First Global Machine Learning Model to Predict the Rate of TEC Index (ROTI) Response to X‐Class Solar Flares

Abstract Solar flares are bursts of electromagnetic radiation originating in the Sun's atmosphere. Solar flares cause a rapid increase in ionization in the ionosphere, resulting in radio signal interference. This paper aims to predict the ionospheric response to the solar flare of various characteristics in all latitudes around the dayside ionosphere. X‐ray flux measured by the Geostationary Operational Environmental Satellite (GOES) satellite associated with 84 solar flare events between 2000 and 2017 are obtained. Global total electron content (TEC) data from more than 5,000 ground Global Navigation Satellite System receivers are used. The rate of the TEC Index (Rate of TEC Index (ROTI)) is calculated to examine the time evolution of ionospheric response. Three selected events are studied in detail by eliminating the ROTI associated with stations on the nightside. A nonlinear response of the ionosphere associated with solar flare characteristics including rise/fall time and maximum amplitude is discussed. The first global machine learning (ML) model to predict solar flare impact on Earth's ionosphere through ROTI parameter is developed. Solar flare parameters measured by the GOES satellite along with solar radiation angle and ROTI data from 5‐degree latitude ranges are selected as an input to the ML model. Thek‐nearest neighbors and random forest algorithms are used. Quantitative and qualitative results show that the random forest provides better accuracy in predicting the time evolution of ionospheric response to X‐class solar flare. The coefficient of determination () and the Pearson Correlation Coefficient ( r ) are used to provide a quantitative comparison of the model prediction with the actual data.