Power spectrum estimation for magnetotelluric signal based on continuous wavelet transform

This research aims to improve the accuracy of power spectrum estimation of magnetotelluric signal by using continuous wavelet transform. Three statist ic method, global average, wavelet coefficients shrinkaging and significanc e test, are used to estimate the power spectrum. Empirical formulas are giv en for the transformation between wavelet scale and Fourier frequency. The algorithm is also described in detail. The numerical tests show that the es timation is robust to stronger white noise and local correlated-noise, and significantly decreases the requirement for the recoding length.