GNSS-RO Residual Ionospheric Error (RIE): A New Method and Assessment

Abstract. GNSS radio occultation (RO) observations play an increasingly important role in monitoring climate changes and numerical weather forecasts in the upper troposphere and stratosphere. The magnitudes of the RO bending angle are small at these altitudes, and therefore residual ionospheric error (RIE) is critical to retrieve vertical profiles of atmospheric temperature and refractivity. The latter represent the state variables of the weather and climate models. RIEs remain poorly characterized in terms of the global geographical distribution and its variations with the local time and altitude influenced by the solar cycle and solar-geomagnetic disturbances. In this study we developed a new method to determine RIE from the RO excess phase measurement on a profile-by-profile basis. The method, called Φ_ex-gradient method, is self-sufficient and based on the vertical derivative of the RO excess phase (Φ_ex) profile, which can be applied to individual RO bending angle observations for RIE correction. In addition to the RIE in bending angle measurements, RIEs are found in the RO Φ_ex measurements in the upper atmosphere where an exponential dependence is expected and in small-scale temperature variance of the RO retrieval. We found that the RIE values derived from the Φ_ex-gradient method can be both positive and negative, which is fundamentally different from the k-method that produces only the positive RIE values. The new algorithm reveals a latitude-dependent diurnal variation with a larger daytime negative RIE (up to ~3 μrad) in the tropics and subtropics. Based on the observed RIE climatology, a local-time dependent RIE representation is used to evaluate its impacts on reanalysis data. We examined these impacts by comparing the data from the Goddard Earth Observing System (GEOS) data assimilation (DA) system with and without the RIE. The RIF impact on GEOS DA temperature is mainly confined to the polar regions of stratosphere. Between 10 hPa and 1 hPa the temperature differences are ~1 K and exceed ~3–4 K in some cases. These results further highlight the need for RO RIE correction in the modern DA systems.