I. C. Prentice




Max Planck Institut for

Biogeochemislry fen a, Germany

D. Raynaud

Laboratoire de Glaciologie et de

Carbon Dioxide


Scientific Challenges Posed by the Ice-Core Records


Géophysique de ï Environnement St. Martin d/ Hères. France


Toward an Integrated Research Strategy for Paleobiogeochemistry



Paleobiogeochemistry—this word, as far as we know, did not previously exist. It defines a newly emerging research field that, we believe, will within the next decade come to play a central role in our understanding of the earth system and of how human activities are modifying that system.

Paleobiogeochemistry draws its inspiration and challenge from the polar ice-core records, which have provided a window on the natural dynamics of atmospheric composition and the relationship of atmospheric composition to an ever-changing climate. Technological developments in ice-core drilling systems, together with the associated systems that had to be developed to allow the secure transport and laboratory sampling of ice, have made the ice-core records possible. The increasing refinement of analytical methods has allowed us to determine concentrations of the impurities and gases incorporated in ice. These include not only the relatively abundant atmospheric components such as C02 and mineral dust, but also many other atmospheric constituents present in far lower amounts such as CH4, N20, light carboxylic acids, and isotopes such as l3C02, l80l60, and even l70l60 and l 3CH4.

High-precision measurements in ice cores now provide a rich source of information about most of the significant, radiatively active constituents of the atmosphere—greenhouse trace gases and aerosols—and, along with this information, data on numerous tracers that help to elucidate mechanisms associated with natural changes in the abundances of these constituents. The overarching challenges posed by the ice-core records of the changing atmospheric composition can be summarized as follows.

• The greenhouse gases CO,, CH4, and N20 and aerosols containing SO.,'-, volatile organic compounds, and mineral dust have varied in abundance during the past half-million years in a systematic manner, showing periodicities characteristic of the earth's orbital variations (the Milankovitch frequencies ~ 20, ~ 40, ~ 100 ka; 1 ka = 1000 years) and a clear association with glacial-interglacial cycles on the 100-ka time scale (Fig. 1) (Petit etai, 1999).

• The changes in greenhouse gases and aerosols could in principle represent either effects or drivers of climate change. Calculations suggest that the total contribution of natural variations in some atmospheric constituents, such as CH4, to global radiative forcing of climate cannot be large. The observed natural variations in CH4 concentration prior to the present human perturbation of the CH4 cycle, represent a response to climate change, and not a significant driver of climate change (Lorius and Oeschger, 1994). Glacial-interglacial C02 changes are large enough to be significant drivers of climate (Raynaud et al, 2000, and references therein). On the other hand, temporal patterns of change in atmospheric C02 concentration over the past four glacial-interglacial cycles suggest that C02 (as well as CH4) concentrations respond to climate change (Fischer et al, 1999). Atmospheric dust loading clearly responds to climate change, but inclusion of dust as a radiatively active atmospheric constituent may be needed to produce an accurate simulation of glacial climates (Kohfeld and Harrison, 2000; Claquin et al, submitted). In other words, these atmospheric constituents (C02 and dust) represent interactive components of the earth system that both influence and are influenced by the changing climate (Prentice, in press). We cannot properly understand the dynamics of climate unless we understand not only how these atmospheric constituents influence climate, but also how climate change influences their atmospheric concentrations by altering the strengths of their natural sources and sinks.

• Despite the richness of the records, ice-core data alone can give only clues and not definitive answers concerning the mechanisms of the observed changes. This is because climate and atmospheric composition are multidimensional phenomena that cannot be adequately indexed by measurements from one or two regions. Changes in climate have distinctive spatial sig-


Copy right v 2001 by - ■ ■ i ■-:: ■ i1;, ■ . All rights o! K'pRxhKlion in Liny ionn reserved. 81

Dust concentration

0 2500 5000 7500 10000

FIGURE 1 Top: Records of CO,, CH4, and dust concentrations from the Vostok Antarctic ice core (Petit et al., (1999) covering four glacial-interglacial cycles. Bottom: Holocene records of C02 and CH4 concentrations. CO, from Taylor Dome, Antarctica (Indermuehle et ai, 1999); CH4 from the Greenland summit (Chapellaz et al., 1997a).

0 2500 5000 7500 10000

FIGURE 1 Top: Records of CO,, CH4, and dust concentrations from the Vostok Antarctic ice core (Petit et al., (1999) covering four glacial-interglacial cycles. Bottom: Holocene records of C02 and CH4 concentrations. CO, from Taylor Dome, Antarctica (Indermuehle et ai, 1999); CH4 from the Greenland summit (Chapellaz et al., 1997a).

natures that cannnot be revealed by records confined to the polar regions (or extremely high elevations) where ice can persist through interglacial periods. The spatial heterogeneity of climate change is an important determinant of the effects of climate change on ecosystem activity and trace-gas production; different ecosystems in different climatic zones can react in quite distinct ways to a given change in climate (Melillo et al, 1993; Cramer et ai, in press). Similarly, important atmospheric constituents with short lifetimes in the atmosphere, including reactive trace gases and aerosols, have heterogeneous spatial distributions and their spatial interactions are crucially important in determining the nature of their effect on climate (Dentener etal, 1996).

