Developments Of OPTRAN And Preparation For Its Use To High Spectral Resolution Sounders

Yoshihiko Tahara


NCEP/EMC,UCAR visiting Scientist from Japan Meteorological Agency

Paul V. Delst


NCEP/EMC-Cooperative Institute for Meteorological Satellite Studies

John C. Derber


NCEP/EMC

Larry M. McMillin, Xiaozhen Xiong, Thomas J. Kleespies


National Environmental Satellite and Data Information Service
NCEP/EMC

Abstract:

The development of earth observing techniques from space provides NWP with a large amount of radiance data observed by current and planned high spectral resolution sounders, such as AIRS, IASI, CrIS and GIFTS. Optical Path TRANsmittance (OPTRAN), which is a fast code to calculate layer to space transmittances, has been used in the NCEP NWP system to assimilate satellite radiance observations. OPTRAN uses a large number of equations and coefficients to calculate absorption by atmospheric gas species. The need for a lot of computing resources prevents OPTRAN from being further developed to be able to use additional absorbers and atmospheric predictors. The limitation also prevents the use of the high spectral resolution sounders. In order to reduce the number of equations and coefficients, and to improve the numerical stability of the transmittance computation, the algorithm to calculate the atmospheric absorption has been modified. As a result, the number of the equations has been reduced by a factor of 1/300 and the volume of the coefficients has been decreased more than 95%. In addition, by introducing a stability index, the selection of predictors used in the regression equations and their verification has been simplified and automated. Off-line validation for the modified OPTRAN shows that standard deviations for infrared brightness temperatures are 0.1 or less degrees, where they reach 0.3 degrees except for the frequency range where ozone absorption dominates. The results from an impact study on the NCEP global model will be presented. This study shows that the impact of the modified OPTRAN on the forecast skill is neutral or slightly positive.