Evolution of spectroscopic information over the last decade and its effect on line-by-line calculations for validation of radiation codes for climate models

Five HITRAN's editions (since 1992) and three radically new H2O spectral databases have been tested for their usefulness in line-by-line benchmark calculations for radiation code validations. The longwave calculations for the mid-latitude summer (MLS) atmosphere with the different HITRAN's editions revealed a good convergence (maximal discrepancies since similar to2.5 W/m(2) decreased to similar to0.5 W/m(2) in our flux calculations with 1992-2002 and 1996-2002 HITRAN's editions, respectively). The shortwave calculations with different HITRAN's editions revealed a good agreement: discrepancies in downward fluxes absorbed by the MLS atmosphere are less than 1 W/m(2), i.e. 0.5% for the all HITRAN's editions since 1992. It has been determined that it is unreasonable to replace HITRAN by any new H2O spectral database in benchmark calculations, although they contain essentially more spectral lines than HITRAN (up to similar to10 times). Moreover, it has been found that no new database could be useful for explaining the "anomalous absorption" problem. Two latest water vapor continuum models have been tested as well. It has been found out that these models are the main source of uncertainties in the present and previous calculations: discrepancies for the MLS atmosphere reached similar to3 and similar to2 W/m(2) in the longwave and shortwave flux calculations, respectively. (C) 2003 Published by Elsevier Ltd.