loading page

On the Variation of Column $O/N\textsubscript{2} $ in the upper atmosphere using Principal Component Analysis in 2-dimensional images
  • +4
  • Divyam Goel,
  • Yen-Jung Joanne Wu,
  • Brian J Harding,
  • Colin Triplett,
  • Thomas J. Immel,
  • Cullens Chihoko,
  • Scott L England
Divyam Goel
Department of Electrical Engineering and Computer Science, University of California Berkeley
Author Profile
Yen-Jung Joanne Wu
University of California, Berkeley

Corresponding Author:[email protected]

Author Profile
Brian J Harding
University of California, Berkeley
Author Profile
Colin Triplett
University of California, Berkeley
Author Profile
Thomas J. Immel
University of California, Berkeley
Author Profile
Cullens Chihoko
University of Colorado at Boulder / LASP
Author Profile
Scott L England
Virginia Polytechnic Institute and State University
Author Profile

Abstract

Day-to-day variability in thermospheric composition is driven by solar, geomagnetic and meteorological drivers. The ratio of the column density of atomic oxygen and molecular nitrogen (O/N\textsubscript{2}) is a useful parameter for quantifying this variability that has been shown to exhibit close correspondence to F-region electron density, total electron content and upper atmospheric transport. Therefore, understanding the variability in O/N\textsubscript{2} gives an insight into the geophysical variability of other relevant ionospheric and thermospheric parameters. The relative contributions of these drivers for thermospheric variability is not well known. Here we report a new analysis of the variability in O/N\textsubscript{2} to identify the sources of variability in a 55-day time period. Principal Component Analysis (PCA) was performed on thermospheric O/N\textsubscript{2} column density ratio from days 81 to 135 of 2020 from NASA’s Global-scale Observations of the Limb and Disk (GOLD) mission. We find that geomagnetic activity is the major source of variability in O/N\textsubscript{2} column density ratio, followed by solar-driven transport and meteorological driving from the lower atmosphere. The first component (PC1) showed a strong correlation to Kp index and IMF, and geomagnetic storm effects are seen in the wavelet analysis of PC1’s weights. The fifth component (PC5) showed a strong quasi-6-day oscillation(Q6DO). The higher explained variance ratio of PC1 suggests a stronger effect of geomagnetic activity relative to meteorological forcing from planetary scale waves. The methodology of the present study also demonstrates how PCA can be used to isolate and rank different sources of variability in other IT parameters.