Time‐correlated patterns from spherical harmonic expansions: Application to geomagnetism. Issue 12 (15th December 2015)

• Record Type: Journal Article
• Authors: Pais, M. A.; Alberto, P.; Pinheiro, F. J. G.
• Is Part Of: Journal of geophysical research. Volume 120:Issue 12(2015:Dec.)
• Page Start: 8012
• Page End: 8030
• Publication Date: 2015-12-15
• Subjects: main magnetic field -- Gauss coefficients -- principal component analysis -- natural orthogonal components -- jerks -- empirical orthogonal functions; Geomagnetism -- Periodicals; Geochemistry -- Periodicals; Geophysics -- Periodicals; Earth sciences -- Periodicals; 551.1
• DOI: 10.1002/2015JB012399
• Languages: English
• ISSNs: 2169-9313

Abstract

Abstract: We use empirical orthogonal function analysis (EOFA) directly on sets of Schmidt spherical harmonic (SH) coefficients modeling the internal geomagnetic field or its time derivatives at different epochs. We show how to properly use the method such that the application of EOFA to either spatial or spectral domains leads to the same results, bypassing the need to work on snapshots of field charts synthesized from SHs. In case a spatial grid is required, we point out which is the best grid to use. We apply the method to the CM4 geomagnetic field model to illustrate the differences in EOFA modes obtained with and without corrections. Once the corrected main modes of secular acceleration (SA) have been singled out, we retrieve previous results showing that the 1969, 1978, and 1991 geomagnetic field acceleration jumps have the same spatial pattern. A new finding in this study is that the same spatial pattern is present in principal modes of secular variation which, once inverted, may provide the flow responsible for the jerk sequence. Another finding is the unveiling of a different spatial structure common to a second group of jerks with SA pulses around 1985 and 1996, displaying a localization very similar to SA pulses identified in 2006 and 2009 using recent satellite data. Finally, if properly handled, the EOFA can be directly applied to a grid of data values of the geomagnetic field in order to produce SH models of decorrelated modes which may help to separate different sources of the field. Key Points: Empirical orthogonal function analysis in both spatial and spectral domains Optimal space grid for empirical orthogonal function analysis Analysis of two uncorrelated jerk patterns in the CM4 model