The well-mixed greenhouse gases have lifetimes long enough to be relatively homogeneously mixed in the troposphere. In contrast, O3 (Section 6.5) and the NMHCs (Section 6.6) are gases with relatively short lifetimes and are therefore not homo-geneously distributed in the troposphere.
Spectroscopic data on the gaseous species have been improved with successive versions of the HITRAN (Rothman et al., 1992, 1998) and GEISA databases (Jacquinet-Husson et al., 1999). Pinnock and Shine (1998) investigated the effect of the additional hundred thousands of new lines in the 1996 edition of the HITRAN database (relative to the 1986 and the 1992 editions) on the infrared radiative forcing due to CO2, CH4, N2O and O3. They found a rather small effect due to the additional lines, less than a 5% effect for the radiative forcing of the cited gases and less than 1.5% for a doubling of CO2. For the chlorofluorocarbons (CFCs) and their replacements, the uncertainties in the spectroscopic data are much larger than for CO2, CH4, N2O and O3, and differ more among the various laboratory studies. Christidis et al. (1997) found a range of 20% between ten different spectroscopic studies of CFC-11. Ballard et al. (2000) performed an intercomparison of laboratory data from five groups and found the range in the measured absorption cross-section of HCFC-22 to be about 10%.
Several previous studies of radiative forcing due to well-mixed greenhouse gases have been performed using single, mostly global mean, vertical profiles. Myhre and Stordal (1997) investigated the effects of spatial and temporal averaging on the globally and annually averaged radiative forcing due to the well-mixed greenhouse gases. The use of a single global mean vertical profile to represent the global domain, instead of the more rigorous latitudinally varying profiles, can lead to errors of about 5 to 10%; errors arising due to the temporal averaging process are much less (~1%). Freckleton et al. (1998) found similar effects and suggested three vertical profiles which could represent global atmospheric conditions satisfactorily in radiative transfer calculations. In the above two studies as well as in Forster et al. (1997), it is the dependence of the radiative forcing on the tropopause height and thereby also the vertical temperature profile, that constitutes the main reason for the need of a latitudinal resolution in radiative forcing calculations. The radiative forcing due to halocarbons depends on the tropopause height more than is the case for CO2 (Forster et al., 1997; Myhre and Stordal, 1997).
Not all greenhouse gases are well mixed vertically and horizontally in the troposphere. Freckleton et al. (1998) have investigated the effects of inhomogeneities in the concentrations of the greenhouse gases on the radiative forcing. For CH4 (a well-mixed greenhouse gas), the assumption that it is well-mixed horizontally in the troposphere introduces an error much less than 1% relative to a calculation in which a chemistry-transport model predicted distribution of CH4 was used. For most halocarbons, and to a lesser extent for CH4 and N2O, the mixing ratio decays with altitude in the stratosphere. For CH4 and N2O, this implies a reduction in the radiative forcing of up to about 3% (Freckleton et al., 1998; Myhre et al., 1998b). For most halocarbons, this implies a reduction in the radiative forcing up to about 10% (Christidis et al., 1997; Hansen et al., 1997a; Minschwaner et al., 1998; Myhre et al., 1998b) while it is found to be up to 40% for a short-lived component found in Jain et al. (2000).
Trapping of the long-wave radiation due to the presence of clouds reduces the radiative forcing of the greenhouse gases compared to the clear-sky forcing. However, the magnitude of the effect due to clouds varies for different greenhouse gases. Relative to clear skies, clouds reduce the global mean radiative forcing due to CO2 by about 15% (Pinnock et al., 1995; Myhre and Stordal, 1997), that due to CH4 and N2O is reduced by about 20% (derived from Myhre et al., 1998b), and that due to the halocarbons is reduced by up to 30% (Pinnock et al., 1995; Christidis et al., 1997; Myhre et al., 1998b).
The effect of stratospheric temperature adjustment also differs between the various well-mixed greenhouse gases, owing to different gas optical depths, spectral overlap with other gases, and the vertical profiles in the stratosphere. The stratospheric temperature adjustment reduces the radiative forcing due to CO2 by about 15% (Hansen et al., 1997a). CH4 and N2O estimates are slightly modified by the stratospheric temperature adjustment, whereas the radiative forcing due to halocarbons can increase by up to 10% depending on the spectral overlap with O3 (IPCC, 1994).
Radiative transfer calculations are performed with different types of radiative transfer schemes ranging from line-by-line models to band models (IPCC, 1994). Evans and Puckrin (1999) have performed surface measurements of downward spectral radiances which reveal the optical characteristics of individual greenhouse gases. These measurements are compared with line-by-line calculations. The agreement between the surface measurements and the line-by-line model is within 10% for the most important of the greenhouse gases: CO2, CH4, N2O, CFC-11 and CFC-12. This is not a direct test of the irradiance change at the tropopause and thus of the radiative forcing, but the good agreement does offer verification of fundamental radiative transfer knowledge as represented by the line-by-line (LBL) model. This aspect concerning the LBL calculation is reassuring as several radiative forcing determinations which employ coarser spectral resolution models use the LBL as a benchmark tool (Freckleton et al., 1996; Christidis et al., 1997; Minschwaner et al., 1998; Myhre et al., 1998b; Shira et al., 2001). Satellite observations can also be useful in estimates of radiative forcing and in the intercomparison of radiative transfer codes (Chazette et al., 1998).
Other reports in this collection