Assessing Model Performance of SAMI3 and IRI-2020 Against GNSS-TEC Observations Under Quiet and Disturbed Space Weather Conditions at Low Latitudes

The Equatorial Ionization Anomaly (EIA) introduces strong spatial and temporal variability in the ionosphere over low-latitude regions, posing significant challenges for Global Navigation Satellite System (GNSS)-based positioning, navigation, and communication systems. This study presents a comparative analysis of ionospheric Total Electron Content (TEC) derived from the physics-based Sami3 is Another Model of the Ionosphere (SAMI3) and the empirical International Reference Ionosphere 2020 (IRI-2020) models against ground-based GNSS-TEC observations from the International GNSS Service (IGS) network across contrasting space weather periods, including quiet, moderate, and geomagnetically disturbed intervals, over a low-latitude region. SAMI3 output, generated in apex coordinates, was regridded to a geographic coordinate system and vertically integrated to obtain TEC, with model time converted to local time to enable direct comparison with GNSS observations. Diurnal and day-to-day variability, EIA crest location and strength, and storm-time TEC enhancements and depletions were examined across the selected periods. During quiet-time conditions, the observed daytime peak TEC reached 63 TECU, whereas SAMI3 and IRI-2020 underestimated the daytime maximum, with peak values of 41 and 40 TECU, respectively; IRI-2020 also exhibited a phase lag in peak timing. Near the pre-dawn minimum, IRI-2020 reproduced the observations (4–5 TECU) more closely than SAMI3 (2–3 TECU). In the post-sunset sector, SAMI3 decayed markedly faster than the observations, underestimating TEC by more than 50% during quiet-time evening hours, whereas IRI-2020 exhibited a more gradual, observation-consistent decline into the night. These results indicate that SAMI3 better reproduces the timing of daytime EIA development, whereas IRI-2020 better captures evening and nighttime TEC persistence, reflecting differences between the physics-based photochemical processes represented in SAMI3 and the smoother climatological formulation of IRI-2020. The comparison provides insight into model-specific biases relevant to regional GNSS error mitigation and future SAMI3–GNSS-TEC data assimilation efforts.