope temperature (dD) record. atures apparently started to rise at 20 ky B.P.,
strain the prediction of an anthropogenic More important is the relative dating of as recorded in the Antarctic Byrd and the
global warming. This procedure, however, ice and air at a certain depth. The ice age–air Greenland GRIP ice core (13). Again, CO2
requires the separation of systematic varia- age difference (Dage) was calculated with a concentrations in the Byrd record increase
tions representative for all climatic cycles climatological firn densification model (11) ;2000 6 500 years later than those in the TD
from those specific for each event, as well as and varies between about 2000 and 6000 data. In view of the excellent agreement for
a more detailed knowledge of the leads and years for warm and cold periods, respective- the rest of the CO2 records, these discrepan-
lags between greenhouse gas concentrations ly. The accuracy of the model is better than cies can be attributed to the insufficient age
and climate proxies. 100 years for recent periods but on the order constraint during the onset of termination I
To resolve short-term changes in the at- of 1000 years for glacial conditions (11), induced by the different Greenland reference
mospheric carbon reservoir, to constrain the which has to be kept in mind when interpret- cores. No such dating uncertainties are en-
onset and end of major variations in CO2 ing the phase shift between ice and gas countered for the interval between 10 and 15
concentrations, and to test whether these vari- records of the ice core archive. In the case of ky B.P. Maximum CO2 concentrations of 270
ations are temporally representative, we ex- termination I, recently published age scales ppmv are reached at 10.5 ky B.P. (9), 600 to
panded the Antarctic Vostok CO2 record over derived by synchronization of CH4 variations 1000 years after the isotope temperature max-
the transition from marine isotope stage in central Greenland and Antarctic ice cores imum in the Byrd record (20). The CO2 peak
(MIS) 8 to MIS 7 [about 210 to 250 thousand (13, 14) were used. The precision of the CH4 is followed by a decrease of 5 to 10 ppmv
years (ky) before present (B.P.)] and ana- correlation is about 200 years for periods of until 8 ky B.P., after which CO2 concentra-
lyzed the time interval around the penulti- substantial CH4 change but is not very well tions gradually rise to the preindustrial value
constrained in the interval between 17 and 25 of 280 ppmv (9). A delay in the increase of
Scripps Institution of Oceanography, Geosciences Re-
ky B.P., when only subtle CH4 changes oc- CO2 concentrations with respect to the warm-
search Division, University of California San Diego, La curred. The uncertainty of Dage varies be- ing during deglaciation is also indicated by a
Jolla, CA 92093– 0220, USA. tween 100 and 300 years for central Green- brief 10-ppmv decline in CO2 concentrations

1712 12 MARCH 1999 VOL 283 SCIENCE www.sciencemag.org

 REPORTS
found in seven samples during the interval 14 three terminations, connected to a climate- nal climate forcing by CO2 emissions due
to 13 ky B.P. This distinct feature lags the driven net transfer of carbon from the ocean to combustion of fossil fuel leads climate
Antarctic Cold Reversal (ACR) in the Ant- to the atmosphere (6). The time lag of the rise variations, so the application of the CO2-
arctic isotope temperatures (21) by 300 to in CO2 concentrations with respect to tem- climate relation deduced from the past on a
500 years but occurs 1000 years before the perature change is on the order of 400 to 1000 recent global warming seems not to be
Younger Dryas cooling event. years during all three glacial-interglacial tran- straightforward.
A dip in CO2 concentrations at 135 ky B.P. sitions. Considering the uncertainties in Dage The situation is even more complicated
precedes the start of the increase in CO2 con- (between 100 and 1000 years for recent and for the interglacial and glaciation periods.
centrations during termination II, which reaches glacial conditions), such a lag can still be During the extended Holocene and Eemian
a maximum of 290 ppmv at 128 ky B.P. Like in explained by an overestimation of Dage for warm periods, atmospheric CO2 concentra-
the Holocene, CO2 concentrations decrease af- glacial conditions. The good agreement of the tions drop by ;10 ppmv after an initial max-
ter this initial maximum to ;275 ppmv. The Dage model with the measured value for the imum, attributable to a substantial increase in
onset of the major warming during termination present supports the idea that at least the lag the terrestrial biospheric carbon storage ex-
II is hard to define, but during the penultimate at the beginning of the warm periods is real. tracting CO2 from the atmosphere. In the case
warm period, CO2 concentrations reach their The size of this lag is on the order of the of the Eemian, CO2 concentrations remain
maximum 400 6 200 years later than Antarctic ocean mixing time (for a well-ventilated constant after the initial maximum in MIS 5.5
temperatures. In the following 15,000 years of ocean like today), which is the major control despite slowly decreasing temperatures; dur-
the Eemian warm period, CO2 concentrations for the time constant of equilibration within ing the Holocene, atmospheric CO2 concen-
do not show a substantial change despite dis- the deep ocean–atmosphere carbon system trations even increase during the last 8000
tinct cooling over the Antarctic ice sheet. Not after climate-induced changes. In the case of years. Application of a carbon cycle model to
until 6000 years after the major cooling in MIS a recent anthropogenic warming, the exter- CO2 and d13CO2 ice core data for the Holo-

