3-day Forecasting of Global TEC Map Using a Deep Learning Model
Abstract
In this study, we develop a global Total Electron Content (TEC) forecasting using a deep learning method. Our deep learning model is based on Pix2PixHD (Wang et al. 2018) which is image-to-image translation method. For training the deep learning model, we use the International GNSS Service (IGS) TEC maps from 2003 to 2012. For further research and comparison of model results, we used same data as previous study (lee et at. 2021). Our model uses two input data: current time and 12 hours ago on the global TEC map. Output data are 6 TEC maps with 12-hour interval (t+12, t+24, ..., t+72). Our model tested with solar maximum period (20132014) and minimum period (20172018). We evaluated the model using Pearson Correlation Coefficient (CC), Root Mean Square Error (RMSE), bais and Standard Deviation (STD). In solar maximum period, CC, RMSE, bias, STD between our model and IGS TEC map are 0.96 0.97 TECU, 4.23 5.65 TECU, -0.31 2.20 TECU and 3.92 4.85 TECU, respectively. In solar minimum period, CC, RMSE, Bias, STD between our model and IGS TEC map are 0.96 0.97 TECU, 1.82 2.20 TECU, -0.43 -0.01 TECU respectively. Our study shows using deep learning method successfully generate long term TEC map. This work was supported by Institute for Information & communications Technology Promotion (IITP) grant funded by the Korea government (MSIP) (2018-0-01422, Study on analysis and prediction technique of solar flares).
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFMNG45B2335Y