A non-linear optimal estimation inverse method for radio occultation measurements of temperature, humidity and surface pressure

摘要

An optimal estimation inverse method is presented which can be used toretrieve simultaneously vertical profiles of temperature and specific humidity,in addition to surface pressure, from satellite-to-satellite radio occultationobservations of the Earth's atmosphere. The method is a non-linear, maximum{\it a posteriori} technique which can accommodate most aspects of the realradio occultation problem and is found to be stable and to converge rapidly inmost cases. The optimal estimation inverse method has two distinct advantagesover the analytic inverse method in that it accounts for some of the effects ofhorizontal gradients and is able to retrieve optimally temperature and humiditysimultaneously from the observations. It is also able to account forobservation noise and other sources of error. Combined, these advantages ensurea realistic retrieval of atmospheric quantities. A complete error analysis emerges naturally from the optimal estimationtheory, allowing a full characterisation of the solution. Using this analysis aquality control scheme is implemented which allows anomalous retrievalconditions to be recognised and removed, thus preventing gross retrievalerrors. The inverse method presented in this paper has been implemented for bendingangle measurements derived from GPS/MET radio occultation observations of theEarth. Preliminary results from simulated data suggest that these observationshave the potential to improve NWP model analyses significantly throughout theirvertical range.