Statistical Analysis of Medium‐Scale Traveling Ionospheric Disturbances Over Japan Based on Deep Learning Instance Segmentation

Abstract

The plasma density in the ionospheric F region keeps fluctuating in various temporal and spatial scales. The quasi-wavelike ionospheric disturbances are called as traveling ionospheric disturbances (TIDs), which can be further divided into small-scale TIDs (SSTIDs) (Yin et al., 2019), medium-scale TIDs (MSTIDs) and large-scale TIDs (LSTIDs) (Georges, 1968) by their scales. In terms of MSTIDs, since Sydney Radio Research Board first gave the experimental existence proof (Munro, 1950), observation methods for MSTIDs have been varying all the time (Evans et al., 1983;Jacobson et al., 1995) but the most common methods have been two-dimensional observation since 1990s due to the large observable range. One two-dimensional observation method, High Frequency (HF) radar (Samson et al., 1990), is mostly used to observe daytime MSTIDs. Another method, all-sky airglow imager was also proposed (Mendillo et al., 1997), but it can only work at nighttime because the intensity of atmospheric airglow is too low compared with that of sunlight. Besides, the completion of the first Global Navigation Satellite System (GNSS) in USA, Global Positioning System (GPS), made it possible to carry out

Occurrence and Characteristics of Medium‐Scale Traveling Ionospheric Disturbances Observed by BeiDou GEO Satellites Over Hong Kong

“…Three‐dimensional electron density is the ideal method for studying TIDs, but it is challenging to use it for statistical analysis of MSTIDs due to the large amount of ionospheric tomography data and low computational efficiency. Currently, the most commonly used method for statistical analysis of MSTIDs is the two‐dimensional Detrended TEC (DTEC) map, which has been employed in numerous studies (Cheng et al., 2021; Figueiredo et al., 2018; P. Liu et al., 2022; Otsuka et al., 2013; Tang, 2023; Tsugawa et al., 2007). However, most of these studies have focused on regions with a large number of densely distributed GNSS receiver networks, such as Japan, Europe, America, and so on (Otsuka et al., 2011, 2013; Tsugawa et al., 2007).…”

Spatiotemporal Sequence Prediction of Global Ionospheric Total Electron Content Map Based on Deep Learning Recurrent Neural Network

“…Machine learning technologies [1] that develop rapidly nowadays are leveraged to predict upcoming frames of temporal and spatiotemporal sequences. Recurrent Neural Network (RNN) can restore current output and state of the network as the input of prediction on the next timestamp, thus RNN and its improved version (for example, bi-directional multilayer RNN) are used for temporal sequence prediction including global TEC map prediction.…”