For the inversion we used trapezoid source time functions
with different duration and maximum rupture velocity of
2.8 km/s. We found a trapezoid with a rise time of 4 s as
shown in Fig. 7a to anticipate the best fitting synthetic
waveforms. To estimate the best fit focal depth we calculated
the sum of misfit errors keeping the source at different
hypocentral depths. Fig. 7b illustrates the best fit depth at
46 km with the minimum sum of misfit error of 0.1386.
Our results show that the first subevent is followed by
the second subevent after 8 s (Fig. 7c). The signature of
the second subevent in broadband seismograms at a lag
of 18 s is due to the 10 s moment rate functions used in
the inversion. Seismic moment of the two subevents is
8.37E + 18Nm and 3.97E + 18Nm, respectively. Total seismic
moment of this earthquake is calculated to be
1.09E + 19Nm consistent with HARVARD CMT Solution
and is accountable for a moment magnitude of 6.63.
During the inversion each subevent was set free to have
a different mechanism noticing that change in focal mechanism
of different subevent generating a deep focused
earthquake is not uncommon (e.g., Sterlitz, 1980; Kikuchi
and Kanamori, 1986). We found both of the nodal planes