Global Ionospheric ROTI Prediction Based on DeepONet

This study extends our previous research published in IEEE Transactions on Geoscience and Remote Sensing [1], which focused on predicting the Slant Total Electron Content (STEC) of the ionosphere using the Deep Operator Network (DeepONet) method. We now advance this work by implementing short-term forecasts of the Rate of Total Electron Content Index (ROTI), broadening the application of the DeepONet operator network in forecasting ionospheric irregularities. This work uses data from February 2023 for training and predicts the ROTI for March 1, 2023, achieving an average Root Mean Square Error (RMSE) accuracy of approximately 0.1 TECU/min. Additionally, the study found that the model's prediction deviations mainly occur near the equator and the poles. This investigation confirms DeepONet's capabilities for effectively forecasting ionospheric conditions, which could significantly improve the operational efficiency of global satellite navigation systems.