A multiresolution inversion for imaging the ionosphere

Abstract

Ionospheric tomography has been widely employed in imaging the large‐scale ionospheric structures at both quiet and storm times. However, the tomographic algorithms to date have not been very effective in imaging of medium‐ and small‐scale ionospheric structures due to limitations of uneven ground‐based data distributions and the algorithm itself. Further, the effect of the density and quantity of Global Navigation Satellite Systems data that could help improve the tomographic results for the certain algorithm remains unclear in much of the literature. In this paper, a new multipass tomographic algorithm is proposed to conduct the inversion using intensive ground GPS observation data and is demonstrated over the U.S. West Coast during the period of 16–18 March 2015 which includes an ionospheric storm period. The characteristics of the multipass inversion algorithm are analyzed by comparing tomographic results with independent ionosonde data and Center for Orbit Determination in Europe total electron content estimates. Then, several ground data sets with different data distributions are grouped from the same data source in order to investigate the impact of the density of ground stations on ionospheric tomography results. Finally, it is concluded that the multipass inversion approach offers an improvement. The ground data density can affect tomographic results but only offers improvements up to a density of around one receiver every 150 to 200 km. When only GPS satellites are tracked there is no clear advantage in increasing the density of receivers beyond this level, although this may change if multiple constellations are monitored from each receiving station in the future.

Performance of MIDAS Over East African Longitude Sector: Case Study During 4–14 March 2012 Quiet to Disturbed Geomagnetic Conditions

“…It has been shown that tomographic imaging captures the ionosphere better when an “adequate” and even GPS rays distribution traverses the voxels of the reconstruction grids radially (Chartier et al, ; Spencer & Mitchell, ). Tomographic results using observations from a clustered network of ground stations could also equally capture an “unrealistic” ionospheric structures as those obtained from a scarce distribution of ground stations (Chartier et al, ; Yin et al, ). Yin et al () recently reported the impact of the density of ground stations on tomographic imaging results and revealed that improvements could only be offered when receivers are to within 150 km to 200 km distance from each other.…”

“…Tomographic results using observations from a clustered network of ground stations could also equally capture an “unrealistic” ionospheric structures as those obtained from a scarce distribution of ground stations (Chartier et al, ; Yin et al, ). Yin et al () recently reported the impact of the density of ground stations on tomographic imaging results and revealed that improvements could only be offered when receivers are to within 150 km to 200 km distance from each other. Furthermore, the authors pointed out that better improvements could be achieved with the inclusion of multiple constellations besides the GPS constellation.…”

“…Usually the inverse problem is underdetermined; hence, certain basis functions are often implemented to compensate for the missing information by selecting the appropriate mapping matrix, thereby representing the spatial and temporal distribution of the ionospheric electron density separately (Bust & Mitchell, ; Yin et al, ). For the horizontal variation of the electron density, MIDAS uses a grid (defined in latitude and longitude) with various resolution, while a set of empirical orthogonal functions is generated from certain ionospheric models (such as IRI and Chapman function) for the radial direction.…”

“…The TEC processing assumes that the ionosphere is an evenly distributed thin shell at an altitude of 450 km and generate global TEC data with the resolution of 2.5° by 5° in latitude and longitude every 2 h intervals (refer to Desai et al, , and the references therein). Similarly to Yin et al (), GIM VTEC data, set to same resolution of MIDAS reconstructions through linear interpolation, were compared to reconstruction results for validation purposes.…”

Reconstruction of Storm‐Time Total Electron Content Using Ionospheric Tomography and Artificial Neural Networks: A Comparative Study Over the African Region

“…VTEC from MIDAS was computed by vertical integration of the electron density obtained following the above procedure. Detailed theory about MIDAS can be found in a number of literature sources (Bust et al, 2007;Jayawardena et al, 2016;Meggs et al, 2005;Mitchell & Spencer, 2003;Spencer & Mitchell, 2007;Yin et al, 2017) and references therein. For more information about the recent MIDAS algorithm used in this study, readers are referred to Spencer and Mitchell (2007), Giday et al (2016), and Yin et al (2017).…”