The sensitivity of radiative transfer calculations to the changes in the HITRAN database from 1982 to 2004

Abstract

Over the last quarter century, improvements in the determination of the spectroscopic characteristics of the infrared-active trace species have enhanced our ability to retrieve quantitative distributions of temperatures, clouds, and abundances for various trace species within the Earth's atmosphere. These improvements have also allowed for refinements in the estimates of climatic effects attributed to changes in the Earth's atmospheric composition. Modeling efforts, however, have frequently experienced significant delays in assimilating improved spectroscopic information. Such is the case for highly parameterized models, where considerable effort is typically required to incorporate any revisions. Thus, a line-by-line radiative transfer model has been used to investigate the magnitude of the effects resulting from modifications to the spectroscopic information. Calculations from this line-by-line model have demonstrated that recent modifications to the HITRAN (High Resolution Transmission) line parameters, the continuum formulation, and the CO 2 line-mixing formulation can significantly affect the interpretation of the high spectral resolution radiance and brightness temperature retrievals. For certain moderate-resolution satellite remote sensing channels, modifications to these spectroscopic parameters and formulations have shown the capacity to induce changes in the calculated radiances equivalent to brightness temperature differences of 1–2 K. Model calculations have further shown that modifications of the spectroscopic characteristics tend to have a modest effect on the determination of spectrally integrated radiances, fluxes, and radiative forcing estimates, with the largest differences being of order 1 W m −2 for the total thermal infrared fluxes, and of order 2–3% of the calculated radiative forcing at the tropopause attributed to the combined doubling of CO 2, N 2O, and CH 4. The results from this investigation are intended to function as a guide to differentiate between cases where older parameterizations provide acceptable results, within specified accuracy bounds, and cases where upgrades to the latest spectroscopic database are necessary.