Statistical validation of a solar wind propagation model from 1 to 10 AU

Abstract

[1] A one-dimensional (1-D) numerical magnetohydrodynamic (MHD) code is applied to propagate the solar wind from 1 AU through 10 AU, i.e., beyond the heliocentric distance of Saturn's orbit, in a non-rotating frame of reference. The time-varying boundary conditions at 1 AU are obtained from hourly solar wind data observed near the Earth. Although similar MHD simulations have been carried out and used by several authors, very little work has been done to validate the statistical accuracy of such solar wind predictions. In this paper, we present an extensive analysis of the prediction efficiency, using 12 selected years of solar wind data from the major heliospheric missions Pioneer, Voyager, and Ulysses. We map the numerical solution to each spacecraft in space and time, and validate the simulation, comparing the propagated solar wind parameters with in-situ observations. We do not restrict our statistical analysis to the times of spacecraft alignment, as most of the earlier case studies do. Our superposed epoch analysis suggests that the prediction efficiency is significantly higher during periods with high recurrence index of solar wind speed, typically in the late declining phase of the solar cycle. Among the solar wind variables, the solar wind speed can be predicted to the highest accuracy, with a linear correlation of 0.75 on average close to the time of opposition. We estimate the accuracy of shock arrival times to be as high as 10-15 hours within ±75 d from apparent opposition during years with high recurrence index. During solar activity maximum, there is a clear bias for the model to predicted shocks arriving later than observed in the data, suggesting that during these periods, there is an additional acceleration mechanism in the solar wind that is not included in the model.

Planetary period oscillations in Saturn's magnetosphere: Examining the relationship between abrupt changes in behavior and solar wind‐induced magnetospheric compressions and expansions

“…Based on the solar wind observations by the Ulysses spacecraft at~5 AU from the Sun, Tao et al [2005] estimated that the arrival prediction error of the MHD simulation is at most 2 days for an Earth-Sun-observer angle less than 60°. A statistical analysis by Zieger and Hansen [2008] of the propagated upstream solar wind data from 1 AU near Earth to 10 AU estimated the accuracy of shock arrival times to be as high as 10-15 h within ±75 days from apparent opposition during years with high recurrence index (so when Earth and the spacecraft were separated by less than ±75°in longitude). The error estimates from Zieger and Hansen [2008] and Tao et al [2005] were calculated when they used only one Earth-based solar wind monitor.…”

“…A statistical analysis by Zieger and Hansen [2008] of the propagated upstream solar wind data from 1 AU near Earth to 10 AU estimated the accuracy of shock arrival times to be as high as 10-15 h within ±75 days from apparent opposition during years with high recurrence index (so when Earth and the spacecraft were separated by less than ±75°in longitude). The error estimates from Zieger and Hansen [2008] and Tao et al [2005] were calculated when they used only one Earth-based solar wind monitor. For this study we use solar wind measurements both from an Earth-based monitor and also from the two STEREO spacecraft, significantly increasing the number of days for which the solar wind monitor is within a conducive narrow angular range of Saturn.…”

Solar Energetic Particles (SEP) and Galactic Cosmic Rays (GCR) as tracers of solar wind conditions near Saturn: Event lists and applications

“…Finally, the use of models that predict the solar wind conditions at the two several days (Tao et al, 2005;Zieger and Hansen, 2008;Witasse et al, 2017). space (Cane et al, 1988;Reames, 1999).…”

Colliding winds in low-mass binary star systems: wind interactions and implications for habitable planets

“…The VAC is a freely available general purpose magnetohydrodynamic (MHD) code that has been used extensively in the past for simulations of the solar wind (Keppens & Goedbloed 1999;Zieger & Hansen 2008;Jacobs & Poedts 2011). By only simulating the wind in 1D using an adaptive mesh that is refined close to the solar surface and becomes increasingly coarse far from the Sun, we are able to simulate the acceleration and propagation of the wind from the stellar surface to 1 AU.…”