Uncertainty quantification of the ionosphere-thermosphere conditions associated with the solar wind
 

Information of the uncertainty in the ionosphere-thermosphere (IT) system is important to make probabilistic predictions of space weather events (i.e., equatorial ionospheric irregularities), satellite or debris obits (i.e, neutral density and atmospheric drag), and impacts on HF communications (i.e., HmF2). In this study, we apply uncertainty quantification (UQ) methods to quantify the variability of the IT conditions associated with the uncertainty of the solar wind by using the whole atmosphere model with ionosphere-plasmasphere dynamics (WAM-IPE) model. To apply the UQ method and conduct sensitivity analysis, historical measurements of solar wind velocity, density, and interplanetary magnetic field (IMF) Bz are fed into a neutral network (variational autoencoder, VAE) to reduce the dimensions of input parameters and generate new solar wind samples to drive WAM-IPE. Polynomial chaos expansion (PCE) is then applied to quantify the uncertainty of the quantities of interest (QoIs) from the outputs of WAM-IPE. Electron densities, plasma drifts, and neutral winds from 6 magnetic equatorial and low latitude locations are selected as QoIs. The uncertainties of these QoIs will be compared with previous observations and numerical simulations. Sensitivity analysis will be conducted to identify the drivers that contribute the most to the uncertainty of the QoIs.