Assessing ionospheric models with COSMIC-2 F2 peak data as a first step toward an Equal Weight Multi Model Ensemble

Kurzfassung

Space weather events can significantly alter electron density in the ionosphere, affecting signal transmission, navigation accuracy, and satellite lifetime. These activities necessitate reliable predictions for the thermosphere-ionosphere system (TI). With the growing number of TI models, ensemble modeling techniques are becoming increasingly practical, but they depend on rigorous evaluation of the individual models. In this study, the results of a multi-model ensemble for ionospheric F2 peak parameters and a weighting scheme based on model performance comparisons with COSMIC-2 data are presented. We evaluate the quality of the NmF2 and hmF2 results from an empirical model (IRI20) and four physics-based models (TIE-GCM3.0, CTIPe, WACCM-X, and PBMOD) by comparing them with the F2 peak values determined from the radio occultation profiles of COSMIC-2, using a total of four 20-day comparison periods from 2020. An ensemble with equal weighting is created based on the predictions from the physics-based models. Four metrics are used to evaluate the model predictions: residuals, correlation coefficients (CC), root mean square errors (RMSE), and prediction efficiencies (PE). The ensemble results for NmF2 show a maximum improvement in RMSE of 27.3\% compared to the performance of the individual models, and for hmF2, the improvement in RMSE ranges from 10.7\% to 35.4\%. The uncertainty of the ensemble results from its spread (indicated by the standard deviation). A one-day case study shows that the observations are largely within the spread of the ensemble. Based on this evaluation, it is possible to create a non-equal weighting scheme to improve the multi-model ensemble in future work.