Limitations of the equivalent CO₂ approximation in climate change simulations

The "equivalent CO₂" approximation often used to simulate the climatic effects of a suite of trace greenhouse gases is investigated using a recent version of the NCAR Community Climate Model. We performed present-day and preindustrial equilibrium climate simulations. The climate sensitivity is lower by ∼20% in the equivalent CO₂ case compared to the control case in which the individual trace gases were treated explicitly. This is reflected in similar percentage differences in global- and annual-mean surface temperature, precipitation, precipitable water, and sea ice. The temperature changes are also different regionally in the tropical and subtropical troposphere and in the stratosphere. This difference in climate sensitivity originates from differences in the spatial pattern of radiative forcing. The equivalent CO₂ forcing pattern differs from the control case forcing pattern for several reasons, but the dependence on temperature of the Planck emission spectrum appears to be fundamentally most important. The primary absorption bands of CH₄ and N₂O are found at wavelengths more sensitive to the temperature-related shift of the wavelength of maximum emission than the absorption bands of CO₂. This leads to stronger spatial variations in absorption by trace gases than by CO₂. We conclude that because of differences in the pattern of radiative forcing, the equilibrium response of global climate to increases in trace gases is larger than the response to an "equivalent" increase in CO₂, and the patterns of response are also different.