|Projektname:||AV-BENDAPP - Feasibility of generating long-term RO refractivity climatologies without using statistical optimization|
Kent B. Lauritsen, Danish Meteorological Institute (DMI), Kopenhagen, DK
|Fördergeber:||DMI, ROM SAF I|
|Dauer:||Okt. 2013 - Feb. 2014|
The retrieval of geophysical information from the observed ionosphere-corrected bending angles requires the use of a priori information. For example, the retrieval of a refractivity profile from neutral bending angles requires an extrapolation of the bending angle profile to infinity. The observations have a limited extension in altitude, and the signal to noise level decreases rapidly with altitude from the upper stratosphere. The standard procedure to retrieve refractivity from noisy bending angles is to smooth and merge the observed bending angle profiles individually with a priori data taken from a climatology. The upper-level bending angle initialization may add to the structural uncertainty of the retrieved variables. This source of structural uncertainty is less important for individual profiles, but may still contribute significantly to the monthly means. It has been suggested that rather than handling the noise in the bending angles separately for each profile, it may be advantageous to first suppress the noise by averaging a large number of profiles. This reduces the random errors and raises the altitude where instrumental errors and ionospheric residuals become significant. The monthly mean refractivities can be directly retrieved from an average bending angle, rather than from statistically optimized refractivity profiles. Studies show that the average-bending angle approach can be used to produce zonal monthly mean refractivities from COSMIC data. However, the generation of long-term climate data requires the use of CHAMP data from 2001 to mid-2006. CHAMP data are more sparse and considerably noisier than COSMIC data. Particularly the data quality may pose a problem. In this project we will study whether the average-bending angle approach works also during the CHAMP period, with an emphasis on the impacts of data quality and data numbers.