Thus, explanation of the ice-core records demands a creative interaction with other fields of earth system science. We identify two such fields as crucial. The first is global modeling, which is a standard approach to climate change problems (both present / future and past) and is being actively extended to include interactions of the biogeochemical cycles with climate. The second is Quaternary science in the broad sense, especially insofar as Quaternary paleo-environmental records have been standardized and compiled on a global scale. Quaternary scientists have amassed data from widely distributed natural archives such as marine and lacustrine sediments which, when suitably compiled, yield spatially extensive information about many aspects of climate, ecosystem composition, and even some indicators of atmospheric composition such as the flux density of mineral dust, over time scales (and, in some cases, temporal resolution) similar to those of the ice-core records (Ko-hfeld and Harrison, 2000).

This chapter is not by any means intended to be a review of ice-core research. Our purpose is rather to select a few illustrations of the specific scientific challenges raised by the ice-core records— especially the most recent, high-resolution records from Antarctica and the Greenland Summit—and to outline an interdisciplinary research strategy for tackling these challenges.

2. Methane

Despite its low natural concentration (< 1 ppm), CH4 can be measured in the air bubbles trapped in fossil ice with high precision and reliability. Once isolated from the "scavenging" effect of oxygen-containing free radicals in the atmosphere, CI I., ceases to be oxidized and its concentration becomes constant. The main features of the ice-core CI I,, records are as follows:

• A clear pattern of variation in concentration between glacial ( — 300 ppb) and intcrglacial periods (600-700 ppb), with additional variations at the Milankovitch frequencies of (= 20 and = 40 ka (Petit et al, 1999).

• Higher-frequency ("sub-iVlilankovitch") variability (Chap-pellaz et al, 1993a) associated with climatic fluctuations in the form of Dansgaard —Oeschger events, including the Younger Dryas cold interval that interrupted the last deglaciation.

• An extremely close tie to climate variations as indexed by isotopic signals in the ice cores, so that any possible lead or lag between CH4 and climate is within the — 50-year de-tectability horizon caused by the finite time taken for occlusion of air bubbles in the firn.

• A distinctive pattern (Blunier et al, 1995; Chappellaz et al., 1997a) of relatively smooth changes in concentration during the Holocene (the present interglacial, starting at »11.6 ka), with amplitude —100 ppb. CH4 concentration was relatively high near the beginning of the Holocene, low during the middle Holocene (minimum around 6 ka B.P., before present), and rising again after 6 ka B.P.

• Changes in the interhemispheric gradient, as shown by the difference in concentration between Greenland and Antarctic ice cores, suggesting that both tropical and northern high-latitude sources are involved in the glacial —interglacial, sub-Milankovitch glacial, and Holoccnc changcs in the concentration of CH4 (Chappellaz et al., 1997a; Dallenbach etai, 2000).

This last point could be established because of the high measurement precision that can be achieved for CH4 in ice cores. Its fast atmospheric mixing time and its fast response to climate change allow the qualitative features of the CH4 records to be used as a way of synchronizing records from the two hemispheres. These features have been useful in various other contexts. For ex ample, using CH4 and §180 of 02 measurements from Antarctic ice, it was possible to show that the lowest layers of the GRIP and GISP2 cores from the Greenland summit, representing the last or Eemian interglacial, were not in a correct temporal sequence (Chappellaz etai, 1997b).

3. Carbon Dioxide

C02 concentrations in the atmosphere have generally been 3 orders of magnitude higher than CH4 concentrations, yet paradoxically it has proved far more difficult to obtain a reliable, high-resolution record for CO,. The problem is the presence of mineral and/or organic impurities in the ice. The much higher content of impurities in Greenland ice than in Antarctic ice can lead to in situ C02 production via acid-carbonate interactions or oxidation of organic material and significantly alter the original concentration of C02 in the ancient air (Delmas, 1993; Anklin et al, 1997; Haan and Raynaud, 1998).

The Vostok core, from the central part of East Antarctica, provided conclusive evidence for glacial-interglacial changes in atmospheric C02 concentration (Barnola et al, 1987; Petit et al, 1999), but the low precipitation rate limits the achievable temporal resolution at Vostok. The Byrd ice-core record from coastal Antarctica (Neftel et al, 1988), with higher precipitation rates, yielded an apparent Holocene signal with a sharp C02 concentration peak at the beginning of the Holocene followed by a drawdown and subsequent recovery (by about 6 ka B.P.) to approximately the "preindustrial" concentration of —280 ppm. This record was long regarded as problematic, because the main changes occurred in a part of the core where the ice was exceptionally brittle. More recent measurements from Taylor Dome (In-dermuehle et al, 1999), another high-resolution site lacking the problems of the Byrd core, have shown that the Byrd measurements overestimated the amplitude of Holocene variability in C02 concentration. Based on present knowledge, the major features of the ice-core C02 records are as follows:

• On glacial-interglacial time scales, there is systematic variation between <200 ppm during glacial periods (185 ppm at the last glacial maximum around 21 ka B.P.) and —280 ppm during interglacial periods, with additional variability particularly in the = 40-ka band (Petit et al, 1999).

