Absence of Evidence for a Shallow Magma Chamber Beneath Long Valley Caldera, California, in Downhole and Surface Seismograms

Abstract

A downhole seismometer at 900-m depth and a temporary network of surface stations were deployed to use rays from local microearthquakes to study the upper and middle crust beneath the Long Valley caldera. The downhole seismograms show S waves with high apparent amplitudes from earthquakes located 2–20 km to the south of the downhole seismometer. In contrast, S waves from earthquakes located in the distance range 20–30 km to the south have low apparent amplitudes. If P and S amplitudes are normalized relative to the respective coda amplitudes, the S to coda amplitude ratios appear to remain constant but the P to coda amplitude ratios vary significantly with takeoff angle. A comparison of the calculated radiation patterns for a double couple in a uniform halfspace and focal mechanisms of the recorded earthquakes suggest that the observed variations in P and S amplitudes are caused by radiation pattern effects. Reanalysis of possible travel time delays found by Elbring and Rundle (1986), who used a subset of the borehole data analyzed in this study, shows that they underestimated the epicentral distances to three of the earthquakes and hence generated an artificial kink in the reduced travel time versus depth curve. Synthetic calculations of reduced travel time versus depth suggest that an apparent velocity of 5.7 km/s gives less scatter than 6.0 km/s used by Elbring and Rundle (1986). Plots of ts/tp versus depth show that contrary to the findings of Elbring and Rundle (1986), the Vp/Vs ratio stays fairly constant with depth and a small (<3-km diameter) magma chamber cannot easily be resolved. Furthermore, combined analysis of downhole and surface data shows that neither data set requires a low-velocity zone or a zone of anomalously high Vp/Vs at depth below the Casa Diablo area.