Modeling Space Weather Parameters with Machine Learning: Total Electron Content Forecasting

---

Aydın Ö. F., Can Z.

Turkish Physical Society 41th International Physics Congress, Muğla, Turkey, 1 - 05 September 2025, pp.253, (Summary Text)

• Publication Type: Conference Paper / Summary Text
• City: Muğla
• Country: Turkey
• Page Numbers: pp.253
• Yıldız Technical University Affiliated: Yes

Abstract

Accurate prediction of the Total Electron Content (TEC) in Earth's ionospheric layer,

which constitutes a natural plasma environment, is critical for mitigating the adverse

effects of space weather events on technological systems.

This study aims to develop a high-accuracy prediction model by using input parameters

including the Dst, Kp, and F10.7 Solar Flux indices, the solar position, and historical TEC

data. In this context, mid-latitude TEC measurements obtained from the ionolab service,

along with relevant space weather indices, were used to train ensemble-based machine

learning models.

The performance of the models was evaluated using standard statistical metrics such as

Root Mean Square Error (RMSE) and the Coefficient of Determination (R2). The results

demonstrate that machine learning techniques offer reliable tools with the potential for

operational implementation in TEC forecasting.

Our future work will focus on analyzing the performance of this successful model in high-

latitude polar regions.