Improved estimation of Curie-point depth using IRLS-centroid method for fractal distribution of sources

Curie-point depth is closely related to the magnetic field and is important to study the thermal evolution and geological structure of the lithosphere. In recent studies, the application of the centroid method has been extensive for estimating the Curie-point depth, which offers valuable information regarding the thermal structure of the Earth’s crust. By analyzing the radially averaged power spectrum of magnetic data, the centroid method determines the top and centroid depths of the magnetic layer with fractal magnetization. Specifically, a least-squares fitting is performed to determine the top and bottom depth from the few points in the one-dimensional radially averaged power spectrum. However, the ordinary least squares (OLS) method is sensitive to outliers, which will lead to distorted estimated parameters. This paper proposes an iteratively reweighted least squares (IRLS) fitting based centroid method to eliminate or minimize the impact of possible outliers to improve estimation accuracy of the Curie-point depth. The tests and analysis demonstrate that the proposed method has significant performance gains over conventional centroid method.