Monitoring Space Weather: Using Automated, Accurate Neural Network Based Whistler Segmentation for Whistler Inversion. Issue 2 (16th February 2022)

• Record Type: Journal Article
• Authors: Pataki, Bálint Ármin; Lichtenberger, János; Clilverd, Mark; Máthé, Gergely; Steinbach, Péter; Pásztor, Szilárd; Murár‐Juhász, Lilla; Koronczay, Dávid; Ferencz, Orsolya; Csabai, István
• Is Part Of: Space weather. Volume 20:Issue 2(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-02-16
• Subjects: whistler -- plasmasphere -- neural network; Space environment -- Periodicals; 551.509992
• DOI: 10.1029/2021SW002981
• Languages: English
• ISSNs: 1542-7390

Abstract

Abstract: It is challenging, yet important, to measure the—ever‐changing—cold electron density in the plasmasphere. The cold electron density inside and outside of the plasmapause is a key parameter for radiation belt dynamics. One indirect measurement is through finding the velocity dispersion relation exhibited by lightning induced whistlers. The main difficulty of the method comes from low signal‐to‐noise ratios for most of the ground‐based whistler components. To provide accurate electron density and L $L$ ‐shell measurements, whistler components need to be detectable in the noisy background, and their characteristics need to be reliably determined. For this reason, precise segmentation is needed on a spectrogram image. Here, we present a fully automated way to perform such an image segmentation by leveraging the power of convolutional neural networks, a state‐of‐the‐art method for computer vision tasks. Testing the proposed method against a manually, and semi‐manually segmented whistler data set achieved < $< $ 10% relative electron density prediction error for 80% of the segmented whistler traces, while for the L $L$ shell, the relative error is < $< $ 5% for 90% of the cases. By segmenting more than 1 million additional real whistler traces from Rothera station Antarctica, logged over 9 years, seasonal changes in the average electron density were found. The variations match previously published findings, and confirm the capabilities of the image segmentation technique. Plain Language Summary: When lightning strikes on the Earth, electromagnetic waves are generated, that can travel along the magnetic field lines of the Earth and can be observed in the other hemisphere. As the waves are in the 1–30 kHz range, they can be measured with radio antennas and exhibit a whistling sound, thus their name, whistlers. As the electromagnetic waves travel, they are distorted due to the velocity dispersion in the ionosphere and the magnetosphere. As the rate of dispersion depends on the physical parameters of the above‐mentioned regions, the precise measurement of the whistlers can be a method to monitor the electron density on different magnetic field lines. The cold electron density inside and outside of the plasmasphere is a key parameter for radiation belt dynamics, that can affect satellites negatively, their precise measurement can be fruitful in many applications. Whistlers have been measured and collected for decades, however, due to the noisy nature of the detection data, the precise, automated estimation of the physical parameters through whistler traces has been a challenging task. We trained a convolutional neural network based model to accurately mark the whistlers on a frequency‐time spectrogram. The trained model can find and segment the whistlers, which allows accurate physical parameter estimation, within 10% error in electron density for 80% of the time. Key Points: Monitoring the electron density along different L‐shells by measuring the dispersion relation via whistlers Convolutional neural network based method to accurately segment the whistler traces from spectrogram images Seasonal variation of plasmaspheric equatorial electron density based on more than 1 million analyzed whistler traces over 9 years