Benchmark Climatologies from Radio Occultation

Accurate, consistent, long-term data are required for any attempt to detect, understand and at-tribute climate variability and change. Our knowledge about changes in the free atmosphere is still incomplete, despite notable efforts to build long-term upper air records. A new data source for climate monitoring, which can overcome some of the problems of established ones, is the Global Navigation Satellite System (GNSS) Radio Occultation (RO) technique. It has the potential to deliver climate benchmark measurements of the upper troposphere and lower stratosphere (UTLS), since RO data can be traced, in principle, to the international standard for the second. In previous work we could indeed show an amazing consistency of climatologies derived from RO data from different satellites (better than 0.1 K). The value of RO data for climate monitoring is therefore increasingly recognized by the scientific community, but there is also concern about potential residual systematic errors in RO climatologies, which might be common to data from all satellites. We have compiled a list of possible reasons for such systematic errors and will analyze them in detail. This work will lead to a deeper understanding of remaining (small) systematic errors in RO climatologies, which will allow to avoid or remove some of these errors, and to provide quantitative error estimates for those, which cannot be removed. We will use this knowledge to improve the retrieval of RO, and to provide RO climatologies of bending angle, microwave refractivity, density, pressure, geopotential height, and temperature in the UTLS region, with unprecedented accuracy and consistency. Due to their quality, and rigorous characterization, these RO climatologies are expected to qualify as a benchmark for global UTLS climatologies of thermodynamic variables. The dataset will be particularly suitable for climate variability and change monitoring in the UTLS region, e.g., changes in the height of the tropopause, changes in the transition height between tropospheric warming and stratospheric cooling, or changes following any major volcanic eruption potentially occurring during the coming years. The improved understanding of systematic errors will also be useful in the field of Numerical Weather Prediction (NWP), where RO data are assimilated. The overarching goal of BENCHCLIM is to quantify and (if possible) remove potential causes of residual systematic errors in Radio Occultation climatologies of the UTLS - especially those, which could pretend false trends in atmospheric parameters - and to provide high-quality climatologies of the UTLS to the scientific community.

Our knowledge about changes in the free atmosphere is still incomplete, despite notable efforts to build long-term upper air records. A new data source for climate monitoring, which can overcome some of the problems of established ones, is the Global Navigation Satellite System (GNSS) Radio Occultation (RO) technique. It has the potential to deliver climate benchmark measurements of the upper troposphere and lower stratosphere (UTLS), since RO data can be traced, in principle, to the international standard for the second. We could demonstrate that RO climatologies of the UTLS show an amazing degree of consistency (within 0.05 % in refractivity and dry temperature) when RO data from different satellites are processed the same retrieval, but there might be potential residual systematic errors in RO climatologies, common to data from all satellites. We have carefully analyzed reasons for such systematic errors, with the overarching goal to quantify and (if possible) remove potential causes of residual systematic errors in RO climatologies of the UTLS especially those, which could pretend false trends in atmospheric parameters and to provide high-quality climatologies of the UTLS to the scientific community. We could show that a proper treatment of the geometry of RO profiles significantly reduces representativity errors, which had a remarkable positive impact when applied to numerical weather prediction. We have identified the removal of (apparent) outliers in particularly cold regions as a potential source for systematic sampling errors which can be avoided with a sophisticated quality control. We have found a small systematic residual due to incomplete ionospheric correction, which depends on the electron density in the ionosphere together with a potential way to remove this residual. We could demonstrate that the potential problem of systematic errors due to uneven local time sampling can be overcome with data from constellations like COSMIC (consisting of six receiving satellites in six different orbit planes) that actually sample and allow to determine the diurnal cycle. We have developed a new method for the high-altitude initialization of RO profiles and we have identified the atmospheric domain in the UTLS, where it is currently safe to use dry temperature as proxy for the physical temperature, and how this domain is expected to change due to the projected water vapor increase, caused by global warming. Furthermore we used RO data to study multiple tropopauses and sudden stratospheric warmings, and we could demonstrate that RO density profiles are well suited to derive parameters of atmospheric gravity waves.