•  
  •  
 

Subject Area

Civil and Environmental Engineering

Article Type

Original Study

Abstract

This study introduces an operational evaluation of a deep learning–driven ionospheric modeling system tailored for Egypt, designed to overcome the limitations of conventional empirical approaches in data-sparse regions. The 3D-DNN Ion-EG model was trained on 737,110 COSMIC-2 radio occultation profiles (2020–2024), integrating spatial temporal coordinates with geophysical drivers to generate high-resolution threedimensional electron density distributions. Validation against the MANS CORS station (Mansoura University) demonstrated strong predictive skill (R2 = 0.9337; r = 0.970) under independent testing. Performance analysis across diverse geophysical regimes confirmed the model’s advantage in capturing storm-time and solar-driven disturbances, surpassing empirical models that perform best only under quiet conditions. Furthermore, the system effectively traced Solar Cycle 25 F2-layer variability from minimum to increasing activity without parameter recalibration. Its computational efficiency enables real-time deployment with modest hardware. These findings establish Egypt as a pioneer in AI-based ionospheric modeling and provide a transferable framework for low-latitude regions.

Keywords

Deep neural networks; COSMIC-2 radio occultation; GNSS positioning accuracy; space weather monitoring; regional atmospheric modeling; solar cycle variability

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS