Accurate parameterizations of photolysis rate coefficients (J-values) are required to represent properly the photochemical transformations, and have direct effects on calculated concentrations of key species such as O3, OH, and NOx.  The calculation of the J-values is complicated by the natural variability of the atmospheric radiation field, and by the need to have detailed spectroscopic data (cross sections,  quantum yields) for the many photo-reactions of interest.  All of these quantites are wavelength-dependent (UV and visible range), in some cases strongly so.

Most current models use one of two approaches:

(1) A look-up table of J values is calculated, off-line, as a function of  the major controlling parameters such as solar zenith angle, altitude,  overhead ozone column, and surface albedo; and

(2) On-line calculations of the radiation field as a function of wavelength,  for the specified local optical environment, e.g. clouds and aerosols, in addition to the variables listed under item (1).

On-line calculations are in principle preferrable because the computed  radiation field is more consistent with the local meteorology (e.g. clouds). However, radiative transfer calculations can add substantial computational cost. Much of this computational cost arises from the need to compute the actinic flux at relatively small wavelength intervals in the spectral range 280-700 nm.  This high spectral  resolution is required because of the often strong spectral variations in molecular  cross sections, extraterresterrestrial flux, and atmosperic transmission (esp. in the 280-330 nm range which is sensitive to absorption  by stratospheric O3).

Several recent studies suggest that the computational time for on-line calculations can be reduced by pre-computing some quantities at high spectral resolution, while computing atmospheric transmission at only a few wavelengths.  The status of such models will be reviewed.