Atmospheric CO₂ concentration has varied on all time-scales during the Earth’s history (Figure 3.2). There is evidence for very high CO₂ concentrations (>3,000 ppm) between 600 and 400 Myr BP and between 200 and 150 Myr BP (Figure 3.2f). On long time-scales, atmospheric CO₂ content is determined by the balance among geochemical processes including organic carbon burial in sediments, silicate rock weathering, and vulcanism (Berner, 1993, 1997). In particular, terrestrial vegetation has enhanced the rate of silicate weathering, which consumes CO₂ while releasing base cations that end up in the ocean. Subsequent deep-sea burial of Ca and Mg (as carbonates, for example CaCO₃) in the shells of marine organisms removes CO₂. The net effect of slight imbalances in the carbon cycle over tens to hundreds of millions of years has been to reduce atmospheric CO₂. The rates of these processes are extremely slow, hence they are of limited relevance to the atmospheric CO₂ response to emissions over the next hundred years.

Figure 3.3: Fossil fuel emissions and the rate of increase of CO2 concentration in the atmosphere. The annual atmospheric increase is the measured increase during a calendar year. The monthly atmospheric increases have been filtered to remove the seasonal cycle. Vertical arrows denote El Niño events. A horizontal line defines the extended El Niño of 1991 to 1994. Atmospheric data are from Keeling and Whorf (2000), fossil fuel emissions data are from Marland et al. (2000) and British Petroleum (2000), see explanations in text.

It is pertinent, however, that photosynthesis evolved at a time when O₂ concentrations were far less than at present. O₂ has accumulated in the atmosphere over geological time because photosynthesis results in the burial of reduced chemical species: pyrite (FeS₂) derived from sulphur-reducing bacteria, and organic carbon. This accumulation has consequences for terrestrial and marine ecosystems today. Primary production is carbon limited in terrestrial ecosystems in part because of (geologically speaking) low CO₂ concentrations, and in part because Rubisco (the enzyme that fixes CO₂ in all plants) also has an affinity for O₂ that reduces its efficiency in photosynthesis (see Section 3.2.2.4). Primary production is iron limited in some marine ecosystems mainly because of the extreme insolubility of Fe(III), the predominant form of iron in the present, O₂-rich environment. These difficulties faced by contemporary organisms represent a legacy of earlier evolution under very different biogeochemical conditions.

In more recent times, atmospheric CO₂ concentration continued to fall after about 60 Myr BP and there is geochemical evidence that concentrations were <300 ppm by about 20 Myr BP (Pagani et al., 1999a; Pearson and Palmer, 1999, 2000; Figure 3.2e). Low CO₂ concentrations may have been the stimulus that favoured the evolution of C₄ plants, which increased greatly in abundance between 7 and 5 Myr BP (Cerling et al., 1993, 1997; Pagani et al., 1999b). Although contemporary CO₂ concentrations were exceeded during earlier geological epochs, they are likely higher now than at any time during the past 20 million years.

The purity of Antarctic ice allows the CO₂ concentration in trapped air bubbles to be accurately measured (Tschumi and Stauffer, 2000). The CO₂ record from the Vostok ice core is the best available for the glacial/inter-glacial time-scale and covers the past four glacial/inter-glacial cycles (420 kyr) with a resolution of 1 to 2 kyr (Petit et al., 1999; Fischer et al., 1999). The general pattern is clear (Figure 3.2d): atmospheric CO₂ has been low (but 180 ppm) during glacial periods, and higher (but 300 ppm) during interglacials. Natural processes during the glacial-interglacial cycles have maintained CO₂ concentrations within these bounds, despite considerable variability on multi-millenial time-scales. The present CO₂ concentration is higher than at any time during the 420 kyr period covered by the Vostok record.

The terrestrial biosphere stores 300 to 700 Pg more carbon during interglacial periods than during glacial periods, based on a widely accepted interpretation of the ¹³C record in deep-sea sediments (Shackleton, 1977; Bird et al., 1994; Crowley, 1995). Terrestrial modelling studies (e.g., Friedlingstein et al., 1995b; Peng et al., 1998) have reached the same conclusion. Thus, the terrestrial biosphere does not cause the difference in atmospheric CO₂ between glacial and interglacial periods. The cause must lie in the ocean, and indeed the amount of atmospheric change to be accounted for must be augmented to account for a fraction of the carbon transferred between the land and ocean. The mechanism remains controversial (see Box 3.4). In part this is because a variety of processes that could be effective in altering CO₂ levels on a century time-scale can be largely cancelled on multi-millenial time-scales by changes in CaCO₃ sedimentation or dissolution, as discussed in Section 3.2.3.1.

Orbital variations (Berger, 1978) are the pacemaker of climate change on multi-millenial time-scales (Hays et al., 1976). Atmospheric CO₂ is one of many Earth system variables that show the characteristic “Milankovitch” periodicities, and has been implicated as a key factor in locking natural climate changes to the 100 kyr eccentricity cycle (Shackleton, 2000). Whatever the mechanisms involved, lags of up to 2,000 to 4,000 years in the drawdown of CO₂ at the start of glacial periods suggests that the low CO₂ concentrations during glacial periods amplify the climate change but do not initiate glaciations (Lorius and Oeschger, 1994; Fischer et al., 1999). Once established, the low CO₂ concentration is likely to have enhanced global cooling (Hewitt and Mitchell, 1997). During the last deglaciation, rising CO₂ paralleled Southern Hemisphere warming and was ahead of Northern Hemisphere warming (Chapter 2).

During glacial periods, the atmospheric CO₂ concentration does not track the “fast” changes in climate (e.g., decade to century scale warming events) associated with Dansgaard-Oeschger events, although there are CO₂ fluctuations of up to 20 ppm associated with the longer-lived events (Stauffer et al., 1998; Indermühle et al., 2000) (see Chapter 2 for explanations of these terms). During the last deglaciation, atmospheric CO₂ concentration continued to increase, by about 12 ppm, through the Younger Dryas cold reversal (12.7 to 11.6 kyr BP) seen in Northern Hemisphere palaeoclimate records (Fischer et al., 1999; Smith et al., 1999). Palaeo-oceanographic evidence shows that the Younger Dryas event was marked by a prolonged shut-down of the thermohaline circulation, which is likely to have been triggered by the release of melt water into the North Atlantic. Similar behaviour, with a slight rise in CO₂ accompanying a major Northern Hemisphere cooling and shutdown of North Atlantic Deep Water production, has been produced in a coupled atmosphere-ocean model (Marchal et al., 1998). The observed CO₂ rise during the Younger Dryas period was modest, suggesting that atmospheric CO₂ has, under natural conditions, been well buffered against abrupt changes in climate, including thermohaline collapse. This buffering is a direct consequence of the large reservoir of DIC in the ocean.

IPCC Homepage