VELOCITY-INDEPENDENT EPICENTER LOCATION

An epicenter location procedure based on relative phase arrival times for local and regional earthquakes is demonstrated. Using the difference in arrival time of the same phase at different stations, the most consistent epicenter is determined without assuming any seismic velocity model. This is accomplished by determining the location giving relative epicentral distances to at least four stations that best matches the relative arrival times at those stations. For four stations A, B, C, and D with arrival times t_a<t_b<t_c<t_d, an epicenter is determined such that the epicentral distances r_a, r_b, r_c and r_d best fit the identity (r_b- r_a)/( r_c- r_a) = (t_b- t_a)/( t_c- t_a) and similarly for station combinations BCD, ABD and ACD. Phase velocities and event times can then be determined a posteriori. As with velocity-based location techniques, the relative arrival time method assumes azimuthal isotropy in seismic velocities and produces smaller location uncertainties with increasing number of stations. Higher uncertainties can result when an epicenter is equidistant to two or more stations and/or several stations are collinear.

The relative time method used with just four stations can produce locations and origin times matching those published from velocity-model-derived techniques using far more stations. For example, this method located the 3/12/07 Summit Co., Ohio earthquake only 2km from the published epicenter obtained using >10 stations, and with an event time that differed by 0.74 seconds. The inferred P-wave velocity was 6.35 km/s. Similarly, a series of three earthquakes in Lincoln Co., Oklahoma between 11/18/11 and 11/24/11 were located at an average distance of 6.1 km from published epicenters and with an average event-time difference of only 0.36 seconds. These cases demonstrate that this method would be effective in locating earthquakes and origin times in regions where the seismic velocity structure is poorly or even entirely unknown.