
Spatial variations of the effective elastic thickness, T e , using multitaper spectral estimation and wavelet methods: Examples from synthetic data and application to South America
Author(s) -
PérezGussinyé M.,
Swain C. J.,
Kirby J. F.,
Lowry A. R.
Publication year - 2009
Publication title -
geochemistry, geophysics, geosystems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.928
H-Index - 136
ISSN - 1525-2027
DOI - 10.1029/2008gc002229
Subject(s) - multitaper , geology , wavelet , deconvolution , lithosphere , geodesy , seismology , wavelet transform , algorithm , mathematics , computer science , artificial intelligence , tectonics
Improved recovery of the spatial variability in the effective elastic thickness, T e , of the lithosphere is important to interpret Earth's structure and geodynamics. Here we use synthetic topography and gravity data to systematically compare the recovery of T e using the multitaper windowing scheme of Pérez‐Gussinyé et al. (2004, 2007) with the wavelet approach of Kirby and Swain (2008). We find that spuriously high T e estimates previously present in synthetic tests with the multitaper are artifacts arising from Fourier transform edge effects introduced by splitting the data into land and ocean areas prior to load deconvolution. Hence, we adopt the mixed land‐ocean loading scheme of Kirby and Swain (2008). Using this approach we find that underestimation of T e when the transitional wavelength approaches the window size is largely related to the limited size of the analysis window and not to random correlations of initial loading surfaces as occurs in the wavelet method. To attenuate this bias we apply a correction factor to our resulting multitapered T e and find that this technique more accurately recovers steep T e gradients and small‐scale T e anomalies than the wavelet method. Finally, we recalculate T e of South America using the multitaper and the mixed land‐ocean approach. The new T e distribution is very similar to that of Pérez‐Gussinyé et al. (2007, 2008), suggesting that Fourier transform edge effects are less severe in the real Earth than in synthetic data. This difference arises because coastlines are much less rough in the real Earth than in our synthetic data. On the basis of the synthetic results, we interpret that most of the small‐scale features present in the multitaper but not in wavelet T e estimates of South America presented by Tassara et al. (2007) are real. These include high T e over the Andean flat subduction zones and low T e along the dyke swarms of the Paraná flood basalts, the Tacutu graben, and Amazonian basin.