Propagation of Non-Gaussian Uncertainties from Storm-Time Neutral Density to Orbital Position

Satellites in low Earth orbit (LEO) are subject to atmospheric drag driven by neutral density, which is highly variable and amplified during geomagnetic storms. Accurate drag modeling during storms is critical for collision avoidance and space situational awareness (SSA). While orbit prediction errors are often modeled under Gaussian assumptions and represented with covariance ellipsoids, prior work suggests that density distributions are non-Gaussian especially during geomagnetic storms. The impact of this variability on orbital position uncertainty has not been quantified.In this brief report, we propagate circular LEO orbits through Monte Carlo density fields, obtained from a particle filter neutral density estimation framework that employs the Thermosphere Ionosphere Electrodynamics General Circulation Model (TIEGCM), during both quiet and storm periods. Reference orbits without drag are compared against drag-included trajectories, and position differences are analyzed in the in-track direction. The statistics of density and in-track position differences are presented.The results show that storm-time density exhibits larger mean and variance compared to quiet-time conditions, as well as elevated skewness and kurtosis, confirming its non-Gaussian structure. These properties drive in-track position differences to be larger with a higher variance and to deviate from Gaussian distributions. The magnitude of storm-time in-track position differences resulting from these density uncertainties is comparable to position uncertainties caused by initial orbital state parameter errors typically considered in SSA orbital determination applications.This study suggests the importance of accounting for realistic neutral density variability in drag modeling to improve the reliability of orbit predictions during geomagnetic storms.