The effect of the Curtis‐Godson approximation on the accuracy of radiative heating‐rate calculations

Some exact heating rates are calculated for the water vapour pure rotation band, the 9·6 μ ozone band, and the 15 μ carbon‐dioxide bands, using the appropriate spectral models with Lorentz line shape to take into account line overlap and the linear dependence of line width on pressure. These ‘exact’ heating rates are then compared with those calculated under the same conditions but using the Curtis‐Godson approximation (CG). It is shown that the errors introduced by CG are always less than a few per cent for all except the 9·6 μ ozone band. The reasons for CG failure are very complex, but tentative conditions for its validity are suggested. This leads to a delineation of regions of the spectrum for which more accurate methods may need to be used.