Evaluation of Ionospheric Models for Central and South Americas

This work shows a 20-month statistical evaluation of different Total Electron Content (TEC) estimators for the Central and South America regions. The TEC provided by the International GNSS Service (IGS) in the area covered around the monitoring GNSS stations are used as reference values, and they are compared to TEC estimates from the physics-based (Sheffield University Plasmasphere Ionosphere Model—PIM) and the empirical (Neustrelitz TEC Model-Global—NTCM-GL) models. The mean TEC values show strong dependence on both solar activity and seasonal variation. A clear response was noticed for a period close to 27 days due to the mean solar rotation, as seen in the solar flux measurements. Consistently, the mean TEC values present an annual variation with maxima during December solstices for southern stations with geographic latitudes greater than 25° S. Semi-annual dependence has been observed in TEC for the sector between ±25° of geographical latitude but with modulations caused by fluctuation in the solar radiation. We observed a high correlation between solar radio flux F10.7 and NTCM-GL outputs. The fast increases in F10.7 index have caused significant differences between IGS data and NTCM-GL results mainly for equatorial and low latitudes. For the initial months of the evaluated period (January–April, 2016), the errors of the physics-based model were considerably larger, mainly near the equatorial ionization anomaly. The discrepancies observed in SUPIM results are mainly due to inputs of solar EUV flux. The EUVAC model has underestimated EUV flux between January and April, 2016, when the solar activity was moderated and Solar2000 model has overestimated such flux during low solar cycle period between May and August, 2017. In relation to IGS data, the two assessed models presented smaller differences during the June solstice season of 2016.