• There is some higher-frequency variability (about ± 10 ppm) associated with Heinrich iceberg-discharge events during the last glacial period. However, the CO, record essentially lacks the imprint of the faster Dansgaard-Oeschger events (Stauffer et al, 1998). During the Younger Dryas, deglacial warming of the Northern Hemisphere was interrupted while CO, concentration continued to rise (Blunier etai, 1997).

• There has been relatively small variability (range 260-280 ppm) during the Holocene, but with a distinctive pattern: a slight fall in concentration to a minimum of — 260 ppm at

~8 ka B.P., followed by a steady rise toward 280 ppm in preindustrial time (Indermuehle et al, 1999). Note that although this pattern shares some features of the Holocene changes in CH4, its amplitude (relative to the average concentration of the two gases) is smaller, the pattern for CO, is less temporally symmetric than that for CH.„ and the timings of the Holocene minima for the two gases are 2000 years apart.

• A higher-resolution record of the past 1000 years provides evidence for a slight (amplitude < 10 ppm), temporary lowering of CO; concentration during the Little Ice Age (Etheridge et al, 1996).

We have no information on past changes in the interhemi-spheric gradient of atmospheric CO; because accurate determination of past CO, concentrations from Northern-Hemisphere ice is still impossibile.

It is possible to measure the ¿' "C of CO, in ice. This measurement is potentially important because the l3C fractionation associated with CO, exchanges between the atmosphere and ocean is very different from the fractionation due to the predominant C, pathway of photosynthesis (Ehleringer, this volume; Kaplan and Buchmann, this volume). Recent measurements suggest that whereas the increase of CO, concentration following the last glacial maximum has the isotopic signature characteristic of a predominantly oceanic source of atmospheric C02 (Smith et al., 1999), the changes after 8 ka B.R bear a distinct signature consistent with progressive loss of carbon (amounting to = 200 Pg C altogether) from the terrestrial biosphere (Indermuehle et al, 1999). However, the measurement errors in the published S'^C measurements are uncomfortably large compared to the signal, indicating the importance of improving the repeatability of these measurements as well as increasing the sampling density for the Holocene. The SI3C record of the Little Ice Age is also consistent with the small CO, anomaly being due to a temporary increase in carbon storage on land (Francey et al, 1999).

The distinctive spatial pattern of lsO in precipitation, caused by fractionation in evaporation and condensation, applies to atmospheric C02 when C02 dissolves in leaf water; this fact has been exploited to provide "top-down" estimates of global gross primary productivity (Farquhar et al, 1993; Ciais et al, 1999). The oxygen isotope composition of C02 is not preseved in ice cores because C02 exchanges oxygen atoms with the ice. However, 02 propagates the l80 signature of CO, because photosynthesis and respiration (including photorespiration) uniquely transfer oxygen atoms from CO, to O, (Bender et al, 1996). The difference between the <5I80 of seawater (determined mainly by continental ice volume) and the <5I80 of O, in air bubbles trapped in ice is known as the "Dole effect" and show promise as a palaeotracer for primary productivity. Additional, independent information on productivity is provided by the l70 content of O, (Barkan et al, 1999). Due to differences in the mechanisms of fractionation of the 3 stable oxygen isotopes in reactions involving stratospheric O,, these isotopes together can be used to infer 02 turnover by the biosphere (terrestrial and marine).

4. Mineral Dust Aerosol

Aerosols include both soluble and insoluble components. The insoluble component, which consists of mineral dust of terrestrial origin, can be measured directly in ice cores. The soluble component, which includes components from sulfate aerosol and sea salt, can be assessed by measuring the concentration of base cations such as Na" and Mg2" and anions such as S042-and CI-.

Unlike CH4 and C02, terrigenous dust is not a well-mixed atmospheric constituent; atmospheric concentrations of dust vary spatially and temporally by several orders of magnitude. Similarly, ice-core records show enormous variation compared to that showed by trace gases (Thompson and Mosley-Thompson, 1981; De Angelis et al, 1997; Petit et al, 1999). Also, what is measured in the ice cores is not directly the atmospheric concentration, but rather the concentration in the ice, which (taking into account the variable precipitation rate) can be used to infer the flux density of mineral dust reaching the ice surface (Mahowald et al, 1999). The main features of the ice-core dust records may be summarized as follows.

• There is a very strong glacial-interglacial pattern, with glacial periods typically showing an order of magnitude higher dust deposition in both Antarctica and Greenland. The dust records from Antarctica show spectral power in all of the Milankovitch bands (Petit et al, 1999).

• A trend during glacial periods, such that higher dust deposition occurs during the latter (colder) part of each glacial period. During the last glacial period, the highest dust deposition rates did not occur until—70 ka B.P. (marine isotope stage 4).

• High variability within glacial periods, associated with the recorded rapid climate changes and also showing considerable amplitude on interannual time scales.

• Very low deposition rates during the Holocene.

Heavy-element abundances and isotopic composition of dust in polar ice cores were compared with signatures of aeolian dust from different regions to assign likely source areas for the glacial dust. This exercise pointed to Central Asia as a likely major source region for glacial-age dust in Greenland (Biscaye et al, 1997), and to Patagonia for glacial-age dust in Antarctica (Basile et al, 1997). Particle size distribution is a useful ancillary statistic. Particle size spectra from Greenland dust support the hypothesis that average wind speeds during dust transport at the last glacial maximum were not much greater than the present, because the mean particle size hardly changed, although the data do indicate a slight increase in the amount of dust in the largest size classes (Steffensen, 1997).

