Main

As Parker mentions, the only available direct measurements show a variation of just 0.15% in solar irradiance S over one solar magnetic cycle of 11 years. Greater variations on longer timescales are possible, but other reports5,6 give figures that are less than the 0.5% that Parker quotes. Even ΔS/S≈0.5% implies global mean radiative forcing: (ΔS/4)(1−α)≈1 W m−2, where α≈0.3 is Earth's albedo. For comparison, the Intergovernmental Panel on Climate Change (IPCC) estimates the change in global mean radiative forcing from 1861 to 1990 as 2.0 to 2.8 W m−2 from greenhouse gases, −0.2 to −2.3 W m−2 from aerosols, and 0.1 to 0.5 W m−2 from solar variability7,8. There has been much speculation about additional solar effects but, with regard to Parker's remark about cosmic-ray effects on clouds, we note that the cited work9 claims that this effect would only amplify solar radiative forcing to less than 2 W m−2.

We are surprised by Parker's suggestion that solar brightening is responsible not only for the observed increase in twentieth-century surface temperatures, but also for increased carbon dioxide in “the same way that a carbonated drink expels most of its CO2 if warm”. Given the observed surface temperature increase of about 1 K, a simple calculation10 reveals that this effect would increase equilibrium CO2 by less than 5%, which is much less than the 30% increase recorded since the beginning of the industrial revolution. Records of 13C and 14C isotopes relative to 12C also support the idea that changes in atmospheric CO2 are primarily due to the burning of fossil fuel8.

Although solar effects on this century's climate may not be negligible, quantitative considerations imply that they are small relative to the anthropogenic release of greenhouse gases, primarily carbon dioxide. Figure 8.4 of the IPCC 1995 assessment report3 (also shown in refs 7, 11) makes that point clearly, even though it assumes solar variability near the upper end of its uncertainty range7.

On longer timescales, it is not a “historical fact” that climate “responds to variations of the Sun's magnetic activity, with substantial warming and cooling with the rise and fall of activity over the centuries”. Several recent palaeoclimate compilations12,13 indicate only that weak responses to putative solar variations have occurred over the past 500–1,000 years. The data suggest that global mean temperatures during the 1990s may have been the warmest of the millennium. Together with climate model simulations (which, contrary to Parker, quantify the combined effects of cloud, winds, ocean currents, solar effects and anthropogenic effects14,15,16, however imperfectly), the data indicate that there is probably a strong anthropogenic component to twentieth-century global warming.