As radiation plays a key role in the determination of the near-surface thermal environment, great accuracy is required in the computation of radiative fluxes, especially because a small error in the fluxes can lead to large errors in estimated cooling rates. A new code that employs a novel numerical scheme for making precise estimates of longwave fluxes and cooling rates near the surface of the earth, for arbitrary surface emissivities, is presented here. The code is a development of the infrared band model of Chou, Ridgway, and Yan. Unacceptable oscillations found near the surface in the cooling rates provided by this code have been overcome through the new numerical scheme. The new code gives results in excellent agreement with available results from Intercomparison of Radiation Codes used in Climate Models (ICRCCM) test cases and with the line-by-line calculations of Clough, Iacono, and Moncet. The code has no restriction on the number of grid points, yields fluxes accurate to a prescribed tolerance, and permits a discontinuity in temperature at the surface (although this is not used in the results presented). The computing times are comparable (for given accuracy) to those demanded by current codes in use elsewhere. It is found that, as surface emissivity $\varepsilon_ {g}$ departs from unity, the cooling rate rises dramatically near the surface, reaching values as high as nearly 40 K $day^{-1}$ at $\varepsilon_ {g}$ = 0.8 in the midlatitude summer atmosphere, and the effect of the surface is noticeable at heights of up to about 1 km. An analysis of spectral distribution shows that, when the surface is not radiatively black, the major contributions to near-surface cooling rates (due to water vapor) come from the two wavenumber bands, 340–540 $cm^{-1}$ and 1215–1380 $cm^{-1}$ (located on either side of the atmospheric window), in which both absorption and radiative flux are significantly high.
Read full abstract