Ionospheric Error Contribution To GPS Radio Occultation Temperature Retrievals

The NRC Decadal Survey calls for long-term climate records. GPS radio occultation (GPS RO) is viewed as a promising measurement technique that fulfills the need for high accuracy measurements of upper atmospheric temperature and pressure with robust systematic error bounds known on orbit. Geophysical observables are derived by measuring propagation delay induced by the atmosphere, a measurement whose fundamental unit--the second--is absolutely determined by calibration against atomic clocks. Agreement between collocated pairs of observations obtained from the COSMIC GPS RO constellation confirm the method is very precise and inter-satellite retrieval biases are less than 0.1 K. Close agreement between nearly collocated measurements does not directly establish an upper bound for all sources of systematic error that may affect individual retrievals. An important example arises from the ionospheric contribution to signal delay that is largely common to pairs of collocated soundings. Ionospheric impact on GPS RO temperature retrievals must be thoroughly understood before on-orbit SI-traceability can be established at the desired accuracy level (approximately 0.1 K or better). We present results of a realistic simulation study to determine the impact of large-scale ionospheric structure on geophysical retrieval error. GPS transmits signals at two frequencies to compensate for ionospheric error. The two signals do not follow identical paths within the ionosphere possibly leading to incomplete calibration of ionospheric delays and a biased atmospheric retrieval. We use a three-dimensional ray-tracing code that computes the paths of GPS signals through the ionosphere to estimate this impact of the ionosphere. Resultant GPS phase and amplitude data are input to the GPS Occultation Analysis System (GOAS) at JPL to retrieve the geophysical quantities of interest. A new feature of this approach is using JPL's Global Assimilative Ionosphere Model (GAIM) to provide realistic electron density profiles for the ray-tracing calculation. GAIM is a new generation of ionospheric "weather model" that uses data from an extensive measurement network of ground-based GPS receivers to produce realistic electron density maps over a wide range of ionospheric conditions, from quiescent to disturbed. Simulation results and estimates of temperature error are presented for a variety of geophysical conditions such as quiet, geomagnetically disturbed, and relatively quiet periods of enhanced density levels. We conclude with a discussion of implications for on-orbit SI-traceability.