Long Term Variations in Magnetotelluric Apparent Resistivities Observed near the San Andreas Fault in Southern California

Since early 1975, we have monitored magnetotelluric apparent resistivities at three sites close to the Los Angeles metropolitan area. We estimate the repeatability of our calculated apparent resistivities to be 10-20%. A major problem has been the removal of bias in the resistivity estimates caused by cultural noise, mainly on the magnetic field components. We have attempted to remove this bias using a variety of techniques including singular value decomposition of the signal matrix. None of these procedures was entirely successful when the signal predictabilities fell below 80%. The only major earthquake to occur in our study area was the Santa Barbara shock on August 13, 1978 (ML=5.2). No significant changes were observed in apparent resistivities measured at a site situated 80km to the east of the epicenter both 25 days before and 7 days after this shock. Significant changes in low frequency apparent resistivities were observed at a site in West Antelope Valley, 50km northwest of Palmdale. Between April 1976 and July 1978 the rotated minimum apparent resistivities in the frequency range 0.001-0.1Hz increases 10-15%, while the maximum resistivities remained almost constant. The direction of the maximum rotated apparent resistivity also changed by as much as 20. These changes cannot be correlated with any obvious change in seismicity. However, they may be related to the uplift and subsidence which is known to have occurred in this area. We have also observed changes in the high frequency (>1Hz) apparent resistivities following a heavy rainfall which occurred in the spring of 1978. At one site, the apparent resistivities at 10Hz dropped two orders of magnitude. This effect was modelled by adding a surface layer with a thickness of 10m and a resistivity of 10m.