Modeling Ionograms and Critical Plasma Frequencies With Neural Networks

Enrique Rojas Villalba ^1* Jhassmin Aricoche ² Marco Milla ³

• ¹ Haystack Observatory, Massachusetts Institute of Technology, Cambridge, United States
• ² Cornell University, Ithaca, New York, United States
• ³ Pontifical Catholic University of Peru, Lima, Peru

Ionosondes provide superior spatial coverage and a more comprehensive sampling of the lower ionosphere than other alternatives. In this work, we used neural networks (NN) to forecast ionograms across two solar activity periods. The ionosonde data was obtained from the digisonde at the Jicamarca Radio Observatory (JRO). Each NN comprises one NN that estimates the ionogram trace and another one that estimates the critical frequency. Furthermore, two forecasting models were implemented. The first one was trained with all available data and was optimized for accurate predictions along that time range. The second one was trained using a rolling-window strategy with just three months of data to predict one day of ionograms. Our results show that both models are comparable and can often outperform predictions by empirical and numerical models.The hyperparameters of both models were optimized using a specialized library. Moreover, we found that a few months of data was enough to produce predictions of comparable accuracy to the reference models. We argue that this high accuracy is obtained because the NN captures the dominant periodic drivers. Finally, we provide suggestions for improving this model.