This paper addresses the often surprisingly difficult challenges that arise in conceptually simple validations of global models of the ionosphere.

AFRL has been tasked with validating the Utah State University GAIM (Global Assimilation of Ionospheric Measurements) model of the ionosphere, which is run in real time by the Air Force Weather Agency.

The USU-GAIM model currently assimilates, in addition to the voluminous GPS TEC data, in situ densities from DMSP satellites, UV radiances from SSUSI sensors on the DMSP satellites, and vertical profiles provided by a limited number of Digisondes. AFRL has performed a large number of USU-GAIM validations, using as ground truth values of foF2 and M(3000)F2 from non-assimilated ionograms, the in situ electron density at ~400 km provided by CHAMP, and the vertical TEC provided over ocean areas by TOPEX and JASON.

USU GAIM runs at AFRL in about one-third real time. For validations against ionogram characteristics, AFRL usually works with a full month of GAIM and Digisonde data, which takes ~10 days to run. The long run times make it difficult to address essential “what if” scenarios, except for limited time intervals. Compounded with the problem of long run times is the fact that the UV observations are from a satellite that is only very rarely in near conjunction with the ground-truth satellites such as CHAMP and JASON, or near ground-based ionosondes. Exacerbating this problem even further is the fact that the most reliable assimilated UV data is from the evening equatorial ionosphere. It is often not possible to obtain useful ionogram characteristics for the evening equatorial ionosphere because of the occurrence of irregularities that lead to spread F echoes on the ionograms.

We will discuss the impact of these various challenges on the lessons that can be learned from validation studies of global ionospheric models.