Assessing the Impact of New GNSS Radio Occultation Measurements with an Ensemble of Data Assimilations Approach

GNSS radio occultation (GNSS-RO) measurements are now a key component of the global observing system. They complement the information provided by satellite radiance measurements because they can be assimilated without bias correction to the NWP model, and they have good vertical resolution. In the coming years, new GNSS-RO measurements from the Eumetsat Polar System Second Generation (EPS-SG) and the joint United States/Taiwanese COSMIC-2 mission will form the basis of a high quality "core constellation" of around 18,000 profiles per day. These core measurements may be supplemented with a large number of additional "medium-quality" GNSS-RO measurements, which are characterised by larger errors in the stratosphere. We have used an Ensemble of Data Assimilations (EDA) approach to estimate the impact of the additional medium-quality data when added to the core GNSS-RO constellation. This talk will describe the EDA approach for assessing new observations, and compare it to the more conventional OSSEs. The EDA results will be interpreted as a 4D-Var information content study. We will then assess the impact of adding up to 64,000 globally distributed medium-quality GNSS-RO observations, when added to a core constellation of 18,000 high-quality GNSS-RO observations and all other available observing systems. The limitations of the EDA results will be discussed, and future work will be outlined.