FAST YET ACCURATE NET FLUX CALCULATION FOR REALISTIC ATMOSPHERES WITHVARIABLE AEROSOL LOADINGS

Citation
T. Trautmann et Ma. Box, FAST YET ACCURATE NET FLUX CALCULATION FOR REALISTIC ATMOSPHERES WITHVARIABLE AEROSOL LOADINGS, JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES, 100(D1), 1995, pp. 1081-1092
Citations number
19
Categorie Soggetti
Metereology & Atmospheric Sciences
Volume
100
Issue
D1
Year of publication
1995
Pages
1081 - 1092
Database
ISI
SICI code
Abstract
Using a perturbation approach computational results have been obtained for the net radiative flux at selected wavelengths in realistic aeros ol atmospheres. A simple interpolation scheme is presented that employ s three base model computations and allows one to obtain quasi-exact f lux results from precomputed perturbation data in a highly efficient m anner. It is shown how the adjoint transfer equation can be employed t o probe the atmosphere for certain radiative effects, i.e., in the pre sent case for the net radiative flux as a function of height. Results for three different aerosol profiles were obtained and compared with e xact calculations. Contour plots showing the relative error of the net flux underline that, over a wide range of aerosol perturbations (comp are 0.1 less than or similar to tau(c) less than or similar to 0.4 for continental-type aerosols) and for all solar zenith angles, perturbat ion theory produces relative errors below +/-0.25 %. Compared to commo nly used two-stream approximations, the perturbation approach not only offers the possibility to obtain the required model parameters off-li ne, that is, in a precomputed manner, but also enables one to study ra diative effects for thin and moderately thick optical media in the sen se of a functional sensitivity approach. Thus it can clearly be conclu ded that the methodological approach as presented in this study is sup erior to the commonly employed two-stream approximations for radiative transfer computations and in the future may be exploited to overcome the shortcomings of flux models.