Routine fault plane solutions for local networks: A test with synthetic data

Abstract For local networks aimed at analysis of microseismicity, the classic method for focal mechanism determination by use of first motion directions can be of limited value because often very few clear first motions are observed. A number of alternative methods making use of amplitudes have been proposed. The radiated far field, low-frequency spectral amplitudes of body waves depend on the seismic moment and orientation of the earthquake source. This can be used to obtain fault plane solutions from amplitude observations. The observed spectral amplitudes are corrected for the free surface effect, attenuation, instrument response and geometrical spreading. Searching over the entire parameter space for strike, dip and rake, the source orientations that satisfy the polarity observations and give “sufficiently” small least squares misfits between observed and predicted amplitudes are accepted as possible focal mechanisms for the earthquake. The reflectivity method was used to generate synthetic seismograms for earthquake sources within a local network in Iceland. The amplitude inversion algorithm was applied to the synthetic seismograms. Inverting noise free amplitudes at eight stations, the solutions obtained deviate less than 10° from the correct mechanism in all source angles for all mechanisms tested. Adding noise and scaling the synthetic data to simulate events of varying magnitude indicates that the method gives fault plane solutions for local events down to magnitude M L = 0.5 correct to within ± 15° in all three source angles.