Comparative testing of empirical, data assimilation and physics-based ionospheric/thermospheric models

This paper compares and describes the ability of empirical, data assimilation and physics based ionospheric/thermospheric models to provide profiles of electron density below the peak of the F2 layer. A median empirical model, three data assimilation models and two different approaches to physics-based modelling are considered. A modified Taylor diagram is used to compare the second-order statistics of the foF2, hmF2 and h(0.8foF2) parameters simultaneously. The paper provides a preliminary study in understanding the effectiveness of using physics-based models as the background in Kalman filter data assimilation schemes. It is shown that the data assimilation models perform best followed by the median model and finally the physics based models.