Sensitivity of Stratospheric Retrievals from Radio Occultations on Upper Boundary Conditions

The main uncertainty in the stratospheric retrievals from GPS radio occultation (RO) measurements comes from the lack of reliable measurements in the upper stratosphere and above where the bending due to the neutral atmosphere is weak and residual ionospheric effects are strong. In this work, we quantify the bias and uncertainty of the refractivity and temperature retrievals due to different upper boundary strategies using a simulation study. We use refractivity profile derived from lidar pressure and temperature profiles as the input states in generating the synthetic occultations. Random noise levels commensurate with the CHAMP RO measurements are added to the simulated data. We examine the sensitivity of stratospheric retrievals to two different upper boundary methods, one based on exponential extrapolation and the other on MSIS climatology. The simulation results show that both methods lead to comparable levels of temperature bias (less than 0.5 K below 30 km altitude), provided that the upper boundary heights are set above 55 km.