Estimating the Hypocenter and Mechanism of the August 15, 1991 Centre Hall, Pennsylvania Earthquake Using Single-Station Data

Abstract

Abstract On August 15, 1991 a small (mbLg = 3.0) earthquake occurred near the town of Centre Hall, Pennsylvania. Based on early reports of felt effects and earthquake-generated sounds, the epicenter was placed somewhere ENE of State College, Pennsylvania. Three-component short-period digital data from the DWWSSN station SCP were analyzed to determine the hypocenter. Often, for small earthquakes in regions without dense seismic networks, information about an event must be obtained from single-station data. In this case, since no shallow velocity model exists for the area, simple ideas of wave propagation are invoked to estimate the distance and back-azimuth to the event. The horizontal P-wave particle motion constrained the back-azimuth, after calibrating the horizontal components by measuring the back-azimuth of quarry blast P-waves of known location. Distance determination was hampered by lack of a detailed upper crustal velocity model. Using iterative forward waveform modeling, a velocity model was generated that fit the observed S-minus-P and Rg-minus-P times and Rg-wave dispersion, and which was consistent with known upper crustal velocities in the area. A source depth of less than 1 km was inferred from the Rg-to-S ratio, the depth phase sP, and reports of earthquake-generated sounds. Estimates of the focal mechanism were obtained by a grid search procedure using Green’s functions computed with wavenumber integration for shear dislocation sources. Theoretical and observed amplitudes of sP, direct SH and SV (taken as ratios to the direct P), along with P polarity were compared for all possible combinations of strike, dip, and rake. Though fault plane orientation is poorly constrained, E-W to WNW-ESE P-axis orientation is a robust result of the search. Normal-faulting mechanisms are inconsistent with the data. However, the theoretical SV-to-P ratio is up to a factor of two larger than the observed ratio. This is probably related to an inadequate structure model and waveform sensitivity to source depth. Mechanism P-axis trends are consistent with other regional stress field indicators in the area.