Abstract
The disposition of mean July clear-sky solar radiation in the Canadian Climate Centre second-generation general circulation model (CCC-GCMII) was analyzed by comparing top of the atmosphere (TOA) net fluxes with earth radiation budget experiment (ERBE) data and atmospheric and surface net fluxes with values inferred from Li's algorithm using ERBE data and European Centre for Medium-Range Weather Forecasts precipitable water data. GCMII tended to reflect ˜5 W m−2 too much to space. Corresponding atmospheric and surface absorption, however, tended to be too low and high, respectively, by ˜30 W m−2 over much of the Northern Hemisphere. These results were echoed when GCMII atmospheric absorption was compared to estimated results from Li's algorithm driven by GCMII TOA albedo and precipitable water.
The latest version of the CCC-GCM (GCMIII) has numerous upgrades to its clear-sky solar radiative transfer algorithm, the most important of which involve water vapor transmittances and aerosols that tend to enhance atmospheric absorptance. GCMIII's water vapor transmittance functions derive from Geophysical Fluid Dynamics Laboratory line-by-line results, whereas GCMII's were based on Air Force Geophysical Laboratory data. GCMIII includes crude distributions of background tropospheric aerosols, whereas GCMII neglected aerosols.
Li's algorithm was then driven by GCMIII data, and atmospheric absorption of solar radiation by GCMIII was assessed. Differences between GCMIII's and Li's atmospheric absorption over land were almost always within 5 W m−2. Over oceans, differences were mostly between −5 W m−2 and −15 W m−2. This apparent underestimation over GCMIII's oceans probably stems from the algorithm's use of a thin, highly absorbing aerosol.