5. Scientific Challenges Posed by the Ice-Core Records

Although the ice-core records illustrate the pervasive Mi-lankovilch periodicities and has yielded a great deal of information about natural changes in climate and biogeochemical cycles occurring at the Milankovitch and higher frequencies, the underlying causal sequences are hardly known. Herein lie the challenges of paleobiogeochemistry. We present some illustrative examples here, focusing as above on CH4, CO,, and mineral dust aerosol.

5.1 Methane

Natural variations in atmospheric CH4 concentration, as observed to date in ice cores, show no evidence for catastrophic CH4 hydrate release (Raynaud et al., 1998), and consequently do not support the speculation (Nisbet, 1992) that CH4 releases from marine clathrates were implicated in triggering the last deglaciation. Normally, the main natural sources of CH4 arc on land, and the largest source component is due to methanogene-sis under anaerobic conditions in seasonal or permanent wetlands (Melillo et al., 1996). The glacial-interglacial changes in CH4 concentrations are too large to be fully accountcd for by plausible variations in the atmospheric chemical sink, and must therefore be explained, at least in part, in terms of the changing areas and activities of natural CH, sources (Pinto and Khalil, 1991, Chappellaz et al., 1993b, Crutzen and Bruhl, 1993, Thompson et al., 1993, Martinerie et al., 1995). It is plausible that the wetland extent was less during glacial times and also that net primary productivity of terrestrial wetlands was less, allowing the production of less substrate for methanogenic microorganisms. But it is not clear whether the rapid climate response of atmospheric CH4 concentration is due to effects of temperature on substrate formation and methanogenesis or to rapid changes in the areas of wetlands. Holocene changes in CH4 concentration were at first attributed to a balance of declining tropical monsoons (implying a reduction in the area of tropical wetlands) and later to an increasing build-up of boreal wetlands (e.g., on the Hudson Bay lowlands, which became exposed due to isostatic uplift only during the latter part of the Holocene) (Blunier et al., 1995). Analysis of the changing inter-hemispheric gradient of CH4 adds important information to these, showing that this picture is probably too simplistic (Chappellaz et al., 1997a). Other quantitative constraints may in future be brought to bear on this issue by measuring the iso-topic composition of CH4 in ice, while better quantification and modeling is needed for the impacts of climate change and atmospheric C02 concentration on wetland extent, CH4 release from wetlands, and sources and sinks of other biogenic trace gases (NO,., CO, and volatile organic compounds) that also affect the strength of the atmospheric sink for CH4.

5.2 Carbon Dioxide

Ice-core records have confirmed Svante Arrhenius' prescient hypothesis that variations in atmospheric C02 concentration were associated with glacial-interglacial cycles, but we are still uncertain about the primary cause of these variations. The first-order explanation must come from the ocean, where more than 90% of the total inventory of carbon in the ocean-atmosphere-terresterial biosphere system. (According to several lines of evidence, carbon storage on land was substantially less during glacial periods, e.g., Shackleton, 1977; Crowley, 1995; Friedlingstein et al., 1995; Bird et al, 1994; Peng et al., 1998.) One family of hypotheses to explain glacial-interglacial C02 variations relies on changes in the dissolution of C02 in the ocean. The effect of increased solubility of C02 in the ocean at low temperatures is insufficient and is counteracted by the effect of higher salinity during glacial periods. Stephens and Keeling (2000) proposed that the extended winter sea ice cover around Antarctica prevented the outgassing of up-welled, C02-rich water in glacial times. This is an attractive hypothesis in that it explains the synchroneity of increasing C02 concentration and Antarctic warming during déglaciations, as shown in Antarctic ice-cores. On the other hand, it postulates far less upwelling in low latitudes than most of the current ocean models allow. A second family of hypotheses relies on changes in nutrient supply or the efficiency of its utilization to increase marine biological productivity, thereby increasing the sinking flux of organic carbon and maintaining a stronger gradient of dissolved inorganic concentration away from the sea surface. The currently popular explanations along these lines invoke increased mineral dust aerosol input as an external source of Fe, which has been shown to limit production in the equatorial Pacific and Southern Oceans and which in addition may be generally limiting for nitrogen fixation in the open ocean (Martin, 1990; Falkowski, 1997; Broecker and Henderson, 1998; Pedersen and Bertrand, 2000). These explanations provide a putative link between C02 and mineral dust, over and above the fact that both have significant radiative forcing effects. A third family of hypotheses relies on various mechanisms that alter the alkalinity of the ocean. Both the nutrient and alkalinity hypotheses have problems to explain the full magnitude of the change without violating the constraints revealed by other information about marine sediments: proxy data for nutrients do not support great increases in productivity while calcium carbonate dissolution patterns do not support a large alkalinity change. It seems likely that more than one mechanism may be involved, but all of them need to be quantified better than they are at present. Useful additional information on the ice-core records is likely to come from geochemical and isotopic proxies for marine productivity, while a better quantification of the processes will require improved modeling of physical and biological processes in the ocean.

