TEC variations and IRI-2016, IRI-2020 and IRI-Plas performance in Mexico

The present work discusses TEC variations over Mexico and their representation by the advanced ionospheric models IRI-2016, IRI-2020 and IRI-Plas during the years of low and increasing solar activity, 2018-2022. The statistical estimates Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Normalized Root Mean Square Error (NRSME) were used to assess model performances. The earlier revealed TEC features were confirmed and new features were revealed. The hour of the diurnal TEC maximum varies in a wide range within 10-20 LT. In summer, it occurs at earlier hours than in other seasons. During the solar minimum, the annual variation (two minima and two maxima during a year) was found absent and a seasonal dependence of the lowest night TEC value is revealed. The accuracy of TEC representation depended on the model, hour of the day, season and solar activity level and suffered a notable change in 2022. In the considered period, IRI-2016 overestimated the daytime and underestimated the nighttime TEC in most cases. Essentially, IRI-2016 and IRI-2020 gave similar results. IRI-Plas overestimated the daytime TEC during the low solar activity and mostly gave adequate estimates at night. Its accuracy at day hours increased with the growth of solar activity. Quantitative estimates of the accuracy of the models are as follows. The yearly average of RMSE for IRI-2016 and IRI-2020 was minimal in 2018 and maximal in 2022. In contrast, for IRI-Plas it was minimal in 2022 and maximal in 2021. At nighttime, RMSE typically varied as follows. For IRI-2016, within (1-5) TECU during low and within (2-12) TECU during increasing solar activity; for IRI-2020 within (1-3) TECU and (1-9) TECU; and for IRI-Plas within (0.5-4) TECU and (2-9) TECU, respectively. At daytime, RSME were higher: for IRI-2016 within (4-11) TECU under low and (6-19) TECU under increasing solar activity; for IRI-2020 (4-9) TECU and (4-17) TECU; and for IRI-Plas (10-19) TECU and (6-15) TECU, respectively. The discrepancies between the modeling results and observations at low latitudes are due to the poor representation of EIA structures and topside and plasmaspheric parts in TEC. TEC annual variation is always represented by the models, but, it was practically absent during the solar minimum according to the experimental results.