Skip to main content

Low Latitude Detection and Tracking of Plasma Bubbles Using Global-Scale Observations of the Limb and Disk (GOLD) 135.6 nm Data With and Without Convolution Neural Network

Vincent Adkins, Virginia Polytechnic Institute and State University
Scott England, Virginia Polytechnic Institute and State University
First Author's Affiliation
Virginia Polytechnic Institute and State University
Abstract text:

Plasma bubbles are regions of depleted plasma within the upper thermosphere/ionosphere that form during post-sunset hours near the magnetic equator. These structures tend to align with local geomagnetic field lines, extend upwards hundreds of kilometers along geomagnetic longitudes, and thousands of kilometers along geomagnetic latitudes. These large scale plasma density gradients can attenuate lower frequency radio waves. Small scale structures along the walls can interfere with centimeter scale wavelengths. Statistical analysis of this phenomenon can further understanding of their occurrence and subsequent behavior. The current study utilizes Global-Scale Observations of the Limb and Disk (GOLD) 135.6 nm nightglow data from October 5, 2018 to September 30, 2022. GOLD has a constant and consistent view of nightglow and structures over the Americas and Atlantic due to its geostationary orbit. A plasma bubble detection method is developed and used to generate a database of plasma bubble occurrences. Occurrences are used to calculate plasma bubble drift speeds and separations. Clear seasonality in plasma bubble occurrence rate is evident. Overall occurrences peak during December solstice months and minimize during June solstice for longitudes seen by GOLD. Within GOLD's field of view, higher occurrences are seen to the west during December solstice and east during June solstice. Plasma bubble drift speeds and separations show consistent distributions regardless of magnetic region, geographic region, season, or local time. This suggests plasma bubbles behave consistently and regularly once formed, at least on spatial scales seen by GOLD. The plasma bubble occurrence database is also used to train a convolution neural network to detect and classify plasma bubbles. This has the potential to detect and classify this phenomenon in real time as new data becomes available.

Student in poster competition
Poster category
EQIT - Equatorial Ionosphere or Thermosphere