Reconstruction of the Regional Total Electron Content Maps Over the Korean Peninsula Using Deep Convolutional Generative Adversarial Network and Poisson Blending

Abstract

The Global Navigation Satellite System (GNSS) is a generic term for the Earth-orbiting satellite constellation that transmits positioning and timing data to users. Besides providing precise positioning and timing information to applications requiring them, GNSS products offer valuable resources for ionospheric research and space weather. One of the key products of GNSS is the total electron content (TEC), the electron column number density from GNSS satellites to receivers. Although the GNSS satellite orbits are above 20,000 km, the TEC is predominantly determined by the dense electron density in the ionospheric F region at a height of 200-1,000 km. Because the GNSS TEC provides diagnostics of ionospheric conditions, TEC measurements on the ground and from space

Deep Learning‐Based Regional Ionospheric Total Electron Content Prediction—Long Short‐Term Memory (LSTM) and Convolutional LSTM Approach

“…The final DCGAN-PB TEC maps are generated by integrating the original KASINet TEC data into the synthetic TEC maps using the PB algorithm, thereby preserving observed ionospheric structures. Further details pertaining to the DCGAN-PB TEC maps can be found in Jeong et al (2022).…”

“…KASINet GNSS data are available on the website (http://gnss.kasi.re.kr). Total electron content maps reconstructed by DCGAN-PB are available at Jeong et al (2022), and IRI TEC maps are produced using the IRI-2016 model data available from https://irimodel.org/IRI-2016/. The codebase for deep learning models was built using the Tensor-Flow Library (https://www.TensorFlow.org).…”

“…This method integrates temporal and spatial data information. For training and evaluating the ConvLSTM model, we employ TEC maps generated by the Deep Convolutional Generative Adversarial Network‐Poisson Blending (DCGAN‐PB) method (Jeong et al., 2022). The ConvLSTM model is trained with the DCGAN‐PB TEC maps in 2002–2018, and its performance is assessed using 2019 data.…”

Tandem Observations of Nighttime Mid‐Latitude Topside Ionospheric Perturbations

“…However, in the case of ground‐based TEC measurements using GNSS signals, including GPS, it is not easy to observe MSTIDs in remote areas such as the oceans, deserts, and polar regions because the terrain where the receiver is to be installed is lacking (Yang et al., 2021). Several techniques have been developed to estimate the Vertical Total Electron Content (VTEC) in areas without GNSS receivers (Jeong et al., 2022; Nematipour et al., 2022; Scharroo & Smith, 2010). However, these estimation techniques are generally unsuitable for investigating ionospheric irregularities, such as MSTID.…”

A High Accuracy Spatial Reconstruction Method Based on Surface Theory for Regional Ionospheric TEC Prediction

“…Therefore, the regional model has been the mainstream of research in recent years, and the proposal of the regional model is also the inevitable demand of future communication system applications (Wang, Yang, & Yan, 2021). In decades, researchers have gradually proposed specialized TEC regional prediction models in China (Xiong et al., 2021), Japan (Mallika et al., 2019), India (Sivakrishna et al., 2022), Korean Peninsula (Jeong et al., 2022), South Africa (Ssessanga et al., 2019), Antarctic region (Yao et al., 2021), low latitude region (Zewdie et al., 2021), and ocean (Ren et al., 2022). On the other hand, from the perspective of the methods used for modeling, the spatial reconstruction of TEC can be divided into mathematical and machine learning‐based methods.…”