Modeling Low-frequency Magnetic-field Precursors to the Loma Prieta Earthquake with a Precursory Increase in Fault-zone Conductivity

Abstract

—The 1989 M _s = 7.1 Loma Prieta earthquake was preceded for 12 days by what have been claimed as precursory ultra-low-frequency (ULF) magnetic noise anomalies ten times background, and by a very high peak up to 100 times background just 3 hours before the earthquake. We propose that these anomalous fields could have been due to the formation of a long thin highly-conductive region along the earthquake fault, which magnified the external electromagnetic waves incident on the earth’s surface. We use a simplified quantitative model, assuming a highly-conductive elliptic cylinder embedded in a layered resistivity structure, which we base on independent magnetotelluric measurements. The magnetic-field anomaly observed 3 hours before the main shock can be modeled by assuming an elliptic conductor extending from the surface to the hypocenter with a conductivity of 5 S · m⁻¹. Our computed anomaly matches the observed anomaly to within a deviation of 35% over an observed frequency range of over 2 orders of magnitude, over which the measured anomaly varies from only about twice background (at 5 Hz) to about 100 times background (at 0.01 Hz). In addition, other anomalies recorded up to 12 days before the earthquake, can be modeled in detail by varying only the size of the elliptic conductor.¶We show that such an increase in conductivity could be caused by a precursory reorganization of the geometry of fluid-filled porosity in the fault-zone, which we call a dilatant-conductive effect. The extreme observed magnetic anomalies can be modeled using the high fault-zone porosity (c. 10%) and fluid conductivity (equivalent to 2 M NaCl) implied by other workers’ magneto-telluric measurements, but without requiring the large-scale precursory fluid flow characteristic of other published models for the magnetic-field precursors.