An Automated Auroral Detection System Using Deep Learning: Real-time Operation in Tromsø, Norway

Nanjo, Sota; Nozawa, Satonori; Yamamoto, Masaki; Kawabata, Tetsuya; Johnsen, Magnar Gullikstad; Tsuda, Takuo T.; Hosokawa, K.

doi:10.21203/rs.3.rs-1090985/v1

Cited by 1 publication

(1 citation statement)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Omega bands propagate at 0.3–2 km/s, and while the activity can persist continuously for several hours, each bump/torch/tongue structure may take a few to a few tens of min to pass a given magnetometer. The numbers above were obtained manually by scientists, and can further be improved by recent application of machine learning to auroral images which makes an automatic identification and tracking of various auroral forms possible (for example, Clausen & Nickisch, 2018; Kvammen et al., 2020; Nanjo et al., 2022).…”

Section: Discussionmentioning

confidence: 99%

Auroral Drivers of Large dB∕dt During Geomagnetic Storms

et al. 2022

View full text Add to dashboard Cite

Forecasting geomagnetically induced currents (GICs) remains a difficult challenge, and open questions hindering our understanding include when and where GICs become large and what magnetospheric and ionospheric processes are responsible. This paper addresses these questions by determining the auroral drivers of large dB∕dt (>100 nT/min, a proxy for GICs) on the ground during geomagnetic storms. We study auroras because, although the current system driving dB∕dt is at times challenging to reconstruct, the accompanying auroras are routinely measured in high resolution. For various types of auroras, our community has already acquired a deep understanding of the driving mechanisms and spatiotemporal characteristics. Using coordinated observations from THEMIS and Geophysical Institute Magnetometer Array magnetometers and THEMIS all‐sky imagers, we statistically examine large dB∕dt intervals during storms from 2015 to 2016. A variety of auroral drivers have been identified, including poleward expanding auroral bulges, auroral streamers, poleward boundary intensifications, omega bands, pulsating auroras, etc. The onset, spatial variability, and duration of large dB/dt are well explained by those of the auroras. For example, poleward expanding auroral bulges drive large dB/dt that spread progressively poleward, and periodic injections of streamers drive large dB/dt that occur in periodic bursts. By referring to the magnetospheric source of the auroras, the magnetospheric source of large dB/dt can be inferred, whether it be dipolarization of the tail magnetic field, bursty bulk flows, instability, or wave‐particle interaction. Our results suggest that auroras can exert significant leverage on GIC research and forecast.

show abstract

Section: Discussionmentioning

confidence: 99%