A simulation study into constructing of the sample space for ionospheric imaging

Ionospheric imaging usually involves solving an underdetermined inversion problem. The inversion is performed involving additional constraints to enforce realistic profiles in the vertical. One way to incorporate those vertical profile constraints is to perform the inversion using Empirical Orthogonal Functions (EOFs). The need of defining a sample space spanned by EOFs to obtain ionospheric images yields the possibility to employ ionosonde measurements in ionospheric tomography based on stochastic inversion of GPS data. Here we present a simulation study based on an existing network of GPS ground receivers and ionosondes across Europe. The locations of the transmitters used in the simulation are actual satellite positions. Simulated GPS data, constructed assuming that the ionosphere is the international reference ionosphere, are inverted via the Multi Instrument Data Analysis System. The sample space of this stochastic inversion is constructed employing ionosonde measurements simulated from the same model ionosphere. Such use of ionosonde data to construct the sample space produces better results than without ionosonde data.