Holocene variations in C02 are also poorly understood. Even if it is true that terrestrial ecosystems lost carbon progressively since 8 ka B.P., the required magnitude of loss cannot be accounted for by the disappearance of vegetation in the Sahara (Indermuehle et al., 1999)—although this is by far the most extensive vegetation change that has occurred during the past 6000 years, according to pollen and plant macrofossil data assembled by the BIOME 6000 project (Prentice ct al., 2000). It is not clear a priori whether terrestrial carbon storage would be expected to increase or decline in response to the changing orbital configuration during the Ilolocene; this too requires quantification. More (and if possible, more precise) measurements of §13C in CO, from ice cores are clearly required, and the possible contribution of changes in marine chemistry through the I Iolocene needs to be more exactly calculated by ocean models.

5.3 Mineral Dust Aerosol

Analyses of dust from the polar ice cores have yielded information about candidate source areas. Although the changes observed in Greenland and Antarctica are large in a relative sense, in an absolute sense even the highest deposition rates to these remote areas are tiny compared to the contemporary rates in regions close to major dust sources, such as the Sahara or the Chinese loess plateau. Modeling is essential to establish links between the ice-core records and dust distribution outside the polar regions (Krogh-Andersen et al, 1998). Dust records have now been obtained from tropical ice cores and show mutually contradictory results, apparently reflecting predominance of large nearby sources with different histories and pointing to a high degree of spatial heterogeneity in the change of atmospheric dust content between glacial and interglacial periods. Such heterogeneity is evident in spatially distributed record of mineral dust deposition in the ocean, as compiled by the DIRTMAP project (Kohfeld and Harrison, 2000). Spatial heterogeneity is particularly crucial to determining the climatic effect of dust, because the sign of the radiative forcing due to dust is a function of the underlying surface albedo and is usually opposite over oceans and land (Tegen and Fung, 1995; Claquin et al, 1998). The lofting of dust from the land surface is itself dependent on climate and atmospheric C02 concentration, insofar as these variables control soil moisture and vegetation structure (Mahowald et al, 1999); therefore, a reciprocal relationship between mineral dust and atmopsheric C02 may exist. Advances in our understanding of the controls on mineral dust aerosol may occur in part through systematic "sourcing" of the dust in ice cores and marine sediments and will also require improved models for changing dust source areas and emission strengths.

6. Toward an Integrated Research Strategy for Paleobiogeochemistry

We propose the following general strategy to test hypotheses about the natural dynamics of the earth system.

1. The first step is to define clear data targets in the ice-core records. For example, the glacial-interglacial difference in CH4 concentration, the increase in C02 concentration since the early Holocene, and the various time scales of response and time sequences of changes in atmospheric CO,, CH4, and mineral dust.

2. The second step is to define model experiments that could be performed by (preferably several) earth system models. The models should include whatever components and linkages are hypothesized to be crucial to explaining the target data. (The models should also be applied without the specified components or linkages, to test their importance in the modeled world.) The target data will provide an immediate assessment of the extent to which the model experiments are successful.

3. Given that complex models can easily be right for the wrong reasons, especially when a global three-dimensional model is called upon to predict a single number, the third and essential step is to define the ancillary tests of model performance against (a) spatially distributed data (such as vegetation types, dust deposition fields, marine biogeochemical tracers) and (b) additional data from ice cores (such as isotopic measurements) that are relevant to the modeled processes. For example, a model to explain CH4 changes should be shown to generate realistic spatial distributions of wetlands; a model to account for C02 changes should also be able to hindcast ice-core measurements of <5L,C and the Dole effect; a model to account for temporal changes in atmospheric dust loading at the poles should be called upon to reproduce global patterns of dust fluxes to the oceans and loess accumulation on land. We note that this strategy implies running considerably more highly coupled models of the earth system than is possible today. Harrison et al (this volume) and Claussen (this volume) show that the incorporation of both ocean-atmosphere and vegetation-atmosphere feedbacks into atmospheric general circulation models is a prerequisite for the correct simulation of Holocene palaeoclimates. Without physical coupling of at-mopshere, oceans, and land, climate models cannot simulate past climates with sufficient accuracy to make our strategy viable. Furthermore, biogeochemical interactions through the carbon cycle and atmospheric chemistry have to be coupled to climate models, and specifically modules describing the sources and sinks of all key reactive chemical species and aerosol precursors have to be coupled to models of terrestrial and marine ecosystems. No existing model is sufficiently comprehensive to do all of these things; yet rapid progress is being made toward the development of true earth system models through separate activities such as coupled climate-carbon modeling, trace-gas source modeling in ecosystem models, and coupled climate - atmospheric-chemistry-transport models. Our strategy therefore relies on the continuation of a trend that already exists in global modeling and is being strongly promoted by the International Geosphere-Biosphere Programme (IGBP).

Earth system models in the limited form that they exist today fall into two major categories, namely, full three-dimensional

(3-D) models (based on atmospheric and ocean general circulation models) and reduced-form models or "models of intermediate complexity" in which the spatial resolution is generally lower (Kutzbach ct al., this volume) and atmospheric and ocean dynamics are represented in a parameterized, computationally efficient form (Schellnhuber, 1999). (Hybrids between these two types of model are beginning to appear, but this does not affect our argument.) The strategy we envisage allows an immediate role for both types of model, because for computational reasons experiments with 3-D models are likely to focus mainly on quasi-equilibrium conditions centered on canonical "time slices" while reduced-form models can far more readily perform multiple, transient simulations of long periods (e.g., the whole Ilolocene: Clanssen et al., 1999). For certain time slices, data already exist as convenient global summaries for comparison with model output (Kohfeld and Harrison, 2000). Transient analyses pose additional challenges to the Quaternary data community, to process the data into a suitable, synthetic form.

