Potential field separation into its deep component (regional) and its shallow component (residual) was carried out many years ago using the least-squares curve fitting. The curve fitting process fits equations of approximating curves to the raw field data. Nevertheless, for a given set of data, the fitting curves of a given type are generally not unique. Thus, an optimum curve with a minimal deviation from all data points is desired. This best-fitting curve can be obtained by the method of least squares.
This work addresses the technical improvement of the results of least squares method technique for separating local and regional component of the potential field data. The data points that are related only to the smoother parts of the observations, as indicated by a second horizontal gradient filter then are used to fit by low order polynomial. The residual data is simply obtained by subtraction. The present method requires the observations to be taken over a very large area such that the observations on the edges of the 2D profile can be assumed to be due to the regional only and no causative bodies exist below them. The proposed method is tested using a synthetic example and satisfactory results are obtained. The advantage of this method is that it causes no distortion to the shape of the original field. Most of this work is concerned with gravity data, but can be extended to magnetic data (RTP).