IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis Absorbing Aerosol

For absorbing aerosols, the simple ideas of a linear forcing-response relationship and efficacy can break down (Hansen et al., 1997; Cook and Highwood, 2004; Feichter et al., 2004; Roberts and Jones, 2004; Hansen et al., 2005; Penner et al., 2007). Aerosols within a particular range of single scattering albedos have negative RFs but induce a global mean warming, that is, the efficacy can be negative. The surface albedo and height of the aerosol layer relative to the cloud also affects this relationship (Section 7.5; Penner et al., 2003; Cook and Highwood, 2004; Feichter et al., 2004; Johnson et al., 2004; Roberts and Jones, 2004; Hansen et al., 2005). Studies that increase BC in the planetary boundary layer find efficacies much larger than 1.0 (Cook and Highwood, 2004; Roberts and Jones, 2004; Hansen et al., 2005). These studies also find that efficacies are considerably smaller than 1.0 when BC aerosol is changed above the boundary layer. These changes in efficacy are at least partly attributable to a semi-direct effect whereby absorbing aerosol modifies the background temperature profile and tropospheric cloud (see Section 7.5). Another possible feedback mechanism is the modification of snow albedo by BC aerosol (Hansen and Nazarenko, 2004; Hansen et al., 2005); however, this report does not classify this as part of the response, but rather as a separate RF (see Section 2.5.4 and Most GCMs likely have some representation of the semi-direct effect (Cook and Highwood, 2004) but its magnitude is very uncertain (see Section 7.5) and dependent on aspects of cloud parametrizations within GCMs (Johnson, 2005). Two studies using realistic vertical and horizontal distributions of BC find that overall the efficacy is around 0.7 (Hansen et al., 2005; Lohmann and Feichter, 2005). However, Hansen et al. (2005) acknowledge that they may have underestimated BC within the boundary layer and another study with realistic vertical distribution of BC changes finds an efficacy of 1.3 (Sokolov, 2006). Further, Penner et al. (2007) also modelled BC changes and found efficacies very much larger and very much smaller than 1.0 for biomass and fossil fuel carbon, respectively (Hansen et al. (2005) found similar efficacies for biomass and fossil fuel carbon). In summary there is no consensus as to BC efficacy and this may represent problems with the stratospherically adjusted definition of RF (see Section 2.8.3). Other Forcing Agents

Efficacies for some other effects have been evaluated by one or two modelling groups. Hansen et al. (2005) found that land use albedo RF had an efficacy of roughly 1.0, while the BC-snow albedo RF had an efficacy of 1.7. Ponater et al. (2005) found an efficacy of 0.6 for contrail RF and this agrees with a suggestion from Hansen et al. (2005) that high-cloud changes should have smaller efficacies. The results of Hansen et al. (2005) and Forster and Shine (1999) suggest that stratospheric water vapour efficacies are roughly one.