Perspectives on monitoring resistivity changes with telluric signals at Parkfield, California: 1988–1999

An array designed to detect changes of resistivity in the upper 10 km of the crust has been in continuous operation since 1988. Long period (T=300–7200 s) telluric signals are recorded digitally on 5–16 km long dipoles. Telluric transfer functions are computed between field dipoles and two reference dipoles and then decomposed to examine daily fluctuations of those functions. Based on analysis of these daily fluctuations, stabilities of 0.1–0.3% are achieved. No intermediate term resistivity changes are seen associated with any of the 5 M_b>4.0 earthquakes that have occurred since 1989. Short term fluctuations are seen associated with most of these earthquakes, however. Minimum bounds on resistivity changes that could have caused the fluctuations in the telluric responses range from 1.7 to 13.0%. While these are within the ranges seen in other field studies, the statistical significance of these fluctuations is low because they also occur at many times that cannot be correlated to earthquakes. Comparison of the one potential resistivity change associated with a strain change results in a strain sensitivity that is unrealistically high. This possible resistivity change coincides spatially with a temporary variation in seismic travel time in May–June 1989, however. It is likely that many of the fluctuations in the telluric responses are random variations except for the one in 1989, and recording through the next M6 earthquake will be needed in order to detect a statistically significant resistivity change.