Noise In the Processing And Application of Magnetic Gradients

The increased use of magnetic gradients brings about the need for reliable noise characterization in order to understand the assumptions made during data processing. We present a statistical approach to characterize noise in magnetic gradient data. We use an equivalent source technique and finite difference calculations coupled with statistical analysis to observe the distribution of noise within gradient measurements. Our methods are applied to real airborne magnetic data where by we compare measured and calculated gradients. The results show that noise in gradient measurements can be characterized by a skew-t distribution rather than the readily assumed exponential or normal distributions. A secondary consequence from the presented methodology indicates that inherent errors associated with gridding may be detrimental to the use of calculated gradients.