Downloaded from http://science.sciencemag.org/ on February 9, 2021

5.4 does a substantial decline in CO2 concen-
tration occur. Another 4000 to 6000 years is
required to return to an approximate in-phase
relation of CO2 with the temperature variations.
Finally, termination III starts with a CO2
concentration of 205 ppmv at 244 ky B.P.,
slightly higher than that for the beginnings of
terminations I and II. At that time, tempera-
tures had already increased since the glacial
temperature minimum at ;260 ky B.P. CO2
concentrations rise slowly from 244 to 241 ky
B.P. and then rapidly to more than 300 ppmv
at 238 ky B.P. Keeping the rather coarse
resolution of the dD record before 238 ky
B.P. in mind, the major increase in CO2 tends
to lag temperature during the transition,
reaching a maximum CO2 concentration
600 6 200 years after the peak in dD. In
contrast to the case for the Eemian, high CO2
concentrations are not sustained during MIS
7 but follow the rapid temperature drop into
MIS 7.4. Minimum CO2 concentrations as
low as 210 ppmv are reached 1000 to 2000
years after the minima in isotope temperature
during MIS 7.4. A short, warm event during
the mild glacial interval at 224 to 228 ky B.P.
appears to be reflected in a 30-ppmv increase
in atmospheric CO2 concentrations with a
phase lag of about 1000 6 600 years relative
to temperature. Another warm event at the
beginning of the warm period MIS 7.3 is
accompanied by a 30-ppmv increase in CO2
concentration, which appears to be in phase
with the temperature record. The variations in
CO2 concentrations during these events are
much larger than anticipated from the Vostok
isotope temperature changes and do not have
any counterparts during MIS 5.
Comparison of the sequence of events for
the three time intervals described above sug- Fig. 1. Records of atmospheric CO2 concentrations and isotope temperature records derived from
the Antarctic Byrd, Vostok, and TD ice cores during the deglaciation and glaciation events around
gests that the carbon cycle–climate relation the last three glacial terminations. Error bars in CO2 concentration data represent 1s of replicate
should be separated into (at least) a deglacia- measurements at the same depth interval. The long-term trend in CO2 concentrations is indicated
tion and a glaciation mode. Atmospheric CO2 by a cubic spline approximation (P 5 5 3 1029) of our data set. For convenience, marine isotope
concentrations show a similar increase for all stages (22) are indicated as referred to in the text.

