Recovering relative locations of the 2005-2006 Mukacheve earthquakes from similarity of their waveforms at a single station

An algorithm was designed for inverting the set of cross-correlation coefficients between the records of recurrent earthquakes by a single station for their relative locations based on an assumption of Menke (1999) that the cross-correlation is only a function of separation between the sources and decays exponentially with it. Tested on synthetic data sets, the algorithm easily recovered the true relative locations, the initial ones seeded randomly. Inversion of the cross-correlation coefficients between the records of the two groups of the Mukacheve 2005–2006 recurrent earthquakes by the station Mukacheve steadily produced an almost flat shape of hypocenter cloud for each of the groups, final relative locations of hypocenters being almost the same for different initial seedlings. The combined set of cross-correlation coefficients for both groups was also inverted, and the shape of the combined hypocenter cloud was almost flat again. These results compare well with ones obtained earlier (Gnyp 2010) by relocating the Mukacheve earthquakes with the use of differential arrivals and source specific station terms.