P3.8
Quantification of errors in satellite data assimilation
Dongsoo Kim, NOAA/FSL and CIRES/Univ. of Colorado, Boulder, CO; and S. E. Koch and P. Van Delst
Accurate specification of error covariances of variables of assimilation system is very important in the variational assimilation as they control fittings to observations and control variables. In satellite data assimilation, errors of forward model must be specified. We use NCEP Community Radiative Transfer Model (CRTM) to compute radiances using various input profiles, both NWP forecast and RAOB. For evaluation of forward model error, we used ARM's Balloon-borne Sounding System (BBSS) profiles projected on 64 layer GFS grids, 60 layer Eta grid, 50 level RUC grids. The differences among simulated radiances from measured radiances of GOES sounder channel guide error bound of forward model.
Poster Session 3, Data Assimilation
Tuesday, 21 September 2004, 9:30 AM-11:00 AM
Previous paper Next paper