The complexity of radiative transfer in a purely gaseous atmosphere resurfaced under the Intercomparison of Radiation Codes in Climate Models (ICRCCM), [ Ellingson et al. 1991] and the special issue of J. Geophys. Res., 96,(D5) [1991] in which this paper is contained is largely devoted to this topic). Since that time a number of observational studies, in part supported under the auspices of the U.S. Department of Energy's Atmospheric Radiation Measurement (ARM) Program, have been carried out to help improve our understanding of absorption by gases in the atmosphere and to validate our present compilations of molecular line absorption spectra [e.g. Ellingson et al., 1993; Dutton, 1993; English et al., 1994]. Important contributions, not so much for their advancement of new knowledge but more for placing our existing understanding on a firmer basis, are the data-bases of spectroscopic parameters of atmospheric gases, such as the High Resolution Transmission Model (HITRAN) and GEISA (Gestion et Etude des Informations Spectroscopiques Atmospheriques) data-bases of Rothman et al., [1992] and Husson et al. [1992] respectively and the microwave line absorption data of Liebe and Hufford [1993]. These data bases continually undergo constant revision.
A persistent and troublesome problem of molecular absorption arises
through the need to define the absorption by water vapor between
absorption lines in the so-called absorption continuum. The basic
problem is that a rigorous theoretical treatment of continuum
absorption does not exist although attempts to address the
theoretical characteristics of the continuum based on line absorption
effects are described in Ma and Tipping [1992a and b].
Despite the lack of a theoretical foundation, we have been able to
consolidate our understanding, albeit more or less empirically, in
terms of two component contributions to the continuum: the self
broadened continuum (i.e., the effects on absorption when two water
molecules collide) and a foreign broadened component (effects of
collision between water and other molecules, chiefly O
and
N
). This recognition implies that the continuum derives from
overlapping line wing effects rather than from water polymers. A
`semi-empirical' model of the continuum introduced by Clough and
colleagues [e.g. Clough et al., 1992] and further discussed
by Clough [1994] has provided acceptable results when tuned
against high resolution spectra measurements available from recent
field programs [ Kisby et al., 1992; Ellingson et al.,
1993, Rudman et al., 1994].
We have known for some time that an incorrect treatment of the continuum absorption presents a number of significant problems. Continuum absorption provides significant increases in the opacity in spectral regions that are otherwise relatively transparent---in all atmospheric windows that exist from the microwave to the visible. The recent ICRCCM comparisons [e.g. Ellingson et al, 1991] indicate that the source of the largest uncertainty between models (of all types and complexities) of the clear-sky longwave flux lies in how these models treat the continuum. This absorption is important from the point of view of radiative transfer since radiant energy exchanges between clouds, the surface and the atmosphere are largest in these spectral regions. It is also important to those remote sensing problems that seek to exploit measurements in these windows [e.g. Taylor, 1992].