www.sciencemag.org SCIENCE VOL 283 12 MARCH 1999 1713

 REPORTS
cene (9) shows that no equilibrium in the 4. C. D. Keeling, T. P. Whorf, M. Wahlen, J. van der described by J. Jouzel et al. [Nature 364, 407
carbon cycle is established and that the wax- Pflicht, Nature 375, 666 (1995). (1993)], is in preparation.
5. M. Leuenberger, U. Siegenthaler, C. C. Langway, ibid. 13. T. Blunier et al., Nature 394, 739 (1998).
ing and waning of the terrestrial biosphere, 357, 488 (1992). 14. E. J. Steig et al., Science 282, 92 (1998).
possibly related to subtle climate variations 6. H. J. Smith, H. Fischer, M. Wahlen, D. Mastroianni, B. 15. A. Neftel, H. Oeschger, T. Staffelbach, B. Stauffer,
and early human land use, are the most im- Deck, in preparation. Nature 331, 609 (1988).
7. C. Lorius, J. Jouzel, D. Raynaud, J. Hansen, H. Le Treut, 16. B. Stauffer et al., ibid. 392, 59 (1998).
portant factors controlling atmospheric CO2 Nature 347, 139 (1990). 17. J. Jouzel et al., Clim. Dyn. 12, 513 (1996).
concentrations over the last 10,000 years. 8. M. Wahlen, D. Allen, B. Deck, A. Herchenroder, Geo- 18. J. R. Petit et al., Nature 387, 359 (1997).
During further glaciation in MIS 5.4, CO2 phys. Res. Lett. 18, 1457 (1991). Air samples were
19. S. J. Johnsen, W. Dansgaard, H. B. Clausen, C. C.
extracted from Vostok 5G and TD ice with a dry
concentrations remain constant, although extraction technique, and CO2 concentrations were
Langway Jr., ibid. 235, 429 (1972).
temperatures strongly decline. We suggest determined with laser spectroscopy. The accuracy of 20. Phase relations were determined by comparison of
a single measurement (as essentially determined by maxima and minima in the long-term trend of CO2
that this reflects the combination of the in- concentrations and isotope temperatures as repre-
the standard deviation of multiple frequency tunings
creased oceanic uptake of CO2 expected for sented by spline approximations. Given errors reflect
of the diode laser) is better than 5 ppmv. The laser
colder climate conditions and CO2 release spectroscopic method enables the use of very small the uncertainty in the actual positions of the ex-
caused by the net decline of the terrestrial samples (;4 g), allowing us to pick crack-free ice and trema, which are weakly dependent on the degree of
to measure replicate samples at the same depth smoothing. They do not take into account the uncer-
biosphere during the glaciation and possibly tainty in Dage. This additional error is treated sepa-
interval. In general, all given CO2 concentrations
by respiration of organic carbon deposited on correspond to the average and standard deviation of rately in the discussion of the data.
increasingly exposed shelf areas. These pro- at least three replicate samples. On average, the 21. T. Blunier et al., Geophys. Res. Lett. 24, 2683 (1997).
variability of such replicate measurements is 7.5 22. D. G. Martinson et al., Quat. Res. 27, 1 (1987).
cesses, however, should terminate (with some
ppmv (1s). 23. J. Imbrie et al., in Milankovitch and Climate, A. Berger
delay) after the lowest temperatures are 9. A. Indermühle et al., Nature, in press. et al., Eds. (Reidel, Hingham, MA, 1984), pp. 269 –
reached in MIS 5.4 and ice volume is at its 10. J. Schwander et al., J. Geophys. Res. 98, 2831 (1993). 305.
maximum at 111 ky B.P. (22). In agreement 11. J.-M. Barnola, P. Pimienta, D. Raynaud, Y. S. Korot- 24. We thank J.-M. Barnola and D. Raynaud for helpful
kevich, Tellus Ser. B 43, 83 (1991). comments and for sharing with us their unpublished

