perature (SST) forcing, which is the change in TOA radiative forcing computed in a global model with fixed sea surface temperatures but letting land and atmospheric temperatures relax to the new equilibrium. This relaxation is relatively rapid (on the order of years); hence the calculation in a GCM is computationally expedient. One resolves in this manner the short-term components of the climate response, such as hydrological perturbations associated with changes in lapse rate. Of particular interest, this approach allows calculation of a meaningful radiative forcing from the indirect or semidirect aerosol effects. Hansen et al. (2002) show that there is good agreement between the stratospheric adjusted radiative forcing and the fixed SST forcing for a range of climate forcing factors (e.g., 2 × CO₂, stratospheric aerosols) and that for changes in ozone, more reasonable forcings result from the fixed SST simulations.

Shine et al. (2003) extended the fixed SST approach to what they call “adjusted troposphere and stratosphere forcing.” Shine et al. not only fix sea surface temperatures, but also fix land surface temperatures because temperatures over land and ocean are related. Therefore, it is more consistent to fix surface temperatures globally. Using a global climate model they show that the adjusted troposphere and stratosphere radiative forcing is consistent with the stratospheric adjusted forcing for more uniform forcings such as doubling CO₂ and solar constant changes. They also show that for forcings due to absorbing aerosols, their newly defined forcing is more meaningful than the stratospheric adjusted forcing, in that the climate sensitivity parameter is largely independent of how the absorbing aerosols are vertically distributed, unlike the standard stratospheric adjusted approach.

These two studies are important contributions to the debate on radiative forcing, but the approach is subject to most of the limitations associated with the traditional radiative forcing calculation. Also, forcings calculated in this manner are not as easy to compute as conventional radiative forcings, nor are they as comparable among different GCMs because of differences in model dynamics and hydrology.

The TOA radiative forcing might not be directly related to surface temperature if a forcing agent changes the vertical distribution of heating in the atmosphere. Well-known examples of such cases are the direct radiative forcing of black carbon (BC) and other absorbing aerosols and the changes in latent and sensible heat fluxes due to land-use modifications. For example, BC causes an increase in atmospheric heating, accompanied by a decrease in solar heating of the surface. For average cloudiness, Indian Ocean Experiment (INDOEX) data reveal that the TOA direct forcing when BC is present can be close to zero, while the surface forcing can be on