A new longwave radiation model for application to atmospheric problems

The accuracy of radiation models has become a critical issue in climate studies. In this dissertation, eight existing research and operation radiation models are compared against the accurate surface observations with high spectral resolution to determine the absolute accuracy of the current models. Moreover, a new accurate narrow band longwave radiative transfer model for clear-sky conditions is developed. The model is used to calculate the case study sensitivities previously studied by the international InterComparison of Radiation Codes used in Climate Models (ICRCCM), particularly the sensitivities of the important greenhouse gases water vapor and carbon dioxide.; In the model development we first show that traditional techniques for estimating Malkmus statistical model parameters from the line compilation and line-by-line models can not be trusted to give accurate transmittance function. We then describe a new technique that calculates water vapor line transmittances with good agreement with line-by-line (LBLRTM) calculations (i.e., with rms errors less than 0.01 for more than 97% of the intervals). The water vapor continuum is included in a manner consistent with the water vapor line absorption. The modeling of the gases CO{dollar}sb2{dollar}, O{dollar}sb3{dollar}, N{dollar}sb2{dollar}O and CH{dollar}sb4{dollar} adopts the Malkmus formula with the parameters fit to LBLRTM transmittances. Fluxes calculated with the model agree with LBLRTM to about 1 W/m{dollar}sp2{dollar} for the entire vertical range of the atmosphere for several test cases. The heating rate errors are reduced by as much as 0.25{dollar}spcirc{dollar}C/day below the tropopause for the test cases compared with the original narrow band model.; The model is validated with surface observations, and is compared with the other models. Both LBLRTM and the new model are documented for ICRCCM test cases, and the performances of the 'average' ICRCCM climate models are evaluated with respect to the LBLRTM calculations.