Global TEC Map Fusion Through a Hybrid Deep Learning Model: RFGAN

Abstract

Total electron content (TEC) is a quantitative measurement of ionospheric columnar electron content within the distance from satellite to ground receiving station. The TEC is not only the key characteristic of ionospheric morphology, but also generally utilized in ionospheric correction related to precise positioning, navigation and radio wave science. Since 1998, the Massachusetts Institute of Technology (MIT) has been collecting and processing TEC observations with high spatial and temporal resolution (MIT-TEC) from more than 6,000 GPS observatories around the world (Vierinen et al., 2016). Due to the limited coverage of GPS receiver stations that measure TEC data (A. J. Mannucci et al., 1998), the original MIT-TEC map is incomplete, limiting the study in ocean areas. The missing portion of the MIT-TEC data is about 52% of the world. Over the past few decades, Global Navigation Satellite System (GNSS) has become one of the main technical means of earth ionosphere detection due to its advantages of all-weather, receiver's easy-to-deploy and high precision (Davies, 1990;A. Mannucci et al., 1999). Since the establishment of the Ionospheric Working Group by International GNSS Services (IGS) in 1998, the Global Ionospheric TEC Grid (IGS-TEC), as an important part of the GNSS precision products released Abstract Timely, reliable and comprehensive global observation information is essential for space weather research. However, limited observation technology hinders the consecutive global coverage of observation data. For the integrity and continuity of the global observation data, deep learning can obtain a global Ionospheric total electron content (TEC) map by fusing multi-source TEC maps. Different from the previous methods, in the study, a deep learning hybrid model (RFGAN) based on Dual-Discriminator Conditional Generative Adversarial Network (DDcGAN) and Free-Form Image Inpainting with Gated Convolution (Deepfill v2) is proposed to fuse the Massachusetts Institute of Technology (MIT)-TEC, International Global Navigation Satellite System TEC (IGS-TEC) and altimetry satellite TEC. Throughout the RFGAN structure, we use an autoencoder model with gated convolution to inpaint the missing parts of MIT-TEC and altimetry satellite TEC. Meanwhile, DDcGAN fuses the inpainted MIT-TEC (MIT'-TEC) and IGS-TEC to get a global TEC map with high accuracy. To a certain extent, we inpainted the ocean area of MIT-TEC through RFGAN. At the same time, RFGAN keeps the consistency of RFGAN-TEC and MIT-TEC in the continent area. Our proposed deep learning hybrid model can be easily extended and widely applied to other fields of space science, especially in addressing observational data loss and multi-source data fusion.Plain Language Summary Ionospheric total electron content (TEC) maps are very important for satellite navigation, shortwave communications and space weather research. There are many TEC map data from different sources, and their TEC show significant difference. Massachusetts Institute of Technology (MIT)-TEC and altimetry ...

An Index Description of the General Characteristics of Thermospheric Density Based on the Two‐Line‐Element Data Sets and the Spectral Whitening Method

“…(2014). Currently, an index group $${J}_{x}^{Y}$$ characterizing the solar to ionospheric disturbances in the Sun‐Earth space has been established based on the SWM (Chen et al., 2014, 2023a, 2023b; Wang et al., 2014; Zhao et al., 2022). This method can effectively remove the periodic components (background) in the data, making the non‐periodic components (disturbances) of interest distinguishable in the final results.…”

Forecasting 24‐Hr Total Electron Content With Long Short‐Term Memory Neural Network

“…Since the correctness of the input data was out of scope in these investigations it may be beneficial to have a closer look at the accuracy of GIMs and perhaps including other techniques proposed in Chen et al. (2023) in the future. The rapid UQRG GIMs contain global TEC maps with a 5° × 2.5° resolution of longitude and latitude, respectively.…”

Storm‐Time Characteristics of Ionospheric Model (MSAP) Based on Multi‐Algorithm Fusion

“…Chen et al. (2019, 2023) further predict missing data of TEC map by a proposed deep learning algorithm: regularized deep convolutional generative adversarial network (RDCGAN), which suggests that deep learning can well extract the spatial feature of the TEC map. Moses et al.…”

“…It avoids the situation where traditional neural networks get stuck in local minima when training data. Chen et al (2019Chen et al ( , 2023 further predict missing data of TEC map by a proposed deep learning algorithm: regularized deep convolutional generative adversarial network (RDCGAN), which suggests that deep learning can well extract the spatial feature of the TEC map. Moses et al (2020) constructed the African Regional Ionospheric TEC Model with deep learning technology to model TEC in the African region.…”