• Data assimilation;
  • Numerical weather prediction;
  • Satellite data;
  • Variational analysis


A methodology is described for the direct assimilation of TOVS (TIROS-N (Television Infra-Red Observation Satellite) Operational Vertical Sounder) radiance data that have not undergone the pre-processing (such as limb adjustment and cloud clearing) that is usually applied to make radiances suitable for use in linear retrieval algorithms. Inherent nonlinearities in the observations can be treated successfully within a variations analysis scheme such as Spectral Statistical Interpolation (SSI) thus making radiance pre-processing unnecessary. The benefit of using unprocessed (or level-lb) data is demonstrated in terms of an improved ability to quality control and bias correct the observations within the analysis. The assimilation of level-1b radiance data in the SSI has shown a significant positive impact on the quality of the analysis compared with the use of pre-processed radiances, and a significant improvement in forecast skill for the tropics and southern hemisphere.