Downloaded from http://science.sciencemag.org/ on February 9, 2021

with this hypothesis, CO2 concentrations start 12. Expanded ice age and air age time scales were Vostok CO2 record of the last four glacial-interglacial
to decrease in the Vostok record at about 111 kindly provided by J. Jouzel and J.-R. Petit. Ages cycles during our sample selection process. This study
ky B.P. Another possibility to explain this were assigned to sample depths after slight depth was funded by NSF grants OPP9615292, OPP9196095,
corrections for the Vostok 5G core (17) by linear and OPP9118534. Financial support of H.F. has been
delayed response of CO2 to the cooling dur- provided by Deutsche Forschungsgemeinschaft.
interpolation of the depth-age scale. A publication
ing MIS 5.4 would be an inhibited uptake of describing the calculation of the expanded time
CO2 by the ocean. In any case, about 5°C scales, which is essentially based on the procedure 30 November 1998; accepted 29 January 1999
lower temperatures on the Antarctic ice sheet
during MIS 5.4 (17) are difficult to reconcile
with the full interglacial CO2 forcing encoun-
tered at the beginning of this cold period and Present-Day Deformation
again question the straightforward applica-
tion of the past CO2-climate relation to the Across the Basin and Range
recent anthropogenic warming.
Another scenario is encountered during
MIS 7, in which no prolonged warm period is
Province, Western
observed. Although temperatures at the end
of termination III are comparable to those at
United States
the end of termination II and CO2 concentra- Wayne Thatcher,1* G. R. Foulger,2 B. R. Julian,1 J. Svarc,1
tions are even slightly higher, a much shorter
lag in the decrease of CO2 relative to the
E. Quilty,1 G. W. Bawden1
Antarctic temperature decrease in MIS 7.4 is
found. Comparison with the SPECMAP The distribution of deformation within the Basin and Range province was
record (23) shows that during the preceding determined from 1992, 1996, and 1998 surveys of a dense, 800-kilometer–
interglacial MIS 7.5, ice volume was much aperture, Global Positioning System network. Internal deformation generally
larger than during the Holocene and the Ee- follows the pattern of Holocene fault distribution and is concentrated near the
mian warm periods. Accordingly, the buildup western extremity of the province, with lesser amounts focused near the
of the terrestrial biosphere during MIS 7.5 is eastern boundary. Little net deformation occurs across the central 500 kilo-
expected to be much less and sea level chang- meters of the network in western Utah and eastern Nevada. Concentration of
es smaller, leading to a smaller net release of deformation adjacent to the rigid Sierra Nevada block indicates that external
CO2 into the atmosphere during the following plate-driving forces play an important role in driving deformation, modulating
glaciation, which is not able to fully counter- the extensional stress field generated by internal buoyancy forces that are due
balance the CO2 uptake by the ocean. to lateral density gradients and topography near the province boundaries.

References and Notes The northern Basin and Range province is an The province has extended (increased in area)
1. J. M. Barnola, D. Raynaud, Y. S. Korotkevich, C. Lorius, actively deforming intracontinental plateau by about a factor of 2 over the past ;20
Nature 329, 408 (1987).
2. A. Neftel, E. Moor, H. Oeschger, B. Stauffer, ibid. 315,
lying between the stable blocks of the Sierra million years (1, 2), and extension continues
45 (1985). Nevada and the Colorado Plateau (Fig. 1). with ongoing seismic activity and slip along
3. D. Raynaud et al., Science 259, 926 (1993). Recent numerous faults distributed across a zone
investigations in central Greenland have reported an
in situ production of CO2 in the ice, possibly related
;800 km wide (3–5). Constraints on the
1
U.S. Geological Survey, MS/977, Menlo Park, CA
to carbonate or organic species reactions (or both), 94025, USA. 2Department of Geological Sciences,
internal deformation of the province are lim-
and have strongly compromised the validity of the University of Durham, Science Laboratories, South ited. Geologic studies delineate regions of
determined CO2 concentrations. However, Antarctic Road, Durham DH1 3LE, UK. Holocene and late Quaternary fault slip (3, 4).
ice cores are (if at all) much less affected by this
effect because of the very low abundance of reactive *To whom correspondence should be addressed. E- Space geodetic measurements broadly define
carbon species dissolved in Antarctic ice. mail: [email protected] movements across the province (6–8), and

1714 12 MARCH 1999 VOL 283 SCIENCE www.sciencemag.org

Ice Core Records of Atmospheric CO2 Around the Last Three Glacial Terminations
Hubertus Fischer, Martin Wahlen, Jesse Smith, Derek Mastroianni and Bruce Deck

Science 283 (5408), 1712-1714.

DOI: 10.1126/science.283.5408.1712

Downloaded from http://science.sciencemag.org/ on February 9, 2021

ARTICLE TOOLS http://science.sciencemag.org/content/283/5408/1712

REFERENCES This article cites 18 articles, 1 of which you can access for free
http://science.sciencemag.org/content/283/5408/1712#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS.
Copyright © 1999 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science.
No claim to original U.S. Government Works.