In conclusion, by insisting that the model results are routinely tested against the full spectrum of available palaeodata (from ice cores and other natural archives), we suggest that the study of past biogeochemical cycles can provide both a unique means to test complex earth system models and a powerful stimulus to their further development.


We thank the many scientists with whom we have discussed these matters, including Jérôme Chapellaz, Torben Christensen, Frank Dentener, Sandy Harrison, Ivar Isaksen, Sylvie Joussaume, Karen Kohfeld, Corinne Le Quéré, Particia Martinerie, Nathalie de No-blet, Henning Rodhe, Doug Wallace and many more. Karen Kohfeld and Jean-Robert Petit commented on an earlier draft.


Anklin, M., Schwander, J., Stauffer, B., Tschumi, J., Fuchs, A., Barnola J. M., and Raynaud, D. (1997). C02 record between 40 and 8 kyr BP from the Greenland Ice Core Project ice core. /. Geophys. Res-Oceans 102, 26539-26545.

Barnola, J. M„ Raynaud, D„ Korotkevich, Y. S., and Lorius, C. (1987). Vostok ice core provides 160,000-year record of atmospheric CO,. Nature 329, 408-414.

Balkan, E., Bender, M. L., Thiemens, M. H., Boering, K. A., and Luz, B. (1999). Triple-isotope composition of atmospheric oxygen as a tracer of biosphere productivity. Nature 400, 547-550.

Basile, L, Grousset, F. F.„ Revel, M„ Petit, J. R„ Biscaye, P. F.„ and Barkov, N. I. (1997). Patagonian origin of glacial dust deposited in East Antarctica (Vostok and Dome C) during glacial stages 2, 4 and 6. Earth Planet. Sci. Lett. 146, 573-589.

Bender, M„ Ellis, T„ Tans, P., Francey, R„ and Lowe, D. (1996). Variability in the 0,/N, ratio of southern hemisphere air, 1991-1994—implications for the carbon cycle. Global Biogeochem. Cycles 10, 9-21.

Biscaye, P. E., Grousset, F. E., Revel, M., van der Gaast, S., Zielinski, G. A., Vaars, A., and Kukla, G. (1997). Asian provenance of glacial dust (stage 2) in the Greenland Ice Sheet Project 2 Ice Core, Summit, Greenland. /. Geophys. Res. 102, 26765-26781.

Bird , M. I., Lioyd. J., and Farquhas, F. d. (1994). Tereshial carbon storage at the LGM. Nature 37, 556.

Blunier, T„ Chappellaz, J., Schwander, J., Stauffer, B., and Raynaud, D. (1995). Variations in the atmospheric methane concentration during the Holocene epoch. Nature 374,46-49.

Blunier, T„ Schwander, J., Stauffer, B„ Stocker, T„ Dällenbach, A„ Indermühle, A., Tschumi, J., Chappellaz, J., Raynaud, D., and Barnola, J.-M. (1997). Timing of the Antarctic Cold Reversal and the atmospheric CO, increase with respect to the Younger Dryas event. Geophys. Res. Lett. 24, 2683-2686.

Broecker, W. S., and Henderson, G. M. (1998). The sequence of events surrounding Termination II and their implications for the cause of glacial-interglacial C02 changes. Paleoceanography 13,352-364.

Chappellaz, J., Blunier, T., Kints, S., Diillenbach, A., Barnola, J.-M., Schwander, J., Raynaud, D„ and Stauffer, B. (1997a). Changes in the atmospheric CH4 gradient between Greenland and Antarctica during the Holocene. /. Geophys. Res. 102, 15987-15997.

Chappellaz, J., Brook, E„ Blunier, T„ and Malaize, B. (1997b). CH4 and S,sO of O, records from Antarctic and Greenland ice: A clue for stratigraphic disturbance in the bottom part of the Greenland Ice Core project and the Greenland Ice Sheet Project 2 ice cores./. Geophys. Res. 102, 26547-26557.

Chappellaz, J., Blunier, T., Raynaud, D., Barnola, J.-M., Schwander, J., and Stauffer, B. (1993a). Synchronous changes in atmospheric CH4 and Greenland climate between 40 and 8 kyr BP. Nature 366,443-445.

Chappellaz, J. A., Fung, I. Y„ and Thompson, A. M. (1993b). The atmospheric CH4 increase since the Last Glacial Maximum—(1) Source estimates. Tellus 45B, 228-241.

Ciais, P., Friedlingstein, P., Schimel, D. S., and Tans, P. P. (1999). A global calculation of the delta C-13 of soil respired carbon: Implications for the biospheric uptake of anthropogenic CO,. Global Biogeochem. Cycles 13, 519-530.

Claquin, T„ Roelandt, C., Kohfeld, K. E„ Harrison, S. P., Prentice, I. C., Balkanski, Y„ Bergametti, G„ Hansson, M„ Mahowald, N„ Rodhe, N„ and Schulz, M„ (submitted). Radiative forcing of climate by ice-age dust.

