Team ID: O3B-01 

Team Title: Deep Learning-Based Prediction Model for Equatorial Ionospheric Irregularities

 

Team Leads:

Lin Min Min (King Monkut's Institute of Technology Ladkrabang (KMITL), linminmin.my@kmitl.ac.th

Anan Kruesubthaworn (King Monkut's Institute of Technology Ladkrabang (KMITL), anankr@kku.ac.th
 

Keywords (Activity Type):
Understanding, Modeling, Forecasting, Data Utilization, Training
 

Keywords (Impact):
Navigation and/or Communications


Cluster with overlapping topics:
G1: Geomagnetic environment, G2B: Ionosphere variability

 

Introduction:

The equatorial ionosphere often develops irregularities, particularly equatorial plasma bubbles (EPBs), which cause trans-ionospheric signal scintillation and degradation in GNSS applications. EPBs originate near the magnetic equator and then extend to higher latitudes along magnetic field lines while drifting zonally. Consequently, the impact of these irregularities can cover wide regions across multiple countries. Accurate prediction of such events is vital for mitigating their effects on navigation and communication systems. This team focuses on developing a deep learning–based prediction model for equatorial ionospheric irregularities using GNSS-derived parameters such as total electron content (TEC) and magnetometer data from magnetic equatorial regions, integrated with interplanetary solar wind parameters. By utilizing comprehensive space weather inputs and cutting-edge deep learning techniques, we aim to enhance short-term forecasting capabilities for GNSS users, including aviation, and to provide insight into the underlying physical drivers of irregularity formation.

Objectives:

1.    To collect and integrate multi-source datasets including total electron content (TEC) from both ground-based and in-situ receivers, magnetometer data from equatorial stations, and interplanetary solar wind parameters for model development.
2.    To develop and optimize a deep learning–based prediction model capable of identifying and forecasting equatorial ionospheric irregularities and EPB occurrences.
3.    To enhance and extend the model for predicting irregularity scale, spatial impact area, and duration, and to expand its applicability to other longitudinal sectors toward global prediction capability. 
4.    To evaluate model performance through comparison with ground-based and satellite observations, such as ionosonde and VHF radar data, to validate prediction accuracy across different geomagnetic and seasonal conditions.