Removing Milky Way from airglow images using principal component analysis

• The Milky Way causes spurious waves and overwhelms gravity waves on airglow images.
• The Milky Way can be effectively extracted by principal component analysis (PCA).
• The Milky Way is removed by subtraction of PCA reconstructed background.
• The PCA method does not affect wave parameters with normal amplitudes and duration.

Airglow imaging is an effective way to obtain atmospheric gravity wave information in the airglow layers in the upper mesosphere and the lower thermosphere. Airglow images are often contaminated by the Milky Way emission. To extract gravity wave parameters correctly, the Milky Way must be removed. The paper demonstrates that principal component analysis (PCA) can effectively represent the dominant variation patterns of the intensity of airglow images that are associated with the slow moving Milky Way features. Subtracting this PCA reconstructed field reveals gravity waves that are otherwise overwhelmed by the strong spurious waves associated with the Milky Way. Numerical experiments show that nonstationary gravity waves with typical wave amplitudes and persistences are not affected by the PCA removal because the variances contributed by each wave event are much smaller than the ones in the principal components.