Comparison of clear-sky surface radiative fluxes simulated with radiative transfer models

Abstract

The surface fluxes of several important radiatively active gases, including H2O, CO2, CH4, N2O, O3, and the chlorofluorocarbons CFC11 and CFC12, were simulated with the radiation band models from the National Center for Atmospheric Research (NCAR) community climate model 3 (CCM3), the single-column community atmospheric model (SCAM), and the Canadian global climate model 3 (GCM3). These results were compared with the measured fluxes for a very cold winter day and with the simulated results for other standard atmospheres using the line-by-line radiative transfer model (LBLRTM). The comparison shows that the total surface radiative flux contributed by all the greenhouse gases combined is well simulated by the SCAM and GCM3 radiation band models. The two models generally agree within about 1% of the line-by-line result for all the atmospheric conditions studied. The error in the total flux simulated by the older CCM3 code, however, can be as large as 7% depending on the atmospheric conditions. The SCAM code consistently models H2O better than the CCM3 and GCM3 codes, typically displaying errors of less than 1 W/m2 for all atmospheric conditions. All of the models have difficulty in modelling accurately the radiative flux of CH4 and N2O. In general, the inaccuracy increases, by as much as 200% in some cases, as the amount of H2O in the atmosphere increases. The source of the problem appears to be related to the overlapping bands of other gases. The error in the ozone flux varies from 5% to 15% for the CCM3 and SCAM models, and it can be as large as 30% for the GCM3 code. The CCM3 and SCAM models simulated the chlorofluorocarbon fluxes to within 0.06 W/m2, but this leads to relative errors of 20%–40% for the various atmospheric scenarios. The errors for the CFCs are even larger in the case of the GCM3 model.