Premium
Using Principal Component Analysis to Characterize the Variability of VLF Wave Intensities Measured by a Low‐Altitude Spacecraft and Caused by Interplanetary Shocks
Author(s) -
Bezděková B.,
Němec F.,
Parrot M.,
Kruparova O.,
Krupar V.
Publication year - 2021
Publication title -
journal of geophysical research: space physics
Language(s) - English
Resource type - Journals
eISSN - 2169-9402
pISSN - 2169-9380
DOI - 10.1029/2021ja029158
Subject(s) - principal component analysis , interplanetary spaceflight , intensity (physics) , physics , spacecraft , shock (circulatory) , computational physics , earth's magnetic field , shock wave , altitude (triangle) , geodesy , solar wind , geophysics , geology , optics , mechanics , magnetic field , astronomy , statistics , mathematics , geometry , medicine , quantum mechanics
Very low frequency wave intensity measurements provided by the French low‐altitude DEMETER spacecraft are studied using the principal component analysis (PCA). We focus on both the physical interpretation of the first two principal components and their application to real physical problems. Variations of the first principal component (PC1) coefficients due to the geomagnetic activity and seasonal/longitudinal changes are studied. It is shown that their distribution corresponds to the wave intensity dependences obtained in previous studies. Moreover, the variations of PC1 coefficients around interplanetary shock arrivals are analyzed. The study is performed for fast forward (FF), fast reverse, slow forward, and slow reverse shocks separately. It shows that the most significant effect on the wave intensity is displayed in the FF case. Furthermore, it turns out that the wave intensity variations depend on the wave intensity detected before the shock arrival. Finally, the shock strength and interplanetary magnetic field orientation are also important. The performed analysis shows that PCA can be successfully applied to characterize large data sets of spacecraft measurements by limited sets of numbers—principal component coefficients (typically first one or two are enough), which still maintain a sufficient amount of information.