Daily Arxiv

This page organizes papers related to artificial intelligence published around the world.
This page is summarized using Google Gemini and is operated on a non-profit basis.
The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.

Forecasting the Ionosphere from Sparse GNSS Data with Temporal-Fusion Transformers

Created by
  • Haebom

Author

Giacomo Acciarini, Simone Mestici, Halil Kelebek, Linnea Wolniewicz, Michael Vergalla, Madhulika Guhathakurta, Umaa Rebbapragada, Bala Poduval, At{\i}l{\i}m G une\c{s} Baydin, Frank Soboczenski

A time-series transformer-based machine learning framework for ionospheric TEC prediction.

Outline

In this study, we present a machine learning framework utilizing Temporal Fusion Transformers (TFTs) to predict ionospheric variability, which is difficult to predict due to the nonlinear coupling between solar, geomagnetic, and thermospheric drivers. We predict the total electron number density (TEC) derived from GNSS observations, using a variety of inputs including solar radiation, geomagnetic indices, and GNSS-based vertical TEC. Experimental results from 2010 to 2025 demonstrate a low root mean square error (RMSE) of 3.33 TECU for predictions up to 24 hours in advance, with solar EUV radiation providing the strongest predictive signal. Reproducibility is enhanced through the open-source toolkit \texttt{ionopy}.

Takeaways, Limitations

•
Takeaways:
◦
Development of a machine learning framework for predicting sparse ionospheric data.
◦
Predicting ionospheric TEC using various input sources.
◦
Excellent accuracy (3.33 TECU RMSE) for up to 24-hour forecasts.
◦
Providing interpretability through attention-based analysis.
◦
Encouraging reproducibility and community contributions through open source toolkit distribution.
•
Limitations:
◦
There is no direct mention of Limitations in the paper. (However, Limitations can be inferred from outside the paper.)
◦
Additional datasets are needed to verify the model's generalization ability.
◦
Further analysis of forecast performance under strong space weather conditions is needed.
👍
Made with Slashpage