oa Validation Of Gps Data-Driven Ionospheric Specification Models

Dual-frequency transmissions from the Global<br>Positioning System satellites have been used for<br>many years to measure and map ionospheric total<br>electron content (TEC) on global scales. JPL's Global<br>Ionospheric Mapping (GIM) software uses<br>observations from ~100 GPS sites, an extended<br>"shell" model, and Kalman filtering to compute<br>global maps of vertical TEC with 15-minute time<br>resolution and ~5-degree spatial resolution. Recently,<br>the GIM algorithms have been enhanced to solve for<br>multiple parameters on a shell grid, instead of just the<br>single value of vertical TEC. Such multi-parameter<br>models are designed to improve the accuracy of slant<br>TEC retrievals and the ability to calibrate slant TEC<br>delays for arbitrary raypaths. The extra parameters<br>allow GIM to better model horizontal gradients and<br>variations in peak height, while still retaining a<br>simple, constrained fitting model as compared to a<br>full tomographic density solution which is vastly<br>underdetermined. A fully 3-dimensional Global<br>Assimilative Ionosphere Model (GAIM) is currently<br>being developed by a joint University of Southern<br>California and JPL team. GAIM uses a firstprinciples<br>ionospheric physics model ("forward"<br>model) and Kalman filtering and 3DVAR techniques<br>to solve for densities on a 3D grid. Although GAIM<br>will ultimately use multiple datatypes and many data<br>sources, one can perform a first test of quantitative<br>accuracy by ingesting GPS-derived TEC observations<br>and then comparing the accuracy of the retrieval to a<br>corresponding run of the enhanced GIM model. A<br>series of such GAIM versus GIM comparisons will<br>be presented and the accuracy of both retrievals will<br>be validated by comparisons to several kinds of<br>independent ionospheric observations: vertical TEC<br>data from the TOPEX altimeter, slant TEC data from<br>GPS sites that were not included in the assimilation<br>runs, and global ionosonde data (F0F2, HMF2, and<br>bottom-side profiles where available). By presenting<br>animated movies of the GAIM densities and TEC<br>maps, and their errors computed as differences from<br>the independent observations, we will characterize<br>the reasonableness and physicality of the climatology<br>derived from the GAIM forward model and the<br>quantitative accuracy of the ionospheric "weather"<br>specification provided by the assimilation retrievals.