Improvements of One-dimensional Variational Assimilation Algorithm Based on Occultation Technique

Utilizing the occultation data observed by LEO (low earth orbit) satellites, we can invert the profiles of air pressure, water vapor, temperature of terrestrial atmosphere and so on, and they are the valuable data resources for studying the meteorology and atmospheric sciences. The assimilation techniques with the occultation data can effectively improve these profiles of meteorological parameters, thus they can improve the accuracy of current numerical weather forecasting. The greatest difficulty is the huge amount of calculation to enter into the meteorological operational processes for applying the parameter profiles observed by occultation with the method of variational assimilation. With the improvements of the function of one-dimensional variational assimilation and the new design for iterative process, the defects of repeated calculation of the large dimension matrix can be avoided, thereby improving the computational efficiency of variational assimilation. In the discussion of applicability, it is used as the true value of the vectors of the background field plus one white Gaussian noise to test the variational assimilation results on the occultation data of satellite CHAMP.