Global Ionospheric Total Electron Content Completion with a GAN-Based Deep Learning Framework

The ionosphere serves as a critical medium for radio signal propagation in outer space. A good morphology of the global TEC distribution is very useful both ionospheric studies and their relative applications. In this work, deep learning framework was constructed better spatial estimation TEC. Both DCGAN WGAN-GP were considered, performances evaluated with completion regional using correlation coefficient, RMSE, MAE. Moreover, IAAC rapid products used to make comparisons. results show that outperformed CORG products. clearly goes well solar activity trend. RMSE differences had maximum 0.5035 TECu between 2009 2014 0.9096 WGAN-GP. Similarly, MAE 0.2606 0.3683 CORG, DCGAN, also verified two selected strong geomagnetic storms 2017. RMSEs 1.8354 2.2437 storm on 18 February 2014, respectively, 1.3282 1.4814 7 September GAN-based can extract detailed features daily morphologies responses during storms.