Claquin, T., Schulz, M., Balkanski, Y., and Boucher, O. (1998). Uncertainties in assessing radiative forcing by mineral dust. Tellus Series B— Client. Phys. Meteorol. 50,491-505.

Claussen, M„ Kubatzki, C., Brovkin, V., Ganopolski, A., Hoelzmann, P., and Pachur, H. J. (1999). Simulation of an abrupt change in Saharan vegetation in the mid-Holocene. Geophys. Res. Lett. 26, 2037-2040.

Cramer, W„ Bondeau, A., Woodward, F. I., Prentice, I. C., Betts, R. A., Brovkin, V., Cox, P. M., Fisher, V., Foley, J. A., Friend, A. D., Kucharik, C., Lomas, M. R., Ramankutty, N., Sitch, S., Smith, B., White, A., and Young-Moiling, C. (in press). Global response of terrestrial ecosystem structure and function to CO, and climate change: resutls from six dynamic global vegetation models. Global Change Biol.

Crowley, T. J., (1995). Ice-age terrestrial carbon changes revisited. Global Biogeochem. Cycles 9, 377-389.

Crutzen, P. J. and Brühl, C. (1993). A model study of atmospheric temperatures and concentrations of ozone, hydroxyl, and some other photochemi-cally active gases during the glacial, the preindustrial Holocene, and the present. Geophys. Res. Lett. 20, 1047-1050.

Dällenbach, A., Blunier, T., Fltickiger, J., Stauffer, B., Chappellaz, J., and Raynaud, D. (2000). Changes in the atmospheric CH4 gradient between Greenland and Antarctica during the Last Glacial and the transition to the Holocene, Geophys. Res. Lett. 27, 1005-1008.

De Angelis, M., StelTensen, J. P., Legrand, M. R., Clausen, H. B., and Hammer, C. U. (1997). Primary aerosol (sea salt and soil dust) deposited in Greenland ice during the last climatic cycle: Comparison with east Antarctic records./. Geopliys. Res. 102,26681-26698.

Delmas, R. J., (1993). A natural artefact in Greenland ice-core CO, measurements. Tellus 45B, 391-396.

Dcntcncr, F. J„ Carmichael, G. R., Zhang, Y., Lelieveld, J., and Crutzen, P. J., (1996). Role of mineral aerosol as a reactive surface in the global troposphere./. Geopliys. Res. 101, 22869-22889.

Etheridge, D. M„ Steele, L. P., Langenfelds, R. L„ Francey, R. J., Barnola, J. M„ and Morgan, V. I. (1996). Natural and anthropogenic changes in atmospheric CO, over the last 1000 years from air in Antarctic ice and firn. /. Geopliys. Res-Atmos. 101, 41 15-4128.

Falkowski, 1'. G. (1997). Evolution of the nitrogen cycle and its influence on the biological sequestration of CO, in the ocean. Nature 387, 272-275.

Farquhar, G. D„ Lloyd, J., Taylor, J. A., Flanagan, L. B„ Svvertsen,). 1'., Hubick, K. T., Wong, S. C., and Ehleringer, J. R. (1993). Vegetation effects on the isotope composition of oxygen in atmospheric CO,. Nature 363,439-443.

Fischer, H„ Wahlin, M„ Smith, J., Mastroianni, D„ and Deck, B. (1999). Ice core records of atmospheric C02 around the last three glacial terminations. Science 283, 1712-1714.

Francey, R. J., Allison, C. E„ Etheridge, D. M„ Trudinger, C. M„ Enting, I. G., Leuenberger, M„ Langenfelds, R. L„ Michel, E„ and Steele, L. P. (1999). A 1000-year high precision record of delta C-13 in atmospheric C02. Tellus Series B—Cliau. Pliys. Meleorol. 51, 170-193.

Friedlingstein, P., Prentice, K. C., Fung, I. Y., John, J. G., and Brasseur, G. P. (1995). Carbon-biosphere-climate interactions in the last glacial maximum climate. /. Geopliys. Res-Atmos. 100, 7203-7221.

Haan, IX, and Raynaud, I). (1998). Ice core record of CO, variations during the two last millennia: Atmospheric implications and chemical interactions within the Greenland ice. Tellus 50B, 253-262.

Indermuehle, A., Stacker, T. F., Joss, F., Fischer, H., Smith, H. J.> Wahlen, M„ Deck, B„ Mastroianni, D., Tschumi, J., Blunier, T., Meyer, R„ and Stauffer, B. (1999). Holocene carbon-cycle dynamics based on C02 trapped in ice at Taylor Dome, Antarctica. Nature 398, 121 - 126.

Kohfeld, K. E„ and Harrison, S. P., (2000). How well can we simulate past climates? Evaluating the models using global palaeoenvironmental datasets. Qual. Sci. Rev. 19, 321-346.

Krogh-Andersen K„ Armengaud, A„ and Genthon, C. (1998). Atmospheric dust under glacial and interglacial conditions. Geopliys. Res. Lett. 25, 2281-2284.

Lorius, C., and Oeschger, H. (1994). Paleo-perspectives-Reducing uncertainties in global change. Ambio 23, 30-36.

