ABSTRACT To fully exploit the multi-epoch data processing in GNSS parameter estimation, a thorough understanding of the parameters’ temporal behavior is required. This behavior is expressed by a dynamic model whose uncertainty is captured by the parameters’ process noises. The more randomly the temporal behavior deviates from the dynamic model, the larger the process noise variance should be considered. In GNSS positioning, the ionospheric parameters are known to be time-varying, requiring a reliable choice of process noise. In this contribution, we present a data-driven method with which one can estimate ionospheric process noise variances, minimizing the risk of having a suboptimal estimation process. We study the dependency of such process noise variance on solar and geomagnetic activities across various geographic positions. By analyzing several globally distributed datasets, a considerable difference in the ionospheric process noise estimates is observed between daytime and nighttime, emphasizing the importance of modeling ionospheric temporal characteristic in GNSS positioning.