Evaluating the Effects of Magnetic Susceptibility in UXO Discrimination Problems

: Using numerical simulations based on magnetic susceptibility properties observed at Kaho'olawe, Hawaii, we have examined the effect of magnetic soil on static magnetic method and time-domain electromagnetic (TEM) method in UXO discrimination problems. We have demonstrated that the static magnetic susceptibility can be effectively modeled using a correlated random process, and that Wiener optimal filter can be used as a preprocessing tool to remove the effect of soil response and improve the reliability of dipole inversions. The frequency-dependent susceptibility in TEM method can be modeled using a complex susceptibility having a broad range of relaxation times. A layer of soil with such susceptibility produces the characteristic t(-1) decay of the voltage measured in TEM at Kaho'olawe. The horizontal component of central loop TEM data is not sensitive to the presence of magnetic soil if it is sufficiently 1D. This provides a preprocessing tool for removing the soil effect from the vertical component data and improving the result of two-dipole inversion. The research project has therefore accomplished the four stated goals of (1) developing and verifying software for simulating soil responses, (2) characterizing the effect of soil susceptibility, (3) determining the applicability of two inversion algorithms, and (4) developing methods for removing the soil effect.