Evaluation of Satellite Sampling of the Middle Atmosphere Using the GFDL Skyhi General Circulation Model

Research has been performed to quantify differences between meteorological diagnostics defined from measurements of satellite radiance and those defined from a four-dimensional verification data set, NOAA's Geophysical Fluid Dynamics Laboratory's SKYHI general circulation model (GCM). A complete simulation of the calculation of vertical temperature profiles from asynoptic, nadir-viewing, polar orbiting satellite data was performed. Channel radiances for the High Resolution Infrared Sounder (HIRS-2), the Stratospheric Sounding Unit (SSU) and the Microwave Sounding Unit (MSU) instruments which compose the Tiros-N Operational Vertical Sounder were defined for the SKYHI atmosphere. These radiances were then used to calculate vertical temperature profiles using the technique of statistical linear regression analysis. Computer codes modeling this mathematical technique as well as the definition of channel radiances were supplied by NOAA/NESDIS in Suitland, MD. NOAA/NESDIS currently uses these programs in an operational mode to process measurements of atmospheric radiance obtained via the satellite platform. The inferred satellite temperatures were subsequently used to define meteorological diagnostics which include geopotential heights, gradient zonal winds and geostrophic meridional winds, meridional fluxes of momentum and temperature, Eliassen-Palm flux divergences and isentropic potential vorticity. Comparisons were made between the satellite's definition of these diagnostics and the diagnostics as defined directly from the SKYHI GCM temperature and wind data. The technique of statistical linear regression analysis requires the use of an initialization data set comprised of 'near-real time' and/or climatological vertical temperature and water vapor mixing ratio profiles. Results of this study reveal that the fidelity of the satellite -inferred temperature profiles relative to the verification data set is severely biased towards the information comprising the initialization data set. This is particularly evident during highly transient meteorological episodes such as the stratospheric sudden warming. An examination of a hierarchy of issues surrounding the use of data from the satellite platform as a meteorological investigative tool is provided.