A New Method for Generating Radio Occultation Climatologies

Although temperature change measurements at the surface are well established, the knowledge about the vertical structure of temperature changes in the free atmosphere is still incomplete. The Global Navigation Satellite System (GNSS) Radio Occultation (RO) technique, however, has the potential of providing benchmark quality data, suitable for climate monitoring in the upper troposphere and lower stratosphere. The RO principle is based on excess phase measurements, from which atmospheric profiles of, e.g., density, pressure and temperature can be derived with high accuracy and high vertical resolution in a long-term stable manner. Nevertheless, in the process of retrieving geophysical parameters of the Earths atmosphere from the initially observed measurements, introduction of background information is necessary by performing a so-called high-altitude initialization of bending angle profiles. This step reduces noise in the data and is commonly realized by a statistical optimization (SO). However, the high-altitude initialization has been identified by the ROtrends collaboration as a major source for structural uncertainties between the RO climate data products from different processing groups. Another important error source in RO data arises from an incomplete correction of an additional ionospheric excess phase experienced by the GNSS signals. This systematic error depends on the ionization, and therefore, increases with solar activity. Recently a new approach for the production of climatologies has been proposed, having the decisive advantage of circumventing the sophisticate SO step. The idea is to perform the averaging of individual profiles already in bending angle space, and retrieving climatological data products of density, pressure, and temperature directly. This reduces the noise in the RO data, leads to a cleaner and easier computation, a clear error characteristic, and avoids a potential error source. This approach has been introduced using COSMIC satellite data. In a follow-up study we showed it is also applicable to (noisier) CHAMP satellite data. In the NEWCLIM project I want to enhance this new averaging approach. The central aim is to perform a careful validation of the new climatologies, in order to test if the new retrieval scheme can be seen as a full valid alternative to the standard retrieval scheme. First, I will compare the new climatologies between three processing centres, including derived parameters like temperature. Furthermore, I will investigate the residual influence of the ionosphere in these new climatologies. The goal is to reduce the ionospheric residual error by using a new and promising model correction. Finally, I will test possible limits of the new climatologies. NEWCLIM aims to target two central problems of RO data in order to achieve benchmark-quality climatologies, the high-altitude initialization and residual ionospheric errors. The new climatologies have the potential to push current upper limits in altitude, enabling to study stratospheric climate processes.

For almost 20 years, the satellite-based radio-occultation (RO) technique has been providing high-quality atmospheric data from the upper troposphere to the lower stratosphere (5 km to 35 km). The RO data are used e.g., for meteorological applications, climate analysis and monitoring. The original measurement of the phase shift scans a vertical profile up to a height of approximately 80 km. However, the data quality decreases with increasing altitude, due to two major error sources, which are (i) the noise in the data and the different ways of handling it in the retrieval process, and (ii) the residual ionospheric errors (RIEs) in the data. The overall goal of the project NEWCLIM was to improve the quality of the data obtained in the middle and upper stratosphere, by handling the problems of noise and ionospheric influences. This enables a more precise analysis of processes in the stratosphere, which are decisive for our climate system. An alternative approach was investigated, which tackles the problem of random noise in the RO data, by a very simple and interesting idea. The so-called average-profile-inversion (API) performs the climatological averaging already at bending angle space, and hence suppresses the random noise in the data, rather than smoothing and merging individual profiles with background information. In a next step, the bending angle climatologies are directly processed to refractivity, pressure, and temperature climatologies. This simple idea is suited for climatological studies, having the decisive advantage of using only observational data up to an altitude of 80 km. As initially planned, the study on the API approach was conducted together with the Danish Meteorological Insitute, where datasets from different processing centers were compared. Results showed a robustness between the processing centers up to an altitude of about 35 km. The second research objective was to handle RIEs in RO data. An easy-to-use model was prior developed in a collaboration with the European Center for Medium-Range Weather Forecasts. The special thing about this model is that only the originally measured phase shift and the F10.7 index, which is a measure of solar activity, are required to calculate the ionospheric correction. Other previously known corrections required much more additional information, like the electron density at the time of the measurement. This model, the so-called kappa correction, was tested for the first time with observed RO data. Detailed information about the sensitivity of the kappa-correction and the impact of RIEs were provided for all RO parameters. In summary, the two introduced methods, the average-profile-inversion and the kappa-correction, are both important tools for dealing with residual errors in RO data. They increased the data usage in the stratosphere to a height of up to about 40 km altitude.