Complex Seismic Source Inversion Method with the Data Covariance Matrix: Application to the 2010 Haiti Earthquake
Abstract
In the finite fault source inversion, seismic source area has usually been approximated by simple fault plane model for simplicity. This approximation, however, may generate the correlated modeling errors originated from the focal mechanism variation in a rupture process, which contributed to biased results in the seismic waveform analysis. This effect becomes predominant for analysis of seismic data around the nodal planes. From CMT inversion analysis, the January 12, 2010 Haiti earthquake may accompany both strike and dip slip on different fault planes (Nettles and Hjörleifsdóttir, 2010, GJI). In addition, one single fault plane model cannot explain teleseismic body wave well due to complex source process and existence of many mechanism-sensitive stations. For waveform analysis of this earthquake, we developed inversion method that estimates moment tensor component for each space knot in seismic source area and applied it to teleseismic P-wave data recorded at FDSN network stations and Global Seismograph Network stations. In general, such high flexibility source model had caused the unstable and unrealistic result. To avoid this problem, we applied new formulation that considers the data covariance components of observed errors and modeling errors originated from uncertainty of Green's function (Yagi and Fukahata, 2010, AGU). It has already been shown that the new formulation can derive plausible solution without non-negative constraint. For inversion, we arranged space knots on the plane of which strike and dip are same as that of the USGS finite fault model. We confirmed that result is robust against change of strike, dip and knot interval. The result shows that P-axes in main rupture area are north-south direction, which is consistent with stress field of the region. Main rupture area can be divided into 3 patches, near the hypocenter, east and west side of the hypocenter patch, which have different focal mechanisms. Reverse fault is dominant in the hypocenter patch whereas strike slip component is dominant in east and west patches. The CMT inversion analysis shows that many reverse aftershocks are observed west side of the west patch. It seems that rupture stopped at the focal mechanism transition area. Our result is consistent with analysis of InSAR data and the CMT analysis.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2010
- Bibcode:
- 2010AGUFM.S43A2024K
- Keywords:
-
- 7203 SEISMOLOGY / Body waves;
- 7215 SEISMOLOGY / Earthquake source observations