IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis

6.6.4 Consistency Between Temperature, Greenhouse Gas and Forcing Records; and Compatibility of Coupled Carbon Cycle-Climate Models with the Proxy Records

It is difficult to constrain the climate sensitivity from the proxy records of the last millennium (see Chapter 9). As noted above, the evidence for hemispheric temperature change as interpreted from the different proxy records, and for atmospheric trace greenhouse gases, inferred solar forcing and reconstructed volcanic forcing, is to varying degrees uncertain. The available temperature reconstructions suggest that decadally averaged NH temperatures varied within 1°C or less during the two millennia preceding the 20th century (Figure 6.10), but the magnitude of the reconstructed low-frequency variations differs by up to about a factor of two for different reconstructions. The reconstructions of natural forcings (solar and volcanic) are uncertain for this period. If they produced substantial negative energy balances (reduced solar, increased volcanic activity), then low-to-medium estimates of climate sensitivity are compatible with the reconstructed temperature variations (Figure 6.10); however, if solar and volcanic forcing varied only weakly, then moderate-to-high climate sensitivity would be consistent with the temperature reconstructions, especially those showing larger cooling (see also Chapter 9), assuming that the sensitivity of the climate system to solar irradiance changes and explosive volcanism is not different than the sensitivity to changes in greenhouse gases or other forcing agents.

The greenhouse gas record provides indirect evidence for a limited range of low-frequency hemispheric-scale climate variations over the last two millennia prior to the industrial period (AD 1–1750). The greenhouse gas histories of CO2, CH4 and N2O show only small changes over this time period (MacFarling Meure et al., 2006; Figure 6.4), although there is evidence from the ice core record (Figures 6.3 and 6.7), as well as from models, that greenhouse gas concentrations react sensitively to climatic changes.

The sensitivity of atmospheric CO2 to climatic changes as simulated by coupled carbon cycle-climate models is broadly consistent with the ice core CO2 record and the amplitudes of the pre-industrial, decadal-scale NH temperature changes in the proxy-based reconstructions (Joos and Prentice, 2004). The CO2 climate sensitivity can be formally defined as the change in atmospheric CO2 relative to a nominal change in NH temperature in units of ppm per °C. Its strength depends on several factors, including the change in solubility of CO2 in seawater, and the responses of productivity and heterotrophic respiration on land to temperature and precipitation (see Section 7.3). The sensitivity was estimated for modest (NH temperature change less than about 1°C) temperature variations from simulations with the Bern Carbon Cycle-Climate model driven with solar and volcanic forcing over the last millennium (Gerber et al., 2003) and from simulations with the range of models participating in the Coupled Carbon Cycle-Climate Model Intercomparison Project (C4MIP) over the industrial period (Friedlingstein et al., 2006). The range of the CO2 climate sensitivity is 4 to 16 ppm per °C for the 10 models participating in the C4MIP intercomparison (evaluated as the difference in atmospheric CO2 for the 1990 decade between a simulation with, and without, climate change, divided by the increase in NH temperature from the 1860 decade to the 1990 decade). This is comparable to a range of 10 to 17 ppm per °C obtained for CO2 variations in the range of 6 to 10 ppm (Etheridge et al., 1996; Siegenthaler et al., 2005b) and the illustrative assumption that decadally averaged NH temperature varied within 0.6°C.