BTTB–RRCG method for downward continuation of potential field data

Abstract This paper presents a conjugate gradient (CG) method for accurate and robust downward continuation of potential field data. Utilizing the Block-Toeplitz Toeplitz-Block (BTTB) structure, the storage requirement and the computational complexity can be significantly reduced. Unlike the wavenumber domain regularization methods based on fast Fourier transform, the BTTB-based conjugate gradient method induces little artifacts near the boundary. The application of a re-weighted regularization in a space domain significantly improves the stability of the CG scheme for noisy data. The synthetic data with different levels of added noise and real field data are used to validate the effectiveness of the proposed scheme, and the computed results are compared with those based on recently proposed wavenumber domain methods and the Taylor series method. The simulation results verify that the proposed scheme is superior to the existing methods considered in this study in terms of accuracy and robustness. The proposed scheme is a powerful computational tool capable of applications for large scale data with modest computational cost.