Mahowald, N„ Kohfeld, K. E„ Hansson, M„ Balkanski, Y„ Harrison, S. P., Prentice, I. C., Rodhe, H., and Schulz, M. (1999). Dust sources and deposition during the Last Glacial Maximum and current climate: a comparison of model results with paleodata from ice cores and marine sediments. /. Geopliys. Res. 104, 15895-16436.

Martin, J. (1990). Glacial-interglacial CO, change: the iron hypothesis. Paleo-ceanography 5, I - 13.

Martinerie, P., Brasseur, G. 1'., and Granier, C. (1995). The chemical composition of ancient atmospheres: a model study constrained by ice core data. /. Geopliys. Res. 100,14291-12304.

Melillo, J. M„ McGuire, A. D„ Kicklighter, D. W„ C.J. Vorosmarty III, B. M„ and Schloss, A. L. (1993). Global climate change and terrestrial net primary production. Nature 363, 234-240.

Melillo, J. M„ Prentice, I. C., Farquhar, G. D„ Schulze, E.-D., and Sala, O. E.

(1996). Terrestrial biotic response to environmental change and feedbacks to climate, In "Climate Change 1995. The Science of Climate Change." (Houghton, J. T„ Filho, L. G. M„ Callander, B. A., Harris, N., Kattenberg, A. and Maskell, K. Eds.), pp. 449-481. University Press, Cambridge.

Neftel, A., Oeschger, H„ Staffelbach, T„ and Stauffer, B„ (1988). CO, record in the Byrd ice core 50,000-5,000 years BP. Nature 331, 609-611.

Nisbet, E. G. (1992). Sources of atmospheric CH4 in early postglacial time. /. Geopliys. Res. 97, 12859-12867.

Pedersen, T. F. and Bertrand, P. (2000). Influences of oceanic rheostats and amplifiers on atmospheric C02 content during the Late Quaternary. Quat. Sci. Rev. 19,273-283.

Peng, C. H., Guiot, J., and Van Campo, E. (1998). Estimating changes in terrestrial vegetation and carbon storage: using palaeoecological data and models. Quat. Sci. Rev. 17, 719-735.

Petit, J. R„ Jouzel, J., Raynaud, D„ Barkov, N. I., Barnola, J. M„ Basile, I., Bender, M„ Chappellaz, J., Davis, M„ Delaygue, G., Delmotte, M„ Kotlyakov, V. M., Legrand, M., Lipenkov, V. Y., Lorius, C., Pepin, L., Ritz, C., Saltzman, E., and Stievenard, M. (1999). Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 439-436.

Pinto, J. P. and Khalil, M. A. K. (1991). The stability of tropospheric Oh during Ice Ages, inter-glacial epochs and modern times. Tellus Series B- Chem. Phys. Meteorol. 43, 347-352.

Prentice, I. C„ Jolly, D., and BIOME 6000 participants. (2000). Mid-Holocene and glacial-maximum vegetation geography of the northern continents and Africa. J. Biogeography. 27, 507-519.

Raynaud, D„ Barnola, J-M., Chappellaz, J., Blunier, T„ Indermuhle, A., and Stauffer, B. (2000). The ice record of greenhouse gases: a view in the context of future changes. Quat. Sci. Rev. 19,9-17.

Raynaud, D., Chappellaz, J., and Blunier, T. (1998). Ice core record of atmospheric methane changes: relevance to climatic changes and possible gas hydrate sources. Geol. Soc. Special Publ. 137,327-331.

Schellnhuber, H. J. (1999). 'Earth system' analysis and the second Copernican revolution. Nature 402, C19-C23.

Shackleton, N. J.> (1977). Carbon-13 in Uvigerina: tropical rainforest history and the equatorial Pacific carbonate dissolution cycles. In "The Fate of Fossil Fuel C02 in the Oceans." Anderson, N. R. and Malahoff.A. (Eds.). Marine Sci. 401-427.

Smith, H. J., Fischer, H„ Wahlen, M„ Mastroianni, D„ and Deck, B. (1999). Dual modes of the carbon cycle since the Last Glacial Maximum. Nature 400,248-250.

Stauffer, B„ Blunier, T„ and Dallenbach, A. (1998). Atmospheric CO, concentration and millennial-scale climate change during the last glacial period. Nature 392, 59-62.

Steffensen, J. P. (1997). The size distribution of microparticles from selected segments of the Greenland Ice Core Project ice core representing different climatic periods./. Geopliys. Res. 102, 26755-26763.

Stephens, B. B., Keeling, R. F. (2000). The influence of of Antarctic sea ice on glacial-interglacial C02 variations, Nature 404, 171-174.

Tegen, I., and Fung, I. (1995). Contribution to the atmospheric mineral aerosol load from land surface modification. }. Geopliys. Res. 100, 18707-18726.

Thompson, A. M„ Chapellaz, J. A., Fung, I. Y„ and Kuscera, T. L. (1993). The atmospheric CH4 increase since the last glacial maximum: 2. Interaction with oxidents. Tellus 45B, 242-257.

Thompson, L. G. and Mosley-Thompson, E. (1981). Microparticle concentration variations linked with climatic change: evidence from polar ice. Science 212, 812-815.

0 0

Post a comment