Mean radiation fluxes in the near-IR spectral range: Algorithms

spectral interval in question into Nt subintervals according to the spectral resolution specified. The mean spectral fluxes are then calculated with the Monte Carlo method by assuming constant cloud optical characteristics within each subinterval. This last method (algorithm 1) accurately accounts for the spectral behavior of optical characteristics of clouds and atmospheric gases and can therefore be regarded as a reference. The mean spectral flux computations of high spectral resolution (say, Av = 10-20 cm -I ) may require several hundreds of spectral intervals. Therefore the method of dependent tests (algorithm 2) is proposed, which effectively uses notable features of the spectral dependence of cloud optical parameters, thus allowing significant simplifications and extra savings of computer time. Comparison of results from algorithmg 1 and 2 shows that the mean radiant fluxes agree to within the relative computation error (3 %). This indicates that algorithm 2 is reasonably accurate; in addition, its efficiency is several orders of magnitude better than that of the reference algorithm 1. An important problem, among others, connected with establishing the quantitative relationships between the Earth's radiation field and Earth's atmospheric physical parameters is the problem of radiation energetics. This is because the difference between influxes of tie incoming solar and Earth's outgoing radiation serves as a basic source of energy for the processes occurring in the Earth's atmosphere. Studies in this direction are of fundamental importance for developing atmospheric general circulation theory as well as for models based on this theory and employed in weather forecasting and Earth's climate research, particularly in the mesoscale atmospheric process treatments. At present, general circulation models (GCMs) are numbered in several tens, differing not only in the numerical methods for handling equations of the atmospheric thermohydrodynamies and in techniques for describing subgrid processes, but also in the calculated results as well. The disagreement in results is in part due to different radiation heat exchange parameterizations applied because the errors due to these can affect significantly the description of processes for which the radiative characteristics are input. In this regard, the international Intercomparison of Radiation Codes in Climate Models (ICRCCM) project was started several years ago (Worm Meteorological Organization, 1984,'