Effect of Grid Boundary Expansion to Include One Additional Data Source on Ionospheric Imaging Accuracy

Four-dimensional ray tomography of ionospheric electron concentration using the Global Navigation Satellite System data is now a well-established technique. Since its advent, there have been a few studies of practical principles for optimizing crucial yet basic aspects of the problem for real experiments. For instance, optimal grid boundaries, voxel numbers, and voxel sizes must be determined case by case. This experiment examines the consequences of a small (<; 11%) increase in the number of voxels in a grid which has its locally horizontal boundary expanded to include one extra ground receiver station. Three different internal division definitions are used for each boundary, making six imaging runs in total. It is found that, when the maximum electron concentration of the ionospheric F-2 layer (N
_m
F2) parameter is compared with independent measurements from an Incoherent Scatter Radar (ISR) over a 12-month period, the expansion of the grid to include the extra data improves the accuracy of the imaging algorithm, regardless of which internal division definition is used. A convolution technique is used to obtain quantitative information about the differences between algorithm runs with different internal divisions and boundaries and the observed ISR values. Examination of the distribution of observed and reconstructed N
_m
F2 values shows that the imaging algorithm does not produce as many low-valued results as the observed data.