Data assimilation techniques have been widely used in meteorology and oceanography for several decades, and since their introduction there has been a continued improvement in weather prediction. However, only recently has there been a sufficient number of data sources and types for data assimilation to be practical for space weather applications. In anticipation of a major increase in space weather data, the DoD funded a Multidisciplinary University Research Initiative (MURI) program to develop a data assimilation model for the upper atmosphere. As part of this DoD program, Utah State University (USU) developed two physics-based, Kalman filter, data assimilation models of the ionosphere under a project termed Global Assimilation of Ionospheric Measurements (GAIM), and these models are briefly described in what follows. The models include a Gauss-Markov Kalman Filter model and Full Physics-Based Kalman Filter model and both models provide ionospheric specifications and forecasts.

The Gauss-Markov Kalman Filter (GMKF) model is based on the Ionosphere Forecast Model (IFM), which covers the E-region, F-region, and topside ionosphere up to 1500 km, and takes account of six ion species (NO⁺, O₂⁺, N₂⁺, O⁺, He⁺, H⁺). However, the output of the model is a 3-dimensional electron density distribution at user specified times. In addition, auxiliary parameters are also provided, including N_mF₂, h_mF₂, N_mE, h_mE, slant and vertical TEC. In the Gauss-Markov Kalman filter, the ionospheric densities obtained from the IFM constitute the background ionospheric density field on which perturbations are superimposed based on the available data and their errors. To reduce the computational requirements, these perturbations and the associated errors evolve over time with a statistical model (Gauss-Markov process) and not, as in the case of the Full-Physics-Based Model, rigorously with the physical model. As a result, the Gauss-Markov Kalman filter can be executed on a single CPU workstation. Like all assimilation techniques, the Gauss-Markov Kalman filter uses the errors on the observations and model in the analysis, and computes the errors in the match. The Gauss-Markov Kalman filter model is a global model that can support regional, higher definition assimilation windows within the model specification. These regions would be used to provide higher resolution within regions of higher density observations, allowing the model resolution to be adjusted to the number of data sources.

The most sophisticated of the USU models is the Full Physics-Based Kalman Filter. This model rigorously evolves the ionospheric (and plasmaspheric) electron density field and its associated errors using the full physical model. Advantages of this rigorous approach are expected to be most significant in data-sparse regions and during times of “severe weather.” Necessary approximations to make the model computationally tractable capitalize on the newest developments in oceanographic data assimilation. The model is based on a new physics-based model that is composed of an Ionosphere-Plasmasphere Model (IPM) that covers low and mid-latitudes and an Ionosphere-Polar Wind Model (IPWM) that covers high latitudes. These new physics-based models are state-of-the-art and include six ion species (NO⁺, O₂⁺, N₂⁺, O⁺, He⁺, H⁺), ion and electron temperatures, and plasma drifts parallel and perpendicular to the geomagnetic field. These models use the International Geomagnetic Reference Field, which accurately describes the relative positions of the geographic and geomagnetic equators and the declination of the magnetic field lines. The physics-based models cover the altitude range from 90 to 20,000 km, which includes the E-region, F-region, topside ionosphere, plasmasphere, and polar wind. The different real-time data sources are assimilated via a Kalman filter technique and quality control algorithms are provided as an integral part of the full Kalman filter model.

GAIM could potentially assimilate a wide range of data types from numerous ground-based locations and space-based platforms. The data sources include in situ electron densities from NOAA and DoD operational satellites, bottomside electron density profiles from a network of 100 digisondes, line-of-sight Total Electron Content (TEC) measurements between as many as several thousand ground stations and the GPS satellites, TECs between low-altitude satellites with radio beacons and several ground-based tomography chains, TECs via occultations between various low-altitude satellites and between low and high altitude satellites, and line-of-sight UV emission data. It is important to note that GAIM assimilates the data the way they are taken, i.e., slant TEC, bottomside N_e profiles, UV radiances, etc. Also, GAIM has a modular construction so that, in principle, it is straightforward to add new data types. In practice, however, the uncertainty in the data must be known, and this is true for every instrument associated with the various data types. Also, when new data are added, a careful analysis must be conducted of its effect on the Kalman filter. For applications where ionospheric specifications and forecasts are desired, the data must be in real time or in near real time (within 90 minutes of the specification).