Solar variability does not explain late-20th-century warming

Abstract

The hypothesis of Nicola Scafetta and Bruce West (see their Opinion piece, PHYSICS TODAY, March 2008, page 50), that most of the observed global warming trend since 1950 is due to variations in total solar irradiance (TSI), is at odds with observations and theory. They argue that near-surface temperatures are strongly influenced by both the well-known 11-year cycle in TSI and shorter-time-scale TSI variations. As for the 11-year cycle in TSI, satellite measurements of TSI since 1979 show variations that are very small (about 0.1%) and cyclical (see figure 1). If Earth’s surface temperature were highly sensitive to those variations, the 11-year cycle in TSI would produce an 11-year cycle in observed surface temperatures. The filtered and smoothed temperature record presented by Scafetta and West appears to show a significant 11-year cycle. However, an exhaustive analysis of the latter half of the 20th century by Gerald North and his collaborators found a temperature cycle with an amplitude of no more than a few hundredths of a degree Celsius, with a significance level of only 25%.1 Even that small signal, which is roughly consistent with the expected signal that North and colleagues estimated using an energy-balance model, may be in part an artifact of fortuitous spacing of major volcanic eruptions (see the letter by Peter Foukal, PHYSICS TODAY, October 2008, page 10). Hence, the historical surface temperature record does not support the hypothesis of strong temperature sensi tivity to solar irradiance changes. Additionally, that hypothesis predicts greater-than-observed temperature variations during prehistoric times (see the letter by Diedrich Schmidt, PHYSICS TODAY, October 2008, page 10). For example, strong sensitivity to solar variations would have resulted in a much colder climate during the Maunder Minimum (a period of reduced solar activity during the 17th century), when the TSI is thought to have been less than today by anywhere from 0.2% 2 to 1.2%.3 In any case, as is evident from figure 1, TSI variations since 1979 are cyclical, with no discernible multidecadal trend. Hence, the variations cannot explain the observed warming trend since 1979, which has continued over more than two 11-year cycles (see figure 2). That conclusion is supported by thorough analyses by Foukal and his collaborators4 and by others. Is it possible that TSI variations, although having little direct effect on Earth’s energy balance, nonetheless control climate by somehow regulating more energetic processes? (An analogy would be a transistor.) The “complexity matching” argument of Scafetta and West is along those lines. Their argument seems unlikely to be correct, however, since observations by Sydney Levitus and his collaborators show that the climate system as a whole, particularly the ocean, has warmed along with the atmosphere; 5 the energy content of the climate system increased between about 1955 and about 1995 by about 2 × 1023 J, equivalent to an energy imbalance at the top of the atmosphere of 0.3 W/m2. Because incoming solar energy—that is, TSI—has not changed, the imbalance must result from increased absorption of outgoing energy, such as by increased greenhouse gases. Indeed, satellite measurements by John Harries and coworkers confirm that absorption of outgoing terrestrial radiation by several greenhouse gases became stronger 6 between 1970 and 1997. An additional failing of Scafetta and West’s hypothesis is that it requires greenhouse gases to have very little ef1376