Tikhonov's regularization for deconvolution in the empirical Green function method and vertical directivity effect

Abstract

Application of Tikhonov's technique, using input errors for the parameter of regularization estimation, enhances the accuracy and stability of the reconstruction of a source time function (STF) by the empirical Green function (EGF) method that gives us an opportunity to use simultaneously for analysis body and surface waves data, and to estimate the horizontal and vertical directivity effects. Knowledge of the last is particularly useful for the choice of an active nodal plane of earthquakes with the dip slip fault orientation that allows us to classify these earthquakes to the interplate or intraplate types and thereby to reach the better understanding of tectonic processes in the region of interest. By way of illustration, an attempt to estimate average parameters of faulting in a first approximation is made herein for two Russian Far East large events with opposite types of focal mechanism orientation, strike slip and dip slip. The former is not a matter of interest in the context of vertical directivity effect but enables us to test the method. The directivity analysis of pulse durations and inverse amplitudes of the relative source time functions (RSTFs) restored at eight globally distributed stations IRIS indicates that the destruction in the source of the Neftegorsk earthquake (05/27/1995 M W=7.1) propagated roughly horizontally in the direction 8±11° during 19.2±0.4 s along the rupture extending 35.5±4.9 km. The calculated slip distribution along the rupture coincides within the error with the results of field geological measurements on the causal surface fault that proves that the Neftegorsk earthquake source is well described by the model of the linear unilateral fault and gives a good assessment of the method applied. The average parameters of faulting in the Kamchatka earthquake (03/08/1999 M W=6.9) have been determined from data of 13 station IRIS. It was shown that the destruction in its source propagated downward at an angle of about 60° with horizon, in the direction about S156° E, during 13.4±0.2 s, along the rupture totaling 25.5±2.3 km in length. Therefore, the nodal plane, steeply dipped to the SE, was active and this event can be regarded as an intraplate type. Two asperities can be selected; the first with the maximum slip 3.3 m located at a distance of about 7 km from the onset of rupture, and the second with the maximum slip about 0.9 m centered at approximately 19 km from that.