Figure 8.20: (a) Radiative forcing (Wm-2) since 1979 due to changes in stratospheric aerosols, ozone, greenhouse gases and solar irradiance.
(b) to (d) Observed global annual mean surface, tropospheric and stratospheric temperature changes and GISS GCM simulations as the successive radiative forcings are cumulatively added one by one. An additional constant 0.65 Wm-2 forcing was included in all their simulations, representing the estimated disequilibrium forcing for 1979. The base period defining the zero mean observed temperature was 1979 to 1990 for the surface and the troposphere and 1984 to 1990 for the stratosphere. Taken from Hansen et al. (1997).
Coupled model control runs have a general tendency to underestimate the variability found in both the instrumental and palaeo-proxy record, especially over the oceans (the converse is true over land when compared to the instrumental record). With the exception of Santer et al. (2000), none of the aforementioned simulations examined the potential climatic effects of stratospheric aerosols associated with volcanic emissions. On the longer time-scales, none of these studies included variability in solar forcing. Crowley (2000) has estimated that changes in solar irradiance and volcanism may account for between 41 and 64% of pre-industrial, decadal-scale surface air temperature variations.
Cubasch et al. (1997) demonstrated that when solar variations were included in the ECHAM3/LSG model, their simulation from 1700 through the 20th century showed enhanced low-frequency variability associated with variability in solar irradiance (see also Lean and Rind, 1998). The implication of climate change detection and attribution studies (Chapter 12; Hegerl et al., 1997; Tett et al., 1999) for the reproduction of 20th century climate by AOGCMs, is that changes in solar irradiance may be important to include if one wants to reproduce the warming in the early part of the century. As noted earlier, it is conceivable that this early warming may also be solely a result of natural internal climate variability (Delworth and Knutson, 2000). Energy balance/upwelling diffusion climate models and Earth system models of intermediate complexity, when forced with volcanic and solar variations for the past 400 years, capture the cooling associated with the Little Ice Age (Betrand et al., 1999; Crowley and Kim, 1999; Free and Robock, 1999), although they are not capable of assessing regional climatic anomalies associated with local feedbacks or changes in atmospheric dynamics. These same models produce the observed warming of the past century when additionally forced with anthropogenic greenhouse gases and aerosols.
Hansen et al. (1997) conducted a systematic study of the climate system response to various radiative forcings for the period 1979 to 1995 (over which period Nimbus 7 satellite data were available) using the Goddard Institute for Space Studies (GISS) AGCM coupled to a mixed-layer ocean model. A series of ensemble simulations, with each ensemble consisting of five experiments, were conducted by cumulatively adding, one-by-one, radiative forcing effects due to stratospheric aerosols (associated with volcanic emissions), decreases in upper level ozone, increases in anthropogenic greenhouse gases, and changes in solar irradiance (Figure 8.20). While changes in tropospheric aerosols, either via direct or indirect effects, were not included in their calculations, over the short record a reasonable agreement with observations was obtained. Internal climate variability (e.g., the warming associated with the El Niño of 1983 and the cooling associated with the La Niña of 1989 in Figure 8.20b,c) is not well resolved in the model. These experiments point out that while solar irradiance changes caused minimal changes over the period (consistent with the analysis of Hegerl et al., 1997; Tett et al., 1999), stratospheric aerosols associated with volcanic emissions and changes in upper level ozone are important components which need to be included if one hopes to accurately reproduce the variations in the instrumental record.
Changing land-use patterns affect climate in several ways (see Chapter 6, Section 6.11). While the impact of land-use changes on radiative forcing is small (e.g., Hansen et al., 1998) changes in roughness, soil properties and other quantities may be important (see Chapter 7, Section 7.4). Brovkin et al. (1999) demonstrated that CLIMBER (Climate Biosphere Model) was able to capture the long-term trends and slow modulation of the Mann et al. (1998) reconstruction of Northern Hemisphere temperatures over the past 300 years provided changes in land-use patterns (as well as changes in atmospheric CO2 and changing solar forcing) were taken into account. Some recent model results suggest that land cover changes during this century may have caused regional scale warming (Chase et al., 2000; Zhao et al., 2001) but this remains to be examined with a range of climate models.
Other reports in this collection