We investigate the role of spatial and temporal resolution for estimation of radiative forcing due to SF6 and a range of halocarbons as well as CO2. A broadband model, which is used in the calculations, is described. Some comparative calculations have also been performed with a line-by-line model. The most detailed horizontal resolution used is 2.5¿¿2.5¿ in latitude and longitude. A variety of resolutions up to global averages are investigated. The effects of variations on diurnal, monthly, and seasonal scales are also studied. Spatial and temporal variation in the radiative forcing due to variations in temperature, humidity, and cloudiness has been taken into account on the basis of observed data. Inaccuracies due to temporal variations are small in all cases (up to about 1%). Deviations in forcings due to spatial averaging are also small (less than 1%) as long as latitudinal variations are resolved, but significant inaccuracies are introduced when global averaged conditions are assumed. The forcing due to CO2 responds somewhat differently to spatial averaging compared to SF6 and the halocarbons, so global warming potential (GWP) values for SF6 and halocarbons with CO2 as a reference gas are less accurate. Resolution of latitudinal variations in input parameters is shown to increase the accuracy of the GWP values for SF6 and the halocarbons. The choice of tropopause level, where radiative forcing is determined, is shown to be crucial, with differences up to 10% in the global average radiative forcing for different assumptions. ¿ 1997 American Geophysical Union