A 3D homogenous locator for magnetic anomalous sources: algorithm development and testing

Abstract:

When a function of spatial coordinates is a homogenous function of order n (n is an integer), each of its derivatives with respect to x, y, or z is also a homogenous function but of order (n-1). Based on this proposition, a point-based algorithm, called a homogenous locator (HL), has been developed to invert simultaneously the position (horizontal position and depth) and structural index (SI) of the magnetic sources that generate the total magnetic intensity (TMI) and its gradient and full tensor magnetic gradiometry (FTMG). The algorithm is then tested against point sources (pole and dipoles) and compared with standard Euler deconvolution (SED) using synthetic and field data sets. Cases with sole induction and remnant dipole moment are also investigated and compared. Results show that: (1) Both techniques can predict very accurately the horizontal positions and with less accuracy in depth prediction, but the homogenous locator has higher accuracy; (2) the homogenous locator can produce much few solutions on average for locating each magnetic object; (3) Since homogenous locator inverts SI simultaneously, clusters of SI around different centers can classify the source types accurately. This is a main advantage of the homogenous locator. While for each value of SI as an input to SED, magnetic sources of different types (pole and dipole) are all located, resulting in almost 100% of both commission and omission errors in classification of the types of sources; and (4) the homogenous locator is more sensitive to remnant than SED.