Error Propagation in Ionospheric Image-based Parameter Estimation

S. Datta-Barua, G. Bust, G. Crowley

Abstract:	Abstract—Using GNSS data effectively in the estimation of ionospheric drivers requires consideration and rigorous treatment of the error propagation of the measurements through the model. In this paper, we develop an error propagation model for one such novel algorithm, known as Estimating Model Parameters from Ionospheric Reverse Engineering (EMPIRE). EMPIRE deduces drivers of ionospheric dynamics from timeevolving 3-dimensional images of ionospheric density. In the algorithm, the ion continuity equation is discretized in spacetime and formulated as a linear system. The system is solved to estimate the drivers that produced an increase or decrease in the time variation of the density over a region. The EMPIRE error analysis begins with the error covariances of the input densities, which in this case are provided by Ionospheric Data Assimilation 4-Dimensional (IDA4D). We apply error covariances from the IDA4D imaging method and propagate them through the spatial and temporal grid interpolation, as well as the linear system inversion, for covariance estimates on the parameters. For the quantities of interest – drifts parallel to the magnetic field, and perpendicular to the field in the meridional and zonal directions – the errors are at minimum 50-100 m/s, which is, for the high-speed storm-time flows, 30% of the stormtime maximum drifts. The chi-squared statistic is also computed and compared to its probability density function to determine goodness-of-fit. There are three distinct periods identifiable in the stormtime data analyzed. Prior to the storm, the chi-squared value is too low, indicating that the input covariances are too conservative. During the initial period of the storm, the chi-squared value indicates that the measurements are reasonably consistent with the model. After the initial couple of hours of the storm, the chi-squared value indicates that the measurements are no longer consistent with the model, so that the fits and estimates are unreliable.
Published in:	Proceedings of the 23rd International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2010) September 21 - 24, 2010 Oregon Convention Center, Portland, Oregon Portland, OR
Pages:	1039 - 1048
Cite this article:	Datta-Barua, S., Bust, G., Crowley, G., "Error Propagation in Ionospheric Image-based Parameter Estimation," Proceedings of the 23rd International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2010), Portland, OR, September 2010, pp. 1039-1048.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In