Narbut, S.M., 1983, Determination of fault-related stress changes using the piezomagnetic effect: University Park, Pennsylvania, Pennsylvania State University, Ph.D. dissertation, 168 p., illust., maps.
Current methods of measuring crustal stresses are not adequate for determining tectonic stress changes associated with faulting. The piezomagnetic effect can be used to measure these dynamic stress variations. To utilize the piezomagnetic technique, it is necessary to show that laboratory measurements of stress-induced changes in remanent magnetization and magnetic susceptibility are consistent with observed magnetic field changes caused by stress release. Magnetic measurements on Amchitka Island in the Aleutian Arc, taken before and after the 1972 CANNIKAN explosion, present an opportunity to test the applicability of the piezomagnetic method. Data recorded on a rectangular grid show changes of 13 to 20 gammas over the major fault covered by the grid. Laboratory measurements of various confining pressures have been made on rock cores from Amchitka to determine the relationship between changes in magnetic susceptibility and remanent magnetism and differential stress. These relationships have been combined with the stress field calculated for a dip-slip, vertical fault model to predict subsurface changes in magnetic properties. A numerical integration over subsurface, stress-induced magnetic changes gives a predicted variation in the surface magnetic field for comparison with the actual field observations. The best-fit model is composed of two solutions, one for a long wavelength component of the observed data and another for a short wavelength component. The match to the long wavelength component is piezomagnetic and results in an average stress change of 3 bars. This low value probably reflects three different situations; (1) the presence of low, initial, ambient stresses; (2) recent tectonic stress release due to two previous nuclear explosions; (3) an assumed homogeneous fault plane where the stress change is averaged over the entire fault plane. The solution to the short wavelength component is not piezomagnetic. This component appears to be related to extensional movement normal to the fault plane. This study shows the feasibility of utilizing the piezomagnetic effect to determine changes in in situ stress given adequate field measurements, complete laboratory calibrations and a realistic fault model.
Theses and Dissertations
