Robust ionospheric tomography using sparse regularization

The ionosphere is a dynamic medium which can interact with electromagnetic waves and cause delay and refraction/diffraction effects. Ionospheric structures occur with different scale in different locations. Therefore, the correct imaging of the location and scale of those structures is of particular importance. Computerized ionospheric tomography uses observations from ground-based receivers to reconstruct the state of the ionosphere and is denoted as an inverse problem. Globally, ground-based receivers tend to be unevenly and sparsely distributed and produce limited-angle observations. Therefore the definition of the problem in a regularized form is required. Furthermore, there is a need to accommodate inconsistencies in observations as the data rate increases, e.g. due to representativity error or residual dispersive offsets. This is the main focus of the paper. A new sparse regularization form, for ionospheric tomography, was introduced. It promotes sparsity in the reconstruction and it is tailored for wavelet representation due to their localization properties. An experiment using simulated data is proposed comparing sparse regularization with a standard approach for 3D ionospheric tomography. Results indicate sparse regularization as a promising technique showing a higher reliability of reconstructed ionospheric structures and robustness to observational inconsistencies compared to a more standard approach.